PROGNOSTIC FACTORS FOR PATIENTS WITH LOCALIZED PRIMARY MALIGNANT FIBROUS HISTIOCYTOMA - A MULTICENTER STUDY OF 216 PATIENTS WITH MULTIVARIATE-ANALYSIS

BACKGROUND. The purpose of this study was to determine the independent prognostic variables in a well documented subset of 216 patients with localized primary malignant fibrous histiocytomas (MFH). METHODS. Bet ween the years 1980 and 1989, 216 patients with localized, primary (In ternational Union Against Cancer [UICC]/American Joint Committee on Ca ncer [AJCC] Stage I-IVA) MFH were evaluated and treated in 10 particip ating centers of the sarcoma group of the French Federation of Cancer Centers (FNCLCC). Clinicopathologic factors were collected retrospecti vely and entered into a cooperative database. Tissue slides of all cas es were jointly reviewed microscopically by the pathology subcommittee . Surgical treatment was performed on all but 6 (3%) patients. One hun dred ninety-five patients (90%) were free of gross disease, with compl ete local control at the end of the initial treatment. The adjuvant tr eatment was radiotherapy in 78 patients (36%), chemotherapy in 19 pati ents (9%), and both in 61 patients (28%). RESULTS. The median follow-u p was 3.5 years (range, 45 days to 12 years). Five-year actuarial rate s of disease specific (DSS), metastasis free (MFS), and local recurren ce free (LRFS) survival were 70%, 63.3%, and 62.7%, respectively. Mult ivariate analyses showed that the adverse prognostic factors independe ntly associated with decreased disease specific survival were UICC/AJC Stage III + IVA (P < 0.00001; relative risk [RR], 3.27; 95% confidenc e interval [CI], 1.6-6.58), residual macroscopic disease following pri mary local therapy (P = 0.00024; RR, 3.99; Cl, 2.04-7.82), deep tumor location (P = 0.0045; RR, 3.37; Cl, 1.21-9.38), non-myxoid histology ( P = 0.0056; RR, 9.28; Cl, 1.03-83.41), and age older than 50 years (P = 0.037; RR, 2.19; Cl, 1.04-4.61). Two factors were significantly rela ted to MFS in the patients with the poorest prognosis: histopathologic Grade 3 (P < 0.0001, RR, 3.46; Cl, 2.02-5.91) and tumor size greater than 8 cm in largest dimension (P = 0.0012; RR, 2.78; Cl, 1.36-3.66), With regard to LRFS, patients who did not undergo radiotherapy had red uced local control (P = 0.0043; RR, 2.36; Cl, 1.46-3.83). CONCLUSIONS. Resection of all macroscopic disease was independently associated wit h improved disease specific survival and adjuvant radiotherapy signifi cantly decreased the local relapse risk. Histopathologic grade was the most important prognostic factor for DSS and MFS. (C) 1996 American C ancer Society.