Daily Intravoxel Incoherent Motion (IVIM) In Prostate Cancer Patients During MR-Guided Radiotherapy-A Multicenter Study.

Abstract

PURPOSE

Daily quantitative MR imaging during radiotherapy of cancer patients has become feasible with MRI systems integrated with linear accelerators (MR-linacs). Quantitative images could be used for treatment response monitoring. With intravoxel incoherent motion (IVIM) MRI, it is possible to acquire perfusion information without the use of contrast agents. In this multicenter study, daily IVIM measurements were performed in prostate cancer patients to identify changes that potentially reflect response to treatment.

RESULTS

In total, 726 fractions were analyzed. Pre-treatment average values, measured on the first fraction before irradiation, were 1.46 × 10^-3 mm²/s, 0.086, and 28.7 × 10^-3 mm²/s in the non-cancerous prostate and 1.19 × 10^-3 mm²/s, 0.088, and 28.9 × 10^-3 mm²/s in the tumor, for D, f, and D*, respectively. The repeatability coefficients for D, f, and D* in the non-cancerous prostate were 0.09 × 10^-3 mm²/s, 0.05, and 15.3 × 10^-3 mm²/s. In the tumor, these values were 0.44 × 10^-3 mm²/s, 0.16, and 76.4 × 10^-3 mm²/s. The mixed effects analysis showed an increase in D of the tumors over the course of treatment, while remaining stable in the non-cancerous prostate. The f and D* increased in both the non-cancerous prostate and tumor.

CONCLUSIONS

It is feasible to perform daily IVIM measurements on an MR-linac system. Although the repeatability coefficients were high, changes in IVIM perfusion parameters were measured on a group level, indicating that IVIM has potential for measuring treatment response.

MATERIALS AND METHODS

Forty-three patients were included, treated with 20 fractions of 3 Gy on a 1.5 T MR-linac. IVIM measurements were performed on each treatment day. The diffusion coefficient (D), perfusion fraction (f), and pseudo-diffusion coefficient (D*) were calculated based on the median signal intensities in the non-cancerous prostate and the tumor. Repeatability coefficients (RCs) were determined based on the first two treatment fractions. Separate linear mixed-effects models were constructed for the three IVIM parameters.