Raman Spectroscopy–Based Modelling of Immune Cell Infiltration In Prostate Cancer Treated with HDR Brachytherapy
Abstract
Purpose
To quantify treatment-associated immune density changes in prostate cancer following HDR brachytherapy, and to determine whether Raman spectroscopic signatures of tissue biochemistry may serve as potential biomarkers of immune response.
Methods
Immunohistochemistry and Raman spectroscopy were employed to measure immune cell densities and tissue biochemical signatures within needle-core biopsies obtained from patients undergoing curative HDR brachytherapy of prostate adenocarcinoma. Biopsies were collected before the first fraction and again two weeks later, before the second fraction was administered, representing baseline and post-radiation timepoints (0 Gy and 13 Gy, respectively). Using in-house developed algorithms, Raman spectra were decomposed into weighted scores reflecting the relative abundance of known biochemical constituents. These biochemical scores were correlated with immune cell counts in adjacent tissue cross-sections. Additionally, combined biochemical and immune data were used to generate predictive models of patient-level inflammatory immune cell densities, including macrophages (CD68+), total T cells (CD3+), and cytotoxic T cells (CD3+CD8+).
Results
Quantitative analysis of immune cell densities demonstrated a two-fold increase in macrophage (CD68+) density and a 1.5-fold decrease in cytotoxic T cell (CD3+CD8+) density two weeks following the first fraction of HDR brachytherapy. Among the predictive approaches evaluated, gradient-boosted tree models achieved the highest performance in estimating overall immune infiltration (CD3+ and CD68+), with a root-mean-squared error of 163 cells/mm² and an R² of 0.65. Biochemical scores assigned to glycerol, citric acid, and DNA ranked among the strongest contributors to model performance, as measured by feature-importance metrics.
Conclusion
These findings suggest that Raman spectroscopic profiling of tissue biochemistry may provide insight into the biochemical processes associated with immune cell infiltration. By integrating Raman spectroscopy with immunohistochemical analysis, this work highlights the potential for developing tissue-based biomarkers of baseline and radiation-induced immune infiltration in prostate cancer.