Can we use pre-therapy PSMA PET/CT as a predictor for Lutetium-177-PSMA-617 therapy? Early experience from the Canadian Cancer Trials Group PR21 trial (NCT 04663997)

Introduction: The recent FDA-approval of ¹⁷⁷Lu-PSMA-617 will likely be a game changer for the management of advanced, metastasized prostate cancer, and widely increasing therapy administrations can be expected. Pre-therapeutic predictive parameters can help to personalize radiopharmaceutical therapies in the future and optimize patient selection. The aim of this work was to determine if standardized uptake values (SUVs) from ⁶⁸Ga-PSMA PET/CT are correlated with SUVs from post-therapeutic ¹⁷⁷Lu-PSMA-617 SPECT/CT over the treatment course of up to 6 cycles.

Methods: A total of 73 therapy cycles of 15 mCRPC patients treated with ¹⁷⁷Lu-PSMA-617 within the PR21 trial (NCT 04663997) at the BC Cancer Center were analyzed. Volumes of interest (VOIs) were generated for kidneys, liver, spleen, salivary glands (parotid glands and submandibular glands), skeleton (as an approximation for bone marrow), and total tumor burden (TTB) on the ⁶⁸Ga-PSMA PET/CT and separately on the 48h post injection (p.i.) ¹⁷⁷Lu-PSMA-617 SPECT/CT. Organs were delineated automatically on the CTs using an artificial intelligence workflow available with MIM software, while lesions were delineated using the qPSMA approach. Manual adjustments were made by an experienced physician when deemed necessary. Descriptive statistics were extracted for each VOI, image, and patient. A correlation analysis was performed for the mean SUVs per VOI and the coefficients of determination R² were compared for different organs.

Results: Average SUVs across patients from ⁶⁸Ga-PSMA PET were 18.77 ± 6.45 (kidneys), 4.29 ± 1.04 (liver), 5.72 ± 3.64 (spleen), 11.47 ± 3.96 (salivary glands), 1.30 ± 1.43 (skeleton), and 13.06 ± 4.80 (TTB). Average SUVs for the ¹⁷⁷Lu-PSMA-617 SPECT across patients were 3.06 ± 1.14 (kidneys), 0.57 ± 0.27 (liver), 0.32 ± 0.26 (spleen), 4.21 ± 1.65 (salivary glands), 0.50 ± 0.88 (skeleton), and 8.20 ± 2.44 (TTB). The lesion SUV (TTB) decreased on average by 28% from cycle 1 to cycle 6, while the kidney SUV increased by 35%. Interestingly, the salivary gland SUV decreased by 40%. Strong correlation between PET and first cycle SPECT with an R² of 0.94 was found for the skeleton. Intermediate correlations with R² of 0.48, 0.45, and 0.44 were found for liver, spleen, and TTB, respectively. Low correlations were found for salivary glands (R² between 0.19 and 0.33). No correlation was found for the kidneys. Overall, the correlations decreased towards cycle 6.

Conclusions: Strong and intermediate correlations between pre-therapy PET and post-therapy SPECT were found for low accumulating organs such as the liver, spleen and skeleton. No correlation was found for kidneys, and intermediate to low correlation was found for salivary glands and lesions, which can likely be attributed to the difference in imaging time points between PSMA PET and PSMA SPECT. PSMA PET scans are usually performed 1h p.i., while PSMA SPECT scans are performed 48h p.i., thus especially influencing the SUVs of lesions, which show slower uptake and washout than the healthy organs. Future studies using ⁶⁴Cu or ⁸⁹Zr labelled PSMA compounds (and/or long axial field-of-view PET scanners) could allow for later PET imaging time points and potentially better correlation with post-therapeutic ¹⁷⁷Lu-PSMA-617 SPECTs. While the SUV for the TTB generally decreased over cycles, the SUV for kidneys increased which can be attributed to the tumor sink effect. In a next step, we plan to investigate if correlations between the pre-therapeutic PET SUVs with the absorbed doses across therapy cycles exist. The inclusion of more patients in the analysis will aid in determining further possible correlations.

• Download figure
• Open in new tab
• Download powerpoint