Abstract 154


Follicular lymphoma (FL) has an indolent course during which patients (pts) receive multiple lines of active treatments ranging from single-agent chemotherapy (CHT) or monoclonal antibodies (MoAb), to high-dose therapy with stem cell transplantation (HSCT). In relapsed patients the efficacy of salvage treatments may be affected by the type and the intensity of the previous treatments and it is currently not known whether a definite sequence of treatments throughout the course of the disease could optimize the outcome of patients. Since randomized trials can hardly be designed to answer this important clinical question, the Fondazione Italiana Linfomi (FIL) launched an observational, multicenter, retrospective study (REFOLL) to analyze if combinations of different first-line and salvage treatments could be identified, able to achieve a better long-term outcome.

Patients and Methods

Of 582 pts with FL at first relapse between 2000 and 2008 registered from 25 Institutions, 548 were included in the study. They had received either alkylating- (AA) (22%), or anthracycline- (AC) (61%) or nucleoside analogues-based (NA)(17%) CHT as first-line treatment, with the addition of rituximab (R) in 284 (52%). AA pts were older (P<.001), but they had less stage III-IV (P=.011) and FLIPI high-risk disease (P=.02), and had received maintenance R after first-line CHT more often (P<.001) compared to pts receiving AC- or NA-based CHT. At relapse 21 pts received radiotherapy and 4 palliation and were excluded. Salvage treatment of the other 523 pts were either the same first-line CHT programs (AA: 20%; AC 18%; NA: 14% of the 523 pts, respectively) in 271 pts, associated with R in 73%, or a program including autologous HSCT in 151 (29%), or MoAb/radioimmunoconjugate without CHT in 101 (19%). The primary endpoint of the study was time to next treatment (TTNT) after relapse. Progression-free survival after relapse (PFS) and overall survival (OS) were secondary endpoints.


The median follow-up after first relapse was 41 months (range 1–122) and the median TTNT (50%) was 41 months (CI95% 35–47 months).

Among first-line treatments, AC+/−R was associated with a better TTNT after any salvage than AA+/−R or NA+/−R (HR: 0.71, P=0.007; HR: 0.73, P=0.021 after adjustment by age, stage and year of diagnosis). The addition of R to first-line CHT did not adversely impact on the efficacy of salvage treatments (P=0.156).

As salvage treatment, auto-HSCT had a significantly better outcome when compared to any other program +/−R (HR ranging from 1.79 to 2.40). Patients receiving auto-HSCT were younger (P<0.001), had a shorter duration of first remission (P=0.012), but the efficacy of HSCT was not affected by these two parameters. ASCT program included R in 63% and was followed by R maintenance in 7% of pts.

Considering the combinations of first-line and salvage treatments received, using multiple Cox regression, the TTNT after relapse was significantly better in pts receiving first-line anthracycline CHT +/−R followed by auto-HSCT as salvage (HR ranging from 1.88 to 3.3 compared to any other sequence, including AA/NA CHT followed by auto-HSCT). Other factors independently influencing TTNT after HSCT were R maintenance (HR: 0.66), duration of first remission > 12 months (HR: 0.8), and stage III-IV at diagnosis (HR: 1.89).

PFS after relapse was 35% at 5 years, superimposable to TTNT (36% at 5 years); in fact treatment sequences had the same effects on PFS as on TTNT. Overall survival from diagnosis was 89% at 5-yr (CI95% 85–91%) and the sequences CHT->HSCT and CHT->CHT with R showed a better outcome than CHT ->CHT without R (HR:0.55, P=0.015; HR 0.61, P=0.047 after adjusting by age and stage at diagnosis).


Auto-HSCT obtained the best TTNT after relapse compared to any other regimen but its efficacy was maximized when anthracycline-containing CHT was used as first-line treatment, compared to other CHT programs. The addition of R to first-line CHT did not adversely impact on the results of any salvage. This study supports the concept that different sequences of active treatments do not necessarily obtain similar long-term results, which is important in the management of indolent diseases like FL and warrants further studies. The sequence of AC-CHT at diagnosis and auto-HSCT at relapse may be tested as the reference sequence of active treatments in FL patients.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.