Research Information System
Annals of Nuclear Medicine, 2026 (SCI-Expanded, Scopus)
Objective: To investigate the correlation between standardized uptake value (SUV) parameters derived from Ga-68 PSMA PET/CT and post-therapeutic Lu-177 PSMA SPECT/CT, and to evaluate whether SPECT derived quantitative SUV metrics are associated with lesion-based interim treatment response in patients with metastatic castration resistant prostate cancer (mCRPC). Methods: Fifty-five mCRPC patients who received 2–6 cycles of Lu-177 PSMA therapy and underwent both baseline Ga-68 PSMA PET/CT and 24-h post-therapy Lu-177 PSMA SPECT/CT were retrospectively analyzed. SUVmax, SUVmean, and SUVmin values were measured on PET (PSUV) and SPECT (QSUV) in up to 10 index metastatic lesions per scan and in predefined physiological regions. A patient-level aggregate metric (mean SUV across the selected metastatic lesions) was also calculated. Lesion-based response was defined by changes in PSUVmax on follow-up PET/CT and categorized as response, stable, or progressive. Correlation analyses, ROC curves, and logistic regression models were applied to examine relationships between QSUV parameters, PET derived metrics, and lesion response. Results: A total of 915 volumes of interest (VOIs) were analyzed, with 664 being pathological and 251 physiological. QSUV parameters showed strong positive correlations with PSUV parameters (rho: 0.71, p < 0.01), particularly in lesions > 3 mL and prostate primary lesions. Responding lesions had significantly higher QSUV values compared to progressing lesions (p < 0.05). ROC analysis yielded AUC values of 0.696 for QSUVmax, 0.699 for QSUVmean, and 0.701 for QSUVmin. QSUVmax, QSUVmean, and QSUVmin were independently associated with lesion-based response in logistic regression models (p < 0.01). In the subgroup analysis excluding stable lesions, PSUV outperformed QSUV in predicting response (p < 0.01 via DeLong test). However, when stable lesions were included (Responders vs. Non-responders), ROC analysis yielded comparable AUC values for PSUV (0.710–0.716) and QSUV (0.615–0.620), with no statistically significant differences observed (p > 0.05). Furthermore, a multimodality approach combining PSUV and QSUV metrics via binary logistic regression did not provide a synergistic improvement in predictive accuracy over individual parameters (p > 0.05). Conclusion: Quantitative SUV parameters from Lu-177 PSMA SPECT/CT correlate significantly with Ga-68 PSMA PET/CT and are associated with therapeutic response. While PET remains the superior modality for high-precision quantification, the comparable performance of QSUV metrics in a clinically representative setting suggests that SPECT-derived parameters can serve as a reliable and accessible tool for interim response assessment. These metrics may serve as accessible tools for interim response assessment and lesion-level prognostication in PSMA-targeted radioligand therapy.