Purpose: Low extracellular volume fraction (ECVf) correlates with poor vasculature and possible necrosis in pancreatic adenocarcinoma. In this work we investigate the potential of using contrast enhanced dual-source dual-energy CT (DECT) to map out ECVf spatial distribution for pancreatic tumor delineation.
Methods: Contrast enhanced DECT acquired using a dual-source DECT scanner (SOMATOM Drive, Siemens) at 80 and 140 kVp during standard radiotherapy (RT) simulation for one pancreatic cancer patient was analyzed. An ECVf map was generated from the low and high energy CT series acquired simultaneously during the late arterial contrast phase in the following steps: (1) calibrating DECT iodine contrast using a phantom with known iodine densities; (2) constructing the patient-specific iodine map based on the iodine calibration; and (3) computing the ECVf map using the patient-specific iodine map and the patient’s hematocrit level measured on the day of simulation. A region of interest (ROI) of low ECVf was delineated using an ECVf threshold of 0 to 7% from the obtained ECVf map. The volume of the ECVf ROI was compared with the clinical gross tumor volume (GTV) using the Dice similarity coefficient (DSC), and the overlap of the corresponding contour pair of the GTV and ECVf ROI on each image slice was measured using distance to agreement (MDA) and Hausdorff distance (HD).
Results: The low ECVf ROI reasonably correlated to the clinical GTV. The DSC, MDA and HD between the ECVf ROI and the clinical GTV were 0.6±0.1, 2.0±1 mm, and 5.0±0.8 mm, respectively.
Conclusion: This work demonstrated the potential of using the extracellular volume fraction map, derived from contrast enhanced dual-source DECT, to help define the pancreatic GTV for RT planning. Further studies with more patient data are required to determine the sensitivity, specificity, and limitations of this ECVf technique.
Funding Support, Disclosures, and Conflict of Interest: Research partially supported by Siemens
Dual-energy Imaging, Contrast, Treatment Planning