Page 140 - 18F-FDG PET as biomarker in aggressive lymphoma; technical and clinical validation

Chapter 6a
in our study), baseline MTV and ∆SUVmax were the only independent outcome predictors [8,28]. We could not confirm these findings; possible explanations are the different PET-timing (HOVON-84: I-PET4) or patient characteristics (HOVON-84: median age 3 y higher; advanced stage, 82% vs. 58% in PETAL). We chose a higher ∆SUVmax because the PET timing was different (I-PET4 vs. I-PET2) and to validate a formerly presented cutoff [10,20]. This choice does not explain the difference in added value of MTV, since the positivity percentages were the same (10.4% vs. 9.6% in PETAL), as was the 2-y PFS for the positive (46.9% and 46.7%) and negative (80.2% and 82.5%) groups according to the ∆SUVmax criteria for HOVON-84 and PETAL, respectively. Recently, Vercellino et al. showed that a combination of high baseline MTV and high performance status (≥2) identifies an ultra-risk DLBCL population [29]. We could not confirm this extra risk in our study.
There were several strengths to our study. First, to our knowledge, there are no other large, randomized trials with a homogeneous first-line treatment regimen and an observational I-PET after 4 R-CHOP14 cycles. Another strength was the central review procedure for Deauville scoring, with 2 independent reviewers and a strict DS5 definition, which allowed for an analysis to determine the optimal I-PET4 response criteria [13].
On the basis of the relatively low values for PPV, escalation of treatment for the I-PET4 positive group is not yet recommend for clinical practice, but evidence for I-PET adapted treatment is clearly growing (11,30-32). The GAINED randomized clinical trial [30] enrolled 670 DLBCL patients (aged 18-60 y, aaIPI ≥1); I-PET2-positive/I-PET4-negative patients (n=87) were scheduled to receive high-dose chemotherapy with autologous stem cell transplantation and had no statistically different PFS from the I-PET2-negative/I-PET4-negative patients (n=401) who continued standard treatment. However, no firm conclusions can be made, because there was no randomization within these I-PET-adapted groups.
Because the NPV is acceptable (>80% for all criteria), reduction of treatment based on I-PET4 could be of interest, especially for low-risk and elderly patients. The randomized FLYER trial showed that in a group of 592 DLBCL patients (aged 18-60 y, no aaIPI risk factors, no bulky disease), 4 cycles of R-CHOP21+ 2 cycles of rituximab was not inferior to 6 cycles of R-CHOP21 [6], and in an
138