Machine-learning prediction model for acute skin toxicity after breast radiation therapy using spectrophotometry

Aug 5, 2026

Upload by Jiven

Radiation-induced skin toxicity is a common and distressing complication after breast radiotherapy. This study evaluates quantitative spectrophotometric skin markers—melanin and erythema indices—as input features to build supervised machine learning models for predicting acute skin toxicity. Data from 129 patients receiving adjuvant whole-breast radiotherapy were assessed at baseline, end of treatment, and 1 and 6 months post-treatment, with toxicity graded using RTOG criteria.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Purpose

Methods and materials

Study population and feature extraction

Modeling approach

Results

Conclusions

TYPE Original Research PUBLISHED 06 January 2023 DOI 10.3389/fonc.2022.1044358 OPEN ACCESS EDITED BY Yang Sheng, Duke University Medical Center, United States REVIEWED BY Yibo Xie, Winship Cancer Institute, Emory University, United States Javad Hassannataj Joloudari, University of Birjand, Iran Dongrong Yang, Duke University Medical Center, United States *CORRESPONDENCE Savino Cilla [savino.cilla@gemellimolise.it](savino.cilla@gemellimolise.it) [savinocilla@gmail.com](savinocilla@gmail.com) †These authors share senior authorship SPECIALTY SECTION This article was submitted to Breast Cancer, a section of the journal Frontiers in Oncology RECEIVED 10 October 2022 ACCEPTED 08 December 2022 PUBLISHED 06 January 2023 CITATION Cilla S, Romano C, Macchia G, Boccardi M, Pezzulla D, Buwenge M, Castelnuovo AD, Bracone F, Curtis AD, Cerletti C, Iacoviello L, Donati MB, Deodato F and Morganti AG (2023) Machine-learning prediction model for acute skin toxicity after breast radiation therapy using spectrophotometry. Front. Oncol. 12:1044358 . doi: 10.3389/fonc.2022.1044358 COPYRIGHT © 2023 Cilla, Romano, Macchia, Boccardi, Pezzulla, Buwenge, Castelnuovo, Bracone, Curtis, Cerletti, Iacoviello, Donati, Deodato and Morganti. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Machine-learning prediction model for acute skin toxicity after breast radiation therapy using spectrophotometry Savino Cilla 1*, Carmela Romano 1, Gabriella Macchia 2, Mariangela Boccardi 2, Donato Pezzulla 2, Milly Buwenge 3, Augusto Di Castelnuovo 4, Francesca Bracone 5, Amalia De Curtis 5, Chiara Cerletti 5, Licia Iacoviello 5,6, Maria Benedetta Donati 5, Francesco Deodato 2,7† and Alessio Giuseppe Morganti 3,8† 1 Medical Physics Unit, Gemelli Molise Hospital, Campobasso, Italy, 2 Radiation Oncology Unit, Gemelli Molise Hospital, Campobasso, Italy, 3 Radiation Oncology, Istituti di Ricovero e Cura a Carattere Scientiﬁco (IRCCS) Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy, 4 Mediterranea Cardiocentro, Napoli, Italy, 5 Department of Epidemiology and Prevention, IRCCSNeuromed, Pozzilli, Italy, 6 Department of Medicine and Surgery, Research Center in Epidemiology and Preventive Medicine (EPIMED), University of Insubria, Varese, Italy, 7 Istituto di Radiologia, Universit´a Cattolica del Sacro Cuore, Rome, Italy, 8 Department of Experimental, Diagnostic, and Specialty Medicine-DIMES, Alma Mater Studiorum Bologna University, Bologna, Italy Purpose: Radiation-induced skin toxicity is a common and distressing side effect of breast radiation therapy (RT) . We investigated the use of quantitative spectrophotometric markers as input parameters in supervised machine learning models to develop a predictive model for acute radiation toxicity. Methods and materials: One hundred twenty-nine patients treated for adjuvant whole-breast radiotherapy were evaluated. Two spectrophotometer variables, i. e . the melanin (IM) and erythema (IE) indices, were used to quantitatively assess the skin physical changes . Measurements were performed at 4-time intervals: before RT, at the end of RT and 1 and 6 months after the end of RT. Together with clinical covariates, melanin anderythema indices were correlated with skin toxicity, evaluated using the Radiation Therapy Oncology Group (RTOG) guidelines. Binary group classes were labeled according to a RTOG cut-off score of ≥ 2. The patient’s dataset was randomly split into a training and testing set used for model development/ validation and testing (75%/25% split) . A 5-tim

FAQs

What spectrophotometric inputs were used to predict acute skin toxicity?

Melanin index (IM) and erythema index (IE) were measured and used as quantitative input parameters, along with relevant clinical covariates.

How was skin toxicity labeled for model training and evaluation?

Patients were grouped using Radiation Therapy Oncology Group (RTOG) guidelines, with binary labels based on a cut-off score of RTOG ≥ 2.

Which machine learning model performed best in the reported results?

The support vector machine (SVM) with an RBF kernel showed the strongest performance, including the highest reported diagnostic accuracy and F-score.

Aug 5, 2026

Upload by Jiven