Breaking News
November 15, 2018 - EU and industry-funded project aims to personalize diabetes treatment
November 15, 2018 - NIH researchers shed light on causes of HBV-associated acute liver failure
November 14, 2018 - FDA Alert: Implanted Pumps: Safety Communication
November 14, 2018 - Weight loss & acute Porphyria
November 14, 2018 - Researchers identify three sub-types of depression
November 14, 2018 - The puzzle of a mutated gene lurking behind many Parkinson’s cases | News Center
November 14, 2018 - The mystery viruses far worse than flu
November 14, 2018 - Research highlights physical changes in the brain of self-injuring teen girls
November 14, 2018 - Speed and error rate of DNA synthesis influenced by DNA structure
November 14, 2018 - Cranberry consumption modifies impact of animal-based diet on gut health
November 14, 2018 - £500,000 grant could pave way for new antibiotic to battle against drug-resistant superbugs
November 14, 2018 - Trump Administration Finalizes Birth Control Coverage Opt-Out
November 14, 2018 - Modern life offers children almost everything they need, except daylight
November 14, 2018 - Getting better: A patient is more than a collection of numbers
November 14, 2018 - 20 Americans Die Each Day Waiting For Organs
November 14, 2018 - First bifacial molecule can invade double-stranded DNA or RNA
November 14, 2018 - Study finds lack of safety data for using flowers in cooking
November 14, 2018 - Statistical methods play key role in predicting efficacy of new drugs
November 14, 2018 - Health Tip: Limit Fat, Sugar and Salt in Your Child’s Diet
November 14, 2018 - CA 19-9 Blood Test (Pancreatic Cancer): MedlinePlus Lab Test Information
November 14, 2018 - Old drug could have new use helping sick premature babies
November 14, 2018 - Surgery, not antibiotics, should remain first-line treatment for appendicitis | News Center
November 14, 2018 - Researchers to develop sports-specific classification system for blind football
November 14, 2018 - Preschool children show awake responses to naptime nonsense words
November 14, 2018 - Survey shows negative effect of vulvovaginal atrophy symptoms on quality of life for women
November 14, 2018 - Study sheds light on mechanisms that prevent autoimmune attack
November 14, 2018 - Sleep quality found to be worse for women who undergo surgical menopause
November 14, 2018 - One-hour cognitive behavioral therapy session reduces insomnia symptoms in prisoners
November 14, 2018 - New study provides deeper insight into chromosome segregation during mitosis
November 14, 2018 - Surgical menopause leads to more disrupted sleep than natural menopause
November 14, 2018 - Inhibition of one protein clears toxic clumps seen in Parkinson’s disease, study finds
November 14, 2018 - Appendix removal is linked to lower risk of Parkinson’s
November 14, 2018 - Lifting weights for less than an hour a week may reduce cardiovascular disease risk
November 14, 2018 - Pulmonary rehabilitation rarely received by hospitalized COPD patients despite health benefits
November 14, 2018 - New anti-HER2 drug shows promising anti-tumor activity in gullet, stomach and bowel cancers
November 14, 2018 - Regular head circumference assessment of preterm babies can help identify long-term IQ problems
November 14, 2018 - Brigham investigators examine opioid use among Massachusetts adolescents, prescription trends
November 14, 2018 - Study defines biomarker in response to treatment of castration-resistant prostate cancer
November 14, 2018 - Study identifies potential therapeutic strategy for patients with clear cell renal cancer
November 14, 2018 - Bausch Health Announces U.S. Launch of Bryhali (halobetasol propionate) Lotion, 0.01%, for Plaque Psoriasis In Adults
November 14, 2018 - Alpha Fetoprotein (AFP) Tumor Marker Test: MedlinePlus Lab Test Information
November 14, 2018 - Researchers evaluate controversial treatment for Parkinson’s disease psychosis
November 14, 2018 - AI could help veterinarians code their notes
November 14, 2018 - Pre-schoolers with autism thrive in mainstream classroom settings
November 14, 2018 - Individual and work-related factors may help promote hospital physician engagement, finds study
November 14, 2018 - Complementary and alternative medicine is widely used by general population in England
November 14, 2018 - Study reveals link between tobacco availability and smoking during pregnancy
November 14, 2018 - Purdue researchers develop translucent base for silicon patches to deliver exact doses of biomolecules
November 14, 2018 - New technology based on moths and magnets could help treat genetic diseases
November 14, 2018 - Concussion-Related Biomarkers Vary Based on Sex, Race
November 14, 2018 - One more year of high school may shape waistlines later in life
November 14, 2018 - Dissecting high drug costs – Scope
November 14, 2018 - Study shows novel strategy to reduce breast cancer bone metastasis
November 14, 2018 - Empowering the NHS through Industry Partnerships
November 14, 2018 - One size does not fit all in obesity treatment, study finds
November 14, 2018 - Seeking ways to prevent ‘secondary cataracts’
November 14, 2018 - Change Within the Eye May Be Early Warning for Macular Degeneration
November 14, 2018 - Study of 500,000 people clarifies the risks of obesity
November 14, 2018 - Ultrasound releases drug to alter activity in targeted brain areas in rats | News Center
November 14, 2018 - Umass Amherst researchers battle against youth suicide in rural Alaska Native communities
November 14, 2018 - Cancer stem cells depend on amino acid metabolism, and it’s proving to be their Achilles’ heel
November 14, 2018 - Epigenetic link found between prenatal exposure to maternal smoking and offspring’s cardio-metabolic health
November 14, 2018 - Meditation, music may change biomarkers of cellular aging and Alzheimer’s disease in older adults
November 14, 2018 - Multidisciplinaryresearch teams selected to study age-related brain disorders
November 14, 2018 - The Current issue of “The view from here” is concerned with Informatics
November 14, 2018 - Researchers identify tool to help transgender women have a more authentic voice
November 14, 2018 - Four faculty members appointed to endowed professorships | News Center
November 13, 2018 - Research finds strongest evidence yet that obesity causes depression
November 13, 2018 - Researchers compare stools of breastfed and formula-fed infants
November 13, 2018 - Entasis Therapeutics Announces Zoliflodacin Phase 2 Results Published in The New England Journal of Medicine
November 13, 2018 - Gene changes driving myopia reveal new focus for drug development
November 13, 2018 - $6 million grant to support study of preeclampsia, atherosclerosis links | News Center
November 13, 2018 - Beneficial gut microbes metabolize high-fiber diet to improve heart health in mouse model
November 13, 2018 - Excessive use of social media through visual postings linked to increase in narcissistic traits
November 13, 2018 - Study finds why obesity both fuels cancer growth and helps immunotherapy to kill tumors
November 13, 2018 - Women prefer and invest more in daughters, while men favor sons
November 13, 2018 - With hospitalization losing favor, judges order outpatient mental health treatment
November 13, 2018 - Transgenic rat model may provide new insights into cerebral amyloid angiopathy
November 13, 2018 - Study identifies factors tied to greater risk of advanced liver disease in cystic fibrosis patients
November 13, 2018 - Risk of blindness among premature babies with low levels of blood platelets
New automated model identifies dense breast tissue in mammograms

New automated model identifies dense breast tissue in mammograms

image_pdfDownload PDFimage_print

Researchers from MIT and Massachusetts General Hospital have developed an automated model that assesses dense breast tissue in mammograms — which is an independent risk factor for breast cancer — as reliably as expert radiologists.

This marks the first time a deep-learning model of its kind has successfully been used in a clinic on real patients, according to the researchers. With broad implementation, the researchers hope the model can help bring greater reliability to breast density assessments across the nation.

It’s estimated that more than 40 percent of U.S. women have dense breast tissue, which alone increases the risk of breast cancer. Moreover, dense tissue can mask cancers on the mammogram, making screening more difficult. As a result, 30 U.S. states mandate that women must be notified if their mammograms indicate they have dense breasts.

But breast density assessments rely on subjective human assessment. Due to many factors, results vary — sometimes dramatically — across radiologists. The MIT and MGH researchers trained a deep-learning model on tens of thousands of high-quality digital mammograms to learn to distinguish different types of breast tissue, from fatty to extremely dense, based on expert assessments. Given a new mammogram, the model can then identify a density measurement that closely aligns with expert opinion.

“Breast density is an independent risk factor that drives how we communicate with women about their cancer risk. Our motivation was to create an accurate and consistent tool, that can be shared and used across health care systems,” says second author Adam Yala, a PhD student in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL).

The other co-authors are first author Constance Lehman, professor of radiology at Harvard Medical School and the director of breast imaging at the MGH; and senior author Regina Barzilay, the Delta Electronics Professor at CSAIL and the Department of Electrical Engineering and Computer Science at MIT.

Mapping density

The model is built on a convolutional neural network (CNN), which is also used for computer vision tasks. The researchers trained and tested their model on a dataset of more than 58,000 randomly selected mammograms from more than 39,000 women screened between 2009 and 2011. For training, they used around 41,000 mammograms and, for testing, about 8,600 mammograms.

Each mammogram in the dataset has a standard Breast Imaging Reporting and Data System (BI-RADS) breast density rating in four categories: fatty, scattered (scattered density), heterogeneous (mostly dense), and dense. In both training and testing mammograms, about 40 percent were assessed as heterogeneous and dense.

During the training process, the model is given random mammograms to analyze. It learns to map the mammogram with expert radiologist density ratings. Dense breasts, for instance, contain glandular and fibrous connective tissue, which appear as compact networks of thick white lines and solid white patches. Fatty tissue networks appear much thinner, with gray area throughout. In testing, the model observes new mammograms and predicts the most likely density category.

Matching assessments

The model was implemented at the breast imaging division at MGH. In a traditional workflow, when a mammogram is taken, it’s sent to a workstation for a radiologist to assess. The researchers’ model is installed in a separate machine that intercepts the scans before it reaches the radiologist, and assigns each mammogram a density rating. When radiologists pull up a scan at their workstations, they’ll see the model’s assigned rating, which they then accept or reject.

“It takes less than a second per image … [and it can be] easily and cheaply scaled throughout hospitals.” Yala says.

On over 10,000 mammograms at MGH from January to May of this year, the model achieved 94 percent agreement among the hospital’s radiologists in a binary test — determining whether breasts were either heterogeneous and dense, or fatty and scattered. Across all four BI-RADS categories, it matched radiologists’ assessments at 90 percent. “MGH is a top breast imaging center with high inter-radiologist agreement, and this high quality dataset enabled us to develop a strong model,” Yala says.

In general testing using the original dataset, the model matched the original human expert interpretations at 77 percent across four BI-RADS categories and, in binary tests, matched the interpretations at 87 percent.

In comparison with traditional prediction models, the researchers used a metric called a kappa score, where 1 indicates that predictions agree every time, and anything lower indicates fewer instances of agreements. Kappa scores for commercially available automatic density-assessment models score a maximum of about 0.6. In the clinical application, the researchers’ model scored 0.85 kappa score and, in testing, scored a 0.67. This means the model makes better predictions than traditional models.

In an additional experiment, the researchers tested the model’s agreement with consensus from five MGH radiologists from 500 random test mammograms. The radiologists assigned breast density to the mammograms without knowledge of the original assessment, or their peers’ or the model’s assessments. In this experiment, the model achieved a kappa score of 0.78 with the radiologist consensus.

Next, the researchers aim to scale the model into other hospitals. “Building on this translational experience, we will explore how to transition machine-learning algorithms developed at MIT into clinic benefiting millions of patients,” Barzilay says. “This is a charter of the new center at MIT — the Abdul Latif Jameel Clinic for Machine Learning in Health at MIT — that was recently launched. And we are excited about new opportunities opened up by this center.”

Source:

http://news.mit.edu/2018/AI-identifies-dense-tissue-breast-cancer-mammograms-1016

Tagged with:

About author

Related Articles