Breaking News
December 15, 2017 - NIH-funded study to explore hearing loss risk in Detroit firefighters
December 15, 2017 - Caris Life Sciences reveals identification of new mechanism of action to treat NHL
December 15, 2017 - Loyola Medicine study finds high success rate for diabetic Charcot foot surgery
December 15, 2017 - Bone marrow edema does not increase due to intense physical activity, study finds
December 15, 2017 - Human ‘common cold’ virus kills healthy chimpanzees in Uganda
December 15, 2017 - Experience of reflex walking refines perception of biological motion during early infancy
December 15, 2017 - FDA Approves Admelog (insulin lispro) Rapid-Acting “Follow-On” Insulin Product to Treat Diabetes
December 14, 2017 - Friday Feedback: Good Idea for Ex-Pharma Exec to Run HHS?
December 14, 2017 - More than 200 people sickened onboard Ovation of Seas cruise
December 14, 2017 - FDA announces new approach for sharing updates on antibiotics, antifungal drugs to physicians
December 14, 2017 - Steroid study provides new insights into medicines’ side effects
December 14, 2017 - Government announces plans to include eye test reminder during driving license renewal
December 14, 2017 - Non-ionizing radiation from magnetic fields could have adverse biological impacts on health
December 14, 2017 - Bi-annual MRI beats mammograms in detecting breast cancer among women with genetic risk
December 14, 2017 - Researchers develop new method for quickly detecting signs of multiple sclerosis
December 14, 2017 - In era of increased competition, hospitals fret over ratings
December 14, 2017 - Female veterans experience improvement in low back pain with course of chiropractic care
December 14, 2017 - Relieving Symptoms of Cystic Fibrosis with Exercise
December 14, 2017 - FDA Alert: Blue Pearl All Natural Male Enhancement Supplement: Recall
December 14, 2017 - CardioBrief: In Defense Of ORBITA
December 14, 2017 - Definition of High Blood Pressure Drops: MedlinePlus Health News
December 14, 2017 - Drug may help surgical patients stop opioids sooner
December 14, 2017 - Researchers develop biosensor that enables development of new health tests
December 14, 2017 - Radiation therapy can be used to treat patients with life-threatening heart rhythm
December 14, 2017 - UVA researchers developing tool to help prostate cancer patients weigh treatment options
December 14, 2017 - Experts tell Congress how to cut drug prices
December 14, 2017 - Researchers use cryptographic techniques to decode activity of motor neurons
December 14, 2017 - Study finds changes in the heart after spinal cord injury
December 14, 2017 - Health Highlights: Dec. 12, 2017
December 14, 2017 - Pelzman’s Picks: How States Can Cut Disparities in Care and Costs
December 14, 2017 - New Hemophilia Treatment Stems Bleeding Episodes: MedlinePlus Health News
December 14, 2017 - Onetime ‘world’s heaviest man’ has second surgery in Mexico
December 14, 2017 - Belgian researchers create transplantable artificial ovary prototype
December 14, 2017 - Using atraumatic needles for lumbar punctures decreases risk of complications
December 14, 2017 - Outpatient total knee replacement surgery linked to higher rates of complications
December 14, 2017 - Social impairments can be corrected by brain stimulation
December 14, 2017 - Studies reveal possibility for memory T cells to serve a dual purpose
December 14, 2017 - Antibody-Drug Conjugate Ups PFS in Untreated Hodgkin’s
December 14, 2017 - Study finds reading information aloud to yourself improves memory
December 14, 2017 - Researchers use RNA nanotechnology to program exosomes for delivering effective cancer therapies
December 14, 2017 - Living Lyme disease bacteria found months after antibiotic treatment
December 14, 2017 - These annual checkups help seniors not only survive but thrive
December 14, 2017 - Study reveals impact of diabetes during pregnancy on baby’s heart
December 14, 2017 - Hydraulic fracturing is harmful to infants health, study states
December 14, 2017 - Huntington’s disease drug clears initial hurdles
December 14, 2017 - TPU researchers create 3D-printed models of children’s hearts
December 14, 2017 - Brain responses of children with inherited dyslexia risk predict their future reading speed
December 14, 2017 - People diagnosed with schizophrenia or bipolar disorder may actually have treatable condition
December 14, 2017 - Study: New Furosemide Formulation Simplifies Administration for HF
December 14, 2017 - Discrimination harms your health—and your partner’s
December 14, 2017 - Having older brothers may increase the likelihood of being gay
December 14, 2017 - New scientific yardstick released to help early detection of Alzheimer’s disease
December 14, 2017 - New finding demonstrates what happens at cellular level during onset of type2 diabetes
December 14, 2017 - Study identifies potassium as key to circadian rhythms in red blood cells
December 14, 2017 - Good friends might be your best brain booster as you age
December 14, 2017 - NIH expected to award up to $70 million to launch Alzheimer’s Clinical Trials Consortium
December 14, 2017 - Pitting pathogens against each other could prevent drug resistance emerging
December 14, 2017 - Study provides new insights into development of Sonic Hedgehog medulloblastoma
December 14, 2017 - Dr. Reddy’s Announces Approval of Impoyz (clobetasol propionate) Cream for Plaque Psoriasis
December 14, 2017 - Gene Screens Can Alter Perception, Behavior
December 14, 2017 - Can Scrotal Vein Condition Hike Heart Risks?: MedlinePlus Health News
December 14, 2017 - Molecules in spit may be able to diagnose and predict length of concussions
December 14, 2017 - Children’s Colorado and RxRevu partner to help prescribers better meet needs of pediatric patients
December 14, 2017 - Researchers discover new way to attack drug-resistant prostate cancer cells
December 14, 2017 - Scientists develop new, high resolution method for identifying microbial species and strains
December 14, 2017 - Declining trend of salmonellosis cases has leveled off in the EU
December 14, 2017 - Death receptors in the blood can help measure risk of cardiovascular diseases, type 2 diabetes
December 14, 2017 - How to Perk Up the Holidays for Hospital Patients
December 14, 2017 - Prolonged Sedation May be Bad for Baby’s Brain
December 14, 2017 - The pediatric submersion score predicts children at low risk for injury following submersions
December 14, 2017 - Video game helps doctors to quickly recognize trauma patients who need high levels of care
December 14, 2017 - Younger persons newly-diagnosed with type 2 diabetes have poorer health than older patients
December 14, 2017 - Clinician re-examines evidence on re-use of catheters and UTIs in people with spinal cord injuries
December 14, 2017 - UK and Russian researchers join forces against AMR
December 14, 2017 - Results of Bariatric Surgery Hold Up Over Time
December 14, 2017 - High-intensity exercise delays Parkinson’s progression
December 14, 2017 - Protein structure could pave way for effective drugs to treat cystic fibrosis
December 14, 2017 - Minority people less likely to see dermatologist for psoriasis treatment
December 14, 2017 - Study indicates decline in use of chemotherapy for early stage breast cancer patients
December 14, 2017 - Chagas disease presents real public health problem to Canadians

Improving clinical trials with machine learning

image_pdfDownload PDFimage_print

 “Current statistical models are too simple. They fail to capture complex biological variations across people, discarding them as mere noise. We suspected this could partly explain why so many drug trials work in simple animals but fail in the complex brains of humans. If so, machine learning capable of modelling the human brain in its full complexity may uncover treatment effects that would otherwise be missed,” said the study’s lead author, Dr Parashkev Nachev (UCL Institute of Neurology).

To test the concept, the research team looked at large-scale data from patients with stroke, extracting the complex anatomical pattern of brain damage caused by the stroke in each patient, creating in the process the largest collection of anatomically registered images of stroke ever assembled. As an index of the impact of stroke, they used gaze direction, objectively measured from the eyes as seen on head CT scans upon hospital admission, and from MRI scans typically done 1-3 days later.

They then simulated a large-scale meta-analysis of a set of hypothetical drugs, to see if treatment effects of different magnitudes that would have been missed by conventional statistical analysis could be identified with machine learning. For example, given a drug treatment that shrinks a brain lesion by 70%, they tested for a significant effect using conventional (low-dimensional) statistical tests as well as by using high-dimensional machine learning methods.

The machine learning technique took into account the presence or absence of damage across the entire brain, treating the stroke as a complex “fingerprint”, described by a multitude of variables.

“Stroke trials tend to use relatively few, crude variables, such as the size of the lesion, ignoring whether the lesion is centred on a critical area or at the edge of it. Our algorithm learned the entire pattern of damage across the brain instead, employing thousands of variables at high anatomical resolution. We used well-established methods of machine learning, teaching the algorithm on subsets of data and then testing its performance on other subsets it had not seen,” explained the study’s first author, Tianbo Xu (UCL Institute of Neurology).

The advantage of the machine learning approach was particularly strong when looking at interventions that reduce the volume of the lesion itself. With conventional low-dimensional models, the intervention would need to shrink the lesion by 78.4% of its volume for the effect to be detected in a trial more often than not, while the high-dimensional model would more than likely detect an effect when the lesion was shrunk by only 55%.

“Conventional statistical models will miss an effect even if the drug typically reduces the size of the lesion by half, or more, simply because the complexity of the brain’s functional anatomy—when left unaccounted for—introduces so much individual variability in measured clinical outcomes. Yet saving 50% of the affected brain area is meaningful even if it doesn’t have a clear impact on behaviour. There’s no such thing as redundant brain,” said Dr Nachev.

The researchers say their findings demonstrate that machine learning could be invaluable to medical science, especially when the system under study—such as the brain—is highly complex.

“The real value of machine learning lies not so much in automating things we find easy to do naturally, but formalising very complex decisions. Machine learning can combine the intuitive flexibility of a clinician with the formality of the statistics that drive evidence-based medicine. Models that pull together 1000s of variables can still be rigorous and mathematically sound. We can now capture the complex relationship between anatomy and outcome with high precision,” said Dr Nachev.

“We hope that researchers and clinicians begin using our methods the next time they need to run a clinical trial,” said co-author Professor Geraint Rees (Dean, UCL Faculty of Life Sciences).

The study was funded by Wellcome and the National Institute for Health Research University College London Hospitals Biomedical Research Centre.

Tagged with:

About author

Related Articles