Other Cancers in Long-Term Survivors of Chronic Lymphocytic Leukemia: Incidence and Prognostic Relevance

Introduction. The clinical course of chronic lymphocytic leukemia (CLL) is mostly indolent. About one third of the patients are managed with lifelong watch-and-wait (WW) and those who receive therapy often achieve a durable remission. As a result, the majority of patients with CLL will live with their disease for long periods of time, and be exposed to several complications, including the occurrence of other cancers (OC). Patients with CLL may have an increased incidence in OC. Published reports indicate an incidence of 3-27%, mostly in treated patients, however, very little is known on OC in patients with CLL not requiring therapy. Furthermore, observation time in published studies is limited to <5 years, and the incidence of OC in patients followed for longer than 10 years is unknown. We, therefore, studied the incidence and prognostic impact of OC in treatment-naïve patients with CLL followed for ≥10 years.

Methods. We reviewed our database and identified all patients with CLL untreated at the time of referral. We selected long-term survivors (LTS), defined as patients with a follow-up ≥10 years, and analyzed the incidence and prognostic impact of OC in this population. Non-melanoma skin cancers were excluded since these were diagnosed and treated promptly in virtually all cases and felt not to have prognostic impact. Standardized incidence ratios (SIR) were calculated for OC occurring after the diagnosis of CLL that were reportable to the Surveillance, Epidemiology and End Results program.The estimated overall survival (OS) according to the presence of OC was plotted considering OC as a time-dependent covariate.

Results. We identified 797 LTS of CLL seen at our institution between 1957 and 2003. Median age was 56 years (24-88). 57% of patients were males. Median follow-up for the entire population is 154 months (120-485). We recorded 383 OC in 286 (36%) patients. 76/286 (26%) patients had >1 OC (62 had 2 OC, 10 had 3, 2 had 4, 1 had 5 and 1 had 6).The firstOC preceded or was diagnosed concomitantly with CLL in 100 patients (35%), while in the remaining 186 (65%) it occurred later during the course of the disease. 570 patients (71%) required treatment for CLL. Median time to treatment was 18 months (0-454). In treated patients, the cumulative frequency of OC was 205/570 (36%) and in WW patients 81/227 (36%). The SIR for all OC was 1.2 (p = .034). Males and patients younger than 60 years had a significantly higher incidence of OC (SIR 1.31 and 1.27, respectively). Among OC types, secondary leukemia, melanoma and head and neck cancers had the highest observed-to-expected ratio. Surprisingly, lung, digestive tract, and bladder cancer had a lower-than-expected incidence (table). 474 patients (59%) are alive. 222/570 (39%) treated patients and 101/227 (44%) WW patients have died. The median OS was longer in patients without OC (279 months) vs. those with OC (189 months). Independent predictors of shorter survival in multivariate analysis included higher creatinine, the presence of OC, and older age.

Discussion. This is the first study to address the incidence of OC in LTS of CLL, including WW patients. In our population, the frequency of OC is similar in treated and WW patients. Although the incidence of OC in LTS of CLL is higher compared to matched general population, the incidence of lung, digestive and bladder cancer is lower than expected. Reasons of this finding remain to be identified.The occurrence of OC is an independent predictor of shorter survival, thus constituting a relevant competing risk of mortality in LTS of CLL.