Encrypted login | home

Program Information

Patterns of Failure for Laryngeal Cancer Patients Treated with Definitive IMRT: Comparing Two Different Methods for Determining the Origin of Recurrence From Follow-Up PET/CT Scans

no image available
P Brodin

P Brodin1*, R Kabarriti2 , I Vogelius3 , C Guha1 , S Kalnicki2 , M Garg2 , W Tome1 , (1) Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, New York, (2) Montefiore Medical Center, Bronx, New York, (3) Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark


SU-F-J-223 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To determine patterns of failure in laryngeal cancer treated with definitive IMRT by comparing two different methods for identifying the recurrence epicenter on follow-up PET/CT.

Methods: We identified 20 patients treated for laryngeal squamous cell carcinoma with definitive IMRT who had loco-regional recurrence diagnosed on PET/CT. Recurrence PET/CT scans were co-registered with the original treatment planning CT using deformable image registration with the VoxAlign deformation engine in MIM Software. Recurrence volumes were delineated on co-registered follow-up scans using a semi-automatic PETedge tool and two separate methods were used to identify the recurrence point of origin:

a) Finding the point within the recurrence volume for which the maximum distance to the surface of the surrounding recurrence volume is smaller than for any other point.

b) Finding the point within the recurrence volume with the maximum standardized uptake value (SUVmax), without geometric restrictions.

For each method the failure pattern was determined as whether the recurrence origin fell within the original high-dose target volumes GTV70, CTV70, PTV70 (receiving 70Gy), intermediate-risk PTV59 (receiving 59.4Gy) or low-risk PTV54 (receiving 54.1Gy), in the original treatment planning CT.

Results: 23 primary/nodal recurrences from the 20 patients were analyzed. The three-dimensional distance between the two different origins was on average 10.5mm (std.dev. 10mm). Most recurrences originated in the high-dose target volumes for both methods with 13 (57%) and 11 (48%) in the GTV70 and 20 (87%) and 20 (87%) in the PTV70 for method a) and b), respectively. There was good agreement between the two methods in classifying the origin target volumes with 69% concordance for GTV70, 89% for CTV70 and 100% for PTV70.

Conclusion: With strong agreement in patterns of failure between two separate methods for determining recurrence origin, we conclude that most recurrences occurred within the high-dose treatment region, which influences potential risk-adaptive treatment strategies.

Contact Email: