Presence of tumor necrosis is not a significant predictor of survival in clear cell renal cell carcinoma: higher prognostic accuracy of extent based rather than presence/absence classification.

Abstract

The presence of necrosis has been proposed as an adverse prognostic factor in clear cell renal cell carcinoma. However, classification based on a presence/absence basis ignores its heterogeneity, which may be associated with other important pathological factors and prognosis. We performed the first prospective study of necrosis in clear cell renal cell carcinoma to our knowledge and tested the traditional presence/absence classification vs an alternative extent based classification. We studied the presence and extent of tumor necrosis, pathological features and cancer specific survival of 343 consecutive patients. Tumor necrosis was present in 227 tumors (66%) and was associated with more advanced tumors. However, the predictive accuracy for cancer specific survival was low (64.6%) and the presence of necrosis was not retained as an independent prognostic factor on multivariate analysis (p = 0.299). There was significant heterogeneity among tumors with necrosis. Increasing extent of necrosis was associated with poorer performance status, higher T, N, M stages and grades, vascular invasion and sarcomatoid features. Extent based classification predicted cancer specific survival better than presence alone (74.5% vs 64.6%) and was retained as an independent prognostic factor on multivariate analysis (p = 0.029). For clinical use a cutoff of 20% was identified for further prognostic subclassification of tumors with necrosis (c-index 71.7%). Tumor necrosis is an adverse prognostic factor in clear cell renal cell carcinoma but prospective evaluation of necrosis on a presence/absence basis shows that it does not provide independent prognostic information. The predictive accuracy of an extent based classification is superior and is retained as an independent prognostic factor. We recommend the scoring of necrosis according to its extent with further subclassification based on a 20% cutoff.