Thomas Neville's Dynamic Analysis of Multiple PSA Results Leads to Better Reliability, Insights | OncLive

Thomas Neville's Dynamic Analysis of Multiple PSA Results Leads to Better Reliability, Insights

March 28, 2015

A new phase begins for Thomas Neville, PhD, a Yale and Stanford-trained engineer and economist who worked in energy before a prostate cancer diagnosis drew him into medical research.

Thomas Neville, PhD

The data took years to acquire and months to prepare, and the actual analysis took weeks and sometimes months for the largest datasets. The progression of PSA levels correlated so strongly with the danger from prostate cancer that it virtually jumped off the screen.

Now, a new phase begins for Thomas Neville, PhD, a Yale and Stanford-trained engineer and economist who worked in energy and helped build an international industrial combustion engineering company before a prostate cancer diagnosis drew him into medical research.

“When I first got involved with prostate cancer research, the results were so strong that I assumed they would have an immediate impact on screening practices,” Neville said. “I was naïvely optimistic, but that’s probably a good thing because it kept me going.”

Neville started down his current path in 2002, when his doctor called to say that a prostate biopsy had tested positive for cancer.

His doctor suggested several options, but Neville lacked the expertise to compare them, so he did what any well-educated individual would do: He reviewed the existing research.

Neville’s initial investigation provided far less certainty than he’d expected. Some evidence supported radical prostatectomy. Some supported radiation. Some supported active surveillance. Different experts further muddied the water with conflicting recommendations.

Still, Neville collected all the evidence he found and synthesized it as best he could, noting areas of consensus and evaluating the relative strengths of conflicting studies. He posted his findings online, updating his material as he went along. His latest material on prostate cancer can be found at ProstateSmart.info.

By the time Neville chose to undergo a radical prostatectomy, his work had reached a surprising number of grateful readers, readers who inspired Neville to supplement the treatment information with similar information about screening. Eventually, the lack of persuasive answers to relatively basic questions about screening inspired Neville to undertake new research.

With his wife Kelly, he also decided to turn his avocation into a business, and they founded a start-up company called Soar BioDynamics, Inc., near their home in North Lake Tahoe.

PSA

The observation that cancer tends to increase production of a protein called prostate-specific antigen (PSA) has made blood tests the primary way of finding the disease in asymptomatic men for the past 20 years. Indeed, during much of that time, many medical organizations recommended yearly PSA tests for men, starting at age 50.

Normal PSA levels vary considerably among men, but doctors have traditionally regarded PSA levels above 4.0 ng/mL with enough suspicion to consider performing a biopsy and making a definitive diagnosis.

Recent research, however, has challenged this reliance on PSA testing in a number of different ways. Studies have demonstrated that some prostate cancer victims have PSA levels below 4.0 ng/mL, while many cancer-free men have much higher concentrations of PSA in their blood. Other studies have found that early prostate cancer detection often does more harm than good. Men frequently overreact and choose to fight indolent tumors with treatments that can potentially result in serious side effects such as impotence and incontinence.

Such findings have led some experts to recommend against widespread yearly PSA testing, but it led Neville (and others) to ask a question: If doctors used data from many years’ of PSA tests rather than considering one test in isolation, could they detect cancer more reliably and gain some insight into its likely speed of progression?

Neville realized that he’d be regarded as a crazy dilettante unless he could attract some established collaborators, so he reached out to many of the experts who were doing related research.

“I was extremely fortunate to have had superb physicians care for me during my time as a patient, and we have stayed in touch. They are also top researchers. Peter Carroll is now associate dean of the School of Medicine and chief of Urology at UCSF, and Eric Klein is now chairman of the Glickman Urological and Kidney Institute at Cleveland Clinic,” Neville said. “But a lot of it is just knocking on doors and saying, ‘Hi. I’m Tom Neville.’ Sometimes folks slam the door in my face, but sometimes they let me in and I end up with a new resource.”

Continuing Success

Few people take on a new field as complicated as medical research after they become eligible for AARP membership, but Neville has demonstrated a wide range of talents over the course his life.

During his time at Yale University, Neville performed well enough as captain on the football field to be drafted by the Baltimore Colts and well enough in the classroom to graduate first in his Engineering and Applied Science class and be selected as a Rhodes Scholar. He chose Oxford over Baltimore and, in addition to earning a degree in economics and politics, learned rugby quickly enough to make the university’s top squad within months of his arrival, playing twice in the annual Oxford-Cambridge match.

Neville returned to the US in 1973 to attend Stanford University just as the first oil embargo was getting underway, and it attracted his interest. While at Stanford, he landed a summer job as an energy policy analyst at a precursor to the US Department of Energy, where he later worked as a policy office director upon the department’s creation in 1979. At Stanford, he worked on energy policy, studied with pioneers of Bayesian decision-making under uncertainty, and earned a PhD in economics and finance, awarded in 1979.

Over the next three decades, Neville’s career took him around the energy sector and then into industrial combustion engineering. By the time he received his prostate cancer diagnosis, he was the chief technology officer and co-owner of North American Manufacturing Co., a privately held firm that sold ultra-low-emission combustion systems to customers in heavy industry around the world.

Room for Improvement

Neville and his partners sold their firm to the Fives Group in 2008, and he began working full-time to further the ongoing revolution in prostate cancer care.

That revolution has seen prostate cancer mortality fall more than 40% in the past 25 years. The American Cancer Society says the 5-year survival rate is now close to 100%, and the 15-year survival rate is 94%.

Those numbers reflect a long series of innovations: breakthroughs such as nerve-sparing radical prostatectomy, high-
energy radiation treatment and, perhaps, the PSA test.

There is, however, plenty of room for further improvement. Prostate cancer still kills nearly 30,000 American men each year.

Neville’s initial question about the potential value of dynamic analysis of multiple PSA tests arose from pioneering work done by Ballentine Carter, MD, professor of Urologic Oncology at Johns Hopkins University and now director of the Division of Adult Urology, and Chris Morrell, PhD, professor of Statistics at Loyola University in Maryland.

Carter spent nearly 2 decades investigating possible strategies for extracting valuable data from changes in PSA levels, including initial work with records from the Baltimore Longitudinal Study of Aging. But Neville figured that additional analy-
tical techniques, like those employed by Morrell to find hidden relationships in many different fields, might work to extract value from a man’s PSA values over time.

Many research projects by Neville and colleagues showed significant correlations between PSA progression and patient health. For example, Neville worked with a team from the University of Innsbruck to follow 817 Austrian men with elevated PSA levels and at least 5 PSA test results in a 4-year period. Analysis of subsequent biopsies and prostatectomies showed that a man whose cancer PSA increased 50% per year had a 35% chance of having an aggressive tumor, while a man whose PSA increased 5% per year had a 12% chance of having an aggressive tumor.

Higher Sights

Neville’s Innsbruck colleagues presented the study at the American Urology Association’s annual meeting in 2010. But he sensed that the analysis would be far more compelling if it could be substantiated on a very large data set.

Neville, therefore, set his sights on a data pool that would satisfy even the most skeptical reviewer, the one maintained by the US Department of Veterans Affairs with 33 million PSA test values for 14 million men.

“I knew the VA shared its data with medical researchers because I’d seen a lot of studies that used it, so I figured it would be pretty easy for us to get access,” Neville said. “It ended up taking 4 years, which I thought was an eternity, but researchers have come up to me since then and asked how we moved so fast.”

Even once the data was obtained, work proceeded slowly at first, because the sheer amount of data the VA maintains on each veteran’s medical records made it important to systematically separate relevant material from the rest and get it into properly structured fields.

Fundamental Insights

Neville describes some of the project’s fundamental insights:

When doctors have access to several years of PSA values for each man, the VA data indicate that sophisticated analysis of PSA growth above a no-cancer baseline provides better predictive power about the presence of cancer and the aggressiveness of that cancer than any single PSA test can provide. The time patterns are valuable in helping to determine both the probability of finding cancer and the deadliness of that cancer. Neville commented, “If you take these powerful dynamic screening results and add to them some of the other valuable tools that are available, there should be way of devising a system of “smart screening” that can bring enormous value to clinical decision-making.”

When screening analysis considers both dynamic analysis of PSA patterns and many other factors, such as age, race, family history, digital rectal exam result and prostate volume, the VA data suggests that screening can become much more effective than relying on a single PSA test. Indeed, Neville and his colleagues estimate that, if such “smart screening” became the norm, it could reduce prostate cancer deaths by 40% while also reducing biopsies by 65% and treatment by 50% while saving $5 billion per year in US medical costs.

He goes on to explain that so many good outcomes are possible because of the divergence in screening in the US — some men are not screened or they are under-screened, and other men are over-screened. The VA results suggest that many prostate cancer deaths are associated with men who were not screened or screened inadequately and diagnosed at high PSA levels, and many biopsies and subsequent diagnoses and treatment are associated with different men who were biopsied prematurely at low risk.

Neville and his urologist colleagues presented three posters at the most recent AUA meeting along with a press briefing and will present additional confirmatory findings on another significant data set at the upcoming AUA meeting in May. He and his colleagues are preparing journal articles on a range of new findings.

He hopes that the work on dynamic screening and “smart screening” will lead to a revolution in screening practices. But, as he says, “I’ve learned that changes in clinical practice don’t happen quickly or easily, so it’s important to stay at it, one step at a time — especially when what’s at stake is the chance to improve the health and well-being of 100 million men around the world. For now, that’s enough to keep me going.”


x