When 58-year-old Kano Okada was diagnosed with ALS at a clinic in Nagoya, his doctors could offer little certainty about how quickly the disease would advance or which abilities would fade first. Now, thanks to a new AI tool developed at Nagoya University, that uncertainty may soon ease for patients like him. DiSPAH, an artificial intelligence system created by researchers at Nagoya University, has uncovered six distinct patterns of ALS progression and can predict both the speed and sequence of functional decline—using only data from a patient’s first visit.
ALS, a devastating neurodegenerative disease, robs people of movement, speech, and eventually breathing—but not in the same way for everyone. Until now, doctors lacked tools to disentangle how fast the disease moves from the order in which functions deteriorate. DiSPAH changes that. By analyzing data from 2,829 limb-onset ALS patients across two datasets—one with 264 patients used for training, the other with 2,565 for validation—the AI identified six reproducible progression patterns. In some, gross motor skills like walking declined before fine motor skills; in others, the reverse was true. Crucially, the speed of progression and the pattern of decline were found to be independent—meaning a patient could follow a severe pattern slowly or a mild one rapidly, a nuance previous models missed.
The real breakthrough? Prediction from day one. Using just initial clinical assessments and genetic information, DiSPAH can forecast both progression speed and likely pattern early in the disease course. This could transform care: families could plan ahead, clinicians could tailor monitoring, and clinical trials could enroll patients with similar progression profiles, increasing the chances of detecting treatment effects.
One genetic clue stood out. Patients with a mutation in the C9orf72 gene showed faster disease progression. Lab-grown motor neurons from these patients revealed disruptions in protein production and signs of cellular stress—offering a biological explanation for rapid decline and a promising target for future therapies.
While DiSPAH is still a prototype, its creators see broad potential. "It's a promising first step and better than anything that existed before for this specific purpose," says Dr. Yuichiro Yada, associate professor at Nagoya University Graduate School of Medicine. The team aims to refine the tool for all ALS types and eventually adapt it for other neurodegenerative diseases like Alzheimer’s and Parkinson’s. For a disease with no cure and limited treatments, DiSPAH offers something invaluable: clarity in the face of uncertainty.