Comparison of in vivo skin and in vitro blood lymphocyte models for the prediction of late normal tissue responses in breast radiotherapy (RT) patients.

Abstract

11116 Background: Critical opinions for the lack of success of DNA double-strand break (DSB) repair as a predictive marker of normal tissue radiosensitivity include the argument that in vitro cellular responses correlate poorly with in vivo responses due to the modifying influence of tissue environment. In this study, we test the hypothesis that a DNA damage assay based on in vivo irradiated skin tissues better predicts clinical responses in human skin, as opposed to the same assay performed in ex vivo irradiated lymphocytes. Methods: DSB levels (24 h post-4 Gy) were quantified using γH2AX/53BP1 immunostaining in irradiated skin tissues and G0 lymphocytes of 35 breast RT patients. Patients were selected on the basis of late RT effects in their breast and individuals with marked or minimal effects were classified as cases and controls, respectively. Risk factors of late effects established from multivariate analyses of outcomes of two breast RT trials were also considered in patient selection. They were 1) total RT dose, 2) RT dosimetry, 3) tumour bed boost, 4) breast size, 5) surgical cavity, and 6) axillary treatment. Results: Clinical parameters were balanced in both patient groups. Residual foci levels in skin epidermis and dermis were comparable between cases (n = 20) and controls (n = 15). Mean foci per cell were 3.29 in cases, 2.80 in controls for dermal fibroblasts (p = 0.07); 3.28 in cases, 2.60 in controls for endothelial cells (p = 0.08); 2.87 in cases, 2.41 in controls for superficial keratinocytes (p = 0.45); 2.32 in cases, 2.35 in controls for basal keratinocytes (p = 0.27). Residual foci levels in lymphocytes were however significantly higher among cases (foci per cell = 12.1) compared to controls (foci per cell = 10.3, p = 0.01). Of the different cell types, only residual foci levels of dermal fibroblasts and lymphocytes correlated with clinical severity (R = 0.722, p < 0.001; 0.593, p = 0.01, respectively). Interestingly, foci levels were not correlated between skin cells and lymphocytes of the same patients. Conclusions: DSB repair of ex vivo irradiated lymphocytes appears to be a better predictive marker of late effects to breast RT than DSB repair of in vivo irradiated skin.