Factors predicting survival in peripheral T-cell lymphoma in the USA: a population-based analysis of 8802 patients in the modern era

Document Type

Article

Abstract

Current prognostic models for peripheral T-cell lymphoma (PTCL) have multiple limitations, and questions exist regarding applicability to current patients. We utilized the Surveillance Epidemiology and End Results (SEER)-18 database to evaluate factors affecting overall survival (OS) of PTCL in the modern era and identified 8802 patients between 2000-2010. Most subtypes of PTCL increased in incidence during the study period. In univariate analyses, age >55 years, black race, advanced stage, absence of extra-nodal disease, omission of radiation therapy (RT) and high-risk histology each predicted inferior OS (P < 0·0001). Multivariate analysis (MVA) demonstrated that hepatosplenic, enteropathy-associated and extra-nodal Natural Killer/T cell histologies, each had hazard ratios >1·5 (P ≤ 0·0001) for death. Further, age ≥55 years, black race and advanced stage maintained their significance in the MVA (P < 0·0001 each). Based on the significant factors, a prognostic model was constructed and subsequently validated in an independent cohort. The new model incorporated age, stage, histology and race, with an OS ranging from 9 months (highest risk group) to 120 months (lowest risk group). In summary, this is the largest study of PTCL patients in the modern era that provides risk stratification utilizing a new prognostic model that can be incorporated into future prospective clinical trials.

Medical Subject Headings

Adult; Aged; Aged, 80 and over; Disease-Free Survival; Female; Humans; Incidence; Killer Cells, Natural (pathology); Lymphoma, T-Cell, Peripheral (mortality, pathology, therapy); Middle Aged; Models, Biological; Risk Factors; Survival Rate; T-Lymphocytes (pathology); United States (epidemiology)

Publication Date

3-1-2015

Publication Title

British journal of haematology

E-ISSN

1365-2141

Volume

168

Issue

5

First Page

708

Last Page

18

PubMed ID

25382108

Digital Object Identifier (DOI)

10.1111/bjh.13202

Link to Full Text

Share

COinS