External validation of prognostic indices for aggressive adult T-cell leukemia/lymphoma (ATL-PI/JCOG-PI) in Okinawa.

Abstract

e19036 Background: Aggressive adult T-cell leukemia/lymphoma (aATL) has an extremely poor prognosis (median OS, 8-10 months). Okinawa, Japan’s only subtropical region, is hyperendemic for aATL. Recently, we demonstrated poorer outcomes among aATL patients in Okinawa compared with patients elsewhere in Japan, and a possible association of strongyloidiasis with ATL-related death. Two prognostic indices (PIs)—ATL-PI and JCOG-PI—have been developed using a database of national surveys and clinical trials. However, aATL patients in Okinawa were not included. This study aimed to validate these PIs using an Okinawa database. We also investigated the impact of strongyloidiasis on aATL patient survival. Methods: We constructed a clinical database of aATL patients from 7 institutions in Okinawa diagnosed between January 2002 and December 2011. The study endpoint was OS. Standard survival analysis methods (Kaplan-Meier method, log-rank test, and Cox proportional-hazards model) were used. Results: The study involved 433 evaluable patients (median OS, 6 months). Risks according to the two PIs in each patient were not always consistent (Table), but both PIs stratified aATL patients by risk. Three-year OS rates for ATL-PI were 35.9% (low-risk, n=66), 10.4% (intermediate-risk, n=256), and 1.6% (high-risk, n=111); rates for JCOG-PI were 22.4% (moderate-risk, n=176) and 5.3% (high-risk, n=257). Strongyloidiasis had little impact on OS (HR and 95% CI from univariate Cox analysis, 1.22 and 0.90-1.66, respectively). Multivariable Cox analysis returned almost the same factors as had been screened out in the previous studies to construct the two PIs. Conclusions: ATL-PI and JCOG-PI were well reproducible in Okinawa database. Strongyloidiasis did not affect prognosis in aATL patients. ATL-PI identifies low-risk aATL patients more clearly than JCOG-PI, and both identify high-risk patients with extremely poor prognosis. These will be useful to devise novel treatment strategies based on risk stratification of all Japan/world aATL patients. [Table: see text]