ESTRO 2024 - Abstract Book

S5616

RTT - Patient care, preparation, immobilisation and IGRT verification protocols

ESTRO 2024

clinical trials recording of breast/trunk lymphoedema is often reduced to a single measure, often only physician reported, on a 4-point subjective scale (3) .

The aim of this study is to evaluate the system capability of a commercially available 3D surface scanning system for detecting small surface area changes; to identify lymphoedema in real-time within the existing breast cancer radiotherapy pathway. The focus was on measuring system sensitivity and consistency to identify small changes in body shape (using existing available software) that may be indicative of developing breast or trunk lymphoedema.

Material/Methods:

A phantom study was conducted using silicone and cotton breast forms with varying sizes (equivalent to UK B, C, E and G-cup breasts). Phantoms were scanned across different skin tones to test system capability across pale skin through to dark skin to simulate different ethnicities. Phantoms were scanned on a CIVCO C-Qual breastboard, with the board angle adjusted to accommodate the slope and shape of the bust forms. A synthetic silicone pad to simulate lymphoedema (Figure 1) was placed superiorly (SUP), inferiorly (INF) and laterally (LAT) to the breast tissue. Images were taken using the VisionRT, AlignRT® system. Two images were taken at baseline (without the fake lymphoedema pad), 2 images with the lymphoedema pad placed superiorly, 2 images with the lymphoedema pad placed inferiorly and 2 images with the lymphoedema pad placed laterally. Images were exported to an image editing software tool (MeshMixer-Autodesk Inc, USA) and assessed for reliability by 3 raters, all blinded. Forty scans across different sizes of chest and breast and across different skin tones were assessed to determine reliability. Images were refined to focus on the breast/trunk region of interest (ROI). The surface area (mm 2 ) was obtained from each scan and inputted into SPSS version 26 to assess the inter and intra-rater reliability using intraclass correlation (ICC).

Descriptive data have been used to identify differences in surface area between the baseline scans and subsequent scans simulating lymphoedema. Repeatability has also been assessed between the breast form sizes.

Results:

Inter-rater reliability for 3 independent raters across 4 different breast cup sizes and 2 different skin tones (n= 40 images per rater) was excellent (intraclass correlation coefficient ICC = 0.9993 95% CI 0.986 to 0.997). Intra-rater reliability also showed excellent within-rater agreement across the same breast cup sizes and skin tones; ICC was 0.999, 0.999 and 0.994 for rater 1, 2 and 3 respectively (95% confidence intervals for all 3 raters was between 0.98 to 1). Consistency of measuring surface area differed between the breast cup sizes. Breast cup size 4 (G cup) showed the least consistency in the measurement of surface area across the 3 raters (inter-rater ICC=0.140) although across 2 raters ICC=0.865 reflecting the difficulty in gaining consistent assessment where there maybe data loss because of the very rigid large breast phantom shape. Similarly, consistency in measuring across the 3 raters was lower for the phantom with the darkest skin tone (Figure 2) (ICC=0.565 3 raters, ICC 2 raters 0.848).

The system showed ability to detect small changes in surface area across all 4 breast cup sizes and across the 2 different skin tones detecting surface area changes of 1%.