Impact of Multi-Energy CT On Tumor Delineation of Primary Head and Neck Tumors
J Miller1*, L DiMaso-Myers1, P Harari1, B Morris1, E Merfeld1, J Shah2, A Burr1, M Lawless1, (1) University of Wisconsin-Madison, Madison, WI,(6) Siemens Healthineers, Durham, NC
Presentations
TU-D1000-IePD-F5-1 (Tuesday, 7/12/2022) 10:00 AM - 10:30 AM [Eastern Time (GMT-4)]
Exhibit Hall | Forum 5
Purpose: To investigate the ability of split-filter multi-energy CT (MECT) to enhance head and neck (HN) tumor segmentation for radiation therapy treatment planning.
Methods: A contouring study was performed for a cohort of sixteen patients with HN cancer who received split-filter MECT, PET, and T1-post-contrast MR imaging during radiation therapy simulation. A Composed 120 kVp-equivalent image (C120) and a monoenergetic-50 keV image (M50) were created for each patient. Four physicians were provided with contours derived from PET/CT or PET/MR for each patient and asked to adjust the contour based on the C120, M50 and MR image, sequentially, to mimic the utilization of scans in a clinical workflow. Intra-physician and inter-physician variability were compared using several metrics (dice coefficient, mean-distance-to-agreement, Hausdorff distance, and contour volume). Qualitative ratings were assigned based on tumor conspicuity, edge sharpness, contouring confidence, and overall image quality.
Results: The addition of the M50 to the C120 image resulted in a statistically significant increase in contour volume. The average contour volume continued to increase with the inclusion of the MR (V(C120)=24.5cc, V(M50)=26.5cc, V(MR)=30.8cc, p<0.05, p<0.05 ). The intra-physician variability demonstrated better agreement between the MR and M50 contours as compared to the MR and C120 contours. Inter-physician variability showed no statistical change between M50 and C120 but increased with the addition of MRI. The largest deviations were seen with average Hausdorff distance (HD(C120)=8.62, HD(M50)=9.01, HD(MR)=10.63, p=0.2, p=0.04). The qualitative results showed an increase in ratings for each category with the addition of the M50 and the MR image.
Conclusion: The addition of the M50 image resulted in changes to tumor segmentation compared to the 120 kVp image and higher physician confidence in segmentation. While the MR was the most impactful for contour adjustment, inter-physician agreement between contours decreased, suggesting subjectivity in the contouring of non-FDG avid enhancing tissue.
Funding Support, Disclosures, and Conflict of Interest: This research was supported, in part, by research funding from Siemens Healthineers
Keywords
Dual-energy Imaging, Segmentation
Taxonomy
IM- CT: Dual Energy and Spectral
Contact Email