Academic Radiology最新文献

{"title":"Enhanced Image Quality and Comparable Diagnostic Performance of Prostate Fast Bi-MRI with Deep Learning Reconstruction","authors":"Liting Shen ,&nbsp;Ying Yuan ,&nbsp;Jin Liu ,&nbsp;Yue Cheng ,&nbsp;Qian Liao ,&nbsp;Rongchao Shi ,&nbsp;Tianyu Xiong ,&nbsp;Hui Xu ,&nbsp;Liang Wang ,&nbsp;Zhenghan Yang","doi":"10.1016/j.acra.2025.06.059","DOIUrl":"10.1016/j.acra.2025.06.059","url":null,"abstract":"<div><h3>Rational and Objectives</h3><div>To evaluate image quality and diagnostic performance of prostate biparametric MRI (bi-MRI) with deep learning reconstruction (DLR).</div></div><div><h3>Materials and Methods</h3><div>This prospective study included 61 adult male urological patients undergoing prostate MRI with standard-of-care (SOC) and fast protocols. Sequences included T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC) maps. DLR images were generated from FAST datasets. Three groups (SOC, FAST, DLR) were compared using: (1) five-point Likert scale, (2) signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), (3) lesion slope profiles, (4) dorsal capsule edge rise distance (ERD). PI-RADS scores were assigned to dominant lesions. ADC values were measured in histopathologically confirmed cases. Diagnostic performance was analyzed via receiver operating characteristic (ROC) curves (accuracy/sensitivity/specificity). Statistical tests included Friedman test, one-way ANOVA with post hoc analyses, and DeLong test for ROC comparisons (P&lt;0.05).</div></div><div><h3>Results</h3><div>FAST scanning protocols reduced acquisition time by nearly half compared to the SOC scanning protocol. When compared to T2WI_FAST, DLR significantly improved SNR, CNR, slope profile, and ERD (P &lt; 0.05). Similarly, DLR significantly enhanced SNR, CNR, and image sharpness when compared to DWI_FAST (P &lt; 0.05). No significant differences were observed in PI-RADS scores and ADC values between groups (P &gt; 0.05). The areas under the ROC curves, sensitivity, and specificity of ADC values for distinguishing benign and malignant lesions remained consistent (P &gt; 0.05).</div></div><div><h3>Conclusion</h3><div>DLR enhances image quality in fast prostate bi-MRI while preserving PI-RADS classification accuracy and ADC diagnostic performance.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 5964-5974"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144668917","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Differentiation of Postoperative Tumor Recurrence and Pseudoprogression in Gliomas: A Comparative Study of Six Diffusion Models","authors":"Peiquan Liu ,&nbsp;Yujie Chen ,&nbsp;Jing Zhao ,&nbsp;Ning Zheng ,&nbsp;Yue Hu ,&nbsp;Tengfei Chao ,&nbsp;Jiaxuan Zhang ,&nbsp;Wenzhen Zhu","doi":"10.1016/j.acra.2025.07.036","DOIUrl":"10.1016/j.acra.2025.07.036","url":null,"abstract":"<div><h3>Rationale and Objectives</h3><div>The purpose of this study is to investigate whether six diffusion models derived from multi-b-value diffusion-weighted imaging can enhance the differentiation between pseudoprogression (PsP) and postoperative tumor recurrence (TR) in glioma patients, with the aim of providing clinical insights.</div></div><div><h3>Materials and Methods</h3><div>A retrospective study was conducted on 82 patients with WHO grade 2–4 gliomas who underwent surgery at our hospital, with MRI sequences including T₁WI, T₂WI, T₂FLAIR, contrast-enhanced T₁WI, and multi-b-value DWI. Postoperative follow-up or secondary surgery pathology confirmed 46 cases of TR and 36 cases of PsP. Six diffusion models were fitted based on multi-b-value DWI sequences, including monoexponential DWI (Mono_DWI), intravoxel incoherent motion (IVIM), diffusion kurtosis imaging (DKI), stretched-exponential model (SEM), fractional-order calculus (FROC), and continuous-time random-walk (CTRW) model. ROIs were manually outlined to calculate the average values of each parameter. Differences between the two groups were compared using T-tests or Mann–Whitney U tests. The diagnostic performance of individual parameters was analyzed using ROC curve analysis, and the diagnostic performance of each model was compared using multivariate logistic regression.</div></div><div><h3>Results</h3><div>Among the 14 parameter maps, significant differences were found in all models (P&lt;0.0036) except for IVIM_D*, IVIM_f. ROC curve analysis showed that CTRW_D demonstrated the highest AUC of 0.8484 (0.7549–0.9240). Further analysis of the diffusion models showed that CTRW performed the best among all models, with an AUC of 0.8635 (0.7816–0.9454), slightly higher than the FROC model, which had an AUC of 0.8629 (0.7839–0.9420).</div></div><div><h3>Conclusion</h3><div>The various diffusion models derived from multi-b-value DWI sequences can effectively distinguish between postoperative recurrence and pseudoprogression in gliomas. Among these models, the CTRW and FROC models are the two optimal models, demonstrating comparable diagnostic performance.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 6181-6193"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144800874","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Association Between Intracranial Plaque Characteristics and Stroke Recurrence Risk Stratified by Stenosis Severity","authors":"Lin Han ,&nbsp;Ziang Li ,&nbsp;Jie Wang ,&nbsp;Yu Gao ,&nbsp;Liulei Zhang ,&nbsp;Shi Han ,&nbsp;Lin Li ,&nbsp;Ziqing Yang ,&nbsp;Xinhui Ma ,&nbsp;Haipeng Liu ,&nbsp;Shufeng Peng ,&nbsp;Hongling Zhao ,&nbsp;Hongkai Cui","doi":"10.1016/j.acra.2025.07.049","DOIUrl":"10.1016/j.acra.2025.07.049","url":null,"abstract":"<div><h3>Rationale and Objectives</h3><div>This study aims to analyze the association between plaque characteristics and stroke recurrence across different degrees of stenosis, and to evaluate the predictive value of combining culprit plaque characteristics with the modified Essen score for stroke recurrence.</div></div><div><h3>Materials and Methods</h3><div>This retrospective analysis included 368 intracranial atherosclerotic stenosis (ICAS) patients from two centers, categorized into mild-to-moderate stenosis (&lt;70%, <em>n</em> = 155) and severe stenosis (≥70%, <em>n</em> = 213) groups. High-resolution vessel wall images and clinical data were analyzed. Univariate and multivariate Logistic regression analyses were conducted to determine risk factors for stroke recurrence. The predictive performance was evaluated using the area under the receiver operating characteristic curve.</div></div><div><h3>Results</h3><div>In the mild-to-moderate stenosis group, a higher Normalized Wall Index (NWI) (OR = 1.088, 95% CI: 1.009-1.186, <em>P</em> = .04) and the presence of intraplaque hemorrhage (IPH) (OR = 4.049, 95% CI: 1.227-16.065, <em>P</em> = .03) were risk factors for stroke recurrence, but not in severe stenosis. A combined model (IPH, NWI, modified Essen score) improved predictive performance over the modified Essen score alone, particularly in mild-to-moderate stenosis.</div></div><div><h3>Conclusion</h3><div>In patients with luminal stenosis of less than 70%, elevated NWI and IPH were validated as important and effective indicators for predicting stroke recurrence, demonstrating significant predictive performance. Conversely, plaque characteristics showed limited predictive utility in severe stenosis, highlighting stenosis-dependent differences in plaque-related stroke risk. Integrating imaging biomarkers with modified Essen score optimizes recurrence prediction in mild-to-moderate ICAS.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 6205-6215"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144812677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"“Invisible” but Important: Radiomics to Identify Microscopic Disease and Recurrence in HCC","authors":"Jordan H. Chamberlin MD,&nbsp;Andrew Hardie MD","doi":"10.1016/j.acra.2025.08.041","DOIUrl":"10.1016/j.acra.2025.08.041","url":null,"abstract":"","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 5772-5773"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144977420","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Beneficial value of Tumor Markers for Enhancing ESCC Grading Prediction By Combining CT Radiomics and Inflammatory Immune Model","authors":"Fei Wang,&nbsp;Shaokun Zheng","doi":"10.1016/j.acra.2025.05.028","DOIUrl":"10.1016/j.acra.2025.05.028","url":null,"abstract":"","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Page 6353"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144235860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improved Beam-hardening Reduction in CT - Impact on Image Quality and Diagnostic Certainty in Emergency Imaging of the Brain","authors":"Foroud Aghapour Zangeneh MD ,&nbsp;Gonçalo G. Almeida MD ,&nbsp;Susanne Bensler MD ,&nbsp;Reza Omidi Varmezani MD ,&nbsp;Thomas Sartoretti MD ,&nbsp;Frank Johannes Ahlhelm MD ,&nbsp;Tilo Niemann MD, MHBA, EBCR ,&nbsp;André Euler MD, MHBA, EBCR","doi":"10.1016/j.acra.2025.05.054","DOIUrl":"10.1016/j.acra.2025.05.054","url":null,"abstract":"<div><h3>Rationale and Objectives</h3><div>To assess the impact of an optimized beam-hardening artifact reduction algorithm on image quality and diagnostic certainty in emergency unenhanced brain CT.</div></div><div><h3>Materials and Methods</h3><div>Retrospective study of consecutive patients referred for unenhanced brain CT to rule out traumatic brain injuries in 2023. Images were reconstructed using both (A) a standard and (B) an optimized iterative beam-hardening correction algorithm (iBHC). CT attenuation, image noise and SNR were measured in the cortex of supratentorial and infratentorial regions and in the pons. Posterior Fossa Artifact Index (PFAI) and Subcalvarial Artifact Index (SAI) were calculated. Two neuroradiologists and two emergency radiologists independently compared artifacts and diagnostic certainty between both algorithms using 5-point Likert scales and side-by-side comparisons. A paired Wilcoxon-test with correction for multiple testing was used.</div></div><div><h3>Results</h3><div>100 consecutive patients (55 women; 64.1 ± 20 years) were included. CT attenuation was significantly lower for B (all P &lt;.0001). SNR was significantly lower supratentorial (frontal region: 10.5 vs. 12.9, P&lt;.0001) and significantly higher in the pons (5.9 vs. 5.5, P &lt;.0001) for B. PFAI was significantly reduced for B (5.5 vs. 6.4, P &lt;.0001), while there was no significant difference in SAI (P = 0.304). The optimized algorithm was selected as superior in 100%, 100%, 99%, 99% of supratentorial and in 100%, 99%, 99%, 86% of infratentorial cases for readers 1 to 4, respectively.</div></div><div><h3>Conclusion</h3><div>An optimized iBHC algorithm demonstrated significantly improved image quality, reduced artifacts and improved diagnostic certainty in emergency unenhanced brain CT.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 6072-6079"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144287030","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Usefulness of [18F]FAPI-04 and [18F]FDG PET/CT for the Detection of Peritoneal Carcinomatosis: A Comparative Study","authors":"Chengzhi Jiang ,&nbsp;Kai Zheng ,&nbsp;Yanyin Zhang ,&nbsp;Xiang Peng ,&nbsp;Jian Yang ,&nbsp;Hui Ye ,&nbsp;Yue Chen","doi":"10.1016/j.acra.2025.05.067","DOIUrl":"10.1016/j.acra.2025.05.067","url":null,"abstract":"<div><h3>Rationale and Objectives</h3><div>In this study, we aimed to compared the performance of [¹⁸F]FAPI-04 and [¹⁸<span>F]fluorodeoxyglucose (FDG) positron emission tomography/computer tomography (PET/CT) in the evaluation of peritoneal carcinomatosis.</span></div></div><div><h3>Materials and Methods</h3><div><span>71 patients with suspected peritoneal malignancies were enrolled in our study. All the participants underwent both [</span>¹⁸F]FAPI-04 and [¹⁸<span>F]FDG PET/CT imaging within 7 days. The detection rates, diagnostic accuracies, semiquantitative parameters of the tracers<span>, peritoneal cancer index (PCI) scores, and tumor markers were evaluated and compared.</span></span></div></div><div><h3>Results</h3><div><span>Among the 71 patients, 40 patients were diagnosed with peritoneal carcinomatosis, and 31 were true-negative patients. The sensitivity and accuracy of [</span>¹⁸F]FAPI-04 PET/CT were higher than those of [¹⁸F]FDG PET/CT (sensitivity: 92.50% vs. 72.50%, <em>p<!--> </em>=<!--> <!-->0.003; accuracy: 91.55% vs. 80.28%, <em>p</em>&lt;0.001), particularly in patients with gastric cancer. The SUVmax, tumor-to-liver background ratio (TBR-L), tumor-to-descending aorta ratio (TBR-A), and PCI score were significantly higher for [¹⁸F]FAPI-04 PET/CT than [¹⁸F]FDG PET/CT (all <em>p</em>&lt;0.05). In the [¹⁸F]FAPI-04 PET/CT group, the PCI score, TBR-L, TBR-A, TBR-M and SUVmax were higher in the high level group than the low level group (all <em>p</em><span>&lt;0.05). The carbohydrate antigen 125 (CA 125) levels were strongly correlated with the PCI of both [</span>¹⁸F]FAPI-04 and [¹⁸F]FDG PET/CT.</div></div><div><h3>Conclusion</h3><div>[¹⁸F]FAPI-04 PET/CT outperformed [¹⁸F]FDG PET/CT in the evaluation of peritoneal carcinomatosis, particularly in patients with gastric cancer. Furthermore, [¹⁸F]FAPI-04 PET/CT may be used for the assessment of peritoneal carcinomatosis in patients, especially FAPI-PCI.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 5897-5906"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144337204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Study on the Selection of Surgical Methods for ≤2cm Lung Adenocarcinomas Based on High-resolution CT Features Combined with Solid Component Size on MPVR","authors":"Mao-lu Tan BSc ,&nbsp;Ke Zhang BSc ,&nbsp;Yun-dan Zhang BSc,&nbsp;Ling-qi Gao BSc,&nbsp;Fajin Lv","doi":"10.1016/j.acra.2025.06.011","DOIUrl":"10.1016/j.acra.2025.06.011","url":null,"abstract":"&lt;div&gt;&lt;h3&gt;Rationale and Objectives&lt;/h3&gt;&lt;div&gt;The precise treatment methods for small early-stage lung cancer remain unclear. This study aims to analyze the high-resolution computed tomography (HRCT) imaging features and the size of the solid component in multiplanar volume reconstruction (MPVR) in depth, to select the optimal surgical approach (lobar resection or sublobar resection) for patients with lung adenocarcinomas ≤2cm, thereby optimizing clinical treatment strategies and improving patient prognosis.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Methods&lt;/h3&gt;&lt;div&gt;&lt;span&gt;&lt;span&gt;A retrospective cohort of 657 consecutive patients with surgically resected lung&lt;span&gt; adenocarcinoma was analyzed at the First Affiliated Hospital of Chongqing Medical University (2014–2023), comprising 345 lobar resection (52.5%) and 312 sublobar resection (47.5%) cases. All participants underwent standardized preoperative and postoperative HRCT examinations with consistent imaging protocols. The differences between the two groups of patients in terms of demographics, preoperative HRCT imaging features, clinical manifestations, postoperative pathological grading, and the 5-year recurrence-free survival rate and 5-year &lt;/span&gt;&lt;/span&gt;overall survival&lt;span&gt; rate were compared. Based on the presence of spiculation and air bronchogram, pulmonary nodules were classified into four categories: nodules with both signs, nodules with spiculation but without air bronchogram, nodules with air bronchogram but without spiculation, and nodules without either sign. Significant variables (&lt;/span&gt;&lt;/span&gt;&lt;em&gt;P&lt;/em&gt;&lt;span&gt;&lt;0.05) from univariate survival analysis were used to construct a multivariate Cox proportional hazards model, and Kaplan-Meier curves and log-rank tests were employed to compare survival rates between different groups.&lt;/span&gt;&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Results&lt;/h3&gt;&lt;div&gt;There was no statistically significant difference in the 5-year overall survival (OS, &lt;em&gt;P&lt;/em&gt;=0.68) and 5-year recurrence-free survival (RFS, &lt;em&gt;P&lt;/em&gt;=0.34) between the lobar and sublobar resection groups. Cox regression analysis revealed that spiculation (&lt;em&gt;P&lt;/em&gt;&lt;0.05), MPVR solid component &gt;5.6 mm (&lt;em&gt;P&lt;/em&gt;&lt;0.05), and sublobar resection (&lt;em&gt;P&lt;/em&gt;&lt;0.05) were predictors of poorer survival, while air bronchogram was associated with better survival (&lt;em&gt;P&lt;/em&gt;&lt;0.05). Statistical results showed that patients with spiculation had a worse prognosis than those with air bronchogram (RFS: &lt;em&gt;P&lt;/em&gt;&lt;0.05, OS: &lt;em&gt;P&lt;/em&gt;&lt;0.05).&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Conclusion&lt;/h3&gt;&lt;div&gt;&lt;span&gt;Spiculation, air bronchogram, and the diameter of the solid component in MPVR were significantly associated with patient prognosis. A tripartite management framework is proposed: 1) High-risk group (spiculation + solid component &gt;5.6 mm): Lobectomy&lt;span&gt; with systematic lymph node dissection. 2) Low-risk group (air bronchogram + solid component ≤5.6 mm): Sublobar resection or &lt;/span&gt;&lt;/span&gt;stereotactic body radiotherapy (SBRT). 3) Intermediat","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 6284-6294"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144545931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated Detection of Focal Bone Marrow Lesions From MRI: A Multi-center Feasibility Study in Patients with Monoclonal Plasma Cell Disorders","authors":"Markus Wennmann MD ,&nbsp;Jessica Kächele MSc ,&nbsp;Arvin von Salomon ,&nbsp;Tobias Nonnenmacher MD ,&nbsp;Markus Bujotzek MSc ,&nbsp;Shuhan Xiao MSc ,&nbsp;Andres Martinez Mora MSc ,&nbsp;Thomas Hielscher MSc ,&nbsp;Marina Hajiyianni MD ,&nbsp;Ekaterina Menis PhD ,&nbsp;Martin Grözinger MD ,&nbsp;Fabian Bauer MD ,&nbsp;Veronika Riebl MD ,&nbsp;Lukas Thomas Rotkopf MD, MSc ,&nbsp;Kevin Sun Zhang MD ,&nbsp;Saif Afat MD ,&nbsp;Britta Besemer MD ,&nbsp;Martin Hoffmann MD ,&nbsp;Adrian Ringelstein MD ,&nbsp;Ullrich Graeven MD ,&nbsp;Klaus Maier-Hein PhD","doi":"10.1016/j.acra.2025.06.034","DOIUrl":"10.1016/j.acra.2025.06.034","url":null,"abstract":"<div><h3>Rationale and Objectives</h3><div>To train and test an AI-based algorithm for automated detection of focal bone marrow lesions (FL) from MRI.</div></div><div><h3>Materials and Methods</h3><div>This retrospective feasibility study included 444 patients with monoclonal plasma cell disorders. For this feasibility study, only FLs in the left pelvis were included. Using the nnDetection framework, the algorithm was trained based on 334 patients with 494 FLs from center 1, and was tested on an internal test set (36 patients, 89 FLs, center 1) and a multicentric external test set (74 patients, 262 FLs, centers 2–11). Mean average precision (mAP), F1-score, sensitivity, positive predictive value (PPV), and Spearman correlation coefficient between automatically determined and actual number of FLs were calculated.</div></div><div><h3>Results</h3><div>On the internal/external test set, the algorithm achieved a mAP of 0.44/0.34, F1-Score of 0.54/0.44, sensitivity of 0.49/0.34, and a PPV of 0.61/0.61, respectively. In two subsets of the external multicentric test set with high imaging quality, the performance nearly matched that of the internal test set, with mAP of 0.45/0.41, F1-Score of 0.50/0.53, sensitivity of 0.44/0.43, and a PPV of 0.60/0.71, respectively. There was a significant correlation between the automatically determined and actual number of FLs on both the internal (r<!--> <!-->=<!--> <!-->0.51, p<!--> <!-->=<!--> <!-->0.001) and external multicentric test set (r<!--> <!-->=<!--> <!-->0.59, p&lt;0.001).</div></div><div><h3>Conclusion</h3><div>This study demonstrates that the automated detection of FLs from MRI, and thereby the automated assessment of the number of FLs, is feasible.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 6012-6026"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144610258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep Learning-Based Prediction of Microvascular Invasion and Survival Outcomes in Hepatocellular Carcinoma Using Dual-phase CT Imaging of Tumors and Lesser Omental Adipose: A Multicenter Study","authors":"Shidi Miao ,&nbsp;Mengzhuo Sun ,&nbsp;Xuemeng Li ,&nbsp;Mingxuan Wang ,&nbsp;Yuyang Jiang ,&nbsp;Zengyao Liu ,&nbsp;Qiujun Wang ,&nbsp;Xuemei Ding ,&nbsp;Ruitao Wang","doi":"10.1016/j.acra.2025.07.015","DOIUrl":"10.1016/j.acra.2025.07.015","url":null,"abstract":"<div><h3>Background</h3><div>Accurate preoperative prediction of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) remains challenging. Current imaging biomarkers show limited predictive performance.</div></div><div><h3>Purpose</h3><div>To develop a deep learning model based on preoperative multiphase CT images of tumors and lesser omental adipose tissue (LOAT) for predicting MVI status and to analyze associated survival outcomes.</div></div><div><h3>Materials and Methods</h3><div>This retrospective study included pathologically confirmed HCC patients from two medical centers between 2016 and 2023. A dual-branch feature fusion model based on ResNet18 was constructed, which extracted fused features from dual-phase CT images of both tumors and LOAT. The model's performance was evaluated on both internal and external test sets. Logistic regression was used to identify independent predictors of MVI. Based on MVI status, patients in the training, internal test, and external test cohorts were stratified into high- and low-risk groups, and overall survival differences were analyzed.</div></div><div><h3>Results</h3><div>The model incorporating LOAT features outperformed the tumor-only modality, achieving an AUC of 0.889 (95% CI: [0.882, 0.962], P<!--> <!-->=<!--> <!-->0.004) in the internal test set and 0.826 (95% CI: [0.793, 0.872], P<!--> <!-->=<!--> <!-->0.006) in the external test set. Both results surpassed the independent diagnoses of three radiologists (average AUC<!--> <!-->=<!--> <!-->0.772). Multivariate logistic regression confirmed that maximum tumor diameter and LOAT area were independent predictors of MVI. Further Cox regression analysis showed that MVI-positive patients had significantly increased mortality risks in both the internal test set (Hazard Ratio [HR]<!--> <!-->=<!--> <!-->2.246, 95% CI: [1.088, 4.637], P<!--> <!-->=<!--> <!-->0.029) and external test set (HR<!--> <!-->=<!--> <!-->3.797, 95% CI: [1.262, 11.422], P<!--> <!-->=<!--> <!-->0.018).</div></div><div><h3>Conclusion</h3><div>This study is the first to use a deep learning framework integrating LOAT and tumor imaging features, improving preoperative MVI risk stratification accuracy. Independent prognostic value of LOAT has been validated in multicenter cohorts, highlighting its potential to guide personalized surgical planning.</div></div>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 10","pages":"Pages 5789-5801"},"PeriodicalIF":3.9,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144709764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}