Predictive Metrics for Moist Desquamation in Treatment Planning for Breast Radiotherapy

Abstract: The objective of this thesis was to establish techniques for predicting and modelling moist desquamation (MD). The work in this thesis is based on the clinical studies of the novel Carbon Fibre Adjustable Reusable Accessory (CARA) breast support device. The epidermal dose measurements, treatment plans, and skin assessments from clinical studies of the CARA device were used to develop models to be applied during treatment planning to reduce the occurrence of MD in breast radiotherapy.

In the first study of this thesis, in vivo film dosimetry was used to establish a relationship between the epidermal surface area receiving various levels of radiation dose and the corresponding skin reactions. 

The results of this study indicated potential dose-area based constraints on the skin for use during treatment planning.

For the second study, the in vivo film dosimetry was compared against the Eclipse(Varian Medical Systems) treatment planning system’s Analytical Anisotropic Algorithm (AAA) and AcurosXB (AXB) dose calculation algorithms. This study produced recommendations for skin rind definitions for the two dose calculation algorithms to improve consistency in epidermal dose reporting.

In the final study, a metric was developed to predict MD based on the spatial distribution of dose across the skin. This metric maps the risk of MD across the skin’s surface based on the treatment planning system’s skin dose calculations. The predictive model was validated against a second dataset, and showed promise in its performance.