Exploring the value of multiparametric quantitative magnetic resonance imaging in avoiding unnecessary biopsy in patients with PI-RADS 3–4

Purpose

To validate the performance of three multiparametric quantitative sequences, including amide proton transfer-weighted (APTw) MRI, diffusion kurtosis imaging (DKI) and apparent diffusion coefficient (ADC), in the enhancement of biopsy decision-making in patients with Prostate Imaging-Reporting and Data System (PI-RADS) core of 3–4.

Materials and methods

A total of 96 participants who scored as PI-RADS 3–4 according to PI-RADS v2.1 and pathologically confirmed were enrolled. The mean of APT, ADC, mean diffusivity (MD) and mean kurtosis (MK) values were acquired by two radiologists and a senior radiologist. The univariate and multivariate logistic regression analyses were performed for parameters selection. Receiver operating characteristic (ROC) analysis was employed to assess the performance of models. The Delong test was applied to compare the area under the ROC curves (AUCs) between models. The proportion of unnecessary biopsies avoided (specificity) and csPCa missed (1-sensitivity) were calculated for each model.

Results

In this PI-RADS 3–4 cohort, the AUCs for the diagnosis of csPCa were 0.730 (95%CI: 0.626, 0.834), 0.682(95%CI: 0.562, 0.802), 0.610(95%CI: 0.482, 0.739) and 0.706(95%CI: 0.593, 0.819) for APT, MD, MK and ADC values, respectively. The combined model yielded the significantly higher AUC of 0.851 (95%CI: 0.767, 0.934) (P < 0.05), and achieved a best trade-off between missing csPCa and avoiding unnecessary biopsy than other models (Youden = 0.615).

Conclusion

A novel diagnostic model incorporating APTw-MRI, DKI and ADC has been shown to provide incremental diagnostic value in stratifying biopsy necessity for patients with PI-RADS 3–4 lesions.

Graphical abstract