Radiation and Cancer Physics
SS 29 - Physics 9 - Imaging for Treatment Planning
207 - Utilization of 18FDG-PET-Based IMRT Planning to Spare the Most Metabolically Active Subvolume of Parotid Glands During Definitive Radiation Therapy for Head and Neck Cancer
Tuesday, October 23
4:55 PM - 5:05 PM
Location: Room 304
Utilization of 18FDG-PET-Based IMRT Planning to Spare the Most Metabolically Active Subvolume of Parotid Glands During Definitive Radiation Therapy for Head and Neck Cancer
H. Marley1, Y. M. Mowery2, Q. Wu2, and D. M. Brizel2; 1Duke University School of Medicine, Durham, NC, 2Duke University Medical Center, Durham, NC
Purpose/Objective(s): To evaluate 18FDG-PET-based IMRT planning to reduce radiation dose to the most metabolically active regions of the parotid glands during radiation therapy for head and neck squamous cell carcinoma (HNSCC) with the goal of reducing xerostomia.
Materials/Methods: 20 patients with HNSCC (19 oropharynx; 1 nasopharynx) underwent 18FDG-PET and contrast-enhanced CT simulation for definitive radiation treatment planning. The original IMRT treatment plan used clinically was generated based on the pretreatment CT with volumetric-based dose constraints for each parotid gland. Prescription was 44 or 50 Gy (low risk PTV) and 70 Gy (high risk PTV) at 2 Gy/fraction. Patients were retrospectively replanned with added metabolically-based constraints for each parotid gland. SUV thresholding was performed in Eclipse to generate a contour encompassing the top one-third most metabolically active region of each parotid gland (SUV-High). SUV-High contours were transferred to the registered CT for the replanning. The original IMRT plans were re-optimized to reduce dose to SUV-High within each parotid while maintaining target coverage and meeting initial dose constraints for all other organs at risk. Median doses (Dmedian) to SUV-High and the entire gland for each parotid were calculated for the original and revised plans. Wilcoxon matched-pairs signed rank test (2-tailed, p<0.05) was performed to compare Dmedian in the initial vs. metabolically-based plan for each structure.
Results: SUV-High represented a median of 44% (IQR 38.9 – 52.4%) and 40.7% (IQR 21.7 – 54.8%) of the left and right parotid gland volumes, respectively. Median dose differed significantly between the metabolically-based (replan) and clinical (original) IMRT plans for the entire gland and SUV-High for each parotid (p < 0.0001). The table shows Dmedian from each plan, as well as relative percent change in Dmedian for the metabolically-based plan relative to the original plan for each structure.
| || SUV-High, Left Parotid || Whole Gland, Left Parotid || SUV-High, Right Parotid || Whole Gland, Right Parotid |
| Original Median Dmedian (IQR), Gy || 14.3 (11.6, 25.8) || 16.3 (13.8, 22.3) || 18.3 (10, 27.6) || 18.8 (14.8, 22.7) |
| Replan Median Dmedian (IQR), Gy || 13.7 (10.7, 21.5) || 15 (13.2, 20.1) || 15.5 (9.3, 24.4) || 16.8 (13.9, 19.6) |
| Median % Change (IQR) || -8.1% (-12.8, -3.9%) || -4.6% (-10.2, 0.1%) || -13.7% (-18, -7.1%) || -6% (-11.8, -0.2) |
Conclusion: We have demonstrated that it is feasible to utilize 18FDG-PET information in IMRT planning to significantly reduce radiation dose to the most metabolically active sub-volume and whole gland for both parotids. Prospective evaluation is necessary to determine whether this approach can reduce late treatment-related xerostomia for head and neck cancer patients.
Author Disclosure: H. Marley: None. Y.M. Mowery: Employee; Duke Health. Research Grant; Radiological Society of North America (RSNA), SARC. Q. Wu: None. D.M. Brizel: Royalty; Up To Date. Committee Chair; ASTRO.