Radiology advances最新文献

{"title":"Deep learning pipeline for fully automated myocardial infarct segmentation from clinical cardiac MR scans.","authors":"Matthias Schwab, Mathias Pamminger, Christian Kremser, Markus Haltmeier, Agnes Mayr","doi":"10.1093/radadv/umaf023","DOIUrl":"10.1093/radadv/umaf023","url":null,"abstract":"<p><strong>Background: </strong>Artificial intelligence (AI) has demonstrated promise in cardiovascular magnetic resonance (CMR) imaging, particularly in myocardial infarct segmentation, where it may help reduce variability and workload in clinical practice.</p><p><strong>Purpose: </strong>To develop and evaluate a deep learning-based model that performs myocardial infarct segmentation in a fully automated way.</p><p><strong>Materials and methods: </strong>For this retrospective study, a cascaded framework of 2- and 3-dimensional convolutional neural networks (CNNs), specialized in identifying ischemic myocardial scars on late gadolinium enhancement (LGE) CMR images, was trained on an in-house training dataset of 144 examinations acquired using a 1.5 Tesla Siemens scanner collected between 2006 and 2022. On a separate test dataset from the same institution, comprising images from 152 examinations, a quantitative comparison was conducted between AI-based segmentations and manual segmentations. Further, segmentation accuracy was assessed qualitatively for both human and AI-generated contours by 2 CMR experts in a blinded experiment. Most cases underwent single human assessment, with double reading conducted only on a subset of 20 cases.</p><p><strong>Results: </strong>Excellent agreement was found between manually and automatically calculated infarct volumes (ρ_c = 0.9). The qualitative evaluation showed that compared to human-based measurements, the experts rated the AI-based segmentations as better representing the actual extent of infarction (<i>P</i> < 0.001) and preferred them more often (33.4% AI, 25.1% human, 41.5% equal). On the contrary, for segmentation of microvascular obstruction (MVO), manual measurements were still preferred (<i>P</i> < 0.001; 11.3% AI, 55.6% human, 33.1% equal).</p><p><strong>Conclusion: </strong>This fully automated segmentation pipeline enables the calculation of CMR infarct size without requiring any pre-processing of the input images while matching the segmentation quality of trained human observers. As automated infarct segmentation is preferred over manual segmentation, further development of this workflow toward clinical application is warranted to improve efficiencies.</p><p><strong>Summary: </strong>We developed and evaluated an algorithm that performs myocardial infarct segmentation from cardiac MR images without requiring pre-processing and that outperforms trained human observers on qualitative expert judgment.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 4","pages":"umaf023"},"PeriodicalIF":0.0,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429236/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145245860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MindMap: the intricate connections of the human brain.","authors":"Sahar Ahmad, Pew-Thian Yap","doi":"10.1093/radadv/umaf006","DOIUrl":"https://doi.org/10.1093/radadv/umaf006","url":null,"abstract":"","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 4","pages":"umaf006"},"PeriodicalIF":0.0,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429220/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145246137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning-based algorithm for automatic detection of incidental pulmonary embolism on contrast-enhanced CT: a multicenter multivendor study.","authors":"Hana Farzaneh, Jacqueline Junn, Yasmina Chaibi, Angela Ayobi, Angelo Franciosini, Marlene Scudeler, Daniel Chow, Brent Weinberg","doi":"10.1093/radadv/umaf021","DOIUrl":"10.1093/radadv/umaf021","url":null,"abstract":"<p><strong>Background: </strong>Incidenal pulmonary embolism (iPE) is increasingly detected on contrast-enhanced computed tomography (CECT) performed for non-PE indications, reflecting the growing volume and complexity of cross-sectional imaging. These findings, although unexpected, carry important clinical implications and may be underreported due to the primary diagnostic focus of the study. Artificial intelligence (AI) applications offer the potential to augment radiologist workflow bt training exams and highlighting cases suspicious for iPE, thereby improving detection accuracy and timeliness in routine clinical practice.</p><p><strong>Purpose: </strong>Likelihood of incidental pulmonary embolism (iPE) increases with increased body computed tomography CT) imaging. This study evaluates the diagnostic performance and effectiveness of triage of a standalone AI solution for detecting iPE in contrast-enhanced CT (CECT) exams obtained for non-PE clinical indications.</p><p><strong>Materials and methods: </strong>A commercially available deep learning-based software, CINA-iPE (Avicenna.AI, La Ciotat, France), analyzes CECT images to highlight suspected incidental PE cases. Consecutive retrospective CECTs from 5 clinical centers, not performed for PE evaluation, were collected until a selected balanced dataset between positive and negative cases was obtained. The reference standard was established by three independent U.S. board-certified radiologists reviewing the same images. Diagnostic performance and the time-to-notification (from data acquisition to processing of results) were computed.</p><p><strong>Results: </strong>A total of 381 anonymized CECT cases were acquired on 39 different scanner models from GE, Philips, Siemens, and Canon. The algorithm correctly identified 159/181 exams positive for PE (sensitivity 87.8% [95% CI: 82.2%-92.2%]) and 184/200 exams negative for PE (specificity 92.0% [95% CI: 87.3%-95.4%]), yielding an accuracy of 90.0% [95% CI: 86.6%-92.8%]. Of 16 detected false positive cases, 50% were complex CECTs subject to disagreement among the reference read radiologists. The device missed 22 pulmonary embolisms, with 45.5% of them being complex cases and subject to disagreement among reviewers. The time from data acquisition to processing results was 1.5 ± 0.5 (mean ± SD, 95% CI: 1.4%-1.5%) minutes.</p><p><strong>Conclusion: </strong>The CINA-iPE application accurately identified incidental PE in studies not performed specifically for evaluation of PE with high sensitivity and specificity. Automatically processed results were available to interpreting physicians within minutes, which could be used to prioritize interpretation of studies. This may be useful for increasing the accuracy or speed of detection of iPE.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 4","pages":"umaf021"},"PeriodicalIF":0.0,"publicationDate":"2025-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429201/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145245867","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep-learning-accelerated T1-MPRAGE MRI for quantification and visual grading of cerebral volume in memory loss patients.","authors":"Nelson Gil, Azadeh Tabari, Dominik Nickel, Wei-Ching Lo, Bryan Clifford, Stephen Cauley, Min Lang, Sittaya Buathong, Azadeh Hajati, Shohei Fujita, Seonghwan Yee, John Conklin, Susie Huang","doi":"10.1093/radadv/umaf022","DOIUrl":"10.1093/radadv/umaf022","url":null,"abstract":"<p><strong>Purpose: </strong>To evaluate a physics-based deep-learning-accelerated super-resolution T1-weighted MPRAGE sequence (DL-MPRAGE) against standard 3-dimensional T1-weighted MPRAGE (STD-MPRAGE) for quantitative and qualitative regional cortical volume assessment.</p><p><strong>Materials and methods: </strong>This prospective single-center study included patients undergoing evaluation for memory loss on 3T MRI scanners (MAGNETOM Vida, Siemens Healthineers, Forchheim, Germany) from October 2023 to January 2024. The absolute symmetrized percent change in cortical volume and thickness was assessed on DL- and STD-MPRAGE images using the FreeSurfer brain segmentation algorithm. Bland-Altman analysis evaluated the agreement in volumetrics for each anatomical region. Additionally, 2 blinded radiologists independently qualitatively rated image quality metrics and cortical volume loss for anatomical regions based on standardized scales.</p><p><strong>Results: </strong>A total of 64 participants (29 women [45%], mean age 62 years ±16 [SD]) were evaluated. DL-MPRAGE increased spatial resolution from 1 mm to 0.5 mm while reducing scan time by more than half (2:11 vs. 5:21). Mean regional volumes for DL-MPRAGE were systematically lower than for STD-MPRAGE (eg, 17 226 ± 2011 vs. 17 923 ± 2185 mm³, corresponding to an absolute difference between the means of 697 mm³, for the cingulate gyrus, <i>P < </i>.004). Corresponding absolute symmetrized percent change values averaged 2.8% across brain regions, with the largest mean value being 5.08% for the cingulate gyrus. Bland-Altman analysis demonstrated high agreement in quantitative measurements for both volume and thickness. On reader assessment, DL-MPRAGE was noninferior to STD-MPRAGE across image quality metrics (<i>P < </i>.01) and equivalent in assessing volume loss.</p><p><strong>Conclusions: </strong>DL-MPRAGE offers quantitatively and qualitatively equivalent volumetric estimation compared to STD-MPRAGE while improving spatial resolution and acquisition speed for patients undergoing evaluation for memory loss.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 4","pages":"umaf022"},"PeriodicalIF":0.0,"publicationDate":"2025-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12255235/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144628822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Whimsy.","authors":"Andria M Powers","doi":"10.1093/radadv/umaf005","DOIUrl":"https://doi.org/10.1093/radadv/umaf005","url":null,"abstract":"","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 3","pages":"umaf005"},"PeriodicalIF":0.0,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429264/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145246333","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning diagnosis of adult tibial plateau fractures: multicenter study with external validation.","authors":"Tongtong Huo, Pengran Liu, Mingdi Xue, Jiayao Zhang, Yi Xie, Honglin Wang, Hong Zhou, Zineng Yan, Songxiang Liu, Lin Lu, Jiaming Yang, Wei Wu, Zhewei Ye","doi":"10.1093/radadv/umaf020","DOIUrl":"10.1093/radadv/umaf020","url":null,"abstract":"<p><strong>Purpose: </strong>To evaluate a MobileNetV3-YOLOv8 deep learning (DL) model for detecting tibial plateau fractures (TPFs), including occult TPFs (OTPFs), on knee radiographs. We hypothesized that the DL model would improve diagnostic performance and reduce interpretation time, particularly for less experienced physicians.</p><p><strong>Materials and methods: </strong>This retrospective, multicenter study, included 1543 adult patients from 5 tertiary hospitals in China. A total of 3547 radiographs were included: 2837 for training/validation and 710 from a single external center for testing. In the test set, 267 (37.6%) were normal, 282 (39.7%) were obvious TPFs, and 161 (22.7%) were OTPFs. Performance metrics comprised sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), F1-score, and the area under the receiver operating characteristic curve (AUROC). Eleven physicians (6 experienced, 5 inexperienced) interpreted 70 test images with and without DL assistance. Interreader agreement (Fleiss' κ) and interpretation time were evaluated.</p><p><strong>Results: </strong>For obvious TPFs, the model achieved 89.4% sensitivity (95% confidence interval [CI], 85.7-92.3), 92.5% specificity (95% CI, 89.9-95.1), 88.7% PPV, 92.9% NPV, 89.0% F1-score, and 91.9% accuracy (95% CI, 89.7-94.1). For OTPFs, it achieved 85.7% sensitivity (95% CI, 81.2-89.4), 91.3% specificity (95% CI, 88.5-93.2), 74.2% PPV, 95.6% NPV, 79.5% F1-score, and 88.2% accuracy (95% CI, 86.4-89.8). The overall AUROC was 0.949 (95% CI, 0.935-0.963). DL assistance improved OTPF sensitivity of less experienced readers (67.5% to 83.8%), increased interreader agreement (κ) (0.58 [95% CI, 0.52-0.64] to 0.71 [95% CI, 0.65-0.76] and reduced mean interpretation time (55.8 seconds to 34.3 seconds).</p><p><strong>Conclusion: </strong>The MobileNetV3-YOLOv8 model accurately detected both obvious and occult TPFs, substantially improving diagnostic sensitivity, interreader agreement, and efficiency. These findings suggest that AI assistance can enhance diagnostic performance and reduce interpretation time, offering considerable benefits for emergency departments where rapid and accurate fracture detection is paramount.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 3","pages":"umaf020"},"PeriodicalIF":0.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429206/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145246168","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intervertebral disc degeneration of the lumbar spine assessed in vivo with 3T magnetic resonance tomoelastography.","authors":"Rolf Reiter, Pinkas Mürdel, Florian N Loch, Mehrgan Shahryari, Rebecca Strehle, Christian Bayerl, Jürgen Braun, Ingolf Sack, Patrick Asbach, David Kaufmann","doi":"10.1093/radadv/umaf013","DOIUrl":"10.1093/radadv/umaf013","url":null,"abstract":"<p><strong>Introduction: </strong>Assessment of intervertebral disc (IVD) degeneration on conventional magnetic resonance imaging (MRI) is limited by large inter-reader variability and lack of stratification in clinical trials and their assessment of treatment responses. Therefore, we aimed to introduce and diagnostically validate multifrequency magnetic resonance elastography (MRE) with tomoelastography processing for the assessment of lumbar IVD degeneration in healthy volunteers and patients with low back pain.</p><p><strong>Methods: </strong>In this prospective single-center study, 60 participants (30 volunteers without low back pain and 30 patients with low back pain, 41 ± 17 years, body mass index 23.9 ± 3.7 kg/m², 25 women) underwent multifrequency MRE using vibration frequencies from 40 to 70 Hz using a custom-built MRE setup in a 3T MRI scanner (Magnetom Skyra, software version XA30, Siemens Healthineers, Erlangen, Germany). Maps of shear wave speed (SWS in m/s) and loss angle (<i>φ</i> in rad), representing stiffness and viscous properties, respectively, were generated using tomoelastography data processing. The Pfirrmann score was used as reference standard to assess lumbar IVD degeneration on sagittal T2-weighted images. Inter-reader agreement (3 readers) and repeatability were assessed using the intraclass correlation coefficient (ICC).</p><p><strong>Results: </strong>Area under the receiver operating characteristic curve (AUC) analysis showed good diagnostic performance for detecting IVD degeneration (Pfirrmann score I/II/III/IV/V with <i>n</i> = 7/18/9/18/9, respectively) based on SWS (AUC: ≥II: 0.83, ≥III: 0.91, ≥IV: 0.96, V: 0.97) and <i>φ</i> (AUC: ≥II: 0.88, ≥III: 0.93, ≥IV: 0.98, V: 0.95). Good and excellent inter-reader agreements were found for Pfirrmann score (ICC = 0.87), SWS (ICC = 0.87), and <i>φ</i> (ICC = 0.92), respectively. Good repeatability was demonstrated for SWS (ICC = 0.88) and <i>φ</i> (ICC = 0.88).</p><p><strong>Discussion: </strong>Multifrequency MRE with tomoelastography processing allows effective IVD assessment and shows promise as a quantitative clinical imaging modality for assessing IVD degeneration.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 3","pages":"umaf013"},"PeriodicalIF":0.0,"publicationDate":"2025-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429253/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145246212","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantifying changes in steatotic liver disease after bariatric surgery using ultrasound-derived fat fraction.","authors":"Nanda Deepa Thimmappa, Ayman Gaballah, Talissa Altes, Deepthi Rao, Andrew W Wheeler, Norbert Richardson, Joanne Cassani, Jacqueline Bailey, Yassin Labyed","doi":"10.1093/radadv/umaf018","DOIUrl":"10.1093/radadv/umaf018","url":null,"abstract":"<p><strong>Background: </strong>A noninvasive, readily accessible, and quantitative methods are needed to evaluate changes in hepatic fat over time in patients undergoing bariatric surgery.</p><p><strong>Purpose: </strong>To measure the change in ultrasound-derived fat fraction (UDFF) in severely obese adults before and after bariatric surgery.</p><p><strong>Materials and methods: </strong>Participants undergoing bariatric surgery were prospectively enrolled at a single center and evaluated at 3 time points: pre-surgery (before a 2-week pre-operative liquid diet), day of surgery (immediately following the liquid diet), and 6 months post-surgery. Measurements included UDFF (Siemens Healthineers ACUSON SEQUOIA), skin-to-liver capsule thickness, serum metabolic tests, liver function tests, and anthropometric data. Changes in these variables were analyzed using 2-sided paired <i>t</i>-tests. Pearson correlation analysis was used to evaluate associations between changes in UDFF and changes in skin-to-capsule thickness, waist circumference, and BMI.</p><p><strong>Results: </strong>Seventeen participants (mean age 47.41 ± 11.11 years, 15 women) completed the study. UDFF (mean ± SD) decreased significantly from 20.4 ± 8.41% pre-surgery to 11.08 ± 6.33% at 6 months post-surgery (<i>P</i> < .001). Skin-to-liver capsule thickness decreased from 4.06 ± 0.66 cm to 2.88 ± 0.55 cm (<i>P</i> < .001). BMI decreased from 46.27 ± 6.65 kg/m² to 35.5 ± 6.64 kg/m² (<i>P</i> < .001), and waist circumference decreased from 130.29 ± 13.49 cm to 111.98 ± 16.03 cm (<i>P</i> < .001). No significant changes were observed in UDFF, skin-to-liver capsule thickness, or waist circumference after the liquid diet phase. Strong positive correlation between UDFF and BMI reduction was observed (r = 0.75, <i>P</i> < .001).</p><p><strong>Conclusions: </strong>Bariatric surgery results in significant reductions in hepatic steatosis, as measured by UDFF, and reductions in skin-to-liver capsule thickness, BMI, and waist circumference at 6 months post-surgery. These findings suggest that UDFF could meet the clinical need for noninvasive monitoring of hepatic steatosis following bariatric surgery.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 3","pages":"umaf018"},"PeriodicalIF":0.0,"publicationDate":"2025-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429188/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145246255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Physics-informed deep learning reconstruction for ultrafast clinical 3D fluid-attenuated inversion recovery brain MRI.","authors":"Shohei Fujita, Dominik Nickel, Wei-Ching Lo, Bryan Clifford, Stephen Cauley, Sittaya Buathong, Azadeh Hajati, Florence L Chiang, John Conklin, Susie Y Huang","doi":"10.1093/radadv/umaf016","DOIUrl":"10.1093/radadv/umaf016","url":null,"abstract":"<p><strong>Background: </strong>Physics-informed deep learning (DL) reconstructions show promise in accelerating MRI yet have not been extensively validated, particularly for 3D fluid-attenuated inversion recovery (FLAIR) sequence.</p><p><strong>Purpose: </strong>To evaluate the diagnostic quality and interchangeability of DL-based 3D FLAIR with a state-of-the-art acceleration technique (wave-controlled aliasing in parallel imaging [Wave-CAIPI] FLAIR) in a clinical setting with 3 T brain MRI.</p><p><strong>Materials and methods: </strong>Participants undergoing evaluation for demyelinating disease between October and December of 2023 were prospectively enrolled at a single center. For each participant, state-of-the-art Wave-CAIPI FLAIR and a resolution-matched 6-fold-under-sampled Cartesian FLAIR acquisition with DL reconstruction were performed at 3-T system (MAGNETOM Vida, Siemens Healthineers, Erlangen, Germany). Four neuroradiologists evaluated overall image quality, anatomic conspicuity, lesion conspicuity, and imaging artifacts. Lesion count, volume, and regional brain volume were compared between imaging methods. Inter-reader agreement was assessed using quadratic weighted Cohen's kappa and Kendall's correlation coefficient. Agreement of continuous metrics was evaluated using intraclass correlation coefficients (ICCs), linear regression, and Bland-Altman analysis. Interchangeability regarding the quantitative metrics was evaluated with individual equivalence index (IEI).</p><p><strong>Results: </strong>Totally, 88 participants (61 women [69%], 47 ± 13 years) were evaluated. DL-FLAIR reduced scan time (1:53 vs. 2:50) and showed higher overall image quality, anatomic conspicuity, lesion conspicuity, and imaging artifacts compared with state-of-the-art technique (all <i>P</i>s < .001). DL-FLAIR also demonstrated higher signal-to-noise ratio and contrast-to-noise ratio compared to Wave-CAIPI-FLAIR, with high agreement in lesion and regional brain volumes between both methods (ICC(2, k) range, 0.91 to 0.99). DL-FLAIR proved interchangeable with Wave-CAIPI-FLAIR for lesion count (IEI: 0.10, acceptable proportion: 0.977, 95% CI: [0.943, 1.000]) and for lesion volume (IEI: 0.32, acceptable proportion: 0.966, 95% CI: [0.930, 1.000]).</p><p><strong>Conclusion: </strong>Deep learning reconstruction of 3D-FLAIR provides higher image quality compared to a state-of-the-art technique with 30% less scan time while maintaining excellent agreement and interchangeability in quantitative evaluation.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 3","pages":"umaf016"},"PeriodicalIF":0.0,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429252/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145246245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving radiologist detection of meniscal abnormality on undersampled, deep learning reconstructed knee MRI.","authors":"Natalia Konovalova, Aniket Tolpadi, Felix Liu, Zehra Akkaya, Johanna Luitjens, Felix Gassert, Paula Giesler, Rupsa Bhattacharjee, Misung Han, Emma Bahroos, Sharmila Majumdar, Valentina Pedoia","doi":"10.1093/radadv/umaf015","DOIUrl":"10.1093/radadv/umaf015","url":null,"abstract":"<p><strong>Background: </strong>Accurate interpretation of meniscal anomalies on knee MRI is critical for diagnosis and treatment planning, with artificial intelligence emerging as a promising tool to support and enhance this process through automated anomaly detection.</p><p><strong>Purpose: </strong>To evaluate the impact of an artificial intelligence (AI) anomaly detection assistant on radiologists' interpretation of meniscal anomalies in undersampled, deep learning (DL)-reconstructed knee MRI and assess the relationship between reconstruction quality metrics and anomaly detection performance.</p><p><strong>Materials and methods: </strong>This retrospective study included 947 knee MRI examinations; 51 were excluded for poor image quality, leaving 896 participants (mean age, 44.7 ± 15.3 years; 472 women). Using 8-fold undersampled data, DL-based reconstructed images were generated. An object detection model was trained on original, fully sampled images and evaluated on 1 original and 14 DL-reconstructed test sets to identify meniscal lesions. Standard reconstruction metrics (normalized root mean square error, peak signal-to-noise ratio, and structural similarity index) and anomaly detection metrics (mean average precision, F1 score) were quantified and compared. Two radiologists independently reviewed a stratified sample of 50 examinations unassisted and assisted with AI-predicted anomaly boxes. McNemar's test evaluated differences in diagnostic performance; Cohen's kappa assessed interrater agreement.</p><p><strong>Results: </strong>On the original images, the anomaly detection model achieved the following: 70.53% precision, 72.17% recall, 63.09% mAP, and a 71.34% F1 score. Comparing performance among the undersampled reconstruction datasets, box-based reconstruction metrics showed better correlation with detection performance than traditional image-based metrics (mAP to box-based SSIM, <i>r</i> = 0.81, <i>P</i> < .01; mAP to image-based SSIM, <i>r</i> = 0.64, <i>P</i> = .01). In 50 participants, AI assistance improved radiologists' accuracy on reconstructed images. Sensitivity increased from 77.27% (95% CI, 65.83-85.72; 51/66) to 80.30% (95% CI, 69.16-88.11; 53/66), and specificity improved from 88.46% (95% CI, 83.73-91.95; 207/234) to 90.60% (95% CI, 86.18-93.71; 212/234) (<i>P</i> < .05).</p><p><strong>Conclusion: </strong>AI-assisted meniscal anomaly detection enhanced radiologists' interpretation of undersampled, DL-reconstructed knee MRI. Anomaly detection may serve as a complementary tool alongside other reconstruction metrics to assess the preservation of clinically important features in reconstructed images, warranting further investigation.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 2","pages":"umaf015"},"PeriodicalIF":0.0,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12021832/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144060178","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}