News Detail

Jakarta-BRIN Public Relations, The use of artificial intelligence (AI) in the health sector continues to grow, including to help diagnose tropical diseases such as malaria, which is still a serious threat in several regions of Indonesia.

In the Media Lounge Discussion (MELODI) activity held on Wednesday (7/5), Researcher at the Research Center  for Artificial Intelligence and Cybersecurity of the National Research and Innovation Agency (BRIN), Anto Satriyo Nugroho, explained the potential of AI technology in accelerating and increasing the accuracy of malaria diagnosis through microscopic image analysis of blood smears.

So far, malaria diagnosis has generally been carried out manually through microscopic examination. This process is not only time-consuming, but also highly dependent on the expertise and accuracy of health workers. Challenges such as fatigue, limited number of experts, and variations in parasite morphology in various stages of its life cycle, often become obstacles in ensuring a correct diagnosis.

Responding to these challenges, BRIN has developed an AI-based diagnostic system to help health workers identify the presence of malaria parasites. This system works by analyzing microscopic photos of thin and thick blood smears to recognize signs of infection.

"Our main goal is to create a Computer Aided Diagnosis system that can automatically recognize malaria status from blood smear images," as explained by Anto Satriyo Nugroho, a researcher with a research background in image processing and biometrics.

In its development, this system uses a dataset containing 1,388 blood smear microphotos collected from various endemic areas in Indonesia. "The research data was obtained from various endemic areas in Indonesia, such as Kalimantan, Papua, and Sumba, in collaboration with the Eijkman Molecular Biology Research Center, BRIN," as explained by Anto, who is now the Head of PR-KAKS BRIN.

The samples include various types of malaria parasites, such as Plasmodium falciparum, P. vivax, P. malariae, and P. ovale, including one case of mixed infection and one negative sample.

"Based on the analysis of 35 micrographs from real cases in endemic areas of Indonesia (covering 3,362 cells), this system shows significant ability in identifying malaria parasites," explained Anto.

In the development of this artificial intelligence (AI)-based malaria diagnostic system, the initial test results showed promising performance.

"The system's sensitivity reached 84,37% in distinguishing healthy and infected cells, with an accuracy value (F1-score) of 80.60% and a positive predictive value (PPV) of 77,14% in identifying parasite species and stages," added Anto. This means that this system is quite reliable in distinguishing healthy blood cells from infected ones.

Opening up Opportunities for Remote Diagnosis

This diagnostic system is also designed to support the implementation of mass blood surveys in the field. "In endemic conditions, one blood smear can require observation of 500 to 1,000 erythrocytes or 200 leukocytes. AI can speed up this process without sacrificing accuracy," said Anto.

In addition to efficiency, this system also opens up opportunities for remote diagnostics so that it is very relevant for use in remote areas. Microscopic knowledge and experience are also stored in the AI ​​system so that it can assist health workers with limited training.

Anto also emphasized that the development of AI in the biomedical field requires special attention to dataset characteristics, data quality, model selection, and appropriate performance evaluation methods. In other words, AI cannot work alone. Collaboration between computing experts and biomedical researchers is an absolute requirement for this technology to be reliable.

"We cannot just rely on computing capabilities. Understanding the medical context is key to making the diagnosis results truly beneficial for patients," he said.

With great potential to improve the accuracy of diagnosis and efficiency of health services in endemic areas, BRIN is optimistic that AI technology will become a strategic partner in national malaria control efforts. BRIN is committed to continuing to improve the system through more comprehensive collaborative research and field trials, as part of efforts to support the malaria elimination program in Indonesia. (mfs/translator:akb)