Academic Radiology最新文献

{"title":"Comparison of Perfusion CT and Conventional Thin-slice Multiphase CT in the Diagnosis of Pancreatic Adenocarcinoma.","authors":"Juan Li, Xin-Yue Chen, Yu-Hong Wang, Xiao Wang, Yue-Juan Cheng, Meng-Chao Liu, Liang Zhu, Xin Gao, Wen-Yi Deng, Jing-Yi Liu, Xi-Juan Lin, Zheng-Yu Jin, Hua-Dan Xue","doi":"10.1016/j.acra.2025.03.049","DOIUrl":"https://doi.org/10.1016/j.acra.2025.03.049","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>Perfusion CT parameters are reported to correlate with pancreatic adenocarcinoma's histopathological response to radiochemotherapy, yet research on morphological diagnosis of perfusion CT for the diagnosis of pancreatic adenocarcinoma is lacking. This study compares mean temporal (MT) post-processed perfusion CT with conventional thin-slice multiphase CT in visualizing tumors, small pancreatic arteries, and assessing tumor resectability.</p><p><strong>Materials and methods: </strong>60 patients (mean age 61.3 ± 8.8, 36 males) underwent perfusion and conventional CT sequentially from December 2021 to April 2024 were retrospectively included. MT images were calculated from perfusion CT and compared with conventional images for tumor depiction (qualitative 5-point scale, quantitative analysis), small pancreatic arteries display (qualitative 4-point scale) and concordance in tumor resectability. Radiation doses were also evaluated.</p><p><strong>Results: </strong>MT images showed superior tumor display scores (5 (4,5) vs. 4 (4,5)), better tumor contrast (99.54 (81.88, 117.29) vs. 51.90 ± 18.85), higher signal-to-noise ratio (4.46 ± 1.75 vs. 3.10 ± 0.98), and contrast-to-noise ratio (5.13 (3.84, 6.77) vs. 3.03 ± 1.24), with all p values < 0.001. Qualitative scores for small pancreatic arteries were higher in MT images, with most p values <0.05 (range from <0.001 to 0.018). Both radiologists showed good resectability consistency, with κ values of 0.740 and 0.785, respectively. Effective radiation doses were 11.86 (9.45, 15.57) mSv for perfusion CT and 12.47 ± 4.01 mSv for conventional CT (p=0.958).</p><p><strong>Conclusion: </strong>Perfusion CT employing MT post-processing outperforms conventional CT in depicting tumors and small pancreatic arteries, with consistent resectability results between the two examinations.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144112672","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":"Radiologists are Human Too: Addressing the Impacts of Occupational Exhaustion, Travel History, and Hunger on Mammography Reading Performance.","authors":"Abdulaziz S Alshabibi, Sultan F Alhujaili, Basel Qenam, Areej Aloufi, Salman M Albeshan, Meaad M Almusined, Abdulmajeed Alotabibi, Nuha A Khoumais","doi":"10.1016/j.acra.2025.04.067","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.067","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To investigate how occupational exhaustion, travel, and hunger affect mammography reading performance among radiologists and radiology trainees.</p><p><strong>Methods: </strong>Thirty participants (22 radiologists, eight radiology trainees) completed mammography reading assessments using the DetectedX platform during two radiology conferences in Saudi Arabia. Each participant interpreted 30 de-identified mammographic cases (15 abnormal, 15 normal) under standardized conditions. Performance was measured using jackknife alternative free-response receiver operating characteristic (JAFROC), lesion sensitivity, area under the receiver operating characteristic curve (ROC AUC), sensitivity, and specificity. Three independent variables were self-reported using a questionnaire completed immediately before the mammography reading session: occupational exhaustion (low to moderate or high, assessed by the Emotional Exhaustion subscale of the Maslach Burnout Inventory-General Survey), recent travel (traveler or non-traveler), and hunger status (hungry or not hungry). Mann-Whitney U tests were used to examine differences in reading performance associated with each variable.</p><p><strong>Results: </strong>Participants with high occupational exhaustion had significantly lower JAFROC scores compared to those with low to moderate exhaustion (0.213 vs 0.383; p=0.041). Recently traveled participants had significantly lower ROC AUC scores (0.681 vs. 0.772; p=0.03) and lower sensitivity (70.0% vs. 80.0%; p=0.04) than non-travelers. Hungry participants exhibited higher sensitivity (85.0% vs. 70.0%; p=0.04) but lower specificity (40.0% vs. 65.0%; p=0.02) compared to non-hungry participants.</p><p><strong>Conclusion: </strong>Occupational exhaustion was associated with poorer JAFROC scores, recent travel lowered sensitivity and overall diagnostic accuracy, and hunger increased sensitivity at the cost of specificity. These findings highlight the importance of addressing radiologists' well-being by optimizing workloads, allowing recovery periods after travel, and ensuring structured meal schedules. Future research should explore real-world clinical settings and targeted interventions, including AI integration.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144095173","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":"Evaluating the Performance of Reasoning Large Language Models on Japanese Radiology Board Examination Questions.","authors":"Takeshi Nakaura, Hiroto Takamure, Naoki Kobayashi, Kaori Shiraishi, Naofumi Yoshida, Yasunori Nagayama, Hiroyuki Uetani, Masafumi Kidoh, Yoshinori Funama, Toshinori Hirai","doi":"10.1016/j.acra.2025.04.060","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.060","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>This study evaluates the performance, cost, and processing time of OpenAI's reasoning large language models (LLMs) (o1-preview, o1-mini) and their base models (GPT-4o, GPT-4o-mini) on Japanese radiology board examination questions.</p><p><strong>Materials and methods: </strong>A total of 210 questions from the 2022-2023 official board examinations of the Japan Radiological Society were presented to each of the four LLMs. Performance was evaluated by calculating the percentage of correctly answered questions within six predefined radiology subspecialties. The total cost and processing time for each model were also recorded. The McNemar test was used to assess the statistical significance of differences in accuracy between paired model responses.</p><p><strong>Results: </strong>The o1-preview achieved the highest accuracy (85.7%), significantly outperforming GPT-4o (73.3%, P<.001). Similarly, o1-mini (69.5%) performed significantly better than GPT-4o-mini (46.7%, P<.001). Across all radiology subspecialties, o1-preview consistently ranked highest. However, reasoning models incurred substantially higher costs (o1-preview: $17.10, o1-mini: $2.58) compared to their base counterparts (GPT-4o: $0.496, GPT-4o-mini: $0.04), and their processing times were approximately 3.7 and 1.2 times longer, respectively.</p><p><strong>Conclusion: </strong>Reasoning LLMs demonstrated markedly superior performance in answering radiology board exam questions compared to their base models, albeit at a substantially higher cost and increased processing time.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144095245","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":"Application of Quantitative CT and Machine Learning in the Evaluation and Diagnosis of Polymyositis/Dermatomyositis-Associated Interstitial Lung Disease.","authors":"Kai Yang, Yanrong Chen, Liyu He, Yadan Sheng, He Hei, Jingping Zhang, Chenwang Jin","doi":"10.1016/j.acra.2025.04.012","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.012","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To investigate lung changes in patients with polymyositis/dermatomyositis-associated interstitial lung disease (PM/DM-ILD) using quantitative CT and to construct a diagnostic model to evaluate the application of quantitative CT and machine learning in diagnosing PM/DM-ILD.</p><p><strong>Method: </strong>Chest CT images from 348 PM/DM individuals were quantitatively analyzed to obtain the lung volume (LV), mean lung density (MLD), and intrapulmonary vascular volume (IPVV) of the whole lung and each lung lobe. The percentage of high attenuation area (HAA %) was determined using the lung density histogram. Patients hospitalized from 2016 to 2021 were used as the training set (n=258), and from 2022 to 2023 were used as the temporal test set (n=90). Seven classification models were established, and their performance was evaluated through ROC analysis, decision curve analysis, calibration, and precision-recall curve. The optimal model was selected and interpreted with Python's SHAP model interpretation package.</p><p><strong>Results: </strong>Compared to the non-ILD group, the mean lung density and percentage of high attenuation area in the whole lung and each lung lobe were significantly increased, and the lung volume and intrapulmonary vessel volume were significantly decreased in the ILD group. The Random Forest (RF) model demonstrated superior performance with the test set area under the curve of 0.843 (95% CI: 0.821-0.865), accuracy of 0.778, sensitivity of 0.784, and specificity of 0.750.</p><p><strong>Conclusion: </strong>Quantitative CT serves as an objective and precise method to assess pulmonary changes in PM/DM-ILD patients. The RF model based on CT quantitative parameters displayed strong diagnostic efficiency in identifying ILD, offering a new and convenient approach for evaluating and diagnosing PM/DM-ILD patients.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144095244","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":"The Clinical and Neuroimaging Factors of Stroke Outcome of Unilateral Moyamoya Disease.","authors":"Hongtao Zhang, Mingming Lu, Shitong Liu, Dongqing Liu, Heguan Fu, Cong Han, Baobao Li, Fugeng Sheng, Jianming Cai","doi":"10.1016/j.acra.2025.04.045","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.045","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>The clinical feature and long-term prognosis of unilateral moyamoya disease (MMD) have not been fully described and studied. The study aimed to investigate independent risk factors for stroke in unilateral MMD patients during a long-term follow-up.</p><p><strong>Materials and methods: </strong>A total of 393 unilateral MMD patients (median age, 40 years) were assessed at baseline and followed for an average time of 68.9 months. Ischemic and hemorrhagic stroke incidence rates were determined. Multiple demographic, clinical and neuroimaging factors at baseline were considered as potential predictors of stroke during the follow-up period. Hazard ratios (HR) and corresponding 95% confidence interval (CI) for stroke were calculated by univariable and multivariable Cox proportional hazards models. Cumulative risk of stroke was estimated by the Kaplan-Meier product-limit method.</p><p><strong>Results: </strong>During the follow-up period, 43 patients experienced stroke events (10.9%). 5 children experienced stroke events (5/46, 10.9%) and 38 adults experienced stroke events (38/347, 11.0%) (P>0.05). 21 patients with encephaloduroarteriosynangiosis (EDAS) experienced stroke events (21/254, 8.3%) and 22 patients with conservative treatment experienced stroke events (22/139, 15.8%) (P<0.05). After adjustment for clinical characteristics, multivariable analysis showed that involvement of posterior cerebral artery (HR, 2.199; 95% CI, 1.100-4.398), decreased cerebral blood flow (CBF) (HR, 2.292; 95% CI, 1.182-4.446) and concentric enhancement of the arterial wall (HR, 3.093; 95% CI, 1.617-5.915) were significantly associated with stroke, and EDAS (HR, 0.385; 95% CI, 0.203-0.730) and compensatory blood supply by anterior communicating artery (HR, 0.413; 95% CI, 0.206-0.830) were protective factors for stroke.</p><p><strong>Conclusion: </strong>Involvement of posterior cerebral artery, decreased CBF, concentric enhancement of the arterial wall, EDAS and compensatory blood supply by anterior communicating artery may help stratify the risk of stroke and improve therapeutic decisions in unilateral MMD. Unilateral MMD could benefit from EDAS and have a lower risk of future stroke.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144095179","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":"Development of a Radiology Department Art Gallery: Theory Into Practice.","authors":"Benjamin Park, Kaitlyn Ortgiesen, Erin A Cooke","doi":"10.1016/j.acra.2025.04.059","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.059","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>The benefits of art in medicine have been well described but integration has remained elusive due to the lack of sustainable initiatives. This study highlights the development of an art gallery to address the lack of arts programming in the field of radiology.</p><p><strong>Materials and methods: </strong>A radiology department art gallery initiative was formalized in 2021 at a tertiary-care university academic medical center. Artwork was displayed in department common areas. To assess the response to the arts gallery, an anonymous REDCap survey was distributed to interested participants. Ancillary processes developed to support the arts gallery included creation of a formal submission procedure, mini-grants to decrease cost as a barrier, and an arts committee to review submissions comprised of faculty, staff, and residents.</p><p><strong>Results: </strong>Among 59 participants in the survey, the majority were either faculty radiologists or staff, with residents making up a minority. Levels of artistic ability were variable with the majority having none (36%) or at the level of novice/dabbler (42%). Despite this, the impact of the art gallery was positive including beneficial effects on interpersonal interactions, wellness, happiness, stress reduction, creativity, a broader range of expression and/or viewpoints in the workplace, diversity and inclusion, and synergy with other initiatives. Submissions of art were associated with significantly higher perceived benefits of art gallery on happiness, stress reduction, interpersonal interactions, and wellness (p<0.05).</p><p><strong>Conclusion: </strong>The benefits of a sustainable radiology department art gallery may be substantial, and most participants noted positive effects that were further increased by participation through art submissions. Ultimately, sustainable arts initiatives with objective measurement mechanisms may support humanistic factors in radiology.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144086935","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":"Developing and Evaluating a Nomogram Model Predicting Axillary Lymph Node Metastasis of Triple-Negative Breast Cancer Based on Multimodal Imaging Characteristics.","authors":"Yantong Jin, Xingyuan Liu, Xingda Zhang, Yang Wang, Xiaoying Cheng, Siwei Cao, Wuyue Zhang, Mingming Zhao, Ye Ruan, Bo Gao","doi":"10.1016/j.acra.2025.04.031","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.031","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>Breast cancer is the most frequently diagnosed cancer among women worldwide, with axillary lymph nodes being common sites of metastasis, particularly triple-negative breast cancer (TNBC), which is the subtype with the poorest prognosis. This study aimed to develop a nomogram model to predict axillary lymph node metastasis (ALNM) in TNBC patients based on mammography (MG), multimodal ultrasound (US), and clinical pathological characteristics.</p><p><strong>Patients and methods: </strong>A retrospective study was performed on 291 patients diagnosed with TNBC from two centers. Patients from the Center 1 were randomly divided into a training cohort (n = 159) and a internal test cohort (n = 68) using a 7:3 ratio, while patients from the Center 2 served as an external test cohort. Each group was further divided into an ALNM group and a non-ALNM group based on the presence or absence of ALNM. Predictors were selected via least absolute shrinkage and selection operator (LASSO) regression and multivariable logistic analysis. The predictive performance of the nomogram model was evaluated by the receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis (DCA).</p><p><strong>Results: </strong>Notable predictors included MG_reported_margin, MG_reported_suspicious malignant calcifications, MG_reported_abnormal ALN, elastography score, and US_reported_abnormal ALN. The area under the receiver operating characteristics curve (AUC) value of the nomogram model was 0.931 (95%CI: 0.890-0.973) for the training cohort, AUC=0.929 (95%CI: 0.871-0.986) for the internal test cohort and AUC=0.891 (95%CI: 0.794-0.987) for the external test cohort. Calibration curves and DCA both suggested that the nomogram exhibited favorable calibration and clinical utility.</p><p><strong>Conclusion: </strong>The predictive model combined with multimodal US and MG characteristics developed in this study is highly accurate, serves as a powerful tool for clinical assessment, and shows promise for predicting ALNM in patients with TNBC.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144086926","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":"Correlating Economic Cycles with Job Placement in Academic Radiology: Implications for Graduating Residents' Recruitment and Mentorship.","authors":"Khushboo Jhala, Sana Majid, Sara Durfee, Glenn C Gaviola, Zhou Lan, William W Mayo-Smith","doi":"10.1016/j.acra.2025.04.061","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.061","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To assess the correlation between US economic cycle and radiology residents' entry into academic vs nonacademic practice models.</p><p><strong>Materials and methods: </strong>US Gross Domestic Product (GDP) data from 2002-2023 identified two discrete economic recessions: the 2008 mortgage crisis and the 2020 COVID-19 pandemic. An alumni database was created for graduating residents from a large academic tertiary care center 2years prior to, during, and after these recessions: 2006-2012 (n=52) and 2018-2024 (n=58). Fisher's Exact Test compared alumni ratios in academic vs other practice models (ie, private, teleradiology, community hybrid) pre- and post-GDP decelerations.</p><p><strong>Results: </strong>35% (6/17) of alumni graduating during the mortgage crisis recession currently practice in academics vs 6% (1/16) who graduated prior to the recession (P=.05). 67% (12/18) of alumni graduating during the COVID recession currently practice in academics vs 35% (7/20) prior to (P=.05). In post-COVID economic expansion, only 25% (5/20) of 2023-2024 graduates practice in or have signed in academics compared to during the recession (P<.05).</p><p><strong>Conclusion: </strong>There is an inverse relationship between phases of economic growth and graduating radiology resident job placement with academic practices. Understanding economic cycles can help academic practices develop timely recruitment strategies to maintain workforce stability and provide resident mentorship.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144086803","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":"External Validation of a CT-Based Radiogenomics Model for the Detection of EGFR Mutation in NSCLC and the Impact of Prevalence in Model Building by Using Synthetic Minority Over Sampling (SMOTE): Lessons Learned.","authors":"Andres A Kohan, S A Mirshahvalad, R Hinzpeter, R Kulanthaivelu, L Avery, C Ortega, U Metser, A Hope, P Veit-Haibach","doi":"10.1016/j.acra.2025.04.064","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.064","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>Radiogenomics holds promise in identifying molecular alterations in nonsmall cell lung cancer (NSCLC) using imaging features. Previously, we developed a radiogenomics model to predict epidermal growth factor receptor (EGFR) mutations based on contrast-enhanced computed tomography (CECT) in NSCLC patients. The current study aimed to externally validate this model using a publicly available National Institutes of Health (NIH)-based NSCLC dataset and assess the effect of EGFR mutation prevalence on model performance through synthetic minority oversampling technique (SMOTE).</p><p><strong>Materials and methods: </strong>The original radiogenomics model was validated on an independent NIH cohort (n=140). For assessing the influence of disease prevalence, six SMOTE-augmented datasets were created, simulating EGFR mutation prevalence from 25% to 50%. Seven models were developed (one from original data, six SMOTE-augmented), each undergoing rigorous cross-validation, feature selection, and logistic regression modeling. Models were tested against the NIH cohort. Performance was compared using area under the receiver operating characteristic curve (Area Under the Curve [AUC]), and differences between radiomic-only, clinical-only, and combined models were statistically assessed.</p><p><strong>Results: </strong>External validation revealed poor diagnostic performance for both our model and a previously published EGFR radiomics model (AUC ∼0.5). The clinical model alone achieved higher diagnostic accuracy (AUC 0.74). SMOTE-augmented models showed increased sensitivity but did not improve overall AUC compared to the clinical-only model. Changing EGFR mutation prevalence had minimal impact on AUC, challenging previous assumptions about the influence of sample imbalance on model performance.</p><p><strong>Conclusion: </strong>External validation failed to reproduce prior radiogenomics model performance, while clinical variables alone retained strong predictive value. SMOTE-based oversampling did not improve diagnostic accuracy, suggesting that, in EGFR prediction, radiomics may offer limited value beyond clinical data. Emphasis on robust external validation and data-sharing is essential for future clinical implementation of radiogenomic models.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144086954","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":"Utilizing Clinicopathological and Radiomic Features for Risk Stratification of Lung Cancer Recurrence.","authors":"Wai Lone J Ho, Nikolai Fetisov, Lawrence O Hall, Dmitry Goldgof, Matthew B Schabath","doi":"10.1016/j.acra.2025.04.062","DOIUrl":"https://doi.org/10.1016/j.acra.2025.04.062","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To predict recurrence risk in patients with surgically resected non-small cell lung cancer (NSCLC) using radiomic analysis and clinicopathological factors.</p><p><strong>Materials and methods: </strong>293 patients with surgically resected stage IA-IIIA NSCLC were analyzed. Patients were randomly stratified into development and test cohorts. The development cohort was further divided into training and validation subsets for feature selection and model building, then applied to the test cohort. Pre-treatment computed tomography were segmented and 107 pyRadiomics features were extracted from intratumoral and peritumoral regions. Feature selection was performed using the maximum relevance minimum redundancy algorithm and Lasso regression. Clinical covariates were selected using univariable Cox regression. Radiomic, clinical, and radiomic-clinical models were constructed using a logistic regression classifier and evaluated using area under the curve (AUC). Kaplan-Meier curves for 3-year recurrence-free survival were compared between high-risk and low-risk groups using the log-rank test.</p><p><strong>Results: </strong>20 percent of patients experienced recurrence within 3 years. The radiomic-clinical model (AUC 0.77) outperformed the radiomic, clinical, and TNM stage models (AUC 0.76, 0.71, and 0.70, respectively) on the test set. Recurrence risk was five times higher in the high-risk group than the low-risk group (p<0.01) after stratification with the radiomic-clinical model. The most important features were regional lymph node metastases, the \"GLDM Large Dependence Emphasis\" texture, and the \"Elongation\" shape feature.</p><p><strong>Conclusion: </strong>Radiomics analysis can be used in combination with clinicopathological features for effective recurrence risk stratification in patients with surgically resected NSCLC.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144086961","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}