Significant Inter-Fractional Variations of Lumpectomy Cavity in Partial Breast Irradiation
X Chen*, J White, N Morrow, J Wilson, X Li, Medical College of Wisconsin, Milwaukee, WISU-C-213CD-3 Sunday 1:30:00 PM - 2:15:00 PM Room: 213CD
Inter-fractional variation in partial breast irradiation (PBI) using external beams is a significant issue. Image guided RT (IGRT), by imaging and repositioning the patient prior to the delivery based on rigid-body registration, has been introduced to reduce inter-fractional setup and translational variations. However, IGRT is not capable of addressing organ deformation and rotation. The purpose of this work is to quantitatively characterize those variations in PBI in supine position.
A total of 100 diagnostic-quality CT sets acquired using an in-room CT during IGRT at each fraction for 10 breast cancer patients treated with PBI in supine position were analyzed. Targets and organs at risk, including the lumpectomy cavity, treated breast, lung and heart, were delineated by populating the contours from the planning CT to each fraction CT using an auto-segmentation tool (ABAS, CMS/Elekta) with manual editing. Various parameters, including volume, maximum overlap rate (MOR), dice's coefficient, and distance between centers of mass (COM) were calculated to quantify the inter-fractional variations.
The shape and volume of the lumpectomy cavity in the fractional CTs, particularly in the first fraction CT, change significantly from those in the planning CT. On average, the volume of the lumpectomy cavity decreases by 23%. The MOR varies from 50% to 90% with the mean value of 72%. The dice's coefficient is shown in a similar range. The variation of the treated breast is relatively small. The changes in the distances between the COMs of the lumpectomy cavity and the heart and of the lumpectomy cavity and the treated breast are up to 3 mm.
The inter-fractional lumpectomy cavity deformation is significant during PBI. Strategies that can address not only the translational motion but also the deformation, such as adaptive RT, should be explored for PBI.