Assessing Prostate Cancer Response to MR-Guided Adaptive Radiotherapy Using Apparent Diffusion Coefficient (ADC) on a 0.35T MR-Linac
Abstract
Purpose
To investigate whether the apparent diffusion coefficient (ADC) from diffusion-weighted imaging (DWI) acquired at a 0.35T MR-Linac can determine prostate and intraprostatic lesion changes caused by adaptive radiotherapy performed on this treatment modality.
Methods
Four prostate cancer patients treated at a 0.35T MRIdian MR-Linac underwent DWI (b=100, 350, 600 s/mm2; 1.2x1.2x3 mm2 voxels) during their planning session and last treatment fraction. DW images were denoised using Local (LPCA) or Marchenko-Pastur Principal Component Analysis (MPPCA) in Python. The ADC and coefficient of determination R2 were obtained by pixel-wise mono-exponential fitting of denoised trace-weighted (dTW) images. dTW of the first and last imaging session were fused to planning TrueFISP images, transferring clinical contours on respective ADC and R2 maps. Prostate-only contours were created by excluding lesions and overlap with bladder, urethra or seminal vesicles. ADC histograms of the lesion and prostate-only contours were produced for each timepoint by considering pixels with R2>0.8 only.
Results
The R2 threshold enabled the selection of contour pixels with reliable ADC values. Median ADC was lower for the intraprostatic lesion than the rest of the prostate, signifying adequate lesion detection on ADC maps, for patients 1-3. Lesion median ADC increased between treatment simulation and last fraction for the first 3 subjects, by 28%, 60%, and 14%, respectively, indicating favorable response to irradiation. ADC in the prostate-only contour also increased after treatment for the first 2 patients (by 5% and 15%, respectively) but decreased by 6% for the third one. The fourth patient demonstrated ADC reduction for both the lesion (13%) and rest prostate (5%), suggesting he may have not responded to radiotherapy.
Conclusion
DWI at a 0.35T MR-Linac may reveal ADC changes in the prostate and its malignant tumors between treatment beginning and completion. These results may differentiate responders and non-responders to adaptive radiotherapy.