Encrypted login | home

Program Information

Predicting Treatment Response According to Metabolic Imaging Parameters Using Deformable Registration of Intra-Treatment 18F-FDG-PET-CT for Head and Neck Squamous Cell Carcinoma (HNSCC)

I Vergalasova

I Vergalasova1*, Y Mowery2 , D Brizel2 , S Das3 , (1) Rutgers Robert Wood Johnson Medical School, New Brunswick, NJ, (2) Duke University Medical Center, Durham, NC, (3) University of North Carolina, Chapel Hill, NC


TH-AB-FS4-7 (Thursday, August 3, 2017) 7:30 AM - 9:30 AM Room: Four Seasons 4

Purpose: To assess the best metabolic imaging parameters for correlating treatment response using deformable registration of intra-treatment to pre-treatment 18F-FDG-PET-CT for patients undergoing radiation therapy for head and neck squamous cell carcinoma (HNSCC).

Methods: 18F-FDG-PET-CT was acquired pre-treatment and repeated after 20Gy. Primary and nodal volumes (GTVPRI & GTVLN) were contoured on the pre-treatment scan by the same physician across all patients. Deformable registration was applied to transfer the volumes to the intra-treatment scan. The following parameters were absolutely extracted per GTV on the intra-treatment scan: Standardized Uptake Values (SUVMAX, SUVMEAN, SUV20%, SUV40%, SUV60%, SUV80%), Metabolic Tumor Volume (MTV50) and Tumor Lesion Glycolysis (TLG50, TLGTOTAL). SUVX% is the minimum SUV to the highest-intensity X% volume. SUV was normalized to blood pool uptake. Treatment response was determined via 12 week post-treatment 18F-FDG-PET-CT, serial history, physical exam, and biopsy if required. Wilcoxon Rank Sum (WRS) evaluated significant differences between the complete and incomplete response groups (CR vs. ICR). Cox Proportional Hazards Regression (CPHR) was computed as a function of progression-free survival with SUV, MTV, and TLG analyzed as continuous variables.

Results: Seventy-five patients (CR=58; ICR=17) were analyzed. Median follow-up time was 2.4 years. For GTVPRI, WRS determined significant differences (p<0.05) between CR vs. ICR groups across all parameters except TLGTOTAL. For GTVLN, WRS demonstrated significance for SUVMAX, SUVMEAN, SUV40%, SUV60%, SUV80% and TLG50. Median (interquartile range) values of SUV60% (p<0.05 for GTVPRI & GTVLN) for CR vs. ICR were as follows: GTVPRI=1.6 (1.3, 2.2) vs. 2.1 (1.9, 3.1) and GTVLN=1.3 (1.0, 1.6) vs. 1.6 (1.3, 2.3). CPHR revealed significant hazard ratios of SUV40%=1.26, SUV60%=1.37, SUV80%=1.47 and TLG50=1.01 for GTVPRI, but none for GTVLN.

Conclusion: These data suggest that metabolic imaging parameters derived from intra-treatment 18F-FDG-PET-CT for primary and nodal GTVs using deformable registration are capable of predicting treatment outcomes for HNSCC patients.

Contact Email: