Back to all blog posts
Infectious diseases come from microorganisms, such as bacteria, viruses, fungi, or parasites. The pathogens are transmitted, directly or indirectly, and can lead to epidemics or even pandemics, as seen with coronavirus. While some infected individuals remain asymptomatic, many infectious diseases have severe consequences and can be fatal. Infectious diseases are a leading cause of death worldwide, especially in low-income countries. Over time, scientists have received access to better tools to predict epidemics, understand the specificity of each pathogen, and identify potential targets for drug development. Artificial intelligence is an extremely useful tool for understanding the pathophysiology of infectious diseases and how to treat them.
The power of AI is particularly valuable in image analysis, where classical tools can not identify early signs of disease. AI can help improve the accuracy of pathological analysis, thus improving diagnosis and treatment. This could potentially limit the transmission of infectious diseases and the development of future epidemics and pandemics.
This article collects together case studies related to hantavirus, malaria, COVID-19, and tuberculosis.
Hantavirus
Hantaviruses are human pathogens that can cause severe disease, typically infecting the lungs or kidneys. A research group at the University of Helsinki studied a group of puumalavirus-infected (a species of hantavirus) patients with acute hemorrhagic fever with renal syndrome (HFRS). Stained kidney biopsy specimens from acute HFRS and unrelated kidney diseases as controls were scanned. The team used Aiforia for the quantification of HLA-DR, CD14, CD16, and CD68 positive cells as compared to a total number of cells. The AI training process included ~2500 iterations based on ~500 annotations. They found increased numbers of cells expressing monocyte and macrophage markers in the kidneys of patients with HFRS.
Learn more about related publications:
- Vangeti, S. et al. (2021, March 10). Monocyte subset redistribution from blood to kidneys in patients with Puumala virus caused hemorrhagic fever with renal syndrome. PLOS Pathogens 17(8). https://doi.org/10.1371/journal.ppat.1009400
- Hepojoki, J. et al. (2021, August 11). Hantavirus infection-induced B cell activation elevates free light chains levels in circulation. PLOS Pathogens. https://doi.org/10.1371/journal.ppat.1009843
Malaria
Malaria is a life-threatening disease caused by Plasmodium parasites and transmitted through the bites of infected mosquitoes. In 2019, nearly half of the world's population was at risk of malaria, with an estimated 229 million malaria cases globally. Detecting malaria parasites in blood smear can be time-consuming, laborious, and error-prone with conventional microscopy. Current diagnostic methods are not enough for the large number of malaria cases in the world, particularly in developing countries. Deep learning AI-based methods have the potential to automate the process, aiding lab technicians in optimizing their time and increasing accuracy and efficiency.
AI models can be trained using annotated datasets to find parasites in blood smears, such as Aiforia's AI software demonstrates in the video below. Through simple annotation of infected cells to train the algorithm, the automation process includes quantifying parasite density faster and more accurately than human-based microscopy. This reduces turnaround time, improves diagnostic performance, and provides consistent and accurate care to patients in a timely manner. Read more about AI-assisted analysis in malaria detection.
Aiforia's AI software is able to detect and quantify malaria parasites in blood samples
Related publication: Holmström, O. et al. (2020, November 17). A novel deep learning-based point-of-care diagnostic method for detecting Plasmodium falciparum with fluorescence digital microscopy. Plos One. https://doi.org/10.1371/journal.pone.0242355
COVID-19
Dr. Robert Mozayeni from the Foundation for the Study of Inflammatory Disease (FSID) in the United States has used Aiforia’s AI models to analyze high-resolution images of blood and skin samples stained for COVID-19 RNA. Inflammatory markers were collected from hundreds of patients, and to put those results in the context of other testing, clinical and medical history data were collected about those individuals to understand better the mechanisms and impact of this disease. Using AI in this process is novel because AI can detect differences that are not obvious to humans.
AI has the potential to increase the throughput of specimens to augment the human technologist’s visual analysis of thousands of microscope slides while reducing the inherent variabilities in the process. Finding COVID-19 and related inflammatory responses anywhere in the body will contribute to the understanding of this virus. Using AI to analyze images of either the novel coronavirus RNA in blood or any other tissues, as well as potentially co-locating such viral RNA along with markers for immune response, would be a very significant contribution to our understanding of the virus’ behavior, pathogenicity, and potential treatment interventions. Read more about this project here.
Tuberculosis
Mycobacterium tuberculosis, the pathogen causing tuberculosis (TB), can persist in hosts for years, evading their immune responses through many mechanisms. One such tactic is deregulating lipid metabolism, resulting in the formation of foamy macrophages. These altered immune cells are, therefore, a significant indicator of the presence of TB in a patient.
Gillian Beamer, pathologist and assistant professor at Tufts University, has been studying experimental TB for 15 years. “Part of my work as a scientist and pathologist is to rank, score, grade, count, or otherwise quantify visual features. Typically, this means evaluation of tissue sections using a microscope and identification of differences amongst treatment groups,” Dr. Beamer describes. This work is manual and very time-consuming. “For 15 years, I have been looking to train a computer to automatically recognize and count visual information,” she explains.
Aiforia was deployed to create deep learning AI models to recognize and quantify foamy macrophage areas and count the cells within these identified spaces. Only 25 slides were needed for training with Aiforia® Create. Samples from over 400 lungs of M. tuberculosis-infected mice were then scanned and analyzed with the AI tool. Read more about Gillian’s experience working with Aiforia in her tuberculosis research.