RESUMO
Uptake segmentation and classification on PSMA PET/CT are important for automating whole-body tumor burden determinations. We developed and evaluated an automated deep learning (DL)-based framework that segments and classifies uptake on PSMA PET/CT. We identified 193 [18F] DCFPyL PET/CT scans of patients with biochemically recurrent prostate cancer from two institutions, including 137 [18F] DCFPyL PET/CT scans for training and internally testing, and 56 scans from another institution for external testing. Two radiologists segmented and labelled foci as suspicious or non-suspicious for malignancy. A DL-based segmentation was developed with two independent CNNs. An anatomical prior guidance was applied to make the DL framework focus on PSMA-avid lesions. Segmentation performance was evaluated by Dice, IoU, precision, and recall. Classification model was constructed with multi-modal decision fusion framework evaluated by accuracy, AUC, F1 score, precision, and recall. Automatic segmentation of suspicious lesions was improved under prior guidance, with mean Dice, IoU, precision, and recall of 0.700, 0.566, 0.809, and 0.660 on the internal test set and 0.680, 0.548, 0.749, and 0.740 on the external test set. Our multi-modal decision fusion framework outperformed single-modal and multi-modal CNNs with accuracy, AUC, F1 score, precision, and recall of 0.764, 0.863, 0.844, 0.841, and 0.847 in distinguishing suspicious and non-suspicious foci on the internal test set and 0.796, 0.851, 0.865, 0.814, and 0.923 on the external test set. DL-based lesion segmentation on PSMA PET is facilitated through our anatomical prior guidance strategy. Our classification framework differentiates suspicious foci from those not suspicious for cancer with good accuracy.
RESUMO
PURPOSE: Soluble epoxide hydrolase (sEH) is an enzyme that shapes immune signaling through its role in maintaining the homeostasis of polyunsaturated fatty acids and their related byproducts. [18F]FNDP is a radiotracer developed for use with positron emission tomography (PET) to image sEH, which has been applied to imaging sEH in the brains of healthy individuals. Here, we report the test-retest repeatability of [18F]FNDP brain PET binding and [18F]FNDP whole-body dosimetry in healthy individuals. METHODS: Seven healthy adults (4 men, 3 women, ages 40.1 ± 4.6 years) completed [18F]FNDP brain PET on two occasions within a period of 14 days in a test-retest study design. [18F]FNDP regional total distribution volume (VT) values were derived from modeling time-activity data with a metabolite-corrected arterial input function. Test-retest variability, mean absolute deviation, and intraclass correlation coefficient (ICC) were investigated. Six other healthy adults (3 men, 3 women, ages 46.0 ± 7.0 years) underwent [18F]FNDP PET/CT for whole-body dosimetry, which was acquired over 4.5 h, starting immediately after radiotracer administration. Organ-absorbed doses and the effective dose were then estimated. RESULTS: The mean test-retest difference in regional VT (ΔVT) was 0.82 ± 5.17%. The mean absolute difference in regional VT was 4.01 ± 3.33%. The ICC across different brain regions ranged from 0.92 to 0.99. The organs with the greatest radiation-absorbed doses included the gallbladder (0.081 ± 0.024 mSv/MBq), followed by liver (0.077 ± 0.018 mSv/MBq) and kidneys (0.063 ± 0.006 mSv/MBq). The effective dose was 0.020 ± 0.003 mSv/MBq. CONCLUSION: These data support a favorable test-retest repeatability of [18F]FNDP brain PET regional VT. The radiation dose to humans from each [18F]FNDP PET scan is similar to that of other 18F-based PET radiotracers.