Comparison of the effects of IMRT and IMPT on MRI features of normal brain tissue: a retrospective study

This study aims to analyze magnetic resonance imaging (MRI) feature changes across different dose regions to investigate the differences in imaging feature alterations induced by Intensity-Modulated Radiation Therapy (IMRT) and Intensity-Modulated Proton Therapy (IMPT) in normal brain tissue. And evaluate the potential application value of radiomic features in dose optimization and biological dose assessment.

A retrospective analysis included 113 patients undergoing brain-target radiotherapy: 44 receiving IMPT (average total dose: 5713.63 ± 774.32 cGy) and 69 receiving IMRT (average total dose: 5489.86 ± 627.05 cGy). There was no significant difference in prescribed doses between groups (p < 0.05). MRI data were collected pre-treatment and one month post-treatment, including T1, enhanced T1, T2, T2-FLAIR, and ADC sequences. Images underwent preprocessing and rigid registration on the 3D Slicer platform. Dose regions (90%, 70%, 50%, 30% of the total dose) were segmented and classified into high (70-90%), medium (50-70%), and low (30-50%) dose regions as volumes of interest (VOI). Radiomic features were extracted from pre- and post-treatment MRI datasets, and Delta features were calculated as the difference between pre- and post-treatment data. Non-parametric tests, t-tests, and Mann-Whitney U tests were used to identify significant feature changes, with effect size analysis to compare the magnitude of changes. Multiple testing correction was applied to reduce false positives. Finally, box plots and the Mapping of Dose Distribution and Differences in Radiomics Feature Variation were utilized to visually illustrate the extent of brain tissue changes in each dose region following radiotherapy.

Non-parametric tests indicated that more radiomic features exhibited significant changes in the IMRT group (p < 0.001). Effect size analysis also showed that the majority of features demonstrated greater and more pronounced changes within dose regions in the IMRT group (p < 0.05). Further correlation analysis revealed that most features (33/34) had weak correlations with dose-volume parameters (r < 0.3), suggesting that treatment modality itself may be the primary driver of these alterations rather than dose distribution alone. This finding highlights the need to refine traditional dose-based radiotherapy evaluation by incorporating biological effect considerations. Additionally, Moreover, the feature Delta_wavelet-HHL.65 exhibited the most significant change, with an effect size of -0.8 in the IMRT group. It demonstrated a moderate-to-high correlation with dose-volume parameters in the medium-to-high dose regions, with a maximum correlation coefficient of 0.65 (p < 0.001). This suggests that Delta_wavelet-HHL.65 may serve as a potential biomarker for reflecting local dose effects, contributing to the optimization of radiotherapy dose distribution.

This study revealed that IMRT and IMPT induce distinct radiomic feature changes under comparable dose conditions, and these differences are not solely determined by dose distribution. This suggests that radiotherapy evaluation should extend beyond physical dose metrics to incorporate biological effect dimensions. Furthermore, Delta_wavelet-HHL.65 demonstrated the most pronounced change in the IMRT group and exhibited correlations with dose-volume parameters in the low-to-medium dose regions, highlighting its potential for radiotherapy optimization.

Not applicable.