Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Follicle size indicates oocyte maturity and blastocyst formation but not blastocyst euploidy following controlled ovarian hyperstimulation of oocyte donors

Follicle size indicates oocyte maturity and blastocyst formation but not blastocyst euploidy... Abstract STUDY QUESTION Is there is an association between follicle size and the quality of oocytes retrieved from them as judged by ability to achieve the blastocyst stage, blastocyst grades and blastocyst ploidy? SUMMARY ANSWER Although follicle size is a valuable predictor of oocyte maturity and is a significant predictor of the ability of a fertilized oocyte to become a quality blastocyst, the ploidy of each quality blastocyst is not related to the size of the follicle from which its oocyte was retrieved. WHAT IS KNOWN ALREADY It is unclear whether the oocytes within larger follicles are the best oocytes of the cohort. Although there have been studies examining follicle size in relation to embryo quality, there has been no study relating the incidence of euploidy in embryos to follicle size. STUDY DESIGN, SIZE, DURATION The purpose of this study was to examine follicle sizes and the oocytes from those follicles (and the embryos that result from those oocytes) to see if there is an association between follicle size and the quality of oocytes as judged by ability to achieve the blastocyst stage, blastocyst grades and blastocyst ploidy. Follicle sizes for oocytes were assessed both as diameters (mm) and as Z values (expressed as their size relative to the mean and standard deviation of that donor’s follicular cohort). Comparisons were made using cumulative histograms, rolling averages and receiver operator characteristic (ROC) curves and its AUC. PARTICIPANTS/MATERIALS, SETTING, METHODS Twenty-two oocyte donors (ages: 24.5 ± 3.5 years) whose recipients would use ICSI for insemination were enrolled in this study. Follicles were aspirated one-at-a-time to be certain that the aspirated oocyte was from the same follicle measured. The follicle measurement (size) was noted in the embryology records. Oocytes were cultured individually throughout their time in the embryology laboratory so that follicle sizes could be uniquely associated with each oocyte. Oocytes and embryos were analyzed according to the size of the follicle from which the oocyte was retrieved. MAIN RESULTS AND THE ROLE OF CHANCE Three hundred seventeen oocytes (96.1%) had an associated follicle size. Of the oocytes with follicle sizes, 255 (80.4%) had a polar body (MII), and 60 (18.9%) were immature: 31 (9.8%) with a visible germinal vesicle (GV stage) and 29 (9.1%) with neither a polar body nor a visible germinal vesicle (MI). The incidence of MII oocytes was significantly associated with larger follicle size using either mm (ROC’s AUC = 0.87; P < 0.0001) or Z values (ROC’s AUC = 0.86; P < 0.0001). Among MII oocytes there was no association with follicle size for the appearance of 228 oocytes with two pronuclei (2 PN). Among 2 PN’s, the development of 94 quality blastocysts that underwent trophectoderm biopsy (TE Bx) exhibited a significant association with larger follicles using either mm (ROC’s AUC = 0.59; P = 0.01) or Z values (ROC’s AUC = 0.57; P = 0.01). The use of follicle diameter as a feature to distinguish between fertilized oocytes that would ultimately become blastocysts versus those that would not become blastocysts resulted in an enrichment for blastocyst formation from 20 to 40%. Of the 94 quality blastocysts, 51 were determined by next generation sequencing (NGS) to be euploid.Although oocyte maturity and the incidence of blastocyst formation were associated with follicle size, the incidence of euploidy among biopsied blastocysts was not. Follicles measured by two different methods (mm or Z values) led to predominantly the same conclusions. LIMITATIONS, REASONS FOR CAUTION This study investigated the relationship between follicle size and measures of oocyte/embryo quality when donors were treated similarly. Therefore, this study does not investigate the effects of triggering and retrieving oocytes when the follicle cohorts are of different sizes or lead follicles are of different sizes. Although no association was found between follicle size and euploid blastocysts, the fact that blastocyst ploidy is not entirely dependent upon oocyte ploidy (e.g. aneuploidies derived from mitotic errors or from the fertilizing sperm) makes it difficult to infer the relationship between follicle diameter and oocyte ploidy. WIDER IMPLICATIONS OF THE FINDINGS It is confirmed that follicle diameter is predictive of oocyte maturity. However, once oocyte maturity is known, the diameter of the follicle from which the oocyte was retrieved is not instructive. Embryos generated through fertilization and development of the mature oocytes from any observed follicle diameter were equally likely to become euploid blastocysts. STUDY FUNDING/COMPETING INTEREST(S) This study was funded by ReproART: Georgian American Center for Reproductive Medicine. None of the authors declare any actual conflicts of interest. D.H.M. received compensation from ReproART, Biogenetics Corporation and the Sperm and Embryo Bank of New York and honoraria and travel funding from Ferring Pharmaceuticals and from Granata Bio. S.M. received compensation from Cooper Genomics and an honorarium and travel funding from Ferring Pharmaceuticals. L.C. is the founder of LTD Ovamedi, the organization that represents Cooper Genomics in Georgia, and received travel funding from the European Society for Human Reproduction and Embryology. TRIAL REGISTRATION NUMBER N/A. Preimplantation genetic testing for aneuploidy (PGT-A), Follicle diameter, Oocyte maturity, Quality blastocyst formation, Blastocyst ploidy Introduction Although the largest follicle during controlled ovarian hyperstimulation is generally thought to be the follicle that would have been the dominant follicle in an unstimulated menstrual cycle, it is unclear whether the oocyte within that follicle is the best oocyte of the cohort. There have been publications indicating that oocytes from larger follicles exhibit a higher incidence of mature and/or fertilized oocytes (Quigley et al., 1982; Wittmaack et al., 1994; Rosen et al., 2008; Kahraman et al., 2017; Wirleitner et al., 2018), a higher incidence of normal cleavage (Kahraman et al., 2017) and a higher incidence of blastocyst formation (Kahraman et al., 2017; Wirleitner et al., 2018). Others suggest that oocytes from larger follicles are suboptimal (Nivet et al., 2016) with medium size follicles more likely to yield oocytes that develop into better quality embryos. Associations between the size of the largest follicle in the cohort and quality of oocytes have had mixed results with larger lead follicles having better cleavage and implantation and less fragmentation (Miller et al., 1996) or no significant difference in fertilization or cleavage (Rosen et al., 2008). It remains unclear whether follicle size is related to the ploidy status of the oocyte within or may predict its subsequent developmental competence and embryonic ploidy. Two prior studies have found that euploidy rates for donor blastocysts are significantly different for different centers (Munné et al., 2017) and for different physicians within the same center (McCulloh et al., 2019). One possible difference between different centers and between different physicians within the same center that we wished to examine in this study is whether the diameter of follicles selected for retrieval may affect the ploidy of blastocysts that arise from oocytes retrieved from those follicles. The purpose of this study was to examine follicle sizes and the oocytes from those follicles (and the embryos that result from those oocytes) to see if there is an association between follicle size and the quality of oocytes as judged by ability to achieve the blastocyst stage, blastocyst grades and blastocyst ploidy. Although more than one euploid embryo can result from controlled ovarian hyperstimulation (COH), the superiority of larger follicles has remained a thought pattern among practitioners of assisted reproductive technologies. Whether larger follicles are more likely to contain oocytes that will not undergo flawed meiosis with flawed chromatid and chromosome segregations is not clear. Further, it is unclear whether the oocyte from a dominant follicle, once fertilized, will become an embryo that is more competent in its ability to undergo mitotic segregation of chromosomes. To date, there has been no study relating follicle size to the incidence of aneuploidy in embryos. The aneuploidies detected in blastocysts may be attributed to meiotic or mitotic errors (Rabinowitz et al., 2012; Fragouli et al., 2013) or to aneuploidy in the fertilizing sperm (Levron et al., 2000; Tempest, 2011; McWilliams et al., 2015). However, the availability of preimplantation genetic testing of ploidy makes it possible to determine whether there is a relationship between the size of a Graafian follicle and the ploidy of the oocyte/embryo that comes from that follicle, specifically its ability to progress from the oocyte stage and become a euploid blastocyst. We have tracked individual oocytes and the embryos created using these oocytes from follicles of known diameter and related their outcomes to the size of their follicles. Our approach is unique in that it examines the distribution of follicle sizes rather than relying on analyses of mean follicle sizes or follicles grouped into arbitrary bins (whereby the selection of bin limits can create an artificial bias in the analysis). Materials and Methods The subjects of this study were 22 anonymous oocyte donors between the ages of 21 and 34 years undergoing COH and oocyte retrieval for the purpose of donating oocytes to unknown recipients at our center. Donors underwent informed consent indicating that they were aware that the oocyte retrieval would involve collection of information about follicle sizes and ultimate outcomes of the oocytes and that the measurement procedure would lead to a more prolonged oocyte retrieval requiring long administration of anesthesia. GnRH analogs were administered to donors to avoid an uncontrolled mid-cycle luteinizing hormone surge. GnRH antagonist (Cetrotide, Merck Serono, Germany) was administered to most of the donors. It was administered daily beginning the evening when one follicle attained a diameter of 14 mm. Exogenous gonadotropin was administered daily with doses specified by the monitoring physician. All donors received recombinant FSH (Gonal-F, Merck Serono, Germany), and some donors received highly purified human menopausal gonadotropin (hMG, Menopur, Ferring Pharmaceuticals, Switzerland). The relative proportions of FSH/hMG were determined by the physician managing the stimulation of the donor. The ratio of hMG to total gonadotropin dose ranged from 0 to 100%. Doses administered on the first 2 days of injections were twice the amount anticipated to be the dose necessary for follicular growth (McCulloh et al., 2012; Charkviani et al., 2014). Donor monitoring (serum levels of FSH, LH, estradiol, progesterone and ultrasound assessment of follicular sizes) was performed after the fourth night of gonadotropin administration and every day or two thereafter as decided by the monitoring physician. Dose adjustments were made as dictated by the donor’s response during monitoring. Two donors received a triggering dose of 10 000 IUs of human chorionic gonadotropin. Two donors received a triggering dose of 2 mg of triptorelin acetate (Decapeptyl, Ferring Pharmaceuticals, Switzerland). Fourteen donors received triggering doses of GnRH agonist (triptorelin, 2 mg) and 1500 IU of human chorionic gonadotropin. The trigger was administered in the evening on the day when 20% of the follicles greater than 12 mm had achieved diameters greater than 16 mm. The retrieval was scheduled 35 h after the trigger injection. Donors underwent follicular aspiration in the dorsal lithotomy position under conscious sedation. Follicular aspirates were obtained using a transvaginal approach with ultrasound guidance using a 17-gauge needle using suction pressure differential of 120 mm Hg. Follicles were scanned more carefully than in a typical retrieval. Each follicle was measured immediately prior to its puncture and aspiration. Measurements were made in two dimensions, and the arithmetic average of the two diameters was calculated for use. Follicles were aspirated one at a time to be certain that the aspirated oocyte was from the same follicle measured. The follicle measurement (size) was noted by the embryologist and the oocyte’s location in the culture dish was noted in association with the follicle size. In rare cases, where a follicle was not aspirated or when multiple oocytes were aspirated from one aspiration, these follicle sizes and oocytes were excluded from the analysis. In cases with large numbers of follicles, only one ovary’s follicles were measured. Oocytes were retrieved from the other ovary without follicle measurements. Oocytes from the unmeasured follicles were segregated from the oocytes that were retrieved from measured follicles and were not used in the analysis. Oocytes were cultured individually throughout their time in the embryology laboratory so that follicle sizes could be uniquely associated with each oocyte. Oocytes were inseminated by ICSI as this was deemed necessary by the semen analysis of the recipient’s partner. Oocytes were cultured in Quinn’s Advantage Fertilization medium (ref no.: ART-1020, Origio, Netherlands) until the fertilization assessment. Fertilized oocytes (those observed with one or two pronuclei (PN) between 16 and 18 h after insemination) were moved to drops of Quinn’s Advantage Cleavage (ref no.: ART-1026, Origio, Netherlands) culture medium for individual embryo culture immediately after fertilization scoring. On the third day of embryo development, embryos underwent zona perforation with a laser and then were moved to drops of Quinn’s Advantage Protein Plus Blastocyst Medium (ref no.: ART-1529, Origio, Netherlands) for extended individual embryo culture. Embryos were examined on Days 5 and 6. Blastocysts were graded according to the Gardner grading method (Gardner et al., 2000). Embryos achieving the blastocyst stage/grade were considered quality blastocysts when they had grades of 2BC or 2CB or better on Day 5 or day 6 of culture. Quality blastocysts underwent biopsy of roughly five to seven trophectoderm cells for preimplantation genetic testing of aneuploidy (PGT-A). The biopsy was incised using a minimum number of laser pulses directed at regions showing particular cell-to-cell adherence. Immediately following biopsy, each blastocyst was vitrified. Biopsy specimens in small-capped reaction tubes were sent to Reprogenetics/Cooper Genomics (New Jersey, USA or UK) for comprehensive chromosomal analysis by next generation sequencing (NGS). Upon receipt of the NGS results, data were assembled for each oocyte and follicle size pair. The data included maturity of the oocyte after removal of cumulus and corona cells in preparation for ICSI, in addition to whether the oocyte attained the following stages subsequently (fertilization with 2 PN, attainment of the blastocyst stage, the grades of its inner cell mass and its trophectoderm and its ploidy as measured by PGT-A). Data about oocyte maturity, its fertilization, blastocyst formation, blastocyst grades and the ploidy of blastocyst biopsies were evaluated with reference to the diameter of the follicle from which the oocyte was retrieved. Follicle diameters were expressed in two ways: the diameter of the follicle (mm) and as Z values. The use of Z values was employed in order to minimize differences in the mean follicle diameter and result in the homogeneity of variance among different donors. Therefore, follicles sizes assessed as Z values were independent of the donor in which they were measured. Receiver operator characteristic curves were used to evaluate whether developmental stages were associated with follicle diameters. Results Characteristics for 22 donors are summarized in Table I. These values indicate that donors were young, anticipated to respond well with adequate downregulation prior to starting COH and with good responses. Table I Demographics and response characteristics for 22 donors. Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 1Mean ± standard deviation (N) 2Oral contraceptive pills used for synchronization of donor with recipient 3Six donors had measurements made on only one of the two ovaries retrieved 4The fraction of total gonadotropin administered that was provided by human menopausal gonadotropin (hMG) Open in new tab Table I Demographics and response characteristics for 22 donors. Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 1Mean ± standard deviation (N) 2Oral contraceptive pills used for synchronization of donor with recipient 3Six donors had measurements made on only one of the two ovaries retrieved 4The fraction of total gonadotropin administered that was provided by human menopausal gonadotropin (hMG) Open in new tab Three hundred thirty oocytes were retrieved from the 22 donors. Two follicles yielded empty zonae pellucidae and were excluded from further analysis. Three hundred seventeen oocytes (96.0%) were retrieved from follicles with measured follicle diameters. Of the oocytes with follicle sizes, 255 (80.4%) had a polar body (MII) and 60 (18.9%) were immature: 31 (9.8%) with a visible germinal vesicle (GV stage) and 29 (9.1%) with neither a polar body nor a visible germinal vesicle (MI). Of the 255 MII oocytes, 228 (89.4%) were fertilized (with two pronuclei), 94 became blastocysts that were biopsied (36.8% of the MII oocytes or 41.2% of the two PN oocytes) and 51 were euploid blastocysts (20.0% of the MII oocytes or 54.3% of the biopsied blastocysts). Three hundred seventeen follicles from which oocytes were retrieved averaged 17.4 ± 2.9 mm in diameter (N = 317 follicles). The average and standard deviation of each of the 22 donor’s mean follicle diameters was determined. This mean follicle diameter and standard deviation averaged among donors was 17.6 ± 1.6 mm (N = 22 donors). The largest follicle diameter for each donor’s follicles averaged 21.1 + 1.3 mm (N = 22 donors). Follicle size and oocyte maturity We determined the maturity status of each oocyte subjected to cleanup prior to the performance of ICSI. The size of the follicle from which each oocyte was retrieved was known. Follicle sizes for oocytes observed with a germinal vesicle (n = 31) were 12.5 ± 1.6 mm in diameter (mean + standard deviation). Follicle sizes for oocytes observed without a germinal vesicle and without a polar body (MI oocytes; n = 29) were 15.3 ± 3.2 mm in diameter. Follicle sizes for oocytes observed with a polar body (MII oocytes; n = 255) were 18.3 ± 2.2 mm in diameter. Although there were significant differences between follicle sizes for the three different maturities, there was significant overlap of the distributions (Fig. 1A). Follicles smaller than 12 mm yielded only GV oocytes (100% GV oocytes; 4/4). Follicles between 12 and 14.6 mm yielded oocytes of all three maturities: GV oocytes (48%; 27/56), MI oocytes (29%; 16/56) and MI oocytes (23%; 13/56). Follicles larger than 14.6 mm never yielded GV oocytes. However, follicles between 14.7 and 21.8 mm yielded either MI (5%; 13/238) or MII (95%; 225/238) oocytes. Follicles larger than 21.2 mm yielded only mature MII oocytes (100% MII oocytes; 17/17). The incidence of a mature MII oocyte increased as follicle size increased, greatly influenced by the high prevalence of mature MII oocytes. Follicle size and oocyte maturity. (A) Cumulative histograms of follicle diameters (mm) for oocytes with a polar body (MII) (black), oocytes with neither a germinal vesicle nor a polar body (MI) (blue) and oocytes containing a germinal vesicle (GV) (red). (B) Cumulative histograms of follicle diameters expressed as Z values (from each cycle) for MII (black), MI (blue) and GV (red) oocytes. (C) Incidence (P) of occurrence of MII (black), MI (blue) and GV (red) oocytes according to Z value. Incidences were calculated using a rolling average of 21 adjacent Z values plotted at the median Z value. Above Z values of −0.5, roughly 95% of the retrieved oocytes were mature, MII oocytes. Below Z values of −0.5, the fraction of retrieved oocytes that were MI and GV oocytes increased. Figure 1 Open in new tabDownload slide Figure 1 Open in new tabDownload slide Follicle sizes (mm) differed significantly among donors (one-way ANOVA, P = 0.00013). An additional method of comparison was applied to determine each follicle and its oocyte in terms of its size relative to the follicles retrieved in that particular donor’s cycle. In order to perform this analysis each follicle’s size was expressed as a Z value:|$Z$| = {follicle diameter (mm) − mean diameter (mm)}/S.D. (mm)where mean diameter is the mean diameter of all follicles measured for the same donor, and S.D. is the standard deviation of all follicles measured for the same donor. Conversion to Z values results in each follicle having a Z value such that the mean follicle diameter within each donor’s cohort of follicles has a Z value of 0, and each follicle that has a diameter that is one standard deviation above or below that donor’s cohort’s mean value has a Z value of 1 or −1, respectively. This transformation resulted in Z values for follicle diameters with no significant differences among donors (one-way ANOVA; P = 1.0). Therefore, we assumed that each follicle or oocyte from that follicle was an independent variate in our analyses, since the Z values for follicle diameters were not dependent on the donors in which they were measured. Using this Z value method for expressing follicle size, the follicles from which MII oocytes were retrieved had mean Z values of 0.26 ± 0.80. The use of Z values also resulted in no significant difference among donors in the diameters of follicles from which mature (MII) oocytes were retrieved (one-way ANOVA; P = 0.68). Follicles from which MI oocytes were retrieved had mean Z values of −0.81 ± 0.90, and follicles from which GV oocytes were retrieved had mean Z values of −1.60 ± 0.54 (see Fig. 1B). Oocyte maturity varied according to the diameters of the follicles from which they were retrieved (Fig. 1C). Follicles with Z values increasingly larger than −1.25 had oocytes with increasingly higher percentages of mature oocytes with a polar body (MII). Above Z values of −0.5, the incidence of mature (MII) oocytes was maximal, and MII oocytes were retrieved from greater than 90% of the follicles. Follicles of sizes decreasingly smaller than Z values of −0.5 were increasingly more likely to yield immature oocytes with a GV. Follicles with Z values near −1.5 reached a peak probability of yielding an immature oocyte (MI). However, MI oocytes were found in follicles with a wide range of follicle sizes spanning all Z values between −1.8 to 1.7. The ability of follicle size to predict oocyte maturity was examined using receiver operator characteristic (ROC) curve analysis (Fig. 2). As the criterion diameter was decreased from 24 to 12 mm, more and more follicles had diameters larger than the criterion and were included. With the criterion at 16 mm, 90% of the follicles from which mature oocytes were retrieved were identified (90% true positives = 90% sensitivity) and only 18% of the follicles from which immature oocytes were identified (18% false positives = 82% specificity). The area under the curve (AUC) for the ROC curve using follicle sizes larger than a criterion value to predict MII oocytes was 0.87 for follicles measured in mm and was 0.86 for follicles measured using Z values. These ROC curves and their AUCs indicated that follicle size was a significant predictor of MII oocytes (Mann–Whitney U = 13 754 or 13 438; z = 9.5; P < 0.0001; Table II). Use of a Z value criterion greater than −0.32 resulted in a sensitivity of 0.78 and a specificity of 0.84. Receiver operator characteristic (ROC) curves indicating the ability of follicle diameter (in mm) to predict oocyte characteristics. Each ROC curve was created by sorting all the follicle diameters by size. All follicles above (or below) a selected criterion (threshold) diameter were examined to determine if they were true positives (the criterion correctly predicted the outcome) or false positives (the criterion incorrectly predicted the outcome). For example, in the ROC curve for GV oocytes, the point on the curve at (1 − specificity) = 0.1 and sensitivity = 1.0 indicates that of the oocytes from follicles smaller than the selected criterion diameter, 100% of the GV oocytes were identified as true positives and only 10% of the non-GV oocytes were identified as false positives. Each ROC curve examines the effect of changing the criterion throughout the range of all follicle diameters, with a different point on the plot representing the sensitivity and specificity for a different criterion diameter. ROC curves near to the diagonal reference are considered incapable of predicting the outcome whereas ROC curves that approach the upper left corner (at 0, 1) are considered excellent predictors of the outcome. The area under the ROC curve (AUC) is used as an estimate of the quality of the test ranging from 0.5 (the area under the diagonal) to 1.0 (the area of the entire plot area). Figure 2 Open in new tabDownload slide Figure 2 Open in new tabDownload slide Table II Areas under the curve for receiver operator characteristic (ROC) curves. Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 1ROC curves used simple monotonic criteria (all follicles either larger than or smaller than the criterion level were assessed for sensitivity and specificity). (Larger) or (smaller) is displayed to indicate whether diameters larger the criterion or smaller than the criterion were used to predict the outcome. 2The area under the ROC curve (AUC) is an estimate of how well diameter predicts the listed outcome. The significance of the AUC was compared to a value of 0.5 to determine statistical significance using the Mann-Whitney U Test. 3An oocyte with a polar body (MII). 4AUC was significantly greater than 0.5; Mann–Whitney U test; P < 0.0001. 5An oocyte with a germinal vesicle (GV) 6AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.01. 7Inner cell mass (ICM). 8Trophectoderm (TE) 9AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.03. Open in new tab Table II Areas under the curve for receiver operator characteristic (ROC) curves. Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 1ROC curves used simple monotonic criteria (all follicles either larger than or smaller than the criterion level were assessed for sensitivity and specificity). (Larger) or (smaller) is displayed to indicate whether diameters larger the criterion or smaller than the criterion were used to predict the outcome. 2The area under the ROC curve (AUC) is an estimate of how well diameter predicts the listed outcome. The significance of the AUC was compared to a value of 0.5 to determine statistical significance using the Mann-Whitney U Test. 3An oocyte with a polar body (MII). 4AUC was significantly greater than 0.5; Mann–Whitney U test; P < 0.0001. 5An oocyte with a germinal vesicle (GV) 6AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.01. 7Inner cell mass (ICM). 8Trophectoderm (TE) 9AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.03. Open in new tab The AUC for ROC curves is evaluated by criteria that are generally accepted. The AUC as a percentage is evaluated like a grade on an examination. Percentages above 90% are considered ‘Grade A’ tests; between 80 and 90% are considered ‘Grade B’ tests; between 70 and 80% are considered ‘Grade C’ tests; between 60 and 70% are considered ‘Grade D’ tests; and percentages lower than 60% are considered poor tests. Therefore, the use of follicle size to predict MII oocytes is considered a Grade B test. The AUC for the ROC curve using follicles sizes smaller than a criterion value to predict GV oocytes was 0.96 for follicles measured in millimeters and was 0.94 for follicles measured using Z values. This indicates that follicle diameter is a Grade A test and able to predict very well when a GV oocyte will be retrieved from a follicle. These ROC curves and the AUC indicated that follicle size was a significant predictor of GV oocytes (Mann–Whitney U = 8511 or 8334; z = 8.4; P < 0.0001; Table II). Use of a criterion less than 14.6 mm resulted in a sensitivity of 1.0 and a specificity of 0.90. Use of a Z value criterion of less than −1.0 resulted in a sensitivity of 0.90 and a specificity of 0.88. Follicle size and fertilization The sizes of follicles from which 228 oocytes (that were fertilized with two pronuclei) came (18.1 ± 2.3 mm) were indistinguishable from the sizes of follicles from which mature MII oocytes came (18.2 ± 2.2 mm; P = 0.35). Further, the sizes of follicles from which oocytes that were fertilized originated were indistinguishable from the sizes of follicles from which the 37 oocytes that did not have two pronuclei (<2 or >2 pronuclei) originated (18.3 ± 2.3 mm; P = 0.54). Expressed as Z values, follicles from which MII oocytes were retrieved that had two pronuclei (0.19 ± 0.84) following ICSI were not significantly different from follicles with MII oocytes that did not have two pronuclei (0.35 ± 0.83; P = 0.28). The ability of follicle diameter to predict whether a MII oocyte will become a 2 PN oocyte (fertilized) was examined using ROC curve analysis (Table II). The AUC for the ROC curve examining the predictive value of follicle diameters (mm) greater than a criterion was 0.53. The AUC for the ROC curve examining the predictive value of follicle diameters’ Z values greater than a criterion was 0.54. These AUCs indicated that follicle diameter is a poor predictor of fertilization both because the AUCs for the ROC curves are too small to qualify as good tests using the ‘rule of thumb’ method and because the AUCs were not significantly different from 0.5 (Mann–Whitney U tests not significant). Follicle size and formation of blastocysts The sizes of follicles from which 94 oocytes that became quality blastocysts originated (18.6 ± 2.1 mm) was significantly different than the sizes of the follicles from which 181 oocytes originated that were fertilized but did not become quality blastocysts (17.9 ± 2.4 mm; P = 0.028). Expressed as Z values, follicles from which oocytes that became quality blastocysts arose (0.36 ± 0.73) were not significantly different than follicles from which oocytes originated that were fertilized but did not become quality blastocyts (0.16 ± 0.88; P = 0.06). The divergence of the two cumulative histograms (Fig. 3A and B) suggests that oocytes from smaller follicles (Z values less than 0–0.5) were less prevalent for biopsied blastocysts than for 2 PN embryos that did not become biopsied blastocysts. The sizes of the follicles from which oocytes that became quality blastocysts originated was significantly different than the sizes of follicles from which all oocytes that were fertilized (2 PN) originated regardless of whether or not they became quality blastocysts (18.3 ± 2.3 mm; P = 0.048). Examination of the incidence of quality blastocyst formation relative to follicle size indicated that the incidence of blastocyst formation increased as follicle size increased (R = 0.19; R2 = 0.036). This was particularly apparent for 17 2-pronuclear embryos that originated in follicles less than 15 mm in diameter. There were no blastocysts from these 2-pronuclear embryos. Formation of quality blastocysts in relation to follicular Z value. (A) Cumulative histogram of embryos containing two pronuclei that became quality blastocysts undergoing trophectoderm biopsy (TE Bx) or embryos containing two pronuclei that did not become quality blastocysts undergoing trophectoderm biopsy (2 PN no TE Bx). (B) The incidence of becoming a quality blastocyst is displayed as a rolling average of 29 adjacent oocytes (vertical axis) sorted by their follicular Z values and plotted at the median Z value (horizontal axis). Figure 3 Open in new tabDownload slide Figure 3 Open in new tabDownload slide The ability of follicle diameter to predict whether a 2 PN oocyte will become a quality blastocyst was examined using ROC curve analysis (Table II). The AUC for the ROC curve examining the predictive value of follicle diameters (mm) greater than a criterion was 0.59 (Fig. 2). The AUC for the ROC curve examining the predictive value of follicle diameters’ Z values greater than a criterion was 0.57. The general method interpreting these AUCs indicates that follicle diameter is a poor predictor of 2 PN oocytes that will become quality blastocysts. Despite this, the ROC curves and the AUC indicate that follicle size larger than a criterion is a significant predictor of 2 PN oocytes becoming quality blastocysts (Mann–Whitney U = 7386; z = 2.31; P = 0.01 for mm; or U = 7132; z = 1.81; P = 0.035 for Z values; Table II). Although a criterion of 15–17 mm for follicle diameter provided sensitivities of 1.0–0.71, respectively, they resulted in specificities of only 0.12–0.33, respectively. Z value criteria of between −1.0 and −0.5 provide sensitivities of 0.95–0.87, respectively, but result in specificities of only 0.14–0.16. The ability of follicle diameter to predict whether a mature (MII) oocyte will become a quality blastocyst was also examined using ROC curve analysis (Table II). The AUC for the ROC curve was 0.53 for both follicle diameters measure both in mm and using Z values. Neither AUC was significantly different than 0.5 (Mann–Whitney U tests: U = 566, z = 1.2, P = 0.11 for mm; U = 440, z = 0.77, P = 0.22 for Z values). These indicate that follicle diameter is a poor predictor for formation of quality blastocysts from MII oocytes. Follicle size and blastocyst grades Among quality blastocysts there was no significant difference between follicle diameters for oocytes that became blastocysts with the inner cell mass graded A (18.6 ± 2.1 mm; N = 50), B (17.1 ± 5.5 mm; N = 35) or C (18.4 ± 1.7 mm; N = 12) (ANOVA; F = 1.6; P > 0.05). Expressed as Z values there was no significant difference between follicle diameters for oocytes that became blastocysts with the inner cell mass graded A (0.37 ± 0.65), B (0.35 ± 0.68) or C (0.14 ± 0.98) (ANOVA; F = 0.51; P > 0.05). ROC curves and the AUCs indicated that follicle size is not a significant predictor of ICM grades (Mann–Whitney U = 65.8 or 45.9; not significant) (Table II). For quality blastocysts, there was no significant difference between follicle diameters for oocytes that became blastocysts with trophectoderm grades A (17.9 ± 2.2 mm; N = 22), B (17.8 ± 4.5 mm; N = 58) or C (19.0 ± 1.9 mm; N = 17) (ANOVA; F = 0.70; P > 0.05). Expressed as Z values there was no significant difference between follicle diameters for oocytes that became blastocysts with the trophectoderm grades A (0.19 ± 0.60), B (0.41 ± 0.71) or C (0.28 ± 0.81) (ANOVA; F = 1.41; P > 0.05). ROC curves and the AUC provided conflicting results. Follicle size in mm smaller than a criterion value was a significant predictor of blastocysts with a trophectoderm grade of A (Mann–Whitney U = 212.3; z = 1.87; P = 0.031) whereas follicle size in Z values was not a significant predictor (Mann–Whitney U = 179.6; z = 1.58; not significant) (Table II). Follicle size and blastocyst euploidy The incidence of euploidy for quality blastocysts was 54.3% (51/94). The sizes of follicles from which 51 oocytes that became euploid blastocysts were retrieved (18.7 ± 2.1 mm) was indistinguishable from the sizes of follicles from which 30 oocytes that became aneuploid blastocysts came (18.7 ± 2.2 mm; P = 0.88). Expressed as Z values, follicles from which euploid blastocysts came (0.39 ± 0.70) were indistinguishable from follicles from which aneuploid blastocysts came (0.29 ± 0.75; P = 0.58). The ROC curves and the AUCs indicated that follicle size was not a significant predictor of euploid blastocysts (among diagnosed blastocysts) (Mann–Whitney U = 15.3; z = 0.15; not significant) (Fig. 2; Table II). Further, the ROC curves examining the diameters for follicles from which mature MII oocytes were retrieved were examined to determine if diameter could predict which of these oocytes would become quality blastocysts with a euploid biopsy. The AUCs for the ROC curves were 0.55 and 0.53 for follicle diameters measured in mm and using Z values, respectively. Neither AUC was significantly different than 0.5 (Mann–Whitney U tests: U = 566, z = 1.2, P = 0.11 for mm; U = 340, z = 0.72, P = 0.24 for Z values). Both AUCs are indicative of poor ability of follicle diameter to predict which MII oocytes would become quality euploid blastocysts due to the low values of the AUC and their lack of statistical significance. We considered whether a selection bias existed for follicles from our donors. There was no significant association (as measured by correlation coefficient) between each donor’s incidence of euploidy and their mean follicle diameter (mm), the median follicle diameter (mm), the number of follicles, the standard deviation of follicle diameters (mm) or a measure of skewness (follicle diameter median − follicle diameter mean). Discussion Maturity Follicle sizes are significantly different between GV, MI and MII oocytes. The significant AUCs of ROC curves for MIIs (~0.86) and especially for GV oocytes (~0.95) indicate that follicle size is a strong predictor of maturity. In contrast to many of the prior studies, the use of ROC curves in this work investigates a wide spectrum of thresholds or criteria for predicting maturity using follicle diameter (or follicle Z value). Supporting a long held notion, oocytes from follicles smaller than 12 mm were more likely to be GV oocytes and oocytes from follicles larger than 17 mm were more likely to be mature MII oocytes (Wittmaack et al., 1994; Rosen et al., 2008; Kahraman et al., 2017; Wirleitner et al., 2018). However, the large degree of overlap seen in this study between follicles sizes of GV and MI oocytes and between follicle sizes of MI and MII oocytes indicates that follicle size is not an absolute predictor of oocyte maturity. Follicle Z values less than −1.7 were associated with a higher incidence of GV oocytes and follicle Z values greater than −0.9 were associated with a higher incidence of MII oocytes. Although the plots using Z values were much smoother, both the plots of follicle diameter and Z values showed significant overlap of follicles size measures for follicles from which oocytes of different maturities were retrieved. The incidences of oocytes of the three maturity levels in the different regions of follicles sizes supports the notion that follicle size does not absolutely predict the maturity of the oocyte it yields. The ability of smaller follicles to predict the presence of a GV (AUC ~ 0.95) is excellent. The ability of larger follicles to predict the presence of a polar body (MII oocyte; AUC ~0.86) is quite good. In both cases, this predictive value applies for follicle diameters measured following an exogenous ovulatory trigger (hCG and/or GnRH agonist). The higher AUC value for GV oocytes as compared with MII oocytes suggests that follicle size is a better predictor of GV breakdown (the transition for GV oocyte to MI oocyte) than of polar body extrusion (the transition from MI oocyte to MII oocyte) after administration of the trigger. Fertilization Low values for the AUC (0.53, 0.54) for the ROC curves comparing the appearance of 2 PN in MII oocytes following ICSI indicate that follicle size is not predictive of fertilizability for MII oocytes. This is supported by the observation of no significant difference in the sizes of follicles from which oocytes that became two pronuclear embryos versus MII oocytes that did not become two pronuclear embryos. Without an association, there is no reason to believe that follicle size is predictive of fertilizability for an MII oocyte that came from the follicle. Follicular size is reported to be predictive of oocyte fertilization (Quigley et al., 1982; Rosen et al., 2008) particularly when standard insemination was performed on oocytes of unknown maturity. In this study, maturity was known prior to ICSI and follicle size was not related to fertilization, similar to previous findings with ICSI (Rosen et al., 2008; Kahraman et al., 2017; Wirleitner et al., 2018). Blastocyst formation The incidence of blastocyst formation (quality blastocysts arising from 2 PN oocytes) increased with increasing follicle diameter. In contrast, the incidence of blastocyst formation (quality blastocysts arising from MII oocytes) was not related to follicle diameter. However, the wide diversity in blastocyst formation for two pronuclear embryos and the low value of the correlation coefficient (R2 = 0.036) that arose from follicles of different sizes suggests that despite the presence of a significant association, follicle size alone is unlikely to be the determinant of blastocyst formation. Significant differences in blastocyst formation have been reported (Kahraman et al., 2017; Wirleitner et al., 2018) when considering the formation of blastocysts from oocytes without regard to maturity or fertilization. In this study, the significant AUC values (~0.58) for the ROC curves of follicle size’s (both mm and Z value) ability to predict that a 2 PN oocyte will become a quality blastocyst indicate that follicle size is a valid predictor of 2 PN oocyte’s competence to become a quality blastocyst. This observation supports the significant differences seen according to follicle size in blastocyst formation per 2 PN oocyte (Kahraman et al., 2017). However, the low AUC values for ROC curves indicate that a significant trade-off of specificity for sensitivity is necessary when selecting a threshold follicle diameter to predict blastocyst formation for oocytes with 2 PN. The poor quality of follicle diameter as a test to predict blastocyst formation is consistent with lack of significance seen by others per MII oocyte (Wirleitner et al., 2018). One anecdotal observation from ROC curves was that blastocyst formation occurred preferentially for 2 PN oocytes arising from larger follicles; whereas, blastocyst formation occurred preferentially for MII oocytes arising from smaller follicles (although this preference was not significant). Similarly, fertilization of MII oocytes (the incremental transition from MII oocyte to 2 PN oocyte) occurred preferentially for oocytes arising from smaller follicles (although, this preference, too, was not significant). Is it possible that different incremental transitions of oocytes/embryos may occur preferentially for smaller or for larger follicles? If this is possible, then incremental preferences could cancel for longer periods of development spanning multiple incremental steps leading to an apparent lack dependence on follicle size. Blastocyst grades The sizes of follicles from which oocytes that became quality blastocysts originated were not significantly different in relation to grades of either the inner cell mass or trophectoderm. Despite this, the ROC curve indicated that better trophectoderm scores occurred in association with smaller follicle sizes (when measured in mm). In contrast, measurement of follicle size in Z values was not predictive. These conflicting results compare with the conflicting observations that larger follicles led to a higher incidence of top quality and good-quality blastocysts (Kahraman et al., 2017) and that there is no significant association between follicle size and blastocyst grades (Wirleitner et al., 2018). Euploidy The follicle sizes and distribution of follicle sizes from which oocytes that became euploid blastocysts were indistinguishable from the follicle sizes and distribution of follicles from which oocytes that became aneuploid blastocysts. The AUCs of the ROC curves (~0.51) indicate that follicle size is not predictive of euploidy for quality blastocysts. AUC values were low (0.53) for ROC curves examining the ability of diameters of follicles from which MII oocytes were retrieved to predict which oocytes would become quality blastocysts. This low value indicates that follicle size is unable to predict which oocytes will become quality euploid blastocysts. The observation that no significant correlations were found between donors’ euploidy rates and donor-specific measures of follicle size, number, dispersion and skewness support the notion that no particular bias of follicle size distribution confounded our conclusions. Maturity was significantly associated with follicle diameter. However, the subsequent steps: fertilization (by ICSI), formation of quality blastocysts, blastocyst grades and blastocyst euploidy each occurred in oocytes/embryos across a wide range of follicle diameters and with no apparent preference for smaller or larger follicles. From this, we conclude that there is no preferential follicle size to achieve euploid blastocysts. Euploid blastocysts may arise following fertilization of oocytes from follicles regardless of size as long as the follicle yields a mature oocyte (any follicle diameter ≥12 mm). Association between follicle size and oocyte quality? There is little evidence that follicle size is predictive of the quality of the oocyte within. However, there is a suggestion that oocytes from smaller follicles are less likely to be mature (more likely to be GV oocytes) and are less likely to become quality blastocysts (more likely not to be quality blastocysts despite having been fertilized). This knowledge may be helpful in estimating the number of mature oocytes that will be retrieved especially for programs performing oocyte cryopreservation or oocyte banking. In the hypothesis of follicular dominance for natural menstrual cycles (Speroff and Fritz, 2005; Strauss and Barbieri, 2014), the dominant follicle is generally believed to be the most developed (largest?) follicle with the most receptors for follicle stimulating hormone. The follicle with the most follicle-stimulating hormone receptors is capable of continuing to grow as follicle stimulating hormone levels decline from Day 3 onward. During COH with exogenous application of follicle-stimulating hormone, follicle size hierarchy seems relatively unchanged, with the largest follicle continuing as the largest follicle and smaller follicles continuing to grow albeit as smaller follicles. The largest follicle at trigger is typically a follicle that was the largest follicle or among the largest follicles throughout COH. The observation that larger follicles are not more likely to yield oocytes that develop into euploid blastocysts suggests that the follicular dominance is unrelated to the ploidy of the oocyte within. Selection of follicles for retrieval by size is unlikely to result in a higher or lower incidence of euploidy for blastocysts resulting from the retrieved oocytes. Limitations The study was limited to oocyte donors so that it could address differences in euploidy between centers (Munné et al., 2017) and between physicians within the same center (McCulloh et al., 2019). It is not clear whether these observations can be generalized to non-donor patients who may have similar or lower incidences of euploidy. Follicle size was determined by direct ultrasound measurement during oocyte retrieval. We chose to perform ultrasound measurements rather than determination of follicular aspirate volumes because we wished to decrease the time required to determine the follicular volume, during which oocytes might be exposed to less controlled conditions. Others have demonstrated that direct measurements are well correlated (r2 = 0.79; r = 0.89) with measurements of follicular volume, indicative that ultrasound measurements are comparable to measurements of follicle volume (Wittmaack et al., 1994) with 79% of the variability in follicle diameter attributed to variability of follicle volume and only 21% of the variability remaining as unexplained (residual) variability. Therefore, we believe that the follicle diameter measurements are sufficient to estimate follicle volume with reasonable errors. We must temper our interpretations based on the direct observations made in our work. Follicle diameters were measured, and for the retrieved oocytes, maturity and developmental measures of embryos resulting from their fertilization were determined. Among the developmental measures was the ploidy of biopsies sampled from embryos that became blastocysts. We directly assessed neither the ploidy of oocytes nor the ploidy of embryos that were not quality blastocysts. It is tempting to speculate that the ploidy of blastocysts is attributable to meiotic errors in oocytes from which the blastocysts arose since neither fertilization, nor blastocyst formation, nor blastocyst grades, nor ploidy of blastocyst resulting from mature oocytes were associated with the size of the follicles from which the oocytes were retrieved. However, we must acknowledge that aneuploidy, as diagnosed from several biopsied trophectoderm cells, could arise from sperm. The relatively low incidence of aneuploidies originating in the sperm (McWilliams et al., 2015) can be a significant contributor to aneuploidy, especially when using donor oocytes (noted for their low incidence of aneuploidy) and when performing ICSI using sperm from men with elevated sperm aneuploidy in association with severe male factor (Levron et al., 2000; Tempest, 2011). Aneuploidy detected in trophectoderm biopsies analyzed by NGS can arise from mitotic errors (mosaicism) (Fragouli et al., 2013). The number of euploid embryos on Day 3 is indistinguishable from the number of euploid blastocysts on Days 5 and 6 (Adler et al., 2014; Demko et al., 2016), suggesting that extended culture of embryos to Days 5 and 6 preferentially selects euploid embryos. However, the mechanisms leading to similar numbers of euploid embryos may be more complex (Rabinowitz et al., 2012; Fragouli et al., 2013). Therefore, concluding that the observed ploidy of trophectoderm biopsies was directly attributable to the ploidy of oocytes retrieved remains disputable. Nevertheless, we have found no evidence in our study to support the notion that differences in the sizes of follicles retrieved could explain the differences in euploidy for different centers (Munné et al., 2017) or for different physicians within the same center (McCulloh et al., 2019). The number of larger follicles in a cycle as examined by Wittmack et al (1994) has not been evaluated in this study. In our study, the criterion for triggering was applied uniformly for all donors. This uniformity is supported by our observation that diameters of the largest follicles varied less than diameters of the mean follicles (Table I). Further investigation with more donors will be needed to determine if the diameters of the lead follicles are associated with the ploidy for the donor’s cohort of eggs retrieved or the embryos/blastocysts arising from those oocytes. Follicle size was analyzed by two methods: direct measurement of diameter and by Z values that expressed the follicle size relative to the mean follicle size of that donor’s cycle. These two methods generally led to concordant results, indicating that either method is useful for prediction. ROC curves are relatively independent of the distribution of data and most transformations of data (Kumar and Indrayan, 2011), since the data are dichotomized into true positives (sensitivity) and false positives (1-specificity) using a criterion value. However, in predicting trophectoderm scores, direct measurement led to a significant ability to predict whereas the Z values did not. It is unclear whether this discrepancy of results is due to a lack of independence that may be present in diameters (mm) but that is adjusted through the use of Z values. We believe that the use of Z values is the preferred method of analysis since Z values correct for the follicle sizes’ dependence on the donor from which oocytes were retrieved. Neither method should be considered a good predictor (AUCs 0.55 for mm and 0.52 for Z values). Therefore, we consider this a concordant observation that follicle size is a poor predictor of trophectoderm score. These data suggest that any follicle size is capable of yielding a euploid oocyte at least using the stimulation methodology in this study. It is unclear if pushing follicles to larger sizes or harvesting oocytes when follicles are smaller will influence the incidence of euploid oocytes or will affect the size of follicles from which euploid oocytes arise since only one stimulation paradigm was employed in this study. It must be noted that follicle sizes on the day of oocyte retrieval are not clearly predicted by the size of follicles on the day of triggering. Follicles may increase in diameter or decrease in diameter, and it is even further unclear as to whether some of the triggered follicles in an individual patient may increase in diameter whereas others decrease in diameter between triggering and retrieval. Summary The sizes of follicles from which oocytes were retrieved were examined. Mature oocytes are more likely to arise from follicles larger than 16 mm or with Z values larger than −0.32 with quite good sensitivity and specificity. GV oocytes are more likely to arise for follicles smaller than 14.6 mm or with Z values smaller than −1.0 with excellent sensitivity and specificity. The progression of fertilized oocytes to become quality blastocysts was more likely for larger follicles (using a similar threshold size), but follicle size is not a strong predictor of blastocyst formation. For either MII oocytes or 2 PN oocytes achieving a quality blastocyst stage there was no indication that euploidy or blastocyst grades were strongly associated with size of the follicles from which the oocytes arose that developed into these euploid blastocysts. Therefore, we feel confident to state that physicians may retrieve smaller follicles with confidence that if they are mature and fertilized, they will become euploid blastocysts with the same incidence as mature oocytes from larger follicles. Authors’ roles D.H.M. participated in conceiving the project, analysis of the data and writing of the manuscript. N.K., T.C., T.Z. and T.B. participated in collection and assembly of the data as well as in writing of the manuscript. S.M. and L.C. participated in conceiving the project as well as in writing the manuscript. Funding ReproART: Georgian American Center for Reproductive Medicine. Conflict of interest None of the authors declare any actual conflicts of interest. However, D.H.M. declares that he receives compensation from ReproART, Biogenetics Corporation and the Sperm and Embryo Bank of New York. In addition, he received honoraria and travel funding from Ferring Pharmaceuticals and from Granata Bio during the period of the study. S.M. declares that he received compensation from Cooper Genomics during the period of the study. In addition, he received an honorarium and travel funding from Ferring Pharmaceuticals during the period of the study. L.C. declares that she is the founder of LTD Ovamedi, the organization that represents Cooper Genomics in Georgia, and received travel funding from the European Society for Human Reproduction and Embryology during the period of the study. References Adler A , Lee H-L, McCulloh DH, Ampeloquio E, Clarke-Williams M, Hodes-Wertz B, Grifo J. Blastocyst culture selects for euploid embryos: comparison of blastomere and trophectoderm biopsies . Reprod Biomed Online 2014 ; 28 : 485 – 491 . Google Scholar Crossref Search ADS PubMed WorldCat Charkviani T , Chkonia L, Kurashvili N, Kutchukhidze N, Zhorzholadze T, McCulloh D. Administration of the first 2 doses of gonadotropin at twice the dose during controlled ovarian hyperstimulation decreases total gonadotropin administration during in vitro fertilisation(IVF) cycle . Fertil Steril 2014 ; 102 : e225 . Google Scholar Crossref Search ADS WorldCat Demko AP , Simon AL, McCoy RC, Petrov DA, Rabinowitz M. Effects of maternal age on euploidy rates in a large cohort of embryos analyzed with 24-chromosome single-nucleotide polymorphism-based preimplantation genetic screening . Fertil Steril 2016 ; 105 : 1307 – 1313 . Google Scholar Crossref Search ADS PubMed WorldCat Fragouli E , Alfarawati S, Spath K, Jaroudi S, Sarasa J, Enciso M, Wells D. The origin and impact of embryonic aneuploidy . Hum Genet 2013 ; 132 : 1001 – 1013 . Google Scholar Crossref Search ADS PubMed WorldCat Gardner DK , Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: towards a single blastocyst transfer . Fertil Steril 2000 ; 73 : 1155 – 1158 . Google Scholar Crossref Search ADS PubMed WorldCat Kahraman S , Pirkevi Cetinkaya C, Cetinkaya M, Yelke H, Kumtepe Colakoglu Y, Aygun M, Montag M. The effect of follicle size and homogeneity of follicular development on the morphokinetics of human embryos . J Assist Reprod Genet 2017 ; 34 : 895 – 903 . Google Scholar Crossref Search ADS PubMed WorldCat Kumar R , Indrayan A. Receiver operator (ROC) characteristic curve for medical researchers . Ind Pediat 2011 ; 48 : 277 – 287 . Google Scholar Crossref Search ADS WorldCat Levron J , Aviram-Goldring A, Mdgar I, Wiessenberg R, Bachar A, Dor J. The prevalence of sperm chromosome aneuploidies in severe male factor infertility patients before and after sperm cell sorting under light microscopy . Fertil Steril 2000 ; 74 : S108 . Google Scholar Crossref Search ADS WorldCat McCulloh DH , Alikani M, Norian J, Kolb B, Arbones JM, Munné S. Controlled ovarian hyperstimulation (COH) parameters associated with euploidy rates in donor oocytes . Eur J Med Genet 2019 ; 62 : Article 103681 . OpenURL Placeholder Text WorldCat McCulloh DH , Colon JM, McGovern PG. Modeling follicle stimulating hormone levels in serum for controlled ovarian hyperstimulation III: improved gonadotropin administration . Curr Pharm Biotechnol 2012 ; 13 : 454 – 463 . Google Scholar Crossref Search ADS PubMed WorldCat McWilliams K , McWilliams TK, Wyatt M, Hughes M. Aneuploidy parent of origin in blastocyst biopsies using karyomapping technology . Fertil Steril 2015 ; 104 :e283. OpenURL Placeholder Text WorldCat Miller KF , Goldberg JM, Falcone T. Follicle size and implantation of embryos from in vitro fertilization . Obst Gynec 1996 ; 88 : 583 – 586 . Google Scholar Crossref Search ADS WorldCat Munné S , Alikani M, Ribustello L, Cols P, Martinez-Ortiz P, Referring Physician Group , McCulloh D. Euploidy rates in donor egg cycles significantly differ between fertility centers . Hum Reprod 2017 ; 32 : 743 – 749 . Google Scholar Crossref Search ADS PubMed WorldCat Nivet AL , Léveillé MC, Leader A, Sirard MA. Transcriptional characteristics of different sized follicles in relation to embryo transferability: potential role of hepatocyte growth factor signaling . Molec Hum Reprod 2016 ; 22 : 475 – 484 . Google Scholar Crossref Search ADS PubMed WorldCat Quigley MM , Wolf DP, Maklad NF, Dandekar PV, Sokoloski JE. Follicular size and number in human in vitro fertilization . Fertil Steril 1982 ; 38 : 678 – 681 . Google Scholar Crossref Search ADS PubMed WorldCat Rabinowitz M , Ryan A, Gemelos G, Hill M, Baner J, Cinnioglu C, Ganjevic M, Potter D, Petrov DA, Demko Z. Origins and rates of aneuploidy in human blastocysts . Fertil Steril 2012 ; 97 : 395 – 401 . Google Scholar Crossref Search ADS PubMed WorldCat Rosen MP , Shen S, Dobson AT, Rinaudo PF, McCulloch CE, Cedars MI. A quantitation assessment of follicle size on oocyte developmental competence . Fertil Steril 2008 ; 90 0015-0282/08/$34.00 . doi: 10.1016/j.fertnstert.2007.02.011 . OpenURL Placeholder Text WorldCat Crossref Speroff L , Fritz MA. Clinical Gynecologic Endocrinology and Infertility , 7th edn. Philadelphia : Lippincott Williams and Wilkins , 2005 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Strauss JF , Barbieri RL. Yen & Jaffe’s Reproductive Endocrinology E-Book : Physiology, Pathophysiology, and Clinical Management , 7th edn. Philadelphia : Elsevier Saunders , 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Tempest HG . Meiotic recombination errors, the origin of sperm aneuploidy and clinical recommendations . Syst Biol Reprod Med 2011 ; 57 : 93 – 101 . Google Scholar Crossref Search ADS PubMed WorldCat Wirleitner B , Okhowat J, Vistejnova L, Kralickova M, Karlikova M, Vanderzwalmen P, Ectors F, Hradecky L, Schuff M, Murtinger M. Relationship between follicular volume and oocyte competence, blastocyst development and live-bith rate: optimal follicle size for oocyte retrieval . Ultrasound Obstet Gynecol 2018 ; 51 : 118 – 123 . Google Scholar Crossref Search ADS PubMed WorldCat Wittmaack FM , Kreger DO, Blasco L, Tureck RW, Mastroianni L, Lessey BA. Effect of follicular size on oocyte retrieval, fertilization, cleavage, and embryo quality in in vitro fertilization cycles: a 6-year data collection . Fertil Steril 1994 ; 62 : 1205 – 1210 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author(s) 2020. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Human Reproduction Oxford University Press

Follicle size indicates oocyte maturity and blastocyst formation but not blastocyst euploidy following controlled ovarian hyperstimulation of oocyte donors

Download PDF
12 pages

Loading next page...
 
/lp/oxford-university-press/follicle-size-indicates-oocyte-maturity-and-blastocyst-formation-but-oxKJLtW07P

References (25)

Publisher
Oxford University Press
Copyright
© The Author(s) 2020. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For permissions, please e-mail: [email protected].
ISSN
0268-1161
eISSN
1460-2350
DOI
10.1093/humrep/dez291
Publisher site
See Article on Publisher Site

Abstract

Abstract STUDY QUESTION Is there is an association between follicle size and the quality of oocytes retrieved from them as judged by ability to achieve the blastocyst stage, blastocyst grades and blastocyst ploidy? SUMMARY ANSWER Although follicle size is a valuable predictor of oocyte maturity and is a significant predictor of the ability of a fertilized oocyte to become a quality blastocyst, the ploidy of each quality blastocyst is not related to the size of the follicle from which its oocyte was retrieved. WHAT IS KNOWN ALREADY It is unclear whether the oocytes within larger follicles are the best oocytes of the cohort. Although there have been studies examining follicle size in relation to embryo quality, there has been no study relating the incidence of euploidy in embryos to follicle size. STUDY DESIGN, SIZE, DURATION The purpose of this study was to examine follicle sizes and the oocytes from those follicles (and the embryos that result from those oocytes) to see if there is an association between follicle size and the quality of oocytes as judged by ability to achieve the blastocyst stage, blastocyst grades and blastocyst ploidy. Follicle sizes for oocytes were assessed both as diameters (mm) and as Z values (expressed as their size relative to the mean and standard deviation of that donor’s follicular cohort). Comparisons were made using cumulative histograms, rolling averages and receiver operator characteristic (ROC) curves and its AUC. PARTICIPANTS/MATERIALS, SETTING, METHODS Twenty-two oocyte donors (ages: 24.5 ± 3.5 years) whose recipients would use ICSI for insemination were enrolled in this study. Follicles were aspirated one-at-a-time to be certain that the aspirated oocyte was from the same follicle measured. The follicle measurement (size) was noted in the embryology records. Oocytes were cultured individually throughout their time in the embryology laboratory so that follicle sizes could be uniquely associated with each oocyte. Oocytes and embryos were analyzed according to the size of the follicle from which the oocyte was retrieved. MAIN RESULTS AND THE ROLE OF CHANCE Three hundred seventeen oocytes (96.1%) had an associated follicle size. Of the oocytes with follicle sizes, 255 (80.4%) had a polar body (MII), and 60 (18.9%) were immature: 31 (9.8%) with a visible germinal vesicle (GV stage) and 29 (9.1%) with neither a polar body nor a visible germinal vesicle (MI). The incidence of MII oocytes was significantly associated with larger follicle size using either mm (ROC’s AUC = 0.87; P < 0.0001) or Z values (ROC’s AUC = 0.86; P < 0.0001). Among MII oocytes there was no association with follicle size for the appearance of 228 oocytes with two pronuclei (2 PN). Among 2 PN’s, the development of 94 quality blastocysts that underwent trophectoderm biopsy (TE Bx) exhibited a significant association with larger follicles using either mm (ROC’s AUC = 0.59; P = 0.01) or Z values (ROC’s AUC = 0.57; P = 0.01). The use of follicle diameter as a feature to distinguish between fertilized oocytes that would ultimately become blastocysts versus those that would not become blastocysts resulted in an enrichment for blastocyst formation from 20 to 40%. Of the 94 quality blastocysts, 51 were determined by next generation sequencing (NGS) to be euploid.Although oocyte maturity and the incidence of blastocyst formation were associated with follicle size, the incidence of euploidy among biopsied blastocysts was not. Follicles measured by two different methods (mm or Z values) led to predominantly the same conclusions. LIMITATIONS, REASONS FOR CAUTION This study investigated the relationship between follicle size and measures of oocyte/embryo quality when donors were treated similarly. Therefore, this study does not investigate the effects of triggering and retrieving oocytes when the follicle cohorts are of different sizes or lead follicles are of different sizes. Although no association was found between follicle size and euploid blastocysts, the fact that blastocyst ploidy is not entirely dependent upon oocyte ploidy (e.g. aneuploidies derived from mitotic errors or from the fertilizing sperm) makes it difficult to infer the relationship between follicle diameter and oocyte ploidy. WIDER IMPLICATIONS OF THE FINDINGS It is confirmed that follicle diameter is predictive of oocyte maturity. However, once oocyte maturity is known, the diameter of the follicle from which the oocyte was retrieved is not instructive. Embryos generated through fertilization and development of the mature oocytes from any observed follicle diameter were equally likely to become euploid blastocysts. STUDY FUNDING/COMPETING INTEREST(S) This study was funded by ReproART: Georgian American Center for Reproductive Medicine. None of the authors declare any actual conflicts of interest. D.H.M. received compensation from ReproART, Biogenetics Corporation and the Sperm and Embryo Bank of New York and honoraria and travel funding from Ferring Pharmaceuticals and from Granata Bio. S.M. received compensation from Cooper Genomics and an honorarium and travel funding from Ferring Pharmaceuticals. L.C. is the founder of LTD Ovamedi, the organization that represents Cooper Genomics in Georgia, and received travel funding from the European Society for Human Reproduction and Embryology. TRIAL REGISTRATION NUMBER N/A. Preimplantation genetic testing for aneuploidy (PGT-A), Follicle diameter, Oocyte maturity, Quality blastocyst formation, Blastocyst ploidy Introduction Although the largest follicle during controlled ovarian hyperstimulation is generally thought to be the follicle that would have been the dominant follicle in an unstimulated menstrual cycle, it is unclear whether the oocyte within that follicle is the best oocyte of the cohort. There have been publications indicating that oocytes from larger follicles exhibit a higher incidence of mature and/or fertilized oocytes (Quigley et al., 1982; Wittmaack et al., 1994; Rosen et al., 2008; Kahraman et al., 2017; Wirleitner et al., 2018), a higher incidence of normal cleavage (Kahraman et al., 2017) and a higher incidence of blastocyst formation (Kahraman et al., 2017; Wirleitner et al., 2018). Others suggest that oocytes from larger follicles are suboptimal (Nivet et al., 2016) with medium size follicles more likely to yield oocytes that develop into better quality embryos. Associations between the size of the largest follicle in the cohort and quality of oocytes have had mixed results with larger lead follicles having better cleavage and implantation and less fragmentation (Miller et al., 1996) or no significant difference in fertilization or cleavage (Rosen et al., 2008). It remains unclear whether follicle size is related to the ploidy status of the oocyte within or may predict its subsequent developmental competence and embryonic ploidy. Two prior studies have found that euploidy rates for donor blastocysts are significantly different for different centers (Munné et al., 2017) and for different physicians within the same center (McCulloh et al., 2019). One possible difference between different centers and between different physicians within the same center that we wished to examine in this study is whether the diameter of follicles selected for retrieval may affect the ploidy of blastocysts that arise from oocytes retrieved from those follicles. The purpose of this study was to examine follicle sizes and the oocytes from those follicles (and the embryos that result from those oocytes) to see if there is an association between follicle size and the quality of oocytes as judged by ability to achieve the blastocyst stage, blastocyst grades and blastocyst ploidy. Although more than one euploid embryo can result from controlled ovarian hyperstimulation (COH), the superiority of larger follicles has remained a thought pattern among practitioners of assisted reproductive technologies. Whether larger follicles are more likely to contain oocytes that will not undergo flawed meiosis with flawed chromatid and chromosome segregations is not clear. Further, it is unclear whether the oocyte from a dominant follicle, once fertilized, will become an embryo that is more competent in its ability to undergo mitotic segregation of chromosomes. To date, there has been no study relating follicle size to the incidence of aneuploidy in embryos. The aneuploidies detected in blastocysts may be attributed to meiotic or mitotic errors (Rabinowitz et al., 2012; Fragouli et al., 2013) or to aneuploidy in the fertilizing sperm (Levron et al., 2000; Tempest, 2011; McWilliams et al., 2015). However, the availability of preimplantation genetic testing of ploidy makes it possible to determine whether there is a relationship between the size of a Graafian follicle and the ploidy of the oocyte/embryo that comes from that follicle, specifically its ability to progress from the oocyte stage and become a euploid blastocyst. We have tracked individual oocytes and the embryos created using these oocytes from follicles of known diameter and related their outcomes to the size of their follicles. Our approach is unique in that it examines the distribution of follicle sizes rather than relying on analyses of mean follicle sizes or follicles grouped into arbitrary bins (whereby the selection of bin limits can create an artificial bias in the analysis). Materials and Methods The subjects of this study were 22 anonymous oocyte donors between the ages of 21 and 34 years undergoing COH and oocyte retrieval for the purpose of donating oocytes to unknown recipients at our center. Donors underwent informed consent indicating that they were aware that the oocyte retrieval would involve collection of information about follicle sizes and ultimate outcomes of the oocytes and that the measurement procedure would lead to a more prolonged oocyte retrieval requiring long administration of anesthesia. GnRH analogs were administered to donors to avoid an uncontrolled mid-cycle luteinizing hormone surge. GnRH antagonist (Cetrotide, Merck Serono, Germany) was administered to most of the donors. It was administered daily beginning the evening when one follicle attained a diameter of 14 mm. Exogenous gonadotropin was administered daily with doses specified by the monitoring physician. All donors received recombinant FSH (Gonal-F, Merck Serono, Germany), and some donors received highly purified human menopausal gonadotropin (hMG, Menopur, Ferring Pharmaceuticals, Switzerland). The relative proportions of FSH/hMG were determined by the physician managing the stimulation of the donor. The ratio of hMG to total gonadotropin dose ranged from 0 to 100%. Doses administered on the first 2 days of injections were twice the amount anticipated to be the dose necessary for follicular growth (McCulloh et al., 2012; Charkviani et al., 2014). Donor monitoring (serum levels of FSH, LH, estradiol, progesterone and ultrasound assessment of follicular sizes) was performed after the fourth night of gonadotropin administration and every day or two thereafter as decided by the monitoring physician. Dose adjustments were made as dictated by the donor’s response during monitoring. Two donors received a triggering dose of 10 000 IUs of human chorionic gonadotropin. Two donors received a triggering dose of 2 mg of triptorelin acetate (Decapeptyl, Ferring Pharmaceuticals, Switzerland). Fourteen donors received triggering doses of GnRH agonist (triptorelin, 2 mg) and 1500 IU of human chorionic gonadotropin. The trigger was administered in the evening on the day when 20% of the follicles greater than 12 mm had achieved diameters greater than 16 mm. The retrieval was scheduled 35 h after the trigger injection. Donors underwent follicular aspiration in the dorsal lithotomy position under conscious sedation. Follicular aspirates were obtained using a transvaginal approach with ultrasound guidance using a 17-gauge needle using suction pressure differential of 120 mm Hg. Follicles were scanned more carefully than in a typical retrieval. Each follicle was measured immediately prior to its puncture and aspiration. Measurements were made in two dimensions, and the arithmetic average of the two diameters was calculated for use. Follicles were aspirated one at a time to be certain that the aspirated oocyte was from the same follicle measured. The follicle measurement (size) was noted by the embryologist and the oocyte’s location in the culture dish was noted in association with the follicle size. In rare cases, where a follicle was not aspirated or when multiple oocytes were aspirated from one aspiration, these follicle sizes and oocytes were excluded from the analysis. In cases with large numbers of follicles, only one ovary’s follicles were measured. Oocytes were retrieved from the other ovary without follicle measurements. Oocytes from the unmeasured follicles were segregated from the oocytes that were retrieved from measured follicles and were not used in the analysis. Oocytes were cultured individually throughout their time in the embryology laboratory so that follicle sizes could be uniquely associated with each oocyte. Oocytes were inseminated by ICSI as this was deemed necessary by the semen analysis of the recipient’s partner. Oocytes were cultured in Quinn’s Advantage Fertilization medium (ref no.: ART-1020, Origio, Netherlands) until the fertilization assessment. Fertilized oocytes (those observed with one or two pronuclei (PN) between 16 and 18 h after insemination) were moved to drops of Quinn’s Advantage Cleavage (ref no.: ART-1026, Origio, Netherlands) culture medium for individual embryo culture immediately after fertilization scoring. On the third day of embryo development, embryos underwent zona perforation with a laser and then were moved to drops of Quinn’s Advantage Protein Plus Blastocyst Medium (ref no.: ART-1529, Origio, Netherlands) for extended individual embryo culture. Embryos were examined on Days 5 and 6. Blastocysts were graded according to the Gardner grading method (Gardner et al., 2000). Embryos achieving the blastocyst stage/grade were considered quality blastocysts when they had grades of 2BC or 2CB or better on Day 5 or day 6 of culture. Quality blastocysts underwent biopsy of roughly five to seven trophectoderm cells for preimplantation genetic testing of aneuploidy (PGT-A). The biopsy was incised using a minimum number of laser pulses directed at regions showing particular cell-to-cell adherence. Immediately following biopsy, each blastocyst was vitrified. Biopsy specimens in small-capped reaction tubes were sent to Reprogenetics/Cooper Genomics (New Jersey, USA or UK) for comprehensive chromosomal analysis by next generation sequencing (NGS). Upon receipt of the NGS results, data were assembled for each oocyte and follicle size pair. The data included maturity of the oocyte after removal of cumulus and corona cells in preparation for ICSI, in addition to whether the oocyte attained the following stages subsequently (fertilization with 2 PN, attainment of the blastocyst stage, the grades of its inner cell mass and its trophectoderm and its ploidy as measured by PGT-A). Data about oocyte maturity, its fertilization, blastocyst formation, blastocyst grades and the ploidy of blastocyst biopsies were evaluated with reference to the diameter of the follicle from which the oocyte was retrieved. Follicle diameters were expressed in two ways: the diameter of the follicle (mm) and as Z values. The use of Z values was employed in order to minimize differences in the mean follicle diameter and result in the homogeneity of variance among different donors. Therefore, follicles sizes assessed as Z values were independent of the donor in which they were measured. Receiver operator characteristic curves were used to evaluate whether developmental stages were associated with follicle diameters. Results Characteristics for 22 donors are summarized in Table I. These values indicate that donors were young, anticipated to respond well with adequate downregulation prior to starting COH and with good responses. Table I Demographics and response characteristics for 22 donors. Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 1Mean ± standard deviation (N) 2Oral contraceptive pills used for synchronization of donor with recipient 3Six donors had measurements made on only one of the two ovaries retrieved 4The fraction of total gonadotropin administered that was provided by human menopausal gonadotropin (hMG) Open in new tab Table I Demographics and response characteristics for 22 donors. Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 Parameter . Mean value1 . Age 24.5 ± 3.5 AMH 4.0 ± 2.0 Days of OCP2 treatment prior to start 18.9 + 2.6 Antral follicle count 24.7 ± 7.6 Follicle-stimulating hormone at downregulation (mIU/mL) 3.5 ± 6.3 Estradiol level at downregulation (pg/mL) 10.4 ± 8.6 Total gonadotropin administered (IU) 3203 ± 536  Total FSH administered (IU) 2116 ± 1040  Total hMG administered (IU) 1088 ± 800  Fraction of gonadotropin comprising hMG (F(hMG))4 0.36 ± 0.31 Estradiol on day of trigger (pg/mL) 5654 ± 2759 Days of gonadotropin (days) 10.1 ± 1.5 Follicle diameter at retrieval (mm) 17.6 ± 1.6 Largest follicle at retrieval (mm) 21.1 + 1.3 Total number of oocytes retrieved 18.1 ± 6.7 Number of Oocytes with Follicle Measurement3 15.0 ± 7.7 1Mean ± standard deviation (N) 2Oral contraceptive pills used for synchronization of donor with recipient 3Six donors had measurements made on only one of the two ovaries retrieved 4The fraction of total gonadotropin administered that was provided by human menopausal gonadotropin (hMG) Open in new tab Three hundred thirty oocytes were retrieved from the 22 donors. Two follicles yielded empty zonae pellucidae and were excluded from further analysis. Three hundred seventeen oocytes (96.0%) were retrieved from follicles with measured follicle diameters. Of the oocytes with follicle sizes, 255 (80.4%) had a polar body (MII) and 60 (18.9%) were immature: 31 (9.8%) with a visible germinal vesicle (GV stage) and 29 (9.1%) with neither a polar body nor a visible germinal vesicle (MI). Of the 255 MII oocytes, 228 (89.4%) were fertilized (with two pronuclei), 94 became blastocysts that were biopsied (36.8% of the MII oocytes or 41.2% of the two PN oocytes) and 51 were euploid blastocysts (20.0% of the MII oocytes or 54.3% of the biopsied blastocysts). Three hundred seventeen follicles from which oocytes were retrieved averaged 17.4 ± 2.9 mm in diameter (N = 317 follicles). The average and standard deviation of each of the 22 donor’s mean follicle diameters was determined. This mean follicle diameter and standard deviation averaged among donors was 17.6 ± 1.6 mm (N = 22 donors). The largest follicle diameter for each donor’s follicles averaged 21.1 + 1.3 mm (N = 22 donors). Follicle size and oocyte maturity We determined the maturity status of each oocyte subjected to cleanup prior to the performance of ICSI. The size of the follicle from which each oocyte was retrieved was known. Follicle sizes for oocytes observed with a germinal vesicle (n = 31) were 12.5 ± 1.6 mm in diameter (mean + standard deviation). Follicle sizes for oocytes observed without a germinal vesicle and without a polar body (MI oocytes; n = 29) were 15.3 ± 3.2 mm in diameter. Follicle sizes for oocytes observed with a polar body (MII oocytes; n = 255) were 18.3 ± 2.2 mm in diameter. Although there were significant differences between follicle sizes for the three different maturities, there was significant overlap of the distributions (Fig. 1A). Follicles smaller than 12 mm yielded only GV oocytes (100% GV oocytes; 4/4). Follicles between 12 and 14.6 mm yielded oocytes of all three maturities: GV oocytes (48%; 27/56), MI oocytes (29%; 16/56) and MI oocytes (23%; 13/56). Follicles larger than 14.6 mm never yielded GV oocytes. However, follicles between 14.7 and 21.8 mm yielded either MI (5%; 13/238) or MII (95%; 225/238) oocytes. Follicles larger than 21.2 mm yielded only mature MII oocytes (100% MII oocytes; 17/17). The incidence of a mature MII oocyte increased as follicle size increased, greatly influenced by the high prevalence of mature MII oocytes. Follicle size and oocyte maturity. (A) Cumulative histograms of follicle diameters (mm) for oocytes with a polar body (MII) (black), oocytes with neither a germinal vesicle nor a polar body (MI) (blue) and oocytes containing a germinal vesicle (GV) (red). (B) Cumulative histograms of follicle diameters expressed as Z values (from each cycle) for MII (black), MI (blue) and GV (red) oocytes. (C) Incidence (P) of occurrence of MII (black), MI (blue) and GV (red) oocytes according to Z value. Incidences were calculated using a rolling average of 21 adjacent Z values plotted at the median Z value. Above Z values of −0.5, roughly 95% of the retrieved oocytes were mature, MII oocytes. Below Z values of −0.5, the fraction of retrieved oocytes that were MI and GV oocytes increased. Figure 1 Open in new tabDownload slide Figure 1 Open in new tabDownload slide Follicle sizes (mm) differed significantly among donors (one-way ANOVA, P = 0.00013). An additional method of comparison was applied to determine each follicle and its oocyte in terms of its size relative to the follicles retrieved in that particular donor’s cycle. In order to perform this analysis each follicle’s size was expressed as a Z value:|$Z$| = {follicle diameter (mm) − mean diameter (mm)}/S.D. (mm)where mean diameter is the mean diameter of all follicles measured for the same donor, and S.D. is the standard deviation of all follicles measured for the same donor. Conversion to Z values results in each follicle having a Z value such that the mean follicle diameter within each donor’s cohort of follicles has a Z value of 0, and each follicle that has a diameter that is one standard deviation above or below that donor’s cohort’s mean value has a Z value of 1 or −1, respectively. This transformation resulted in Z values for follicle diameters with no significant differences among donors (one-way ANOVA; P = 1.0). Therefore, we assumed that each follicle or oocyte from that follicle was an independent variate in our analyses, since the Z values for follicle diameters were not dependent on the donors in which they were measured. Using this Z value method for expressing follicle size, the follicles from which MII oocytes were retrieved had mean Z values of 0.26 ± 0.80. The use of Z values also resulted in no significant difference among donors in the diameters of follicles from which mature (MII) oocytes were retrieved (one-way ANOVA; P = 0.68). Follicles from which MI oocytes were retrieved had mean Z values of −0.81 ± 0.90, and follicles from which GV oocytes were retrieved had mean Z values of −1.60 ± 0.54 (see Fig. 1B). Oocyte maturity varied according to the diameters of the follicles from which they were retrieved (Fig. 1C). Follicles with Z values increasingly larger than −1.25 had oocytes with increasingly higher percentages of mature oocytes with a polar body (MII). Above Z values of −0.5, the incidence of mature (MII) oocytes was maximal, and MII oocytes were retrieved from greater than 90% of the follicles. Follicles of sizes decreasingly smaller than Z values of −0.5 were increasingly more likely to yield immature oocytes with a GV. Follicles with Z values near −1.5 reached a peak probability of yielding an immature oocyte (MI). However, MI oocytes were found in follicles with a wide range of follicle sizes spanning all Z values between −1.8 to 1.7. The ability of follicle size to predict oocyte maturity was examined using receiver operator characteristic (ROC) curve analysis (Fig. 2). As the criterion diameter was decreased from 24 to 12 mm, more and more follicles had diameters larger than the criterion and were included. With the criterion at 16 mm, 90% of the follicles from which mature oocytes were retrieved were identified (90% true positives = 90% sensitivity) and only 18% of the follicles from which immature oocytes were identified (18% false positives = 82% specificity). The area under the curve (AUC) for the ROC curve using follicle sizes larger than a criterion value to predict MII oocytes was 0.87 for follicles measured in mm and was 0.86 for follicles measured using Z values. These ROC curves and their AUCs indicated that follicle size was a significant predictor of MII oocytes (Mann–Whitney U = 13 754 or 13 438; z = 9.5; P < 0.0001; Table II). Use of a Z value criterion greater than −0.32 resulted in a sensitivity of 0.78 and a specificity of 0.84. Receiver operator characteristic (ROC) curves indicating the ability of follicle diameter (in mm) to predict oocyte characteristics. Each ROC curve was created by sorting all the follicle diameters by size. All follicles above (or below) a selected criterion (threshold) diameter were examined to determine if they were true positives (the criterion correctly predicted the outcome) or false positives (the criterion incorrectly predicted the outcome). For example, in the ROC curve for GV oocytes, the point on the curve at (1 − specificity) = 0.1 and sensitivity = 1.0 indicates that of the oocytes from follicles smaller than the selected criterion diameter, 100% of the GV oocytes were identified as true positives and only 10% of the non-GV oocytes were identified as false positives. Each ROC curve examines the effect of changing the criterion throughout the range of all follicle diameters, with a different point on the plot representing the sensitivity and specificity for a different criterion diameter. ROC curves near to the diagonal reference are considered incapable of predicting the outcome whereas ROC curves that approach the upper left corner (at 0, 1) are considered excellent predictors of the outcome. The area under the ROC curve (AUC) is used as an estimate of the quality of the test ranging from 0.5 (the area under the diagonal) to 1.0 (the area of the entire plot area). Figure 2 Open in new tabDownload slide Figure 2 Open in new tabDownload slide Table II Areas under the curve for receiver operator characteristic (ROC) curves. Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 1ROC curves used simple monotonic criteria (all follicles either larger than or smaller than the criterion level were assessed for sensitivity and specificity). (Larger) or (smaller) is displayed to indicate whether diameters larger the criterion or smaller than the criterion were used to predict the outcome. 2The area under the ROC curve (AUC) is an estimate of how well diameter predicts the listed outcome. The significance of the AUC was compared to a value of 0.5 to determine statistical significance using the Mann-Whitney U Test. 3An oocyte with a polar body (MII). 4AUC was significantly greater than 0.5; Mann–Whitney U test; P < 0.0001. 5An oocyte with a germinal vesicle (GV) 6AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.01. 7Inner cell mass (ICM). 8Trophectoderm (TE) 9AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.03. Open in new tab Table II Areas under the curve for receiver operator characteristic (ROC) curves. Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 Follicle size used to predict1: . ROC curve AUC2(diam) . ROC curve AUC2(Z value) . Oocyte will be MII3 (larger) 0.874 0.854 Oocyte will be GV5 (smaller) 0.964 0.944 MII will become 2 PN (smaller) 0.53 0.54 2 PN will become quality blastocyst (larger) 0.596 0.576 Quality blastocyst will have ICM7 Grade A (smaller) 0.53 0.50 Quality blastocyst will have ICM Grade C (smaller) 0.56 0.60 Quality blastocyst will have TE8 Grade A (smaller) 0.639 0.61 Quality blastocyst will have TE Grade C (larger) 0.55 0.52 Quality blastocyst will be euploid (smaller) 0.51 0.51 MII will become a quality blastocyst (smaller) 0.53 0.53 MII will become a euploid quality blastocyst (larger) 0.55 0.53 1ROC curves used simple monotonic criteria (all follicles either larger than or smaller than the criterion level were assessed for sensitivity and specificity). (Larger) or (smaller) is displayed to indicate whether diameters larger the criterion or smaller than the criterion were used to predict the outcome. 2The area under the ROC curve (AUC) is an estimate of how well diameter predicts the listed outcome. The significance of the AUC was compared to a value of 0.5 to determine statistical significance using the Mann-Whitney U Test. 3An oocyte with a polar body (MII). 4AUC was significantly greater than 0.5; Mann–Whitney U test; P < 0.0001. 5An oocyte with a germinal vesicle (GV) 6AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.01. 7Inner cell mass (ICM). 8Trophectoderm (TE) 9AUC was significantly greater than 0.5; Mann–Whitney U test; P = 0.03. Open in new tab The AUC for ROC curves is evaluated by criteria that are generally accepted. The AUC as a percentage is evaluated like a grade on an examination. Percentages above 90% are considered ‘Grade A’ tests; between 80 and 90% are considered ‘Grade B’ tests; between 70 and 80% are considered ‘Grade C’ tests; between 60 and 70% are considered ‘Grade D’ tests; and percentages lower than 60% are considered poor tests. Therefore, the use of follicle size to predict MII oocytes is considered a Grade B test. The AUC for the ROC curve using follicles sizes smaller than a criterion value to predict GV oocytes was 0.96 for follicles measured in millimeters and was 0.94 for follicles measured using Z values. This indicates that follicle diameter is a Grade A test and able to predict very well when a GV oocyte will be retrieved from a follicle. These ROC curves and the AUC indicated that follicle size was a significant predictor of GV oocytes (Mann–Whitney U = 8511 or 8334; z = 8.4; P < 0.0001; Table II). Use of a criterion less than 14.6 mm resulted in a sensitivity of 1.0 and a specificity of 0.90. Use of a Z value criterion of less than −1.0 resulted in a sensitivity of 0.90 and a specificity of 0.88. Follicle size and fertilization The sizes of follicles from which 228 oocytes (that were fertilized with two pronuclei) came (18.1 ± 2.3 mm) were indistinguishable from the sizes of follicles from which mature MII oocytes came (18.2 ± 2.2 mm; P = 0.35). Further, the sizes of follicles from which oocytes that were fertilized originated were indistinguishable from the sizes of follicles from which the 37 oocytes that did not have two pronuclei (<2 or >2 pronuclei) originated (18.3 ± 2.3 mm; P = 0.54). Expressed as Z values, follicles from which MII oocytes were retrieved that had two pronuclei (0.19 ± 0.84) following ICSI were not significantly different from follicles with MII oocytes that did not have two pronuclei (0.35 ± 0.83; P = 0.28). The ability of follicle diameter to predict whether a MII oocyte will become a 2 PN oocyte (fertilized) was examined using ROC curve analysis (Table II). The AUC for the ROC curve examining the predictive value of follicle diameters (mm) greater than a criterion was 0.53. The AUC for the ROC curve examining the predictive value of follicle diameters’ Z values greater than a criterion was 0.54. These AUCs indicated that follicle diameter is a poor predictor of fertilization both because the AUCs for the ROC curves are too small to qualify as good tests using the ‘rule of thumb’ method and because the AUCs were not significantly different from 0.5 (Mann–Whitney U tests not significant). Follicle size and formation of blastocysts The sizes of follicles from which 94 oocytes that became quality blastocysts originated (18.6 ± 2.1 mm) was significantly different than the sizes of the follicles from which 181 oocytes originated that were fertilized but did not become quality blastocysts (17.9 ± 2.4 mm; P = 0.028). Expressed as Z values, follicles from which oocytes that became quality blastocysts arose (0.36 ± 0.73) were not significantly different than follicles from which oocytes originated that were fertilized but did not become quality blastocyts (0.16 ± 0.88; P = 0.06). The divergence of the two cumulative histograms (Fig. 3A and B) suggests that oocytes from smaller follicles (Z values less than 0–0.5) were less prevalent for biopsied blastocysts than for 2 PN embryos that did not become biopsied blastocysts. The sizes of the follicles from which oocytes that became quality blastocysts originated was significantly different than the sizes of follicles from which all oocytes that were fertilized (2 PN) originated regardless of whether or not they became quality blastocysts (18.3 ± 2.3 mm; P = 0.048). Examination of the incidence of quality blastocyst formation relative to follicle size indicated that the incidence of blastocyst formation increased as follicle size increased (R = 0.19; R2 = 0.036). This was particularly apparent for 17 2-pronuclear embryos that originated in follicles less than 15 mm in diameter. There were no blastocysts from these 2-pronuclear embryos. Formation of quality blastocysts in relation to follicular Z value. (A) Cumulative histogram of embryos containing two pronuclei that became quality blastocysts undergoing trophectoderm biopsy (TE Bx) or embryos containing two pronuclei that did not become quality blastocysts undergoing trophectoderm biopsy (2 PN no TE Bx). (B) The incidence of becoming a quality blastocyst is displayed as a rolling average of 29 adjacent oocytes (vertical axis) sorted by their follicular Z values and plotted at the median Z value (horizontal axis). Figure 3 Open in new tabDownload slide Figure 3 Open in new tabDownload slide The ability of follicle diameter to predict whether a 2 PN oocyte will become a quality blastocyst was examined using ROC curve analysis (Table II). The AUC for the ROC curve examining the predictive value of follicle diameters (mm) greater than a criterion was 0.59 (Fig. 2). The AUC for the ROC curve examining the predictive value of follicle diameters’ Z values greater than a criterion was 0.57. The general method interpreting these AUCs indicates that follicle diameter is a poor predictor of 2 PN oocytes that will become quality blastocysts. Despite this, the ROC curves and the AUC indicate that follicle size larger than a criterion is a significant predictor of 2 PN oocytes becoming quality blastocysts (Mann–Whitney U = 7386; z = 2.31; P = 0.01 for mm; or U = 7132; z = 1.81; P = 0.035 for Z values; Table II). Although a criterion of 15–17 mm for follicle diameter provided sensitivities of 1.0–0.71, respectively, they resulted in specificities of only 0.12–0.33, respectively. Z value criteria of between −1.0 and −0.5 provide sensitivities of 0.95–0.87, respectively, but result in specificities of only 0.14–0.16. The ability of follicle diameter to predict whether a mature (MII) oocyte will become a quality blastocyst was also examined using ROC curve analysis (Table II). The AUC for the ROC curve was 0.53 for both follicle diameters measure both in mm and using Z values. Neither AUC was significantly different than 0.5 (Mann–Whitney U tests: U = 566, z = 1.2, P = 0.11 for mm; U = 440, z = 0.77, P = 0.22 for Z values). These indicate that follicle diameter is a poor predictor for formation of quality blastocysts from MII oocytes. Follicle size and blastocyst grades Among quality blastocysts there was no significant difference between follicle diameters for oocytes that became blastocysts with the inner cell mass graded A (18.6 ± 2.1 mm; N = 50), B (17.1 ± 5.5 mm; N = 35) or C (18.4 ± 1.7 mm; N = 12) (ANOVA; F = 1.6; P > 0.05). Expressed as Z values there was no significant difference between follicle diameters for oocytes that became blastocysts with the inner cell mass graded A (0.37 ± 0.65), B (0.35 ± 0.68) or C (0.14 ± 0.98) (ANOVA; F = 0.51; P > 0.05). ROC curves and the AUCs indicated that follicle size is not a significant predictor of ICM grades (Mann–Whitney U = 65.8 or 45.9; not significant) (Table II). For quality blastocysts, there was no significant difference between follicle diameters for oocytes that became blastocysts with trophectoderm grades A (17.9 ± 2.2 mm; N = 22), B (17.8 ± 4.5 mm; N = 58) or C (19.0 ± 1.9 mm; N = 17) (ANOVA; F = 0.70; P > 0.05). Expressed as Z values there was no significant difference between follicle diameters for oocytes that became blastocysts with the trophectoderm grades A (0.19 ± 0.60), B (0.41 ± 0.71) or C (0.28 ± 0.81) (ANOVA; F = 1.41; P > 0.05). ROC curves and the AUC provided conflicting results. Follicle size in mm smaller than a criterion value was a significant predictor of blastocysts with a trophectoderm grade of A (Mann–Whitney U = 212.3; z = 1.87; P = 0.031) whereas follicle size in Z values was not a significant predictor (Mann–Whitney U = 179.6; z = 1.58; not significant) (Table II). Follicle size and blastocyst euploidy The incidence of euploidy for quality blastocysts was 54.3% (51/94). The sizes of follicles from which 51 oocytes that became euploid blastocysts were retrieved (18.7 ± 2.1 mm) was indistinguishable from the sizes of follicles from which 30 oocytes that became aneuploid blastocysts came (18.7 ± 2.2 mm; P = 0.88). Expressed as Z values, follicles from which euploid blastocysts came (0.39 ± 0.70) were indistinguishable from follicles from which aneuploid blastocysts came (0.29 ± 0.75; P = 0.58). The ROC curves and the AUCs indicated that follicle size was not a significant predictor of euploid blastocysts (among diagnosed blastocysts) (Mann–Whitney U = 15.3; z = 0.15; not significant) (Fig. 2; Table II). Further, the ROC curves examining the diameters for follicles from which mature MII oocytes were retrieved were examined to determine if diameter could predict which of these oocytes would become quality blastocysts with a euploid biopsy. The AUCs for the ROC curves were 0.55 and 0.53 for follicle diameters measured in mm and using Z values, respectively. Neither AUC was significantly different than 0.5 (Mann–Whitney U tests: U = 566, z = 1.2, P = 0.11 for mm; U = 340, z = 0.72, P = 0.24 for Z values). Both AUCs are indicative of poor ability of follicle diameter to predict which MII oocytes would become quality euploid blastocysts due to the low values of the AUC and their lack of statistical significance. We considered whether a selection bias existed for follicles from our donors. There was no significant association (as measured by correlation coefficient) between each donor’s incidence of euploidy and their mean follicle diameter (mm), the median follicle diameter (mm), the number of follicles, the standard deviation of follicle diameters (mm) or a measure of skewness (follicle diameter median − follicle diameter mean). Discussion Maturity Follicle sizes are significantly different between GV, MI and MII oocytes. The significant AUCs of ROC curves for MIIs (~0.86) and especially for GV oocytes (~0.95) indicate that follicle size is a strong predictor of maturity. In contrast to many of the prior studies, the use of ROC curves in this work investigates a wide spectrum of thresholds or criteria for predicting maturity using follicle diameter (or follicle Z value). Supporting a long held notion, oocytes from follicles smaller than 12 mm were more likely to be GV oocytes and oocytes from follicles larger than 17 mm were more likely to be mature MII oocytes (Wittmaack et al., 1994; Rosen et al., 2008; Kahraman et al., 2017; Wirleitner et al., 2018). However, the large degree of overlap seen in this study between follicles sizes of GV and MI oocytes and between follicle sizes of MI and MII oocytes indicates that follicle size is not an absolute predictor of oocyte maturity. Follicle Z values less than −1.7 were associated with a higher incidence of GV oocytes and follicle Z values greater than −0.9 were associated with a higher incidence of MII oocytes. Although the plots using Z values were much smoother, both the plots of follicle diameter and Z values showed significant overlap of follicles size measures for follicles from which oocytes of different maturities were retrieved. The incidences of oocytes of the three maturity levels in the different regions of follicles sizes supports the notion that follicle size does not absolutely predict the maturity of the oocyte it yields. The ability of smaller follicles to predict the presence of a GV (AUC ~ 0.95) is excellent. The ability of larger follicles to predict the presence of a polar body (MII oocyte; AUC ~0.86) is quite good. In both cases, this predictive value applies for follicle diameters measured following an exogenous ovulatory trigger (hCG and/or GnRH agonist). The higher AUC value for GV oocytes as compared with MII oocytes suggests that follicle size is a better predictor of GV breakdown (the transition for GV oocyte to MI oocyte) than of polar body extrusion (the transition from MI oocyte to MII oocyte) after administration of the trigger. Fertilization Low values for the AUC (0.53, 0.54) for the ROC curves comparing the appearance of 2 PN in MII oocytes following ICSI indicate that follicle size is not predictive of fertilizability for MII oocytes. This is supported by the observation of no significant difference in the sizes of follicles from which oocytes that became two pronuclear embryos versus MII oocytes that did not become two pronuclear embryos. Without an association, there is no reason to believe that follicle size is predictive of fertilizability for an MII oocyte that came from the follicle. Follicular size is reported to be predictive of oocyte fertilization (Quigley et al., 1982; Rosen et al., 2008) particularly when standard insemination was performed on oocytes of unknown maturity. In this study, maturity was known prior to ICSI and follicle size was not related to fertilization, similar to previous findings with ICSI (Rosen et al., 2008; Kahraman et al., 2017; Wirleitner et al., 2018). Blastocyst formation The incidence of blastocyst formation (quality blastocysts arising from 2 PN oocytes) increased with increasing follicle diameter. In contrast, the incidence of blastocyst formation (quality blastocysts arising from MII oocytes) was not related to follicle diameter. However, the wide diversity in blastocyst formation for two pronuclear embryos and the low value of the correlation coefficient (R2 = 0.036) that arose from follicles of different sizes suggests that despite the presence of a significant association, follicle size alone is unlikely to be the determinant of blastocyst formation. Significant differences in blastocyst formation have been reported (Kahraman et al., 2017; Wirleitner et al., 2018) when considering the formation of blastocysts from oocytes without regard to maturity or fertilization. In this study, the significant AUC values (~0.58) for the ROC curves of follicle size’s (both mm and Z value) ability to predict that a 2 PN oocyte will become a quality blastocyst indicate that follicle size is a valid predictor of 2 PN oocyte’s competence to become a quality blastocyst. This observation supports the significant differences seen according to follicle size in blastocyst formation per 2 PN oocyte (Kahraman et al., 2017). However, the low AUC values for ROC curves indicate that a significant trade-off of specificity for sensitivity is necessary when selecting a threshold follicle diameter to predict blastocyst formation for oocytes with 2 PN. The poor quality of follicle diameter as a test to predict blastocyst formation is consistent with lack of significance seen by others per MII oocyte (Wirleitner et al., 2018). One anecdotal observation from ROC curves was that blastocyst formation occurred preferentially for 2 PN oocytes arising from larger follicles; whereas, blastocyst formation occurred preferentially for MII oocytes arising from smaller follicles (although this preference was not significant). Similarly, fertilization of MII oocytes (the incremental transition from MII oocyte to 2 PN oocyte) occurred preferentially for oocytes arising from smaller follicles (although, this preference, too, was not significant). Is it possible that different incremental transitions of oocytes/embryos may occur preferentially for smaller or for larger follicles? If this is possible, then incremental preferences could cancel for longer periods of development spanning multiple incremental steps leading to an apparent lack dependence on follicle size. Blastocyst grades The sizes of follicles from which oocytes that became quality blastocysts originated were not significantly different in relation to grades of either the inner cell mass or trophectoderm. Despite this, the ROC curve indicated that better trophectoderm scores occurred in association with smaller follicle sizes (when measured in mm). In contrast, measurement of follicle size in Z values was not predictive. These conflicting results compare with the conflicting observations that larger follicles led to a higher incidence of top quality and good-quality blastocysts (Kahraman et al., 2017) and that there is no significant association between follicle size and blastocyst grades (Wirleitner et al., 2018). Euploidy The follicle sizes and distribution of follicle sizes from which oocytes that became euploid blastocysts were indistinguishable from the follicle sizes and distribution of follicles from which oocytes that became aneuploid blastocysts. The AUCs of the ROC curves (~0.51) indicate that follicle size is not predictive of euploidy for quality blastocysts. AUC values were low (0.53) for ROC curves examining the ability of diameters of follicles from which MII oocytes were retrieved to predict which oocytes would become quality blastocysts. This low value indicates that follicle size is unable to predict which oocytes will become quality euploid blastocysts. The observation that no significant correlations were found between donors’ euploidy rates and donor-specific measures of follicle size, number, dispersion and skewness support the notion that no particular bias of follicle size distribution confounded our conclusions. Maturity was significantly associated with follicle diameter. However, the subsequent steps: fertilization (by ICSI), formation of quality blastocysts, blastocyst grades and blastocyst euploidy each occurred in oocytes/embryos across a wide range of follicle diameters and with no apparent preference for smaller or larger follicles. From this, we conclude that there is no preferential follicle size to achieve euploid blastocysts. Euploid blastocysts may arise following fertilization of oocytes from follicles regardless of size as long as the follicle yields a mature oocyte (any follicle diameter ≥12 mm). Association between follicle size and oocyte quality? There is little evidence that follicle size is predictive of the quality of the oocyte within. However, there is a suggestion that oocytes from smaller follicles are less likely to be mature (more likely to be GV oocytes) and are less likely to become quality blastocysts (more likely not to be quality blastocysts despite having been fertilized). This knowledge may be helpful in estimating the number of mature oocytes that will be retrieved especially for programs performing oocyte cryopreservation or oocyte banking. In the hypothesis of follicular dominance for natural menstrual cycles (Speroff and Fritz, 2005; Strauss and Barbieri, 2014), the dominant follicle is generally believed to be the most developed (largest?) follicle with the most receptors for follicle stimulating hormone. The follicle with the most follicle-stimulating hormone receptors is capable of continuing to grow as follicle stimulating hormone levels decline from Day 3 onward. During COH with exogenous application of follicle-stimulating hormone, follicle size hierarchy seems relatively unchanged, with the largest follicle continuing as the largest follicle and smaller follicles continuing to grow albeit as smaller follicles. The largest follicle at trigger is typically a follicle that was the largest follicle or among the largest follicles throughout COH. The observation that larger follicles are not more likely to yield oocytes that develop into euploid blastocysts suggests that the follicular dominance is unrelated to the ploidy of the oocyte within. Selection of follicles for retrieval by size is unlikely to result in a higher or lower incidence of euploidy for blastocysts resulting from the retrieved oocytes. Limitations The study was limited to oocyte donors so that it could address differences in euploidy between centers (Munné et al., 2017) and between physicians within the same center (McCulloh et al., 2019). It is not clear whether these observations can be generalized to non-donor patients who may have similar or lower incidences of euploidy. Follicle size was determined by direct ultrasound measurement during oocyte retrieval. We chose to perform ultrasound measurements rather than determination of follicular aspirate volumes because we wished to decrease the time required to determine the follicular volume, during which oocytes might be exposed to less controlled conditions. Others have demonstrated that direct measurements are well correlated (r2 = 0.79; r = 0.89) with measurements of follicular volume, indicative that ultrasound measurements are comparable to measurements of follicle volume (Wittmaack et al., 1994) with 79% of the variability in follicle diameter attributed to variability of follicle volume and only 21% of the variability remaining as unexplained (residual) variability. Therefore, we believe that the follicle diameter measurements are sufficient to estimate follicle volume with reasonable errors. We must temper our interpretations based on the direct observations made in our work. Follicle diameters were measured, and for the retrieved oocytes, maturity and developmental measures of embryos resulting from their fertilization were determined. Among the developmental measures was the ploidy of biopsies sampled from embryos that became blastocysts. We directly assessed neither the ploidy of oocytes nor the ploidy of embryos that were not quality blastocysts. It is tempting to speculate that the ploidy of blastocysts is attributable to meiotic errors in oocytes from which the blastocysts arose since neither fertilization, nor blastocyst formation, nor blastocyst grades, nor ploidy of blastocyst resulting from mature oocytes were associated with the size of the follicles from which the oocytes were retrieved. However, we must acknowledge that aneuploidy, as diagnosed from several biopsied trophectoderm cells, could arise from sperm. The relatively low incidence of aneuploidies originating in the sperm (McWilliams et al., 2015) can be a significant contributor to aneuploidy, especially when using donor oocytes (noted for their low incidence of aneuploidy) and when performing ICSI using sperm from men with elevated sperm aneuploidy in association with severe male factor (Levron et al., 2000; Tempest, 2011). Aneuploidy detected in trophectoderm biopsies analyzed by NGS can arise from mitotic errors (mosaicism) (Fragouli et al., 2013). The number of euploid embryos on Day 3 is indistinguishable from the number of euploid blastocysts on Days 5 and 6 (Adler et al., 2014; Demko et al., 2016), suggesting that extended culture of embryos to Days 5 and 6 preferentially selects euploid embryos. However, the mechanisms leading to similar numbers of euploid embryos may be more complex (Rabinowitz et al., 2012; Fragouli et al., 2013). Therefore, concluding that the observed ploidy of trophectoderm biopsies was directly attributable to the ploidy of oocytes retrieved remains disputable. Nevertheless, we have found no evidence in our study to support the notion that differences in the sizes of follicles retrieved could explain the differences in euploidy for different centers (Munné et al., 2017) or for different physicians within the same center (McCulloh et al., 2019). The number of larger follicles in a cycle as examined by Wittmack et al (1994) has not been evaluated in this study. In our study, the criterion for triggering was applied uniformly for all donors. This uniformity is supported by our observation that diameters of the largest follicles varied less than diameters of the mean follicles (Table I). Further investigation with more donors will be needed to determine if the diameters of the lead follicles are associated with the ploidy for the donor’s cohort of eggs retrieved or the embryos/blastocysts arising from those oocytes. Follicle size was analyzed by two methods: direct measurement of diameter and by Z values that expressed the follicle size relative to the mean follicle size of that donor’s cycle. These two methods generally led to concordant results, indicating that either method is useful for prediction. ROC curves are relatively independent of the distribution of data and most transformations of data (Kumar and Indrayan, 2011), since the data are dichotomized into true positives (sensitivity) and false positives (1-specificity) using a criterion value. However, in predicting trophectoderm scores, direct measurement led to a significant ability to predict whereas the Z values did not. It is unclear whether this discrepancy of results is due to a lack of independence that may be present in diameters (mm) but that is adjusted through the use of Z values. We believe that the use of Z values is the preferred method of analysis since Z values correct for the follicle sizes’ dependence on the donor from which oocytes were retrieved. Neither method should be considered a good predictor (AUCs 0.55 for mm and 0.52 for Z values). Therefore, we consider this a concordant observation that follicle size is a poor predictor of trophectoderm score. These data suggest that any follicle size is capable of yielding a euploid oocyte at least using the stimulation methodology in this study. It is unclear if pushing follicles to larger sizes or harvesting oocytes when follicles are smaller will influence the incidence of euploid oocytes or will affect the size of follicles from which euploid oocytes arise since only one stimulation paradigm was employed in this study. It must be noted that follicle sizes on the day of oocyte retrieval are not clearly predicted by the size of follicles on the day of triggering. Follicles may increase in diameter or decrease in diameter, and it is even further unclear as to whether some of the triggered follicles in an individual patient may increase in diameter whereas others decrease in diameter between triggering and retrieval. Summary The sizes of follicles from which oocytes were retrieved were examined. Mature oocytes are more likely to arise from follicles larger than 16 mm or with Z values larger than −0.32 with quite good sensitivity and specificity. GV oocytes are more likely to arise for follicles smaller than 14.6 mm or with Z values smaller than −1.0 with excellent sensitivity and specificity. The progression of fertilized oocytes to become quality blastocysts was more likely for larger follicles (using a similar threshold size), but follicle size is not a strong predictor of blastocyst formation. For either MII oocytes or 2 PN oocytes achieving a quality blastocyst stage there was no indication that euploidy or blastocyst grades were strongly associated with size of the follicles from which the oocytes arose that developed into these euploid blastocysts. Therefore, we feel confident to state that physicians may retrieve smaller follicles with confidence that if they are mature and fertilized, they will become euploid blastocysts with the same incidence as mature oocytes from larger follicles. Authors’ roles D.H.M. participated in conceiving the project, analysis of the data and writing of the manuscript. N.K., T.C., T.Z. and T.B. participated in collection and assembly of the data as well as in writing of the manuscript. S.M. and L.C. participated in conceiving the project as well as in writing the manuscript. Funding ReproART: Georgian American Center for Reproductive Medicine. Conflict of interest None of the authors declare any actual conflicts of interest. However, D.H.M. declares that he receives compensation from ReproART, Biogenetics Corporation and the Sperm and Embryo Bank of New York. In addition, he received honoraria and travel funding from Ferring Pharmaceuticals and from Granata Bio during the period of the study. S.M. declares that he received compensation from Cooper Genomics during the period of the study. In addition, he received an honorarium and travel funding from Ferring Pharmaceuticals during the period of the study. L.C. declares that she is the founder of LTD Ovamedi, the organization that represents Cooper Genomics in Georgia, and received travel funding from the European Society for Human Reproduction and Embryology during the period of the study. References Adler A , Lee H-L, McCulloh DH, Ampeloquio E, Clarke-Williams M, Hodes-Wertz B, Grifo J. Blastocyst culture selects for euploid embryos: comparison of blastomere and trophectoderm biopsies . Reprod Biomed Online 2014 ; 28 : 485 – 491 . Google Scholar Crossref Search ADS PubMed WorldCat Charkviani T , Chkonia L, Kurashvili N, Kutchukhidze N, Zhorzholadze T, McCulloh D. Administration of the first 2 doses of gonadotropin at twice the dose during controlled ovarian hyperstimulation decreases total gonadotropin administration during in vitro fertilisation(IVF) cycle . Fertil Steril 2014 ; 102 : e225 . Google Scholar Crossref Search ADS WorldCat Demko AP , Simon AL, McCoy RC, Petrov DA, Rabinowitz M. Effects of maternal age on euploidy rates in a large cohort of embryos analyzed with 24-chromosome single-nucleotide polymorphism-based preimplantation genetic screening . Fertil Steril 2016 ; 105 : 1307 – 1313 . Google Scholar Crossref Search ADS PubMed WorldCat Fragouli E , Alfarawati S, Spath K, Jaroudi S, Sarasa J, Enciso M, Wells D. The origin and impact of embryonic aneuploidy . Hum Genet 2013 ; 132 : 1001 – 1013 . Google Scholar Crossref Search ADS PubMed WorldCat Gardner DK , Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: towards a single blastocyst transfer . Fertil Steril 2000 ; 73 : 1155 – 1158 . Google Scholar Crossref Search ADS PubMed WorldCat Kahraman S , Pirkevi Cetinkaya C, Cetinkaya M, Yelke H, Kumtepe Colakoglu Y, Aygun M, Montag M. The effect of follicle size and homogeneity of follicular development on the morphokinetics of human embryos . J Assist Reprod Genet 2017 ; 34 : 895 – 903 . Google Scholar Crossref Search ADS PubMed WorldCat Kumar R , Indrayan A. Receiver operator (ROC) characteristic curve for medical researchers . Ind Pediat 2011 ; 48 : 277 – 287 . Google Scholar Crossref Search ADS WorldCat Levron J , Aviram-Goldring A, Mdgar I, Wiessenberg R, Bachar A, Dor J. The prevalence of sperm chromosome aneuploidies in severe male factor infertility patients before and after sperm cell sorting under light microscopy . Fertil Steril 2000 ; 74 : S108 . Google Scholar Crossref Search ADS WorldCat McCulloh DH , Alikani M, Norian J, Kolb B, Arbones JM, Munné S. Controlled ovarian hyperstimulation (COH) parameters associated with euploidy rates in donor oocytes . Eur J Med Genet 2019 ; 62 : Article 103681 . OpenURL Placeholder Text WorldCat McCulloh DH , Colon JM, McGovern PG. Modeling follicle stimulating hormone levels in serum for controlled ovarian hyperstimulation III: improved gonadotropin administration . Curr Pharm Biotechnol 2012 ; 13 : 454 – 463 . Google Scholar Crossref Search ADS PubMed WorldCat McWilliams K , McWilliams TK, Wyatt M, Hughes M. Aneuploidy parent of origin in blastocyst biopsies using karyomapping technology . Fertil Steril 2015 ; 104 :e283. OpenURL Placeholder Text WorldCat Miller KF , Goldberg JM, Falcone T. Follicle size and implantation of embryos from in vitro fertilization . Obst Gynec 1996 ; 88 : 583 – 586 . Google Scholar Crossref Search ADS WorldCat Munné S , Alikani M, Ribustello L, Cols P, Martinez-Ortiz P, Referring Physician Group , McCulloh D. Euploidy rates in donor egg cycles significantly differ between fertility centers . Hum Reprod 2017 ; 32 : 743 – 749 . Google Scholar Crossref Search ADS PubMed WorldCat Nivet AL , Léveillé MC, Leader A, Sirard MA. Transcriptional characteristics of different sized follicles in relation to embryo transferability: potential role of hepatocyte growth factor signaling . Molec Hum Reprod 2016 ; 22 : 475 – 484 . Google Scholar Crossref Search ADS PubMed WorldCat Quigley MM , Wolf DP, Maklad NF, Dandekar PV, Sokoloski JE. Follicular size and number in human in vitro fertilization . Fertil Steril 1982 ; 38 : 678 – 681 . Google Scholar Crossref Search ADS PubMed WorldCat Rabinowitz M , Ryan A, Gemelos G, Hill M, Baner J, Cinnioglu C, Ganjevic M, Potter D, Petrov DA, Demko Z. Origins and rates of aneuploidy in human blastocysts . Fertil Steril 2012 ; 97 : 395 – 401 . Google Scholar Crossref Search ADS PubMed WorldCat Rosen MP , Shen S, Dobson AT, Rinaudo PF, McCulloch CE, Cedars MI. A quantitation assessment of follicle size on oocyte developmental competence . Fertil Steril 2008 ; 90 0015-0282/08/$34.00 . doi: 10.1016/j.fertnstert.2007.02.011 . OpenURL Placeholder Text WorldCat Crossref Speroff L , Fritz MA. Clinical Gynecologic Endocrinology and Infertility , 7th edn. Philadelphia : Lippincott Williams and Wilkins , 2005 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Strauss JF , Barbieri RL. Yen & Jaffe’s Reproductive Endocrinology E-Book : Physiology, Pathophysiology, and Clinical Management , 7th edn. Philadelphia : Elsevier Saunders , 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Tempest HG . Meiotic recombination errors, the origin of sperm aneuploidy and clinical recommendations . Syst Biol Reprod Med 2011 ; 57 : 93 – 101 . Google Scholar Crossref Search ADS PubMed WorldCat Wirleitner B , Okhowat J, Vistejnova L, Kralickova M, Karlikova M, Vanderzwalmen P, Ectors F, Hradecky L, Schuff M, Murtinger M. Relationship between follicular volume and oocyte competence, blastocyst development and live-bith rate: optimal follicle size for oocyte retrieval . Ultrasound Obstet Gynecol 2018 ; 51 : 118 – 123 . Google Scholar Crossref Search ADS PubMed WorldCat Wittmaack FM , Kreger DO, Blasco L, Tureck RW, Mastroianni L, Lessey BA. Effect of follicular size on oocyte retrieval, fertilization, cleavage, and embryo quality in in vitro fertilization cycles: a 6-year data collection . Fertil Steril 1994 ; 62 : 1205 – 1210 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author(s) 2020. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For permissions, please e-mail: [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Human Reproduction – Oxford University Press

Published: Mar 27, 2020

There are no references for this article.