One reason that mesothelioma has such a poor survival rate is its late diagnosis and staging: By the time mesothelioma is identified as the cause of a patient’s symptoms, it is often too late for effective treatment. In recent years, various forms of Artificial Intelligence have been used to diagnose and manage the disease. Now, a group of Iranian researchers has reviewed reports of AI’s use in predicting and managing mesothelioma, finding that it holds promise in enhancing early detection and improving patient outcomes.
Artificial Intelligence Used for Diagnosis, Prevention, and Prognosis of Mesothelioma
Writing in the journal Technology in Cancer Research & Treatment, the mesothelioma researchers reviewed the latest research into Artificial Intelligence applications for the management of the rare, asbestos-related disease. With nearly 200 studies identified from database searches, the group chose 19 articles for inclusion in their study. The articles chosen focused on tumor diagnosis and classification, prevention and prognosis, and tumor volumetric measurement of mesothelioma.
Several different Artificial Intelligence models have been used in managing mesothelioma, including neural networks, decision trees, random forests, logistic regression, Naïve Bayes, and support vector machines. The group found that support vector machines, decision trees, and random forests yielded the highest accuracies, ranging from 78.3% to 99.97%.
Researchers Call AI’s Accuracy in Mesothelioma Prediction “Remarkable”
In comparing the accuracy of predictions of mesothelioma among the different artificial intelligence models reviewed, the researchers found that the four times that support vector machines were used delivered 99.97% accurate results, while the random forest model (also used four times) was accurate 93.75% of the time. The decision tree model was used five times, delivering an accuracy rate of 88.36%, and the Naïve Bayes model was used three times, achieving an accuracy rate of 78.3%. Notably, two studies used the logistic regression model and achieved 100% accuracy.
Individual studies reviewed by the researchers noted that AI showed superior predictive abilities for airway disorders over pulmonologists. The researchers also noted the significant expense of the diagnostic procedures, including thoracoscopy and laparoscopy, versus using AI models for prognosis, predicting risk factors, and distinguishing between malignant and benign cell proliferation. The group concluded that AI, and particularly machine learning models, have significant potential in improving mesothelioma prediction and management.
Innovative mesothelioma research and tools provide real hope for people at risk for mesothelioma. If you’d like more information on the resources available to you, contact the Patient Advocates at Mesothelioma.net today at 1-800-692-8608.
