Trends in cancer prognosis in a population-based cohort survey: can recent advances in cancer therapy affect the prognosis?

Abstract

Background: The aim of the study was to investigate trends in cancer prognosis by examining the relationship between period of diagnosis and probability of death from cancer in a population-based cohort.

Methods: Within a cohort of Japanese men and women aged 40-69 years and free of prior diagnosis of cancer and cardiovascular disease at baseline, data from 4403 patients diagnosed with cancer between 1990 and 2006 and followed up until 2012 were analyzed using survival regression models to assess the presence of an effect of the period of diagnosis (before 1998 versus after 1998) on the risk of dying from cancer.

Results: We noted a significant decrease in risk of dying from cancer among individuals diagnosed after 1998 with lung cancer (hazard ratio [HR]=0.676 [0.571-0.800]) or colorectal cancer (HR=0.801 [0.661-0.970]). A decrease in the estimated five-year probability of death from cancer was also noted between the first (before 1998) and the second (after 1998) period of diagnosis for lung and colorectal cancers (e.g., 85.4% vs. 73.3% for lung cancer and 44.6% vs. 37.7% for colorectal cancer, respectively, for stage III in men aged 60 at diagnosis).

Conclusions: This study presented the first scientific evidence of improvement in prognosis for lung and colorectal cancer patients in a population-based cohort in Japan. Our results suggest that recent advances in cancer treatment could have influenced cancer survival differently among lung, colorectal and gastric cancers.

Keywords: Cancer; Population-based cohort; Survival; Trends.