Academic Radiology最新文献

Rationale and objectives: In 2028, the American Board of Radiology will reinstate an oral certifying examination for diagnostic radiology, replacing the multiple-choice format introduced in 2013. This study evaluates program directors' perspectives and their programs' adaptations by surveying members of the Association of Program Directors in Radiology (APDR). The survey focuses on challenges, strategies, and curricular changes in response to the new examination format.

Materials and methods: A voluntary, anonymous 20-question survey was distributed to APDR members in February 2025. The survey included multiple-choice, Likert scale, and open-ended questions covering topics such as resident conferences, mock oral board exams, faculty development, support for external review courses, and anticipated challenges. Descriptive statistics and comparative analyses were performed.

Results: Of 262 APDR members, 87 responded (33.2%), representing residency programs across the country. 76% of program directors supported the transition to the oral boards. Those who personally took the oral boards viewed this change more favorably than those who did not (P = 0.048). To prepare residents, many programs plan to increase hot-seat sessions and decrease didactic and multiple-choice question-based conferences. Over half plan to offer mock oral board exams, primarily to senior residents. Top challenges identified were exam timing, lack of faculty experience with the oral format, and uncertainty about exam content. Faculty shortages and burnout were also noted as barriers to preparation.

Conclusion: Program directors generally support the transition to an oral certifying exam. However, addressing challenges related to exam timing, faculty experience, and content uncertainty will be essential.

RATIONALE AND OBJECTIVES: This study aimed to develop and validate a magnetic resonance imaging (MRI)-based habitat radiomics model to predict recurrence-free survival (RFS) in patients with nonmetastatic clear cell renal cell carcinoma (ccRCC) after surgical resection.

Materials and methods: A retrospective cohort of 630 patients with nonmetastatic ccRCC who underwent surgical resection at the First Medical Center of Chinese PLA General Hospital (2011-2019) was included. Preoperative T2-weighted imaging (T2WI) and contrast-enhanced corticomedullary phase (CP) MRI were used to cluster tumor voxels into homogeneous habitats via K-means algorithm based on signal intensity. Radiomic features were extracted from habitats; after feature selection, these features were integrated with clinicopathological indicators to build a Cox proportional hazards regression model. Model performance was assessed via receiver operating characteristic (ROC) curves, concordance index (C-index), calibration curves, and decision curve analysis (DCA).

Results: Three distinct tumor habitat regions were identified through clustering, from which 13 recurrence-related radiomic features were selected to construct a Habitat Signature (HS). Multivariate Cox regression analysis demonstrated that age (HR = 1.039, 95% CI: 1.016-1.063, P<0.001), sex (male vs female, HR = 2.608, 95% CI: 1.291-5.270, P=0.008), and pathological T stage (T3 vs T1, HR = 4.284, 95% CI: 1.997-9.193, P < 0.001) served as independent predictors of postoperative recurrence. Constructed by combining clinicopathological predictors with the HS score, the clinical-habitat combined model yielded AUC values for 3-year and 5-year postoperative recurrence prediction of 0.80/0.81 in the training set and 0.85/0.81 in the test set, along with C-indices of 0.80 and 0.81, respectively.

Conclusion: The predictive model constructed by combining MRI-based HS score and clinical-pathological features has predictive value for recurrence of nonmetastatic ccRCC.

Rationale and objectives: Accurately predicting tumor response to neoadjuvant immunochemotherapy (NICT) and patient prognosis remains challenging in esophageal cancer. This study aimed to explore the value of ¹⁸F-FDG PET/CT in predicting the pathological response and prognosis of patients with resectable esophageal squamous cell carcinoma (ESCC) undergoing NICT.

Materials and methods: Patients who underwent NICT between January 2020 and April 2024 were retrospectively analyzed. ¹⁸F-FDG PET/CT scans were performed before (scan-1) and after NICT (scan-2). Parameters derived from ¹⁸F-FDG PET/CT and enhanced CT were analyzed for response evaluation and survival prediction. The pathological tumor regression grade served as the gold standard for response evaluation.

Results: Among the 94 patients, 41 patients (43.6%) achieved a major pathological response (MPR). ¹⁸F-FDG PET/CT identified more responders than CT. Compared with non-MPR patients, those achieving MPR demonstrated significantly lower uptake on scan-2 and a greater relative reduction (Δ%) between scan-1 and scan-2. The SUV_max of scan-2 demonstrated the best predictive performance for MPR (AUC = 0.829). The SUL_peak of scan-2 showed the highest sensitivity for predicting MPR (90.2%). Multivariate analysis indicated that CPS, SUV_max-2, and ΔSUV_max% were independent predictors of MPR, while pathological stage, PERCIST, and TLG-2 were independent predictors of PFS; further, the pathological stage, PERCIST, and ΔMTV% were independent predictors of OS. Patients with TLG-2 < 8.1 or ΔMTV% > 75.5% indicated better treatment response and longer survival.

Conclusion: Parameters after NICT and their changes before and after treatment were valuable in identifying patients achieving MPR and predicting prognosis. ¹⁸F-FDG PET/CT is a potentially valuable method for predicting the pathological response to NICT and the prognosis of resectable ESCC.

Take-home message: Metabolic parameters after NICT and their changes from baseline were valuable for identifying MPR and predicting prognosis. Post-treatment SUV_max and SUL_peak may serve as effective predictors of MPR, while a favorable PERCIST or lower TLG-2 relates to improved PFS, and a favorable PERCIST or higher ΔMTV% correlates with prolonged OS.

Rationale and objectives: To clinically implement and evaluate a 4D diamond-shaped pseudo-golden angle stack-of-stars acquisition with k-space weighted image contrast reconstruction for breast dynamic contrast-enhanced MRI (4D-DCE), assessing image quality, diagnostic confidence, and Breast Imaging Reporting and Data System (BI-RADS) agreement across conventional and ultrafast protocols.

Material and methods: This retrospective study included female patients who underwent breast MRI at 3 T using the 4D-DCE sequence. Three protocol types were generated: (1) conventional (4 × 60 s), (2) ultrafast (20 × 3 s), and (3) combined (ultrafast followed by 3 × 60 s). Two readers independently or in consensus rated image quality (overall quality, artifacts, sharpness, lesion conspicuity, and morphology) and diagnostic confidence using a 5-point Likert scale. BI-RADS scores were compared to the final reference standard (histology or ≥2-year imaging follow-up). Agreement was assessed using Cohen's kappa.

Results: A total of 167 patients (mean age: 59 ± 11 years) were included. Despite high temporal resolution of the ultrafast 4D-DCE, image quality was good to excellent and was comparable to the standard-resolution post-contrast T1 mDixon sequence (e.g., overall quality: 4.8 ± 0.4 vs. 4.8 ± 0.3, P =0.99). The combined 4D-DCE protocol yielded the highest diagnostic confidence by BI-RADS assignment in both readers, with the most pronounced improvement observed in patients with high background parenchymal enhancement (e.g., reader 1: 3.2 ± 0.6 [conventional] vs. 4.2 ± 0.4 [ultrafast] vs. 4.9 ± 0.3 [combined], P <0.001). BI-RADS agreement with the final reference standard was good to excellent across all DCE protocols, with the highest agreement achieved using the combined 4D-DCE (e.g., reader 1: κ = 0.89, 95% CI: 0.84-0.95).

Conclusion: The proposed 4D-DCE technique enables robust breast DCE-MRI with high spatial and temporal resolution. Combining ultrafast and conventional acquisitions within a single protocol improves diagnostic confidence and BI-RADS agreement.

Rationale and objectives: The aim of this study was to assess the prognostic value of MIBG metastasis patterns identified by ¹²³I-Metaiodobenzylguanidine (MIBG) single-photon emission computed tomography/computed tomography (SPECT/CT) imaging in pediatric patients with high-risk stage 4 neuroblastoma after induction therapy.

Materials and methods: A retrospective analysis was performed on a cohort of 77 pediatric patients with high-risk stage 4 neuroblastoma who underwent induction therapy followed by ¹²³I-MIBG SPECT/CT imaging at our institution. Progression-free survival (PFS) and survival (OS) were estimated using the Kaplan-Meier method, and differences in survival outcomes were assessed using the log-rank test. Categorical variables were analyzed using the chi-square test. Univariate and multivariate Cox proportional hazards regression models were employed to identify independent risk factors associated with diseases recurrence.

Results: All children were followed up for a median duration of 1.5 years. Among the 77 children with high-risk stage 4 neuroblastoma after induction therapy, 41 experienced endpoint events (53.2%), including 29 cases of disease recurrence or progression (37.6%) and 12 deaths attributable to ineffective treatment or treatment-related complications (15.6%). Univariate survival analysis revealed that patients with diffuse bone metastasis (n=32) exhibited a significantly lower 3-year PFS rates (11.5%±6.8%) compared to those with focal bone metastasis (56.8%±8.8%) (n = 45). (P<0.05). Further analysis demonstrated that the 3-year PFS and 3-year OS were significantly reduced in patients with skull metastasis (n=38) and axial bone metastasis (n=47) relative to those without such involvement (P<0.05). Multivariate Cox regression analysis identified axial bone metastasis, diffuse bone metastasis, Curie score>2, MYCN amplification, and 11q23 deletion as independent prognostic factors significantly associated with poorer prognosis (P<0.05).

Conclusion: The presence of diffuse systemic metastasis, axial bone involvement, or skull metastasis is strongly associated with poorer prognosis in pediatric patients with high-risk stage 4 neuroblastoma. Furthermore, axial bone metastasis, diffuse bone metastasis, Curie score>2, MYCN amplification, and 11q23 deletion constitute independent predictors of unfavorable clinical outcomes.

Background: Ultra-low-field magnetic resonance imaging (ULF-MRI) is highly promising for extremity musculoskeletal (MSK) imaging due to its portability, cost-effectiveness, and rapid deployability. Nevertheless, fat signal hyperintensity in ULF-MRI images often obscures pathological information, and the Dixon method used in our prior study has limitations in ULF settings, including reliance on a priori phase information and increased scan duration.

Methods: To address these challenges, we propose an improved two-point Dixon method for efficient water-fat separation in extremity imaging at 0.05 T. The optimizations include incorporation of T₂* correction to compensate for signal decay induced by long echo time and the use of a region-growing-based algorithm to eliminate dependence on prior information. Additionally, both multi-echo spin-echo and gradient-echo sequences were implemented to enable single-scan data acquisition. Experiments were conducted on a home-built 0.05 T MRI scanner, including phantom evaluations and in vivo extremity imaging.

Results: For phantom evaluation, the scan time was approximately 3 min, and the fat fraction values for Phantom 1 and Phantom 2 were 0.94 ± 0.02 and 0.91 ± 0.02, respectively. For in vivo imaging, the scan time was about 13 min, and clear water-only and fat-only images were successfully generated, which effectively distinguished muscle and fat tissues. Comparative results at 3 T demonstrated close agreement, supporting the validity of the proposed approach at 0.05 T.

Conclusion: These findings suggest a feasible solution for robust fat suppression in ULF-MRI, extend the application potential of ULF systems for quantitative diagnosis of skeletal muscle injury and bedside follow-up of osteoarticular diseases, and lay the groundwork for future clinical studies on MSK disease diagnosis.

Objective: To evaluate the value of combining high-frequency ultrasound, sound touch elastography (STE), and ultramicro angiography (UMA) for the diagnosis and short-term postoperative follow-up of carpal tunnel syndrome (CTS) patients.

Methods: This prospective study enrolled 50 CTS patients (78 wrists), classified by severity into mild (n = 22), moderate (n = 32), and severe (n = 24) groups, along with 35 healthy volunteers (70 wrists) as controls. All participants underwent multimodal ultrasound examination combining high-frequency ultrasound, STE, and UMA to measure median nerve cross-sectional area (CSA), shear wave velocity (SWV), and color pixel percentage (CPP) within 2 cm proximal to the carpal tunnel inlet. Patients with moderate and severe CTS underwent carpal tunnel release surgery followed by repeat ultrasound evaluation at 3 months postoperatively. Statistical comparisons included the following: (1) ultrasound parameters between preoperative CTS patients and controls; (2) parameter differences across severity subgroups; (3) pre-versus postoperative changes in CTS patients; and (4) diagnostic performance of individual and combined parameters for CTS identification.

Results: Significant elevations in median nerve CSA, SWV, and CPP were observed at the carpal tunnel inlet in the CTS group compared to controls (all P < 0.001). Among CTS patients, CPP demonstrated progressive increases with severity across all subgroups (all P < 0.001). CSA was significantly larger in severe cases than in mild and moderate wrists (P < 0.001), while SWV values were higher in both moderate and severe cases compared to mild CTS (all P < 0.001). Receiver operating characteristic (ROC) analysis indicated that the combination of CSA, SWV, and CPP provided optimal diagnostic accuracy for CTS with area under the curve (AUC) of 0.990, with 96.2% sensitivity and 98.6% specificity. Postoperative assessment at 3 months revealed significant reductions in all three parameters compared to preoperative values (all P < 0.001).

Conclusion: Multimodal ultrasound provides clinical value in carpal tunnel syndrome by supporting early diagnosis and severity stratification, and it may serve as a useful tool for short-term postoperative monitoring.

Rationale and objectives: Rapid advancements in multimodal large language models (LLMs) highlight their expanding potential in radiology education and assessment. This study aims to evaluate and compare the performance of three state-of-the-art LLMs; GPT-5, Gemini 2.5 Pro, and Claude 4.5 Sonnet, using a complete, retired version of the European Diploma in Radiology (EDiR) examination.

Materials and methods: The official EDiR examination, consisting of 78 Multiple Response Questions, 24 Short Cases, and 10 Clinically Oriented Reasoning Evaluation (CORE) cases, was administered to the models under standardized, zero-shot conditions. Inputs included text, static images, and videos (MP4 or sequential frames). Responses were scored according to official European Board of Radiology criteria. Passing thresholds were defined based on the historical human cohort mean for both the Weighted Written Score and the CORE component.

Results: In the written component (pass mark 50.9%), Gemini 2.5 Pro achieved the highest score (72.6%), followed by GPT-5 (67.3%), with both models surpassing the passing threshold. Claude 4.5 Sonnet scored 47.5%, failing to pass. In the CORE component (pass mark 55%), GPT-5 (62.5%) and Gemini 2.5 Pro (56.3%) successfully passed, whereas Claude 4.5 Sonnet (50.7%) did not. While models demonstrated high proficiency in text-dominant and static image interpretation, performance dropped in tasks requiring specific coordinate localization and dynamic video interpretation.

Conclusion: GPT-5 and Gemini 2.5 Pro successfully met the passing criteria for the EDiR examination, demonstrating advanced capabilities suitable for educational augmentation. However, persistent limitations in spatial localization and temporal reasoning highlight that while semantic processing has matured, clinical visual grounding remains a challenge for autonomous deployment.

Rationale and objectives: Accurate staging of hepatic fibrosis is essential for guiding immunosuppressive and antifibrotic therapies. However, percutaneous liver biopsy, the current reference standard, remains invasive and is subject to sampling errors and interobserver variability. To address these limitations, we developed and validated a noninvasive deep learning model using routine two-dimensional B-mode ultrasound for fibrosis staging in patients with autoimmune liver disease (AILD).

Materials and methods: We retrospectively enrolled 245 consecutive patients with AILD and randomly assigned them to the training set (60%), validation set (20%), and internal testing set (20%). Additionally, 61 biopsy-confirmed patients with AILD from another hospital were recruited as an external testing set. A deep learning model was constructed using the ResNet34 network architecture based on two-dimensional B-mode ultrasound images to evaluate its performance in liver fibrosis staging. Model performance was assessed using metrics such as macro- and microarea under the curve (AUC). Calibration curves and decision curves were employed to evaluate model goodness-of-fit and clinical utility, and class activation mapping was used for model interpretation.

Results: The model demonstrated robust performance across different datasets. In the internal and external test sets, the macroaverage AUCs were 0.812 (0.692-0.901) and 0.801 (0.688-0.902), respectively, while the microaverage AUCs were 0.819 (0.717-0.900) and 0.847 (0.761-0.911), respectively. The calibration and decision curves indicated favorable goodness-of-fit and clinical utility, and the class activation maps revealed the model's decision-making rationale, enhancing interpretability.

Conclusions: The model demonstrated robust diagnostic performance for the noninvasive staging of hepatic fibrosis in patients with AILD.