U.S. House of Representatives: Committee on Energy and Commerce, RepublicansPrepared Witness Testimony: Papineau, Frank

Chairman Greenwood, Ranking Member Deutsch, and Members of the Subcommittee, I am Frank Papineau, on detail to the Energy & Commerce Committee's staff. I am here today to provide you with background information and key facts and dates surrounding the Food and Drug Administration's decision to end its consideration of ImClone Systems' highly touted cancer drug, Erbitux, and the questionable ImClone stock-selling activity during this turn of events.

My remarks are an oral summary taken from the Committee staff report prepared for today's hearing. I am accompanied today by Dr. Raymond Weiss, Consultant in Oncology and Clinical Professor of Medicine at Georgetown University Medical Center. Dr. Weiss is under contract with the Committee to provide assistance to the staff. Dr. Weiss wrote a report of his findings, which is appended to the Committee staff report.

BACKGROUND

By way of background, ImClone Systems is a small biotech company based in New York City, founded in 1984 by two brothers - Sam and Harlan Waksal. ImClone has never turned a profit in its 18 years of existence and reportedly has spent over $200 million on research of Erbitux. Many people involved in cancer research believe that Erbitux is a promising drug and widely expected it to be on the market this year. Erbitux, however, was not approved for the market because the Food and Drug Administration found so many problems with ImClone's application for approval that it issued a Refusal To File letter, a rare FDA action that effectively turned the drug back to the company for further study. This situation attracted national attention because of the pre-market publicity about the drug, because of ImClone's record-setting $2 billion alliance with Bristol-Myers Squibb to market Erbitux, and because of multi-million dollar stock trades by ImClone insiders in the weeks before FDA's negative decision.

Over the past six months, Committee staff has conducted an extensive investigation into matters surrounding ImClone's cancer drug and related business dealings. The Committee's investigation focused on the FDA drug approval process, Erbitux's clinical trials, Bristol-Meyer's partnership arrangement to acquire commercial rights to Erbitux, and the key events leading up to FDA's Refusal to File letter and trading of ImClone stock by its board members and officers, as well as, several of Sam Waksal's immediate family and friends.

Before we proceed to the business dealings, let me first highlight two of staff's findings regarding FDA's review of Erbitux: One, FDA's initial decision in August 2000 to grant fast-track designation to Erbitux appears to have been based on incorrect information regarding the study protocol submitted by ImClone in support of the proposed cancer treatment, which involved Erbitux and another cancer drug, Irinotecan. Two, the FDA made this initial decision before it had full information about Erbitux's activity when administered in the absence of this other drug; and it was this other information -- requested in a letter by FDA in January 2001 and received from ImClone in October 2001 -- that led agency reviewers to conclude the application was inadequate.

THE BRISTOL-MEYERS SQUIBB DEAL AND IMCLONE'S INTERNAL LOAN

In early 2001, Bristol-Meyers Squibb (BMS) failed in its effort to form an alliance with a biotech company, OSI, that it believed had a promising cancer drug. The company believed it was losing its share of the oncology drug market and decided to re-visit ImClone and its cancer drug Erbitux. On June 1, 2001, after a month of negotiations, Sam Waksal outlined an acquisition plan that would give BMS a 70% majority stake in ImClone. BMS's Board of Directors rejected the deal. Mr. Waksal then told BMS that he was willing to consider alternative proposals provided they include a significant equity investment in ImClone by BMS and he also advised BMS that he believed ImClone's existing stockholders would benefit most if BMS acquired an equity interest through a tender offer to the ImClone's existing stockholders.

During July 2001, after ImClone was virtually assured of the equity deal and in anticipation of the tender offer from BMS ImClone's Board agreed to lend $35.2 million to the Waksal brothers and the Chairman of the Board. The loans provided the opportunity for the three individuals to exercise stock options and warrants they held to purchase a total of approximately 4.5 million shares of ImClone stock. (Sam Waksal and Harlan Waksal's loans were $18.2 and $15.7 respectively. The Chairman's loan was in the amount of $1.2 million.)

On October 29, 2001, thousands of ImClone's shareholders participated in the BMS tender offer to purchase ImClone stock at $70 a share, a $20 premium over the trading price. ImClone's Board of Directors tendered 2.1 million shares to BMS by themselves -- representing approximately 15% of the stock tendered by ImClone shareholders to BMS. Sam and Harlan Waksal tendered 814,674 and 776,450 shares for about $111 million. Simply stated this means that the Waksal brothers received more than 10% of the entire proceeds paid by BMS during the tender offer. Although all ImClone shareholders were allowed to tender shares to BMS, only the Waksals and two other board members borrowed millions of dollars of company funds to purchase the stock and then tender it to BMS.

THE RTF LETTER

On December 28, 2001, the FDA issued its "refuse-to-file" (RTF) letter in response to the ImClone submission. The RTF letter is sent in rare cases when a submission is deemed insufficient. (It is a non-public document containing trade secret or confidential commercial information.) In a December 31, 2001 conference call with investors, ImClone executives said that FDA sent the RTF letter because the Erbitux application was missing certain "train of documentation" information needed by regulators to accept the filing. ImClone said it would be able to answer the FDA questions by the end of the first quarter, leading, hopefully to an approval of Erbitux in the fall.

On January 4, 2002, the Cancer Letter published excerpts of the RTF letter indicating that -- contrary to ImClone statements to investors -- the FDA had a long list of concerns that went far beyond record keeping. The FDA believed ImClone's clinical trial was not adequate and well controlled and that additional studies would be needed. The letter suggested that the FDA had warned ImClone starting in August 2000 that its data would have to demonstrate that Irinotecan, the standard chemotherapy mentioned above, was needed along with Erbitux. But the data submitted by ImClone was not sufficient to distinguish the effects of the two treatments.

TRADING ACTIVITY BY IMCLONE EXECUTIVES AND OTHERS

Adding to the controversy over Erbitux has been the trading of ImClone stock by ImClone insiders a few weeks before the FDA letter, as well as the trading of stock by Waksal family relatives and friends during the 48 hours before the FDA letter was issued.

On December 21, 2001, ImClone issued a Company order stopping its employees from trading in ImClone stock until after the FDA decision on Erbitux was made public. Committee staff believes that no board member or officer of ImClone traded ImClone stock between December 21 and 28, 2001. However, staff found that, except for Sam and Harlan Waksal, members of Sam Waksal's immediate family sold ImClone stock on December 27, 2001 or the next day hours before ImClone announced publicly that FDA had refused to accept the filing of Erbitux.

We found that three officers of ImClone sold stock prior to December 18, 2001 on the advice of their broker. In addition, Harlan Waksal conducted a forward sale of 700 thousand shares on December 6, 2001.

The staff learned that on October 31, 2001, Harlan Waksal notified the ImClone Board Members that he planned to execute a 700,000 share stock transaction. He told the board that the stock would still be under his voting control for the next three years. He also stated that he'd finalize transaction over the next two weeks. He told Committee staff that in early November 2001 he attempted to shop the sale. He told staff he was forced to sell the ImClone stock to come up with enough cash to pay substantial taxes racked up from his prior exercise of stock options and his tendering of shares to BMS. He also stated that because he didn't want to sell shares he entered into a forward sales contract that gives him a percentage of the cash value of the shares up front but still allows him to control the shares and defer tax payments for another two years. In short, Waksal received less than what the stock was worth at the time of the sale, but he also limited the downside risk when ImClone's stock price continued to drop. It should be noted that Harlan Waksal sold the 700,000 shares on the same day that ImClone hit its 52-week high.

AN INQUIRY INTO THE IMCLONE CANCER-DRUG MATTER PRELIMINARY COMMITTEE STAFF REPORT

At the direction of Chairman W.J. "Billy" Tauzin and Subcommittee Chairman James C. Greenwood (later joined by Ranking Minority Member John D. Dingell and Subcommittee Ranking Minority Member Peter Deutsch), Committee staff conducted an investigation into matters surrounding the development by ImClone Systems, Inc., (ImClone) of its colorectal cancer drug Erbitux (also known as C225 or Cetuximab).

ImClone, a small biotechnology company based in New York, was founded by two brothers, Drs. Sam and Harlan Waksal, in 1984. ImClone developed a cancer therapy drug called Erbitux, reportedly spending more than $200 million on research on this drug. ImClone has never turned a profit in its 18 years of existence.

In the spring of 2000, ImClone sought accelerated approval from the Food and Drug Administration (FDA) to market Erbitux to meet the medical need of colorectal cancer patients who have failed to respond to standard chemotherapies. ImClone and Erbitux are internationally known, having been featured on the CBS news program "60 Minutes," and the international cover of the July 30, 2001 issue of Business Week. One reason Erbitux received such attention is that, according to Business Week, this drug was "the furthest along of a handful of new cancer treatments that precisely home in on a growth signal found in up to 50% of all cancer types." In clinical trials, "the drug demonstrated remarkable success in causing colon cancer to regress in patients who had failed to respond to all other treatments." Erbitux also is promising because it is an antibody that targets and blocks off cancer cells, without the high degree of side effects from standard cancer treatment. Such promise apparently prompted thousands of cancer patients to try to obtain Erbitux either through clinical trial enrollment or "compassionate use" access. For example, USA Today reported that ImClone had received 400 calls a day from patients desperate to get Erbitux outside of clinical trials.

In September 2001, Bristol-Myers Squibb (BMS) bought 19.9 percent of ImClone for $1 billion, and agreed to pay as much as $1 billion more to obtain the marketing rights to Erbitux. On October 30, 2001, ImClone submitted its Biologics License Application (BLA) for Erbitux to FDA. On December 17, 2001, ImClone was one of seven biotechnology companies included for the first time in the NASDAQ 100 index. Excitement and confidence in ImClone was reflected in such media reports as an article in the December 26, 2001 Los Angeles Times, which proclaimed, "Erbitux, a colon cancer treatment from ImClone Systems Inc., is set to make one of the biggest splashes of 2002."

Many observers and investors were thus stunned to learn that, on December 28, 2001, FDA issued a "refuse-to-file" (RTF) letter in response to the ImClone license submission. The RTF letter is sent in rare cases when a submission is deemed insufficient, and is a non-public document, since it contains trade secret or confidential commercial information. ImClone publicly announced the FDA decision the evening of December 28th, which prompted a sharp sell off in ImClone shares starting on December 31, 2001.

The Committee's investigation focused on the validity of the claims that were asserted about ImClone's effectiveness, the FDA filing and review process, and evidence uncovered by the Committee that friends and family members of ImClone's founders sold large amounts of ImClone stock just prior to ImClone's receipt of the negative determination from FDA.

To review the above issues, Committee staff conducted hundreds of hours of interviews with officials from ImClone, BMS, and other pharmaceutical companies, FDA, Wall Street firms, patient advocacy groups, oncologists, and representatives of family and friends of Sam and Harlan Waksal. Staff also obtained and reviewed thousands of documents from the above officials, corporations, and FDA. These documents and discussions with officials included, but were not limited to, the FDA drug approval process, clinical trials, the BMS tender offer and milestone payments with ImClone, events leading up to the FDA refusal-to-file letter, stock trading by ImClone officials and Waksal family and friends, and ImClone's filings with the Securities and Exchange Commission (SEC). Staff also reviewed the due diligence activities conducted by seven other major pharmaceutical firms during 1999 and 2000, to determine what they learned about ImClone and its products, and what their rationale was for not entering into an alliance with ImClone, as BMS did in 2001.

The ImClone case highlights the policy question of how to test cancer drugs in a way that balances rapid access to life-saving drugs with the need to ensure that the drugs work, particularly when a publicly traded company is involved. In the standard approval process for a drug, FDA normally requires one or more large clinical trials (usually called Phase III trials) showing that a drug prolongs life compared with a placebo or with an already-approved drug. Such trials can take years, involve thousands of patients, and cost hundreds of millions of dollars to perform.

When a company develops a drug for patients with life-threatening diseases and there are comparatively few treatment options available, FDA sometimes approves the new drug based on smaller trials, without a control group for comparison. The trials normally look at whether tumors are shrinking, which can be determined much faster than whether patients are living longer. Often, these trials are limited to patients who have not responded to existing therapies (known in medical terms as "refractory" patients). If FDA approves a drug based on such small trials, it typically requires companies to conduct additional studies to show more widespread benefit, such as additional survival time.

In ImClone's case, the company was trying to get approval for Erbitux based on a study where the drug was used in combination with an approved chemotherapy, in a universe of approximately 120 patients - a very small patient pool. ImClone's strategy appears to have been unprecedented. According to the BMS Due Diligence Findings, dated June 12, 2001: "No accelerated approval has ever been granted for an oncology drug for use in a combination therapy." It also should be noted that ImClone was seeking FDA's agreement for accelerated approval with a protocol design of a study that already had been conducted.[1]

The Committee's investigation focused on two areas of the FDA process prior to the submission of ImClone's BLA for Erbitux in October 2001: (1) the clinical protocol design and conduct of the pivotal 9923 study, and (2) the single-agent study of Erbitux.

In the spring of 2000, ImClone had two Phase II clinical trials that looked promising for accelerated approval: a study in head-and-neck patients, and a study in colorectal cancer patients. ImClone originally anticipated that it would be the head-and-neck trial that would be the vehicle for possible FDA approval. However, because of faster accrual of patients and promising results, it was the colorectal cancer patient study, known as the 9923 study, that ultimately formed the clinical core of ImClone's BLA. According to ImClone, the results of the 9923 study showed a 22.5% positive response rate in colorectal cancer patients who already failed the standard chemotherapies.

In August 2000, ImClone was scheduled to meet with FDA to discuss, among other things, whether the results of the 9923 study were clinically meaningful and whether 9923 could meet accelerated approval criteria and receive fast-track designation. Prior to the ImClone meeting, FDA officials held an internal "pre-meeting" to prepare. At this pre-meeting, the primary FDA medical review officer indicated her reservations concerning the 9923 study. Her notes from this meeting state: "1) Is ORR [overall response rate] = 15% clinically meaningful for colorectal CPT-11 failure? Only if as a single agent. 2) CP02-9923 meet accel. approval criteria and fast track? No." According to Committee staff interviews, nobody on the FDA staff expressed disagreement with the assessment of the medical review officer at this internal "pre-meeting."

On August 11, 2000, FDA met with ImClone officials and consultants to discuss ImClone's accelerated approval strategy using the 9923 study. According to the minutes of this meeting prepared by FDA, FDA participants described the 9923 study during this meeting as follows:

"This is a Phase 2 open label study of Cetuximab [Erbitux] plus irinotecan in metastatic or recurrent colorectal cancer refractory to irinotecan. Following two courses of irinotecan, patients' tumors are measured and based on the results, divided into the Stable Disease Treatment Group (tumor volume change < 25%) or the Progressive Disease Treatment Group (tumor > increased in volume 25%). Patients then receive irinotecan plus cetuximab until treatment failure."

This description accurately tracks the first version of ImClone's protocol for 9923. According to that August 2, 1999 Version 1.0 of Protocol IMCL CP02-9923, Section 3.1.2, the patient "must have demonstrated progression of disease after completing a minimum of two courses of a regimen containing irinotecan." However, a few months later, when patients were being enrolled into the study, ImClone relaxed the inclusion criteria in an amended protocol. According to the October 18, 1999 Version 2.0 of Protocol IMCL CP02-9923 amended Section 3.1.2 (Inclusion Criteria), the patient "has documented stable disease (must have received a minimum of 12 weeks of irinotecan therapy) or progressive disease at any time after receiving an irinotecan-containing regimen. Copies of scans must be provided to confirm the lack of an objective response to prior therapy." (Emphasis added).

Therefore, FDA was relying on an outdated version of the protocol at the August 2000 meeting with ImClone. Yet nobody from ImClone informed FDA about the amended protocol at this meeting or any time thereafter. Moreover, the minutes of the meeting taken by the company and FDA were exchanged, yet, again, the company did not correct the FDA's misunderstanding on this point.

At the August 11, 2000 meeting with ImClone, the most senior FDA medical officer agreed that "the basic trial design is probably acceptable," -- albeit, relying on the incorrect version of the study protocol -- and, in effect, overruled the view of the primary medical reviewer that had been expressed at the pre-meeting among FDA personnel. The senior FDA officer told Committee staff that her decision to accept the protocol was based on her belief that she should be flexible for a promising drug meeting an unmet medical need, but was also based on representations that ImClone made about the special synergistic effect of Erbitux when used in combination with irinotecan. The senior FDA officer said that ImClone asserted that Erbitux showed no activity when used alone, which would support the claim of synergistic effect. This assertion was based on animal data and one small human trial. In the context that ImClone discussed this point, she assumed the human trial involved human colorectal cancer patients. The senior FDA officer later learned that the human trial involved renal cancer patients, which cannot be used as a basis for determining single-agent activity in colorectal cancer patients. ImClone disputes that the issue of single-agent activity came up at the August 11, 2000 meeting, but the company agrees that the issue was discussed in subsequent phone calls and meetings with FDA.

On January 12, 2001, FDA granted fast-track designation for Erbitux. The FDA fast-track designation appears to be based on the inclusion criteria of the outdated version of the 9923 protocol. According to the January 12, 2001 letter to Nikhil Mehta of ImClone from Glen Jones of FDA: "[W]e are designating as a Fast Track development program the investigation of cetuximab in combination with irinotecan for its effect on durable tumor responses (complete and partial responses) in patients with metastatic colon cancer who are refractory to standard chemotherapy (5 fluorouracil and irinotecan), where refractory is defined as progressive disease during at least two cycles of standard doses of 5-fluorouracil and irinotecan." (Emphasis added).

On January 19, 2001, FDA sent a letter to ImClone requiring them to conduct a small study of 25-50 patients to test the response rate when using Erbitux alone as opposed to being used in combination with the toxic Irinotoecan. As FDA explained:

1. Preclinical and clinical data (including at least 25-50 patients) which excludes the possibility (e.g., through establishment of the upper limit of the 95% confidence interval around the observed response rate and the lower limit of the 95% confidence interval around the observed response rate with combination therapy) that the response rate observed with the combination of irinotecan and Cetuximab [Erbitux] would not be observed with single agent Cetuximab at the dose and schedule proposed. You must provide evidence that continuation of a toxic agent (irinotecan) is necessary to achieve the desired clinical effect. If you do not have such data, you should generate this information in a randomized controlled trial directly comparing the efficacy of single agent Cetuximab (the generic name for Erbitux) to the combination of Cetuximab plus irinotecan to establish the contribution of irinotecan in this setting."

During the winter and spring of 2001, while conducting the single-agent study, ImClone was actively pursuing a joint venture or a sale of the company, or of a majority interest in the company, to several pharmaceutical companies. It appears that, in pursuing such an arrangement, the ImClone leadership attempted to downplay the significance of the single-agent study required by FDA. For example, according to one drug company official's e-mail, dated April 6, 2001:

"They [Imclone] have to complete the pilot trial of C225 [Erbitux] alone in refractory colon cancer patients, 25-40 patients. The FDA has required a final study report from this trial prior to an ODAC [Oncologic Drug Advisory Committee] meeting. Per [ImClone] estimately [sic], they believe a final study report will be sent Oct/Nov, meaning a likely Spring ODAC meeting. According to Harlen [reference to Harlan Waksal], the FDA has agreed that while this study is necessary for filing, it will not impact the approval of the combination in refractory. They need to have the single agent activity per their regulations. They won't use the small trial to compare RR [response rate] of the single agent to the combo, but will use it to help plan further development of C225 as a single agent if appropriate."

On October 12, 2001, ImClone finished its single-agent study. The results of this study showed six responses out of 57 patients, for a response rate of 10.5%. As FDA noted in its December 28, 2001 refusal-to-file letter: "Based on the summary information provided, and assuming that the results can be confirmed, the data do not show that the response rate observed with the combination of Cetuximab and irinotecan could not also be observed with single agent Cetuximab at the dose and schedule proposed."

Even though there was a difference in the response rates (10.5% single agent; 22.5% combination), because both studies had such small populations, the confidence intervals overlapped and, thus, there was still a possibility that a very sick colorectal cancer patient could respond just as well with Erbitux alone as with Erbitux combined with a toxic chemotherapy. As a result, additional studies would be needed to isolate and establish the contributions of each drug. These additional studies would, at a minimum, significantly delay the launch of Erbitux.

However, it appears that ImClone attempted to portray the results of the single-agent study and the prospects for its application in an inaccurate light to BMS, its likely new business partner. According to an October 12, 2001 e-mail from BMS Chief Scientific Officer Peter Ringrose to other BMS executives: "I just had Sam Waksal on the phone re the single-agent data. Apparently it came out at 13% which he feels is half the C225 plus CPT-11 data. They have informed the FDA who were 'pleased' and confirmed that they would be on for the Feb 28 ODAC (FDA's Oncologic Drugs Advisory Committee). He reckons they will be on the market by March. I am planning to meet with Sam in NY week after next."

But, according to Committee staff interviews with FDA personnel, no one at FDA spoke to ImClone about the single-agent data on or around October 12, 2001, and FDA had never placed Erbitux on the agenda for the February 2002 ODAC meeting. The submission of the single-agent study to FDA was not completed until December 4, 2001.

To more closely evaluate these two studies relied upon by ImClone, the Committee hired an expert consultant to review the studies' designs, protocols, and results. The key findings from this review are contained in a Report to the House Committee on Energy and Commerce by Raymond Weiss, MD, FACP (attached as an appendix to this report).

On October 31, 2001, ImClone completed its BLA application for Erbitux by submitting the clinical portion of the BLA to FDA. This clinical portion included the records for the 9923 study and the single-agent study 0141 (except for data on 17 patients, which was submitted on December 4, 2001). Under the fast-track designation of the FDA Modernization Act of 1997, the agency was required to complete its review of Erbitux and determine filability within 60 days of the submission date. Until this submission, FDA had relied on assurances from ImClone and the records in ImClone's Investigative New Drug file. FDA did not actually see the details of the clinical trials for Erbitux until ImClone submitted this portion of its BLA at the end of October 2001. Upon reviewing the clinical portion, FDA reviewers immediately identified significant problems, and the number of problems continued to mount as their review continued in November 2001. According to the FDA reviewers, the Erbitux application, as filed, raised serious questions and lacked needed information that ImClone had been advised on several occasions would be required as part of the application. The FDA reviewers told Committee staff that it was readily apparent that the clinical research was severely deficient and could not meet the legal requirement of an adequate and well-controlled clinical trial.

On November 30, 2001, key FDA reviewers reached the conclusion that problems in the clinical portion were so severe that there was no option but to issue a refusal-to-file (RTF) letter, a rare event. On December 4, 2001, after raising the prospect of an RTF in a conversation with one of the FDA reviewers, ImClone's Regulatory Affairs Vice President formed an impression for the first time that an RTF letter was a realistic possibility, according to her interview with Committee staff. That same day, she reported this conversation and FDA's concerns to Dr. Harlan Waksal. On December 5, 2001, FDA management decided ImClone would receive an RTF letter. On December 7, 2001, a BMS Regulatory Affairs executive reported that she was not sure ImClone fully understood the implications of the comments of a FDA medical reviewer regarding the individual contributions of the drugs in the combination trial. In the e-mail opinion of the BMS executive, based on the FDA reviewer comments, "a refusal to file decision doesn't appear altogether unlikely at this point."

Both FDA and officials from the two companies told Committee staff that the tone of conversations between the agency and ImClone dramatically changed following the early December discussions with FDA. By mid-December 2001, it was clear to both ImClone and BMS that FDA had serious concerns about the Erbitux drug application. After a teleconference with FDA on December 12, 2001, key ImClone executives perceived an increased probability of an RTF letter, according to their interviews with Committee staff. On December 20, 2001, FDA told ImClone and BMS to no longer contact the agency until after they received FDA's letter on filability on December 28, 2001. Some personnel from ImClone and BMS thought from the tone of this conversation that an RTF letter was likely, but some in ImClone still held out hope for a positive FDA response. On December 24, 2001, an outside consultant for BMS was able to get an incidental confirmation from a source at FDA that FDA would be sending an RTF letter to ImClone. The next day, December 25, BMS Senior Vice President for Marketing Brian Markison called Dr. Harlan Waksal, who was vacationing in Colorado, to inform him of this confirmation BMS' consultant had received from an FDA source. Dr. Sam Waksal was vacationing at a Caribbean island and returned to New York on December 26, 2001.

It appears that Sam and Harlan Waksal and other key ImClone and BMS executives knew about the RTF letter by the morning of December 26, 2001. That day, ImClone sent a letter to FDA in an attempt to prevent the RTF by offering to waive its rights to the 60-day deadline that FDA had to meet by December 28, 2001. FDA declined the offer on the grounds that ImClone could not legally waive the deadline. On December 27, 2001, Sam Waksal for the first time personally interacted with FDA with respect to Erbitux, calling a senior official at FDA's Center for Biologics he knew when Waksal worked at the National Institutes of Health. The purpose of this call appears clear. Based on internal notes produced to the Committee by ImClone, dated 12:00 noon on December 27, 2001, "Sam and Harlan [Waksal] are calling FDA to try to stop RTF." The senior FDA official declined to intercede, and on December 28, 2001, at approximately 2:55 p.m., FDA faxed the RTF letter to ImClone. The company in turn publicly revealed the receipt of the letter later that day, at approximately 7:14 p.m.

As discussed above, on December 28, 2001, FDA issued a refusal-to-file letter in response to the ImClone submission. The RTF letter, sent in rare cases when a submission is deemed insufficient, is a non-public document containing trade secret or confidential commercial information. In its December 31, 2001 investors' conference call, ImClone executives said that FDA regulators sent the RTF letter because the Erbitux application was missing certain "train of documentation" information needed by regulators to accept the filing. ImClone said it would be able to answer FDA's questions by the end of the first quarter, leading, hopefully to an approval of Erbitux in the fall of 2002. On the first trading day after the issuance of the RTF letter, ImClone's shares fell $11.15, or 20 percent, to $44.10 per share.

On January 4, 2002, the Cancer Letter published excerpts of the RTF letter, which indicated that FDA had greater concerns about ImClone's data than company executives stated in the December 31 conference call with analysts and investors. The Cancer Letter article reported that the RTF letter detailed a long list of FDA concerns that went far beyond record keeping. The FDA was quoted as saying that ImClone's clinical trial was "not adequate and well controlled," and that additional studies would be needed. Moreover, the letter suggested that FDA had warned ImClone starting in August 2000 that its data would have to demonstrate that irinotecan, a standard chemotherapy, was needed along with Erbitux. But the data submitted by ImClone was not sufficient to distinguish the effects of irinotecan and Erbitux. After the Cancer Letter report appeared, ImClone shares fell sharply further, to open on January 7, 2002, at $34.96 per share.

On January 9, 2002, after ImClone had lost nearly $1.5 billion in market value since December 28, 2001, and after the filing of at least 11 federal class action lawsuits, Sam Waksal, ImClone's president and chief executive officer, attempted to explain the company's situation at the J.P. Morgan H&Q Healthcare conference. "What happened was that we put together a faulty package and we screwed up," Waksal reportedly said. The principal problem, he said, was the company's failure to provide documentation demonstrating that the patients enrolled in ImClone's pivotal trial had met the eligibility criteria.

During 1999 and 2000, ImClone invited BMS, as well as several other major pharmaceutical firms, to meet with representatives of ImClone to conduct due diligence with a view toward acquiring a majority ownership in ImClone. Over this time period, several pharmaceutical firms, including BMS, met with Sam Waksal and conducted preliminary due diligence activities. Each pharmaceutical firm, including BMS, concluded that the price being asked by ImClone was too high to continue discussions at that time.

In early 2001, BMS conducted an extensive internal review of its own biologics business, and evaluated a number of opportunities to expand its biologics capabilities. BMS concluded in April 2001 that ImClone's IMC-C225 compound, Erbitux, could sustain its leadership position in oncology, significantly contribute to its corporate growth strategy, and provide a significant step towards BMS becoming a leader in biologics.

In mid-April 2001, Mr. Brian Markison, BMS Senior Vice President of Marketing, contacted Dr. Sam Waksal to determine whether ImClone would be interested in pursuing a deal involving a significant equity investment in ImClone by BMS. On May 3, 2001, Dr. Waksal, Mr. Markison and Dr. Peter Ringrose, Chief Scientific Officer of BMS, met in New York City to discuss BMS' interest in ImClone. During that meeting, Dr. Waksal outlined the type of deal that would be acceptable to ImClone. Dr. Waksal's preference was that ImClone remain a publicly traded entity after the deal. As a result, Mr. Markison agreed to explore a possible transaction whereby BMS would acquire a majority interest of ImClone in return for BMS common stock, together with a separate agreement providing for the commercial rights to IMC-C225 by BMS.

After further discussions, on May 19, 2001, the two companies entered into a confidentiality agreement, and BMS conducted further due diligence of ImClone. On June 1, 2001, Mr. Richard Lane, President of BMS' Worldwide Pharmaceutical Division, and Dr. Waksal met to discuss an outline of a deal prepared by ImClone's legal advisors that called for an acquisition by BMS of a 70 % stake in ImClone.

On June 5, 2001, BMS' Board of Directors entertained the majority ownership deal with ImClone. However, some BMS board members raised concerns about acquiring majority ownership of ImClone, and suggested that BMS seek an arrangement of less equity in ImClone while still securing the rights to C-225. On June 7, 2001, representatives of the two companies met to discuss BMS' proposed due diligence activities. Shortly thereafter, employees of BMS and representatives of its legal and financial advisors conducted an extensive due diligence review of ImClone in the areas of clinical development, legal matters, information technology, marketing and sales, tax, finance, manufacturing, intellectual property and regulatory affairs.

In late June 2001, BMS concluded that the acquisition of a minority interest in ImClone, together with a separate commercial agreement relating to the co-development, co-promotion, and distribution of ImClone's IMC-C225 compound, would be a preferable structure for a deal with ImClone. Thereafter, Dr. Waksal was contacted by Mr. Peter Dolan, Chief Executive Officer of BMS, and Mr. Lane, who confirmed to Dr. Waksal that BMS no longer had interest in a deal to acquire a majority interest in ImClone where ImClone remained a publicly-traded entity. Mr. Dolan and Mr. Lane reaffirmed BMS' interest in ImClone and BMS' intent to consider other deals that met the economic and business objectives of both companies. Dr. Waksal stated that he was willing to consider alternative proposals, but emphasized that he was not interested in a commercial transaction that did not also include a significant equity investment in ImClone by BMS. Dr. Waksal also advised BMS that he felt ImClone's existing stockholders would benefit most if BMS acquired an equity interest through a tender offer to ImClone's existing stockholders.

On June 26, 2001, BMS provided ImClone with an outline of a proposed commercial transaction for the co-development, co-promotion, and distribution of IMC-C225, and an equity structure that proposed an acquisition of a 19.9% interest in ImClone by BMS. During the end of June and the first two weeks of July 2001, BMS and ImClone, and their respective legal and financial advisors, met several times to discuss terms and conditions of a 19.9% equity investment and a commercial transaction relating to rights to IMC-C225. Also during this time, the two companies and their respective financial advisors discussed the price at which BMS would offer to purchase the ImClone shares, which would be at a significant premium to the publicly-traded stock price.

In mid-July 2001 -- after ImClone was virtually assured of the 19.9% equity deal and in anticipation of the lucrative tender offer from BMS - ImClone's Board of Directors agreed to lend $35 million to Sam and Harlan Waksal and Robert Goldhammer, the Chairman of the Board, to provide them with an opportunity to exercise stock options and warrants they held to purchase a total of approximately 4.5 million shares of ImClone stock. Sam Waksal and Harlan Waksal's loans were $18.2 million and $15.7 million respectively. Mr. Goldhammer's loan was in the amount of $1.2 million. These unsecured loans were at an interest rate equal to the prime lending rate plus 1 percent (7.75 percent on the date of the note).

On July 20, 2001, BMS and ImClone agreed, on a preliminary basis, to a tender offer price of $70.00 per share. On September 17, 2001, the Board of Directors of BMS unanimously approved the ImClone deal. On September 19, 2001, ImClone's Board of Directors approved the deal, and both companies issued separate press releases announcing that BMS would acquire 14.4 million shares, or about a 20 percent stake, of ImClone's common stock for $1 billion through a tender offer of $70 a share, exclusively set aside for ImClone shareholders. At the time of the announcement, ImClone shares were selling at roughly $40 per share. BMS also agreed to pay as much as another $1 billion in milestone payments in return for the marketing rights to Erbitux in the United States.

On October 29, 2001, thousands of ImClone's shareholders participated in the BMS tender offer to purchase ImClone stock at $70 a share, a $20 premium over the increased trading price. Sam Waksal sold 814,674 shares, and Harlan Waksal sold 776,450 shares, or just more than 20% of each of their holdings. Although all ImClone shareholders were allowed to tender their shares of ImClone stock to BMS, only the Waksals, the Chairman of the Board, and one other board member were given loans by ImClone to purchase ImClone stock, at highly discounted prices, and then tender it to BMS at $70 per share.

A number of experts in the financial and biotech areas told Committee staff that there is no precedent in pharmaceutical-biotech alliances for the BMS and ImClone deal, which resulted in the immediate personal enrichment of top executives through a tender offer to existing shareholders. The more typical alliance formed between a major pharmaceutical company and a smaller biotech firm is centered on milestone payments that provide much needed cash to the biotech firm.

The Committee's investigation also focused on BMS' due diligence into the clinical research behind Erbitux prior to its decision to strike a commercial deal with ImClone. In May 2001, BMS scientists were mobilized to examine the clinical research package. On June 14, 2001, BMS Senior Vice President Laurie Smaldone sent an e-mail to her colleagues Peter Ringrose and Beth Seidenberg concerning ImClone, stating: "On the whole this remains a very high risk opportunity." Among the critical outstanding issues she cited:

"Pivotal CRC [colorectal cancer] program issues - Single agent activity. The trial which is ongoing will need to be shared with us. We should attend the FDA meeting with ICE [ImClone] when the data is final. There is no agreement that we could find that is reassuring regarding activity level needed for approval.

"Weak dose selection rationale - They have developed a PK [pharmacokinetic] rationale for dose selection, however the dose is questionable for refractory patients and the safety margin for early stage patients has not been determined. In their phase 3 first line study they are evaluating the same dose used in refractory disease. This is already seen as a problem by the FDA and by us. . .

"Safety - The safety of the product, specifically related to skin toxicity, bleeding, allergy has not been well characterized. This reemphasizes the weakness of the dose selection argument . . . "

Ultimately, concerns about the single-agent study and the 9923 study were not completely resolved before BMS entered into the agreement with ImClone. In a June 12, 2001 "Summary of Key Findings," BMS executives pointed out the risks of the results of the single-agent study:

"FDA has requested that data be provided on the antitumor activity of C225 as a single agent. Preclinical data has thus far been provided to FDA to address this issue, but they have persisted in their interest that clinical data be provided. No accelerated approval has ever been granted for an oncology drug for use in a combination therapy. (emphasis added). In the event that tumor responses are observed in the ongoing single-arm single agent refractory colorectal study then it is possible that this could throw into question the approvability of the combination claim based on nonrandomized antitumor data (given that the value of CPT-11 after CPT-11 might be questioned)."

On September 4, 2001, a BMS Vice President sent an e-mail to other senior BMS executives, stating:

"Based on today's discussions with Susan and Steve our preliminary recommendation is a 'go' decision. We are still trying to obtain data from the mono therapy study from ICE [ImClone]. As of 6:30 PM today we did not have any more information. I will be discussing this with Susan again in the AM."

Despite requests to BMS, Committee staff has not been provided any evidence at this time that shows that BMS obtained the data on the single-agent study prior to making its historic deal with ImClone.

In addition, the BMS independent radiology review of ImClone 9923 study lowered the ImClone reported response rate and the size of the patient pool, both significantly. In an August 30, 2001 e-mail, the BMS independent radiologist noted:

"Attached to this message you will find the latest update of the spread sheet we are using to keep track of our review of the CT's and MRI's of patients enrolled in CP02-9923.

"We are in the process of reviewing a total of 34 cases, 27 of which were initially assigned by the investigator to the PD [progressive disease] cohort and 7 of which were assigned to the SD [stable disease] cohort. To date we have reviewed 23/27 cases from the PD cohort and 6/7 cases from the SD cohort.

"In the PD cohort we can now confirm 14 partial responses. We may have 15, but one case will require adjudication. With 4 more cases to review, and the one case for adjudication, the RR in the PD cohort could be as high as 15 + 4/120 = 15.8%.

"I should mention, however, that in 4 of these confirmed partial responses our radiologists have judged the disease to be only stable at the time of patient's enrollment into the study. If these 4 cases were thrown out, then the highest possible response rate would 11 + 4/120 = 12.5%. However, we have not conducted a strict review of all of the 120 cases, and it is likely that if we carefully reviewed all of the cases we would throw many out on the same basis [emphasis added]. Indeed, it is my understanding that the study sponsor has conducted such an analysis on the basis of its own radiologists' review, and has thereby reduced the denominator of the patient population with radiographically confirmed progressive disease.

"I will review the study sponsor's data and see if I can get at the same denominator [patient pool size] as it did (? N = 89), and calculate the response rate accordingly. More cases and analysis to follow tomorrow... "

It should be noted that, if indeed the denominator in 9923 was below 100 (particularly if it were as low as 89, which the BMS independent radiologist appears to have indicated in the above e-mail), the entire study probably could no longer serve to support an accelerated approval application. As ImClone consultant, Roger Cohen MD, e-mailed to Dr. Harlan Waksal on January 4, 2002:

"9923 is a small study to begin with. It cannot get much smaller and have any hope of serving as a registration study. I think it is clear that it has to have at least 100 fully eligible and evaluable subjects (closer to 100)."

Therefore, although BMS received tentative support from its scientific leadership and outside consultants, it appears that the status of crucial issues were as follows at the time BMS entered into the alliance with ImClone in September 2001:

On the evening of December 28, 2001, ImClone revealed to the public that it had received a refusal-to-file letter from FDA. On December 29, 2001, a Reuters news article reported: "Sam Waksal, ImClone's chief executive officer, told Reuters that the agency first wants more 'annotation' information, about how the company verified that patients enrolled in its trials had indeed failed previous drug regimens and that subsequent tumor reductions attributed to Erbitux were indeed real. Concerns raised by the FDA mainly involve how the data were presented and do not raise outright concerns about safety or efficacy of the drug, the CEO added." An internal BMS e-mail dated December 30, 2001, responding to earlier BMS e-mails on the Reuters article, states: "I agree that some alot [sic] of Sam's comments are misleading and at this point we should continue to be silent. As you heard from yesterday's discussion, there's a lot we don't know."

On that same date, December 30, 2001, another BMS official commented on the draft documents being prepared for the ImClone investor relations conference call: "These draft documents leave me most uncomfortable. They gloss over the seriousness of the RTF letter and make it appear that the integrity of the study results is not in question, when in fact it is.. We will also need to rewrite major portions of the clinical and pharmacology part of the BLA including a new 9923 study report, new 141 (monotherapy) study report, new ISS and ISE based on these revised reports. I know that this is not what ImClone wants to tell their investors, but I think it represents the reality of this situation."

TRADING ACTIVITY OF SAM AND HARLAN WAKSAL, THEIR FAMILY MEMBERS AND CLOSE FRIENDS,AND IMCLONE DIRECTORS

Adding to the controversy over Erbitux has been the trading of ImClone stock by ImClone insiders a few weeks before the FDA refusal-to-file letter, and by Waksal family relatives and friends during the 48 hours before the FDA letter was issued. Committee staff examined public records and conducted interviews with Sam and Harlan Waksal, and with representatives of several of their family members and friends, to determine the degree of trading in ImClone stock by these individuals over the last year. Of particular interest were board members who tendered stock to BMS on October 29, 2001, and whether any board members or officers of ImClone sold stock during the critical month of December 2001. Committee staff also attempted to gather information on those trades of Sam Waksal's immediate family members and close friends that were identified during discussions with Dr. Waksal.

Committee staff found that ImClone board members exercised stock options to acquire 8.1 million shares of ImClone common stock between the period of June 1, 2001 and October 29, 2001. Committee staff examined this time period because it represents the period of negotiations between BMS and ImClone officials regarding an equity purchase of ImClone by BMS. Of these 8.1 million ImClone shares, Sam and Harlan Waksal acquired approximately 4.1 million. Each board member who exercised stock options during this time period is shown in the table below.

STOCK OPTIONS EXERCISED BY IMCLONE BOARD MEMBERS DURING THE PERIOD OF NEGOTIATIONS WITH BMS JUNE 1 THROUGH OCTOBER 29, 2001

It should be noted that ImClone awarded many of these options to the Waksal brothers in 1999 and 2000, and accelerated the vesting of these options with the rise in the stock price. According to ImClone's SEC filings, on May 24, 1999, the stockholders approved the grant of an option to Sam Waksal to purchase 1,000,000 shares and Harlan Waksal to purchase 650,00 shares of Common Stock at a per share exercise price equal to $18.25, the last reported sale price of the Common Stock on the date shareholder approval was obtained at the annual shareholders meeting. The option was to vest no later than six years from the grant date and specified amounts were subject to earlier vesting if specified Company Common Stock price thresholds were met. On May 31, 2000, the stockholders approved amendments to a total of 1,600,000 options that were granted to Sam and Harlan Waksal the year before. The shareholders also approved amendments to a total of 3,300,000 additional options held by Sam and Harlan Waksal. All these options were amended to provide that each tranche vested immediately upon achievement of the relevant stock target price associated with such tranche, without regard to the passage of time that was a requirement in the original options. The options became fully vested and exercisable upon the approval of the amendments. As reported in a previous section, the ImClone board granted the Waksal brothers and two other directors company loans to finance the exercise of their options as part of the tender offer.

In total, Committee staff found that members of ImClone's Board of Directors tendered 2.1 million shares of ImClone common stock at $70 a share to BMS on October 29, 2001. This represents approximately 15 % of the stock tendered by ImClone shareholders to BMS. Sam and Harlan Waksal tendered a total of 1.6 million shares of ImClone stock to BMS for about $111 million. Simply stated, this means that the Waksal brothers received over 10 percent of the entire proceeds paid by BMS during the $1 billion tender offer, and the ImClone Board combined received nearly 15 percent of the proceeds from the BMS tender offer. The table below shows the number of shares tendered and the proceeds for each of ImClone's Board members.

Committee staff also examined trading by ImClone board members and officers during the critical month of December 2001 to determine if any ImClone officials who sold stock had knowledge of discussions with FDA regarding whether the agency would accept the Erbitux filing. We found that, with the exception of Harlan Waksal's disposition of 700,000 shares on December 6, 2001 (discussed below), three officers of ImClone sold stock prior to December 18, 2001. In each case, Committee staff were told that the officials involved were unaware of the details of the FDA review of Erbitux, sold less than 20 percent of their holdings in ImClone, and did so based on their brokers' advice. Even though ImClone has internal rules that require officers of the company to receive pre-clearance before trading in company stock, two of the three trades were not pre-cleared. In one case, the individual was not an officer at the time of the trade, but was since promoted. In the other case, the officer claimed to have simply forgot to pre-clear the trade.

On December 21, 2001, ImClone issued an order prohibiting its employees from trading in ImClone stock until after the FDA decision on Erbitux was made public. ImClone has told Committee staff that no board member or officer of ImClone traded ImClone stock between December 21 and 28, 2001. However, Committee staff discovered that several of Sam Waksal's immediate family members or friends sold ImClone stock on December 27, 2001 -- the day before ImClone announced publicly that FDA had refused to accept the filing of Erbitux. This list of traders included his father, sister, two daughters, and son-in-law. In addition, Committee staff learned from discussions with Sam Waksal that the SEC has questioned him about trades made by three other friends on December 27 or 28, 2001.

With the exception of Sam Waksal's father (who has not yet provided information to the Committee), attorneys for each of the family members admitted that their client sold stock on or around December 27, 2001, but asserted that they received no non-public information about ImClone and each had a reason why they sold the stock on that particular day. Although phone records and logs obtained from Sam and Harlan Waksal, covering the time period December 26-28, 2001, suggest that both men had conversations with each other and may have had conversations with members of their family and friends, both Sam and Harlan Waksal denied that they had tipped off anyone as to their knowledge that ImClone was about to receive a RTF letter from FDA.

On December 6, 2001, Harlan Waksal sold 700,000 shares of ImClone stock. On October 31, 2001, Harlan Waksal notified the ImClone board members that he planned to execute a forward transaction involving 700,000 shares of ImClone common stock:

As a result of my recent option exercise and the sale of stock to Bristol-Myers Squibb I am left with an additional tax burden that I need to meet. As I am averse to having such a great personal liability I plan to meet this obligation (and provide some liquidity), by the sale of additional shares of ImClone stock. I am moving to do this through a prepaid forward contract for the sale of stock. This will be a 700,000 share transaction, the stock will still be under my voting control for the next three years and I will retain some continued upside if the stock continues to perform as we anticipate. I plan on finalizing this transaction over the next two weeks.

Dr. Harlan Waksal told Committee staff that, in November 2001, he attempted to shop the sale of his ImClone stock. Dr. Waksal filed a Form 144 with the SEC, announcing his intention to sell 700,000 shares of ImClone. Dr. Waksal told Committee staff he was forced to sell the ImClone stock to come up with enough cash to pay substantial taxes generated from his prior exercise of stock options and his tendering of shares to BMS. He also stated that, because he did not want to sell shares, he entered into a forward sales contract that gave him a percentage of the cash value of the shares up front but still allowed him to control the shares and defer tax payments for another two years. Simply put, Dr. Waksal received less than what the stock was worth at the time of the sale, but he also limited his downside risk when ImClone's stock price dropped considerably in the month thereafter. It should be noted that Dr. Waksal sold the 700,000 shares on the same day that ImClone's share price hit its 52-week high.

Moreover, in February 2002, Dr. Sam Waksal revealed about 50 unreported stock trades that should have been reported to the SEC and returned to ImClone about $486,000 in profit he made on some sales of company stock because he may have violated an insider-trading regulation.

The key findings from the Committee staff's investigation at this point are as follows:

[1] Some companies meet with FDA before they conduct the clinical trial to seek the agency's input and guidance on the clinical protocol design. Agreements between the company and FDA can be made binding through Special Protocol Assessments. Although FDA's Center for Drugs has used dozens of these assessments for cancer drugs, the FDA's Center for Biologics (the division handling ImClone) had never used one for a biologic product, other than in one instance involving a vaccine.

Raymond B. Weiss, MD, FACP
Independent Consultant in Oncology
Rockville, Maryland
and
Clinical Professor of Medicine
Lombardi Cancer Center
Georgetown University Medical Center
Washington, DC

Since finishing all postgraduate training in internal medicine and the subspecialty of medical oncology, I have spent the past 30 years in clinical academic practice. I have been on the faculty of three medical schools, one of which was full time (in the 1970s) and the other two, part time. I also spent 3.5 years in a senior administrative position at the National Cancer Institute (NCI) and 15 years as the civilian Chief of Medical Oncology at the Walter Reed Army Medical Center (WRAMC). Since 1996 I have been an independent consultant in oncology for a variety of contracts with Federal Government agencies and several private organizations. Since 1981 I have been the Chair of the Data Audit Committee of the Cancer and Leukemia Group B (CALGB), one of the collaborative research organizations funded by the NCI for testing new treatments for a variety of cancers. In the past six years since leaving full-time employment with the Federal Government I have continued this work under a contract with the University of Chicago that is supported by a grant from the NCI. This role involves directing and managing the 20 members of this committee, which makes site visits at least once every 36 months to the nearly 300 institutions participating in CALGB research around the country, auditing the medical records of patients entered on the research protocols. The CALGB accrues 4500+ patients a year to 70+ research protocols ongoing at any one time covering a broad selection of cancers. In addition, I also do this sort of quality assurance work under an NCI contract with the Cancer Trials Support Unit, another entity funded by the NCI to conduct clinical trials with new treatments for cancer. Over the past 21 years in this role as an auditor of scientific research I have personally reviewed the medical records of several thousand patients. I have also been a consultant to review and analyze five instances of fraudulent clinical research involving patients with cancer. Finally, since graduation from medical school I have 37 years of experience caring for patients with all forms of cancer, and I presently work part-time in this capacity at the Walter Reed Army Medical Center and at a private practice with offices in both Westminister, MD and Gettysburg, PA. In the latter venue I personally care for patients with various stages of colon cancer, including widely metastatic disease. For these patients I personally use chemotherapy (including irinotecan) in the hope of palliating their disease and adding some extra months to their survival.

I neither own any stock in ImClone nor know the principals personally. I do not own any stock in Bristol-Myers Squibb (BMS). I am acquainted with several BMS officials involved in development of new oncology drugs. A living trust in my wife's name holds 15 shares of BMS stock (present value about $425), but this account is professionally managed, and she has absolutely no input to any of the financial transactions or decisions to buy or sell shares held in any company in this trust portfolio. In 1993 to 1994 I was the Principal Investigator for a Phase I clinical trial with a new BMS compound while I was employed full time at WRAMC. BMS provided $30,000 to the WRAMC Department of Clinical Investigation for support of this research (and not to me personally). I was supported by BMS to speak at the 13^th Bristol-Myers Squibb Nagoya International Cancer Treatment Symposium sponsored by BMS and held in Nagoya, Japan in October 1997. Other than these items I have had no interaction with BMS. I do know and hold in high esteem some of the oncologists who have been participants in the clinical trials sponsored by ImClone.

The House Committee initiated a review of the clinical trial undertaken by ImClone Systems, Inc. to assess the efficacy and toxicity of a new agent known as cetuximab (Erbitux®, C225) developed by this company for the treatment of metastatic colon cancer. The Committee engaged my services as a physician consultant with experience reviewing clinical research involving new agents for cancer treatment. This Committee investigation grew out of the fact a Refusal to File (RTF) Letter was sent to ImClone by the FDA on 28 December 2001 denying further review of their Biologic Licensing Application (BLA) for accelerated approval of cetuximab. A concern was raised that inappropriate financial transactions by company officials may have taken place in the latter part of 2001, especially around the time of the RTF action. A part of this matter involves the possibility of hyperbole in discussions of this new agent by ImClone officials, especially in regard to its efficacy for colon cancer and how speedily marketing approval would be granted by the FDA.

In my capacity as consultant to the Committee staff I was provided access to a number of the items acquired by the Committee, including case report forms of some of the patients who were said to have responded to cetuximab, confidential communications by ImClone and Bristol-Myers Squibb (BMS) officials, confidential communications by FDA officials, and radiographs of some patients. All items involving patients had names and other identifiers redacted. My charge by the Committee staff was to assess certain aspects of the clinical trial entitled 9923. A copy of the protocol for 9923 was provided as part of my review. One version was 1.0 and was dated 2 August 1999. The other version was 2.0, dated 18 October 1999. In addition, I acquired a number of items myself including media discussions of cetuximab and articles published in the medical literature regarding colon cancer treatment and the preclinical development of cetuximab. My work involved multiple visits to the Committee offices and lengthy discussions with Mr. Alan Slobodin, the lead Committee staff member assigned to this investigation. I reviewed the radiographs of three patients entered on 9923 with a radiologist at WRAMC who is experienced in the interpretation of CT scans of the chest and abdomen.

Colon and rectal cancers are the second (after lung cancer) most common cancer in the U.S. when both sexes are combined. Some 60,000 people will die of this disease in the U.S. this year. Approximately half of all people who develop this cancer will have metastases (spread) of the cancer distant to the primary cancer site in the large bowel or rectum. The liver is the most common site for this cancer to spread. Only a small fraction of patients (perhaps 5%) will not succumb to this cancer once it has spread to other organs. The median survival after the appearance of metastases is approximately nine months with only a rare patient surviving more than three or four years later. The only patients who are cured of metastatic colorectal cancer (CRC) are a very fortunate subset (only some 15% of all patients with metastases) who have isolated metastases in the liver feasible to remove surgically and who survive without developing further metastases (which is only a third of those undergoing such resection surgery).

Since 1959 and until 1996 only one drug had established efficacy for treatment of metastatic CRC, 5-fluorouracil (5-FU). Therapy with this agent has been demonstrated to improve both the overall and symptom-free survival of patients with metastatic CRC when compared to supportive care alone without active antitumor treatment.^1,2 In the 1980s a drug that modulates the metabolism of 5-FU (leucovorin) was found to increase the percentage of patients who achieve significant remissions of their metastatic cancer, but it has not further improved the overall survival of such patients treated with 5-FU. In 1996 another agent, irinotecan (Camptosar®, CPT-11), which was originally developed in Japan, was approved by the FDA for treatment of metastatic CRC. Now two drugs were available to treat patients with metastatic CRC. Not only were there published studies that established the efficacy of 5-FU/leucovorin as initial therapy, but now there were trials^3-6 that established irinotecan as an effective second-line therapy, once a patient's cancer was proven to be unresponsive to further 5-FU/leucovorin therapy. However, unfortunately once both these drugs are no longer effective, there is no other systemic therapy available to treat patients with this ultimately fatal cancer.

Cetuximab is an antibody designed to bind with, and block, the Epidermal Growth Factor Receptor (EFGR), which lies on the surface of cell membranes. The action of the antibody is to block this receptor and therefore ultimately interfere with cell division and proliferation. There is a strong scientific basis for designing drugs to interfere with the action of this receptor.⁷ It is known that these receptors are frequently overexpressed by cancer cells, and it is known that when such overexpression occurs, the cancer tends to proliferate and spread faster. Already one such antibody, known as trastuzumab (Herceptin®), has been established as being a useful drug for the treatment of breast cancer and is marketed for that indication. This agent also has promise for efficacy in a variety of other cancers and is in clinical trial for such evaluation. With receptor-binding antibodies having a strong theoretical basis for anticancer therapy and one instance of high efficacy in the clinical setting, a number of biotechnology companies have developed other antibodies to elements of the family of EGFRs.⁷ Some are probably close to FDA approval for marketing, while others are still in clinical trials.

Approximately 20 years ago Dr. John Mendelsohn created a panel of mouse antibodies to EGFR while he was a faculty member and cancer center director at the University of California at San Diego. He continued his work with this set of antibodies while chief of the Department of Medicine at the Memorial Sloan-Kettering Cancer Center in the 1990s and more recently as CEO of the M.D. Anderson Cancer Center in Houston. The original mouse-derived antibody is immunogenic and likely to induce antibodies against it when administered to human beings that would either cause serious allergic reactions and/or inactivate the antibody so no antitumor effect can occur. Thus, cetuximab was developed as a human:murine chimeric⁸ (meaning two originating elements) of the mouse (murine) antibody that Dr. Mendelsohn created.

All new agents that might have benefit for treatment of human cancers must traverse an extensive process of evaluation in the laboratory setting. Such laboratory work with cetuximab suggested efficacy against colon cancer, and when used in combination with irinotecan, even greater antitumor effect was demonstrated.⁹ It also had antitumor effect against renal (kidney) cancers and cancers of the head and neck area in such laboratory testing. This agent had significant promise for being an effective drug in treatment of human cancers, especially CRC.

This study was an open-label, phase II study designed to "determine the response rate of cetuximab administered in combination with irinotecan to patients with advanced colorectal cancer who are refractory, i.e., have demonstrated stable or progressive disease to treatment with an irinotecan-containing regimen" and "to determine the time to progression, evaluate the safety/toxicity profile of cetuximab in combination with irinotecan [and] assess the Quality of Life" in patients treated with this two-drug combination. "Refractory" to prior therapy means that an adequate attempt to cause tumor regression with a particular therapy has been made, and the cancer progressed despite the treatment. Version 1.0 of the protocol was dated 2 August 1999. Version 2.0 was dated 18 October 1999. It was not originally designed to be a study used as the basis for submission of a BLA for marketing approval. After accrual of most of the patients entered and discussions with the FDA by ImClone officials in 2000, the purpose of the clinical trial was modified so it would serve as a registration study. A meeting was held between ImClone and FDA officials in August 2000 at which time the understanding was that 9923 would be the registration study with a plan for accelerated approval designation. A study used for submission to the FDA for marketing approval (i.e., "a registration study") would have adequate numbers of patients entered to provide proof of the efficacy and safety of the new drug, and the study would be very tightly controlled and conducted so that errors in its conduct or uncertainties about the results would not be issues.

The patient eligibility criteria for any clinical trial are extremely important. Without carefully defining what category of patient is eligible for entry on such a study, any results from the trial will be subject to various biases and likely be meaningless. In addition, the exact method of administering the drugs (dose and schedule) must be defined and adhered to. The eligibility criteria in Version 1.0 stated the patient must have demonstrated "progression of disease [metastases] after completing a minimum of two courses of a regimen containing irinotecan." The definition of a "course" of irinotecan was not stated. These eligibility criteria were changed in Version 2.0 of the protocol. In the newer version the patient had to have "documented stable disease (must have received a minimum of 12 weeks of irinotecan therapy) or progressive disease at any time after receiving an irinotecan-containing regimen." The irinotecan dose and administration frequency were to be the same as was being used for the patient when progressive disease occurred prior to entry on the trial.

Documents regarding the August 2000 FDA meeting indicate FDA officials had concerns about the study design and whether there was sufficient documentation a patient had clearly failed irinotecan therapy before study entry. The whole scientific basis for clinical use of this new drug was that the combination of irinotecan and cetuximab represented a potentially effective, third-line therapy for patients with metastatic CRC after failing prior 5-FU and irinotecan therapy. The ImClone officials stated their belief that "there exists a core of patients who had clearly refractory disease for whom the evidence of antitumor activity is compelling" in the results of the study conducted to that time point. In order to prove that both drugs had to be administered together to obtain an antitumor effect (while accepting the potential for toxicity of both drugs), the patient's cancer had to demonstrate clear resistance to any further therapy with irinotecan. The eligibility criteria as understood by FDA officials (according to minutes of the meeting in August 2000) were that patients would have to have either stable disease defined as <25% volume change in the measurable cancer lesions or progressive disease defined as a >25% change "after two courses of irinotecan." This latter point is what the protocol Version 1.0 stated. There is a difference in the definition of what constitutes an eligible patient between Version 2.0 of the protocol and the understanding of the FDA officials at this meeting. Version 2.0 (dated 18 October 1999) loosened the eligibility requirements to "progressive disease at any time after receiving an irinotecan-containing regimen." No minimum amount or duration of therapy with irinotecan was required in Version 2.0 The final conclusion of the FDA officials was that the study design was "probably acceptable."

Study 9923 Conduct

There were 139 patients entered on 9923, 121 with progressive cancer after irinotecan and 18 with stable disease. Somewhere one patient was deleted, because all reports subsequent to 2000 indicate there were 120 patients with progressive disease who were treated on this study.

During the study a contract research organization (PharmaNet) had research monitors verifying the data collected at the participating institutions. There were a number of deviations from the study protocol cited by these monitors, but in my opinion many were trivial in nature (e.g., being one day off in recording the date of a blood test). However, there were some that appear to be substantive, such as entering patients who did not meet the eligibility criteria. In January 2002 BMS staff reviewed and critiqued the BLA. According to Table 1 of this review an incredible 37 patients (26.6% of the 139 patients entered) "had at least one inclusion/exclusion" criterion "that did not qualify them to be eligible for the study," and eight of them "had more than one reason for ineligibility." Twenty-five of these 37 patients had initial blood counts or serum chemistry values that were outside the range required by the protocol. Another incredible point is the fact that 15 of these 25 patients "were given exemptions to be enrolled in the study." The purpose of eligibility criteria for a study is to define what patients have organ function and disease parameters that make them a suitable candidate for the clinical trial. Once these criteria are set, exemptions are not given to allow patients to be entered who don't meet the established criteria. If this is done, it could invalidate the results of the study. If the criteria are too restrictive, thus making it difficult to enroll patients who meet them, one might amend the study to loosen the criteria or perhaps just stop the study as being unworkable. Eligibility exemptions are forbidden in all the clinical trials with which I have experience. Rates of ineligibility should not be more than single digit percentages in any clinical trial. When an ineligible patient is entered, it is usually an error on somebody's part in which a particular point was overlooked or forgotten and should be an uncommon event.

Another set of major deviations in the study was changing the dose and administration frequency of the irinotecan. This drug was supposed to be administered in the same pattern as had been done before the patient went on the 9923 study. Irinotecan is most often administered in a schedule of four consecutive weekly doses, with a 14- to 21-day break before another series of four consecutive weekly doses is begun. However, it may also be given in a schedule of once every three weeks. Thus, there were variations in the manner patients might receive the irinotecan on the study. There are directions in the 9923 protocol regarding delaying one or more of the weekly cetuximab infusions for any significant toxicity that might occur, but there were no directions for modifying the irinotecan dose or frequency. The treating physician would thus make ad hoc decisions regarding this point, with multiple variations based on the physician's best judgement. It is standard practice in cancer treatment protocols to provide specific directions for changes in the drug doses and/or treatment frequency based on the degrees and kinds of therapy toxicity encountered. In my opinion this point is a design flaw in the 9923 protocol that could lead to problems in interpreting the results.

The protocol deviations in the irinotecan dosing were delineated in Table 3 of the BLA review by BMS staff. Although the protocol specified that the irinotecan was to be given in the same dose and schedule as previously when disease progression occurred, with no dose increases, at least 17 patients had major changes in the irinotecan dose when entered on the study, including dose increases. This fact adds further uncertainty regarding the validity of any results from this study.

Flaws in the design of the 9923 protocol were also expressed publicly by three prominent medical oncologists after the publication of the RTF.¹⁰ For example, one oncologist stated: "Overall, this is a protocol that asks the wrong questions, and then is not tightly written and efficient. The protocol generates far more questions than it could ever answer. It is a blueprint for the production of vague answers."¹⁰ Another oncologist stated that "the entry criteria on the study were so vague it can't be determined whether all the patients in the trial are indeed refractory to prior therapy."¹⁰

An independent panel of two medical oncologists and two radiologists was convened by ImClone to review the case records and the radiographs (the Independent Response Assessment Committee or IRAC) and evaluate the responses, or lack thereof, of all patients entered whether counted as "progressive disease" or "stable disease" on the irinotecan therapy given prior to study entry. Many of these same radiographs were then reviewed by consultants to BMS. A comparison of these two sets of evaluations indicates the subjectivity that can occur in making assessments of the same CT scans. As with much of medicine, assessing response of cancer lesions to therapy is not an exact science. For example, eight patient cases the IRAC had categorized as achieving a Partial Response (PR), which is defined as at least 50% regression of the measurable tumor lesions visible on serial radiographic studies (almost always CT scans), were categorized by the BMS consultants as achieving only stable disease (SD). A total of 23 patients were categorized as achieving a PR by the investigators, while 27 were so categorized by the IRAC. Twenty of these patients were considered to have a PR by both the investigators caring for the patients and the IRAC for an overall response rate of 16.5%. Of the 121 patients coded as having disease progression prior to entry, the IRAC and BMS agreed that a PR had been achieved in only 16 cases. Now the response rate was only 13.2% where both sets of consultants agreed. In addition, three patients whose response after treatment with irinotecan and cetuximab was called SD by the IRAC had it changed to "progressive disease" by the BMS review. The number of responders where both BMS and the IRAC would agree with the interpretation of the scans is now possibly below the point of real meaning. Most clinical oncologists would agree that at least 15% of patients treated with an agent should achieve a PR to be meaningful. In fact, the ImClone officials themselves discussed with the FDA officials in the August 2000 meeting that at least a 15% response rate must be achieved to be "clinically meaningful." A response rate lower than 15% would only be important if a randomized study with half the patients receiving a new treatment and half receiving only supportive care indicated a significantly longer survival for the treated group. Such is indeed the case with irinotecan, as has been established scientifically.^3,5,6 Although the response rate of irinotecan in patients who had failed 5-FU therapy was only approximately 13%,³ overall survival was significantly improved by irinotecan therapy when compared in a randomized study to supportive care only (without systemic anticancer therapy).^5,6 Of course, such could also be the case with cetuximab if it were subjected to the same sort of randomized study as has been accomplished with irinotecan.^5,6

In the context of the disparities regarding which patients achieved a response and to what degree, it is worth quoting the statements made by Dr. Sam Waksal at a conference call to the financial community on 31 December 2001. He said the IRAC came to a similar conclusion about responders as did the investigators. He further stated that all CT scan films had been reviewed "internally by us, they have been reviewed by the sites themselves where the conclusions were made and by the IRAC, and again there is concordance across the board" (italics are mine). Overall, there were 38 patients where the category of disease status prior to study entry was in disagreement between the IRAC and the investigators. In addition, of the 35 patients whose radiographs were reviewed by BMS consultants, there was disagreement between the IRAC and BMS consultants in the response category for 14 cases, which is more than a third of the sample.

The results of this study were published in an abstract¹¹ submitted for the annual meeting of the American Society of Clinical Oncology, held each May. This society is the premier organization worldwide for professionals involved in cancer research and the treatment of cancer in patients. The abstract submission deadline is early in the month of December prior to the meeting (in this case it would have been December 2000). The abstract stated that the patients were refractory to both irinotecan and 5-FU, and 121 patients were said to have been entered.¹¹ Whenever any study data are presented at this meeting, only 10 minutes are allowed for the oral presentation. Thus, only a limited amount of information can be presented. It was stated patients were entered who had "documented progression of metastatic disease on irinotecan" and " no intercurrent chemotherapy could have been given between irinotecan failure and protocol entry." The abstract states 121 patients were entered, but the oral presentation involved only 120 patients. A total of 27 of these 120 patients (22.5%) were said to have achieved a PR, which is the number determined by the IRAC.

I reviewed with the WRAMC radiologist the film sets of the three cases where selected CT scan pictures of metastatic lesions were shown at the ASCO presentation. These were Cases #615, #644, and #683. It is noteworthy that Case #644 was coded as having achieved only a SD status by the IRAC after treatment on the study. Although this patient indeed had regression of some metastatic lesions in the lungs, a pelvic mass was at the same time invading and encroaching more on the urinary bladder in the pelvis. It is also noteworthy that Case #615 had irinotecan therapy only from 15 November 1999 to 6 December 1999, and assuming the drug was given weekly, only four doses could have been given. A chest X-ray done on 4 January 2000, a month after the last irinotecan dose, did indeed show a new nodular density in the right mid lung, indicating cancer progression. Although this patient did meet the revised eligibility criteria of the study in Version 2.0 of the protocol, he would not have been eligible based on the understanding of the FDA of the eligibility criteria where the patient would have had to be treated with "two courses" of irinotecan. If the drug is given weekly in four consecutive weeks, then this would constitute a "course." A second "course" of four weekly doses would then be given after a rest interval without treatment of 14 to 21 days. This patient did not receive two courses of irinotecan therapy prior to entry on the 9923 study.

An example of clear ineligibility for this study is Case #643. This patient received his last dose of irinotecan on 31 March 1999. He was then treated with oxaliplatin (another investigational agent for CRC) between June 1999 and August 1999. It must be recalled that according to the protocol no other chemotherapy should have been administered between the time the patient was last treated with irinotecan and the time of study entry, a point that was reiterated in the ASCO abstract presentation.¹¹ Nonetheless, he was entered on the study on 14 March 2000.

Another example of problems with this study is Case #704. The ASCO abstract¹¹ states the patients were refractory to both irinotecan and 5-FU. The case report form for this particular patient indicates that the only chemotherapy the patient had received prior to study entry was irinotecan. There is no evidence the patient ever received 5-FU.

The serial radiographs of Patient #683 were reviewed. This patient had clear progression of his cancer on irinotecan, so he met the eligibility criteria of 9923 in this regard. In response to the 9923 therapy he had definite shrinkage of the cancer lesions in the liver, which regressed at least 50%, but the response lasted only three months before disease progression occurred once more. This is a rather short interval, but one must recall the patients treated on this study had undergone much prior systemic therapy and sometimes radiation therapy also. Such a short interval of response would be expected from a drug with some modest antitumor efficacy but not one that had been espoused as another blockbuster anticancer agent.

In order to assess the effect of cetuximab as a single agent, this study was initiated in early 2001. A total of 57 patients were entered on this study, the results of which were presented¹² at the May 2002 meeting of ASCO. The patients entered had to have "documented progressive disease at any time after receiving an irinotecan-containing regimen." The title of the abstract¹² presented states the patients were refractory to irinotecan, and six patients were stated to have achieved a PR.

The BMS review of the BLA in January 2002 indicates that there was uncertainty regarding the fact that all patients were truly refractory to irinotecan (meaning they had "documented progressive disease") before being entered on the study. The BMS reviewers stated that "irinotecan refractoriness can be inferred" for 11 of the 57 patients, but "the data collected in this trial are insufficient to determine irinotecan refractoriness for any patient." Two of these 11 patients were verified to have had a PR by the BMS review.

Although six patients were stated to have responded to cetuximab given by itself in this study,¹² the BMS review indicates that one of these patients may not truly have had a response. The final opinion of the BMS staff was "there are five patients....whose data compellingly support a determination of partial response to the single agent" cetuximab. Although it is apparent this agent does have some antitumor activity by itself, the rate of such responses, with a solid assessment of the response, is only five (8.7%) of the 57 patients.

It appears that cetuximab has some antitumor effect for metastatic CRC when used as sole therapy as reported in an ASCO presentation in May this year.¹¹ Cetuximab also appears to have some effect when given in conjunction with irinotecan despite disease progression having occurred when the patient was treated previously with irinotecan, as was reported at the May 2001 ASCO meeting.¹⁰ However, for some patients it is unclear whether or not the irinotecan makes any contribution to the therapy. The irinotecan perhaps only adds toxicity to the therapy and no benefit. The 9923 study has major problems in adherence to the eligibility criteria and the irinotecan dosing. In addition, the assessments of response are subject to considerable variation depending on who reviews the CT scans. After examining a great deal of the information assembled by the House Committee staff, I agree with the assessment of the three oncologists as published.⁹ Based on the results of the 9923 study available for my review, I am unable to determine if this drug has meaningful activity in CRC and adds to patient survival after failure of all available standard therapy. The single-agent study does show the drug has an effect for a rare patient, but a reliable response rate is <10%, a level that possibly provides little patient benefit or improved survival. Unfortunately, further studies will have to be initiated before all these issues are more certain. It is my understanding that BMS and ImClone plan to conduct such further studies.

BACKGROUND

THE BRISTOL-MEYERS SQUIBB DEAL AND IMCLONE'S INTERNAL LOAN

THE RTF LETTER

TRADING ACTIVITY BY IMCLONE EXECUTIVES AND OTHERS

AN INQUIRY INTO THE IMCLONE CANCER-DRUG MATTER PRELIMINARY COMMITTEE STAFF REPORT

Study 9923 Conduct

References