Cancer Medicine (Sep 2019)

Interim PET/CT based on visual and semiquantitative analysis predicts survival in patients with diffuse large B‐cell lymphoma

  • Xiaoqian Li,
  • Xun Sun,
  • Juan Li,
  • Zijian Liu,
  • Mi Mi,
  • Fang Zhu,
  • Gang Wu,
  • Xiaoli Lan,
  • Liling Zhang

DOI
https://doi.org/10.1002/cam4.2404
Journal volume & issue
Vol. 8, no. 11
pp. 5012 – 5022

Abstract

Read online

Abstract Purpose The role of interim 18F‐FDG PET/CT (iPET/CT) in diffuse large B‐cell lymphoma (DLBCL) remains controversial. The purpose of this study was to assess the prognostic value of iPET/CT in patients with newly diagnosed DLBCL according to visual and semiquantitative interpretation methods. Methods A total of 129 newly diagnosed DLBCL patients with baseline PET/CT data were retrospectively screened. The iPET/CT findings were evaluated by the Deauville 5‐point scale (DS) and ΔSUVmax. Furthermore, the reduction in SUVmax incorporated with tumor size (ΔSUVmax*ΔSLD) was calculated. The optimal cutoff values of ΔSUVmax and ΔSUVmax*ΔSLD were determined by receiver operating characteristic (ROC) analysis. Kaplan‐Meier analysis was applied to test for the influence of prognostic values. Univariate and multivariate analyses were conducted to examine the potential independent impacts of iPET/CT. Results Seventy‐seven patients with PET/CT images acquired both at baseline and after four cycles of chemotherapy were finally enrolled. The optimal cutoff values for ΔSUVmax and ΔSUVmax*ΔSLD were 74% and 30%, respectively. After a median follow‐up of 23 months, iPET/CT findings were significant predictors of PFS and OS whenever iPET/CT was interpreted by DS, ΔSUVmax, or ΔSUVmax*ΔSLD methods. ΔSUVmax‐based methods were more accurate than those based on DS. The IPI, DS, ΔSUVmax, and ΔSUVmax*ΔSLD were predictive in univariate analyses. However, in the multivariate analysis, only IPI and ΔSUVmax remained independent predictors of PFS and OS. Conclusions Interim PET/CT may help to identify DLBCL patients with different prognoses. ΔSUVmax analysis shows the best accuracy and the strongest predictive value among these three methods. ΔSUVmax*ΔSLD may be a promising parameter to interpret iPET/CT images, reflecting both the changes in tumor size and metabolic activity.

Keywords