Background: This study is to evaluate the predictive value of FDG-PET (PET) in pediatric and adolescent patients suffering from non-Hodgkin lymphoma (pNHL) in comparison to information provided by conventional imaging methods (CIM).
Methods: Imaging was performed at baseline and at interim (after 2 cycles of chemotherapy). The response assessment in PET was carried out visually and semi-quantitatively, the latter one by use of percentage decrease in SUVmax from baseline to interim (ΔSUVmax). The PET-based results were compared to the findings by CIM. Progression-free survival (PFS) was analyzed using Kaplan-Meier curves (KM) and log-rank test.
Results: The final study included 16 patients (mean follow-up time, 60.2 months (range, 4.0 to 85.7 months)). Relapse occurred in four patients. Visual PET compared to CIM revealed higher sensitivity (3/4 vs 1/4) and NPV (6/7 vs 10/13), and equal PPV (3/9 vs 1/3), but lower specificity (6/12 vs 10/12) and accuracy (9/16 vs 11/16). False-positive findings in PET at interim were predominantly observed in patients presenting bulky disease (5/6), whereas CIM was true-negative in all of these cases. KM analyses revealed no significant differences in 5-year PFS neither for CIM (76.9% vs 66.7%; p = 0.67) nor for visual PET (85.7% vs 66.7%; p = 0.34) nor for ΔSUVmax (88.9% vs 57.1%; p = 0.12).
Conclusions: The predictive value of iPET in pediatric patients suffering from NHL was limited due to considerably high amount of false-positive findings, especially in patients suffering from bulky disease. However, due to our limited sample size, final conclusions cannot be drawn and, thus, call for further evaluation of PET in pNHL in larger and more homogenous patient series.