ΔSUVmax between Baseline and Interim 18f-FDG PET/CT Scans Could Better Assess the Early Response to Induction Therapy and Identify False-Positive Findings at Interim PET/CT in Extranodal NK/T-Cell Lymphoma

Background:

Extranodal NK/T-cell lymphoma (ENKTCL) is generally sensitive to pegaspargase-based induction chemotherapy, and inadequate response to induction therapy will predict dismal prognosis. ¹⁸F-FDG PET/CT scan has been consistently shown valuable in staging, response assessment, and prediction of long-term outcomes in ENKTCL. Patients with five-point Deauville score (5-DS) ≥4 at interim PET-CT scan are usually defined as positive for residual lymphoma lesions and demand subsequent switch of treatment regimens or strategies. However, through re-biopsy of suspected lesions revealed by interim PET-CT scan, we found most were false-positive, and those patients had similar prognosis compared with patients who had 5-DS ≤3 at interim PET-CT scan. Thus, a more accurate assessing method using PET-CT scan is urgently needed. A quantitative approach based on reduction in the standardized uptake values (ΔSUV_max) allows more objective evaluation of the PET response, which may reduce false-positive interim PET/CT interpretations.

Methods:

20 patients with newly diagnosed ENKTCL who underwent baseline and interim ¹⁸F-FDG PET/CT scans (after 3-4 cycles of pegaspargase-based induction chemotherapy) and with 5-DS≥4 at interim PET/CT were included in this retrospective analysis. All patients received re-biopsy of suspected lesions revealed by interim PET-CT scan. Receiver operating characteristic (ROC) analysis was used to determine an optimal cutoff for ΔSUV_max in assessing early response to induction therapy. Kaplan-Meier curves were used to analyze long-term survival outcomes.

Results:

16 (80%) patients had stage II disease, while four (20%) patients were diagnosed with stage IV disease. Nine (45%) patients presented with a NRI score≥3, and 70% of patients were classified as high risk according to the PINK score. After 3-4 cycles of induction chemotherapy, three (15%) patients had a 5-DS score of five at interim PET/CT, while the remaining 85% of patients had 5-DS score of four. Among these 20 patients, two (10%) were confirmed to have residual lymphoma via re-biopsy, while 18 (90%) patients obtained negative results and were defined as complete remission (CR). All patients with a 5-DS score of four exhibited negative results upon re-biopsy, while two out of three cases assigned a 5-DS score of five had positive pathology findings, indicating the potential false positive outcomes within the 5-DS scoring system, particularly among the patients with scores of four. Using ROC analysis, the optimal cut-off value for discriminating CR and non-CR was 66.75% for ΔSUV_max%, which represented good discriminative power with the area under curve (AUC) of 0.985. At a median follow-up time of 27 months, patients with re-biopsy-proven active disease presented a trend towards poorer outcomes compared to those achieving negative biopsy results in terms of PFS (P＜0.001). Utilizing the ΔSUV_max% (> 66.75% vs ≤ 66.75%) between baseline and interim PET/CT scans, it was feasible to identify patients with significantly different PFS (P =0.003). The median PFS was unreached for patients with ΔSUV_max% > 66.75%, whereas it was 10 months for patients with ΔSUV_max% ≤ 66.75%.

Conclusions:

Utilizing ΔSUV_max% (＞66.75%) between baseline and interim PET/CT scans can effectively identify a subset of patients who were visually analyzed as false positives (5-DS ≥ four), which were confirmed by interim biopsy results, thus serving as a more accurate indicator for early assessment of treatment outcomes in patients with ENKTCL.