Tc-99m-MIP-1404 SPECT/CT for Patients With Metastatic Prostate Cancer Interobserver and Intraobserver Variability in Treatment-Related Longitudinal Tracer Uptake Assessments of Prostate-Specific Membrane Antigen-Positive Lesions

摘要

Background Tc-99m-MIP-1404 is a SPECT-suitable prostate-specific membrane antigen (PSMA) ligand for detection of prostate cancer. In patients with metastatic prostate cancer, there are no data as yet on interobserver and intraobserver variability when assessing PSMA-positive lesions for longitudinal changes of tracer uptake. Methods Tc-99m-MIP-1404 SPECT/CT scans of 22 patients with metastatic prostate cancer were analyzed, and each subject was imaged at 2 separate points in time, before and after treatment. Mean interval between scans was 10 months. Three independent observers visually assessed a total of 96 PSMA-positive metastases (bone, 69; lymph node, 22; viscera, 3) or local recurrences (n = 2) for longitudinal changes in tracer uptake on planar scintigraphy and SPECT/CT. All lesions were categorized as regressive, stable, or progressive based on visual findings and on peak SUV (SUVpeak) of quantitative SPECT/CT (progressive, >30% SUVpeak increase; regressive, <30% SUVpeak decrease; or stable, all others). Results Quantitative analysis of PSMA-positive lesions yielded significantly higher interobserver agreement (90.6%; 95% confidence interval [CI], 0.83%-0.96%) than visual assessments by either SPECT/CT (76.0%; 95% CI, 0.66%-0.84%) or planar scintigraphy (56.3%; 95% CI, 0.46%-0.66%). Intermethod comparison of aggregated results yielded significantly higher agreement between quantitative and visual SPECT/CT (85.1%; 95% CI, 0.80%-0.89%), as opposed to quantitative SPECT/CT and planar scintigraphy (53.1%; 95% CI, 0.47%-0.59%) or visual SPECT/CT and planar scintigraphy (54.9%; 95% CI, 0.49%-0.61%). In visual and quantitative analysis of 96 PSMA-positive lesions, the number of discrepancies ranged from 9 (9.4%) for quantitative SPECT/CT to 42 (43.8%) for planar scintigraphy. Overall reader confidence was higher for SPECT/CT than for planar scintigraphy (P < 0.001). Intraobserver agreement was near-perfect for all methods, whether SPECT/CT (visual, all kappa = 0.94-0.97; quantitative kappa = 0.94-0.98) or planar scintigraphy (all kappa = 0.90-0.94). Conclusions Quantitative evaluation of longitudinal change in tracer uptake by PSMA-positive lesions measured via SPECT/CT is superior to visual interpretation of images by planar scintigraphy or SPECT/CT. Compared with visual evaluation, quantitative SPECT/CT is highly reproducible, showing near-perfect agreement among observers and higher reader confidence.