Page 217 - 18F-FDG PET as biomarker in aggressive lymphoma; technical and clinical validation

Optimal timing and response criteria of Interim-PET in DLBCL
(supplemental Table 2). After correcting the survival curves for IPI scores, studies had similar 2-year PFS, 2-year TTP, and 2-year OS survival rates (using the largest study [PETAL] as the reference; Figure 2; supplemental Table 3).
Figure 1. CONSORT diagram of patient inclusion.
I-PET response criteria
In total, 1085 patients had scans after 2 cycles of I-PET (I-PET2) scans and 482 had I-PET4 scans. There were relatively few patients with I-PET1 and I-PET3 scans (Figure 1). The prevalence of positive I-PET scans was lower when I-PET scans were performed later during treatment, independent of PET response criteria (Table 1). A total of 1675 patients were assessed according to DS, and 1533 patients also had a baseline scan available to calculate the ∆SUVmax. There were no differences in baseline characteristics between these groups (supplemental Table 4). ∆SUVmax identified a larger proportion of poor responders at I-PET scanning than did DS5 (supplemental Table 5).
In a multivariable Cox regression analysis, IPI score and I-PET scans (all PET response criteria) were independent predictors of outcome. The univariable HR
215
8