Pub Date : 2025-11-14DOI: 10.1016/j.clinimag.2025.110675
Bryce D. Beutler , Janine Chan , Jay Tsuruda , Danny J. Wang , Cynthia T. Chin , Dakshesh B. Patel , Max Gevorkian , Priya Rajagopalan
The lumbosacral plexus is the human neural network that provides motor and sensory innervation to the pelvis and lower extremities. Injury to the lumbosacral plexus may result due to trauma, compression, diabetes, infection, inflammation, or demyelination. Patients may present with persistent or progressive pain, paresthesia, or paresis. A diagnosis can typically be established based on correlation of clinical history, physical examination, and electromyography. However, although conventional diagnostic modalities provide vital functional information pertaining to motor unit activity and muscle response, it is often challenging to localize lesions and identify specific structural or intrinsic causes of nerve injury. Magnetic resonance neurography harnesses high-resolution imaging to delineate pathology of the lumbosacral plexus and identify sources of nerve compression, fascicular thickening, and inflammation that may be implicated in peripheral neuropathy. In this narrative review, we detail the anatomy of the lumbosacral plexus, describe common and uncommon neuropathology, summarize our early-stage investigation of lumbosacral plexus imaging using ultra-high-field 7 Tesla MRI. In addition, we provide case examples from our institutions that illustrate the value of magnetic resonance neurography for diagnosing lumbosacral plexus pathology and developing individualized treatment strategies.
{"title":"The lumbosacral plexus: Anatomy, clinical imaging, and proof-of-concept insights with ultra-high-field 7 Tesla MRI","authors":"Bryce D. Beutler , Janine Chan , Jay Tsuruda , Danny J. Wang , Cynthia T. Chin , Dakshesh B. Patel , Max Gevorkian , Priya Rajagopalan","doi":"10.1016/j.clinimag.2025.110675","DOIUrl":"10.1016/j.clinimag.2025.110675","url":null,"abstract":"<div><div>The lumbosacral plexus is the human neural network that provides motor and sensory innervation to the pelvis and lower extremities. Injury to the lumbosacral plexus may result due to trauma, compression, diabetes, infection, inflammation, or demyelination. Patients may present with persistent or progressive pain, paresthesia, or paresis. A diagnosis can typically be established based on correlation of clinical history, physical examination, and electromyography. However, although conventional diagnostic modalities provide vital functional information pertaining to motor unit activity and muscle response, it is often challenging to localize lesions and identify specific structural or intrinsic causes of nerve injury. Magnetic resonance neurography harnesses high-resolution imaging to delineate pathology of the lumbosacral plexus and identify sources of nerve compression, fascicular thickening, and inflammation that may be implicated in peripheral neuropathy. In this narrative review, we detail the anatomy of the lumbosacral plexus, describe common and uncommon neuropathology, summarize our early-stage investigation of lumbosacral plexus imaging using ultra-high-field 7 Tesla MRI. In addition, we provide case examples from our institutions that illustrate the value of magnetic resonance neurography for diagnosing lumbosacral plexus pathology and developing individualized treatment strategies.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110675"},"PeriodicalIF":1.5,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145551799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-14DOI: 10.1016/j.clinimag.2025.110677
Yihao Xiong , Ke Deng , Futao Cui , Ling Li , Rong Tian , Yan Ren , Qianrui Li
Objectives
To examine the prevalence of inappropriate 18F-FDG PET/CT use in oncological patients and associated factors.
Methods
We analyzed 900 oncological 18F-FDG PET/CT scans that were performed between July 1 to December 31, 2022 in West China Hospital, China. The appropriateness of 18F-FDG PET/CT use was assessed by two nuclear medicine experts independently using the NCCN guidelines as the reference. We performed univariate and multivariate logistic regression analyses to examine the association between predefined factors and inappropriate 18F-FDG PET/CT use.
Results
The use of 29.8 % 18F-FDG PET/CT oncological scans was potentially inappropriate. Multivariable analysis found that having cancers located in the genitourinary system (OR 37.1, 95 % CI 16.9 to 81.3) and the digestive system (OR 23.0, 95 % CI 11.2 to 47.0) generated higher odds of potentially inappropriate 18F-FDG PET/CT use, comparing to those in the circulatory system. Patients imaged for the purpose of diagnosis (OR 4.0, 95 % CI 2.2 to 7.1) or others (a combination of follow-up and searching for primary disease; OR 68.9, 95 % CI 23.7 to 200.2) generated higher odds of potentially inappropriate use, comparing to screening. Age and gender of patients, inpatient or outpatient setting, and specialty of scan prescribing physician were not significantly associated with inappropriate 18F-FDG PET/CT use.
Conclusion
Potentially inappropriate 18F-FDG PET/CT use was observed in real-world practice. The site of cancer and purpose of imaging was associated with inappropriate practice, particularly in indications with insufficient evidence. Developing high-quality evidence and guideline recommendations might be helpful to promote high-value imaging.
目的：探讨肿瘤患者不适当使用18F-FDG PET/CT的患病率及其相关因素。方法：我们分析了2022年7月1日至12月31日在中国华西医院进行的900例肿瘤18F-FDG PET/CT扫描。两位核医学专家以NCCN指南为参考，独立评估了18F-FDG PET/CT使用的适宜性。我们进行了单变量和多变量逻辑回归分析，以检验预定义因素与不适当的18F-FDG PET/CT使用之间的关系。结果：使用29.8%的18F-FDG PET/CT肿瘤扫描可能是不合适的。多变量分析发现，与循环系统相比，泌尿生殖系统（OR 37.1, 95% CI 16.9至81.3）和消化系统（OR 23.0, 95% CI 11.2至47.0）的癌症患者可能不适当使用18F-FDG PET/CT的几率更高。与筛查相比，以诊断为目的（OR 4.0, 95% CI 2.2至7.1）或其他目的（随访和寻找原发疾病的结合；OR 68.9, 95% CI 23.7至200.2）的患者出现潜在不当使用的几率更高。患者的年龄和性别，住院或门诊情况，以及扫描处方医师的专业与不适当的18F-FDG PET/CT使用无显著相关。结论：18F-FDG PET/CT在实际应用中存在潜在的不当使用。肿瘤部位和影像学目的与不适当的做法有关，特别是在证据不足的适应症中。制定高质量的证据和指南建议可能有助于促进高价值成像。
{"title":"Inappropriate use of oncological 18F-FDG PET/CT scans: A cross-sectional analysis of prevalence and associated factors","authors":"Yihao Xiong , Ke Deng , Futao Cui , Ling Li , Rong Tian , Yan Ren , Qianrui Li","doi":"10.1016/j.clinimag.2025.110677","DOIUrl":"10.1016/j.clinimag.2025.110677","url":null,"abstract":"<div><h3>Objectives</h3><div>To examine the prevalence of inappropriate <sup>18</sup>F-FDG PET/CT use in oncological patients and associated factors.</div></div><div><h3>Methods</h3><div>We analyzed 900 oncological <sup>18</sup>F-FDG PET/CT scans that were performed between July 1 to December 31, 2022 in West China Hospital, China. The appropriateness of <sup>18</sup>F-FDG PET/CT use was assessed by two nuclear medicine experts independently using the NCCN guidelines as the reference. We performed univariate and multivariate logistic regression analyses to examine the association between predefined factors and inappropriate <sup>18</sup>F-FDG PET/CT use.</div></div><div><h3>Results</h3><div>The use of 29.8 % <sup>18</sup>F-FDG PET/CT oncological scans was potentially inappropriate. Multivariable analysis found that having cancers located in the genitourinary system (OR 37.1, 95 % CI 16.9 to 81.3) and the digestive system (OR 23.0, 95 % CI 11.2 to 47.0) generated higher odds of potentially inappropriate <sup>18</sup>F-FDG PET/CT use, comparing to those in the circulatory system. Patients imaged for the purpose of diagnosis (OR 4.0, 95 % CI 2.2 to 7.1) or others (a combination of follow-up and searching for primary disease; OR 68.9, 95 % CI 23.7 to 200.2) generated higher odds of potentially inappropriate use, comparing to screening. Age and gender of patients, inpatient or outpatient setting, and specialty of scan prescribing physician were not significantly associated with inappropriate <sup>18</sup>F-FDG PET/CT use.</div></div><div><h3>Conclusion</h3><div>Potentially inappropriate <sup>18</sup>F-FDG PET/CT use was observed in real-world practice. The site of cancer and purpose of imaging was associated with inappropriate practice, particularly in indications with insufficient evidence. Developing high-quality evidence and guideline recommendations might be helpful to promote high-value imaging.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110677"},"PeriodicalIF":1.5,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145558174","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-14DOI: 10.1016/j.clinimag.2025.110676
Michelle Mai , Grayson Baird , Eric Dietsche , Terrence Healey , Michael Atalay , Saurabh Agarwal
Purpose
This study evaluates the utility of gray-scale inversion imaging for the detection of pulmonary nodules on chest radiography, which are often difficult to appreciate in traditional imaging.
Approach
100 paired chest radiographs and CT scans was provided to six radiologists (three attending and three residents) in a within-subjects study design. Each radiologist analyzed the radiographs with and without gray-scale inversion. Detection rates of pulmonary nodules were compared across conditions. Sensitivity and specificity were analyzed using statistical tests to determine differences based on training levels.
Results
We found that gray-scale inversion did not significantly increase the sensitivity of pulmonary nodule detection. However, it did enhance specificity, particularly among attending radiologists (81.3 (95 % CI [73.2, 87.4]) from 73.1 (95 % CI [63.6, 80.9]), p = 0.0277). No significant improvements were observed in residents.
Conclusions
Gray-scale inversion imaging may not enhance sensitivity but offers potential benefits in specificity, especially for experienced radiologists. The findings suggest that gray-scale inversion could serve as a useful educational tool for improving radiographic interpretation skills, particularly in differentiating between true and false positives. Further research is recommended to explore the impact of training and experience on radiographic interpretation using advanced imaging techniques.
目的：本研究评估灰度反转成像在胸部x线摄影中检测肺结节的效用，这在传统成像中往往难以识别。方法：在受试者内研究设计中，向6名放射科医生（3名主治医生和3名住院医生）提供100张配对的胸部x线片和CT扫描。每位放射科医生分析了有灰度反转和没有灰度反转的x线片。比较不同条件下肺结节的检出率。敏感度和特异度分析采用统计检验，以确定基于训练水平的差异。结果：灰度倒置对肺结节检测的敏感性无明显提高。然而，它确实提高了特异性，特别是在主治放射科医生中(81.3 （95% CI[73.2, 87.4])，从73.1 (95% CI [63.6, 80.9]), p = 0.0277）。在居民中没有观察到明显的改善。结论：灰度反转成像可能不会提高灵敏度，但对特异性有潜在的好处，特别是对经验丰富的放射科医生。研究结果表明，灰度反演可以作为一种有用的教育工具，用于提高放射成像解释技能，特别是在区分真阳性和假阳性方面。进一步的研究建议探讨培训和经验对使用先进成像技术的放射影像解释的影响。
{"title":"Diagnostic value of grayscale inversion imaging for detecting difficult cases of pulmonary nodules","authors":"Michelle Mai , Grayson Baird , Eric Dietsche , Terrence Healey , Michael Atalay , Saurabh Agarwal","doi":"10.1016/j.clinimag.2025.110676","DOIUrl":"10.1016/j.clinimag.2025.110676","url":null,"abstract":"<div><h3>Purpose</h3><div>This study evaluates the utility of gray-scale inversion imaging for the detection of pulmonary nodules on chest radiography, which are often difficult to appreciate in traditional imaging.</div></div><div><h3>Approach</h3><div>100 paired chest radiographs and CT scans was provided to six radiologists (three attending and three residents) in a within-subjects study design. Each radiologist analyzed the radiographs with and without gray-scale inversion. Detection rates of pulmonary nodules were compared across conditions. Sensitivity and specificity were analyzed using statistical tests to determine differences based on training levels.</div></div><div><h3>Results</h3><div>We found that gray-scale inversion did not significantly increase the sensitivity of pulmonary nodule detection. However, it did enhance specificity, particularly among attending radiologists (81.3 (95 % CI [73.2, 87.4]) from 73.1 (95 % CI [63.6, 80.9]), <em>p</em> = 0.0277). No significant improvements were observed in residents.</div></div><div><h3>Conclusions</h3><div>Gray-scale inversion imaging may not enhance sensitivity but offers potential benefits in specificity, especially for experienced radiologists. The findings suggest that gray-scale inversion could serve as a useful educational tool for improving radiographic interpretation skills, particularly in differentiating between true and false positives. Further research is recommended to explore the impact of training and experience on radiographic interpretation using advanced imaging techniques.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110676"},"PeriodicalIF":1.5,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145558144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To evaluate the diagnostic performance of Node-RADS for assessing regional lymph node metastasis (RLNM) in renal cell carcinoma (RCC) by computed tomography (CT) and explore its prognostic value.
Materials and methods
This retrospective study included 250 RCC patients who underwent nephrectomy and RLN dissection (2018–2024). Two radiologists independently evaluated lymph nodes using Node-RADS. Receiver operating characteristic (ROC) curves assessed predictive performance for RLNM. Inter- and intra-observer agreement was analyzed with Cohen's kappa. A nomogram integrating clinical variables and Node-RADS was constructed. Overall survival (OS) was compared between low (Node-RADS 1–3) and high (Node-RADS 4–5) score groups.
Results
Of 250 patients, 61 had RLNM. Subtype, T category, Fuhrman grade, distant metastasis, tumor size, systemic methods, and Node-RADS differed significantly between groups (all P < 0.05). AUC values for Node-RADS cutoffs were as follows: ≥4 (0.88), ≥3 (0.86), ≥5 (0.83), and ≥ 2 (0.76). Inter-observer agreement was 0.75, and intra-observer agreement was 0.82 (both P < 0.001). The newly developed nomogram demonstrated superior predictive performance for RLNM status compared to Node-RADS score, T classification, and distant metastasis, with significantly higher AUC values (0.960 vs. 0.880, 0.810, and 0.700, respectively; all P < 0.001). High Node-RADS scores were associated with shorter median OS (24.2 months vs. not reached; P < 0.001).
Conclusions
CT-based Node-RADS shows high accuracy for RLNM prediction and provides clinically relevant prognostic information in RCC.
{"title":"Node-RADS score on CT imaging predicts lymph node metastasis and survival in renal cell carcinoma: A multicenter diagnostic study","authors":"Donghua Gu, Cheng Shen, Hua Zhu, Jiangang Chen, Zhibo Gu, Jing Chang, Bing Zheng","doi":"10.1016/j.clinimag.2025.110665","DOIUrl":"10.1016/j.clinimag.2025.110665","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the diagnostic performance of Node-RADS for assessing regional lymph node metastasis (RLNM) in renal cell carcinoma (RCC) by computed tomography (CT) and explore its prognostic value.</div></div><div><h3>Materials and methods</h3><div>This retrospective study included 250 RCC patients who underwent nephrectomy and RLN dissection (2018–2024). Two radiologists independently evaluated lymph nodes using Node-RADS. Receiver operating characteristic (ROC) curves assessed predictive performance for RLNM. Inter- and intra-observer agreement was analyzed with Cohen's kappa. A nomogram integrating clinical variables and Node-RADS was constructed. Overall survival (OS) was compared between low (Node-RADS 1–3) and high (Node-RADS 4–5) score groups.</div></div><div><h3>Results</h3><div>Of 250 patients, 61 had RLNM. Subtype, T category, Fuhrman grade, distant metastasis, tumor size, systemic methods, and Node-RADS differed significantly between groups (all <em>P</em> < 0.05). AUC values for Node-RADS cutoffs were as follows: ≥4 (0.88), ≥3 (0.86), ≥5 (0.83), and ≥ 2 (0.76). Inter-observer agreement was 0.75, and intra-observer agreement was 0.82 (both <em>P</em> < 0.001). The newly developed nomogram demonstrated superior predictive performance for RLNM status compared to Node-RADS score, T classification, and distant metastasis, with significantly higher AUC values (0.960 vs. 0.880, 0.810, and 0.700, respectively; all <em>P</em> < 0.001). High Node-RADS scores were associated with shorter median OS (24.2 months vs. not reached; <em>P</em> < 0.001).</div></div><div><h3>Conclusions</h3><div>CT-based Node-RADS shows high accuracy for RLNM prediction and provides clinically relevant prognostic information in RCC.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110665"},"PeriodicalIF":1.5,"publicationDate":"2025-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145497477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-04DOI: 10.1016/j.clinimag.2025.110662
Andrew Huang , Ali Pirasteh , Scott Reeder
Rationale and objectives
Cost often weighs heavily in a physician's decision to request a diagnostic imaging exam. Further, the reimbursement for shortened, i.e. focused or abbreviated MRI exams is currently undefined at a systematic level. The purpose of this study is to examine the reimbursement of common imaging exams, as a surrogate for cost, and its evolution over the past two decades to frame a discussion on the appropriate reimbursement for both focused and abbreviated MRI exams.
Materials and methods
Common indications from the American College of Radiology (ACR) Appropriateness Criteria were used to select common imaging exams. Centers for Medicare & Medicaid Services (CMS) fee schedules from 2003 through 2025 were reviewed, and reimbursement trends analyzed over time. Inflation correction was performed using correction factors provided by the Bureau of Labor Statistics.
Results
Reimbursement for all imaging exams decreased substantially from 2003 through 2025. This effect was most prominent for MRI with a mean decrease of $1004.33 (76.5 %), range $590.01–1629.61 (64.0–83.6 %) from peak reimbursement during this period until 2025. CT also experienced a large but relatively smaller decrease in mean reimbursements of $466.10 (69.1 %), range $61.13–858.56 (38.2–81.7 %). Ultrasound, radiography, and lung ventilation/perfusion imaging experienced comparatively lower decreases in reimbursement.
Conclusion
Large decreases in CMS reimbursement for many imaging modalities has occurred over the past two decades, disproportionately affecting MRI with reimbursement now comparable to alternative modalities. Reimbursement for “focused” MRI exams that are shorter, yet fully diagnostic and sufficient to answer specific clinical questions should not be reduced, considering the dramatic decrease in reimbursement that has already occurred in the past two decades.
{"title":"Evolution of Medicare/Medicaid imaging reimbursements from 2003 to 2025","authors":"Andrew Huang , Ali Pirasteh , Scott Reeder","doi":"10.1016/j.clinimag.2025.110662","DOIUrl":"10.1016/j.clinimag.2025.110662","url":null,"abstract":"<div><h3>Rationale and objectives</h3><div>Cost often weighs heavily in a physician's decision to request a diagnostic imaging exam. Further, the reimbursement for shortened, i.e. focused or abbreviated MRI exams is currently undefined at a systematic level. The purpose of this study is to examine the reimbursement of common imaging exams, as a surrogate for cost, and its evolution over the past two decades to frame a discussion on the appropriate reimbursement for both focused and abbreviated MRI exams.</div></div><div><h3>Materials and methods</h3><div>Common indications from the American College of Radiology (ACR) Appropriateness Criteria were used to select common imaging exams. Centers for Medicare & Medicaid Services (CMS) fee schedules from 2003 through 2025 were reviewed, and reimbursement trends analyzed over time. Inflation correction was performed using correction factors provided by the Bureau of Labor Statistics.</div></div><div><h3>Results</h3><div>Reimbursement for all imaging exams decreased substantially from 2003 through 2025. This effect was most prominent for MRI with a mean decrease of $1004.33 (76.5 %), range $590.01–1629.61 (64.0–83.6 %) from peak reimbursement during this period until 2025. CT also experienced a large but relatively smaller decrease in mean reimbursements of $466.10 (69.1 %), range $61.13–858.56 (38.2–81.7 %). Ultrasound, radiography, and lung ventilation/perfusion imaging experienced comparatively lower decreases in reimbursement.</div></div><div><h3>Conclusion</h3><div>Large decreases in CMS reimbursement for many imaging modalities has occurred over the past two decades, disproportionately affecting MRI with reimbursement now comparable to alternative modalities. Reimbursement for “focused” MRI exams that are shorter, yet fully diagnostic and sufficient to answer specific clinical questions should not be reduced, considering the dramatic decrease in reimbursement that has already occurred in the past two decades.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110662"},"PeriodicalIF":1.5,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145497545","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Large language models such as ChatGPT have shown potential in clinical reasoning and radiologic interpretation. Recent versions with image-analysing capabilities allow for combined visual and textual processing. This study aims to assess the performance and limitations of ChatGPT-4v and ChatGPT-4o in interpreting image-based multiple-choice questions from official national radiology board examinations, which are designed to reflect core radiologic scenarios.
Methods
This prospective study used 222 image-based multiple-choice official questions from a national radiology board examinations administered between 2020 and 2024. Questions were entered into ChatGPT-4v and ChatGPT-4o; Generated answers were compared to the official answer key. Accuracy was further analysed by radiologic subspecialty and the presence or absence of clinical information.
Results
ChatGPT-4o achieved a 59 % (130/222) success rate, while ChatGPT-4v achieved a 54 % (119/222), with both models underperforming relative to the board exam passing standard. No significant difference was found between the two versions (two-tailed P-value = 0.339). Analysis by subspecialty revealed that ChatGPT-4v had a similar success rate across all fields (p = 0.330), whereas the success rate of ChatGPT-4o varied significantly (p = 0.0009). Both models achieved significantly higher success rates on questions that included clinical information. ChatGPT-4v: 63.8 % (60/94) vs. 46.1 % (59/128), p = 0.0099; ChatGPT-4o: 67.0 % (63/94) vs. 52.3 % (67/128), p = 0.0384.
Conclusion
ChatGPT shows potential as a supportive diagnostic tool, but its accuracy remains below the standard required for board-level image interpretation. The variability across subspecialties highlights current limitations and underscores the need for further research before safe clinical integration.
{"title":"Limited performance of ChatGPT-4v and ChatGPT-4o in image-based core radiology cases","authors":"Romi Noy Achiron , Shmuel Kagasov , Rina Neeman , Tamar Peri , Chifra Fenton","doi":"10.1016/j.clinimag.2025.110663","DOIUrl":"10.1016/j.clinimag.2025.110663","url":null,"abstract":"<div><h3>Objective</h3><div>Large language models such as ChatGPT have shown potential in clinical reasoning and radiologic interpretation. Recent versions with image-analysing capabilities allow for combined visual and textual processing. This study aims to assess the performance and limitations of ChatGPT-4v and ChatGPT-4o in interpreting image-based multiple-choice questions from official national radiology board examinations, which are designed to reflect core radiologic scenarios.</div></div><div><h3>Methods</h3><div>This prospective study used 222 image-based multiple-choice official questions from a national radiology board examinations administered between 2020 and 2024. Questions were entered into ChatGPT-4v and ChatGPT-4o; Generated answers were compared to the official answer key. Accuracy was further analysed by radiologic subspecialty and the presence or absence of clinical information.</div></div><div><h3>Results</h3><div>ChatGPT-4o achieved a 59 % (130/222) success rate, while ChatGPT-4v achieved a 54 % (119/222), with both models underperforming relative to the board exam passing standard. No significant difference was found between the two versions (two-tailed <em>P</em>-value = 0.339). Analysis by subspecialty revealed that ChatGPT-4v had a similar success rate across all fields (<em>p</em> = 0.330), whereas the success rate of ChatGPT-4o varied significantly (<em>p</em> = 0.0009). Both models achieved significantly higher success rates on questions that included clinical information. ChatGPT-4v: 63.8 % (60/94) vs. 46.1 % (59/128), <em>p</em> = 0.0099; ChatGPT-4o: 67.0 % (63/94) vs. 52.3 % (67/128), <em>p</em> = 0.0384.</div></div><div><h3>Conclusion</h3><div>ChatGPT shows potential as a supportive diagnostic tool, but its accuracy remains below the standard required for board-level image interpretation. The variability across subspecialties highlights current limitations and underscores the need for further research before safe clinical integration.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110663"},"PeriodicalIF":1.5,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145566208","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-03DOI: 10.1016/j.clinimag.2025.110659
Ali A. Alabdullah , Ahmed Marey , Yupeng Li , Stuti Jha , Campbell Rogers , Jawdat Abdulla , Muhammad Umair
Purpose
To assess the clinical value of FFR-CT (Computed Tomography Fractional Flow Reserve) in guiding revascularization, predicting major adverse cardiac events (MACE), and conducting a cost-effectiveness analysis in patients with low-to-moderate risk coronary artery disease.
Methods
We identified 15 studies evaluating FFR-CT in patients with suspected or known coronary artery disease. Primary outcomes included revascularization rates and MACE occurrence, categorized into FFR-CT negative (FFR-CT > 0.8) and positive (FFR-CT ≤ 0.8) groups. A systematic review of the cost of FFR-CT was also performed.
Results
In 13,224 patients from 11 studies, the revascularization rate was significantly lower in the FFR-CT negative group (5.8 %) than the positive group (41.4 %) (OR 13.02, 95 % CI 7.08–23.94, p < 0.0001). Unplanned revascularization rates were not significantly different (p = 0.09). In 11,814 patients from seven studies, MACE events were lower in the FFR-CT negative group (0.9 %) than the positive group (2.6 %) (OR 3.62, 95 % CI 1.87–6.97, p = 0.0001). Cost analysis, based on 10 studies, suggested potential savings, though results were heterogeneous due to healthcare cost disparities and lack of long-term follow-up.
Conclusion
FFR-CT can predict MACE and revascularization needs, potentially improving outcomes and reducing costs. However, it was not effective in guiding unplanned revascularization, and further research is needed to confirm these findings in larger trials.
目的评价FFR-CT（计算机断层扫描分数血流储备）在指导低至中度危险冠心病患者血运重建、预测主要心脏不良事件（MACE）以及进行成本-效果分析方面的临床价值。方法我们收集了15项研究，评估FFR-CT在疑似或已知冠状动脉疾病患者中的应用。主要结局包括血运重建率和MACE发生率，分为FFR-CT阴性（FFR-CT≤0.8）组和阳性（FFR-CT≤0.8）组。对FFR-CT的成本也进行了系统回顾。结果在11项研究的13224例患者中，FFR-CT阴性组血运重建率（5.8%）明显低于阳性组（41.4%）（OR 13.02, 95% CI 7.08-23.94, p < 0.0001）。非计划血运重建率差异无统计学意义（p = 0.09）。在7项研究的11,814例患者中，FFR-CT阴性组的MACE事件（0.9%）低于阳性组（2.6%）（OR 3.62, 95% CI 1.87-6.97, p = 0.0001）。基于10项研究的成本分析显示了潜在的节省，尽管由于医疗成本差异和缺乏长期随访，结果不尽相同。结论ffr - ct可以预测MACE和血运重建需求，可能改善预后并降低成本。然而，它在指导计划外血运重建方面并不有效，需要进一步的研究来在更大规模的试验中证实这些发现。
{"title":"Clinical value and cost effectiveness of FFR-CT in guiding revascularization and predicting major adverse cardiac events: A meta-analysis","authors":"Ali A. Alabdullah , Ahmed Marey , Yupeng Li , Stuti Jha , Campbell Rogers , Jawdat Abdulla , Muhammad Umair","doi":"10.1016/j.clinimag.2025.110659","DOIUrl":"10.1016/j.clinimag.2025.110659","url":null,"abstract":"<div><h3>Purpose</h3><div>To assess the clinical value of FFR-CT (Computed Tomography Fractional Flow Reserve) in guiding revascularization, predicting major adverse cardiac events (MACE), and conducting a cost-effectiveness analysis in patients with low-to-moderate risk coronary artery disease.</div></div><div><h3>Methods</h3><div>We identified 15 studies evaluating FFR-CT in patients with suspected or known coronary artery disease. Primary outcomes included revascularization rates and MACE occurrence, categorized into FFR-CT negative (FFR-CT > 0.8) and positive (FFR-CT ≤ 0.8) groups. A systematic review of the cost of FFR-CT was also performed.</div></div><div><h3>Results</h3><div>In 13,224 patients from 11 studies, the revascularization rate was significantly lower in the FFR-CT negative group (5.8 %) than the positive group (41.4 %) (OR 13.02, 95 % CI 7.08–23.94, p < 0.0001). Unplanned revascularization rates were not significantly different (p = 0.09). In 11,814 patients from seven studies, MACE events were lower in the FFR-CT negative group (0.9 %) than the positive group (2.6 %) (OR 3.62, 95 % CI 1.87–6.97, p = 0.0001). Cost analysis, based on 10 studies, suggested potential savings, though results were heterogeneous due to healthcare cost disparities and lack of long-term follow-up.</div></div><div><h3>Conclusion</h3><div>FFR-CT can predict MACE and revascularization needs, potentially improving outcomes and reducing costs. However, it was not effective in guiding unplanned revascularization, and further research is needed to confirm these findings in larger trials.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110659"},"PeriodicalIF":1.5,"publicationDate":"2025-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145448694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-01DOI: 10.1016/j.clinimag.2025.110657
Afsaneh Lahooti , Somayeh Farahani , Timothy Stait-Gardner , Yves De Deene , Daniel Moses , Trang T. Pham , William S. Price
Purpose
Breast cancer diagnosis relies on early detection and accurate subtype identification. This meta-analysis evaluated Diffusion-weighted imaging (DWI) and Intravoxel incoherent motion (IVIM)-DWI techniques for differentiating breast lesions, focusing on sensitivity, specificity, and limitations, to address inconsistencies and enhance diagnostic reliability.
Methods
Data from PubMed, Scopus, Embase, and Google Scholar (January 2013–September 2024) were systematically searched to assess the diagnostic accuracy of IVIM for breast lesions. Studies meeting the inclusion criteria underwent QUADAS-2 evaluation. Statistical analyses included SROC curves, heterogeneity assessment, leave-one-out meta-analysis, and publication bias evaluation via R packages and MetaBayesDTA.
Results
A total of 442 articles were identified; 26 were included in the review, and 23 were included in the meta-analysis. The diffusion coefficient (D), perfusion fraction (f), and apparent diffusion coefficient (ADC) values were significantly different between benign and malignant lesions. Among the parameters, the ADC and D exhibited the highest pooled sensitivity at 0.84 (95% CI: [0.77, 0.89]) across 15 studies, and 0.84 (95% CI: [0.79, 0.88]) across 19 studies, respectively. Also, D parameter demonstrated the highest pooled specificity at 0.80 (95% CI: [0.73, 0.85]).
Conclusion
Our meta-analysis identified D and ADC as the most current effective parameters for distinguishing benign from malignant breast lesions, warranting further research for comprehensive validation. Also, we outlined key methodological considerations in IVIM-DWI breast imaging, including sample size determination, b-value selection, ROI delineation, and menstrual status. However, careful planning, including b-value optimization and precise ROI placement, improves reliability in heterogeneous lesions, but the effect of hormonal fluctuations remains unclear.
{"title":"Diagnostic accuracy and limitations of intravoxel incoherent motion diffusion-weighted imaging for differentiating breast tumors: A systematic review and meta-analysis","authors":"Afsaneh Lahooti , Somayeh Farahani , Timothy Stait-Gardner , Yves De Deene , Daniel Moses , Trang T. Pham , William S. Price","doi":"10.1016/j.clinimag.2025.110657","DOIUrl":"10.1016/j.clinimag.2025.110657","url":null,"abstract":"<div><h3>Purpose</h3><div>Breast cancer diagnosis relies on early detection and accurate subtype identification. This meta-analysis evaluated Diffusion-weighted imaging (DWI) and Intravoxel incoherent motion (IVIM)-DWI techniques for differentiating breast lesions, focusing on sensitivity, specificity, and limitations, to address inconsistencies and enhance diagnostic reliability.</div></div><div><h3>Methods</h3><div>Data from PubMed, Scopus, Embase, and Google Scholar (January 2013–September 2024) were systematically searched to assess the diagnostic accuracy of IVIM for breast lesions. Studies meeting the inclusion criteria underwent QUADAS-2 evaluation. Statistical analyses included SROC curves, heterogeneity assessment, leave-one-out meta-analysis, and publication bias evaluation via R packages and MetaBayesDTA.</div></div><div><h3>Results</h3><div>A total of 442 articles were identified; 26 were included in the review, and 23 were included in the meta-analysis. The diffusion coefficient (<em>D</em>), perfusion fraction (<em>f</em>), and apparent diffusion coefficient (ADC) values were significantly different between benign and malignant lesions. Among the parameters, the ADC and <em>D</em> exhibited the highest pooled sensitivity at 0.84 (95% CI: [0.77, 0.89]) across 15 studies, and 0.84 (95% CI: [0.79, 0.88]) across 19 studies, respectively. Also, <em>D</em> parameter demonstrated the highest pooled specificity at 0.80 (95% CI: [0.73, 0.85]).</div></div><div><h3>Conclusion</h3><div>Our meta-analysis identified <em>D</em> and ADC as the most current effective parameters for distinguishing benign from malignant breast lesions, warranting further research for comprehensive validation. Also, we outlined key methodological considerations in IVIM-DWI breast imaging, including sample size determination, <em>b</em>-value selection, ROI delineation, and menstrual status. However, careful planning, including <em>b</em>-value optimization and precise ROI placement, improves reliability in heterogeneous lesions, but the effect of hormonal fluctuations remains unclear.</div></div>","PeriodicalId":50680,"journal":{"name":"Clinical Imaging","volume":"129 ","pages":"Article 110657"},"PeriodicalIF":1.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145520982","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号