Blood test using artificial intelligence paves the way for early diagnosis of leprosy

A new blood test combined with a standardized questionnaire and artificial intelligence may transform the way leprosy is diagnosed in Brazil. Researchers at the University of São Paulo (USP) tested the new test in Brazil using blood samples collected during a COVID-19 population survey. The method showed potential for identifying the disease earlier, in its initial stages, when symptoms are subtle and traditional laboratory tests often fail.

The new diagnostic method was evaluated in a study conducted by researchers from the Department of Clinical Medicine, Biochemistry, Immunology, and Social Medicine at the Ribeirão Preto School of Medicine of the University of São Paulo (FMRP-USP), with support from FAPESP. Coordinated by researcher Marco Andrey Frade, the study was published in the journal BMC Infectious Diseases.

“Leprosy is an ancient disease, but it still faces challenges typical of health issues that aren’t prioritized. There’s still a lack of sensitive laboratory technologies for early diagnosis, and many healthcare professionals aren’t adequately prepared to recognize the early stages of the disease,” explains Filipe Lima, a biomedical scientist and one of the authors of the study. Furthermore, the standard treatment has essentially remained the same for over four decades, contributing to cases of treatment failure and bacterial resistance.

To address this issue, the researchers set out to identify new biomarkers and tests for early diagnosis. To do so, they used blood samples collected during a serological survey conducted in Ribeirão Preto during the COVID-19 pandemic. The goal was to use this existing material to identify individuals who may have been exposed to the leprosy bacillus and detect new cases earlier.

Screening

The study combined two screening tools. The first tool was a clinical questionnaire called the Leprosy Suspicion Questionnaire (LSQ). The LSQ consists of 14 questions focused primarily on neurological signs and symptoms. The questionnaire was enhanced with an artificial intelligence system called MaLeSQs.

The second tool was a blood test that detects antibodies to the Mce1A antigen. This antigen is a key protein of Mycobacterium leprae that facilitates invasion and survival of the bacterium in human cells. Currently, blood tests use the PGL-I antigen, which also facilitates the entry of the bacterium into the nerve but is technically less sensitive.

“Unlike the traditional [anti-PGL-I] test, which assesses the presence of only one type of antibody, the new [anti-Mce1A] test analyzes three different classes of antibodies (IgA, IgM, and IgG), which increases sensitivity and helps differentiate between exposure to the bacillus, active infection, and prior contact,” Lima explains. According to Lima, the traditional test usually only yields a positive result in the most severe forms of the disease, when the bacillus has already proliferated and lesions are present. “Mce1A allows for the identification of contact with the bacillus and active disease at a much earlier stage,” he explains.

Invitation, questionnaire, and test

To obtain the results, the researchers invited approximately 700 people from the COVID-19 population survey to participate in the leprosy study. Of those, 224 agreed to take part and completed the digital questionnaire, and 195 had their blood samples analyzed. All participants were invited to undergo an in-person clinical evaluation with specialist physicians, which is a crucial step for diagnostic confirmation.

Of those individuals, 37 attended the consultation. Cross-referencing the data from the questionnaire, laboratory tests, and clinical evaluation yielded striking results: 12 new cases of leprosy were diagnosed, equivalent to about one-third of those evaluated. “These are people who had no obvious symptoms, didn’t suspect they were sick, and were diagnosed thanks to the project,” Lima notes.

According to the researchers, the IgM antibody test for the Mce1A antigen was the most effective of the laboratory tests, identifying two-thirds of the newly confirmed cases. Combining laboratory analysis with the artificial intelligence tool achieved 100% sensitivity, successfully flagging all suspected cases of leprosy that were confirmed during the in-person consultation.

“The blood test doesn’t confirm a leprosy diagnosis on its own, but it’s an important tool for identifying who truly needs to be evaluated by a specialist,” the researcher explains. According to Lima, the test can strengthen diagnostic screening in the public health system. In terms of cost, the difference compared to existing tests is minimal. “From a laboratory standpoint, these are very similar techniques – low-cost and easy to perform. Any clinical laboratory has the technical capacity to carry them out. In practice, the only thing that changes is the molecule being analyzed.”

In addition to facilitating the early diagnosis of leprosy, the study used georeferenced maps to analyze the spatial distribution of identified cases. The mapping revealed a diffuse pattern of exposure to the bacillus. “This can be explained by the fact that we were unable to clinically evaluate all participants. But our results show that leprosy is randomly distributed throughout the city. There’s no specific region with a higher concentration. Today, we see the disease diagnosed in patients with different socioeconomic profiles,” says Lima.

Public health problem

Leprosy is an infectious disease that primarily affects the skin and peripheral nerves. It can cause pale or reddish patches, loss of sensation, and muscle weakness. According to the World Health Organization (WHO), more than 200,000 new cases are reported each year, primarily in low- and middle-income countries.

Brazil ranks second globally in the number of cases, behind only India, accounting for approximately 90% of reported cases in the Americas. The most common symptoms include tingling, cramps, numbness, and areas of the skin with reduced sensitivity. During this initial phase, common laboratory tests, such as smear tests, which attempt to identify the bacteria directly on the skin, often produce negative results because the bacterial load is still very low. “More than 60% of our patients may test negative, even though they’re sick,” says Lima.

Treatment involves taking antibiotics for six months to a year, depending on the stage of the disease. Currently, the disease is classified as a socially determined disease (SDD), a term replacing “neglected disease” in Brazil.

The next step is to advance the validation of these tools for large-scale use with the goal of incorporating them into Brazil’s national public health system, the SUS (short for “Sistema Único de Saúde”). Meanwhile, the researcher is working on a new phase of the project aimed at increasing the specificity of the Mce1A marker. “Currently, the test uses the entire bacterial protein. Now, we’re studying small parts of that same protein to assess the possibility of developing an even more sensitive and accurate test,” he concludes.

Reference:

Lima, F.R., Simões, M.M.R., Vitiritti, B. et al. Serum geoepidemiology of leprosy biomarkers in a city-wide COVID-19 survey in Brazil. BMC Infect Dis 26, 258 (2026). https://doi.org/10.1186/s12879-025-12483-0