Artificial intelligence in parasitic disease control: A paradigm shift in health care

Abstract

Parasitic diseases, including malaria, leishmaniasis, and trypanosomiasis, continue to plague populations worldwide, particularly in resource-limited settings and disproportionately affecting vulnerable populations. It has limited the use of conventional health-care delivery and disease control approaches and necessitated exploring innovative strategies. In this direction, artificial intelligence (AI) has emerged as a transformative tool with immense promise in parasitic disease control, offering the potential for enhanced diagnostics, precision drug discovery, predictive modeling, and personalized treatment. Predictive AI algorithms have assisted in understanding parasite transmission patterns and outbreaks by analyzing vast amounts of epidemiological data, environmental factors, and population demographics. This has strengthened public health interventions, resource allocation, and outbreak preparedness strategies, enabling proactive measures to mitigate disease spread. In diagnostics, AI-enabled accurate and rapid identification of parasites by analyzing microscopic images. This capability is particularly valuable in remote regions with limited access to diagnostic facilities. AI-driven computational methods have also assisted in drug discovery for parasitic diseases by identifying novel drug targets and predicting the efficacy and safety of potential drug candidates. This approach has streamlined drug development, leading to more effective and targeted therapies. This article reviews these current developments and their transformative impacts on the health-care sector. It also assessed the hurdles that require attention before these transformations can be realized in real-life scenarios.

INTRODUCTION

Artificial intelligence (AI) combines the excellence of science and engineering to create intelligent machines, mostly computer programs. Although the concept first appeared in the 1950s, it was the early 2020s that witnessed the AI spring. Unprecedented advances in computation and development in machine learning (ML) modules, including deep learning (DL), infrastructure, and parallel excellence in human resources, have led us to the middle of the AI revolution. Today, AI has almost reached every sphere of human life and is poised to transform the health-care sector.[1] Along with human intelligence, AI is enabling a cohesive and integrative health-care ecosystem where the latter is largely assistive in improving the diagnosis, treatment of patient monitoring, etc. Fuelled by the demand, the AI-based health-care market stood at USD 9.64 billion in 2022 and is expanding at a CAGR of 51.87%.[2] In India, although in its infancy, the AI market is expected to reach USD 11.78 billion by 2025.[3]

The health-care sector has always prioritized the response and management of infectious diseases.[4] In this, humanity’s exposure to and experiences from outbreaks and pandemics, including the recent COVID-19 pandemic, has stressed the need for a more resilient health-care ecosystem. In recent times, it has been more relevant for the growing threats of emerging and reemerging infections, where the inevitability of frequent outbreaks and pandemics marks the future.[5] Among all these agents of biological threats, parasitic infections hold a higher position. Earlier considered a challenge for low-resource nations, today’s parasites have adapted to climatic changes.[6] Climate change events have altered vector transmissibility and facilitated zoonotic parasites’ emergence with higher transmissions in newer geographies to pose a more significant public health threat. As such, more responsive and resilient health-care innovations and interventions are timely for these global biothreats. AI has brought multiple benefits, from disease forecasting, drug discovery and development and diagnostics.

ARTIFICIAL INTELLIGENCE IN PREDICTING PARASITIC DISEASE OUTBREAKS

Predictive disease modeling is increasingly transforming the approach to outbreak preparedness and response. It is emerging as the centerpiece of public health management attention.[7] By harnessing the power of AI and data analysis, predictive models have identified patterns and trends in disease incidence, enabling timely interventions to mitigate disease spread and protect public health. It facilitated public health officials to take valuable time to make informed decisions, allocate resources effectively, and implement targeted interventions before an outbreak escalates into a widespread crisis.[8] Such models have been tested and employed in various settings to predict disease outbreaks, demonstrating their effectiveness in real-world scenarios. For example, Liu et al. utilized Google Search data to predict influenza epidemics in Hong Kong.[9] The model had a remarkable accuracy of up to 70%, which provided valuable lead time for public health interventions. Smilarly, the case for predicting dengue fever outbreaks in Yucatan, Mexico, and San Juan, Puerto Rico;[10] Zika virus transmission in Africa, and the Asia–Pacific.[11]

Predictive AI-based modeling was also used for parasitic infection outbreak forecasting. A convolutional neural network (CNN) algorithm was trained with 2013–2017 data for three vector-borne diseases, i.e., chikungunya, malaria, and dengue.[12] The trained model predicted disease outbreaks with 88% accuracy. Several other predictive algorithms were developed for forecasting malaria outbreaks.[13] These models factored atmospheric data, recorded cases over the years, and effectively forecasted future incidences. Similarly, Shabanpour et al. used geospatial AI that integrated ML algorithms with geographic information system (GIS)-based approaches for mapping cutaneous leishmaniasis.[14] Isfahan province’s Northern and central areas were predicted to have high disease risk.

ARTIFICIAL INTELLIGENCE-ASSISTED PARASITIC DRUG DISCOVERY

The traditional drug discovery process is extremely lengthy. It often spans a decade or more, from initial target identification to market approval for numerous steps, including target validation, lead compound identification, optimization, preclinical testing, and clinical trials.[15] The process primarily relies on empirical approaches, often lacking predictive models that can accurately assess the likelihood of a drug candidate’s success. This lack of predictive tools leads to a high failure rate, with an estimated 90% of potential drug candidates failing to progress beyond preclinical testing.[16] This high failure rate is due to various factors, including poor target selection, inadequate efficacy, unacceptable toxicity, and unfavorable pharmacokinetic properties.[17] This prolongs the drug discovery process and incurs significant financial losses for pharmaceutical companies.[18] As such, a pressing need exists for escalating drug discovery efforts to address the persistent global health challenges, especially from increasing episodes of emerging and reemerging infectious diseases.

AI has emerged as a transformative force in this context.[19] By harnessing the power of AI algorithms, AI-driven computational methods analyze vast amounts of data, including genomic, proteomic, and structural information, to unravel the complex mechanisms underlying parasitic diseases and identify potential therapeutic targets. Through this, AI can pinpoint critical proteins or enzymes that play essential roles in pathogen survival and replication. These enabled researchers to identify novel drug targets, predict the efficacy and safety of potential drug candidates, optimize the design of new therapies, and even repurpose existing drugs in a much smaller time frame.[20] This can lower the attrition rate and escalate the entire developmental pipeline from decades to months.[21]

Such AI-assisted technologies have been fairly used for antiparasitic drug discoveries. LabMol-167 was identified as a new potential PK7 inhibitor with in vitro antiplasmodial activity. The authors used an AI-assisted virtual screening along with shape-based and machine-learning models. The LabMol-167 exhibited low cytotoxicity in mammalian cells yet inhibited Plasmodium falciparum at nanomolar concentrations.[22] In another case, DeepMalaria, a Graph CNN-based DL process, was developed to find potential antimalarial compounds.[23] The model was trained using the GlaxoSmithKline dataset. It identified potential compounds wherein more than 85% of the compound showed parasite inhibition with 50% and more effectiveness. DC-9237 was identified as the most promising drug candidate against malaria for its fast-acting nature. Companies like Novartis used an ML-based profile-quantitative structure-activity relationship (pQSAR) platform for screening potential drug candidates against malaria.[24] The model ended with a potential compound library after training with blood-stage P. falciparum 3D7 data. The compounds showed desirable pharmacological properties and novelty as potential antimalarial drugs. The pQSAR and other ML platforms are now used to screen drugs for multiple parasites.[25]

The usefulness of AI was also realized in developing medical countermeasures against trypanosomiasis. Leonardi et al. used a neural network-based ML model to optimize the oral absorption of benznidazole, a drug of choice against American trypanosomiasis.[26] The model collectively analyzed multiple process parameters while proposing a perfect strategy for producing benznidazole chitosan microparticles for the most effective outcome. In search of potential drug targets, AI-based DeepMind Technologies was used to predict the target protein structures in Trypanosoma.[27] This discovery has paved the way for developing more effective trypanosomiasis treatments. More and more evidence is emerging on AI-based integration of existing genomics and chemical datasets to prioritize target and focus areas for combating trypanosomes.[28] It is largely expanding drug discovery pipelines.

Finally, AI is increasingly being used for drug repurposing. Repurposing that leverages existing drugs, already approved for other diseases, to tackle new medical challenges offers many benefits for patients, the health-care system, and the pharmaceutical industry. Extensive repurposing efforts can be seen across several emerging infections, including the COVID-19 pandemic. Similar attempts can also be seen in the case of treatment for parasitic infections such as malaria, Chagas, African sleeping sickness, and schistosomiasis.[29] In this direction, Williams et al. developed “Eve,” which uses AI to economically perform drug repurposing by integrating a process pipeline consisting of library screening, hit confirmation, and lead generation.[30] Eve identified that the antimicrobial compound fumagillin has the potential to inhibit the growth of P. falciparum strains. The drug, when tested in a mouse model, was able to inhibit parasitemia.

ARTIFICIAL INTELLIGENCE-BASED DIAGNOSTICS

The burden of parasitic infections is as old as humanity. Even today, with all scientific and technological advances, health care across the nation dwells with multiple such infections. In these, global warming is facilitating the geographical spread of parasites, especially those that are vector borne.[31] The health security stress from these emerging parasitic infections is more for resource-poor settings where traditional diagnostics are either resource intensive, unsuitable for field deployment, or lack sensitivity.

Exactly here, AI, with its power of ML and DL algorithm, is revolutionizing parasitic diagnostics, offering faster, more accurate, and accessible solutions. These algorithms are trained on large datasets of parasite images, enabling them to identify specific morphological features and patterns with remarkable accuracy. Our previous report consolidated the collective benefits of AI-empowered Deep Technologies over the microscopic, serological, and molecular methods.[32] We also discussed the availability and importance of public databases to construct AI modules for parasite diagnostics.

AI has shown tremendous benefits in processing a large number of images for blood smears, stool samples, and tissue biopsies. Using ML tools such as CNNs, the AI has perfectly identified and classified parasitic stages such as eggs, larvae, and adult worms. The diagnostic power of AI has been well explored in detecting different intestinal parasites from stool samples, such as hookworms, roundworms, and whipworms. Similarly, malaria parasites, blood parasites (African sleeping sickness and leishmaniasis), and schistosomiasis were also detected. For example, the development of AIDMAN, an AI-based system for malaria parasite detection in blood smears captured through smartphones, exemplifies the transformative potential of AI in parasitic diagnostics.[33] It had 98.62% and 97% diagnostic accuracy for cells and blood-smear images, respectively. The clinical validation accuracy was 98.44%, comparable to highly experienced microscopists.

Not only does AI offer the benefits of high sensitivity, but it also offers a limit of detection much lower than conventional strategies. It also significantly reduces human error and workload, especially in high-burden areas. Most importantly, these AI-powered tools can be incorporated into smartphone-based applications, making parasite diagnostics readily available even in remote areas with limited health-care infrastructure.[32,33] These benefits warrant its scalability and cost-effectiveness. It paves the way for rapid technology adoption and commercialization by automating labor-intensive tasks and minimizing the need for specialized expertise.

CONCLUSION AND WAY FORWARD

Humanity’s future lies in health-care’s success in responding effectively against emerging public health challenges. In the present situation, the health-care landscape is burdened by several challenges, such as rising costs, limited access, clinician burnout, and inefficient resource allocation. The problem of timely detection and limited medical countermeasures adds to the burden during pathogen management. Fortunately, AI is emerging as a powerful tool with the potential to tackle these issues head-on. Although AI is emerging and mostly centered on ML subsets, the extent of R and D and its related pace indicates a revolution across disease forecasting, drug discovery, diagnosis, treatment, and ultimately patient experience [Figure 1].

Figure 1.

Potentials and challenges in artificial intelligence implementation for parasitic disease control. ML: Machine learning

Despite its transformative potential, the implementation of AI in parasitic disease control faces challenges where data scarcity, particularly in endemic regions, has hindered the development and validation of AI models. Such “data poverty” primarily impacts the ability to implement an unbiased AI/ML model. It delays or something inserts errors while developing accurate diagnostic tools, treatment algorithms, and predictive models for parasite outbreaks. This lack of investment in data collection and AI development perpetuates the cycle of neglect, leaving millions suffering from preventable illnesses. The WHO, in its recent roadmap for neglected tropical diseases, noted data limitations on neglected tropical diseases like parasitic infections. The report recommended a more extensive integration of AI and other new technologies such as digital health, satellite imagery, and drones in the innovation landscape to counter these diseases.[34] In this direction, collaborative efforts are crucial. Data-sharing initiatives between endemic regions and research institutions can pool scarce resources and create robust datasets. Crowdsourcing is a type of collaboration that has shown great promise in fast-tracking developments. Lin et al. demonstrated the benefits of training an AI model using crowdsourced annotated medical images to detect soil-transmitted helminthiasis in microscopy images.[35] As such, investment in building data infrastructure will be necessary to ensure widespread AI developments.

However, any AI development must adhere to proper ethics and ensure equitable access to AI-powered tools in resource-limited settings. Global deliberations on ethics in AI are on the rise. UNESCO’s first global standard was issued in 2021 and adopted by 193 member states.[36] The WHO also prepared a separate guidance focusing on AI in health care. The guidance identified challenges and risks for the indiscriminate use of AI. It provided six sets of recommendations for good AI governance.

In contrast, equitable access to AI-powered tools in resource-limited settings is yet to be fully realized.[37] There is a need to include low- and middle-income countries by the advanced nations to create the global AI ecosystem. In this, the requirements of local communities, including societal and cultural factors, must be prioritized for a more extensive distribution of AI’s benefits.[38] Addressing these challenges will be essential to harness AI’s full potential in alleviating the burden of parasitic diseases worldwide.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.