Zoom-DWI versus general-DWI in gastric cancer: a comparative evaluation of image quality, apparent diffusion coefficient, and entropy
TL;DR
Zoom-DWI may serve only as a complementary sequence—not a replacement—for gastric cancer MRI and is unsuitable for quantitative assessments that rely on established diagnostic ADC thresholds, demonstrating that the two DWI sequences cannot share unified quantitative diagnostic cutoffs.
Abstract
To evaluate and compare image quality, apparent diffusion coefficient (ADC) values, and entropy metrics derived from diffusion-weighted imaging (DWI) acquired using two echo-planar imaging (EPI)-based techniques—zonal oblique multi-slice DWI (Zoom-DWI) and general single-shot EPI DWI with parallel imaging (General-DWI) —in gastric cancer. This single-center, retrospective study included 70 patients with histopathologically confirmed gastric cancer who underwent abdominal MRI between March 2023 and October 2024. DWI was acquired using free-breathing, two-dimensional, fat-suppressed sequences: General-DWI and Zoom-DWI. Image quality was assessed qualitatively using a five-point scale. Quantitative analysis included measurement of mean ADC values and entropy metrics within tumor regions of interest. The Shapiro–Wilk test was applied to assess the normality of continuous variables. Statistical analyses comprised paired t-tests, linear mixed-effects models (LMMs), Pearson correlation coefficients, and Bland–Altman agreement analysis. Inter-reader reproducibility was evaluated using Cohen’s kappa (κ) for categorical assessments and intraclass correlation coefficients (ICCs) for continuous measurements. Zoom-DWI yielded significantly higher tumor boundary delineation scores (3.78 ± 0.77, P < 0.001) but poorer image noise performance (3.76 ± 0.91, P = 0.019) compared with General-DWI. General-DWI generated significantly higher ADC values [(1.43 ± 0.26) ×10⁻³ mm²/s vs. (1.33 ± 0.25) ×10⁻³ mm²/s, P < 0.001], with linear mixed-effects modeling revealing a systematic ADC underestimation bias of − 0.096 × 10⁻³ mm²/s associated with Zoom-DWI. Mean entropy was comparable between the two sequences (4.70 ± 0.73 vs. 4.72 ± 0.68, P = 0.701). ADC values showed an extremely strong correlation ( R = 0.906), and entropy showed a strong correlation ( R = 0.776). Bland–Altman plots revealed consistent systematic ADC deviation and wide variation in entropy measurements, demonstrating that the two DWI sequences cannot share unified quantitative diagnostic cutoffs. Zoom-DWI provides superior visualization of tumor margins but is associated with higher image noise. Although ADC values and entropy metrics demonstrate strong linear correlations between Zoom-DWI and General-DWI, Zoom-DWI yields systematically lower ADC measurements. Owing to this systematic ADC bias, Zoom-DWI may serve only as a complementary sequence—not a replacement—for gastric cancer MRI and is unsuitable for quantitative assessments that rely on established diagnostic ADC thresholds. Not applicable. This study is a retrospective study. This study was approved by the Institutional Review Board (or Ethics Committee) of Army Medical Center of PLA (Approval No. 2022-223-01).