Risk‐adjusted prognostic models for Hodgkin's disease (HD) and grade II non‐Hodgkin's lymphoma (NHL II): validation on 6728 British National Lymphoma Investigation patients

Abstract

Summary. Using significant factors from multivariate analyses, based on 20 putative markers from a consecutive series of 1198 Sheffield Lymphoma Group patients, risk-adjusted prognostic models had been previously derived for Hodgkin's disease (HD) (using age, albumin and lymphocyte count) and non-Hodgkin's lymphoma (NHL) grade II (based on albumin, age, erythrocyte sedimentation rate, lactate dehydrogenase and stage). Data from 6728 patients on the British National Lymphoma Investigation database were used for validation: thus the models were applied to 4411 patients with HD and 2317 patients with NHL grade II. Survival curves derived from these validation groups confirmed our risk models. Keywords: prognostic model, lymphoma, BNLI.Although the malignant lymphomas account for less than 10% of all malignant diseases, they are important because many are potentially curable. They can be simplistically divided into Hodgkin's disease (HD), and non-Hodgkin's lymphoma grade I and II (NHL I/II). Many efforts have been made to define prognostic groupings to enable the selection of appropriate therapies. International collaborations have defined groupings for HD (Hasenclever & Diehl, 1998) and aggressive NHL (The International NHL Prognostic Factors Project, 1993). In a previous study (Low et al, 2001), we assessed the significance of a range of putative prognostic markers on a consecutive series of patients seen at one centre over three decades. A multivariate analysis of factors influencing prognosis showed that age, albumin and lymphocyte count were significant in predicting survival (or death) in HD and albumin, age, serum erythrocyte sedimentation rate (ESR), serum lactate dehydrogenase (LDH) and stage were significant in predicting survival (or death) in NHL II. Risk models were then constructed. The aim of the present study was to validate these original predictive models with a second independent patient data set provided by the British National Lymphoma Investigation (BNLI). MATERIALS AND METHODSThe original models had been constructed on data from a consecutive series of 851 patients presenting with HD (n ¼ 315) or NHL grade II (n ¼ 536) to the Sheffield Lymphoma Group (SLG) between 1970 and 1998. All patients had had histopathologically proven and untreated malignant lymphoma. To test and validate these risk models required a second independent data set. The resulting models were, therefore, applied to a series of 6728 patients (HD n ¼ 4411, NHL n ¼ 2317) from the BNLI. Table IA and B shows the distribution of patient characteristics based on the number of risk factors in the SLG and BNLI samples.As in the previous study (Low et al, 2001), we used the Kaplan-Meier method to derive survival curves for the various risk group. The survival curves of the different risk groups were then formally compared by the log-rank test. RESULTSThe SLG model incorporated three risk factors for HD (age, albumin, lymphocyte count) and five for NHL II (age, stage, ESR, albumin and LDH). Older age ( ‡ 45 years), low albumin (£ 35 ...

Cholesterol and serum albumin as risk factors for death in patients undergoing general surgery

“…After adjusting for diseases associated with inflammation, low serum albumin remained significantly associated with death. These results agree with other findings of an association between serum albumin and 5‐year survival in patients with end‐stage renal disease16 and various neoplasms17–21. In contrast to a previous study, in which an inverse trend between serum albumin and death was noted in current smokers only22, a trend was also found in non‐ or ex‐smokers in the present series.…”

Prognostic Implication of the Absolute Lymphocyte to Absolute Monocyte Count Ratio in Patients With Classical Hodgkin Lymphoma Treated With Doxorubicin, Bleomycin, Vinblastine, and Dacarbazine or Equivalent Regimens

“…Lymphocytopenia, a well-established feature of HL, is much more pronounced in advanced disease. The absolute lymphocyte count (ALC) has been independently associated with worse outcomes by several investigators, using different cutoffs [1,4,5]. Severe lymphocytopenia-defined either as lymphocytes ,8% or as an ALC of ,0.6 3 10 9 /L-is one of the seven independent factors of IPS for advanced HL [1].…”