Analysis of a surgical series of 21 cerebral radiation necroses

Abstract Background There is no standard approach to differentiate cerebral radiation necrosis from tumor recurrence and no standard treatment pathway for symptomatic lesions. In addition, reports on histology-proven radiation necrosis and the underlying pathophysiology are scarce and highly relevant. Methods Our monocentric, retrospective analysis included n=21 histology-proven cerebral radiation necroses. Our study focused on i) potential risk factors for the development of radiation necrosis, ii) on radiological and histopathological features of individual necroses and iii) on the suitability of previously published, MRI-based methods to identify radiation necroses based on specific structural image features. Results Average time between radiation treatment and development of necrosis was 4.68y (95% CI, 0.19y - 9.55y). Matching available MRI data sets with that of patients suffering from tumor lesions, we compared specificity and sensitivity of 3 previously published methods to identify radionecrosis based on imaging criteria. In our hands, none of these methods reached a sensitivity ≥ 70 %. Radionecrosis presented with large edema and showed elevated levels of cell proliferation as inferred by Ki-67 staining. Surgical removal of radiation necrosis proved to be a safe approach with low permanent morbidity ( Conclusions While the overall incidence of cerebral radiation necrosis is low, our data suggest an increasing incidence over the last two decades, which is likely associated with the use of stereotactic radiotherapy. There are no current imaging standards to identify radiation necrosis on standard MRI with structural sequences. Surgical removal of radiation necrosis is associated with low morbidity and mortality.