Athens, Greece – New research has unveiled the development of MaLeSQs, a promising machine learning tool aimed at enhancing the screening process for leprosy, which continues to be underdiagnosed across the globe. The study, conducted by experts at the National Referral Center for Sanitary Dermatology and Hansen’s Disease, utilizes the Leprosy Suspicion Questionnaire (LSQ) as its foundation for identifying new cases of this chronic illness.
Leprosy, caused primarily by the bacteria Mycobacterium leprae, can lead to severe nerve damage if not diagnosed and treated early. Unfortunately, it often mimics other dermatological and neurological conditions, leading to missed diagnoses by healthcare professionals who may lack training. The World Health Organization (WHO) has identified proactive case detection as one of the pillars of its strategy to eradicate leprosy.
The LSQ is composed of 14 straightforward yes/no questions, developed as both a screening tool and educational resource for communities. The findings of the study revealed noteworthy correlations: "The combination of marked questions was related to new case detections," underscoring the importance of specific symptoms associated with the disease.
With the goal of refining the screening process, the research team developed the MaLeSQs tool, leveraging machine learning algorithms to analyze the responses gathered using the LSQ. The proactive approach aims not only to indicate potential cases but also to educate individuals on the symptoms associated with leprosy. To reach optimal predictive power, four classifiers—Support Vectors Machine, Logistic Regression, Random Forest, and XGBoost—were initially tested.
After extensive validation, Support Vectors Machine emerged as the top-performing model, achieving 85.7% sensitivity and 98.3% negative predicted value. These results demonstrate its ability to effectively classify individuals as either likely to have or not have leprosy.
This breakthrough could significantly streamline the process for identifying leprosy cases, especially in regions where healthcare access is limited. By employing algorithms trained on responses from over 1,800 participants, the tool provides both accuracy and efficiency for healthcare practitioners.
Investing resources to train machine learning algorithms can help address the persistent challenges faced by healthcare systems combating leprosy. The researchers noted, "Even though healthy people might be alarmed by a false positive, these will look for health assistance," emphasizing the importance of addressing any potential drawbacks to maximize public health benefits.
The research concludes with the call for broader implementation of such tools across various healthcare settings. While the LSQ already serves as a valuable resource for identifying high-risk individuals, integrating machine learning capabilities will bolster efforts to screen for this often-overlooked disease effectively.
Moving forward, studies will focus on validating MaLeSQs against diverse populations to affirm the tool's applicability and reliability across different geographies and healthcare frameworks. The hope is to transform leprosy screening efforts worldwide, paving the way toward eventual eradication of the disease.