Page 221 - 18F-FDG PET as biomarker in aggressive lymphoma; technical and clinical validation

Optimal timing and response criteria of Interim-PET in DLBCL
Good responders at I-PET4 had a significantly higher survival compared with good responders at I-PET2 for all PET positivity criteria (DS4-5 negative: HR, 0.70; DS5 negative: HR, 0.74; and ∆SUVmax negative: HR, 0.72; supplemental Table 6). There were no significant differences in PFS between good responders at other time points compared with I-PET2. There were no significant differences between poor responders at all time points compared with I-PET2 (supplemental Table 6). The tables for TTP and OS as outcome parameters are similar (supplemental Tables 7-12).
Discussion
I-PET was predictive in all IPI risk groups in this meta-analysis of individual patient data. PET criteria that applied ∆SUVmax and DS5 positivity discriminate good responders from poor responders better than DS4-5 positivity criteria. But the DS5 criterion identified only a very small number of patients. Performing I-PET scans at later time points during therapy improved patient stratification. Limited data for I-PET1 and I-PET3 timings precluded firm conclusions from being drawn about these time points.
We found a univariable HR of 2.20 (95% CI, 1.79-2.69) for I-PET using DS4- 5 positivity criteria, confirming the predictive value of I-PET scans in patients with DLBCL. In this study, the univariable HR was lower than the pooled univariable HR of 3.13 (95% CI, 2.52-3.89) reported in a recent meta-analysis [3]. This difference in HRs can be explained partly because different definitions of outcome parameters and different response criteria were used in various studies. Moreover, we included recent, larger trials (PETAL, HOVON84) that reported lower HRs than the ones included in this meta-analysis. The higher HR using ∆SUVmax positivity criteria is in line with other recent studies, which reported that ∆SUVmax positivity criteria better discriminated poor responders and good responders at I-PET scans compared with currently used DS4-5–positivity criteria [11,20,21].
Our results showed that I-PET scans of patients with DLBCL have an NPV >80% for 2-year PFS, which is in line with previously published results [3,22,23]. In the literature, PPVs at I-PET2 ranged between 37% and 74% using DS4-
219
8