In the earliest days of the pandemic, machine learning showed exceptional promise for COVID-19 diagnosis. Reliably, early machine learning models outperformed doctors in recognizing the telltale COVID-induced pneumonia on CT scans from hospitalized patients. However, more conventional testing methods quickly lapped machine learning-based methods, detecting the onset of COVID well before hospitalization and with greater accuracy. Now, a year later, a team of researchers led by the University of Cambridge has concluded a review of COVID diagnosis ML models, finding that even in 2021, none of the proposed models are suitable for clinical use.

The researchers whittled down 2,212 studies, eventually focusing on 62 studies – most of which were not peer-reviewed – published between January 1st and October 3rd of 2020, all of which presented machine learning models for diagnosing or predicting COVID-19 infection based on X-rays and/or CT scans. These 62 studies collectively described more than 300 such models – and the researchers found all of them substantially lacking.

“The international machine learning community went to enormous efforts to tackle the COVID-19 pandemic using machine learning,” said James Rudd, one of the senior authors of the review and a member of Cambridge’s Department of Medicine. “These early studies show promise, but they suffer from a high prevalence of deficiencies in methodology and reporting, with none of the literature we reviewed reaching the threshold of robustness and reproducibility essential to support use in clinical practice.”

The issues were wide-ranging: some studies suffered from poor data quality, while others were not reproducible and yet more exhibited biases in their design. By way of example, the authors pointed out that some of the datasets used to train some of the machine learning models included scans from children. “Since children are far less likely to get COVID-19 than adults, all the machine learning model could usefully do was to tell the difference between children and adults, since including images from children made the model highly biased,” explained Michael Roberts, a member of Cambridge’s Department of Applied Mathematics and Theoretical Physics. 

Other datasets were too small, some were poorly labeled. Some models used the same data for training and testing. And, overwhelmingly, the designers of the models failed to meaningfully incorporate input from radiologists and clinicians who might have insight into the real-world implications of the data and diagnoses at hand. “Whether you’re using machine learning to predict the weather or how a disease might progress,” Roberts said, “it’s so important to make sure that different specialists are working together and speaking the same language.”

Better late than never, though, and to that end, the reviewers have some recommendations for machine learning model developers working on COVID diagnosis: know the data you’re working with, especially when it comes to public datasets; work with diverse, large datasets; and, crucially, include better documentation to allow for reproducibility.

The post Machine Learning for COVID Diagnosis Falls Short appeared first on Artificial Intelligence.

London: Artificial intelligence (AI) can diagnose Alzheimer’s disease faster and improve prognosis, a new study has revealed.

Alzheimer’s disease is a neurological disorder in which the death of brain cells causes memory loss and cognitive decline.

Scientists at the UK’s University of Sheffield’s Neuroscience Institute examines how the routine use of AI in healthcare could help to relieve the time and economic impact that common neurodegenerative diseases, such as Alzheimer’s and Parkinson’s, put on the National Health Service (NHS).

The main risk factor for many neurological disorders is age, and with populations worldwide living longer than ever before, the number of people with a neurodegenerative disease is expected to hit unprecedented levels. The number of people living with Alzheimer’s alone is predicted to treble to 115 million by 2050, posing a real challenge for the health system, the study noted.

The study, published in the journal Nature Reviews Neurology, highlights how AI technologies, such as machine learning algorithms, can detect neurodegenerative disorders—which cause part of the brain to die—before progressive symptoms worsen. This can improve patients’ chances of benefitting from successful disease-modifying treatment.

Lead author of the study, Dr Laura Ferraiuolo from the University of Sheffield, said: Most neurodegenerative diseases still do not have a cure and in many cases are diagnosed late due to their molecular complexity. Widespread implementation of AI technologies can help, for example, predict which patients showing mild cognitive impairment will go on to develop Alzheimer’s disease, or how severely their motor skills will decline over time.

AI-powered technologies can also be used to help patients communicate their symptoms remotely and in the privacy of their own homes, which will be an enormous benefit to patients with mobility issues.

Machine learning algorithms can be trained to recognise changes caused by diseases in medical images, patient movement information, speech recordings or footage showing patient behaviour, making the AI a valuable diagnostic aid.

For example, it can be used by trained professionals in radiology departments to analyse images more quickly and highlight critical results for an immediate follow-up.

Algorithms can also listen to patients’ speech and analyse their vocabulary and other semantic features to assess their cognitive function. Machine learning can also use information contained within electronic health records or genetic profiles to suggest the best treatments for individual patients.

Monika Myszczynska, another scientist from the University of Sheffield, said: “It is too early to talk about the outcomes in terms of treatments but, in this study, we examined how machine learning methods can be used to identify the best course of treatment for patients based on their disease progression or how it could be used to identify new therapeutic targets and drugs.”

The post Artificial intelligence can speed up and improve Alzheimer’s diagnosis, says study appeared first on Artificial Intelligence.

To understand the benefits that artificial intelligence can bring to the world of human medicine, consider the case of Ayako Yamashita, a 60-year-old Japanese woman, whose condition befuddled doctors in 2015.

Yamashita was thought to be suffering from acute myeloid leukemia. But after several unsuccessful treatment attempts, her doctors decided to search for another answer to her condition. They turned to IBM’s Watson, an AI system capable of analyzing vast amounts of data.

The computer reviewed nearly 20 million previously-published oncological research studies and cross-referenced data points. Watson’s analysis suggested the woman had a rare form of leukemia not detected through conventional methods. This led to a change in treatment and doctors crediting Watson for saving  the woman’s life.

The analysis of such huge amount of data is next to impossible for a human mind, but it’s like a walk in a digital park for AI. And it shows what may be one of the valuable things that AI can do for us. It is “the most practical application in the field of medical and healthcare for artificial intelligence,” said Seiji Yamada, of the National Institute of Informatics and chairman of the Japanese Society for Artificial Intelligence.

The global artificial intelligence market is expected to reach $19.47 billion by 2022, according to the research firm Allied Market Research. As AI is marking its presence, tech giants are working to capitalize on new opportunities. The healthcare sector is a natural fit, according to Sanjay Gupta, managing director, South Asia and Middle East for NICE, a technology firm based in Israel, in an interview with ETHealthworld:

“The development of automation enabled by technologies including robotics and artificial intelligence in healthcare sector brings the promise of higher productivity with increased safety.”

Saving lives and time

Among Google’s many AI ventures is an effort to develop new products targeting the health sector. The company is focusing on applications for life preservation, preventive care and improving health care services.

The company plans to launch a trial in India to test an AI system that scans a person’s eyes to look for signs of diabetic retinopathy. The company aims to license the technology to clinics. The system already has proven itself adept at detecting high blood pressure, or risk of heart disease or stroke, according to a study published in early 2018.

From a story published in the Washington Post:

“This may be a rapid way for people to screen for risk,” Harlan Krumholz, a cardiologist at Yale University who was not involved in the study, wrote in an email. “Diagnosis is about to get turbo-charged by technology. And one avenue is to empower people with rapid ways to get useful information about their health.”

Jeff Dean, the Chief at Google AI, outlined for Boss Magazine how this system will enable doctors to better diagnose and treat patients for a range of diseases. Moreover, this system will also track key events in the patient’s past (including hospital stays) to help doctors more effectively.

The post Can artificial intelligence give us a more efficient health care system? appeared first on Artificial Intelligence.