Head and Neck Squamous Cell Carcinoma of Unknown Primary: Is There a Better Treatment?

Abstract

Materials/Methods: Eleven patients with evidence of local and/or regional failure following curative-intent IMRT for nasopharyngeal carcinoma or cervical node carcinoma from unknown primary between 2004 and 2010 were retrospectively reviewed. Radiologic evident recurrent gross disease (rGTV) was manually delineated on the diagnostic CT scans documenting recurrence (rCT) and co-registered with the planning simulation CT (pCT) using validated atlas-based DIR software. Subsequently, deformation fields were applied to rGTV segmented on the rCT to convert them into deformed rGTV on the pCT. Evaluation of deformed rGTVs relative to original planning TVs and prescribed radiation dose was done using both volume overlap and focal methods. The focal method assumed the center of mass of deformed rGTV was the origin of the recurrence volume (rGTVorg) and its location was compared relative to planning TVs and dose. Rigid image registration (RIR) was also implemented to compare the spatial and dosimetric criteria of mapped rGTVs using both registration methods. Paired-samples Wilcoxon signed rank test was used to compare analysis metrics. Results: A total of thirteen rGTVs were identified. Using DIR, the focal method significantly assigned failures to more central TVs compared to the volume overlap method (p Z 0.009) where 12 (92.3%) rGTVorgs were located at GTV and one (7.7%) rGTVorg at CTV2 while the 95% volume of rGTV overlapped with GTV for 4 failures, CTV1 for 5 failures, and CTV2 for 4 failures. Moreover, failures mapped using DIR were assigned to more central TVs compared to RIR when compared using the volume overlap method (p Z 0.03) and the focal method (p Z 0.02). The dose to 95% failure volume was higher than the 95% prescribed dose for 12 out of 13 deformed rGTVs compared to 9 out of 13 for rGTVs mapped rigidly. Additionally, deformed rGTVs had significantly higher mean doses when compared to rGTVs mapped rigidly (70.93 vs 70.27 Gy, p Z 0.019). Conclusions: The vast majority of failures originated in the high dose target volumes and received full prescribed doses suggesting biological rather than technology-related causes of failure. Workflow methodology involving the use of validated DIR software allows more accurate evaluation of the patterns of failure compared to rigid registration and is recommended for large-scale future implementation. Author Disclosure: A.S. Mohamed: None. M. Awan: None. E. Kocak: None. B.M. Beadle: None. M.E. Kantor: None. G.B. Gunn: None. A.S. Garden: None. D.I. Rosenthal: None. C.D. Fuller: E. Research Grant; National Institutes of Health Cancer Center Support (Core) Grant CA016672 to The University of Texas MD Anderson Cancer Center. Dr. Fuller received/receives grant/funding support from the SWOG/Hope Fo.