2462 - MR vs. CT-Based Cardiac Delineation in Pediatric Patients with Hodgkin Lymphoma
Presenter(s)
S. Ramesh1, Y. Gao2, B. Li2, J. R. Rocchetti3, C. E. Hill-Kayser1,4, G. Kurtz1,4, H. G. Hubbeling1, J. P. Plastaras1, M. M. Kim2, and M. J. LaRiviere1,4; 1Department of Radiation Oncology, Hospital of the University of Pennsylvania, Philadelphia, PA, 2Department of Radiation Oncology, Division of Medical Physics, Hospital of the University of Pennsylvania, Philadelphia, PA, 3Rowan-Virtua School of Osteopathic Medicine, Stratford, NJ, 4Children's Hospital of Philadelphia, Philadelphia, PA
Purpose/Objective(s): Radiation therapy (RT) plays a central role in the management of Hodgkin lymphoma (HL), and pediatric patients often receive PET/MRI, rather than PET/CT staging and response-assessment imaging to reduce radiation exposure. Because long-term overall survival rates with multimodality treatment now exceed 85% for this disease, significant effort is directed toward reducing such late toxicities as cardiotoxicity. We hypothesized standard-of-care response-assessment MRI would allow for more accurate delineation of the heart vs. CT simulation alone, allowing for reductions in RT dose to the heart.
Materials/Methods: 22 pediatric patients, previously treated with RT for mediastinal HL, who had a dedicated MRI of the chest during the workup or restaging of their disease, were identified. 18 of these patients were simulated using deep inspiration breath hold (DIBH) technique. The most recent MRI prior to simulation was rigidly registered to the planning CT, focusing on alignment to the heart and sternum. A new cardiac structure was contoured using the registered MRI and transferred to the planning CT. The maximum (Dmax), mean (Dmean), and median (D50%) doses were then re-calculated using the original RT plan and compared to those of the original CT-based cardiac structure. 2-tailed paired t-tests were performed, and results are reported as mean +/- standard deviation.
Results: Heart mean Dmax was 2564 +/- 464 for MRI vs. 2565 +/- 443 cGy for CT (p=0.96), mean Dmean 924 +/- 589 for MRI vs. 796 +/- 470 cGy for CT (p=0.07), and mean D50% 791 +/- 847 for MRI vs. 555 +/- 700 cGy for CT (p=0.06). Subjectively, delineation of the heart was more challenging using MRI vs. CT, and identification of substructures was more challenging to the extent that comparative dosimetry would not be possible. Registration issues were most evident for patients undergoing CT simulation with DIBH. As reviewed by an experienced pediatric radiation oncologist, only subjectively small differences in heart contours were noted and would not have led to clinically meaningful changes in CTV, iCTV, or PTV.
Conclusion: No statistically significant differences were seen in mean Dmax, Dmean, or D50%, and delineating the heart was more difficult using diagnostic MRI vs. simulation CT. Although we hypothesized that diagnostic MRI would aid in more accurate delineation of heart at the time of simulation and that this could improve cardiac dosimetry, it is possible that CT may already provide an optimal, cost-effective approach. Approaches that are currently standard of care – involved site RT, DIBH, conformal RT with IMRT/VMAT or proton therapy, and patient selection for combined modality vs. systemic therapy-only approach – may offer sufficient improvements in cardiotoxicity. Nonetheless, future directions may explore dedicated cardiac MRI, obtained in the simulation position, at the time of simulation.