Risk score application in patients older than age 70 with newly diagnosed non-Hodgkin lymphoma (NHL): A single-institution experience in midwestern Ireland.

Abstract

e19508 Background: Cytotoxic therapy in elderly patients with cancer is historically controversial; this is often compounded by a lack of robust prospective data. The elderly account for a significant proportion of NHL diagnosed, and determining the appropriateness and level of intervention is best established on an individual basis. To reduce negative bias to potential treatment tolerance, stratification reliant on reproducible risk scores, such as age-adjusted International Prognostic Index (aaIPI), augments clinical assessment. We herein report our retrospective experience of newly diagnosed patients over 70-years and risk score application. Methods: Study period: 01/01/2001 – 31/12/2010. Data obtained from (1) Mid-Western Cancer Centre (MWCC) Oncology Database (2) MWRH pathology records (3) MWRH patient-files. Data was collated and entered into an Access database and exported into Predictive Analytics Software (PASW) for analysis. Results: Forty patients identified. M:F = 19:21. Mean age: 78 years (Range: 70 – 92 years). Median follow: 11.5 months (Range: 1 – 86 months). aaIPI groupings, low-risk, 45% (18/40), intermediate-risk 17.5% (7/40) and high-risk 17.5% (7/40); mean survival for low-risk, intermediate-risk and high-risk was 62.1 months, 40.1 months and 14.7 months (p = 0.044). For our cohort, a MWCC risk score was devised, by assigning score of 1 for patients with anaemia and extra-nodal involvement; this was added to aaIPI score. Patients were then grouped by MWCC risk score, good-risk, 15.6%, moderate-risk, 53.1% and poor-risk, 31.3%; mean survival for, moderate-risk and poor-risk respectively was 53.4 months, and 15.4 months (p = 0.014); no deaths seen in good-risk grouping, hence no mean survival available. Conclusions: Risk score tools are helpful in evaluating and predicting possible outcome. aaIPI certainly correlated well with outcome for patients. MWCC risk score also correlated well but needs validation and multivariate analysis with a larger dataset; further work is needed before this could be adopted. Such models and scoring systems are useful, when applied correctly, but are not an alternative to good clinical judgement.