Quantitative Assessments of Tumor Activity in a General Oncologic PET/CT Population: Which Metric Minimizes Tracer Uptake Time Dependence?

Abstract

In oncologic PET, the SUV and standardized uptake ratio (SUR) of a viable tumor generally increase during the postinjection period. In contrast, the net influx rate (K_i ), which is derived from dynamic PET data, should remain relatively constant. Uptake-time-corrected SUV (cSUV) and SUR (cSUR) have been proposed as uptake-time-independent, static alternatives to K_i Our primary aim was to quantify the intrascan repeatability of K_i , SUV, cSUV, SUR, and cSUR among malignant lesions on PET/CT. An exploratory aim was to assess the ability of cSUR to estimate K_i Methods: This prospective, single-center study enrolled adults undergoing standard-of-care oncologic PET/CT. SUV and K_i images were reconstructed from dynamic PET data obtained before (∼35-50 min after injection) and after (∼75-90 min after injection) standard-of-care imaging. Tumors were manually segmented. Quantitative metrics were extracted. cSUVs and cSURs were calculated for a 60-min postinjection reference uptake time. The magnitude of the intrascan test-retest percent change (test-retest |%Δ|) was calculated. Coefficients of determination (R ²) and intraclass correlation coefficients (ICC) were also computed. Differences between metrics were assessed via the Wilcoxon signed-rank test (α, 0.05). Results: This study enrolled 78 subjects; 41 subjects (mean age, 63.8 y; 24 men) with 116 lesions were analyzed. For both tracers, SUV_max and maximum SUR (SUR_max) had large early-to-late increases (i.e., poor intrascan repeatability). Among [¹⁸F]FDG-avid lesions (n = 93), there were no differences in intrascan repeatability (median test-retest |%Δ|; ICC) between the maximum K_i (K_{i ,max}) (13%; 0.97) and either the maximum cSUV (cSUV_max) (12%, P = 0.90; 0.96) or the maximum cSUR (cSUR_max) (13%, P = 0.67; 0.94). For DOTATATE-avid lesions (n = 23), there were no differences in intrascan repeatability between the K_{i ,max} (11%; 0.98) and either the cSUV_max (13%, P = 0.41; 0.98) or the cSUR_max (11%, P = 0.08; 0.94). The SUV_max, cSUV_max, SUR_max, and cSUR_max were all strongly correlated with the K_{i ,max} for both [¹⁸F]FDG (R ², 0.81-0.92) and DOTATATE (R ², 0.88-0.96), but the cSUR_max provided the best agreement with the K_{i ,max} across early-to-late time points for [¹⁸F]FDG (ICC, 0.69-0.75) and DOTATATE (ICC, 0.90-0.91). Conclusion: K_{i ,max}, cSUV_max, and cSUR_max had low uptake time dependence compared with SUV_max and SUR_max The K_{i ,max} can be predicted from cSUR_max.