Background Accurate protocoling is critical for imaging accuracy. Manual protocoling is time-consuming and error prone. Purpose To evaluate the performance of large language models (LLMs) in automatically assigning protocols for abdominal and pelvic CT scans after optimization with context engineering and fine-tuning and to compare performance with that of radiologists in practice. Materials and Methods This retrospective study included patients with abdominal or pelvic CT scans obtained between January 2024 and June 2024. Requisition data, human-selected protocol, and training level (resident, fellow, or radiologist) were extracted. Reference standard protocols were defined by radiologists in consultation with institutional guidelines. Context engineering involved detailed prompt instructions using a prompt set with GPT-4o (version 2024-08-06; Open AI). A subset of patients was reserved for fine-tuning (training set and validation set) and another for testing (internal test set). Two models were tested (prompting-only and fine-tuned). Model-selected protocols and original human-selected protocols were categorized compared with the reference standard after review by blinded radiologists as follows: exact match, equal alternative, reasonable but inferior, or inappropriate. Exact match and equal alternative were considered optimal. Performance of models and radiologists were compared using the McNemar test. Results This study included 1448 patients (mean age, 61 years ± 17 [SD]; 728 female patients). GPT-4o with prompting only selected optimal protocols more frequently than humans (96.2% [527 of 548 patients] vs 88.3% [484 of 548 patients]; P < .001), but there was no evidence of a difference in inappropriate protocols (1.3% [seven of 548 patients] vs 2.4% [13 of 548 patients]; P = .21). Fine-tuning GPT-4o did not improve the proportion of optimal protocols over prompting only (96.2% [527 of 548 patients] vs 96.2% [527 of 548 patients]; P > .99). In subgroup analyses, the proportion of protocols matching the reference standard was similar among radiologists (79.4% [173 of 218 patients]), fellows (74.9% [164 of 219 patients]), and residents (72.1% [80 of 111 patients]; P = .30). Conclusion For protocoling abdominal and pelvic CT scans, the LLM, GPT-4o, selected optimal protocols more frequently than radiologists when optimized with detailed prompting, and fine-tuning of the model did not further improve performance. © RSNA, 2026 Supplemental material is available for this article.

Purpose To compare published CT-based systems for small solid renal mass (SoRM) assessment, propose modifications that may increase specificity and interreader agreement, and validate the revised system. Materials and Methods Our retrospective study included patients with histologically confirmed SoRMs measuring ≤4 cm who underwent CT imaging (single-institution internal dataset, n = 194; external dataset from The Cancer Imaging Archive, n = 55). Two blinded radiologists (readers 1 [R1] and 2 [R2]) compared four CT systems (CT score, modified CT score, abbreviated CT score, and UCLA CT score) for diagnostic accuracy in clear cell renal cell carcinoma (ccRCC) and papillary RCC (pRCC) and for interreader agreement (Gwet agreement coefficient [AC1]). We also evaluated the addition of two decision rules to the best-performing algorithm (noncontrast CT [NCCT] attenuation ≤ 20 HU and corticomedullary phase-NCCT attenuation at two thresholds, ≤20 HU and ≤30 HU) to create a modified algorithm (CT-Score version 2.0). Results The abbreviated CT score had the best combination of accuracy for ccRCC (R1: 85% [95% CI: 79, 89], R2: 72% [95% CI: 65, 78]) and pRCC (R1: 86% [95% CI: 80, 91], R2: 86% [95% CI: 80, 91]) and interreader agreement (Gwet AC1 = 0.53). CT-Score version 2.0 (derived by adding decision rules to the abbreviated CT score) demonstrated substantial agreement (Gwet AC1 = 0.63). Specificity of CT-Score version 2.0 was higher for ccRCC (R1: 99% [95% CI: 94, 100], R2: 99% [95% CI: 94, 100] vs R1: 92% [95% CI: 84, 96], R2: 81% [95% CI: 72, 89]; P = .02, P < .001) and pRCC (R1: 100% [95% CI: 98, 100], R2: 100% [95% CI: 98, 100] vs R1: 93% [95% CI: 87, 96], R2: 93% [95% CI: 87, 96]; P = .003, P = .003) when compared with the abbreviated CT score. Validation in the external dataset showed similar results: Gwet AC1 = 0.53; specificity for ccRCC (R1: 100% [95% CI: 83, 100], R2: 100% [95% CI: 83, 100]); and specificity for pRCC (R1: 100% [95% CI: 82, 100], R2: 100% [95% CI: 92, 100]). Conclusion Application of CT-Score version 2.0 resulted in modest improvements in interreader agreement and high specificity for ccRCC and pRCC diagnosis. Keywords: CT, Kidney, Urinary, Oncology, Renal Mass, Algorithm, Clear Cell RCC, Papillary RCC Supplemental material is available for this article. © RSNA, 2026.

Renal oncocytic neoplasms present diagnostic challenges, both at imaging and pathologic evaluation. The World Health Organization classification of renal neoplasms defines a spectrum of oncocytic neoplasms, including emerging entities that help define previously uncharacterized or mischaracterized tumors. Low-grade oncocytic tumors and eosinophilic vacuolated tumors are distinguishable from other oncocytic neoplasms at pathologic evaluation and typically demonstrate indolent behavior. Nomenclature regarding hybrid neoplasms has been clarified in reference to hereditary cases associated with Birt-Hogg-Dubé syndrome. Preoperative diagnostic difficulties at imaging contribute to high rates of resected benign renal tumors, the majority being renal oncocytomas. The imaging appearances of oncocytic neoplasms are similar, and the inability to confidently diagnose them at imaging has led to increased resection rates. Preoperative renal mass biopsy may be preventative, but its utilization remains low, diagnoses can be equivocal, and establishing tumor aggressiveness may not always be reliable. Malignant renal oncocytic tumors, including chromophobe renal cell carcinoma, are generally considered the less aggressive subtypes of renal cell carcinoma. However, distinguishing them from the more aggressive clear cell subtype remains challenging, despite imaging frameworks designed to aid categorization. Active surveillance is a safe management option among biopsy-confirmed renal oncocytic neoplasms, but it remains uncertain which patients are suitable for this approach. Diagnostic imaging may assist in risk-stratifying oncocytic neoplasms, with mass enhancement, heterogeneity, and calcification potentially differentiating benign from malignant oncocytic neoplasms. Mass attenuation and heterogeneity may differentiate low-grade and high-grade cancers. Molecular imaging and other emerging techniques, such as MR fingerprinting, may play a role in the future. ©RSNA, 2026 Supplemental material is available for this article.

INTRODUCTION: Automated segmentation using artificial intelligence (AI) has the potential to rapidly perform three-dimensional (3D) segmentation of small renal masses (SRM). The objective of this study was to test for clinically and statistically significant differences in time spent segmenting, accuracy, and reliability when comparing manual and automated segmentation of computed tomography (CT) scans with SRM. METHODS: Patients with a CT scan, SRM <4 cm, and renal neoplasm were identified through an institutional database. Of the 854 patients identified, 184 were excluded. Forty test cases were randomly selected. There were 630 cases for training (using nnU-Net) to which 488 cases from the KiTS23 open-source dataset were added. Each of the test cases was segmented by a radiologist, a urologist, and the AI model. Time to segment and Dice coefficients were compared. Deidentified segmented CTs were provided to two independent radiologists, who attempted to identify the segmentor and rated the acceptability of the segmented images on a five-point Likert scale. RESULTS: There were 39 cases with complete timing data. The median time for the AI model to segment was one-third of the radiologist's (152.4 s, interquartile range [IQR] 120.9-177.8 vs. 450 s, IQR 318.8-551.2) and about one-fifth of the urologist's (800.0 s, IQR 492.0-1538.0) (p<0.001). There was a high degree of inter-rater reliability (median Dice coefficients 0.86-0.90, p=0.09). The scoring radiologists were able to correctly identify the true segmentor in 61.6% of cases (p <0.001). The AI segmentations were scored highest among the three segmentors (median score 4.1/5, standard deviation [SD] 1.0) compared to 3.8 (SD 0.7) for the radiologist, and 3.3 (SD 0.7) for the urologist. CONCLUSIONS: Automated segmentation of CT scans for patients with SRM was efficient, accurate, and acceptable in this study. This approach has the potential to greatly improve the clinical use of radiomics to assess medical images for these patients.

OBJECTIVES:Local tumour progression (LTP) after percutaneous ablation of small renal cell carcinoma (RCC) is suspected when new enhancing or enlarging soft tissue appears within the ablation zone. Benign post-treatment changes can mimic this finding. This study compares the incidence and imaging characteristics of non-malignant changes (NMC) versus LTP after renal ablation. MATERIALS AND METHODS:In this single-center, retrospective study, all patients with RCC treated with radiofrequency ablation (RFA) from February 2004 to May 2016 were identified. Post-ablation imaging reports from through May 2017 were reviewed to detect findings suspicious for LTP. Patients with suspicious findings underwent clinical, imaging, and histopathologic follow-up through May 2025 to determine the reference diagnosis. Imaging features were categorized by morphology, location within the ablation zone, and enhancement pattern. RESULTS:Among 256 patients (mean age 65.6 years ± 10.8, 193 men) with 268 treated tumours, 18 tumours (6.7%) developed suspicious imaging findings. Eight tumours (3.0%) were classified as NMC and 10 tumours (3.7%) as LTP. NMC had significantly lower CT enhancement than LTP (31 vs 152 HU, P < .001). Lesions along the renal parenchymal margin were exclusively associated with LTP (9/9), whereas abnormalities at the extrarenal margin or centrally within the ablation zone were predominantly NMC (8/9). Enhancement with washout was seen only in LTP. CONCLUSION:Non-malignant post-ablation changes can mimic LTP and occur with similar frequency. Imaging features can help differentiate benign changes from local tumour progression and reduce unnecessary re-interventions.

Fat-containing adrenal nodules are frequently encountered in clinical practice and most often represent benign entities, particularly adrenal adenomas or myelolipomas. Although most adrenal nodules are adenomas regardless of imaging appearance, accurate characterization of microscopic and macroscopic fat remains important because rare non-adenomatous and malignant lesions may also contain fat and can create diagnostic pitfalls in selected clinical contexts. Importantly, even when imaging features support a benign adrenal adenoma, hormonal evaluation may still be required because functional status cannot be determined by imaging alone. This review provides an imaging centered overview of fat-containing adrenal nodules on computed tomography (CT) and magnetic resonance imaging (MRI), with emphasis on distinguishing microscopic from macroscopic fat. Microscopic fat is most commonly associated with adrenal adenomas and can be detected using chemical-shift MRI or unenhanced CT attenuation measurements. Homogeneous signal loss on chemical-shift imaging strongly supports adenoma; however, emerging evidence indicates that certain heterogeneous signal loss patterns may also be compatible with benignity, particularly in patients without a history of malignancy. Macroscopic fat most commonly indicates adrenal myelolipoma, although small foci of macroscopic fat may rarely occur in other adrenal neoplasms due to myelolipomatous degeneration. Key imaging features, diagnostic pitfalls, and differential considerations are reviewed, including metastases from lipid-containing primary tumors, collision tumors, adrenocortical carcinoma and retroperitoneal mimics. A practical imaging-based framework integrating CT attenuation, chemical-shift MRI findings, fat proportion, lesion morphology, and clinical context is proposed to facilitate systematic evaluation of fat-containing adrenal nodules across incidental, oncologic and problem-solving scenarios. Integration of imaging findings with guideline-recommended biochemical evaluation and clinical context remains essential for accurate diagnosis and appropriate patient management.

Background Clinical information improves imaging interpretation, but physician-provided histories on requisitions for oncologic imaging often lack key details. Purpose To evaluate large language models (LLMs) for automatically generating clinical histories for oncologic imaging requisitions from clinical notes and compare them with original requisition histories. Materials and Methods In total, 207 patients with CT performed at a cancer center from January to November 2023 and with an electronic health record clinical note coinciding with ordering date were randomly selected. A multidisciplinary team informed selection of 10 parameters important for oncologic imaging history, including primary oncologic diagnosis, treatment history, and acute symptoms. Clinical notes were independently reviewed to establish the reference standard regarding presence of each parameter. After prompt engineering with seven patients, GPT-4 (version 0613; OpenAI) was prompted on April 9, 2024, to automatically generate structured clinical histories for the 200 remaining patients. Using the reference standard, LLM extraction performance was calculated (recall, precision, F1 score). LLM-generated and original requisition histories were compared for completeness (proportion including each parameter), and 10 radiologists performed pairwise comparison for quality, preference, and subjective likelihood of harm. Results For the 200 LLM-generated histories, GPT-4 performed well, extracting oncologic parameters from clinical notes (F1 = 0.983). Compared with original requisition histories, LLM-generated histories more frequently included parameters critical for radiologist interpretation, including primary oncologic diagnosis (99.5% vs 89% [199 and 178 of 200 histories, respectively]; P < .001), acute or worsening symptoms (15% vs 4% [29 and seven of 200]; P < .001), and relevant surgery (61% vs 12% [122 and 23 of 200]; P < .001). Radiologists preferred LLM-generated histories for imaging interpretation (89% vs 5%, 7% equal; P < .001), indicating they would enable more complete interpretation (86% vs 0%, 15% equal; P < .001) and have a lower likelihood of harm (3% vs 55%, 42% neither; P < .001). Conclusion An LLM enabled accurate automated clinical histories for oncologic imaging from clinical notes. Compared with original requisition histories, LLM-generated histories were more complete and were preferred by radiologists for imaging interpretation and perceived safety. © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Tavakoli and Kim in this issue.

PURPOSE:Oncocytic neoplasms account for up to 15% of small renal masses. Active surveillance (AS) is increasingly adopted for these lesions; however, knowledge gaps remain regarding their growth kinetics, concordance with renal tumor biopsy, and impact on renal function. This study reviews these factors in the largest cohort of biopsy-confirmed oncocytic neoplasms managed with AS. MATERIALS AND METHODS:This single-center, retrospective study included patients with biopsy-confirmed oncocytoma or chromophobe renal cell carcinoma (chRCC) on AS (2003-2021). Tumor growth rates and change in renal function (estimated glomerular filtration rate) were analyzed using linear mixed models, while development of chronic kidney disease was assessed using Cox proportional hazards model. Biopsy-to-surgical pathology concordance and triggers for intervention were examined. RESULTS:Among 229 patients (245 lesions; 222 oncocytomas, 23 chRCC), the median time between first and last imaging was 4.7 and 2.5 years, respectively. Predicted growth rates were 0.21 cm/y (oncocytoma) and 0.31 cm/y (chRCC). Larger baseline tumor size correlated with faster growth (P = .04), while age inversely correlated with growth rate (P < .01). Intervention was more common in the chRCC cohort (35% vs 13%). Biopsy-to-surgical pathology concordance rates were 80% for oncocytoma and 86% for chRCC. No significant association was found between tumor size and renal function decline (P = .3) or level (P = .6). There were no events of metastases or kidney cancer-related deaths over a median follow-up of 5.7 years (IQR: 3.0, 8.7). CONCLUSIONS:Oncocytic lesions on AS demonstrate slow growth, low intervention rate, preservation of renal function, and excellent disease-specific survival. Although biopsy limitations exist, histology-guided AS provides effective management.

Objectives To evaluate [18F]F-DCFPyL PET/MRI whole-gland-derived radiomics for detecting clinically significant (cs) prostate cancer (PCa) within the prostate gland and predicting extra-prostatic metastasis (N and M staging).Methods In this single-centre, retrospective study, therapy-na & iuml;ve PCa patients who underwent [18F]F-DCFPyL PET/MRI were included. Whole-prostate segmentation was performed. Feature extraction from each modality was done. The selection of potential variables was made through regularized binomial logistic regression. The oversampled training data were used to train binomial logistic regression for each outcome. The estimates of the models were calculated, and the mean accuracy was reported. The trained models were assessed on the test data for comparative evaluation of performance.Results A total of 103 patients (mean age = 65; mean PSA = 23.4) were studied. Among them, 89 had csPCa and 20 had metastatic disease. There were five radiomics variables selected for the International Society of Urological Pathology Grade Group (ISUP GG) >= 2 from T2w, ADC, and PET. To detect N1, five radiomics variables were selected from the T2w and PET. For M1, four radiomics variables were selected from T2w and ADC. Regarding the performance of models for the prediction of csPCa, the imaging-based hybrid model (T2w + PET) provided the highest AUC (0.98). The performance of N1 models showed the highest AUC (0.80) for T2w + PET. To predict M1, the T2w + ADC model showed the highest AUC (0.93).Conclusions Whole-gland PET/MRI radiomics may provide a reliable model to predict csPCa. Also, acceptable performance was reached for predicting metastatic disease in our limited population. Our findings may support the value of whole-gland radiomics for non-invasive csPCa detection and prediction of metastatic disease.Advances in knowledge Whole-gland PET/MRI radiomics, a less operator-dependent segmentation method, can be potentially used for treatment personalization in PCa patients.Trial Registration NCT03535831. Registered 2018; NCT03149861. Registered 2017

The evolution of modern medicine has been significantly driven by medical and health care research, underscoring the importance of disseminating findings to advance health care. Medical literature, encompassing various publication types such as case reports, review articles, and original research, plays a crucial role in this process by facilitating the communication and discussion of new discoveries. This review article provides a comprehensive guide to understanding and navigating radiologic publications. It examines the various types of radiologic research articles, including case reports and series, pictorial reviews, original research, systematic reviews, and meta-analyses, each of which serve distinct purposes in contributing to the field of radiology. The study adopts the “six honest men” approach—addressing why, who, what, when, where, and how—to elucidate the essential elements of successful radiology research and publication. Key topics include the motivations for publishing, the types of articles suited for different research questions, and strategic considerations for selecting appropriate journals. Additionally, the review highlights the importance of understanding publication timing, journal selection criteria, and the overall publication process, including manuscript preparation and peer review. By offering these insights, the review aims to equip early-career researchers with the knowledge and skills necessary to effectively contribute to radiology literature and advance their academic and professional careers.

BACKGROUND:Clear cell papillary renal cell tumour (CCPRCT) is a clinically indolent neoplasm. The diagnosis of CCPRCT at renal mass biopsy (RMB) may facilitate conservative management; however, this can be challenging in small samples and there are limited data on the outcome of patients managed conservatively. This study aimed to: (1) evaluate concordance between RMB and nephrectomy for the diagnosis of CCPRCT, and (2) determine the clinical outcomes of patients with CCPRCT, particularly those who have undergone conservative management. METHODS AND RESULTS:We reviewed all cases at our institution where CCPRCT was diagnosed or included in the differential diagnosis at RMB. In all, 65 RMB with a diagnosis/differential diagnosis of CCPRCT were identified, including 25 patients who had subsequent surgical resection. Of cases where CCPRCT was the likely or favoured diagnosis on RMB (n = 19), the final diagnosis was concordant in 16/19 (84%). Discordant cases comprised clear cell renal cell carcinoma (ccRCC) (n = 1) and RCC with fibromyomatous stroma (n = 2). In all, 15 patients were managed with active surveillance (median follow-up 34 months) and 26 were treated with ablation (median follow-up 37 months). There was no recurrence, metastasis, or death in the cohort. CONCLUSION:CCPRCT can be reliably diagnosed at RMB in most cases if strict morphologic and immunohistochemical (IHC) criteria are applied. Furthermore, there was no adverse behaviour among patients managed nonsurgically in our cohort. It is important, however, to recognize the limitations of RMB, particularly given that occasional examples of ccRCC can show foci resembling CCPRCT. Pathologists and treating clinicians should be aware of the limitations of RMB when considering conservative management.

Torsion of abdominal and pelvic organs, such as the spleen, gallbladder, omentum, epiploic appendages, fallopian tube, and epididymis/testicular appendix are relatively rare, and yet clinically significant, often imitating more common causes of acute pain. Given the nonspecific presentation and limited value of laboratory tests, imaging drives early recognition and timely, organ-preserving management. This review consolidates pathophysiology, modality-specific appearances, and practical diagnostic strategies for uncommon non-intestinal torsions. The unifying mechanism is rotation around a mesenteric or vascular stalk that first impedes venous outflow, then compromises arterial supply, culminating in edema, ischemia, and possible infarction. Predisposing conditions include congenital ligamentous or mesenteric insufficiency and acquired factors such as trauma, pregnancy, mass effect, and age-related laxity. Ultrasound may demonstrate organ enlargement, asymmetric morphology of paired organs, and, when captured, the twisted pedicle (‘whirlpool’ sign). Doppler findings may support the diagnosis but preserved arterial signals do not exclude torsion. CT allows clearest delineation of organ displacement, pedicle twisting, hypoenhancement, and inflammatory fat stranding. CT is especially helpful in diagnosing splenic and gallbladder torsion and in distinguishing omental torsion from epiploic appendagitis. MRI, as a radiation-sparing alternative, reliably depicts early ischemic change and variations of pelvic anatomy. Management varies from conservative therapy for epiploic appendage torsion to urgent detorsion or definitive resection for splenic, biliary and gonadal torsions. Familiarity with cross-sectional hallmarks, modality-appropriate workflows can shorten time to intervention and enhance organ salvage and outcomes.

Classification systems have been developed to categorize cystic and solid renal masses by likelihood of malignancy and pathologic subtype with MR imaging. While this information is useful, future directions of renal mass imaging will include methods to predict aggressive behavior of renal masses with the aim to improve patient management and disease-specific survival.

AIMS:Renal tumours with oncocytic morphology are among the most difficult to classify at renal mass biopsy (RMB), and a number of emerging entities with low-grade oncocytic morphology have been recently described. This study aimed to evaluate pathological concordance between RMB and subsequent nephrectomy or repeat biopsy for oncocytic renal neoplasms and to identify pathological factors contributing to diagnostic discordance, including the impact of evolving tumour classification. METHODS AND RESULTS:We retrospectively reviewed 145 cases of oncocytic renal neoplasms diagnosed on RMB, including 114 with subsequent nephrectomy and 31 with repeat biopsy only. Overall concordance was 92.9% between RMB and nephrectomy and 96.7% between initial and repeat RMB. Concordance for oncocytoma at nephrectomy was lower (81.4%), likely reflecting selection bias, but was 100% in cases with repeat biopsy. Review of discordant cases (n = 9) revealed that 55% (5/9) were reclassified as emerging tumour entities, specifically low-grade oncocytic tumour (LOT) and eosinophilic vacuolated tumour (EVT). Additional discordant cases were due to heterogeneous tumour morphology in chromophobe renal cell carcinoma (ChRCC) and incomplete immunohistochemical work-up leading to misclassification of rarer renal cell carcinoma subtypes. CONCLUSIONS:Despite inherent diagnostic challenges, there was overall good concordance between RMB and nephrectomy or subsequent biopsy for the diagnosis of oncocytic tumours. Recognition of emerging tumour entities may reduce diagnostic uncertainty, improve classification in challenging cases, and further improve diagnostic concordance over time. Nonetheless, limitations of RMB, particularly related to tumour heterogeneity, highlight the importance of integrating pathological, clinical, and radiologic data to inform patient management.

A hybrid large language model (LLM)–based application, optimized by combining LLM feature classification with deterministic elements, accurately assigned Ovarian-Adnexal Reporting and Data System MRI scores from adnexal lesion descriptions and outperformed originally reporting radiologists.