Skip to Navigation
University of Pittsburgh
Print This Page Print this pages

December 10, 2015

Research Notes

Info sciences prof wins grant

School of Information Sciences (SIS) faculty member Kostas Pelechrinis has won a one-year grant from the Army Research Office for his project, “Models and Metrics for Socio-Spatial Composite Networks.”

His research will explore the use of tensor theory to represent and analyze composite network structures that involve multiple types of nodes, edges and/or time evolution. It aims to develop methods that will enable the analysis of heterogeneous networks.

Said Pelechrinis: “Currently we have a good understanding of and tools for the analysis of homogenous networks, networks that describe a single type of relationship between a single type of entity. However, in the era of big data, more complex structures appear that involve more than one type of entity and more than one type of relationship.”

Heterogeneous networks can include people, locations and companies all interacting on various levels. These types of relationships and connections do not currently have proven methods for analyzing all of the information within them together. Most researchers will choose to focus on just one part of the network, thus contributing to information loss. Pelechrinis’ hope is that this project will help create the foundations of a working methodology for analyzing these composite networks.

“Our goal is not to solve all the problems in the space — this would also be rather unrealistic,” said Pelechrinis. “We hope that we will stimulate further research in the area, which will lead to faster development of the required methods.” Having such a tool can be beneficial to researchers in a variety of fields, as heterogeneous networks appear in many different settings. “Even though networks in a different context might require specialized methods or metrics, we hope that we will be able to provide a basic set of tools for the analysis of heterogeneous networks regardless of the domain and provide insights for developing specialized methods when needed.”

Pelechrinis became interested in pursuing this research after working on a different project that involved the analysis of social and spatial information in social media. He discovered that there was a gap in the literature concerning heterogeneous networks.


Telepathology consultations shown to benefit patients

International telepathology consultations can improve patient care significantly, according to a study by Pitt and KingMed Diagnostics researchers. Their review of more than 1,500 pathology cases submitted electronically to UPMC over three years found that consultation with UPMC pathologists resulted in significantly altered treatment plans for more than half of the cases in which a patient’s primary diagnosis had been provided from referring hospitals in China.

Published in the Journal of Pathology Informatics, the study supports a growing view among health care experts that telemedicine is an effective way to expand access to specialists and other needed medical services.

Said lead author Chengquan Zhao, pathology faculty member in the School of Medicine: “The results of this study indicate, for the first time, that international telepathology can improve patient care in China by facilitating access to pathology expertise.” Previous studies of telepathology have focused more on the technology than on the clinical experience, noted Zhao, who also is with Magee Womens Hospital.

Telepathology refers to the remote practice of pathology through transmission of digital images. Since 2012, UPMC has provided telepathology consultation to KingMed Diagnostics, a network of 27 central laboratories serving more than 13,000 hospitals and clinics in China. Glass slides are scanned in China, and the high-resolution digital slides then are presented to subspecialty pathologists at UPMC in Pittsburgh through a customized, web-based portal. “It is easier and faster to move an image than it is to move a patient or the glass slide containing a patient’s tissue specimen,” said Zhao.

In this retrospective study, the researchers examined 1,561 cases submitted to UPMC for telepathology consultations from January 2012 to December 2014. Hematopathology received the most cases (23.7 percent), followed by bone/soft tissue (21 percent) and gynecologic/breast (20.2 percent) subspecialties. In 82.4 percent of the cases, UPMC pathologists provided a definite diagnosis, usually for malignancies that were rare or difficult to diagnose.

Among 855 cases in which a primary diagnosis was provided by the referring Chinese hospital — usually in smaller to mid-sized cities or rural areas — the final diagnoses from UPMC pathologists were identical in 25.6 percent of cases and different or considerably modified in 50.8 percent of cases. Initially, most of the requests for a second opinion came from pathologists in China, but over time those requests came from treating clinicians and even patients themselves, the researchers noted, indicating the value placed on obtaining an expert pathology opinion. Also of note, the average turnaround time per case fell from 6.8 days in 2012 to five days in 2014.

To explain the significant differences between the final and primary diagnoses, Zhao noted that “our study was not intended to compare diagnostic capabilities between pathologists in China and the U.S., but rather to demonstrate that telepathology is a convenient mechanism to assist those seeking consultation. Most of these cases referred for telepathology consultation were from small or mid-sized hospitals. Hence, the primary diagnostic accuracy in this study does not fully represent the diagnostic skill of all Chinese pathologists.” In addition, he said, immunohistochemical studies requested to subsequently work up cases submitted for teleconsultation often were helpful in the final diagnosis.

“The success of this international digital consultation service was dependent on strong commitment and support from leadership, information technology expertise and dedicated pathologists who understood the language and culture on both sides,” said Zhao. In addition, the web portal was customized to meet the needs of both the referring site and consulting pathologists, including support for bi-directional discussions.

Although the researchers noted some obstacles — such as Internet speed, cultural and language differences, and lack of clinical information in some cases — the results of the UPMC-KingMed partnership should encourage more telepathology practice, they concluded.

In addition to Zhao, the study’s Pitt/UPMC co-authors were Anil V. Parwani, Jeff McHugh, Anthony Picolli, Gonzalo Romero Lauro, Douglas J. Hartman, Raja R. Seethala, Samuel Yousem and Liron Pantanowitz.

The study received funding from Omnyx, a digital pathology company jointly owned by UPMC and GE Healthcare.


Big data may predict cancer outcomes

Big data researchers have received a $5 million, three-year Commonwealth Universal Research Enhancement  (CURE) grant to develop better methods for integrating, analyzing and modeling large volumes of diverse data on cancer patients. The goal is to produce more accurate predictions of patient outcomes and to enable clinicians to tailor care for each patient.

Gregory Cooper, vice chair of biomedical informatics, and Ziv Bar-Joseph at Carnegie Mellon University will lead the Big Data for Better Health (BD4BH) project, which also includes UPMC and the Pittsburgh Supercomputing Center.

Said Cooper: “We will investigate breast and lung cancer as clinical domains to develop the methods and software tools; however, the methods will be generalizable to other diseases. The basic approach will be to process raw data, such as gene sequence and expression data, to derive highly informative biological patterns in the data that are then used to predict patient outcomes. We believe these biological patterns will predict outcomes significantly better than would using the raw data directly, and we plan to test this hypothesis.”

For example, rather than using the set of mutated genes, a type of raw data, in a cancerous tumor as predictors of cancer metastasis, they will infer the cell signaling pathways, a type of biological pattern, that are likely having a significant influence on tumor growth. Those aberrant pathways then will be used to predict clinical outcomes, such as tumor spread or metastasis. The ultimate goal is for such predictions to help inform clinical care.

Machine learning methods, for instance, can analyze large datasets automatically to discover patterns that people cannot discern. These automated discoveries then can enable researchers to identify relationships between the way specific individuals respond to treatment and their DNA to allow more personalized tailoring of treatments.

“We hope that more accurate predictions of clinical outcomes will assist physicians in devising treatment plans and help patients in making health care decisions,” Cooper said. “The patterns contained in the models also may spur biological insights into the diseases being modeled.”

In collaboration with Lincoln University, the BD4BH program also will train underrepresented minority students to work with big data in both the biomedical and data science realms.
The project is funded by the Pennsylvania Department of Health.


Text messages cut binge drinking

A School of Medicine-led trial to test a text message-based program aimed at reducing binge drinking is the first to show that such an intervention can produce sustained reductions in alcohol consumption in young adults.

The findings revealed that the first-of-its-kind program, designed by lead author Brian Suffoletto, emergency medicine faculty member, reduced binge drinking and alcohol-related injuries when compared to a control group and a self-monitoring group. The positive effect continued six months after the program ended. The results have been published in the journal PLOS ONE and are available online.

Said Suffoletto: “Given the low cost to send text messages and the capacity to deliver them to almost every at-risk young adult, a text message-based intervention targeting binge drinking could have a public health impact on reducing both immediate and long-term health problems.”

The 12-week trial randomized into three groups 765 people ages 18-25 who were discharged from four urban emergency departments in western Pennsylvania. The control group received standard care and no text messages. The self-monitoring group received text messages on Sundays asking about drinking quantity but received no feedback. The final group received the full program, which consisted of text messages on Thursdays inquiring about weekend drinking plans and promoting a goal commitment to limit drinking, followed by another text on Sunday to inquire about actual drinking and give tailored feedback aimed at reducing alcohol consumption.

If someone receiving the full intervention program reported anticipating a heavy drinking day (more than five drinks during any 24-hour period for men and more than four for women), he or she received a text message expressing concern about those levels and asking if they would be willing to set a goal to limit their drinking below binge thresholds for the weekend. Those who responded in the affirmative then received messages expressing positive reinforcement and strategies for cutting down. Those who refused to set goals received a text message encouraging them to reflect on the decision (for example, “It’s OK to have mixed feelings about reducing your alcohol use. Consider making a list of all the reasons you might want to change”).

Six months after the end of the trial, participants who were exposed to the full text-message intervention reported an average of one fewer binge-drinking day per month. There also was a 12-percent reduction in binge drinking. The control group and the self-monitoring group both had no reductions in alcohol consumption.

Suffoletto said: “Compared to in-person interventions with a clinician discussing drinking habits with a young adult in the emergency department, which requires time and resources not routinely available, suffers high variability in how it is performed, and has shown limited ability to produce lasting reductions in hazardous drinking among young adults, our text message-based intervention is scalable, provides uniform behavioral materials and seems to produce meaningful, potentially life-saving results. By interacting with these young adults in a way in which they are receptive to communicating, and reducing the stigma associated with traditional face-to-face counseling, text messages can provide the boost they need to control their drinking when they are at their most vulnerable to forget what is healthiest for them.”

Pitt co-authors of the study were Jeffrey Kristan, Tammy A. Chung, Kwonho Jeong, Anthony Fabio and Duncan B. Clark. A colleague from Brown University also contributed.

The study was supported by the Emergency Medicine Foundation.


PInCh picks 6 teams

The 2015 Pitt Innovation Challenge (PInCh) competition awarded $375,000 to six teams that developed fresh health-care ideas. The awards, begun by the Office of the Provost, Clinical and Translational Science Institute and Innovation Institute, offer research assistance to proposals from teams comprising at least one Pitt faculty member. Teams can include non-Pitt members.

Three top prizes of $100,000 each went to:

Michael Balsan, School of Medicine; Jeanne Burns; Fred Kimock; and Emily Hirsch for NEATCAP, a hearing protection device for babies in neonatal intensive care units that blocks noxious noise and aims to reduce infant stress, improve sleep and promote brain development.

Noah Snyder, Andrew Glowacki, James Eles, Kasey Catt and Zhanhong Du, all of the Swanson School of Engineering, for OXI-Dent, a coating developed to solve the problem of dental implant inflammation.

Yadong Wang, Daniel Long, Eric Jeffries, Piyusha Gade, Jennifer Zhuang, Hassan Awada and Chelsea Stowell, all of engineering, and Prabir Roy-Chaudhury, Terry Kim, Begona Campos-Naciff, Diego Celdran-Bonafonte and Robert Allen for Phoenix, a bio-inductive vascular graft that combines the wide applicability of an off-the-shelf prosthetic graft with the in-vivo performance of an arteriovenous fistula.

Three $25,000 awards went to:

Jeffrey Gusenoff, Beth Gusenoff and Kacey Marra of medicine for Feet for Life, which uses a person’s fat to cushion painful feet so they can walk with less pain and suffering.

Richard Debski, engineering, and Volker Musahl, medicine, for PIVOT, a new technology that quantifies the pivot shift test to evaluate knee ligamentous injuries. The pivot shift test can help predict future development of osteoarthritis.

Judy Cameron, Neal Ryan, Dana Bovbjerg and Meryl Butters, medicine; Jeanette Trauth, Graduate School of Public Health; Tom Akiva, School of Education; John Wallace, School of Social Work; Jennifer Brach, School of Health and Rehabilitation Sciences, and Martica Hall, Shannah Tharp-Gilliam and Becky Mercatoris for Working Together for Kids: A Community Solution. This community training program teaches senior citizens, adults and teens how to provide young children in underserved communities with critical skills for school success, and to assess the health benefits to all involved.

PInCh winners were judged on the impact of the problem, innovation and feasibility of the proposed solution and their technical approach.


Study looks at Alzheimer’s biomarkers in those with Down syndrome

With the help of about $12.5 million over five years from the National Institute on Aging, part of the National Institutes of Health (NIH), Pitt Alzheimer’s disease experts will lead a multicenter effort exploring the progression of biomarkers of the disorder and cognitive function in people with Down syndrome.

Adults with Down syndrome typically show symptoms of Alzheimer’s by their 40s and have a 70-80 percent chance of developing clinical dementia by their 60s. People with Down syndrome have three copies of chromosome 21, each containing a copy of the gene associated with the precursor protein for beta-amyloid, which is found in excess in the brains of Alzheimer’s patients.

Said co-principal investigator Ben Handen, psychiatry faculty member in the School of Medicine: “It’s apparent from Down syndrome patients that amyloid deposition is not sufficient to produce dementia. These individuals have these deposits for more than a decade before cognitive decline is apparent. Understanding the relationships between Down syndrome, disease biomarkers and cognitive decline is critically important to help us design better therapies for all people with Alzheimer’s.”

In the Neurodegeneration in Aging Down Syndrome (NiAD) Study, researchers at four sites in the United States and the United Kingdom will monitor for five years 180 people older than 25 with Down syndrome and 40 people without the condition for factors including amyloid in blood and cerebrospinal fluid and on PET scans; structural and functional MRI scans; and cognitive/functional measures.

NiAD is part of the NIH’s Biomarkers of Alzheimer’s Disease in Adults with Down Syndrome Initiative, which supports teams of researchers using brain imaging and fluid and tissue biomarkers in research that may one day lead to effective interventions for all people with dementia.

The other research sites are the University of Wisconsin-Madison; Banner Alzheimer’s Institute in Phoenix; and Cambridge University.

—Compiled by Marty Levine


The University Times Research Notes column reports on funding awarded to Pitt researchers as well as findings arising from University research.

We welcome submissions from all areas of the University. Submit information via email to:, by fax to 412/624-4579 or by campus mail to 308 Bellefield Hall.

For detailed submission guidelines, visit “Deadlines” page.

Leave a Reply