THE IMPACT OF AN AGGRESSIVE LOCAL CONTROL POLICY AND OF ADJUVANT CHEMOTHERAPY IN PRIMARY PRESENTATION EXTREMITY ADULT SOFT-TISSUE SARCOMA IN SEQUENTIAL COHORTS OVER 20 YEARS

O’Sullivan B, Bell R, Davis A, Quirt I, Catton C, Wunder J, Kandel R, Cummings B, Fornasier V, Blackstein M, Panzarella T (Princess Margaret Hospital, University of Toronto, Toronto, Ontario, M5G 2M9, Canada)

The purpose of this study is to evaluate the impact on the rate of distant metastasis (DM) of two strategies from sequential eras in adult extremity soft tissue sarcoma at our institution. All registered cases from Jan 1975 to Jan 1986 (Era 1 = E1) and Feb 1986 to June 1996 (Era 2 = E2) were included excepting registrations with recurrent disease or prior "other cancers". Patients with DM at presentation were also included to control for screening bias due to change in imaging over the period. E1 was characterised by a limb conservation approach, frequently with gross total removal and radiotherapy (RT), and with chemotherapy (CT) (41% of cases) as a local and systemic adjuvant. During E2 the approach intended to maximize local control with more rigorous surgical clearance than E1, including re-operation to obtain clear resection margins whenever possible. RT was also frequently used. Adjuvant CT was only rarely used (4%) in E2. Data prior to 1989 came from previous data bases with follow-up updated by two observers. No change to original baseline data was permitted. From 1989 all data were collected prospectively. There were 592 patients (E1: 206; E2: 386) of median age 52 years; males (54 %), female (46%); upper limb (22%), lower limb (78%); low grade (29%), high grade (71%); superficial (20%), deep (80%). All were balanced over both eras including metastases at presentation (E1 = 8% vs, E2 =10.5%). There was a similar proportion of small tumors (<=5cm, 29%) but there was a greater proportion > 10 cm in E2 (p= 0.03); during E2 there was more local extension (LE) to neurovascular structures, skin, or bone (41% vs, 30%, p = 0.005). The 3-year actuarial local relapse free rate was significantly better for E2 (89% vs 73%, p=0.0001), but the 3 year actuarial DM relapse free rate was similar (E2: 69%, E1: 66%, which for both includes the base rate of DM at registration). A Cox regression with DM cases excluded showed that size as a continuous variable (p=0.0001), LE (p=0.0001), high grade (p=0.0001), and deep location (p=0.055) were independent adverse factors for DM. The adjuvant CT vs no CT (p=0.04) and E2 vs E1 (p=0.004) in the model had independent significant protective influences against DM. These results are consistent with the hypothesis that both adjuvant CT and improved local control are independently protective against the risk of DM and resulted in similar rates of DM in both eras of the study.