SU‐GG‐J‐163: Radiotherapy Response Assessment Using Deformed Serial 18F‐FDG PET/CT

• Published in: Medical Physics 2008-06, Vol.35 (6), p.2717-2717
• Main Authors: Cannon, B, Zhang, L, Mawlawi, O, Schwartz, D, Moeller, B, Dong, L
• Format: Article
• Language: English
• Subjects: Biomedical modeling; Computed tomography; Image analysis; Image registration; Image scanners; Materials analysis; Medical imaging; Positron emission tomography; Radiation therapy
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.2961712

Purpose: To assess the ability of SUV measured on serial 18 F ‐ FDG PET/CT to differentiate radiotherapy (RT) responders from non‐responders. Method and Materials: Between November 2005 and August 2007, 88 eligible patients with AJCC stage III‐IVb HNSCC were enrolled in an IRB approved protocol to receive serial PET/CT imaging studies prior to and following RT. Analysis of similar anatomic volumes on temporally separated images was facilitated with a deformable image registration technique to map pre‐ and post‐RT CT images to a reference image set, namely the RT planning CT images. The resulting deformable transformations were applied to the pre‐ and post‐RT PET images, thereby aligning both PET data sets to the RT planning CT images. SUV reduction factors were created from pre‐ and post‐RT SUV ratios and calculated for RT contours and fixed thresholds of the maximum signal intensity on PET images. Variations in pre‐ and post‐RT PET/CT scanner model and uptake duration were controlled to eliminate possible confounding factors for intrapatient analysis. Results: Cross correlation of patients whose serial PET/CT images were collected with the same scanner model and with uptake duration differences ⩽ 15 min reduced the useable number of patients drastically from 88 to 10. For the controlled patient cohort, the average SUV reduction factor for responders (N = 7) vs. non‐responders (N = 3) calculated for the GTV contour was (1.4 ± 1.6) and (1.0 ± 1.7), respectively. For all contours (N = 4), nonparametric testing revealed [P > 0.05]. Conclusion: The average SUV reduction factor was not able to significantly differentiate responders and non‐responders in all contours studied. The large reduction in useable patients following strict data set control may be an important cautionary note for future studies investigating serial PET/CT for treatment response monitoring. Research sponsored by Varian Medical Systems.