Photo: Jan Doerler
Thanks in large part to local funding from the S.D. Ireland Foundation, Dr. David Krag has built a world-class lab.
By the time of our late afternoon interview, David Krag, S.D. Ireland Professor of Surgical Oncology at the University of Vermont (UVM), needs a pick-me-up, so off we go to hunt down a coffee stand. Krag, a tall, lean man, moves through the labyrinthine Given Building with the barely suppressed energy of a child after dessert. As we take the stairs down four flights he begins – without waiting for a question – by explaining that the mortality rate from breast cancer could drop by 30 percent, if we just ensured that existing “best practices” were followed in all instances. By the time we hit the coffee stand he has moved from the need to systematize care, to the difficulty today of managing the quantity of medical information on even the smallest of subjects, let alone processing it as a physician and providing the most up-to-date care for patients. He pauses to order a large French roast and a double espresso, both for himself. “I get a little tired in the afternoons,” he explains, almost apologetically.
From Research Bench to Bedside
Krag moved to Vermont in 1991 as the state’s first trained surgical oncologist. He received his degree at Loyola Stritch School of Medicine in Chicago and went on to complete a residency in general surgery at the University of California-Davis and a fellowship in surgical oncology research at the John Wayne Cancer Center in Los Angeles.
Shortly after Krag’s arrival, a group of general surgeons all retired who had worked extensively with breast cancer patients. “One of the important things with breast cancer is that many women will have lumpectomies. They are at a higher risk of developing a recurrence, and that needs to be picked up and treated promptly for a conservative approach to be as effective, long-term, as a mastectomy. With these people gone, nobody was minding the shop,” he said.
So in the early 1990s, Krag, in collaboration with staff from other departments including pathology, radiology, and medical oncology, created one of the nation’s first breast cancer centers to help facilitate coordinated care for breast cancer patients. “There were women out there who were not being helped and there was no plan,” Krag said. Today, the Breast Care Center at Fletcher Allen serves more than 4,000 patients a year.
Identifying a problem and collaborating across disciplines to create sustainable solutions is the hallmark of Krag’s working style, which is now often described as translational research. “You’ve got all these PhDs, brilliant people doing amazing work, but it’s not getting to patients. Then you have superb clinicians and they don’t have a clue about the world of basic science research. You need people to translate the languages,” Krag explained. “It’s a term that describes working bench to bedside. Really, I’ve been doing this my whole career and now there’s a name for it.”
Sentinel Node Biopsy and IHC
The results of Krag’s own research, as well as the process he has developed to help move experimental techniques into the mainstream, have consistently broken new ground in breast cancer treatment. He is best known for his work on sentinel node biopsy, a significantly less invasive treatment than the traditional axillary node dissection for cancers affecting the lymphatic system.
“The goal of [axillary node dissection] surgery is removing all the lymph nodes and, because there is no real anatomic boundary, you just take the whole fat pad,” Krag said, pointing to a slide that demonstrated an average size of tissue removal. By analyzing the nodes, physicians can determine if the cancer has begun to metastasize, and if so, suggest additional treatment with radiation and chemotherapy.
Image courtesy of Dr. David Krag
This slide illustrates the dramatic difference in size between the traditional axillary node dissection's
removal of the entire lymphatic fat pad (in this case almost six inches long) and sentinel node biopsy
(here, a single node less than one inch in length is held by the tweezers).
“The problem with taking all that out, aside from being a sizable surgery, is that there are nerves there and all the lymphatics that drain the arm as well, and now they are blocked,” Krag said. According to the National Cancer Institute (NCI), more than 80 percent of women receiving axillary node dissection have at least one complication after the surgery of varying severity, most commonly lymphedema, swelling of the arm caused by excess fluid buildup; numbness; burning sensations; infection; and a limited range of movement in the shoulder.
In the early 1990s, researchers began experimenting with injecting melanoma patients with blue dye in order to identify which nodes drained the relevant site, known as sentinel nodes. However, this method still required lifting large patches of skin to visually follow the course of the dye. “If you look at the veins on the back of your hand, the pattern is going to be different from my hand to your hand to everybody else’s hand. The lymphatics are like that. I know they are there, but their exact pathway is different,” Krag explained. He pioneered the use of injecting a radioactive tracer and using a modified Geiger counter to identify sentinel nodes, requiring a much smaller incision of just a few inches. “The only option [for melanoma] had been a humongous surgery or doing nothing. This was a way to peek.”
Sentinel node biopsy has since become a standard of care for a wide range of cancers – “basically anything that affects the lymphatic system,” Krag noted – but physicians don’t know what its effects are on long-term survival rates compared with axillary node dissection. For that reason, in 1998 Krag spearheaded a study following more than 5,500 women at 80 centers across North America over the course of 15 years to identify a difference in survival as low as two percent. Preliminary results from this study are expected within the next two to four years.
Image courtesy of Dr. David Krag
This diagram illustrates the relationship between the sentinel nodes
(shown in detail on the right) and the larger lymphatic system for the breast area (on the left).
Krag also embedded a second study within this one, to try to determine the correlation between the presence of individual cancer cells in otherwise negative nodes and patients’ survival rates. In a standard procedure to identify metastasized tumors, a pathologist cuts the lymph node in half, removes a thin slice from the cut edge and treats it with hematoxylin–eosin (H&E) dyes. Comparing an egg to a lymph node with cancer, Krag noted that you could slice through an egg several times before hitting the yolk. But with as many as 15 or 20 nodes in a single axillary node dissection, however, time and resources limited pathologists to the single slice.
For Krag’s study, all nodes were cut at intervals of two to three centimeters and examined with H&E. Negative nodes were then sent to his UVM lab to be treated with additional cuts and a different procedure known as immunohistochemistry (IHC). IHC uses antibodies to bind to cytokeratin, a cellular component common in breast cells that should not be part of a normal lymph node, thus identifying individual cancer cells that have already moved from the breast.
IHC is not standard because of the additional expense and complexity, according to Krag, but also because physicians don’t have any data on the impact of individual cancer cells in the lymph nodes as compared with tumors. “We know that if there is a metastasis of a certain size, it portends a worse survival, but one cell – we don’t really know what that means,” Krag said. He estimated that this secondary study will examine 4,000 H&E negative nodes, identifying approximately 600 as positive through IHC to use as a subset for determining if there is any correlation between the presence of individual cancer cells and long-term survival rates.
Creating and Funding the Research Cycle
In the course of studying sentinel node biopsy, Krag developed a system to more efficiently move his research from concept to standard of care that he has since replicated with other topics. First, he conducts small in-house clinical trials to establish seed data supporting his hypothesis. Trials then expand to include a handful of other regional medical centers. Krag uses the regional results to secure funding for creating national trials.
The realities of funding scientific research also contributed to Krag’s model. Over the last 15 years foundations have decreased the number of studies they support, largely due to cutbacks in federal funding. “As the money gets smaller, the competition heats up and they want not only a good idea, but preliminary results to the point where, the joke is, you have to have done the research to get the funding for it. And that’s not far off the mark,” Krag said. “But the big question is, where do you get funds for the first step?”
The answer for Krag came from a former patient, Scott Ireland, who was one of the first individuals to receive a sentinel node biopsy for melanoma in 1992. Ireland, with his wife Kim [see profile on page 21], donated $2 million dollars to create the S.D. Ireland Professor of Surgical Oncology endowment at UVM in 1999. “I view this gift as an investment that will pay off in many ways,” Krag said at the time. “It frees me up to focus more on research [and] ensures that there will be an individual focused on the problem of cancer forever – or at least as long as there is a university and as long as cancer is a problem.”
The Irelands also created the S.D. Ireland Foundation for Cancer Research, funneling 100 percent of local donations and profits from fundraising efforts directly to Krag’s research facility at UVM. “With three young children, we have a strong interest in education. We also like to be associated with cutting-edge initiatives – that’s why we wanted to support the cancer research,” Scott Ireland states on the foundation’s website (sdirelandcanceresearch.org). Krag credits the Irelands’ support for allowing him to continue focusing on innovative research at the earliest stages. In a rough analysis of the relationship between donated funds and grant monies, he calculated that every $1 in donations led to an additional $20 from the federal government. “That’s a phenomenal bang for the buck,” he said. “And that’s where unrestricted funds are so valuable.”
In addition to enabling the earliest stages of research and attracting more federal support, Krag also noted that local donations help mitigate the boom-bust cycle typical of grant-funded work. A business that suddenly had no income for several months or a year would shut down, he said by way of example. And even if that business owner pooled all of her resources to get a new contract and reopened in six months, she would still have to hire and train new staff. “Having sustaining funds has far more implications than keeping that momentary research alive. It really speaks to a collection of intellectual capacity as a researcher,” Krag said. He pointed to his research staff and facility built over the last decade and a half as something that would not be feasible with inconsistent support.
Sustained funding also buffers the strain of continual job insecurity that can affect staff productivity. “It’s hard to overemphasize the value of that emotional stability so that people can remain highly committed and want to come to work and solve these enormous problems. The group we have here would not be as successful as we are without that.”
Even research in the latest stages risks losing funds midstream, given the time between launching a study and releasing results. Krag’s sentinel node study on long-term survival, for example, opened in 1998 but even preliminary results won’t be available for two to four more years. “If they suddenly cut off funding we would never analyze the data,” Krag said. “There are probably ten people just managing the data, so if the funding were cut, we’d just never know.”
In fact, Krag’s grant application to NCI to continue that study was rejected twice in recent years as the funding cut-off receded from the top 25 percent of applications to 16 and finally to just 8 percent. Krag negotiated with NCI based on the potential significance of his results, but he is not optimistic about his abilities to do so in the future. “I can’t fathom what the difference is between a grant that scores 8 percent and one at 14 percent. It’s a whim,” he said.
The other major evolution in funding research is the inclusion of advocates on decision-making boards, which, Krag noted, began in the late 1980s as a result of breast cancer activists at the federal level. “We’ve had this traditional system that was, and still is, an old boys club. Then, you had these women who marched on Congress, got a bunch of money and said we are going to participate in disbursement – that was different,” he recalled. “Now, you don’t have committees without advocates and I think you have to give credit to them that we have so much attention paid to translational issues, moving it from the lab to the clinic.”
Passing the Daughter Test
In addition to overseeing the near completion of the Phase III sentinel node and IHC study, Krag is also pursuing research to create a personalized delivery method for chemotherapy, analyze bone marrow for early evidence of metastasis, and examine blood protein profiles to potentially develop a blood test for breast cancer.
Describing himself as a both a problem solver and the guy in class who isn’t afraid to ask the “dumb” question everyone is thinking, Krag said he is currently considering moving out of the lab, at least partially, to devise a more efficient system for sharing new data between researchers, practitioners, and patients. “If you look up how many articles were published in the past year on sentinel nodes in the breast alone, it is over 350, and it has been like this for years. Who is reading those? Who is keeping up with it? Who is managing that huge amount of information?” he said. “If I can’t do it with a commitment to it, resources, and a higher level of expertise on that narrow topic than the average person, then who can? I am right at the point where I can’t keep up, where I have to stop or consolidate certain projects or take time out to solve this problem.”
Krag, in collaboration with bioinformatics researchers at UVM, is pursuing funding for one method of managing and disseminating cutting-edge medical data to researchers, clinicians, and patients, addressing the last two years of published research on sentinel node biopsies and breast cancer. Despite refocusing, however, Krag won’t be dropping any existing clinical trials. “I’m not going to give those up,” he said. “The beauty of doing [research] in-house first is: if it works, it’s expandable; if it doesn’t, then you don’t.”
While the shift from treatment to data management may seem sudden, it is of a piece with the trajectory of Krag’s career, one marked by pushing ideas that lie beyond the view of conventional wisdom. “I was at a meeting and one of the doctors said, this doesn’t pass the mom test ,” Krag said. “Well, to not do it doesn’t pass the daughter test, meaning if, when our daughters are old enough to get breast cancer, we are still doing the same thing, then we have failed. We must do something different; otherwise we are just hurting people, and it’s amazing how little things can change unless we have that attitude.”
Assistant Editor Mary Fratini lives in Barre.
