An Analysis of Daily Setup Variation in Prone Breast Radiation of Early-Stage Breast Cancer

Abstract

For women with left-sided breast cancer, the most relevant side effect is collateral damage to the heart with radiation therapy. Prone treatment techniques have helped reduce heart dose, but is more unstable and difficult to replicate. Setup uncertainties can lead to higher than expected heart doses or under-dosing of breast tissue. Our objective was to investigate the potential dosimetric benefit of utilizing a low dose cone beam CT (CBCT) in the daily setup of prone left breast radiation.We performed a retrospective review patients treated at our institution between April 2019 and February 2020 in the prone position for left sided early-stage breast cancer. Patients were set up to clinical marks and a custom low dose partial arc CBCT was acquired with each fraction. Using the daily shifts, we calculated the delivered dose if these shifts were not performed. Differences between planned and pre-CBCT shift doses were recorded for the heart and breast tissue. Patients were grouped into 3 outcomes: (A) Significantly high heart dose: More than 10% increase in delivered heart dose from planned; (B) Poor Breast Coverage: More than 5% decrease in the volume receiving 90% of the prescription dose; or (C) Minimal Variation: No significant increase in heart dose or decrease in breast coverage. Patient factors were also collected. Statistical comparisons were made using t-tests.18 patients were included, with 283 daily CBCT shifts and an average of 15.8 CBCT's per patient (range 13-25). The mean age of the patients was 61.6 (39-80) and average BMI 28.3 (20.3 - 45). Breast and heart dose were available for 196 scans, of which 47 (24%) had high heart dose (A), 55 (28%) had poor breast coverage and only 98 (50%) had no more than minimal variation in either. Patients with breast size < 1000 cc had a 64.5% chance of a delivery with poor breast coverage or increased heart dose as opposed to 40% for women with larger breast (P < 0.01). The same relationship was seen for weight (P < 0.01), refer to Table 1. When combining the total dose for each patient, 3 (17%) patients were in group A (increased heart dose), 9 patients (50%) were in group B (poor breast coverage) and 6 patients (33%) were in group C. Patients with minimal variation had significantly higher weight (84 kg vs 66 kg, P = 0.038), BMI (33 vs 26, P = 0.03) and breast volumes (1716 cc vs 992 cc, P = 0.01).Overall, our study demonstrated significant variation in daily setup for patients receiving prone breast radiation. Only half of treatments would have been delivered with minimal variation and a majority of patients would have either received significantly increased heart dose or suboptimal breast coverage if setup errors were not corrected. Likelihood of a significant setup error was seen most in women with smaller breasts or lower weight.