3395 - Dose Accumulation and Geometric Variation of Local Recurrence in Anal Cancer Patients Receiving Concurrent Chemoradiation

Purpose/Objective(s): Local regional recurrence is the primary pattern of failure in anal squamous cell carcinoma patients after definitive chemoradiation. Improved target localization and dose escalation remain potential strategies for better local control. This study explores the relationship between the area of local recurrence and potential variations in dose and anatomy during the course of treatment.

Materials/Methods: Between 2008 and 2013, 101 patients were treated with daily image-guided IMRT at the local institution. Fourteen patients with local recurrence were identified. All were prescribed 63 Gy to primary tumors using sequential phase planning and a 0.7 cm PTV margin. Volume of tumor recurrence (RTV) were manually delineated on planning CT which were rigidly co-registered with diagnostic MR.

Hybrid deformable image registration was used to register the plan-CT and daily CBCT with controlling ROIs of external anal sphincter, rectum, bladder, bone, and rectum POIs. All controlling ROIs and POIs were manually delineated on CBCT by a single user and independently reviewed. Planning CTVs were mapped to daily CBCT (CTV_map) and the residual displacement of the CTV_map centroids following image guidance were calculated. Location of the deformed CTV exceeding planning PTV margin was reviewed. CBCT dose reconstruction was performed to estimate delivered dose to the CTV and RTV.

Results: A total of 16 CTs and 484 CBCTs were evaluated, 2 patients were rescanned and replanned due to substantial tumor shrinkage.

Overall, the median delivered CTV doses were: D98= 60.7 Gy (R: 49.6 - 62.7), D50= 63.1 Gy (R: 59.5 - 63.9), D2= 64.6 Gy (R: 60.6 - 65.6). The median CTV D98 delivered dose was 0.4 Gy or 1% lower than corresponding plans, with the largest deviation of -8.4 Gy or -14%. The mean absolute CTV centroid displacements were 0.1 cm (R: 0.3 right - 0.2 left), 0.4 cm (R: 0.5 inferior - 0.5 superior), and 0.2 cm (R: 0.3 anterior - 0.5 posterior). Seven patients had systematic centroid displacements over the course of treatment.

Median RTV measured 11.0 cm³ (1.0 - 27.0). The median RTV delivered doses were: D98= 61.3 Gy (R: 44.1 - 63.7), D50= 62.8 Gy (R: 58.7 - 64.2), D2= 63.9 Gy (R: 59.4 - 65.0) Gy. The CTV_map outside PTV occurred in 303 CBCTs, median volume 1.8 cm³ (R: 0.01 - 179.5), and the location coincided with tumor recurrence site of six patients based on qualitative image review. Five of the 6 presented with linear trends (¦r ¦= 0.3 - 0.7) of CTV centroid displacements during fractions 1-20. This may contribute to -2% RTV D98 dose in this group of patients.

Conclusion: Implementation of dose escalation requires an improved understanding of dosimetric and geometric variation of targets previously limited by an inability to analyze delivered doses. In this initial small cohort, we observed minimal overall dose deviation. However, individual variation occurs. Timely online or offline treatment adaptation may benefit patients with a consistent positioning and volume variation.