Citations

Abstract Despite growing interest in AI for pancreatic cyst management, no prior study has systematically investigated how lesion characteristics influence model behavior or provided a comprehensive explainability analysis in IPMN malignancy risk stratification. We present the first multi-center study integrating explainable AI with uncertainty quantification to evaluate how cyst type, size, and location affect deep learning performance in IPMN malignancy prediction. Our retrospective study analyzed 170 IPMNs from seven centers using a radiomics-deep learning fusion model. Cases were stratified by dysplasia grade, IPMN type, size, and location; with model interpretability assessed using SHAP, LIME, Grad-CAM, and uncertainty quantification. Overall accuracy was 67.1%, with model uncertainty lower for correct versus incorrect predictions (0.72 vs 0.78, p < 0.001). On subgroup analysis: Low-Risk lesion accuracy exceeded High-Risk (81.3% vs. 42.9%, p < 0.001), BD accuracy was inversely correlated with cyst size while MD remained stable, and body and whole-pancreas cysts stratified more accurately than tail cysts (77.8%, 76.6% vs. 47.8%). AI models face significant, previously uncharacterized performance drops when evaluating complex high-risk IPMNs. Our study shows that integrating uncertainty quantification can successfully flag unreliable predictions, enabling a selective-prediction framework where clinicians can confidently rely on AI for low-risk, low-uncertainty lesions while deferring high-uncertainty cases to expert review.

A 32-fold accelerated real-time cine MRI sequence (radial sampling) demonstrated better temporal resolution and image quality than a 16-fold sequence (Cartesian sampling) in patients with cardiac implantable electronic devices and arrhythmias.

Background: The prevalence of cardiac implantable electronic devices (CIED) is increasing, with over 300,000 new implants annually in the U.S. Stress cardiac magnetic resonance (CMR) has proven as a powerful modality for myocardial characterization, enabling accurate diagnosis of different cardiac conditions. In patients without devices, CMR findings have also demonstrated prognostic value for major adverse cardiovascular events (MACE) and guide clinical decision-making. However, its use in CIED patients has historically been limited due to safety and image quality concerns. Hypothesis: We aimed to evaluate the prognostic value of CMR imaging using wideband sequences in CIED patients. Methods: CIED patients referred for clinical CMR were included prospectively and retrospectively between January 2020 and February 2025. All patients underwent a stress CMR protocol, which included cine, wideband late gadolinium enhancement (LGE), wideband T1 mapping, and wideband stress/rest perfusion imaging. Clinical follow-up is ongoing. The primary endpoint was the occurrence of MACE, defined as the composite of myocardial infarction, heart failure hospitalization, implantable cardioverter defibrillator (ICD) therapy, cardiac death, revascularization, or heart transplant. Results: A total of 89 patients were included (median age 69.0 [58.0–75.5] years; 62.9% male). Device types included pacemakers (33.7%), ICDs (48.3%), and cardiac resynchronization therapy devices (16.9%). Over a median of follow-up of 613 days [355-991], 21 patients experienced a MACE. In unadjusted models, higher left ventricular ejection fraction (LVEF) (per 5%) was significantly associated with a lower risk of MACE (odds ratio [OR] 0.80, 95% confidence interval [CI] 0.65–0.95; p=0.016), as was higher stress myocardial blood flow (MBF) (per 0.25 mL/g/min increase: OR 0.78, 95% CI 0.60–0.98; p=0.05). In contrast, greater LV myocardial mass (per 10 g) was significantly associated with a higher risk of MACE (OR 1.27, 95% CI 1.08–1.52; p = 0.006), figure 1. Conclusion: The use of wideband sequences enabled the assessment of prognostic imaging value in CIED patients. LVEF remains a strong and consistent predictor of MACE in this population and LV mass was significantly associated with higher rates of MACE reflecting the influence of underlying cardiomyopathy. While stress perfusion CMR imaging showed promising trends, further studies with longer follow-up are needed to clarify its prognostic value in this population.