Page 168 - 2020_07-Haematologica-web

P. Strati et al.
(P<0.001)
and prolonged PFS was maintained also after adjusting for pre-treatment SUVmax >18 (HR 2, 95% CI: 1.3-2.9, P<0.001). Further subgroup analysis, to assess the effect of maintenance rituximab after specific regimens according to pre-treatment SUVmax >18 could not be performed, because of small population samples.
Excluding 11 patients who did not progress and were lost to follow-up within 24 months, 67 (20%) patients had a PFS of less than 24 months: PFS <24 months was observed in 51 (18%) patients with baseline SUVmax <18, and 16 (34%) of patients with baseline SUVmax >18 (P=0.02). SUVmax >18 did not associate with a significantly higher rate of PFS <24 months among patients treated with R-CHOP (32% vs. 24%, P=0.39), but it did associate with a higher rate of PFS <24 months among patients treated with other regimens (40% vs. 14%, P=0.05). After excluding patients treated with single agent rituximab, a trend for higher PFS24 rate among patients treated with other frontline regimens and SUVmax >18 was observed (20% vs. 11%, P=0.45); of interest, only five patients with SUVmax >18 were evaluable for PFS24, likely limiting achievement of statistical significance.
Overall survival (OS) and risk of transformation
After a median follow-up of 94 months (95% CI: 88-100 months), median OS has not been reached, and it was sig- nificantly shorter among patients with SUVmax >18 (8-year OS 65% vs. 89%, P=0.001). SUVmax >18 associated with shorter OS both in patients treated with frontline R- CHOP (8-year OS 70% vs. 90%, P=0.02) and in patients treated with other frontline regimens (8-year OS 50% vs. 85%, P=0.001)(Figure 3). A trend for longer OS was observed among patients with SUVmax >18 when compar- ing treatment with frontline R-CHOP to other frontline treatments (8-year OS 70% vs. 50%, P=0.15). The associ- ation between SUVmax >18 and OS was maintained also on multivariate analysis including either FLIPI score (HR 2.6, 95% CI: 1.5-4.6, P=0.001) or FLIPI-2 score (HR 2.2, 95% CI: 1.3-3.9, P=0.006).
At the most recent follow-up, 18 (5%) patients have progressed with histological evidence of transformation to large B-cell lymphoma, after a median time of 23 months (range: 5-139 months). Twelve transformations (4%) occurred among patients with baseline SUVmax <18, and
Figure 1. Complete response (CR) rates by maximum standardized uptake (SUVmax) values according to frontline regimen.
six (11%) among patients with baseline SUVmax >18 (P=0.04). Of these 18 patients, 11 had received frontline R- CHOP (of whom six had baseline SUVmax >18) and seven other frontline regimens (none of whom had baseline SUVmax >18). After excluding patients treated with single agent rituximab, a statistically significant shorter OS among patients treated with other frontline regimens and SUVmax >18 was observed (86 months vs. not reached, P=0.05).
Discussion
Recent guidelines recommend the use of PET-CT in FL for initial staging, evaluation of potential transformation, and at time of response assessment.31,33 However, the impact of baseline PET-based imaging on the outcome fol- lowing specific frontline treatment has not been previous- ly explored.
In our analysis, an SUVmax cut-off of 18 was identified as clinically significant, showing the strongest association with PFS. SUVmax cut-offs of 10, 14 and 17 have been pro- posed in previous retrospective studies to identify patients with FL at higher risk of transformation to DLBCL,11,12,14 with OR for transformation of 1.25 for each increase in unit of SUVmax .10 To this regard, patients with histological evidence of FL grade IIIB or DLBCL were excluded from this study. Although tumor heterogeneity may have caused a false-negative result, elevated SUVmax in the absence of histological evidence of transformation may reflect a more aggressive biology, associating with a worse outcome, as already shown in other forms of low grade B-cell lymphomas.34 It is interesting to note that in our study the only factor significantly associated with SUVmax >18 was the lymph node size ≥6 cm and that more trans- formations occurred in patients with pre-treatment SUVmax >18 (11% vs. 4%, P=0.04). While a core needle biopsy of a large lymph node may miss a diagnosis of DLBCL, the large tumor size may also simply be a surrogate marker of accelerated biology in the absence of transformation. Prospective studies, employing excisional biopsy, as opposed to core biopsy, may shed light on this question.
In our study, patients with SUVmax >18 were more fre- quently treated with R-CHOP, compared to other thera-
1910
haematologica | 2020; 105(7)