Huntington's Disease: Decades of Research, No Approved Treatments

The article discusses the failure to develop effective treatments for Huntington's disease despite decades of research and a deep understanding of the genetic mutation that causes it. It highlights the lack of a learning architecture that can effectively share insights across clinical trials.

Details

The article explains that the genetic mutation responsible for Huntington's disease has been known since 1993, and the pathological mechanisms have been studied for decades. Despite this deep understanding, there are still no FDA-approved disease-modifying therapies for Huntington's. The issue is not a lack of biological knowledge, but rather a failure in the way clinical trials are designed and conducted. The article highlights that while extensive natural history studies and clinical trials have generated a wealth of data, this data is not effectively shared or leveraged across trials. As a result, each trial operates in isolation, missing opportunities to learn from the signals in the data. The article uses the example of the GENERATION HD1 trial of the drug tominersen, where a dose-disease-burden interaction signal was present in the data but not surfaced in time to inform the trial. The article argues that a network-based architecture that can route pre-distilled outcome signals from early-enrollment participants to later-enrollment sites could have helped surface this signal earlier. It also explains why federated learning approaches are insufficient to address the challenges posed by the genetic heterogeneity of Huntington's disease.