The application of artificial intelligence in medicine and healthcare can assist with the optimization of the care trajectory of chronic disease patients; and it can suggest precision therapies for complex illnesses. Furthermore, algorithms can improve subject enrollment into clinical trials, leading to the creation of more efficacious medicines.
To showcase the technology, three examples of artificial intelligence as applied to the medical field are examined.
AI for lung cancer detection
The science company Draper has begun a program designed to make artificial intelligence systems better at detecting lung cancer warning signs. This is by assessing medical images. The reason for this focus is due to the high mortality rate linked to lung cancer, and the complexities involved in detecting the disease early. With lung cancer most of the symptoms only become apparent when the disease has advanced, making treatment difficult.
The Draper study, as PharmaceuticalPhorum discusses, involves applying artificial intelligence and machine learning to improve clinical decision-making by assessing two-dimensional images from CT scans. This is to aid radiologists in checking for suspicious areas on the images.
The technology in development uses a three-dimensional convolutional neural network to classify regions of a scan that appear suspicious and also to calculate the level of statistical confidence in the decision.
The technology has been presented to the journal IEEE Transactions on Medical Imaging, where the research paper is titled “A 3D Probabilistic Deep Learning System for Detection and Diagnosis of Lung Cancer Using Low-Dose CT Scans.”
This application represents the latest step in demonstrating how artificial intelligence is proving to be as reliable as human physicians in diagnosis are.
World’s first drug produced using AI
British drug discovery company Exscientia has produced the first precision engineered drug produced with the aid of artificial intelligence. The medicine is now set to commence clinical trials.
For the development, Exscientia worked in partnership with the Japanese company Sumitomo Dainippon Pharma, to create a medicine intended to treat obsessive-compulsive disorder. The clinical trial will assess the medicine’s efficacy.
The typical time to develop a new medication is five years; in the case of the Exscientia project, the process took just 12 months. The compound selected for the drug was detected through the analysis of 350 synthesised compounds, whittled down from 2,500 compounds by a specially developed algorithm.
Andrew Hopkins, CEO of Exscientia said in a statement: “This is very different from the use of AI to repurpose drugs. Success stories like this will provide us with the hard evidence that AI really will deliver on its transformative potential.”
Hopkins added: “We believe that this entry of DSP-1181, created using AI, into clinical studies is a key milestone in drug discovery.”
Tis news follows the research and development organisation Deep Genomics deploying artificial intelligence to select a therapeutic drug candidate for the first time.
Assessing glucose levels
A research team have put in place an artificial intelligence system to detect low glucose levels via an electrocardiogram readout (for hypoglycaemia detection). This approach obviates the need for a blood test. The method is effective for the detection of diabetes.
The electrocardiogram readout is obtained from a wearable sensor. Trials have shown the approach to have an 82 percent reliability rate and the potential is there for the invasive fingerprick test to be eliminated altogether.
The artificial intelligence scans for a distinctive electrocardiogram waveform which is indicative of a hypoglycaemic reaction. The effectivity is built-in through the algorithm being trained for each patient’s own readings; hence the system looks for unique patterns.
The research has been published in the journal Scientific Reports, where the paper is headed “Precision Medicine and Artificial Intelligence: A Pilot Study on Deep Learning for Hypoglycemic Events Detection based on ECG.”
Essential Science
This article is part of Digital Journal’s regular Essential Science columns. Each week Tim Sandle explores a topical and important scientific issue.
Last week we considered an advanced method for the detection of biosignatures, paving the way for the early detection of tuberculosis. The method allows for TB to be detected in patients, months before symptoms appear.
The week before we looked at a new machine learning system, developed to characterize 800 million-year-old amino acid patterns that had, up until now, puzzled scientists. These protein patterns are of great importance and they are responsible for facilitating protein interactions.
