Author: Resat Aydin, Brett Lewis, Jayesh Mistry, Roland Teboh 👨🔬
Affiliation: HUMC, Hackensack University Medical Center 🌍
Purpose:
We developed a Python-based solution to systematically quantify and detect fluorodeoxyglucose (FDG) uptake beyond a predefined Biologically Tracking Zone (BTZ) in PET images. This method is designed to support biology-guided radiotherapy (BgRT) feasibility studies by identifying whether any voxels beyond the BTZ surpasses a 50% of the BTZ’s maximum standardized uptake value (SUVmax).
Methods:
The solution processes PET DICOM data and utilizes an RT Structure defined as BTZ contour for reference. SUVmax within the BTZ is computed, and a 3D distance transform identifies voxels outside the BTZ within a user-specified radial threshold (e.g., 2cm).
The solution first identifies the maximum SUV in BTZ. Based on the user-defined threshold, it creates a volume to search for any voxel that has an SUV ≥ 50% of the BTZ SUVmax. Once that voxel is found (if any), the solution calculates its distance to the BTZ boundary. In a case no voxel is found within the user-defined region that exceeds 50% of the BTZ’s SUVmax, the solution reports no such voxel is found, however, it still reports the voxel which has the maximum SUV in the user-defined region and that voxel’s distance to the BTZ’s boundary.
Results:
The Python-based solution was clinically tested on six patient datasets, demonstrating its ability to identify the outside voxel with a SUV ≥ 50% of BTZ SUVmax. To verify the solution, its output was compared with clinical workflows employing Eclipse TPS, where a ring (e.g., 2cm) around the BTZ was used to locate regions with elevated FDG uptake. The Python-based solution was consistent with the ring-expansion logic using Eclipse TPS.
Conclusion:
This Python-based solution offers a reliable, flexible approach for evaluating FDG uptake beyond a predefined BTZ contour. The finding could be used as one of the parameters which identifies whether the patient is eligible for BgRT.