AutoSCT and AlloSCT have both been shown to be effective for the treatment of relapsed/refractory FL. However, outcomes with either modality are not well defined in the post-rituximab era, especially regarding autoSCT. The NCCN NHL Outcomes Database was utilized to examine survival for FL patients (pts) refractory or relapsed after rituximab-based therapy who underwent subsequent autoSCT or alloSCT. Further, prognostic factors were investigated.

The NCCN NHL Outcomes Database is a prospective cohort study collecting comprehensive clinical, treatment, and outcomes data for NHL pts at 7 participating NCCN centers. Among all NHL pts in the database (n=5,395), 1,670 had FL, of whom 240 had SCT during the study observation period of 1/1/00 to 12/31/09 (follow-up through May 2011). Pts were excluded if they 1) did not have relapsed/refractory FL (n=13) or 2) had not received prior rituximab (n=11). In total, 216 pts (autoSCT n=158, alloSCT n=58) were included for analysis. Median follow-up was 2.9 years. Univariate Cox proportional hazards regression was used to assess associations of prognostic factors within each type of SCT for overall survival (OS) and failure free survival (FFS). OS was defined as years from SCT to death; FFS was defined as relapse, transformation, disease progression, or death. Variables with a p <0.20 were entered into a multivariate Cox model within each type of SCT. In a similar fashion, Cox regression was used to assess OS and FFS between type of SCT.

There were several notable differences in pt and disease characteristics between SCT modalities. Pts who received autoSCT were significantly older at initial FL diagnosis compared with alloSCT (median 51 vs 46 years, p=0.002) and older at time of SCT (median 55 vs 51 years, p=0.005). However, median time from initial FL diagnosis to SCT was shorter for alloSCT vs autoSCT (3.3 vs 4.1 years, p=0.02). AlloSCT pts received a median of 4 prior therapies vs 3 for the autoSCT cohort (p<0.001). Further, pts who had alloSCT had a higher proportion of resistant disease at time of SCT than autoSCT pts (19% vs 7%, p=0.01), while autoSCT pts were more likely to have grade 3 FL compared with alloSCT (36% vs 9%, p=0.001). There were no comparative differences among gender, race, stage at SCT or performance status. 59% of alloSCT pts had a matched-sibling donor, while 42% had an unrelated donor. The most common conditioning regimens for autoSCT were CBV 50%, BEAM 30%, and TBI-based 16%, while for alloSCT were Flu/Mel 31%, TBI-based 28% (Flu-TBI or Cy-TBI), and Bu/Flu 24%. The cumulative rate of relapse, progression, and/or transformation was 33% for autoSCT pts compared with 16% for alloSCT (p=0.01), while the overall non-relapse mortality (NRM) rate for alloSCT was 33% vs 10% for autoSCT (p<0.0001). There was no difference in FFS between type of SCT (p=0.30) with 3-year FFS of 55% (95%CI 46%-63%) for autoSCT vs 56% (95%CI 42%-70%) for alloSCT (Figure 1). However, OS was significantly different between type of SCT (p<0.001); the 3-year OS was 85% (95%CI 79%-91%) for autoSCT compared with 64% (95%CI 50%-77%) for alloSCT (Figure 2). Factors that predicted survival on univariate analysis for pts who had autoSCT were increasing age (continuous variable, years) and >3 prior therapies, while age > 50 years and resistant disease status at SCT predicted survival for alloSCT. Further, these factors remained significant on Cox regression multivariate analysis (Table 1). Additionally, multivariate analysis including all SCT pts (n=216) was performed to compare survival between autoSCT and alloSCT. After adjusting for age, number of prior therapies, and disease status, alloSCT still showed increased risk of death compared with autoSCT (HR 2.2, 95%CI 1.2–4.1, p=0.01); no significant difference was noted for FFS (p=0.72).

View Large

To our knowledge, this represents the largest analysis of autoSCT and alloSCT among FL pts relapsed/refractory s/p prior rituximab. Within this multicenter prospective cohort analysis, we identified prognostic factors that predicted survival within autoSCT and alloSCT cohorts. Furthermore, autoSCT and alloSCT were associated with comparable FFS, however including adjustment for competing prognostic factors, OS was improved among pts who had autoSCT. Based on these data, we conclude that autoSCT remains a viable therapy for relapsed/refractory FL in the post-rituximab era.

No relevant conflicts of interest to declare.