A prognostic factor analysis for use in development of predictive models for response in adult acute leukemia

M. J. Keating, T. L. Smith, E. A. Gehan, K. B. McCredie, G. P. Bodey, E. J. Freireich

Research output: Contribution to journalArticlepeer-review

85 Scopus citations

Abstract

The pretreatment characteristics of 325 adults with acute leukemia who were treated at the M. D. Anderson Hospital between 1973 and 1977 have been evaluated to assess their value as prognostic indicators. The patient population includes all patients treated with an anthracycline (Adriamycin or rubidazone), cytosine arabinoside, vincristine, and prednisone during the time period. Most patients had one of the variants of acute myelogenous leukemia (75%), and the remaining patients had acute lymphoblastic leukemia (16%) or undifferentiated leukemia (8%). Twenty‐one factors were found to be significantly associated with probability of obtaining a complete response. In addition to characteristics previously known to provide prognostic information such as age, temperature status at the start of treatment, morphology, the presence of Auer rods, sex, and hemoglobin level, we identified the presence of a documented antecedent hematologic disorder and the finding of insufficient metaphases on cytogenetic analysis using the squash technique as being major prognostic variables. In addition, the pre‐treatment biochemical characteristics of hypoalbuminemia and elevated blood urea nitrogen and creatinine were found to adversely influence prognosis. The prognostic significance of factors such as the leukocyte count and platelet count, identified in earlier studies, was not confirmed in this group of patients. From this natural‐history analysis predictive models for response have been developed using multivariate logistic regression techniques. One of these models has been used to evaluate the effect of morphology, treatment, and cytogenetic pattern on response to the combination of drugs used.

Original languageEnglish (US)
Pages (from-to)457-465
Number of pages9
JournalCancer
Volume50
Issue number3
DOIs
StatePublished - Aug 1 1982

ASJC Scopus subject areas

  • Oncology
  • Cancer Research

Fingerprint

Dive into the research topics of 'A prognostic factor analysis for use in development of predictive models for response in adult acute leukemia'. Together they form a unique fingerprint.

Cite this