Cancer Researchers, Colleagues Moving Forward With COVID-19 Projects

UW Carbone's David Beebe, PhD, is part of a team working on a different approach to COVID-19 testing. 

UW Carbone's David Beebe, PhD, is part of a team working on a different approach to COVID-19 testing.


As COVID-19 continues to impact just about every aspect of everyday life, health researchers across the country are engaging in projects to better understand - and treat - this rapidly spreading public health threat.


Researchers at the University of Wisconsin and UW Carbone Cancer Center are no exception. And they're eager to help. Recently, the National Cancer Institute announced it would be offering up one-time grant supplements of up to $250,000 for projects targeting COVID-19, and the virus that causes it, known as SARS-CoV-2.


In a matter of days, nearly 25 proposals came forward from UW Carbone members and colleagues from across campus.


Each cancer center was able to nominate two projects for funding consideration from the NCI. Research projects did not need to be limited to cancer patients, and applicants didn't need to be limited to cancer researchers. While many of the two-dozen proposals will continue in some form, these two were selected to move forward.




Accelerated COVID-19 Testing Enabled by an Integrated RNA Extraction and Detection Platform


The availability of COVID-19 tests in recent weeks has been limited, and demand is only expected to increase going forward.


Many researchers, including UW Carbone's David Beebe, PhD, are working on developing a viable alternative.


Beebe, a professor of Pathology & Laboratory Medicine and Biomedical Engineering at UW-Madison, is an expert in microfluidics, with experience developing low cost assays for infectious disease. A couple of weeks ago, David O'Connor, PhD, and Thomas Friedrich, PhD, two other UW professors and researchers that have been studying various aspects of COVID-19, contacted him with a problem. They had an idea for a different approach to COVID-19 testing, and they had already made progress, but they realized they would likely need a way to concentrate the sample to be successful – an area Beebe specializes in.


"We very much are taking an alternative and complementary approach to the current testing," Beebe said. "If it works, it will not put additional pressure on the supply chain and would expand the number of tests that can be run."


Both O'Connor and Friedrich had been trying an alternative method of COVID-19 testing using an isothermal amplification method to detect viral RNA in swab samples. Typical COVID-19 tests have relied upon what's known as polymerase chain reaction, or PCR, which is standard for this kind of testing. However, demand for the necessary lab equipment and supplies is high.


Using isothermal amplification, which doesn't require cycling a sample through numerous hot and cold cycles, offers an alternative, but in early testing of this method by O'Connor and Friedrich for COVID-19, the sensitivity of the test wasn't as high as it needed to be.


That's where Beebe's lab at UW comes in. By using a combination approach of exclusion-based sample preparation (ESP) and exclusive liquid repellency (ELR), his team should be able to better concentrate viral RNA from a complex mixture found in a swab sample. "By adding that sample prep step, which will only take a minute, we can increase the sensitivity of the test," he said.


Beebe is also exploring a different sample prep approach through work at his private company, Salus Discovery, to give the collaborative group the best chance of success.


Both approaches combine extraction and detection into an integrated workflow. The result is a high throughput, low cost test that can be quickly and broadly implemented, using instruments that are available in nearly all clinical labs. Beebe estimates that about 100 samples could be tested in an hour using this process.


Implementation of this test will begin alongside the current COVID-19 testing method to ensure accuracy. The next step would be to start to use the new test in a research capacity to understand how the virus is spreading in the community.


"We need to ideally be testing everybody a lot more, but until diagnostic testing is more universal there is a need to understand how many people are being missed with current guidelines," Beebe said.



Using machine learning to produce accurate algorithms for early detection, risk stratification and treatment recommendations in patients with COVID-19


Matthew Churpek, MD, MPH, PhDRapid spread, limited testing availability, and lack of evidence regarding how to best treat hospitalized COVID-19 patients have all combined to create massive issues for doctors and health professionals across the country.


To ease that burden, and potentially improve outcomes for patients, some researchers are turning to algorithms and data to help clinicians make the best decisions possible.


Matthew Churpek, MD, MPH, PhD, is drawing on his experience with machine learning to attempt to predict which hospitalized patients are most likely to be diagnosed with COVID-19, and how they can best be treated. Churpek is an associate professor in the Division of Allergy, Pulmonary and Critical Care Medicine at the UW School of Medicine and Public Health. He's working with Mark Craven, PhD, a UW Carbone member and professor in the Department of Biostatistics and Medical Informatics.


"Our proposal will provide clinicians with tools to better identify, risk stratify, and treat patients with COVID-19, which will allow the delivery of early, life-saving care to these patients," Churpek said. "This work will also provide insights into the natural history of disease in hospitalized patients with and without cancer."


Using granular-level data from electronic health records, or EHR, the researchers hope to create automated algorithms that could predict whether a hospitalized patient will test positive for COVID-19. Knowing this earlier, Churpek says, would allow doctors to better prioritize the limited resources they find themselves currently facing.


The proposal would also look to develop risk-stratification tools that would bring critical care resources to the high-risk patients that need them most. Churpek also hopes to develop novel algorithms for personalized treatment recommendations. The idea is that machine learning methods could be used to effectively estimate the effect of a specific treatment, based on a patient's situation.


Follow UW Carbone Cancer Center


Twitter icon Follow UW Carbone on Twitter

Twitter icon Follow UW Carbone on Facebook

Date Published: 04/08/2020

News tag(s):  cancerAdvancescancer researchresearch

News RSS Feed