Jacques Rouëssé

The pathological features of 155 adult patients with soft-tissue sarcomas were studied retrospectively, in an attempt to set up a grading system for these tumors. As the first step, seven histological criteria (tumor differentiation, cellularity, importance of nuclear atypia, presence of malignant giant cells, mitosis count, pattern of tumor necrosis and presence of vascular emboli) were evaluated in a monofactorial analysis. Five of these (tumor differentiation, cellularity, mitosis count, tumor necrosis, and vascular emboli) were correlated with the advent of metastases and with survival. A multivariate analysis, using a Cox model, selected a minimal set of three factors (tumor differentiation, mitosis count, and tumor necrosis) the combination of which was necessary and sufficient to retain all the prognostic information. A grading system was elaborated, which turned out to be correlated with the advent of metastasis and with patients' survival. A second multivariate analysis introducing clinical prognostic features showed that the histological grade was the most important prognostic factor for soft-tissue sarcomas. Thus, this grading system appears to be highly interesting because of its prognostic value and the facility of its elaboration. However, its reproducibility should be tested.

Tumor grade has been proposed as an essential factor in the staging of patients with soft tissue sarcomas. In a previous study, a histopathologic grading system using the evaluation of tumor differentiation, mitosis count, and tumor necrosis was described. The current study was conducted to test its reproducibility. The pathologic sections of 25 soft tissue sarcomas were submitted to a study group composed of 15 pathologists who had not been involved in the development of the grading system. The results were compared with those of a panel group. The crude proportion in agreement observed between the study group and the panel group was 81% for the evaluation of tumor necrosis, 74% for tumor differentiation, and 73% for the mitosis count. The crude proportion in agreement for the tumor grade was 75%, which was significantly better than the crude agreement rate of 61% for the diagnosis of histologic type (P = 0.001). A kappa statistical analysis, to check the possibility of chance-related concordance, showed a proportion in agreement of 68%. A two-way variance analysis showed that the homogeneity of the evaluation of tumor grade is impaired by tumor-related and observer-related factors. However, an improvement may be obtained by better training of pathologists. We conclude that the tumor grading system developed inside the French Federation of Cancer Centers, although perfectible, already provides reliable prognostic information and its use in prospective clinical studies may provide more information about its clinical usefulness.

PURPOSE: A randomized pilot study was undertaken to assess the acute and chronic toxicities of two short intensive chemotherapy regimens, and to evaluate the feasibility of conservative surgery in this setting. Additional aims were to determine the clinical and radiologic response and the degree of histologic necrosis after chemotherapy. With extension of the study, eventual accrual was sufficient to compare disease-free survival (DFS) and overall survival (OS). PATIENTS AND METHODS: Between July 1983 and December 1986, the European Osteosarcoma Intergroup (EOI) entered 198 eligible patients with classic high-grade extremity osteosarcoma onto a randomized trial that compared doxorubicin (DOX) 25 mg/m2/d times three, intravenous (IV) bolus plus cisplatin (CDDP) 100 mg/m2, 24 hour infusion, every 3 weeks times six; the same combination was preceded 10 days earlier by high-dose methotrexate (HDMTX) 8 g/m2, 6-hour infusion, every 4.5 weeks times four. In the majority of patients (179), chemotherapy was commenced after biopsy; definitive surgery was scheduled at 9 weeks in both groups. RESULTS: Toxicities for both regimens did not differ substantially from those that occurred in other trials of adjuvant chemotherapy in osteosarcoma. Local recurrence (9%) and surgical complications (18%) after conservative surgery were acceptable. With a median follow-up of 53 months, DFS at 5 years is superior (P = .02) for DOX/CDDP, 57% versus 41%, although OS, 64% versus 50%, is not different significantly (P = .10). In a subset of 66 patients for whom pathologic data on the resected specimen were available, DFS (P = .003) and OS (P = .008) were better for those who demonstrated > or = 90% necrosis. CONCLUSION: A brief intensive chemotherapy regimen of DOX/CDDP has produced excellent long-term results, which are similar to those that have been achieved in cooperative group studies of longer, more complex multiagent chemotherapy, and provide the basis for a direct comparison in the next EOI study.

In a study of 1,010 patients with solitary, unilateral, nonmetastatic breast cancer, the histologic grade, assessed by a multifactorial analysis (Cox model) to study its significance with other prognostic factors, was found to be an important, independent factor. For 612 operable patients, two laboratory characteristics, the number of histologically positive nodes and the histologic grade, were the most valuable predictors. These two factors alone form a predictive index that may be an excellent and simple guide for the clinical decision of subsequent therapy. For 398 patients with inoperable breast cancer (ie, tumor greater than or equal to 7 cm, N2-3, inflammatory, skin fixation, and clinically rapidly growing forms), the histologic grade (performed on drill or cutting needle biopsy) was again a most important (and with inflammatory forms the most important) predictor of prognosis in these patients. Our data support that performing our modified histoprognostic grading of Scarff and Bloom is simple, reproducible, incurs no additional cost, may be carried out in the simplest histology laboratory, and finally, defines an important risk factor in all patients. It should be routine for all breast cancer specimens. Furthermore, studies of adjuvant therapy should stratify patients for this variable.