Are Areas of Isolated Atypical Cells or Residual Tumor Cartilage Prognostically Significant in Neoadjuvant Treated Osteosarcoma? An Analysis of 106 Consecutive Cases from a Single Institution

Abstract

Treatment of high-grade osteosarcoma consists of neoadjuvant chemotherapy then surgical resection. Generally, < 90% tumor necrosis predicts an unfavorable prognosis and can influence post-operative therapy. Thus, accurate, reproducible, microscopic measurement of tumor necrosis is imperative. Treated osteosarcomas contain a variety of morphologic patterns making quantitative assessment difficult. Whether residual tumor cartilage or widely dispersed atypical cells in fibrous stroma represent prognostically significant "viable" tumor is unknown and, to our knowledge, has not been formally tested. Full cross sections of the post-chemotherapy resection specimens of 106 osteosarcomas were evaluated for the presence and pattern of viable tumor using two methods. First, we counted as viable only dense areas of mitotically-active atypical cells ("stringent"). Second, we also included as viable rare (<25 / 400X field), atypical cells in fibrous stroma and/or tumor cartilage ("inclusive"). Treatment response was judged as good (>90% necrosis) or poor (<90% necrosis) and correlated with disease-free survival using the Kaplan-Meier method. Eleven tumors (10%) had no evidence of residual tumor after chemotherapy. The remainder contained different proportions of tightly-packed, atypical, mitotically-active cells (n=73, 69%), isolated atypical cells in fibrosis or necrosis (n=35, 33%) and/or tumor cartilage (n=10, 8%). Disease-free survival correlated significantly with chemotherapy response using the "stringent" method (p=0.008) but not the "inclusive" method (p=0.34). These data confirm that the percent tumor necrosis after chemotherapy is an important prognostic indicator, but areas of tumor cartilage or widely distributed, mitotically-inactive, atypical cells should not be measured as residual viable tumor.