Presentation Authors: Yoh Matsuoka*, Hiroshi Tanaka, Tomo Kimura, Shingo Moriyama, Sho Uehara, Yosuke Yasuda, Toshiki Kijima, Soichiro Yoshida, Minato Yokoyama, Junichiro Ishioka, Kazutaka Saito, Yasuhisa Fujii, Tokyo, Japan
Introduction: Apparent diffusion coefficient (ADC) is expected to reflect tumor aggressiveness. In prostate cancer (PC), however, the use of ADC value (ADCV) is limited to the peripheral zone (PZ) because of the background signal, especially in the transition zone (TZ). We assessed the usefulness of ADC ratio (ADCR) in predicting ISUP grade groups on radical prostatectomy (RP-GG) and subsequent biochemical recurrence (BCR) by prostatic zone.
Methods: We enrolled 284 clinically localized PC patients who had undergone multiparametric MRI, systematic biopsy, and RP. A region of interest (ROI) was drawn outlining the index tumor on ADC map. Another equal-sized ROI was drawn symmetrically in the contralateral lobe and adjusted to include normal regions if necessary. ADCR was calculated by dividing the tumor ADCV by the contralateral ADCV. ADC metrics were examined for associations with RP-GG and BCR using Spearman&[prime]s coefficient and logistic or Cox regression, and their predictive ability was assessed using the area under the curve (AUC) or concordance (c-) index. Cutoffs of variables were set to obtain their best performance.
Results: In total (TZ/PZ = 103/181, RP-GG1/2/3/â‰¥4 = 25/146/70/43), 76 men (27%) had BCR at a median follow-up of 68 months. ADCV and ADCR inversely correlated with RP-GG in TZ (Ï = -0.42 and -0.55, each p < 0.001) and in PZ (Ï = -0.61 and -0.66, each p < 0.001). In the prediction of RP-GGâ‰¥3, AUCs of ADCV and ADCR were 0.72 and 0.83 in TZ (p=0.01) and 0.86 and 0.88 in PZ (p=0.32), respectively. Independent predictors included ADCR and biopsy-GGâ‰¥3 in TZ (each p < 0.001) and ADCR (p < 0.001), biopsy-GGâ‰¥3 (p=0.02), and PSAâ‰¥10 ng/mL (p=0.02) in PZ. The best cutoff of ADC metrics in TZ/PZ was 0.55/0.64 for ADCV (Ã—10-3mm2/s) and 0.62/0.60 for ADCR. The positive/negative predictive values of ADCR < 0.61 for RP-GGâ‰¥3 were 78%/86% in TZ and 83%/84% in PZ. In the prediction of BCR, independent predictors included ADCR (p=0.04) and PSAâ‰¥10 ng/mL (p < 0.001) in TZ, and ADCR (p=0.01), PSAâ‰¥10 ng/mL (p=0.01), and biopsy-GGâ‰¥3 (p=0.03) in PZ. Postoperative courses of men with ADCRâ‰¥0.59/ < 0.59 were stratified with 5-year BCR-free rates of 86/69% for TZ (p=0.02, c-index 0.59) and 75%/45% for PZ (p < 0.001, c-index 0.65).
Conclusions: ADCR serves as a quantitative marker of grade groups and BCR, with nearly the same cutoffs for TZ and PZ.