Chronic, progressive neurologic disorders such as Alzheimer’s and Parkinson’s Disease affect millions of people worldwide. They have a profound impact on the lives of affected patients and their families.

There are approximately 30 million patients affected worldwide with Alzheimer’s. It is the underlying cause in most of cases of dementia and most commonly found in those over the age of 65. The underlying brain pathology involves the buildup of amyloid protein.

Parkinson’s Disease results from depleted dopamine levels in the brain and affects approximately ten million people worldwide. About sixty thousand new cases are diagnosed annually in the United States. Parkinson symptoms include tremor, slowed movement, muscle rigidity, poor balance as well as speech and writing changes. Although the symptoms can be diminished to varying degrees by medication, the disease is progressive and there is no cure.

Artificial intelligence involves algorithms where computers can learn essentially on their own by being exposed to data sets. In other words, they are not given specific instructions continuously in the form of code. With recent advances in AI algorithms, diagnostic imaging studies may be able to help predict with a fair degree of accuracy who will develop these disorders.  

In a recent study conducted at McGill University in Canada, researchers utilized an AI algorithm to analyze amyloid patterns in the PET (positron emission tomography) scans of patients before and after the development of subtle cognitive impairment. The program learned from this comparative analysis and was then exposed to the PET scans of an entirely new set of patients with mild cognitive impairment. According the study, the algorithm could predict with an 84 percent accuracy, who would develop Alzheimer’s.

Predicting the onset of Parkinson’s Disease using AI assisted diagnostic imaging analysis is more challenging.  Standard MRI imaging of the brain lacks specific diagnostic markers for Parkinson’s. There are also variants of PD that may present with different neurodegenerative changes on imaging studies. Cortical atrophy of the temporal, occipital and subcortical structures of the brain may correlate with the development of dementia associated with traditional Parkinson’s.

For patients with Atypical Parkinsonism, atrophy and signal intensity changes on MRI are often seen in the cortex, basal ganglia, brain stem and cerebellum.  Vascular Parkinsonism should be considered when MRI studies demonstrate gliosis, lacunar infarctions and white matter lesions.

The various forms of Parkinson’s Disease and the non-specific changes associated with each type on imaging studies makes predicting who will develop PD challenging even with an AI algorithmic approach.

Aside from conventional MRI, there are various types of sequencing and imaging modalities that can assist with the diagnosis of PD and AP (Atypical Parkinson’s) ​The DWI sequence can be used to measure the microstructural integrity of white matter and grey matter structures in patients with PD and AP. DTI, or diffusion tensor imaging, can be used to provide information about tissue diffusivity. Studies have indicated that even when traditional MRI does not demonstrate any abnormalities, DWI and DTI measures of brain structures such as the basal ganglia, brainstem and cerebellum appear to accurately identify PD and AP patients.

More advanced imaging techniques such as magnetization transfer imaging and functional MRI as well more powerful MRI magnets have become available. These will lead to enhanced diagnostic accuracy in the clinical setting. In conjunction with more powerful processors, AI algorithms will no doubt improve predictive accuracy for neurodegenerative disorders and other chronic disease states that present with pathology in imaging studies.

With AI enhanced predictive capability, drug development can focus on preventative therapeutics instead of just medications to temper the symptoms of neurodegenerative disorders.