Pretreatment quality-of-life score is a better discriminator of oesophageal cancer survival than performance status

Abstract

OBJECTIVES: Performance status [Eastern Cooperative Oncology Group (ECOG)] is a physician-assigned score indicating a patient’s fitness for treatment. Functional assessment of cancer therapy-esophagus (FACT-E) is a patient-reported, health-related quality-of-life (HRQOL) instrument containing an oesophageal cancer subscale (ECS). Our objective was to assess the discriminative ability of pretreatment FACT-E and ECS when compared with performance status in predicting survival in patients with Stage II–III oesophageal cancer. METHODS: Patient data from four prospective studies were pooled together. These four studies included oesophageal patients who received chemoradiation either as neoadjuvant therapy or as definitive therapy. Three separate Cox regressions were performed considering FACT-E, ECS and ECOG as the main predictors, respectively. Receiver-operating characteristics analyses were performed. RESULTS: Of the 120 curative intent patients, 39.8% (n = 51), 58.6% (n = 75) and 1.6% (n = 2) had ECOG 0, 1 and 2, respectively. On Cox regression analysis, pretreatment FACT-E (P = 0.04) and ECS (P = 0.004) but not ECOG (P = 0.27) were independently associated with overall survival. ECOG could not discriminate between survivors and non-survivors (P = 0.28) with an area under the curve (AUC) of 0.56 [95% confidence interval (CI): 0.45–0.66], whereas FACT-E (P = 0.02) and ECS (P < 0.001) were discriminative with AUC = 0.63 (95% CI: 0.52–0.73) and AUC = 0.69 (95% CI: 0.60–0.79), respectively. CONCLUSIONS: In patients with Stage II–III oesophageal cancer being considered for curative therapy, pretreatment FACT-E and ECS have better discrimination for survival than does ECOG. The majority of patients were ECOG 0/1. Thus, these patient-derived scores were able to discriminate survivors from non-survivors even within this constrained range of clinician-assigned performance status. This highlights the potential utility of FACT-E and ECS as prognostic tools.