Prediction of tumor control probability in prostate cancer radiotherapy using a biophysical model incorporating cancer stem cell and hypoxia.

BACKGROUND

This study aims to bridge the gap between cellular radioresponse mechanisms and clinical interpretation by developing a biophysical model that predicts cell inactivation through integration of cancer stem cell (CSC) characteristics represented by the side population (SP) and the oxygen enhancement ratio (OER) within the tumor microenvironment.

MATERIALS AND

METHODS

SP cells, characterized by high Hoechst 33342 efflux, and the main population (MP) were isolated from the DU145 prostate cancer cell line. X-ray-induced initial DNA double-strand breaks (DSBs) and cell inactivation were evaluated via γ-H2AX/53BP1 colocalized foci and colony formation, respectively. Cellular responses under hypoxia (<0.0% O2) were assessed using a hypoxia culture kit (nBIONIX-2,Sugiyamagen, Japan). The measured cell survivals were analyzed using an updated integrated microdosimetric (IMK) model incorporating MP/SP fractions and OER. The model was further applied to estimate tumor control probability (TCP) under various fractionation regimens (2 Gy, 3 Gy, and 7 Gy/fx) and to predict clinical freedom from biochemical relapse (FFBR) based on literature reported outcomes.

RESULTS

In vitro survival measurements and predictions consistently showed higher radioresistance in SP than in MP. By assuming a population-independent OER and measured MP/SP fractions, the dose-response of the parental population was accurately reproduced under both normoxia and hypoxia. The findings indicate that initial DSB yield is a key determinant of radiosensitivity, influenced by oxygenation status and cellular characteristics. Using intratumoral hypoxic volume as an input parameter, the model accurately reproduced FFBR for 2 Gy/fx, whereas reduced OER values were required to match clinical outcomes for 3 Gy/fx and 7 Gy/fx.

CONCLUSION

The developed IMK model successfully reproduced in vitro cell survival and clinical FFBR by integrating cellular parameters (MP/SP fraction and DSB yields) and tumor oxygenation (OER). The present model supports the use of a common OER for predicting curative effects in both MP and SP and may aid in optimizing alternative radiotherapy strategies, including dose-painting for hypoxic tumor regions.