共找到 20 条结果
Photon counting computed tomography offers spectral capabilities on every scan, significant radiation exposure reductions with improved image quality, and improved spatial resolution. Spectral reconstructions, such as iodine maps and virtual non-contrast images improve the diagnostic capabilities of CT. In the emergency department, photon counting CT has proven benefits in assessing acute conditions, characterizing incidental masses and renal stones. In this review, we highlight some of the benefits of photon counting CT as it pertains to emergency abdominal imaging.
To analyze the intercontinental differences and inequities in the management of hepatocellular carcinoma (HCC), including surveillance, diagnosis, and treatment, through the survey responses from the participants of the Global Abdominal Imaging Forum on HCC, organized by the European Society of Gastrointestinal and Abdominal Radiology (ESGAR). An online anonymous survey was distributed to the attendees of the Global Abdominal Imaging Forum on HCC. The survey consisted of 14 multiple-choice questions, covering demographic, epidemiological and occupational data, and information on the management of HCC. Global differences and differences between European and non-European countries were analyzed. Of 1963 attendees from 110 countries, 408 (20.8%) from 67 countries responded to the survey. Multidisciplinary HCC meetings were reported with the highest frequency in North America (83.3%) and the lowest in Africa (41.2%). The most commonly reported tools for HCC surveillance were ultrasound (86.8%) and serum alpha-fetoprotein (79.4%). Global inequities were reported for contrast-enhanced ultrasound (p < 0.001, lowest access in South America, Africa, and Asia), MRI (p < 0.001, lowest access in Africa) and PET/CT (p < 0.001, lowest access in Africa). LI-RADS was the most common algorithm used for HCC diagnosis (83.6%), with the highest use reported in North America (94.4%). Heterogeneity in treatment availability was observed, with respondents from Africa reporting limited availability of surgery (41.2%), locoregional treatments (23.5%), liver transplantation (0%), and immunotherapy (17.6%). Although the surveillance strategy for HCC is similar at a global level, disparities exist in access to CEUS, MRI, PET/CT, and treatments for HCC. Question What are the global disparities in hepatocellular carcinoma surveillance, diagnosis, and treatment reported by doctors working in different continents? Findings The survey from the Global Abdominal Imaging Forum on HCC revealed significant global variability with disparities regarding access to advanced imaging modalities and treatments, particularly affecting Africa. Clinical relevance Global efforts should focus on improving access to advanced imaging techniques (including MRI and PET/CT) and treatments (including transplantation and immunotherapy). Guidelines for the management of HCC should consider the existing regional disparities in healthcare infrastructure and resource availability to ensure more equitable HCC care worldwide.
The inherently slow acquisition speed of MRI makes abdominal imaging highly sensitive to respiratory motion artifacts. Since the early days of MRI, the development of respiratory motion compensation techniques has been an active research topic, and this field has seen substantial progress. Despite these advances, the majority of these techniques are not used in daily clinical practice, and motion management methods used in clinical abdominal MRI today have changed little over the past decades. This observation points to a significant gap between technical innovation and clinical translation in this area. This review is motivated by this question: why have so many motion management techniques not been adopted into routine clinical workflows? Unlike conventional survey-style reviews that focus on summarizing emerging methods, this article takes a different, and perhaps opposite, perspective to investigate why those technologically sophisticated innovations are misaligned with practical clinical needs. Specifically, we discuss the barriers behind the gap between research advances and clinical practice, clarify the clinical requirements for effective respiratory motion management in abdominal MRI, and highlight research directions with stronger relevance to routine workflows. The review begins with an overview of the clinical impact of respiratory motion in abdominal MRI, followed by a discussion of standard abdominal MRI sequences and their motion sensitivity. We then summarize current clinical strategies and advanced approaches, along with the barriers that hinder their clinical adoption. The article concludes with future directions and broader lessons learned from this translational gap, with the goal of guiding future developments toward improved clinical integration.
To evaluate the consistency and clinical implications of radiology reporting for adrenal lesions detected on cross-sectional imaging in a younger adult population. This IRB-approved minimal risk retrospective study included 253 patients aged 18-40 years with adrenal lesions identified on cross-sectional imaging (CT, MRI, and PET/CT). Radiology reports were reviewed to determine lesion descriptors, inclusion of the adrenal lesion in the report impression, and management recommendations. Logistic regression was used to identify factors associated with inclusion of adrenal lesions in the impression. Among 253 adrenal lesions, 194 (76.7%) were included in the report impression and 59 (23.3%) were omitted. Lesion size was the most frequently reported descriptor (83.7% of reports), whereas attenuation measurements were reported in only 14.3%. Overall, 57.9% of reports contained no management recommendation. Follow-up imaging was recommended in 39.2% of cases, hormonal evaluation in 2.8%, and endocrine referral in 0.8%. Inclusion of adrenal lesions in the impression was independently associated with larger lesion size (OR 1.46 per cm, p = 0.025) but was not associated with imaging context, prior malignancy, or mechanism of discovery. Among 162 incidental lesions in patients without prior malignancy or known genetic predisposition, 5 (3.1%) were non-metastatic pheochromocytomas, 115 (71.0%) were benign, and 42 (25.9%) remained non-characterized. Adrenal lesions identified in younger adults are frequently omitted from radiology impressions and often lack key descriptors or management recommendations. Reporting practices appear to be influenced primarily by lesion size rather than clinical context, suggesting opportunities to improve the consistency and clinical utility of adrenal lesion reporting, particularly in younger patients where standardized evaluation strategies are less well established.
A dual imaging pattern of hyperattenuation on non-enhanced computed tomography and hypointensity on T2-weighted imaging is frequently encountered in various lesions. In the present review, we aimed to explore this imaging combination across a spectrum of abdominal and pelvic lesions to clarify its clinical implications.This pattern is commonly observed in lesions rich in fibrous or smooth muscle components, making it more frequently encountered in benign entities, including fat-poor angiomyolipomas, uterine leiomyomas, and tumors with predominant fibrous stroma. However, similar imaging findings may also be encountered in rare high-cellularity malignancies, selected systemic diseases, and lesions containing melanin, metal deposition, or thyroid tissue. Although not specific to a single disease, this imaging pattern offers valuable diagnostic clues for lesion characterization. When contrast-enhanced imaging is unavailable or contraindicated, combining non-enhanced computed tomography and T2-weighted imaging findings may help narrow the differential diagnosis. Despite its general association with benign lesions, this imaging phenotype requires careful exclusion of important pathologic mimics. Overall, integration with clinical information and other imaging findings remains essential for accurate interpretation.
BackgroundMidlife obesity is considered one of the top modifiable risk factors for dementia and Alzheimer's disease (AD). However, body mass index (BMI) on its own does not fully represent obesity-associated risks and it is crucial to disentangle the role of body adiposity and its localization.ObjectiveTo investigate the relationship of MRI-derived body adiposity metrics with AD-related pathology at midlife.MethodsNinety-seven cognitively normal midlife individuals underwent brain amyloid and tau PET, body MRI, and metabolic and cognitive assessments. Key measures included hepatic fat fraction, visceral (VAT) and subcutaneous adipose tissue (SAT) volumes, and thigh muscle and adiposity. The correlation between adiposity/metabolic measurements and amyloid/tau pathologies was investigated.ResultsThe average age of participants was 49.8 years, 65.3% were female and 53.6% had obesity. Amyloid PET burden in Centiloids correlated with VAT (rho = 0.36, p = 0.002), BMI (rho = 0.33, p = 0.002), SAT (rho = 0.33, p = 0.002), and insulin resistance (IR) (rho = 0.34, p = 0.003) in females and Whites, lower high-density lipoprotein (HDL) cholesterol (rho = -0.36, p = 0.002) irrespective of sex and race, and lower MMSE scores (rho = -0.57, p = 0.043) in only in African-Americans, after correction for age, sex, and education. There was no evidence that HDL nor IR mediated VAT-related amyloid. VAT/SAT ratio was significantly associated with mean cortical tau SUVR (β = 0.138, p = 0.030) after adjustment for age, sex, education, and amyloid.ConclusionsAmong fat depots in our study, visceral fat was more strongly correlated to amyloid pathology, and this association is present even independent from BMI. Also, higher visceral compared to subcutaneous fat is related to higher tau pathology.
To characterize the etiology and clinical course of patients with subcapsular renal hematoma (SRH) in order to better understand and manage patients in the acute and chronic settings. A natural language processing algorithm screened computed tomography and magnetic resonance imaging studies for "subcapsular renal hematoma" across 10 inpatient facilities from 2011 to 2021. Patients with at least three cross-sectional exams were included. Retrospective chart review assessed etiology, clinical course, and management of SRH. A total of 105 patients with acute SRH were included. The underlying etiologies involved traumatic (21%), spontaneous (32%), and iatrogenic causes (47%). Three patients developed Page kidney with hypertension secondary to renin-angiotensin-aldosterone system activation. Notably, two of these patients had solitary kidneys, with one necessitating temporary hemodialysis and the other undergoing hematoma evacuation. Additionally, six patients developed infected hematomas; five required drainage by interventional radiology (IR), and one patient underwent nephrectomy. Angiography by IR was recommended in 21 patients, 17 of whom underwent renal artery embolization. One patient developed intractable pain, requiring operative hematoma evacuation. Another patient developed abdominal compartment syndrome requiring decompressive laparotomy. Repeat imaging was done within 1-10 months in 69 patients, with 11 (16%) demonstrating complete hematoma resolution. In 16 patients who had further imaging between 10 and 15 months, an additional 8 had resolved. The average time to complete resolution was 368 days. Subcapsular hematoma is a rare finding, most commonly occurring after iatrogenic causes. In our series, roughly 1/3 (34%) of all hematomas require intervention, including hemodialysis, embolization, and nephrectomy. One third (33%) of spontaneous SRH required embolization, most commonly those with underlying renal masses. Three patients developed the Page kidney phenomenon, which was more common in patients with a solitary kidney.
Wunderlich syndrome is a rare urological emergency characterized by spontaneous, nontraumatic renal hemorrhage into the subcapsular and perirenal spaces. We present the case of a 37-year-old male with end-stage renal disease on hemodialysis, hypertension, and bilateral inguinal hernias who presented with acute chest and abdominal pain, nausea, and vomiting after missing a dialysis session. The initial assessment identified severe anemia necessitating a massive transfusion, and he was subsequently treated for suspected uremic coagulopathy. Imaging with computed tomography angiography (CTA) demonstrated a massive left retroperitoneal hemorrhage. The patient also tested positive for COVID-19, potentially contributing to a complex coagulopathic state. Given ongoing bleeding, he underwent successful renal artery embolization, achieving hemostasis without surgical intervention. This case highlights the importance of early recognition of Wunderlich syndrome in high-risk patients, the diagnostic value of CTA, and the critical role of interventional radiology in management.
Endometrial cancer is a growing global health concern, and magnetic resonance imaging (MRI) is central to its diagnosis, staging, and management. This study used a descriptive bibliometric approach to map the research landscape, identify emerging themes, and characterize global collaboration in MRI‑related endometrial cancer research over the past four decades. This study searched Web of Science Core Collection for English-language original articles on MRI in endometrial cancer published from 1984 to 2024. Using Microsoft Excel, VOSviewer, CiteSpace, and Bibliometric.com, we described publication trends, main contributors, collaboration networks, and research hotspots. Only network- and frequency-based analyses were performed; no inferential statistics or causal analyses were done. A total of 1,165 publications by 6,067 authors from 3,676 institutions in 256 countries/regions were identified, showing sustained growth in research activity. China, Japan, and the USA were the most productive countries. Key journals included Radiology, European Radiology, and Gynecologic Oncology. Co-authorship and institutional analyses indicated extensive international cooperation, particularly among centers such as the University of Bergen and Kyoto University. Keyword co-occurrence and burst analyses identified four main thematic clusters: (1) diagnostic imaging and pathology, (2) multi-modality imaging and disease management, (3) clinical outcomes and therapeutic approaches, and (4) prognostic assessment and staging. Recent studies have increasingly focused on advanced imaging analytics, with emerging hotspots in "radiomics," "nomogram," and "risk assessment". MRI research in endometrial cancer has grown and become more diverse, with increasing international collaboration. New work on radiomics, artificial intelligence (AI), and prognostic models points toward more personalized imaging but is still mainly experimental. As a descriptive bibliometric study, this work does not judge the quality or clinical readiness of specific methods. Standardized MRI protocols, better access in low-resource settings, and prospective validation of radiomics and AI tools are needed to turn these trends into real benefits for patients.
Accurate preoperative identification of extraprostatic extension (EPE) in prostate cancer (PCa) is important for treatment strategies. This retrospective study aimed to determine which MRI sequence provides the most accurate measurement of length of capsular contact (LCC) and maximum lesion radial distance (MaxRADD) for predicting EPE. We included 63 patients who underwent radical prostatectomy for biopsy-confirmed PCa and had preoperative multiparametric MRI (mpMRI) from 2018 to 2020. Two radiologists measured LCC and MaxRADD on T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) maps, and dynamic contrast-enhanced (DCE) sequences. LCC was evaluated using 10 mm and 15 mm thresholds based on Prostate Imaging-Reporting and Data System (PI-RADS) and European Society of Urogenital Radiology (ESUR) guidelines. Histopathology was the reference standard. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were calculated. 63 patients with 105 evaluable lesions were included. Significant differences were observed between the 10 mm and 15 mm LCC thresholds for predicting EPE. At 10 mm, DCE demonstrated the highest sensitivity (90.1%) and a 80.7% specificity. At 15 mm, DCE achieved higher specificity (94.7%) with 85% sensitivity. ROC analysis identified 14 mm on DCE as the optimal LCC threshold, reaching 90% sensitivity, 96.6% specificity (AUC:0.93). MaxRADD showed variable diagnostic performance, with the highest AUC on DCE (0.83), T2WI (0.80), DWI (0.78), and ADC (0.74), without significant differences. Combining LCC and MaxRADD slightly improved AUC (0.96 vs. 0.93 for LCC alone, p = 0.270). MaxRADD performed best on DCE at an 11 mm threshold (sensitivity 65%, specificity 88.5%). DCE provides the highest diagnostic performance for predicting EPE using LCC measurements, with a 14 mm LCC and 11 mm MaxRADD threshold yielding the best overall accuracy. Combining LCC and MaxRADD numerically increased AUC (0.96 vs. 0.93), however without significant differences in EPE prediction (p = 0.270).
Hepatocellular carcinoma can be treated with liver transplantation, surgical resection, or locoregional therapy. Locoregional therapy encompasses a broad range of therapeutic options, including non-radiation- and radiation-based options. The assessment of treatment response following radiation-based options is often challenging as imaging findings following these therapies evolve over time. In response to this challenge, and with advancements in knowledge of the expected imaging findings following locoregional therapy, the Liver Imaging Reporting and Data System (LI-RADS) treatment response algorithm was updated in 2024 by the American College of Radiology Treatment Response Working Group. The 2024 version was separated into two distinct algorithms, reflecting the differences in the expected imaging findings following non-radiation- and radiation-based options. This article presents a pictorial review of the expected and unexpected imaging appearance of hepatocellular carcinoma following treatment with radiation-based locoregional therapy options using the LI-RADS treatment response algorithm v2024.
Symptomatic postoperative fluid collections (POFCs) can result in significant morbidity and mortality after abdominal surgery requiring timely intervention. EUS-guided drainage is traditionally delayed up to four weeks to allow wall maturation and reduce perforation or peritonitis risk. However, some POFCs may be suitable for earlier intervention. This study compared the efficacy and safety of acute (≤ 15 days), early (16-30 days), and delayed (> 30 days) EUS-guided drainage. A retrospective cohort of patients undergoing EUS-guided drainage for symptomatic POFCs between 2013 and 2023 at a single tertiary center was evaluated. Technical success was defined as accessing and draining a POFC by transmural stent placement on initial endoscopy. Clinical success was defined as radiographically or endosonographically confirmed symptomatic POFC improvement without further percutaneous or surgical intervention. Among 85 patients with POFCs, most (61%) had undergone distal pancreatectomy with splenectomy. 59% required drainage ≤ 30 days after surgery, with 28% managed acutely. Most (83%) received lumen-apposing metal stents. Overall technical and clinical success rates were 94% and 79%, respectively, after a median 2 endoscopies (IQR 2-3). Success did not differ by timing (technical: 92% vs. 96% vs. 94%; clinical: 83% vs. 85% vs. 71%; P = 0.86 and P = 0.37). Adverse event rates were similar across groups (P = 0.85). Transgastric access was associated with clinical success (P < 0.001) and fewer adverse events (P = 0.03). Transduodenal access predicted technical (P = .05) and clinical failure (P = 0.02). In this large single-center experience of symptomatic POFCS, acute and early EUS-guided drainage with lumen-apposing metal stents in carefully selected patients was found to be technically safe and clinically effective, potentially avoiding more morbid interventions such as ERCP, percutaneous drainage, or surgery. Further randomized, prospective studies are needed to define predictors of technical and clinical success as well as adverse events.
Systemic therapy for hepatocellular carcinoma (HCC) has undergone rapid transformation over the past decade, significantly expanding treatment options and improving survival for patients with advanced disease. Guided by the Barcelona Clinic Liver Cancer staging system, systemic therapy is primarily indicated for advanced-stage HCC and select intermediate-stage cases with preserved liver function. Immune checkpoint inhibitor-based combination regimens have redefined first-line therapy, demonstrating substantial improvements in overall survival and response rates. As systemic and locoregional treatments increasingly converge, radiologists are encountering more complex imaging scenarios requiring nuanced interpretation. Assessment of treatment response in HCC relies on integration of imaging findings, biomarkers, and clinical parameters. Although Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 remains the primary response assessment method in clinical trials, its size-based criteria may not fully capture biologic response in HCC, where therapy can induce necrosis or decreased vascularity without substantial tumor shrinkage. Modified RECIST (mRECIST), which evaluates viable enhancing tumor, may better reflect treatment effect in certain contexts; however, higher response rates by mRECIST have not consistently translated into improved overall survival. Atypical response patterns such as pseudoprogression and hyperprogression may further complicate evaluation in the immunotherapy era but remain uncommon in HCC. Emerging evidence suggests that combining systemic therapy with locoregional therapies such as transarterial chemoembolization, transarterial radioembolization, and stereotactic body radiation therapy may improve disease control and survival in selected patients, further increasing imaging complexity. In routine clinical practice, radiology reports should emphasize global disease burden assessment, "the forest, not the trees," while clearly describing mixed responses, tumor-in-vein, and extrahepatic disease. In cases of temporally overlapping systemic and locoregional therapies, a compartmentalized approach integrating RECIST or mRECIST principles with LI-RADS treatment response algorithm for recently treated lesions, "the forest and the trees," is recommended to support multidisciplinary decision-making. Advanced imaging techniques including perfusion imaging, radiomics, and artificial intelligence-based modeling offer promising tools for earlier and more precise response assessment but remain largely investigational. As therapeutic paradigms continue to evolve, radiologists play a central role in guiding management decisions. A comprehensive, multidisciplinary approach that integrates imaging interpretation with clinical context and treatment history is essential to accurately assess response and optimize care in patients with HCC receiving systemic therapy.
Pediatric thoracolumbar spinal injuries (TLSIs) are rare and uncommonly require intervention. In adults, plain x-ray is not adequate to screen for spinal injury. In children, plain film utilization for screening is variable and supportive evidence is lacking. Liberal CT screening for TLSI in children results in significant unnecessary radiation exposure. We investigated the utility of plain x-rays and physical exam (PE) to screen children for TLSI. Children aged 1 to 18 years with MRI or CT-confirmed TLSI presenting to one of five Level 1 pediatric trauma centers between 2017 and 2022, who had a plain film involving the injured spine, were identified. ICD10 codes for injuries, imaging types, intervention (surgery or bracing), age, mechanism, BMI, fracture type (thoracic, lumbar, both), comorbidities, PE findings, and type of radiograph (spine, chest, abdomen) were identified and compared. Two hundred thirty-two children with MRI or CT-confirmed TLSI and plain x-rays were identified (46% thoracic, 34.9% lumbar, 18.1% both). 57.3% (n=133) of patients underwent dedicated spine radiographs, while the others had only chest (n=87, 37.5%) or only abdominal (n=10, 4.3%) radiographs. In total, 13.79% of patients underwent surgery, 43.9% required bracing, and the remaining had no intervention. Of the patients with dedicated spine films, x-rays alone were 82.7% sensitive for injury and 100% sensitive for injury needing surgery. Patients with lumbar spine injuries who had positive spine x-rays and suspicious PE findings were significantly more likely to require intervention (p<0.0001). No child with a normal spinal x-ray in the lumbar region required intervention, regardless of PE findings. In this retrospective study, dedicated spinal x-rays combined with PE reliably excluded >99% of all injuries and 100% of injuries requiring intervention. These findings suggest children can be effectively screened for TLSI with PE and spinal x-rays, reserving cross-sectional imaging for positive x-ray findings and persistent PE findings on repeat exam. (J Trauma Acute Care Surg. 2026;100:915-921. Prognostic/Epidemiological; Level III.
暂无摘要（点击查看详情）
This study aimed to develop and validate a hybrid deep learning-radiomics model that leveraged Cycle-consistent generative adversarial networks (CycleGAN)-synthesized contrast-enhanced computed tomography (CE-CT) images to differentiate advanced from non-advanced hepatic fibrosis. This retrospective study included 410 patients with biopsy-confirmed hepatic fibrosis (2017-2024). A trained CycleGAN model was used to generate synthetic three-phase CE-CT images from the corresponding non-contrast computed tomography (NC-CT) data. Each group of images was randomly split 6:4 ratio into training and test sets. After region of interest (ROI) segmentation, handcrafted radiomic (HCR) features were extracted. Concurrently, eight end-to-end deep learning (DL) models were trained; DL features were extracted from the best-performing model. Feature selection was performed using Spearman's rank correlation and the least absolute shrinkage and selection operator (LASSO). Six machine learning classifiers were developed for each feature type (HCR, DL, and late-fused DL features) using the final selected feature set. The performance of models was assessed by the area under the receiver operating characteristic curve (AUC), accuracy, calibration curves, decision curve analysis (DCA) and the DeLong test. Models utilizing synthetic CE-CT images outperformed those based on NC-CT. DL feature-based models surpassed HCR-based models. A late-fusion hybrid model integrating DL features further improved performance, achieving an AUC of 0.880 (95% CI: 0.819-0.942). The model based on synthetic CE-CT images demonstrated excellent diagnostic performance. Moreover, the hybrid model combining both real NC-CT and synthetic CE-CT images further improved diagnostic performance. The hybrid model can serve as a non-invasive diagnostic method for differentiating advanced from non-advanced hepatic fibrosis.
暂无摘要（点击查看详情）
The diagnosis and management of focal pancreatic lesions is complex with diagnostic uncertainty and inconsistent guidelines. While multidisciplinary teams (MDTs) have emerged as the standard of care to navigate this challenge, a comprehensive synthesis of their impact is absent from literature. This systematic review aims to consolidate evidence on how formal MDT-led care influences diagnostic accuracy, treatment decisions and timelines, guideline compliance, and clinical outcomes for this diverse patient population. This systematic review was conducted following the PRISMA guidelines and was registered on PROSPERO. We searched PubMed, Scopus, and CENTRAL through 14 April 2025 to identify studies evaluating the impact of MDTs on the management of focal pancreatic lesions. Three independent reviewers performed screening and data extraction, with conflicts resolved by a fourth reviewer. Our systematic review included 20 studies encompassing 14,366 patients. MDT review led to a change in primary diagnosis in 17.6% to 37.9% of cases and altered clinical stage or resectability assessment in 18.7% to 31.5% of patients following expert re-evaluation of imaging and pathology. Consequently, the initial management plan was modified in 18% to 72% of cases. This consistently improved patient selection, with increased use of neoadjuvant therapy, higher rates of guideline-concordant care and clinical trial enrollment, and more frequent referral to palliative care, significantly reducing non-therapeutic laparotomies. While MDT management was associated with a significantly higher likelihood of achieving a complete (R0) surgical resection (OR 5.47), its impact on overall survival was inconsistent across the reviewed literature. This systematic review provides substantial evidence that MDT-led care fundamentally improves the management of focal pancreatic lesions. MDT implementation consistently enhanced diagnostic accuracy, optimized treatment planning, and increased guideline-concordant care, leading to improved surgical outcomes. While the evidence is largely retrospective, these benefits strongly support the MDT as an indispensable standard of care. Future research should focus on optimizing the MDT model's structure, quality, and cost-effectiveness to maximize patient benefit globally. PROSPERO (CRD420251142032).
To evaluate the diagnostic value of multislice spiral computed tomography (MSCT) in differentiating pancreatic acinar cell carcinoma (PACC) from pancreatic ductal adenocarcinoma (PDAC). The clinical, pathological, and imaging data of 17 patients with pathologically confirmed PACC and 62 patients with PDAC were retrospectively analyzed. Quantitative variables were compared between groups using the independent samples t-test or the Mann-Whitney U test, as appropriate. Qualitative variables were compared using the Pearson's chi-square test or Fisher's exact test. Variables showing statistical significance in univariate analysis were entered into multivariate logistic regression analysis to identify independent predictors for distinguishing PACC from PDAC. Diagnostic performance was assessed using receiver operating characteristic curve analysis, with calculation of the area under the curve (AUC), sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. Univariate analysis demonstrated significant differences between the two groups in tumor shape, margin, pancreatic atrophy, pancreatic duct transection, maximum tumor diameter, CT attenuation values, and enhancement ratios in the pancreatic parenchymal, portal venous, and delayed phases, all of which showed statistically significant differences. Multivariate logistic regression analysis identified tumor margin, pancreatic duct transection, pancreatic parenchymal phase CT attenuation value as independent predictors for distinguishing PACC from PDAC. The combined diagnostic model incorporating these variables achieved the highest diagnostic performance, with an AUC of 0.968. The model demonstrated a sensitivity of 94.1%, specificity of 88.7%, accuracy of 89.9%, positive predictive value of 69.5%, and negative predictive value of 98.2%. Tumor margin, pancreatic duct transection, and pancreatic parenchymal phase CT attenuation value are significant imaging features for differentiating PACC from PDAC. A combined diagnostic model integrating these imaging features provides excellent diagnostic performance and may aid in improving preoperative differential diagnosis.
To estimate patient-level diagnostic accuracy of deep learning (DL) for MRI-based detection of clinically significant prostate cancer (csPCa), assess heterogeneity and clinical-readiness signals, and compare DL-alone, PI-RADS-alone, and AI-assisted/DL + PI-RADS interpretation where direct comparator data were available. Following PRISMA-DTA, MEDLINE, Embase, and Web of Science were searched from 2010 to June 2025 for studies reporting patient-level 2 × 2 diagnostic accuracy data for DL applied to prostate MRI. Risk of bias was assessed using QUADAS-2. Pooled sensitivity and specificity were estimated using bivariate random-effects and HSROC models, with prespecified subgroup, meta-regression, and sensitivity analyses. Deeks' funnel plot asymmetry test assessed publication bias and small-study effects. A secondary direct three-way comparative analysis was performed in studies reporting DL-alone, PI-RADS-alone, and AI-assisted/DL + PI-RADS data within the same or closely matched cohorts. AI-specific reporting and clinical-readiness signals were mapped using items adapted from STARD-AI, CLAIM, and DECIDE-AI. Thirty-six studies including 9,411 patients were included. Pooled sensitivity was 0.91 (95% CI, 0.89-0.93), specificity was 0.55 (95% CI, 0.46-0.64), LR + was 2.04, and LR - was 0.16. Sensitivity was relatively consistent, whereas specificity varied widely, with a broad HSROC prediction region indicating limited transportability of pooled specificity. Deeks' test showed no significant funnel plot asymmetry (p = 0.393). Sensitivity analyses excluding MRI + clinical-variable hybrid models and small cohorts produced similar estimates. In nine directly comparative studies, sensitivity was similar across groups, while specificity was highest for AI-assisted/DL + PI-RADS and lowest for PI-RADS-alone. DL for prostate MRI shows high sensitivity and low LR-, supporting a rule-out or assistive role. However, moderate and variable specificity limits stand-alone rule-in use. Combined AI/DL + PI-RADS workflows may reduce false positives, but prospective validation, calibration, interpretability evaluation, and patient-level safety studies are needed.