AI tool for Predicting Side Effects in Breast Cancer Patients Post-Surgery and Radiotherapy
Authored by the German Organization for Research and Treatment of Cancer
Credit: Pixabay/CC0 Public Domain
An innovative artificial intelligence (AI) tool has been developed by an international team of researchers to pinpoint breast cancer patients who might be susceptible to side effects subsequent to undergoing surgery and radiotherapy.
During the 14th European Breast Cancer Conference (EBCC14) in Milan, Dr. Tim Rattay unveiled that a clinical trial will be initiated across three countries – France, The Netherlands, and the UK – to evaluate the efficacy of this tool.
This AI application, known for its explainable nature, provides transparent reasoning behind its decisions. Dr. Rattay, a consultant breast surgeon and Associate Professor at the University of Leicester (UK), emphasized that this feature simplifies decision-making for physicians and enables them to furnish evidence-based explanations to their patients.
The PRE-ACT project (Prediction of Radiotherapy Side Effects using explainable AI for patient communication and treatment modification) not only aims to enhance the accuracy of predictions tailored to individual patients but also to offer comprehensible explanations to both doctors and patients. Some of the known factors that contribute to the risk of side effects are being considered in this initiative.
Dr. Rattay articulated, “While the long-term survival rates for breast cancer patients are on the rise, some individuals have to contend with the aftermath of their treatment. This includes physical alterations, scarring, lymphedema (swelling of the arm), and potential heart damage from radiation therapy. To address the risk of severe arm swelling post breast cancer surgery and radiotherapy, we are developing an AI tool. Our goal is to assist healthcare professionals and patients in making informed decisions regarding radiation treatment options and mitigating adverse outcomes.”
The research team, comprising experts from six European nations, harnessed data from three European and French datasets (REQUITE, Hypo-G, and CANTO) encompassing 6,361 breast cancer patients to train diverse machine learning algorithms for predicting arm swelling up to three years post-surgery and radiotherapy.
Prior to EBCC14, Dr. Guido Bologna, Associate Professor at the University of Applied Sciences and Arts of Western Switzerland in Geneva and co-analyst on the project, elucidated, “The most recent and effective model leverages 32 distinct patient and treatment parameters, such as the administration of chemotherapy, the performance of sentinel lymph node biopsy in the armpit, and the type of radiotherapy administered.”
Across the three datasets, 6% of patients encountered significant lymphedema. The AI tool accurately predicted lymphedema in 81.6% of cases and correctly identified its absence in patients in 72.9% of cases, resulting in an overall predictive accuracy of 73.4%.
Dr. Rattay highlighted, “Patients identified as high-risk for arm swelling could benefit from additional supportive measures, like wearing an arm compression sleeve during treatment, which has proven to alleviate long-term arm swelling. This information can guide physicians in discussing lymph node irradiation options with patients, particularly in cases where the benefits may be somewhat limited. The forthcoming clinical trial will evaluate the impact of the prediction model on patient and clinician behavior, along with the use of prophylactic arm sleeves.”
The researchers are developing software that will furnish doctors and patients with assessments and forecasts based on the current AI model. This software will undergo testing once the PRE-ACT-01 clinical trial commences later this year. Additionally, efforts are underway to expand the tool’s predictive capabilities to encompass other side effects like skin and heart damage. Genetic marker and imaging data will be collected during the trial to enhance the accuracy of the AI tools, albeit these data will not be utilized for predictions in the PRE-ACT trial.
Dr. Rattay disclosed, “Our target is to enroll approximately 780 patients by early 2026, with a follow-up period of two years.”
Professor Michail Ignatiadis from Institut Jules Bordet in Brussels, Belgium, who presides over the 14th European Breast Cancer Conference, lauded the collaborative effort showcased by the PRE-ACT project in harnessing the potential of AI to facilitate clinicians in predicting and preempting arm lymphedema, thereby enabling them to communicate treatment options effectively to patients.
More information: Abstract no: 23, “Development of an explainable AI prediction model for arm lymphedema following breast cancer surgery and radiotherapy”, Thursday 21 March, Poster in the Spotlight session, 11:00-11:30 hrs CET, Exhibition Hall. https://cm.eortc.org/cmPortal/Searchable/ebcc14/config/Normal/
Provided by the European Organization for Research and Treatment of Cancer
Citation: AI tool predicts risk of side effects after surgery and radiotherapy in breast cancer patients (2024, March 20) retrieved 20 March 2024 from https://medicalxpress.com/news/2024-03-ai-tool-side-effects-surgery.html
This document is subject to copyright. No part may be reproduced without the prior written consent, aside from any fair dealing made for the purpose of private study or research. The content is only provided for informational purposes.
Explore further
A study suggests that a low-dose radiotherapy boost can enhance the cosmetic outcomes for young breast cancer patients by reducing local recurrence rates and achieving superior cosmetic results.
