Artificial Intelligence-Accelerated vs. Conventional Diffusion-Weighted Imaging for Prostate MRI: Comparing Quality and Quantitative Metrics.
Authors
Affiliations (4)
Affiliations (4)
- Department of Diagnostic and Interventional Radiology, University Medical Center Freiburg, University of Freiburg, Hugstetterstr. 55, 79106 Freiburg, Germany (V.S., O.G., F.B., J.W., M.F.R., H.E.). Electronic address: [email protected].
- Department of Diagnostic and Interventional Radiology, University Medical Center Freiburg, University of Freiburg, Hugstetterstr. 55, 79106 Freiburg, Germany (V.S., O.G., F.B., J.W., M.F.R., H.E.).
- Research & Clinical Translation, Magnetic Resonance, Siemens Healthineers AG, Henkestr. 127, 91052, Erlangen, Germany (W.L., E.W.).
- EMEA Scientific Partnerships, Siemens Healthineers AG, Siemensstr. 3, 91301 Forchheim, Germany (R.S.).
Abstract
Diffusion-weighted imaging (DWI) is central to prostate magnetic resonance imaging (MRI) but lengthens examinations. We evaluated whether an artificial intelligence (AI)-accelerated, reduced-field-of-view diffusion sequence (AI-DWI) could shorten scan time without sacrificing perceived diagnostic image quality, and how it affects quantitative diffusion metrics. This prospective, single-center study of diagnostic accuracy enrolled consecutive men with elevated prostate-specific-antigen levels between March and May 2025. The index AI-DWI sequence was compared against the standard conventional DWI sequence (c-DWI) for each patient. Three radiologists scored subjective image quality. Quantitative analysis involved comparing mean apparent diffusion coefficient (ADC) and seven additional texture features. Wilcoxon signed-rank tests assessed ordinal scores, and paired t-tests were used for quantitative metrics. 62 men (mean age, 68.7 years ± 9) were evaluated. The AI-DWI sequence demonstrated a significantly shorter acquisition time compared to c-DWI (3 min 59 s vs. 4 min 21 s; p<0.01). There was no significant difference in subjective scores for overall image quality, lesion conspicuity, artifacts, or anatomic differentiability (p>0.05 for all). AI-DWI yielded significantly lower mean ADC values (975.92±174.57 vs. 1013.21±189.34; adj. p<0.01) and maximum ADC values (adj. p<0.01). No significant differences were found for standard deviation, coefficient of variation, entropy, kurtosis, minimum, or skewness (adj. p>0.05). The AI-DWI sequence allows for reduced acquisition time while preserving subjective image quality compared to the c-DWI. Quantitatively, it yields lower mean and maximum ADC values, while showing no significant differences in the rest of the quantitative metrics relative to the conventional sequence.