Re: Rising Incidence of Small Renal Masses: A Need to Reassess Treatment Effect

Abstract

The trends in mortality rates according to tumor size for renal cell cancer (RCC) patients that were calculated by Hollingsworth et al. ( 1 ) are misleading because most RCC patients survive beyond the year that they are diagnosed. The prolonged survival means that the mortality rates for patients diagnosed during a confined period (1983–2002) should be low during the initial years and increase over time as more patients diagnosed during the early years die. In Table 1 , we present two sets of total mortality rates based on data from the Surveillance, Epidemiology, and End Results (SEER) incidence-based mortality file ( 2 )—one set for cases diagnosed during 1983–2002 (henceforth referred to as the restricted period) and the other for all cases diagnosed since the beginning of the SEER program (1973–2002, henceforth referred to as the whole period). Using the restricted dataset, the total mortality rates increased from 1.6/100000 in 1983 to 6.5/100000 in 2002, an increase of nearly 300% that is comparable to that reported by the authors. In contrast, based on the whole dataset, the increases in total mortality rates during 1983–2002 were much smaller, from 4.6/100000 in 1983 to 6.8/100000 in 2002, a 48% increase. The discrepancy in the increase in mortality rates was mostly due to higher mortality rates during the 1980s in the whole dataset relative to rates in the restricted dataset; the mortality rates in later years were more comparable. It is noteworthy that the mortality rates for patients with tumors of unknown size in the whole dataset declined steadily during the same time period, while rates in the restricted dataset stayed relatively constant. The difference in the trends was due mainly to omission of case subjects diagnosed before 1983, for whom information on tumor size was not available, and the declining contributions of these subjects to subsequent deaths over time. Similar discrepancies in mortality rates between the whole and restricted datasets emerged when only deaths due to kidney cancer were included ( Table 1 ).