Room: Track 2
Purpose: Intratumoral dose response can be detected at the tumor voxel-level using serial FDG-PET/CT imaging feedbacks during treatment and used to guide dose painting. However, to reliably implement this technique, the effect of deformable image registration (DIR) uncertainty on the response assessment needs to be determined.
Methods: FDG-PET/CT images were obtained at pretreatment and weekly during chemoradiotherapy from 18 HNC patients. Tumor voxel dose response matrices (DRM)s were constructed voxel-by-voxel following an image intensity-based DIR (IM-DIR) and a biomechanical-based DIR (BM-DIR) respectively. The DRM value of a tumor voxel represents the average change rate of its metabolic activity with respect to the pretreatment/baseline activity. The IM-DIR was performed using a research software (ADMIRE). The BM-DIR was performed using a finite element method to improve the displacement regularity in tumors based on tissue mechanical properties. The displacement-vector-fields (DVF)s obtained from the two DIRs were compared and their physical plausibility was evaluated using the Jacobian matrix. The mean and SD of tumor voxel DRM discrepancies induced by the DIR variations were calculated on individual tumors.
Results: Eighty-eight% of tumor voxels have a DVF-vector difference = 2mm (half of a PET voxel size). For all tumor voxels, the mean±SD Jacobians were 0.87±0.24 and 0.88±0.2 for the IM-DIR and the BM-DIR respectively. The tumor voxel DRM discrepancy was 1.7%±9.1%. For individual tumors, the mean of DRM discrepancy was significantly correlated with volume shrinkage with a Spearman correlation coefficient ? = 0.52. The SD of DRM discrepancy was correlated with tumor volume ? = 0.61.
Conclusion: Due to lack of CT image contrast in tumors, IM-DIR can cause non-negligible discrepancies on the quantitative response assessment, particularly for tumors with large volume or considerable shrinkage during treatment. On the other hand, BM-DIR generated more physically plausible deformations and was suggested for the intratumoral dose response assessment.