Machine Learning Protocol to Improve Clinical Diagnosis, Prognosis and Disease Subtyping for Spectrum Disorders

Case ID:

This innovation is a proposed machine learning-based software designed to improve diagnosis, prognosis, and clinical management of rare diseases, a category encompassing over 10,000 conditions with new additions each year. Currently, the field grapples with a scarcity of effective solutions, with only 5% of rare diseases having established therapeutic options. This innovative approach would leverage individual genotypic, phenotypic, and medical data to unravel the complexities of rare diseases, classify patients into distinct endotypes, and identify predictive risk factors for tailored therapeutic interventions. This software has the potential to revolutionize the diagnosis, prognosis, and clinical management of rare diseases with intricate and diverse phenotypic profiles, offering new hope for patients and healthcare providers alike.

The realm of rare diseases presents a complex and challenging landscape, with over 10,000 distinct conditions and 300 new diseases emerging annually, of which 2,227 remain enigmatic. Effective treatment strategies are available for only 5% of these rare diseases, emphasizing the critical need for innovation. One major hurdle is the significant variability in symptoms among individuals, making diagnosis, prognosis, and clinical management intricate. Moreover, these conditions often do not fit into conventional mild-to-severe classifications, showcasing a diverse spectrum of endotypes. For instance, a rare ailment linked to pathogenic SCN8A gene variants exhibits a multifaceted phenotype with developmental delay, seizures, cognitive impairments, and comorbidities. The extensive variation in symptoms among patients further complicates diagnosis and treatment. Existing methodologies often struggle to accommodate the diverse endotypes exhibited by many rare diseases. 


  • Medical diagnostics
  • Personalized treatment strategies for rare diseases


  • Improved accuracy
  • Efficient identification and categorization of diverse endotypes
  • May identify predictive risk factors
Patent Information:
Contact For More Information:
Lewis Humphreys
Licensing Manager, Eller College of Mngmt & OTT
The University of Arizona
Lead Inventor(s):
Michael Hammer
Joshua Hack
Joseph Watkins