Improvement in cancer survival rates constitutes such intuitively forceful evidence of progress in cancer treatment that its corollary, improvement in mortality rate, appears self-evident. However, Welch et al,1 employing data from the nationally compiled Surveillance, Epidemiology, and End Results program, found no overall correlation between increased survival and mortality rates for 20 of the most common types of solid tumors. Survival (5-year) is the percentage of individuals alive 5 years following a cancer diagnosis. Incidence and (cause-specific) mortality rates are age-adjusted, population-based rates. Incidence is the number of new cases per 100,000 population per year, and the (cause-specific) mortality rate is the number of deaths (from the disease) per 100,000 population per year. Increased survival, with no change in mortality rate, is most often attributable to lead-time bias, in which improvements in ascertainment permit diagnosis at an earlier point in time without affecting longevity (outcome). Thus, earlier diagnosis of a highly lethal and untreatable neoplastic disorder would improve survival without affecting mortality rate. Less obviously, if an increase in longevity resulting from intervention in some individuals was offset by a delayed decrease in longevity in others, survival would improve and the mortality would remain unchanged. Survival and mortality are not complementary measures.