Prognostic factors for patients with advanced non-small cell lung cancer: Univariate and multivariate analyses including recursive partitioning and amalgamation

Nagio Takigawa, Yoshihiko Segawa, Masayuki Okahara, Yoshinobu Maeda, Ichiro Takata, Masaaki Kataoka, Masafumi Fujii

Research output: Contribution to journalArticle

76 Citations (Scopus)

Abstract

In an attempt to determine the prognostic significance of pretreatment factors for patients with advanced non-small cell lung cancer (NSCLC), 24 pretreatment clinical variables were analyzed for 185 patients with NSCLC who underwent chemotherapy and/or radiotherapy between 1985 and 1994. Following univariate analysis, we applied two multivariate statistical techniques. In a Cox regression model, independently significant factors influencing patient survival included performance status (PS), disease stage, hemoglobin level, and serum calcium level. Recursive partitioning and amalgamation (RPA) resulted in three distinct prognostic subgroups based on PS, stage, weight loss, and hemoglobin level. The best survival was observed for patients with a good PS and Stage III disease who had a hemoglobin level > 11 g/dl. The worst survival was observed for patients with a poor PS and presence of weight loss irrespective of stage. All other patients had an intermediate prognosis. Median survival times were 95.1 weeks, 17.1 weeks and 39.3 weeks, respectively (P < 0.00005). The results of our analyses show that three important prognostic subgroups could readily be discerned using RPA.

Original languageEnglish
Pages (from-to)67-77
Number of pages11
JournalLung Cancer
Volume15
Issue number1
DOIs
Publication statusPublished - Aug 1996
Externally publishedYes

Keywords

  • Chemotherapy
  • Non-small cell lung cancer
  • Prognostic factor
  • Radiotherapy
  • Recursive partitioning and amalgamation

ASJC Scopus subject areas

  • Oncology
  • Pulmonary and Respiratory Medicine
  • Cancer Research

Fingerprint Dive into the research topics of 'Prognostic factors for patients with advanced non-small cell lung cancer: Univariate and multivariate analyses including recursive partitioning and amalgamation'. Together they form a unique fingerprint.

  • Cite this