Of the quarter- million women diagnosed with breast cancer every year in the United States, about 180,000 undergo surgery to remove the cancerous tissue while preserving as much healthy breast tissue as possible.
However, there’s no accurate method to tell during surgery whether all of the cancerous tissue has been successfully removed. The gold-standard analysis takes a day or more, much too long for a surgeon to wait before wrapping up an operation. As a result, about a quarter of women who undergo lumpectomies receive word later that they will need a second surgery because a portion of the tumour was left behind.
Now, researchers at Washington University School of Medicine in St. Louis and California Institute of Technology report that they have developed a technology to scan a tumour sample and produce images detailed and accurate enough to be used to check whether a tumour has been completely removed.
Called photoacoustic imaging, the new technology takes less time than standard analysis techniques. But more work is needed before it is fast enough to be used during an operation.
"This is a proof of concept that we can use photoacoustic imaging on breast tissue and get images that look similar to traditional staining methods without any sort of tissue processing," said Deborah Novack, MD, PhD, an associate professor of medicine, and of pathology and immunology, and a co-senior author on the study.
The researchers are working on improvements that they expect will bring the time needed to scan a specimen down to 10 minutes, fast enough to be used during an operation. The current gold-standard method of analysis, which is based on preserving the tissue and then staining it to make the cells easier to see, hasn’t gotten any faster since it was first developed in the mid-20th century.
To speed up the process, the researchers took advantage of a phenomenon known as the photoacoustic effect. When a beam of light of the right wavelength hits a molecule, some of the energy is absorbed and then released as sound in the ultrasound range. These sound waves can be detected and used to create an image.
"All molecules absorb light at some wavelength," said co-senior author Lihong Wang, PhD, who conducted the work when he was a professor of biomedical engineering at Washington University’s School of Engineering & Applied Science. He is now at Caltech. "This is what makes photoacoustic imaging so powerful. Essentially, you can see any molecule, provided you have the ability to produce light of any wavelength. None of the other imaging technologies can do that. Ultrasound will not do that. X-rays will not do that. Light is the only tool that allows us to provide biochemical information."
The researchers tested their technique by scanning slices of tumours removed from three breast cancer patients. For comparison, they also stained each specimen according to standard procedures.
The photoacoustic image matched the stained samples in all key features. The architecture of the tissue and subcellular detail such as the size of nuclei were clearly visible.
"It’s the pattern of cells – their growth pattern, their size, their relationship to one another – that tells us if this is normal tissue or something malignant," Novack said. "Overall, the photoacoustic images had a lot of the same features that we see with standard staining, which means we can use the same criteria to interpret the photoacoustic imaging. We don’t have to come up with new criteria."
Having established that photoacoustic techniques can produce usable images, the researchers are working on reducing the scanning time.

Siteman Cancer Center http://tinyurl.com/y87u35l5

Continuously recording the brain’s electrical signals and examining how those impulses evolve over time is a more reliable way to identify infants at risk for brain injury, compared with doing snapshot evaluations, according to a prospective cohort study led by Children’s National Health System research-clinicians.
Amplitude-integrated electroencephalogram (aEEG) is a bedside tool that involves attaching tiny electrodes to the newborn’s scalp to permit clinicians to monitor the complex electrical activity of the child’s brain over time. It’s a positive sign when an aEEG shows babies beginning to sleep and wake normally by the time they are 3 days old. Conversely, severely abnormal aEEG readings in the first days of life predict poor outcomes.
The Children’s team used aEEG with infants born with hypoxic-ischemic encephalopathy (HIE), one of the most severe complications that can affect full-term infants. During pregnancy, birth or shortly after birth, a hypoxic-ischemic event can occur that impedes blood flow and oxygen delivery to the brain, resulting in destruction of brain tissue. Cooling (therapeutic hypothermia) is now standard for newborns with HIE in order to stave off life-long consequences, but deaths and neurodevelopmental disability still can occur.
“We know whole-body cooling─or lowering the body’s temperature by about 3 degrees Celsius─can help vulnerable newborns survive and can protect their brains from suffering profound injuries,” says An N. Massaro, M.D., a Children’s National neonatologist and senior author of the study. “What we were trying to determine with this study is whether evaluating the pattern of evolution of the aEEG as a whole provides more information compared with looking at snapshots in time.”
Eighty infants undergoing therapeutic cooling who met the inclusion criteria were enrolled in the five-year study, one of the largest such studies to date. The babies weighed more than 1,800 grams and were older than 35 weeks’ gestational age at birth, and either needed prolonged resuscitation after birth or had low APGAR scores─ a measure of how well newborns fare outside the womb. Continuous recordings of EEG data occurred from the time of admission up to 12 hours after the infants temperatures were raised to normal and aEEG tracings were calculated.
After the therapeutic cooling blankets were removed, the infants underwent at least one magnetic resonance imaging (MRI) scan prior to discharge. During the routine follow-up check at about 18 months of age, the HIE survivors’ cognitive and motor skills were assessed using validated instruments.
Fifty-six of the infants in the study had favourable outcomes. Twenty-four infants had adverse outcomes, including 15 with severe brain injury detected by MRI and nine infants who died. These children had lower APGAR scores at five minutes, and were more likely to have severe HIE and to have experienced more frequent seizures.
“Infants whose aEEG abnormalities do not improve were at increased risk: Infants who do not reach a discontinuous background pattern by 15.5 hours of life, achieve cycling by 45.5 hours after birth and who fail to achieve continuous normal voltage by 78 hours after birth are most at risk for adverse outcomes,” Dr. Massaro says. “In addition to defining worrisome trends, we found that overall assessment of continuous aEEG readings through the course of hypothermia treatment provide the most meaningful predictive power. This means we can speak with families at the bedside with more confidence about their child’s outcomes after the infant undergoes cooling therapy.”
Children’s National Health System
childrensnational.org/news-and-events/childrens-newsroom/2017/continuous-eeg-better-at-identifying-oxygen-deprived-newborns-most-at-risk

Doctors are often deluged by signals from charts, test results, and other metrics to keep track of. It can be difficult to integrate and monitor all of these data for multiple patients while making real-time treatment decisions, especially when data is documented inconsistently across hospitals.
In a new pair of papers, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) explore ways for computers to help doctors make better medical decisions.
One team created a machine-learning approach called “ICU Intervene” that takes large amounts of intensive-care-unit (ICU) data, from vitals and labs to notes and demographics, to determine what kinds of treatments are needed for different symptoms. The system uses “deep learning” to make real-time predictions, learning from past ICU cases to make suggestions for critical care, while also explaining the reasoning behind these decisions.
“The system could potentially be an aid for doctors in the ICU, which is a high-stress, high-demand environment,” says PhD student Harini Suresh, lead author on the paper about ICU Intervene. “The goal is to leverage data from medical records to improve health care and predict actionable interventions.”
Another team developed an approach called “EHR Model Transfer” that can facilitate the application of predictive models on an electronic health record (EHR) system, despite being trained on data from a different EHR system. Specifically, using this approach the team showed that predictive models for mortality and prolonged length of stay can be trained on one EHR system and used to make predictions in another.
ICU Intervene was co-developed by Suresh, undergraduate student Nathan Hunt, postdoc Alistair Johnson, researcher Leo Anthony Celi, MIT Professor Peter Szolovits, and PhD student Marzyeh Ghassemi. It was presented this month at the Machine Learning for Healthcare Conference in Boston.
EHR Model Transfer was co-developed by lead authors Jen Gong and Tristan Naumann, both PhD students at CSAIL, as well as Szolovits and John Guttag, who is the Dugald C. Jackson Professor in Electrical Engineering. It was presented at the ACM’s Special Interest Group on Knowledge Discovery and Data Mining in Halifax, Canada.
Both models were trained using data from the critical care database MIMIC, which includes de-identified data from roughly 40,000 critical care patients and was developed by the MIT Lab for Computational Physiology.


MIT CSAIL
www.csail.mit.edu/using_machine_learning_to_improve_patient_care%20

Severe acute liver failure (ALF), a rare but life-threatening illness, is associated with high death rates if patients don’t receive timely treatment or a liver transplant. Unlike the heart or the kidneys, there is no established mechanical device to replace the liver’s function. Now, University of Maryland School of Medicine (UM SOM) researchers report that a device that removes toxins from the blood can also effectively provide a bridge to liver transplantation or buy time for a traumatically injured liver to heal, suggesting broader uses for the device than previously thought.
The researchers, present the largest series of cases in the United States in which the Molecular Adsorbent Recirculating System has been used as temporary liver replacement for ALF.
MARS can be likened to a dialysis machine for the liver. It essentially “washes” a patient’s blood with a solution containing albumin – normally produced by healthy livers – to remove toxins such as bile acids, ammonia, bilirubin, copper, iron and phenols from the blood.
“We’ve found in the use of MARS that we’re able to get trauma patients with massive liver injury to recovery and, in patients who are deemed good transplant candidates, get them to transplant with excellent survivals,” says lead researcher, Steven I. Hanish, MD, associate professor of surgery at UM SOM and a liver transplant surgeon at the University of Maryland Medical Center (UMMC).
The US Food and Drug Administration (FDA) has approved the device to clear the liver after overdoses and poisonings, and reduce the effects of brain swelling related to liver failure. However, it is not yet FDA-approved as a bridge to transplant.

University of Maryland School of Medicine http://tinyurl.com/ya32cuxm