Purpose: The survival probability of patients with lung cancer is usually based on the extent of disease as assessed at the time of diagnosis. The discouraging 5-year survival is often reported (< 10%) without taking into account changes in the survival probability as time advances from diagnosis.
Patients and methods: Conditional survival estimates by sex, age, extent of disease, and histology were estimated for patients diagnosed with lung cancer in Denmark from 1943 to 1997. Survival probabilities were calculated by the Kaplan-Meier method, and cumulative survival estimates were used to derive conditional survival estimates.
Results: For every additional year survived, the probability of surviving the next 5 years increases from 33% (men) and 36% (women) after the first year, to 60% (men) and 67% (women) who have survived 5 years. The 5-year survival probability of patients younger than 49 years who had survived the first year was 33%, and increased to 81% after the fifth year. Corresponding estimates for 60- to 69-year-old patients were 23% and 52%. The conditional survival differed greatly among patients with localized and regional disease (29% and 10%, respectively) in the first year, but converged with time (52% and 47%, respectively) after 5 years. The conditional survival is similar in patients with squamous cell carcinoma, adenocarcinoma, and large-cell carcinoma, but is markedly lower in patients with small-cell carcinoma.
Conclusion: For patients who have survived more than 1 year, the conditional survival probability provides a more accurate estimate of survival as compared with the conventional observed survival rates.