To effectively diagnose and treat rare diseases, clinicians and researchers rely on genetic and diagnostic tests and phenotypic data to inform their decisions.
For example, phenotypic data commonly reported by patients, including quantified observable characteristics such as short stature, low-set ears, and blood biochemistry, are often insufficient to yield a definitive diagnosis. However, combined with additional genetic data, phenotypic data has the potential to unlock life-changing diagnoses for patients with rare cancers, pediatric diseases, inherited genetic syndromes and other conditions.
While the proliferation of testing and the abundance of data continues to expand the rare disease knowledge base, it also presents challenges for clinicians and researchers seeking precise information to answer key questions. To reduce time-consuming manual searches and improve their discoveries, many organizations are turning to artificial intelligence techniques such as natural language processing (NLP). This enables organizations to assess and normalize more phenotypic and test data faster and more accurately.
Understanding the challenges of rare disease diagnosis
A rare disease is defined as any disease, disorder, disease or condition that affects fewer than 200,000 people in the United States. Rare Action Network. An estimated 25 to 30 million Americans, or nearly 1 in 10 Americans, have at least one of the approximately 7,000 identified rare diseases. There are more than 500 rare cancers, and all pediatric cancers are considered rare.
In order to start treatment for patients with rare diseases as soon as possible, rapid diagnosis is critical – that’s where NLP comes in. NLP automatically mines complex unstructured data so it can be transformed into curated, well-structured data to inform research and analysis. NLP is able to quickly read, understand, and translate the nuances contained in medical documents, including comments in electronic health record systems and free-form text in test results sections.
In the case of rare diseases, NLP can convert free text into Human Phenotype Ontology (HPO) terms to capture phenotypic data created when patients are referred for testing. Clinicians can then use it to better understand the results of genetic testing based on any phenotypic presentation. The presence of a genetic marker does not ensure that the disease itself will emerge now or in the future.
NLP in action
Clinical and research organizations are adopting NLP to derive unstructured information from EHRs to improve diagnosis and identification. Below are two examples of real-world applications of NLP.
Identify patients with heart disease
Researchers at a large California medical institution sought to understand the prevalence of aortic valve stenosis in its patient population. Because health systems often use procedures or billing codes that lack clinical nuance, researchers can have difficulty accurately identifying complex clinical conditions in patient populations.
To overcome this challenge, the researchers used NLP to comb through more than a million patient records and echocardiogram reports to identify certain acronyms, words, and phrases associated with severe aortic stenosis. In just a few minutes, the technology identified 54,000 patients with the disease, a feat that would have required researchers to use traditional manual search methods.
the organization’s success Identifying previously unrecognized patients with complex diseases shows the promise of NLP to effectively identify other diseases, allowing clinicians to more effectively manage the health of patients with rare diseases.
Advancing Personalized Medicine Research
Personalized medicine considers a patient’s response to treatment based on the patient’s specific phenotypic and genetic characteristics. However, clinicians and researchers seeking to advance personalized medicine have difficulty finding the phenotypic information required for personalization because details are often hidden in EHRs in unstructured formats.
To gain a more complete understanding of individual chronic disease patients, researchers at a medical school lever NLP extracts unstructured information about patients with diseases such as Alzheimer’s disease, breast cancer, lung cancer, diabetes, and obesity. Automatic discovery of key patient information helps the organization advance its biomedical practice, such as identifying patterns in data and developing predictive analytics to determine patient outcomes.
Because rare diseases are by definition uncommon, patients with these complex and unusual conditions are often misdiagnosed and undertreated. Thanks to advances in artificial intelligence-based technologies such as NLP, researchers and clinicians are now able to rapidly gain new insights into rare diseases, offering hope for faster, more accurate diagnosis and personalized treatment.
Photo: Peach_iStock, Getty Images
