Breaking News
February 20, 2019 - Newly licensed nurses work for long hours, also have a second paid job
February 20, 2019 - Physicists identify simple mechanism used by deadly bacteria to fend off antibiotics
February 20, 2019 - FDA Grants Priority Review to Genentech’s Personalized Medicine Entrectinib
February 20, 2019 - Exposure to chemicals before and after birth is associated with a decrease in lung function
February 20, 2019 - Neuroscientists reveal that simple brain region can guide complex feats of mental activity
February 20, 2019 - Study finds new link between food allergies and multiple sclerosis
February 20, 2019 - First gene therapy operation for macular degeneration is a success
February 20, 2019 - Physicians graduated outside the U.S. offer better care for Medicare patients with complex needs
February 20, 2019 - FDA Approves Keytruda (pembrolizumab) for the Adjuvant Treatment of Patients with Melanoma with Involvement of Lymph Node(s) Following Complete Resection
February 20, 2019 - Study identifies brain cells that modulate behavioral response to threats
February 20, 2019 - Researchers take closer look at how viruses bind cells and cause infection
February 20, 2019 - Newly developed gene therapy helps decelerate aging process
February 20, 2019 - Study suggests new treatment strategy for deadly brain cancer
February 20, 2019 - Scientists develop unique hybrid implant that imitates bone structure
February 20, 2019 - Push-ups can be tailored to meet specific needs of individuals
February 20, 2019 - CVD Does Not Modify Depression-Mortality Link in Elderly
February 20, 2019 - Electrical activity early in fruit flies’ brain development could shed light on how neurons wire the brain
February 20, 2019 - Machine learning technique helps predict which asthma patients respond to corticosteroid therapy
February 20, 2019 - Self-reported sleep duration is a useful tool to measure sleep in children, study suggests
February 20, 2019 - T-cells play key role in how the body fights follicular lymphoma
February 20, 2019 - Study shows how 3D organization of genetic material helps perpetuate the species
February 20, 2019 - Researchers engineer stem cell with ‘suicide genes’ to induce cell death in all but beta cells
February 20, 2019 - Study reveals major sex differences in management of cardiovascular risk factors among U.S. adults
February 20, 2019 - Health Tip: Get Your Child to School on Time
February 20, 2019 - Shortcut strategy for screening compounds with clinical potentials for drug development
February 20, 2019 - Common acid reflux drugs tied to elevated risk for kidney disease
February 20, 2019 - Microbiome could be culprit when good drugs do harm
February 20, 2019 - Prenatal exposure to forest fires causes stunted growth in children
February 20, 2019 - Gene therapy restores hearing in mice with congenital genetic deafness
February 20, 2019 - First molecular test predicts treatment response for kidney cancer
February 20, 2019 - New method for improved visualization of single-cell RNA- sequencing data
February 20, 2019 - Researchers capture altered brain activity patterns of Parkinson’s in mice
February 20, 2019 - A possible blood test for detecting Alzheimer’s disease before symptoms show
February 20, 2019 - Primary care physicians associated with longevity, new research finds
February 19, 2019 - New study identifies many key lessons to establish sanctioned safe consumption sites
February 19, 2019 - Single CRISPR treatment can safely and stably correct genetic disease
February 19, 2019 - Multinational initiative to study familial primary distal renal tubular acidosis
February 19, 2019 - Breakthrough study highlights the promise of cell therapies for muscular dystrophy
February 19, 2019 - Subsymptom Threshold Exercise Speeds Concussion Recovery
February 19, 2019 - Midline venous catheters – infants: MedlinePlus Medical Encyclopedia
February 19, 2019 - Searching for side effects
February 19, 2019 - Humanity is all right, probably, although human extinction remains quite possible, researcher says
February 19, 2019 - Having Anesthesia Once as a Baby Does Not Cause Learning Disabilities, New Research Shows
February 19, 2019 - Anti-cancer immunotherapy could be used to fight HIV
February 19, 2019 - Customized Micropatterning for Improved Physiological Relevance
February 19, 2019 - Unique gene therapy approach paves new way to tackle rare, inherited diseases
February 19, 2019 - Activating gene that helps excite neurons reverses depression in male mice
February 19, 2019 - Science Puzzling Out Differences in Gut Bacteria Around the World
February 19, 2019 - Cells that destroy the intestine
February 19, 2019 - On recovery, vulnerability and ritual: An exhibit in white
February 19, 2019 - Scientific Duo Gets Back To Basics To Make Childbirth Safer
February 19, 2019 - COPD patients need more support when understanding new chest symptoms
February 19, 2019 - Using light-based method for production of pharmaceutical molecules
February 19, 2019 - Scientists find link between inflammation and cancer
February 19, 2019 - The High Cost Of Sex: Insurers Often Don’t Pay For Drugs To Treat Problems
February 19, 2019 - Hearing impairment associated with accelerated cognitive decline with age
February 19, 2019 - Researchers identify multiple genetic variants associated with body fat distribution
February 19, 2019 - Influenza and common cold are completely different diseases, study shows
February 19, 2019 - Scientists untangle how microbes manufacture key antibiotic compound
February 19, 2019 - Greater primary care physician supply associated with longer life spans
February 19, 2019 - HIV-1 protein suppresses immune response more broadly than thought
February 19, 2019 - Brain imaging indicates potential success of drug therapy in depressive patients
February 19, 2019 - For 2020 Dem Hopefuls, ‘Medicare-For-All’ Is A Defining Issue, However They Define It
February 19, 2019 - Specialized lung cells appear in the developing fetus much earlier than previously thought
February 19, 2019 - KU professor discusses promise of brain-computer interface to aid, restore communication
February 19, 2019 - Highly effective solution for detecting onset of aggregation in nanoparticles
February 19, 2019 - Early marker of cardiac damage triggered by cancer treatment identified
February 19, 2019 - Antidepressant drug could save people from deadly sepsis, research suggests
February 19, 2019 - CRISPR technology creates pluripotent stem cells that are ‘invisible’ to the immune system
February 19, 2019 - New study establishes how stress favors breast cancer growth and spread
February 19, 2019 - Midlife Systemic Inflammation Linked to Later Cognitive Decline
February 19, 2019 - Therapy derived from parasitic worms downregulates proinflammatory pathways
February 19, 2019 - Antimicrobial reusable coffee cups are less likely to become contaminated with bacteria, study shows
February 19, 2019 - Harnessing the evolutionary games played by cancer cells to advance therapies
February 19, 2019 - AHA News: Heart Transplant Survivor Gets Wedding Proposal at Finish Line
February 19, 2019 - HIV hidden in patients’ cells can now be accurately measured
February 19, 2019 - Research finds reasons for sudden cardiac death in patients with stable ischemic disease
February 19, 2019 - New protocol could help physicians to rule out bacterial infections in infants
February 19, 2019 - Women experiencing miscarriage should be offered treatment choices
February 19, 2019 - New protocol can help identify febrile infants at low risk for serious bacterial infections
AI matched, outperformed radiologists in screening X-rays for certain diseases | News Center

AI matched, outperformed radiologists in screening X-rays for certain diseases | News Center

image_pdfDownload PDFimage_print

“Usually, we see AI algorithms that can detect a brain hemorrhage or a wrist fracture — a very narrow scope for single-use cases,” said Matthew Lungren, MD, MPH, assistant professor of radiology. “But here we’re talking about 14 different pathologies analyzed simultaneously, and it’s all through one algorithm.”

The goal, Lungren said, is to eventually leverage these algorithms to reliably and quickly scan a wide range of image-based medical exams for signs of disease without the backup of professional radiologists. And while that may sound disconcerting, the technology could eventually serve as high-quality digital “consultations” to resource-deprived regions of the world that wouldn’t otherwise have access to a radiologist’s expertise. Likewise, there’s an important role for AI in fully developed health care systems too, Lungren added. Algorithms like CheXNeXt could one day expedite care, empowering primary care doctors to make informed decisions about X-ray diagnostics faster, without having to wait for a radiologist.

“We’re seeking opportunities to get our algorithm trained and validated in a variety of settings to explore both its strengths and blind spots,” said graduate student Pranav Rajpurkar. “The algorithm has evaluated over 100,000 X-rays so far, but now we want to know how well it would do if we showed it a million X-rays — and not just from one hospital, but from hospitals around the world.”

A paper detailing the findings of the study was published online Nov. 20 in PLOS Medicine. Lungren and Andrew Ng, PhD, adjunct professor of computer science at Stanford, share senior authorship. Rajpurkar and fellow graduate student Jeremy Irvin are the lead authors.

Practice makes perfect

Lungren and Ng’s diagnostic algorithm has been in development for more than a year. It builds on their work on a previous iteration of the technology that could outperform radiologists when diagnosing pneumonia from a chest X-ray. Now, they’ve boosted the abilities of the algorithm to flag 14 ailments, including masses, enlarged hearts and collapsed lungs. For 11 of the 14 pathologies, the algorithm made diagnoses with the accuracy of radiologists or better.

Back in the summer of 2017, the National Institutes of Health released a set of hundreds of thousands of X-rays. Since then, there’s been a mad dash for computer scientists and radiologists working in artificial intelligence to deliver the best possible algorithm for chest X-ray diagnostics.

We need to be thinking about how far we can push these AI models to improve the lives of patients anywhere in the world.

The scientists used about 112,000 X-rays to train the algorithm. A panel of three radiologists then reviewed a different set of 420 X-rays, one by one, for the 14 pathologies. Their conclusions served as a “ground truth”— a diagnosis that experts agree is the most accurate assessment — for each scan. This set would eventually be used to test how well the algorithm had learned the telltale signs of disease in an X-ray. It also allowed the team of researchers to see how well the algorithm performed compared to the radiologists.

“We treated the algorithm like it was a student; the NIH data set was the material we used to teach the student, and the 420 images were like the final exam,” Lungren said. To further evaluate the performance of the algorithm compared with human experts, the scientists asked an additional nine radiologists from multiple institutions to also take the same “final exam.”

“That’s another factor that elevates this research,” Lungren said. “We weren’t just comparing this against other algorithms out there; we were comparing this model against practicing radiologists.”

What’s more, to read all 420 X-rays, the radiologists took about three hours on average, while the algorithm scanned and diagnosed all pathologies in about 90 seconds.

Next stop: the clinic

Now, Lungren said, his team is working on a subsequent version of CheXNeXt that will bring the researchers even closer to in-clinic testing. The algorithm isn’t ready for that just yet, but Lungren hopes that it will eventually help expedite the X-ray-reading process for doctors diagnosing urgent care or emergency patients who come in with a cough.

“I could see this working in a few ways. The algorithm could triage the X-rays, sorting them into prioritized categories for doctors to review, like normal, abnormal or emergent,” Lungren said. Or the algorithm could sit bedside with primary care doctors for on-demand consultation, he said. In this case, Lungren said, the algorithm could step in to help confirm or cast doubt on a diagnosis. For example, if a patient’s physical exam and lab results were consistent with pneumonia, and the algorithm diagnosed pneumonia on the patient’s X-ray, then that’s a pretty high-confidence diagnosis and the physician could provide care right away for the condition. Importantly, in this scenario, there would be no need to wait for a radiologist. But if the algorithm came up with a different diagnosis, the primary care doctor could take a closer look at the X-ray or consult with a radiologist to make the final call.

“We should be building AI algorithms to be as good or better than the gold standard of human, expert physicians. Now, I’m not expecting AI to replace radiologists any time soon, but we are not truly pushing the limits of this technology if we’re just aiming to enhance existing radiologist workflows,” Lungren said. “Instead, we need to be thinking about how far we can push these AI models to improve the lives of patients anywhere in the world.”

Other Stanford authors of the study are biostatistician Robyn Ball, PhD; undergraduate student Kaylie Zhu; former research assistant Brandon Yang; data scientist Hershel Mehta; research assistants Tony Duan and Daisy Ding; former research assistant Aarti Bagul; professor of radiology and of medicine Curtis Langlotz, PhD; assistant professor of radiology Bhavik Patel, MD; associate professor of radiology Kristen Yeom, MD; research associate Katie Shpanskaya; associate professor of radiology Francis Blackenberg, MD; clinical assistant professor of radiology Jayne Seekins, MD; clinical associate professor of radiology Safwan Halabi, MD; and clinical assistant professor of radiology Evan Zucker, MD.

Researchers from Duke University and from the University of Colorado also contributed to the study.

Lungren is a member of Stanford Bio-X, the Stanford Child Health Research Institute and the Stanford Cancer Institute.

Stanford’s departments of Radiology and of Computer Science along with the Stanford Center for Artificial Intelligence in Medicine & Imaging supported the work.

Tagged with:

About author

Related Articles