Clinical indicators of prognosis in unresected lung cancer

A prognosis estimate in palliative treatment for inoperable lung cancer is thought to be of value. Performance status is now recognized as being of major importance, but only recently has been consistently available. We examined other simple information not requiring clinical or laboratory tests that is available in the clinical histories of a series of 1,839 patients with unresected lung cancer. Actuarial survivals at 1, 2, and 5 years were 21.9, 7.8, and 2.0 percent, respectively. The median survival rate was 24 weeks. Squamous cell histology and an increasing interval between the first symptoms and diagnosis were associated with a better prognosis. The number of symptoms recorded at the time of assessment had strong negative association with survival; asymptomatic patients had a two-year survival of 26 percent compared with 6 percent or less with four or more symptoms. A similar discrimination is given by Feinstein's index, which combines information on the number and type of symptoms and on the interval between first symptom and diagnosis. The clinical stage was strongly related to survival. Additional statistical analysis showed that the prognostic value of each of the most powerful prognostic factors, the number of symptoms, and Feinstein's index was little altered by the use of data on stage and histology in addition. The data show a range of median survival of 18 to 60 weeks for Feinstein's indices from 6 to 1, suggesting that the symptom index is useful particularly in the relatively well patients and the performance status particularly in those more ill. The combination may be better than either alone, and it is recommended that such information be recorded for all lung cancer patients.