Predictive Value of Synthetic PET Imaging for Assessing Biology-Guided Radiotherapy Feasibility: Results from a Prospective Imaging Study
Abstract
Purpose
To evaluate the ability of synthetic PET images derived from standard-of-care (SOC) 18F-FDG PET/CT images to predict the feasibility of biology-guided radiotherapy (BgRT).
Methods
In this prospective imaging study, patients undergoing SOC 18F-FDG PET/CT subsequently received a PET scan on a BgRT platform (X1). Disease sites include the head & neck region, esophagus, mediastinum, abdomen, and pelvis. Gross target volumes (GTVs) were delineated on SOC PET/CT images for PET-avid lesion. Planning target volumes (PTVs) were defined as 5-mm isotropic expansions of the GTV, and biology tracking zones (BTZs) were defined as 5-mm isotropic expansions of the PTV. Synthetic PET images were generated from SOC PET images to simulate predicted PET Images acquired on the BgRT machine. BgRT treatment planning was attempted on both synthetic PET and X1 PET images. Activity concentration (AC) and normalized target strength (NTS) were assessed to determine BgRT feasibility.
Results
X1 PET scan were performed with an average of 2.0±0.6 hours (range: 1.3 - 3.2 hours) after SOC PET/CT imaging. Of the 17 patients who completed both SOC PET/CT and X1 PET scans, BgRT planning based on synthetic PET images were successful for 13 patients, whereas BgRT planning based on X1 PET images was successful in 9 patients. Among the 8 cases in which X1 PET-based planning failed, 4 also failed synthetic PET-based BgRT planning. Three cases passed synthetic PET-based planning but were correctly predicted to fail X1 PET-based BgRT planning due to reduced AC from decay, and one case failed X1 PET-based planning due to lower-than-predicted AC relative to SOC PET.
Conclusion
These results demonstrate the clinical utility of synthetic PET imaging for BgRT planning. Feasibility outcomes derived from synthetic PET images were predictive of BgRT planning success on actual X1 PET images in the majority of cases.