An evaluation of the predictive value of mid-treatment 18F-FDG-PET/CT scans in pediatric lymphomas and undefined criteria of abnormality in quantitative analysis.

Abstract

Our purpose was to evaluate quantitative mid-treatment fluorine-18-fluorodeoxyglucose (18F-FDG) PET/CT scans in predicting the quantitative result of the end of treatment 18F-FDG PET/CT scan. With approval of Emory's Institutional Review Board, data were extracted from 273 existing 18F-FDG PET/CT scans of 143 pediatric patients performed for evaluation of lymphoma. The inclusion criteria were the availability of an initial staging scan (D0) and a mid-treatment scan after 1 to 3 cycles of chemotherapy (D1) and a post-treatment scan (D2). Absolute and relative changed of D1 compared to D0 were measured and their values in predicting D3 values were determined. Analysis was performed on a lesion basis (N=78) in 18 patients with an average of 4.3 lesions per patients. Results showed that the predictive value depended on the value selected as significant for the predictors (D1 SUV and D1 %SUV), and on the limit between negative and positive selected for the predicted value D2 SUV. If the maximum SUV<2.0 in D2 was the limit for negative, the negative predictive value if D1<4 was 0.84%. If positive was defined as D2>3.0, the positive predictive value of D1>4 was 100%. In that way outcome was predictable with absolute certainty in as many as 71% of the lesions with a single limit for D1 and D2. In conclusion, in this limited retrospective study the positive predictive value of the mid-treatment scan, was high for the post-treatment result for patient and lesion response seen on D2.