Top 5 Major Quantum Leaps of AI in Healthcare

by July 26, 2019 0 comments

As the world’s most significant tech company Google is venturing in the AI-driven healthcare system, the focus of the rest of the industry is drifted towards AI breakthroughs in the medical sector. The reverberating power of AI is genuinely changing lives for better.

Here are some major quantum leaps taken by Artificial Intelligence in the healthcare sector to empower public health with new-age technology.


AI Detects Lung Cancer

Google researchers and Northwestern Medicine have recently worked together to develop an AI system that is capable of detecting lung cancer better than human radiologists. The system is trained with a deep learning algorithm which interprets CT scans to foresee a patient’s likelihood of possessing the disease.

Daniel Tse, product manager at Google Brain and his research team applied DL (AI) to 42,290 low-dose CT (LDCT) scans. The scans were provided by the Northwestern Electronic Data Warehouse and other sources belonging to Northwestern hospitals in Chicago. Further, the images were taken from almost 15,000 patients from a National Institutes of Health study conducted in 2002. Out of these patients, 578 were developing cancer within a year.

The study was funded by Google and researchers employed AI as a diagnostic tool to evaluate images and predict disease eliminating human opinion. The AI model detected lung cancers 5 percent more often than the experts. This particular system will go public via Google Cloud Healthcare API as it is still under trials and additional research.


AI Can Spot Causes of Autism in Uncharted DNA

Recently, a research team has employed AI to unveil novel genetic mutation involved with autism in non-coding areas of DNA which is in common language called a junk region. Scientists implemented deep learning algorithm to examine these regions of the genome which affects how much genes are being produced.

The team of researchers went through 1,20,000 mutations to identify selected ones tied to genetic behavior in an autistic patient. Although the result did not show real cause but revealed the potential noncoding genetic contributor that seems rare.

The algorithm used in this study operates on complex data analysis to reveal which are challenging to recognize by other means. The deep learning algorithms spotted the relevant region of DNA in the examined genomes and further predicted that which sections influence the 2,000+ protein interactions that regulate genes. The system also forecasts if a single mutation in a DNA pairing could affect these protein interactions.


Combination of Machine Learning and Wearable Sensor Can Detect Heart Disease

An AI classifier has been developed recently which is capable of detecting specific cardiovascular disease with the use of wearable biosensors tied on wrists of people. By lodging diagnostic approach using ML and wearable sensor, researchers have developed a non-invasive tool to identify hypertrophic cardiomyopathy which can cause serious complications.

After employing AI/ML technology, researchers concluded that the use of the wearable sensor along with machine learning is successful in spotting patients with oHCM. Even if several devices use PPG sensors to detect the rate of blood flow and heart pumping, AI-driven tool has the potential to detect unrecognizable oHCM. Well, there is still scope for the device to go through future research to analyze a larger population for a better understanding of age and gender factors of disease.


AI System Detects Diabetic Retinopathy

The private AI diagnostics company IDx recently announced that the American Medical Association’s (AMA) Current Procedural Terminology (CPT) Editorial Panel has accepted a new category 1 CPT® code for automated point-of-care retinal imaging. The new code facilitates correct billing of IDx-DR which is an FDA-cleared autonomous AI system that detects diabetic retinopathy causing blindness. The CPT codes are developed and reviewed by clinician experts under a transparent and open process. The code provides a uniform language for submitting healthcare procedures and services for payor reimbursement. It is scheduled to be effective in January 2021 which will streamline the coding and billing process for healthcare providers using IDx-DR.

The policy recommendations provided by AMA for AI in healthcare says “payment and coverage of all health care AI systems that are conditioned on complying with all appropriate federal and state laws and regulations, including but not limited to those governing patient safety, efficacy, equity, truthful claims, privacy, and security as well as state medical practice and licensure laws.”

Additionally, in 2018, IDx-DR became the first autonomous AI diagnostic system to be cleared for use without any involvement of a physician to interpret the result. The system also benefited non-eye care providers to put forward an immediate diagnostic assessment for diabetic retinopathy at the point of care without any specialist review.


AI/ML Can Detect Psychosis Like Schizophrenia

A machine learning approach has been developed by Harvard and Emory University researchers to analyze language for signs of psychosis. ML technology provides an accurate way to detect mental illness through subtle characteristics projected by people in their speaking.

The ML system evaluates the semantic density and use of words relating to sound in one’s speech and hence accurately predicts which individuals are likely to develop psychosis.

Psychosis conditions like schizophrenia often display auditory hallucinations in patients and tend to implicitly talk about voices and sounds. The researchers used a technique called vector unpacking which determines semantic density by dividing sentences into component vectors and measuring the sentence’s richness.

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.

Your data will be safe!Your e-mail address will not be published. Also other data will not be shared with third person.