Investigative Radiology

Objectives The aim of the study is to develop and evaluate the performance of a deep learning (DL) model to triage breast magnetic resonance imaging (MRI) findings in high-risk patients without missing any cancers. Materials and Methods In this retrospective study, 16,535 consecutive contrast-enhanced MRIs performed in 8354 women from January 2013 to January 2019 were collected. From 3 New York imaging sites, 14,768 MRIs were used for the training and validation data set, and 80 randomly selected MRIs were used for a reader study test data set. From 3 New Jersey imaging sites, 1687 MRIs (1441 screening MRIs and 246 MRIs performed in recently diagnosed breast cancer patients) were used for an external validation data set. The DL model was trained to classify maximum intensity projection images as “extremely low suspicion” or “possibly suspicious.” Deep learning model evaluation (workload reduction, sensitivity, specificity) was performed on the external validation data set, using a histopathology reference standard. A reader study was performed to compare DL model performance to fellowship-trained breast imaging radiologists. Results In the external validation data set, the DL model triaged 159/1441 of screening MRIs as “extremely low suspicion” without missing a single cancer, yielding a workload reduction of 11%, a specificity of 11.5%, and a sensitivity of 100%. The model correctly triaged 246/246 (100% sensitivity) of MRIs in recently diagnosed patients as “possibly suspicious.” In the reader study, 2 readers classified MRIs with a specificity of 93.62% and 91.49%, respectively, and missed 0 and 1 cancer, respectively. On the other hand, the DL model classified MRIs with a specificity of 19.15% and missed 0 cancers, highlighting its potential use not as an independent reader but as a triage tool. Conclusions Our automated DL model triages a subset of screening breast MRIs as “extremely low suspicion” without misclassifying any cancer cases. This tool may be used to reduce workload in standalone mode, to shunt low suspicion cases to designated radiologists or to the end of the workday, or to serve as base model for other downstream AI tools.

Introduction Prostate cancer diffusion weighted imaging (DWI) MRI is typically performed at high-field strength (3.0 T) in order to overcome low signal-to-noise ratio (SNR). In this study, we demonstrate the feasibility of prostate DWI at low field enabled by random matrix theory (RMT)-based denoising, relying on the MP-PCA algorithm applied during image reconstruction from multiple coils. Methods Twenty-one volunteers and 2 prostate cancer patients were imaged with a 6-channel pelvic surface array coil and an 18-channel spine array on a prototype 0.55 T system created by ramping down a commercial magnetic resonance imaging system (1.5 T MAGNETOM Aera Siemens Healthcare) with 45 mT/m gradients and 200 T/m/s slew rate. Diffusion-weighted images were acquired with 4 non-collinear directions, for which b = 50 s/mm2 was used with 8 averages and b = 1000 s/mm2 with 40 averages; 2 extra b = 50 s/mm2 were used as part of the dynamic field correction. Standard and RMT-based reconstructions were applied on DWI over different ranges of averages. Accuracy/precision was evaluated using the apparent diffusion coefficient (ADC), and image quality was evaluated over 5 separate reconstructions by 3 radiologists with a 5-point Likert scale. For the 2 patients, we compare image quality and lesion visibility of the RMT reconstruction versus the standard one on 0.55 T and on clinical 3.0 T. Results The RMT-based reconstruction in this study reduces the noise floor by a factor of 5.8, thereby alleviating the bias on prostate ADC. Moreover, the precision of the ADC in prostate tissue after RMT increases over a range of 30%–130%, with the increase in both signal-to-noise ratio and precision being more prominent for a low number of averages. Raters found that the images were consistently of moderate to good overall quality (3–4 on the Likert scale). Moreover, they determined that b = 1000 s/mm2 images from a 1:55-minute scan with the RMT-based reconstruction were on par with the corresponding images from a 14:20-minute scan with standard reconstruction. Prostate cancer was visible on ADC and calculated b = 1500 images even with the abbreviated 1:55-minute scan reconstructed with RMT. Conclusions Prostate imaging using DWI is feasible at low field and can be performed more rapidly with noninferior image quality compared with standard reconstruction.

doi: 10.1097/rli.0000000000000982pmid: 37185253

Objectives Detailed visualization of the arterial runoff is mandatory for the assessment of peripheral arterial occlusive disease. This study aims to compare the performance of a first-generation photon-counting detector computed tomography (PCD-CT) to a third-generation energy-integrating detector CT (EID-CT). Materials and Methods Computed tomography angiographies of 8 upper leg arterial runoffs were performed on human cadaveric models with continuous extracorporeal perfusion. For both PCD-CT and EID-CT, radiation dose-equivalent 120 kVp acquisition protocols (low-/medium-/high-dose: CTDIVol = 3/5/10 mGy) were used. All scans were performed with standard collimation (PCD-CT: 144 × 0.4 mm; EID-CT: 96 × 0.6 mm), a pitch factor of 0.4, and a gantry rotation time of 1.0 second. Reformatting of data included the use of comparable vascular kernels (Bv 48/49), a slice thickness and increment of 1.0 mm, and a field of view of 150 × 150 mm. Eight radiologists evaluated image quality independently using a browser-based pairwise forced-choice comparison setup. Kendall concordance coefficient (W) was calculated to estimate interrater agreement. Signal-to-noise ratio and contrast-to-noise ratio (CNR) were compared based on 1-way analyses of variance and linear regression analysis. Results Low-dose PCD-CT achieved superior signal-to-noise ratio/CNR values compared with high-dose EID-CT (P < 0.001). Linear regression analysis suggested that an EID-CT scan with a CTDIVol of at least 15.5 mGy was required to match the CNR value of low-dose PCD-CT. Intraluminal contrast attenuation was higher in PCD-CT than EID-CT, irrespective of dose level (415.0 ± 31.9 HU vs 329.2 ± 29.4 HU; P < 0.001). Subjective image quality of low-dose PCD-CT was considered superior to high-dose EID-CT (P < 0.001). Interrater agreement was high (W = 0.989). Conclusions Using cadaveric models with continuous extracorporeal perfusion allows for intraindividual image quality comparisons between PCD-CT and EID-CT on variable dose levels. With superior luminal contrast attenuation and denoising in angiographies of the peripheral arterial runoff, PCD-CT displayed potential for radiation saving of up to 83% compared with EID-CT.

Abstract Brain and cardiac MRIs are fundamental noninvasive imaging tools, which can provide important clinical information and can be performed without or with gadolinium-based contrast agents (GBCAs), depending on the clinical indication. It is currently a topic of debate whether it would be feasible to extract information such as standard gadolinium-enhanced MRI while injecting either less or no GBCAs. Artificial intelligence (AI) is a great source of innovation in medical imaging and has been explored as a method to synthesize virtual contrast MR images, potentially yielding similar diagnostic performance without the need to administer GBCAs. If possible, there would be significant benefits, including reduction of costs, acquisition time, and environmental impact with respect to conventional contrast-enhanced MRI examinations. Given its promise, we believe additional research is needed to increase the evidence to make these AI solutions feasible, reliable, and robust enough to be integrated into the clinical framework. Here, we review recent AI studies aimed at reducing or replacing gadolinium in brain and cardiac imaging while maintaining diagnostic image quality.

doi: 10.1097/rli.0000000000000986pmid: 37222527

Objectives In multiple myeloma and its precursor stages, plasma cell infiltration (PCI) and cytogenetic aberrations are important for staging, risk stratification, and response assessment. However, invasive bone marrow (BM) biopsies cannot be performed frequently and multifocally to assess the spatially heterogenous tumor tissue. Therefore, the goal of this study was to establish an automated framework to predict local BM biopsy results from magnetic resonance imaging (MRI). Materials and Methods This retrospective multicentric study used data from center 1 for algorithm training and internal testing, and data from center 2 to 8 for external testing. An nnU-Net was trained for automated segmentation of pelvic BM from T1-weighted whole-body MRI. Radiomics features were extracted from these segmentations, and random forest models were trained to predict PCI and the presence or absence of cytogenetic aberrations. Pearson correlation coefficient and the area under the receiver operating characteristic were used to evaluate the prediction performance for PCI and cytogenetic aberrations, respectively. Results A total of 672 MRIs from 512 patients (median age, 61 years; interquartile range, 53–67 years; 307 men) from 8 centers and 370 corresponding BM biopsies were included. The predicted PCI from the best model was significantly correlated (P ≤ 0.01) to the actual PCI from biopsy in all internal and external test sets (internal test set: r = 0.71 [0.51, 0.83]; center 2, high-quality test set: r = 0.45 [0.12, 0.69]; center 2, other test set: r = 0.30 [0.07, 0.49]; multicenter test set: r = 0.57 [0.30, 0.76]). The areas under the receiver operating characteristic of the prediction models for the different cytogenetic aberrations ranged from 0.57 to 0.76 for the internal test set, but no model generalized well to all 3 external test sets. Conclusions The automated image analysis framework established in this study allows for noninvasive prediction of a surrogate parameter for PCI, which is significantly correlated to the actual PCI from BM biopsy.