Using an advanced computer program and a single drop of blood from patients, researchers from the National Institutes of Health and a private firm, Correlogic Systems Inc., reported that they had accurately diagnosed 50 out of 50 women with ovarian cancer.
The results created a sensation.
Ovarian cancer is especially lethal because it tends to go undetected until it's too late. By the time symptoms appear, 80% of its victims are too ravaged by the disease to be saved. Reliable, early detection and treatment would save lives.
But nearly three years later, the diagnostic breakthrough is not close to reaching patients.
The project stalled while the government's two lead researchers -- Dr. Lance A. Liotta of the National Cancer Institute and Emanuel F. "Chip" Petricoin III of the Food and Drug Administration -- signed on as paid consultants with a rival of Correlogic named Biospect Inc.
Both companies were seeking ways to diagnose diseases, including ovarian cancer, by developing systems that could recognize patterns of proteins in the blood. In order to verify whether such systems would work, both companies were seeking samples of blood from patients.
Correlogic, which was working with the NIH on its product, complained in mid-2003 about Liotta's deal with Biospect. But senior officials at the National Cancer Institute allowed Liotta to continue as a leader of the government's ovarian cancer collaboration with Correlogic and as a paid consultant to Biospect. An administrator said that he saw no reason to believe that Liotta's consulting was overlapping with his government work. Similarly, Petricoin had received permission from the FDA to enter into the paid arrangement with Biospect.
Interviews and government and company records examined by the Los Angeles Times show that Liotta and Petricoin shared information with Biospect executives about their work with Correlogic, sometimes consulting with Biospect directly from their government offices.
Liotta and Petricoin declined through an attorney to be interviewed for this article. The attorney, Charles J. Morton Jr., said that both men had performed their government duties diligently and properly and had shared no information that was confidential. Any delay in developing an ovarian cancer test, Morton said, was not the fault of Liotta or Petricoin.
In sworn testimony this year to the House Energy and Commerce subcommittee on oversight and investigations, the scientists said that their jointly arranged and paid consulting services for Biospect were unrelated to their government work on the ovarian cancer test.
"I would never knowingly engage in any conflict of interest," Liotta, a 28-year NIH employee, testified May 18.
Petricoin, an 11-year FDA employee, testified that his consulting for Biospect was "performed to the highest ethical standards." He added that "at no time did I directly consult with Biospect about the work in our research [collaboration] with Correlogic, or provide them with any secret or nonpublic information."
Spotting a Promising Idea
The idea for using a drop of blood and a computer program to detect ovarian cancer was hatched in June 1999, when Petricoin and Peter J. Levine, Correlogic's president, had brunch with their wives in Alexandria, Va. The women had met at work, and the couples were social friends.
The two men scribbled notes on a cocktail napkin about spotting patterns of protein in blood that might indicate ovarian cancer. The idea seemed promising. Soon, the National Cancer Institute and Correlogic began collaborating in the lab, examining blood samples from women who might be at risk for ovarian cancer and from women who already had the disease.
In February 2002, the company and government scientists published the favorable results in the Lancet, the British medical journal.
At first, progress was swift. In comments posted on the cancer institute's website, Liotta estimated that clinical studies seeking to confirm the results could start in about six months. Liotta, chief of the institute's laboratory of pathology, had already said publicly that the FDA might grant marketing approval for an ovarian cancer test on a fast-track basis.
In April 2002, the government agencies licensed Correlogic to commercialize the test being developed for ovarian cancer or other diseases.
Correlogic and the agencies took another step together, entering into a formal cooperative research and development agreement. Such pacts are intended to rapidly transfer innovations from government labs to the marketplace. The agreements provide for product licensing rights to be shared, through negotiation, between the labs and the participating companies.
At Biospect, a start-up firm based in South San Francisco, researchers were trying to develop their own system for diagnosing diseases based on patterns of protein in the blood.
The leadership of Biospect was well-connected at the NIH. The company's vice president for strategic partnerships, Carol A. Dahl, had been chief of the National Cancer Institute's office of technology and industrial relations; Biospect was co-founded by the former director of the cancer institute, Dr. Richard D. Klausner.
On June 24, 2002, Dahl approached Liotta at an NIH conference, and she described the outcome the next day in an e-mail to Klausner and other colleagues:
"I spoke with Lance Liotta yesterday at the NIH Biosensors meeting. I mentioned the possibility of consulting with us. He is very interested." The e-mail mentioned the ovarian cancer project that the NIH had underway with Correlogic, adding:
"I think [Liotta and Petricoin] could be useful as consultants, in terms of providing more detail about what they have done as well as assisting us in thinking through issues like sample collection."
While Liotta and Petricoin were being courted by Biospect, their government role with Correlogic and the ovarian cancer project continued.
On Sept. 12, 2002 -- seven months after publication of the ovarian cancer testing results -- an official from the National Cancer Institute told Correlogic that the government intended to conduct a clinical trial in partnership with the company. Success in such studies is one way of getting FDA approval for bringing a product to market.
Yet the clinical trial was not launched. Instead, Correlogic met a series of frustrations, according to government and company records and interviews with participants.
At meetings to discuss the status of the research, government scientists at times declined to share information, citing concerns about confidentiality. When Correlogic sought documentation necessary to verify the origins and status of 500 samples of frozen blood serum to be provided by a university researcher working jointly with the company and the government, months passed with no resolution. Correlogic concluded that because of the missing paperwork, the samples were not useable.
"Everything was like pulling rhino teeth," said Levine, Correlogic's president.
Morton, the lawyer for Liotta and Petricoin, said it would be "bizarre" to fault them for the difficulties that Correlogic experienced with the serum samples because the university researcher was not accountable to his clients.
Yet unknown at the time to Correlogic, the two government scientists were going to work for Biospect. On Nov. 6, 2002, Liotta and Petricoin conferred for one hour with Biospect's Dahl at "Lances office," according to Petricoin's daily government calendar.
One month later, NIH officials approved Liotta's arrangement with Biospect. From that point until Liotta stopped consulting for the company in May of this year, Biospect paid him $70,000 in fees, according to government and company documents. (The scientists were paid at the same rate; the total fees collected by Petricoin could not be obtained.)
Levine said that he first heard about Biospect in May 2003, when a friend in the biomedical industry sent him company materials indicating that Biospect was a competitor.
Within weeks, Correlogic had more cause for concern.
On July 7, company executives were meeting in their glass-enclosed headquarters on the third floor of Democracy Plaza, a building in Bethesda about three miles west of NIH headquarters. Correlogic's offices provided a view of the hallway.
Timothy A. Coleman, a company vice president, spotted a couple of familiar faces leaving a nearby office: Petricoin and Liotta.
"I remember looking out and I said, 'Gee, that's Chip and Lance. What are they doing here?' " Coleman said. Levine approached them and spoke briefly with Petricoin, learning that the small office they were leaving belonged to Biospect.
Correlogic would have had little way of knowing earlier that Liotta and Petricoin had arranged side deals with Biospect because neither was required to file income-disclosure reports that were open to the public.
Levine noted the office tower encounter in an e-mail to an NIH lawyer.
"I bumped into Chip and Lance as they were exiting a meeting in my building this afternoon (which raises another issue, but I will leave that for another time)," Levine wrote.
The next day, Levine sent another e-mail to the NIH lawyer, questioning the control that Liotta and Petricoin held over the progress of the cooperative research and development agreement, known as a CRADA. At Correlogic, cooperation on the test for ovarian cancer had become so scant that the arrangement was known as "the un-CRADA."
Later in July, Correlogic complained about Liotta's paid deal with Biospect to a deputy director of the National Cancer Institute. But institute officials authorized Liotta to continue his arrangement with the company.
Liotta and Petricoin subsequently provided Biospect information describing their government work with Correlogic on diagnosing ovarian cancer.
On Jan. 12 of this year, Petricoin e-mailed a Biospect vice president, John T. Stults. Petricoin included information from the government's collaboration with Correlogic. The heading on one page said: "CO-DEVELOPED THROUGH A CRADA WITH CORRELOGIC SYSTEMS, INC." The e-mail contained slides explaining how proteins in blood are analyzed for diagnostic use.
On Jan. 22, Petricoin's government office calendar had a reminder message from Stults about a conference call to be held the next day, saying, "Agenda: Status of the ovarian cancer" clinical trials. Also to be discussed in the call, according to the message, were formal "protocols" for collecting blood serum and "serum stability questions." Copies of the e-mail and calendar entry were obtained by The Times.
The House subcommittee summoned Liotta and Petricoin on May 18 as part of its investigation of NIH conflicts of interest. The two scientists testified that they had not viewed Biospect as a competitor to Correlogic.
Liotta and Petricoin stopped consulting for Biospect days before the hearing. The company was renamed Predicant Biosciences the same month. Company representatives declined to be interviewed for this article.
"I never consulted with Biospect about my CRADAs, never consulted with them about my government work," Liotta testified.
Petricoin told the subcommittee: "[A]t no time did I directly consult with Biospect about the work in our research CRADA with Correlogic, or provide them with any secret or nonpublic information."
At a subsequent hearing on June 22, a Biospect executive, Jonathan C. Heller, was asked about the company's dealings with the two government consultants. He noted that Biospect had "signed a confidentiality agreement with the National Cancer Institute to acquire new data that was publicly available, but not yet published from Drs. Liotta and Petricoin." The data Heller referred to was contained in an article, subsequently published in a scientific journal, and written by Liotta, Petricoin, Correlogic's chief scientific officer and others.
Heller said that the company was "18 to 24 months away from having a diagnostic service ready for sale on the market." But Biospect, Heller said, "would not have sought confidential information pertaining to Correlogic's CRADA because such information would have been of no value to our company."
Rep. Joe Barton (R-Texas), chairman of the House Energy and Commerce Committee, called the financial ties between Biospect and Liotta and Petricoin "a remarkable case." Barton added:
"As a result of those secret deals, progress may have been slowed on the public-private partnership that could have led to prompt commercialization of a life-saving ovarian cancer diagnostic test."
NIH Director Elias A. Zerhouni declined to defend Liotta's dealings with Biospect. Referring to the circumstances surrounding the collaboration with Correlogic, he told the hearing in June that he had reached a "tipping point" in dealing with NIH conflicts of interest.
At Correlogic, Levine said he still hoped to develop a reliable test for early stage ovarian cancer. Although the company's formal agreement with the National Cancer Institute expired in April, Levine left open the possibility that Correlogic could resume work with the institute.
"If we had had a genuine collaboration in the last two years, the science and technology would have advanced," Levine said. "What went on here is an example of the way the public-private partnership should not work."