Predicting pathological tumor volume in prostate cancer lesions: A head-to-head comparison of micro-ultrasound vs. MRI

Abstract

Background and Objective: Our objective was to evaluate the agreement between micro-ultrasound, MRI and pathological tumor and prostate volume. Methods: Retrospective analysis of consecutive prostate cancer patients with MRI and micro-ultrasound diagnostic assessment who subsequently underwent radical prostatectomy. Tumor and prostate volume on micro-ultrasound and MRI imaging calculated by a dedicated software were compared to those of the prostatectomy specimen. Clinical, radiological, and pathological predictors of pathological tumor size were assessed. Results: 65 men with a total of 104 lesions in the final pathology were included. Median micro-ultrasound tumor size was 1.05 ml (IQR 0.35–2.65). On MRI T2WI, DWI and ADC sequences median tumor volume was 0.73 ml (IQR 0.34–1.94), 0.94 ml (IQR 0.38–2.09) and 0.86 ml (IQR 0.42–1.58), respectively. The pathological median tumor size was 1.2 ml (IQR 0.2–3.9). On average, micro-ultrasound underestimated pathological tumor volume by 0.15 ml (P < 0.01) while DWI, the most precise MRI sequence underestimated tumor size by 0.26 ml (P < 0.01). The MRI and micro-ultrasound underestimated the pathological prostate volume by 6 ml (P < 0.01) and 3 ml (P = 0.47), respectively. Conclusions: Both micro-ultrasound and MRI tend to slightly underestimate pathological tumor and prostate volume. Our study shows that both micro-ultrasound and MRI can be useful in the surgical planning although the underestimation of actual tumor size should be considered.