Quantitative imaging allows radiologists to do more than just lesion detection. Hazel Tang reports on how AI-driven NeuroQuant is providing fast, accurate and proven automated brain image analysis

 

An elderly couple walk into the clinic. The husband is anxious that his wife has been experiencing significant memory loss in recent months and is extremely concerned about her wellbeing. Such cognitive complaints are common in neuroradiological practice. However, without added information, it’s very challenging for clinicians to get a sense of what’s behind the memory loss which occurs in normal aging, but also in several mimics of neurodegeneration including side effects from medicine, alcohol consumption, sleep problems, and depression.

“On the clinical side, we always want to reassure patients by telling them their condition is likely to remain stable or only decline slowly over time, says Dr. James Brewer, a neurologist who studies how Alzheimer’s disease changes the brain at the University of California, San Diego. “But in cases like these, we can only say we are not sure what’s going to happen. So, this is a blind spot in practice.” In addition, many of the neurodegenerative illnesses are symmetric or at least affecting both sides of the brain much more symmetrically. Therefore, it’s very difficult to visually pick up abnormalities from MRI scans especially at the very earliest stages.

So when Cortechs.ai, a Californian radiological AI company launched its flagship product NeuroQuant, an AI-driven medical device that makes quantitative analysis of MRI images of the human brain, Dr. Brewer jumped on board and became the first physician to incorporate the tool into his clinical practice.

NeuroQuant automatically detects and quantifies atrophy of brain structures to facilitate the diagnosis of a variety of brain disorders including Alzheimer’s disease, multiple sclerosis, epilepsy, and traumatic brain injury. The quantification helps pick up subtle changes in the anatomy of the brain and underlines the biomarkers showing an individual has the early signs of Alzheimer’s disease.

“In all the mimics of early memory changes, because of medications, lifestyle habits, or mood – if you look into the brain, you would not see a neurodegenerative process going on or any clear objective evidence of atrophy,” Dr. Brewer explains. “Whereas in these other cases, by the time the patients come in with a memory complaint due to neurodegenerative etiology, there’s already been at least five years of buildup atrophy. There is a great deal of scientific evidence that brain structures such as the hippocampus begin to atrophy in the early stages of dementia.

“Reliably characterizing such changes based on MRI data requires the use of quantitative techniques,” continues Dr. Brewer. “But previous methods for deriving numerical information from MRI scans have been too slow and cumbersome for use outside of research studies. Radiologists may use statements like ‘mild evolutional changes’ or ‘moderate atrophy that’s appropriate for age’ to describe the state of a patient’s brain, causing referring physicians to be somewhat unsatisfied when it comes to neurodegeneration diagnosis. But by post-processing MRI data with NeuroQuant, we can now routinely consider quantitative information about the extent and progression of atrophy in a patient’s brain when evaluating treatment options.”

Dr. Brewer went on to become the Chief Medical Advisor at Cortechs.ai, assisting in the studies that led to NeuroQuant’s FDA 510(k) clearance in 2006 and the CE mark thereafter. He believes the most helpful component of NeuroQuant is to segment the brain and provide numbers that allow radiologists to confidently tell patients, “You are relatively healthy” or “There’s something we need to look into”. This not only accelerates radiological interpretation but also for direct use by the referring physicians. “There’s a lot of excitement about providing a numerical format to mark the changes in the brain,” Dr. Brewer adds. “This means that NeuroQuant can also be used as an efficient way for deriving quantitative imaging biomarkers in future clinical trials research.”

NeuroQuant receives MRI scans routed from a scanner or Picture-Archiving and Communication System (PACS) as input and returns age-related atrophy reports and numerical and color-blended anatomical volumes annotated with graphical overlays that comply with DICOM standards. The tool can also be configured to provide automated MRI quantification solutions that meet the needs of independent healthcare providers, imaging centers, and large enterprises conducting multi-site clinical trials.

Founded in 2001, Cortechs.ai is a new brand name undertaken by CorTechs Labs last December to reflect the company’s evolution in the AI imaging software space and vision for the future. In the words of its former President, Michael E. Smith; Cortechs.ai seeks to develop a powerful pipeline of sophisticated, AI-driven analysis processes that can quantify MRIs without human interventions, providing numerical information that’s convenient and easy to read by clinicians, offering immediate comparison to age-appropriate normative data.