Quantitative Thermal Imaging Biomarkers to Detect Acute Skin Toxicity from Breast Radiotherapy Using Supervised Machine Learning.

Quantitative Thermal Imaging Biomarkers to Detect Acute Skin Toxicity from Breast Radiotherapy Using Supervised Machine Learning.:

Related Articles

Quantitative Thermal Imaging Biomarkers to Detect Acute Skin Toxicity from Breast Radiotherapy Using Supervised Machine Learning.

Int J Radiat Oncol Biol Phys. 2020 Jan 23;:

Authors: Saednia K, Tabbarah S, Lagree A, Wu T, Klein J, Garcia E, Hall M, Chow E, Rakovitch E, Childs C, Sadeghi-Naini A, Tran WT

Abstract

PURPOSE: Radiation-induced dermatitis is a common side effect of breast radiotherapy (RT). Current methods to evaluate breast skin toxicity include clinical examination, visual inspection, and patient-reported symptoms. Physiological changes associated with radiation-induced dermatitis, such as inflammation, may also increase body-surface temperature which can be detected by thermal imaging. Quantitative thermal imaging markers were identified using supervised machine-learning to develop a predictive model for radiation dermatitis.

METHODS: Ninety patients treated for adjuvant whole-breast radiotherapy (4250 Gy/fx=16) were recruited to the study. Thermal images of the treated breast were taken at four intervals: prior to RT, then weekly, at fx=5, fx=10, and fx=15. Parametric thermograms were analyzed and yielded 26 thermal-based features which included surface temperature (°C) and texture parameters obtained from 1) grey-level co-occurrence matrix (GLCM), 2) grey-level run-length matrix (GLRLM) and 3) neighborhood grey-tone difference matrix (GTDM). Skin toxicity was evaluated at the end of RT using the Common Terminology Criteria for Adverse Events (CTCAE) guidelines (Ver.5). Binary group classes were labelled according to a CTCAE cut-off score of ≥2, and thermal features obtained at fx=5 were used for supervised machine learning to predict skin toxicity. The dataset was partitioned for model training, independent testing, and validation. Fifteen patients (∼ 17% of the whole dataset) were randomly selected as an unseen test dataset, and 75 patients (∼ 83% of the whole dataset) were used for training and validation of the model. A random forest classifier with leave-one-patient-out cross-validation was employed for modelling single and hybrid parameters. The model performance was reported using receiver operating characteristic analysis on patients from an independent test set.

RESULTS: Thirty-seven patients presented with adverse skin effects, denoted by a CTCAE score ≥2, and had significantly higher local increases in skin temperature, reaching 36.06°C at fx=10 (p=0.029). However, machine-learning models demonstrated early thermal signals associated with skin toxicity after the fifth RT fraction. The cross-validated model showed high prediction accuracy (Acc) on the independent test data (test Acc=0.87) at fx=5 for predicting the skin toxicity at the end of RT.

CONCLUSION: Early thermal markers after five fractions of RT are predictive of radiation-induced skin toxicity in breast radiotherapy.

PMID: 31982495 [PubMed - as supplied by publisher]