Objectives: To evaluate whether deep learning reconstruction (DLR) improves pancreas-referenced image quality in high-b-value field-of-view optimized and constrained undistorted single-shot diffusion-weighted imaging (FOCUS DWI) for pancreatic cancer, and to assess reconstruction-related changes in apparent diffusion coefficient (ADC) measurements in tumor and non-tumorous pancreatic regions. Methods: This prospective single-center study included 68 patients with pathologically confirmed pancreatic cancer who underwent 3.0-T pancreatic MRI. FOCUS DWI was acquired with b values of 0, 1000, 1500, and 2000 s/mm2 and reconstructed with and without DLR. Quantitative metrics included lesion signal-to-noise ratio (SNR), lesion-to-pancreas contrast-to-noise ratio (CNR), lesion-to-pancreas signal intensity ratio (LP-SIR), and ADC values in tumor regions and non-tumorous pancreatic parenchyma. Two abdominal radiologists independently assessed composite image quality and lesion conspicuity using a 5-point ordinal scale. Paired comparisons were performed with Holm correction across b values. Results: DLR significantly increased lesion SNR and lesion-to-pancreas CNR at b = 1000, 1500, and 2000 s/mm2 (all Holm-adjusted p < 0.001). Absolute image quality was highest at b = 1000 s/mm2, whereas the greatest DLR-related lesion SNR gain occurred at b = 1500 s/mm2. LP-SIR remained broadly stable after DLR (Holm-adjusted p = 0.767, 1.000, and 1.000). Tumor regions showed lower ADC values than non-tumorous pancreatic parenchyma across all b values and reconstruction conditions (all p < 0.001), while ADC values did not differ significantly between DLR and non-DLR reconstructions within either region (p = 0.414–0.794). Qualitative scores improved after DLR for both readers at all b values (all Holm-adjusted p < 0.001), with weighted κ ranging from 0.717 to 0.896. Conclusions: DLR improved ROI-based quantitative image-quality metrics and qualitative image assessment in high-b-value pancreatic FOCUS DWI, with the greatest lesion SNR gain at b = 1500 s/mm2. No statistically significant reconstruction-related differences in ADC measurements were observed within tumor or non-tumorous pancreatic regions.
Background/Objectives: This study aimed to evaluate whether artificial intelligence-derived vertebral volumetric bone mineral density (AI-vBMD) and paraspinal intermuscular adipose tissue (IMAT) ratio from routine computed tomography (CT) could identify moderate-to-severe vertebral compression fractures (VCFs) in breast cancer survivors, and whether paraspinal IMAT ratio and routinely available clinical variables improved diagnostic performance. Methods: This retrospective study included 275 women with breast cancer who underwent routine non-contrast CT and lumbar quantitative computed tomography (QCT). Hounsfield unit-derived volumetric bone mineral density (HU-vBMD) was derived using a QCT-referenced HU-to-vBMD conversion equation, whereas AI-vBMD and paraspinal IMAT ratio were extracted using automated software. Moderate-to-severe VCF was defined as Genant grade ≥ 2. Agreement with QCT-vBMD was assessed using correlation, intraclass correlation coefficient (ICC), and Bland–Altman analysis. Model discrimination was evaluated using receiver operating characteristic analysis and DeLong tests. Results: Moderate-to-severe VCF was present in 75 patients (27.3%). HU-vBMD and AI-vBMD showed excellent agreement with QCT-vBMD (ICC, 0.978 and 0.987, respectively). AI-vBMD outperformed HU-vBMD for identifying VCFs (AUC, 0.738 vs. 0.714; p < 0.001). IMAT ratio showed comparable standalone discrimination to AI-vBMD (AUC, 0.760 vs. 0.738; p = 0.604). Adding IMAT ratio to AI-vBMD improved discrimination (AUC, 0.786 vs. 0.738; p = 0.038). The full model incorporating clinical covariates achieved the highest AUC (0.828; 95% CI, 0.778–0.878). Conclusions: AI-vBMD and paraspinal IMAT ratio automatically extracted from routine CT improved the diagnostic assessment of prevalent moderate-to-severe VCFs in breast cancer survivors. This study supports an automated CT-based approach that integrates vertebral bone density and paraspinal muscle–fat information for opportunistic identification of clinically relevant VCFs.