Prognostic impact of the presence or absence of prior cancer in patients with cancer using cure models: A population-based study.

Developing a subsequent cancer is one of the major concerns for cancer survivors; however, whether prior cancer could affect their prognosis is unknown. We therefore aimed to analyze how prognosis varies depending on prior cancer in patients with newly developed cancer, focusing on cancer that had been "cured." We used the record-linked database of the Osaka Cancer Registry and Vital Statistics to select 186,798 patients with stomach, colorectal, or lung cancer aged ≥40 years from 1995 to 2009 in Osaka, Japan. These cancers were defined as index cancers. We classified the patients into two groups according to whether they had a prior cancer diagnosis within 10 years before the index cancer diagnosis. The cured proportion was defined as the proportion of cancer patients with the same mortality as the general population and was estimated using the parametric mixture cure model. The cured proportion of patients with prior cancer by sex and age group was not significantly lower than those without prior cancer, except for patients with stomach cancer aged ≥65 years. According to the index cancer stage in the localized stomach or colorectal cancer, the cured proportion in patients with prior cancer was lower than in those without prior cancer. However, at any stage of lung cancer, the proportion of patients with prior cancer who had been cured was similar to patients without prior cancer, therefore prior cancer had a prognostic impact only in some patient groups based on the characteristics of their index cancer.