Introduction
With an incidence of 5.6 per 100,000 per year, diffuse large B-cell lymphoma (DLBCL) is the most common subtype of non-Hodgkin lymphoma (Teras, 2016) and has a 5-year relative survival of 63.8% (SEER Statistics, 2018). While there is not currently a universally accepted timeline for initiation of therapy in DLBCL, retrospective studies have demonstrated that delays in initiation of treatment can lead to adverse outcomes in DLBCL, such as decreased overall survival (Phipps, 2018). We performed a retrospective review of Veterans Affairs (VA) patients nationwide to assess how frequently delays occurred in treatment initiation and we analyzed the impact of time to treatment on response rates to first-line chemotherapy and on survival outcomes.
Methods
We performed a retrospective chart review of 2036 randomly selected records of patients seen within the VA nationwide who were diagnosed with lymphoma between 1/1/2011 and 12/31/2017. We included patients diagnosed with DLBCL. We excluded other types of lymphoma, patients whose workup and treatment were outside of the VA system, and patients with primary central nervous system (CNS) lymphoma. We determined the time from diagnosis to treatment (TDT), defined as the number of days from the date of pathology results to the date of initiation of treatment for each patient, and divided these into two-week blocks. The Wilcoxon-Mann-Whitney test was used to compare median overall survival between the groups.
Results
971 patients were included in the study. Patients were predominantly male (96.2%), with a median age of 67 (Table 1). The median TDT was 20 days. Those with TDT of 0-14 days, 15-28 days, 29-42 days, and 43+ days had an objective response rate (ORR) of 70.9%, 84.5%, 82.9%, and 77.6%, respectively (Table 2). The same groups had a 2-year overall survival (OS) rate of 61.6%, 77.6%, 80.7%, and 72.7%, respectively. They demonstrated median OS of 36.1 months, 46.9 months, 46.4 months, and 47.1 months, respectively. The 2-year OS rates were significantly lower for the 0-14 day group compared to the 15-28 days group (P = 0.0001), 29-42 group (P <0.0001), and 43+ days (P = 0.020). The ORR and median overall survival were also significantly lower in the 0-14 day group compared to the other TDT subgroups.
Although the rates of high International Prognostic Index (IPI) scores and advanced stage disease were highest in the 0-14 day groups, we found that when we assessed only the patients with IPI 3-5, the 2-year OS rates, ORR, and median OS trends were similar, with inferior outcomes in the earlier treatment initiation groups (Table 3). When we analyzed only the patients with stage III-IV disease, as well as only those who were outpatient at time of diagnosis, we again found similar trends of inferior outcomes in those who began treatment within 0-14 days. The Cox proportional hazards model was employed to explore factors associated with survival time (Figure 1). An increase in age, a higher IPI score, and being an inpatient at the time of diagnosis were statistically significant factors that explained a lower OS in the 0-14 day TDT subgroup. When these were included in the model, TDT no longer made a significant difference in OS.
Conclusion
Our study demonstrated that shorter TDT did not improve survival outcomes for this aggressive lymphoma, contrary to what we expected. In fact, patients who had shorter TDT had statistically significant lower survival rates at 2 years, lower median OS, and lower ORR. We theorized that there was a propensity for patients who have more high-risk disease to be started on treatment more quickly because they are more often symptomatic or gravely ill at presentation. We found that when we controlled for inpatient status, high IPI, and advanced stage disease individually, the survival outcomes remained poorest in the earliest treatment initiation groups. However, when we controlled for all of these variables, there was no longer a statistically significant difference in OS based on TDT. Future studies are needed to better understand why our outcomes differed from prior studies which showed improved survival and response in shorter TDT subgroups, and to control for other possible confounders, such as treatment regimens and patients' comorbid conditions.
Disclosures
No relevant conflicts of interest to declare.
Vulvar cancer in Botswana in women with and without HIV infection: patterns of treatment and survival outcomes
