Analysis of Prognostic Factors and Application of International Prognostic Index Model to Determine the High Risk Group for the Treatment of Aggressive Non - Hodgkin's Lymphoma / 대한내과학회지

ABSTRACT

OBJECTIVE: Although the therapeutic outcome of aggressive non-Hodgkin's lymphoma (NHL) has been considerably improved by the introduction of combination chemotherapy, many patients still fail to achieve complete response(CR) and/or long-term survival. Because the outcome appears to depend on certain prognostic factors, long term prognosis can be predicted by identification of risk group. And also, the patients in high risk group may benefit from new therapeutic modality. In 1993, the international prognostic index model for aggressive NHL as developed far the purpose of predicting outcome and designing of therapeutic trial. Thus, analysis of prognostic factors was performed to identify independent factors for the end points of CR, overall survival, and disease-free survival. METHODS: From 1989 to 1994, total 340 patients were treated with combination chemotherapy and/or radiotherapy for NHL in Korea Cancer Center Hospital. Among 340, informations on eleven prognostic factors(sex, age, performance status, Ann Arbor stage, serum LDH level, tumor size, number of extranodal disease sites, bone marrow involvement, presence of B symptom, sex, time to CR, and histologic grade) were avaliable for 273 patients. Among these, 221 patients with aggressive NHL(NCI clinical schema) were eligible for the prognostic factor analysis for the response and survival. Also, 186 patients were eligible to determine whether International Prognostic Index Model could be applicable for Korean NHL. RESULTS: One hundred fifty patients(68%, 95% CI 62-74%) achieved a complete remission, 43 patients (20%) a partial remission. With a median follow-up of 3,5 years, overall 3 year survival rate was 6396, and 3 year DFS for the 150 CRs was 72%. In a univariate analysis for the CR and survival, Ann Arbor stage, number of extranadal disease, performance status, presence of B symptoms, presence of BM involvement, serum LDH level and histologic grade were found to be statistically significant prognostic factors. Among them, by multivariate analysis, number of extranodal disease(RR 0.2, 95% CI 0.1-0.7), B Symptoms (RR 0.4, 95% CI 0.2-0.9), and histologic grade(RR 0.2, 95% CI 0.08-0.7) showed to be independent adverse prognostic factors for CR. For disease-free survival, Ann Arbor stage(RR 2.6, 95% CI 1.1-6.4) was independent risk factor. For overall survival, number of extranodal involvement(RR 2, 95% CI 1.3-4) and histologic grade(RR 2, 95% CI 1.2-3.7) were independently significant prognostic factors. With these 2 independent prognostic factors for survival, we could establish a prognastic index model which could separate the high risk patients. However, the usefulness of this model should be confirmed in a larger patient population. The dose intensity of cyclophosphamide, during initial 3 months of treatment, was significantly associated with CR rate and overall survival(p=0.01 and 0.03, respectively). When International Prognostic Index Model was applied to our patients, patients in the lower risk groups had significantly better outcome than patients in the higher risk groups(3 year survival and RR 77% and 1 for low risk group, 61% and 1.9 for low-intermediate risk group, 50% and 2.2 for high-intermediate risk group, and 25% and 6 for high risk group). CONCLUSION: In this study, we confirmed that features other than the Ann Arbor stage were independently associated with CR and survival, and the International Prognostic Index Model would be an useful tool for the selection of high-risk patients who could be benefited from more aggressive chemotherapy.