Artificial intelligence in rare disease diagnosis and treatment

Scientific News

The application of artificial intelligence in healthcare has demonstrated its potential to enhance diagnosis and treatment efficiency. AI can overcome the conventional limitations associated with rare diseases by analyzing vast datasets, identifying patterns, and making predictions for personalized treatment recommendations. This optimizes drug research and development costs and enables early detection, genetic analysis, and personalized therapies for rare diseases. In ‘Artificial intelligence in rare disease diagnosis and treatment,’’ a group led by Magda Wojtara at the Stony Brook University (2023) noted that health care is being revolutionized by artificial intelligence and machine learning, particularly in diagnosing and treating rare diseases. AI and ML algorithms are used for diagnosis, patient recruitment, clinical trial optimization, and precision medicine development. These technologies have shown promise in early detection, genetic analysis, and personalized therapies for rare diseases. Government support and advances in AI have led to the approval of drugs for rare diseases. AI systems have enabled a patient-centered approach to treatment and management, improving end-organ function and therapeutic options. AI can contribute to earlier diagnosis, prognosis improvement, and precision optimization for rare diseases. The authors evaluated 45 studies and ultimately concede that “AI models require accurate and large amounts of data to perform well, but rare diseases often lack sufficient data for effective training. Data augmentation and transfer learning are emerging approaches to address this issue, allowing models to learn from smaller datasets. Research on rare diseases needs to collect more data to improve treatment outcomes.”

Wojtara, M., Rana, E., Rahman, T., Khanna, P., & Singh, H. (2023). Artificial intelligence in rare disease diagnosis and treatment. Clinical and Translational Science, 16(11), 2106–2111. Portico.


Health care is one of several industries being revolutionized by rapidly developing technologies known as artificial intelligence (AI) and machine learning (ML).

From enhancing patient outcomes to optimizing hospital operations, these technologies offer various uses in health care.

It is imperative to develop a tool to assist in early diagnosis, improve treatment effectiveness, and monitor conditions to improve care and reduce costs.

These challenges in the diagnosis and treatment of rare diseases affect more than 350 million people worldwide and create a substantial economic burden on the healthcare system, as well as result in poor patient outcomes.

One such tool is AI, which has, in the literature, shown benefits for common and rare disease diagnoses and treatments.

AI has undergone significant improvements because previous iterations allowed for its possible application for less common diseases, which often have smaller datasets.


Diagnostic decision support systems effectively assist the medical practitioner by providing a list of relevant differential diagnoses.These systems have previously been effectively utilized for various well-known use cases.

It has been harnessed for the early detection and diagnosis of coronavirus disease 2019 (COVID-­19) through monitoring of patients’ demographic, clinical, and epidemiological characteristics. Structural imaging data can be harnessed to determine whether a person with Huntington’s disease (HD) will receive a clinical diagnosis within five years or quantifiable assessments of oculomotor function preceding HD.

These use cases show promising potential for future utilization of AI in rare disease diagnoses. Compared with more traditional biometric methods, AI has shown greater flexibility and scalability, which allows it to contribute to understanding complex relationships, improving early detection, and making routine tasks more efficient.

Types of algorithms and associated benefits for disease diagnosis

Different AI algorithms have appreciable benefits in diagnosing RDs and non-R­ Ds.

The ML helps in diagnosis via three types of algorithms:

  1. Unsupervised, which works by identifying patterns.
  2. They are supervised, which classifies or predicts decisions based on former examples.
  3. Reinforcement learning uses reward and punishment to form a blueprint for operating in a definite obstacle.

ML, for example, has recently been utilized to identify which patients with systemic sclerosis are at a high risk of severe complications, early detection of organ involvement, and more. AI is an integral tool for diagnosing RDs, as it can assist in image recognition, genetic analysis, and clinical decision-making. DL can provide superior recognition with extensive and high-­dimensional data.

Artificial intelligence systems utilizing genetic data

Several AI systems have shown effectiveness in analyzing data to provide accurate diagnoses in phenotypic and genetic analysis.

The Xrare tool has been used to identify causative variants of Mendelian diseases by using similarity scores from phenotypic sets, genetic information from variant databases, and guidelines for variant prioritization.

These tools can perform various important functions, including predicting gene pathogenicity, discovering molecular markers, and building prediction models based on gene expression data. Other AI tools and applications have been developed to aid in the initial diagnosis of RDs. Rare Disease Discovery and Genetic Disease Diagnosis based on Phenotypes are examples.

AI aids in diagnosing rare diseases and their subtypes and provides crucial insights for informed clinical decision-making.


RDs have been neglected by the pharmaceutical industry due to their low and variable incidence, leaving individuals with RDs often with few, expensive treatment options.

Based on previous studies and information from the National Institutes of Health (NIH) Office of Rare Diseases Research and the US Food and Drug Administration (FDA) Office of Orphan Products Development, it is clear that recent government support has led to the approval of many drugs for RDs, thanks in part to advances in AI technology.

The passage of the US Orphan Drug Act in 1983 and the European Union Regulation on Orphan Medicines in 2000 has rewarded innovation in rare disease treatment.

The conventional drug discovery pipeline has significantly hindered the research and development of new drugs to treat RDs. The process is further considered challenging because of the potential for low revenue gains. Still, drug repurposing and establishing interdisciplinary centers for RDs have become trending topics to alleviate this challenge.

Artificial intelligence systems for rare disease treatment development

Two subtypes of AI, ML and DL, have proven helpful in drug development. DL enables the creation of more tailored therapies, whereas ML is useful in clinical trials.

Support Vector Machine is a type of DL algorithm that performs supervised learning, and Random Forest makes output predictions by combining outcomes from a sequence of regression decision trees. These are the most commonly used AI approaches in rare and ultra-r­ diseases, given their ability to handle complicated, high-­dimensional data and images.

Quantitative structure-activity relationship (QSAR) modeling and high-­throughput screening can generate large data sets, which AI can utilize for treatment development.

Computational approaches like QSAR modeling may generate novel treatment compounds with more desired properties. High-t­hroughput screening campaigns can generate large amounts of data, leading to the discovery of drugs such as riluzole for treating amyotrophic lateral sclerosis.

Advances in rare disease treatment development

The development of second-g­eneration AI systems has enabled a patient-­centered approach to treating and managing RDs. The development of second-g­eneration AI systems has enabled a patient-­centered approach to treating and managing RDs.

These systems aim to fill the gaps in diagnostic, prognostic, and therapeutic options by using a tailored closed-l­oop system to improve end-organ function and overcome tolerance or loss of effectiveness issues. The system helps patients at three levels: reminding patients of the dose and administration time and incorporating non-pharmacological therapies, such as physiotherapy.

It includes a closed-l­oop system that adjusts the dose and delivery timing based on the patient’s reaction to therapy and an algorithm that identifies disease-r­ elated patterns by determining variability in laboratory results or clinical parameters.

AI has the potential to play a role in the development of treatments for RDs, enabling a patient-­centric approach tailored to each individual’s specific needs.


As nearly 80% of RDs are genetic, AI has great potential. Thirty-­eight of the 45 studies (84%) published and indexed in PubMed on this particular topic were published from 2020 to 2023.


In this mini-­review, the authors investigated AI and its potential for RD diagnosis and treatment.

For RDs, these advancements have great potential to create a more effective, efficient pipeline for RD studies, drug discovery, and therapeutic fine-­tuning.

At the clinical stage, it can aid in patient recruitment, optimizing RCT, real-w­orld data (RWD) analysis, diagnostic imaging, and developing precision medicine approaches. Quantitative model-b­ased approaches, such as disease progression modeling with AI, advanced statistical approaches in natural history data, and RWD, will play pivotal roles in increasing the efficiency of clinical study design in the drug development process. The PLIER framework developed by researchers utilizes an unsupervised transfer learning framework and may be applied to smaller datasets, such as RDs and precision medicine.

AI is likely to contribute to earlier diagnosis of RDs, which can improve prognosis, advance translational research, and help optimize precision.

Was this article helpful?