Diagnosing Rare Diseases In The Digital Age
By Dekel Gelbman, CEO, FDNA
The digital age has transformed how nearly every industry conducts business, and healthcare is no exception. Technological advances often outpace our ability to apply them, but successful integration can mean new efficiencies and improved clinical outcomes for patients. Recent developments have proven Big Data holds the key to powerful diagnostics, and nowhere is this more evident than in the application of that data in identifying rare diseases in patients. The effective use of this data makes a real difference to patients who are seeking answers, meaning every attempt by healthcare to keep up with technological advancements must be made.
The Challenge With Rare Diseases
The diagnostic odyssey of a patient with a rare disease is unmatched in terms of time. It takes an average of over seven years and seven physicians for patients to finally begin receiving answers. For patients and their families, this means years of anguish, waiting, and considerable expense before a diagnosis is ultimately achieved. For geneticists, it is a frustrating process trying to assemble a puzzle with seemingly unrelated pieces. With over 7,000 known rare diseases, and tens-of-thousands of signs and symptoms, diagnosing is a daunting task.
This is where digital technologies are making an unprecedented difference. A patient’s signs and symptoms, along with other clinical phenotypes such as broad neck or wide set eyes, can be used to reveal correlations with rare diseases and disease-causing genetics. With recent advancements in vision science and deep learning technologies, it has become easier than ever to detect these patterns and correlate them with syndromes and genes.
New approaches to structuring phenotype information for use by diagnostic technologies has made this all possible. Human Phenotype Ontology (HPO), a structured and controlled vocabulary for phenotypic features, is now used broadly in the rare-disease space. This has opened the way for research, as thousands of studies now show correlations between HPO phenotypes and diseases. With all this structured data, technologies can now capture phenotypes using HPO and compare the data to known syndromes and genes to highlight likely causes of the symptoms. In the rare disease setting, a suite of phenotyping apps called Face2Gene does just that. By highlighting phenotypic similarities through facial analysis technology and capturing, analyzing and warehousing phenotypic information using HPO, Face2Gene provides a list of the most clinically relevant syndromes and genes to consider. This information can help the clinician make a clinical diagnosis or filter the molecular variants to discover the most likely disease-causing genetics.
Applying New Technologies To Existing Ones
For healthcare to benefit from the advantages made available through genomics and phenotyping, institutions must proactively find ways to accommodate the data and technologies that can use the data. Because most EMRs are not designed to support structured genomic and phenotypic information, new systems need to be introduced that can link into existing EMRs and workflows. Evaluating those technologies should focus on how they were designed, and if they inherently work around the existing workflow of clinicians with minimal or no disruptions. Implementing data warehouses that can manage the data and provide the analysis, yet still integrate with existing EMR systems, can solve the technological hurdles, enabling this new data to bring value to clinicians and labs.
The Promise Of Big Data And Phenotyping
The evolution of big data and digital technology in healthcare is in full swing, giving immense promise to transform the work of clinicians, and most importantly the lives of the patients they treat. Navigating the challenges and opportunities of adopting these technologies will define the next generation of leaders in clinical and molecular diagnostics while paving the way to a diagnosis for patients that are still waiting for answers.