This study aimed to clarify the relationship between tumor redox reaction evaluated by 62Cu-diacetyl-bis (N4-methylthiosemicarbazone) (62Cu-ATSM) PET/computed tomography (CT) and disease-free survival (DFS) in patients with primary diffuse large B-cell lymphoma of the central nervous system (DLBCL-CNS).
Fifteen consecutive patients with histologically confirmed DLBCL-CNS underwent preoperative 62Cu-ATSM PET/CT and 18F-fluorodeoxyglucose (FDG) PET/CT. Statistical features of seven first-order parameters, including the standardized uptake value (SUV); 12 second-order parameters, including gray-level co-occurrence matrices and gray-level zone size matrices; and 5 high-order parameters, including neighborhood gray-tone difference matrices, were calculated from the volume of interest. We compared DFS with parameters, including SUVmax and tumor-to-background (T/B) ratio of FDG, and SUVmax, T/B ratio, and other textural features of 62Cu-ATSM.
The mean follow-up duration after PET/CT was 458 (range, 41–1071) days. The SUVmax of FDG was significantly higher than that of 62Cu-ATSM (P = 0.001), but the T/B ratio was not significantly different between the scans (3.49 ± 2.29 vs 2.48 ± 1.18; P = 0.244). A Mantel–Cox log-rank test revealed no significant association between SUVmax of FDG and DFS (P = 0.641). A high SUVmax of 62Cu-ATSM had a tendency of shorter DFS (P = 0.055). Total lesion reduction, reductive tumor volume, and T/B ratio of 62Cu-ATSM were significantly correlated with poor DFS by univariate analysis (P = 0.049, 0.031, and 0.007, respectively). Neighborhood gray-level co-occurrence matrix dissimilarity was significantly correlated with poor DFS (P = 0.015).
Metabolic and textural features derived from pretreatment 62Cu-ATSM PET/CT could be used for predicting DFS and establishing a novel treatment strategy in DLBCL-CNS patients.