Back to GetFilings.com

SECURITIES AND EXCHANGE COMMISSION

WASHINGTON, D.C. 20549

FORM 10-K

ANNUAL REPORT PURSUANT TO SECTION 13 OR 15(d)

OF THE SECURITIES EXCHANGE ACT OF 1934

MGI PHARMA, INC.

(Exact name of registrant as specified in its charter)

Indicate by check mark whether the registrant (1) has filed all reports required to be filed by Section 13 or 15(d) of the Securities Exchange Act of 1934 during the preceding 12 months (or for such shorter period that the registrant was required to file such reports), and (2) has been subject to such filing requirements for the past 90 days.  Yes  x            No  ¨

Indicate by check mark if disclosure of delinquent filers pursuant to Item 405 of Regulation S-K is not contained herein, and will not be contained, to the best of registrant’s knowledge, in definitive proxy or information statements incorporated by reference in Part III of this Form 10-K or any amendment to this Form 10-K.  x

Indicate by check mark whether the registrant is an accelerated filer (as defined in Exchange Act Rule 12b-2). Yes  x            No  ¨

The aggregate market value of voting and non-voting common equity held by non-affiliates of the registrant as of June 30, 2003, the last business day of the registrant’s most recent second quarter, was approximately $669,000,000 (based on the closing price of the registrant’s common stock as reported by the Nasdaq National Market on such date).

The number of shares outstanding of each of the registrant’s classes of common stock, as of February 20, 2004, was: Common Stock, $.01 par value; 35,040,426 shares.

DOCUMENTS INCORPORATED BY REFERENCE

Pursuant to General Instruction G the responses to Items 10, 11, 12 and 14 of Part III of this report are incorporated herein by reference to certain information contained in the registrant’s definitive Proxy Statement for its 2004 Annual Meeting of Shareholders to be held on May 11, 2004.

PART I

This Form 10-K contains forward-looking statements within the meaning of federal securities laws that may include statements regarding intent, belief or current expectations of the Company and its management. These forward-looking statements are not guarantees of future performance and involve a number of risks and uncertainties that may cause the Company’s actual results to differ materially from the results discussed in these statements. Risks and uncertainties that might affect our results are detailed from time to time in the Company’s filings with the Securities and Exchange Commission, and are included in Item 7. Management’s Discussion and Analysis of Financial Condition and Results of Operations – Cautionary Statement of this Form 10-K. The Company does not intend to update any of the forward-looking statements after the date of this Form 10-K to conform them to actual results.

Item 1.    Business

Overview

MGI PHARMA, INC. (“MGI” or “Company”) is an oncology-focused biopharmaceutical company that acquires, develops and commercializes proprietary pharmaceutical products that meet cancer patient needs. It is our goal to become a leader in oncology through application of our three core competencies of oncology product acquisition, development and commercialization, which we apply toward our portfolio of oncology products and product candidates. We acquire intellectual property or product rights from others after they have completed the basic research to discover the compounds that will become our product candidates or marketed products. This allows us to focus our skills on product development and commercialization rather than directly performing drug discovery.

We currently market several cancer-related products in the United States using our 104 person oncology-focused sales organization. (See Item 7, Management’s Discussion and Analysis of Financial Condition and Results of Operations—Revenues for a breakdown of sales by product over the last three years.) We focus our sales efforts solely within the United States, where we have retained product rights to our currently marketed products and product candidates under development. We create alliances with other pharmaceutical or biotechnology companies for the sales and marketing of our products in other countries. (See Note 13 to the financial statements for further information on revenues attributable to U.S. and foreign customers.)

We began promotion of Aloxi^™ (palonosetron hydrochloride) injection in September 2003 in the United States for the prevention of chemotherapy-induced nausea and vomiting (“CINV”). We market Salagen^® Tablets (pilocarpine hydrochloride) in the United States to oncologists as a treatment for the symptoms of radiation-induced dry mouth in head and neck cancer patients and to rheumatologists as a treatment for dry mouth associated with the autoimmune disease Sjögren’s syndrome. Historically, the cash flow generated from the U.S. sales of Salagen Tablets has been the primary source of funding for our selling, general and administrative activities. Sales of Salagen Tablets accounted for approximately 67 percent of our $39.9 million of product sales in 2003. In March 2001 we began direct promotion, or face-to-face sales calls, of Hexalen^® capsules after acquiring the Hexalen capsules business in November 2000.

Hexalen capsules are a second-line chemotherapy for ovarian cancer patients who are refractory to first-line therapies.

We believe we have a complementary portfolio of oncology product candidates. Our current product candidates are at various stages with an emphasis on advanced stages of development and are intended to have diverse roles in treating cancer patients. Our portfolio includes therapeutic and supportive care product candidates.

On November 17, 2003, we and Helsinn Healthcare SA announced an expansion of our exclusive United States and Canada license and distribution agreement to include rights for the post operative nausea and vomiting (“PONV”) application of Aloxi injection and an oral Aloxi formulation. Phase 2 studies of Aloxi injection for

2

prevention of PONV indicated that Aloxi injection was safe and well tolerated. We expect phase 3 trials to begin in 2004 for use of Aloxi injection in the prevention of PONV and the oral Aloxi formulation.

Irofulven, the lead cancer therapy product candidate from our proprietary family of compounds, called acylfulvenes, is in a series of clinical trials. Based on our analysis of results from our trials of irofulven, we believe that irofulven is adequately tolerated for a chemotherapeutic, with evidence of monotherapy activity against a wide range of cancers including liver, pancreatic, ovarian, and prostate tumors. These results have been seen in patients with refractory tumors, meaning tumors that are unresponsive to first-line chemotherapy. Trials in refractory cancer patients are the usual way to begin development of new chemotherapy agents, followed by trials comparing the activity of currently approved therapies to the investigational chemotherapy agent. If the investigational chemotherapy agent establishes improved safety and/or efficacy compared to the established therapy, regulatory approval may be sought.

We also are conducting a series of phase 1 dose escalation trials of irofulven in combination therapy with currently approved cancer agents. We believe that an advantage of combination therapy is the potential to achieve better anti-cancer benefit with an acceptable side effect profile compared to either agent used as single agents. In preclinical studies, irofulven has demonstrated synergistic activity with a number of marketed chemotherapies, meaning the extent of tumor cell death was greater than would be predicted by simply adding the activity of each agent individually.

The most promising combination therapy clinical activity has been observed in colorectal, hormone-refractory prostate, and thyroid cancers. We plan to initiate phase 2 combination studies of irofulven in patients with hormone refractory prostate cancer during 2004 as a potential path toward U.S food and Drug Administration (“FDA”) registration. Throughout phase 1 and 2 monotherapy and combination therapy trials, consistent evidence of clinical activity in hormone refractory prostate cancer (“HRPC”) has been reported, including activity in patients previously exposed to chemotherapy. Phase 2 data presented at the 39^th Annual Meeting of the American Society of Clinical Oncology (“ASCO”) in 2003 indicated that irofulven in combination with prednisone, a steroid, induced decreases of more than 50 percent in prostate specific antigen (“PSA”) levels, in 25 percent of evaluable patients and disease stabilization in 86 percent of patients evaluable for objective tumor response. PSA is a blood marker used by clinicians for treatment decisions. Activity has also been shown in prostate cancer patients treated with irofulven plus the marketed chemotherapies, capecitabine or cisplatin, in phase 1 trials. Following comprehensive analysis of our available data, we are preparing to initiate a phase 2 combination program for irofulven in HRPC patients previously treated with docetaxel.

Most HRPC patients in the United States who are receiving chemotherapy are treated with docetaxel, and we believe significant unmet need exists among prostate cancer patients who have failed hormone therapy and who have progressed on docetaxel.

3

The following table summarizes the principal indications and commercial rights for our currently marketed products and development programs.

We were incorporated under the name Molecular Genetics, Inc. in Minnesota in November 1979.

Business Strategy

Our goal is to become a leading oncology-focused biopharmaceutical company serving well-defined markets. The key elements of our strategy are to:

4

Cancer Overview

According to the American Cancer Society, or ACS, there are approximately 1.3 million new cases of cancer diagnosed in the United States each year. Cancer is the second leading cause of death in the United States and is projected to result in approximately 556,500 deaths in the United States in 2003.

Cancer is characterized by the uncontrolled growth and spread of abnormal cells. These abnormal or malignant cells accumulate and form tumors that can compress, invade and destroy normal tissue. If malignant cells break away from the primary tumor, they can travel through the bloodstream or lymphatic system to other areas of the body. There they may settle and form new tumors. The spread of a tumor to a new site is called metastasis.

Different types of cancer vary in their rates of growth and patterns of spread, and, consequently, typically respond in varying degrees to different types of treatment. The three most common forms of treatment for cancer in order of their typical use are:

A cancer patient often receives a combination of treatments depending upon the type and progression of the disease. While surgery attempts to remove the cancer from the patient and radiation attempts to kill the cancer cells, there are significant limitations and complications associated with these treatments that result in high rates of treatment failure. This failure is due primarily to metastasis and dose-limiting severe side effects. Chemotherapeutic agents attempt to address the limitations of surgery and radiation, which are local treatments, by interfering with the replication of cancer cells that may have spread to distant sites in the patient. In addition to treatment of their cancer, cancer patients often need supportive care to prevent or treat the side effects of radiation or chemotherapy. Examples include treatment of chemotherapy-induced nausea and vomiting, pain control or stimulation of blood cell growth.

Cancer is a disease characterized by uncontrolled cell replication, which requires cells to first replicate their DNA. Therefore, many chemotherapeutic agents target the cancer cells’ ability to replicate DNA, or following the replication of their DNA, the ability of the cancer cells to divide. Different classes of chemotherapeutic agents are distinguished by their mechanism of action or how they specifically interfere with the cancer cells’ ability to replicate DNA or divide.

When a cancer cell’s DNA is damaged by a chemotherapeutic agent, DNA synthesis is inhibited or cell division is inhibited, and a cellular process known as apoptosis, or programmed cell death, may be activated. Apoptosis of tumor cells can lead to reduction in tumor size or to the arrest of tumor growth. Certain chemotherapeutic agents may act to directly promote apoptosis in tumor cells.

Because different chemotherapeutic agents may target different cellular processes required for DNA replication and cell division or of apoptosis, chemotherapeutic agents are often used in combination to maximize tumor cell death or inhibition of growth, while more effectively managing the side effect profile of the individual agents. Agents that specifically induce apoptosis or interfere with DNA replication or cell division in novel ways are therefore excellent candidates to be used in combination with existing chemotherapy agents.

5

One of the principal causes of chemotherapy treatment failure is the development of drug resistance by cancer cells, where cancer cells become resistant or refractory to the intended cytotoxic action of a variety of conventional chemotherapeutic agents. In many cases, resistance developed to a specific chemotherapeutic agent results in multi-drug resistance where the cancer cell becomes cross-resistant to a wide variety of chemotherapeutic agents. Given the current limitations of chemotherapy, there is a clear need for new therapies that are effective against a broad range of resistant and refractory cancers, as well as chemotherapeutics that act by novel mechanisms that can offer benefits as a combination therapy with existing chemotherapeutic agents.

Standard response criteria are used to report the results of oncology clinical trials. In solid tumor clinical trials, a complete response means that all measurable tumor tissue has disappeared and the patient appears to be disease free. A partial response means that measurable tumor tissue has shrunk by at least 50 percent. Stable disease means that the size of the measurable tumor tissue has not shrunk sufficiently to be considered a partial response, but it has not grown more than 25 percent from its smallest size during treatment. Progressive disease means that the tumor has grown by more than 25 percent from its smallest size during treatment.

Developed Products

Chemotherapy-Induced Nausea and Vomiting Overview

Depending on the type of cancer and treatment goals determined by physicians, patients may receive chemotherapy as part of their treatment regimen. One of the most feared side-effects of most chemotherapy treatments is chemotherapy-induced nausea and vomiting, or CINV. In recent years supportive care products to treat the side-effects of chemotherapy, such as CINV, have emerged to improve patient comfort and compliance with treatment regimens. While there are many types of supportive care products for cancer patients, this discussion is focused on the prevention of CINV.

Efforts to treat tumors such as those found in breast and lung cancer have led physicians to administer more aggressive chemotherapy regimens. These cytotoxic agents often cause CINV by triggering release of serotonin from certain cells in the gastrointestinal tract. The released serotonin stimulates nerve receptors that activate the vomiting center via the chemoreceptor trigger zone. When, and if, serotonin stimulates the vagal afferents through the 5-HT₃ receptors, vomiting (emesis) ensues. Although CINV has been managed to a greater degree in recent years, it is estimated that up to 85 percent of cancer patients receiving chemotherapy will experience some degree of emesis if not prevented with an antiemetic. The severity of emesis is dependent upon the type of chemotherapy administered, the dosing schedule of the chemotherapy, how quickly it was administered, and the patient’s age and gender, among other predisposing factors. If emesis is not properly managed, it can cause dehydration and poor quality of life, eventually leading to interruption or discontinuation of chemotherapy. Most chemotherapy regimens are classified as low or moderately-emetogenic and, although the majority of patients receiving moderately-emetogenic chemotherapy are administered a currently available 5-HT₃ antagonist, there remains a need to improve upon the prevention of acute, within 24 hours of chemotherapy, CINV and, especially, delayed, more than 24 hours after chemotherapy, CINV.

Aloxi Injection for the Prevention of Chemotherapy-Induced Nausea and Vomiting

Aloxi (palonosetron hydrochloride) injection is a potent, highly selective serotonin subtype 3, or 5-HT₃, receptor antagonist differentiated by its strong receptor binding affinity and extended half life for the prevention of CINV. We obtained exclusive U.S. and Canadian Aloxi injection license and distribution rights from Helsinn Healthcare SA in April 2001. On July 25, 2003, approval was received from the FDA to market Aloxi injection for the prevention of acute and delayed CINV. We launched Aloxi injection in September 2003 through our 104-person oncology-focused sales organization. Aloxi injection competes in the growing U.S. 5-HT₃ receptor antagonist market estimated to be $1.5 billion in 2003. Of this amount, the market for prevention of CINV is estimated to be approximately $900 million.

A primary cause of CINV is the release of serotonin in the body in response to chemotherapy. 5-HT₃ receptor antagonists, like Aloxi injection, act by binding to serotonin receptors in the peripheral and possibly central

6

nervous system involved with nausea and vomiting, thereby blocking serotonin stimulation of these nerves and reducing or eliminating CINV. CINV can be characterized as acute nausea and vomiting occurring zero to 24 hours post-chemotherapy, or delayed nausea and vomiting occurring 24 to 120 hours post chemotherapy. Despite the availability of preventative treatments, including first generation 5-HT₃ receptor antagonists, nearly fifty percent of all patients receiving chemotherapy experience nausea and vomiting.

The results of phase 3 clinical trials demonstrate that Aloxi injection is more effective than ondansetron (Zofran^®), the current 5-HT₃ receptor antagonist market leader by sales, and dolasetron (Anzemet^®), the number two product by sales, for treating CINV due to moderately emetogenic chemotherapy. The most frequently prescribed chemotherapies, including those used to treat the most common cancers such as breast, lung and colon, are considered moderately emetogenic or having a moderate to moderately-high potential to cause nausea and vomiting. Based on the phase 3 trials conducted, Aloxi injection is shown statistically to have greater efficacy or to be as effective as ondansetron and dolasetron for prevention of acute and delayed CINV. Overall, the incidence, pattern, duration, and intensity of adverse reactions were similar among patients treated with Aloxi injection and other 5-HT₃ receptor antagonists. In 633 patients treated with Aloxi injection during phase 3 trials, the most common adverse reactions related to the study drug were headache (9%), and constipation (5%). The table below provides a summary of the efficacy results from these pivotal phase 3 clinical trials, where complete response rate is defined as the proportion of treated patients which had no vomiting and no rescue medication.

* Indicates results that demonstrate statistically greater efficacy than the alternative 5-HT₃ product compared in trial.

We believe that Aloxi injection will be competitive in the U.S. CINV market for 5-HT₃ receptor antagonists given that Aloxi injection:

The potency and extended half-life of Aloxi injection may enable patients and their healthcare providers to control both acute and delayed CINV for several days following chemotherapy with a single fixed intravenous dose. We believe this single fixed dose treatment will be more convenient compared to currently marketed 5-HT₃ receptor antagonists, which usually demand patients to adhere to an oral follow-up therapy regimen for several days.

Salagen Tablets for the Symptoms of Xerostomia and Sjögren’s Syndrome

MGI conceived, developed and markets Salagen Tablets (pilocarpine hydrochloride) in the United States. Salagen Tablets’ clinically-proven efficacy and safety allow it to maintain leadership in the market for treatment

7

of chronic dry mouth symptoms associated with head and neck cancer patients treated with radiation and with Sjögren’s syndrome patients.

Salagen Tablets were the first prescription drug approved to treat the symptoms of chronic dry mouth in these patient populations. Chronic dry mouth can be a painful and debilitating condition. Salagen Tablets stimulate the exocrine glands, including the salivary glands, to increase their moisture-producing activity. Saliva is important to oral health and quality of life in general. People with chronic dry mouth can experience difficulty eating and sleeping, rapid tooth decay, periodontal disease and oral infections. Sales of Salagen Tablets in the United States were $26.5 million in 2003. Underlying demand for Salagen Tablets in the United States, as measured by prescription growth, grew at an annual rate of approximately one percent in 2003 over 2002.

Head and neck cancer

Although often effective in treating primary tumors of the head and neck, radiation therapy can permanently damage a patient’s salivary glands, resulting in xerostomia, a significant, chronic reduction of saliva production. Patients using Salagen Tablets for radiation-induced dry mouth typically take one tablet three times per day during the course of radiation therapy, which will last approximately six to eight weeks and may then take it indefinitely to treat the residual dry mouth symptoms that follow radiation therapy.

Salagen Tablets have been shown to stimulate the residual functioning tissue in the damaged salivary glands to increase saliva production and provide patients with a longer-lasting solution for the symptoms of radiation-induced chronic dry mouth. In two 12-week trials that were the primary basis for the approval of Salagen Tablets in 1994, 369 patients who had been treated with radiation therapy for head and neck cancer were assessed for the ability of Salagen Tablets, versus placebo, to relieve the symptoms of dry mouth and to stimulate saliva production. In both studies, patients who received Salagen Tablets experienced significant improvement in their overall condition of dry mouth. Those patients also demonstrated statistically significant improvements in salivary flow compared to the patients receiving placebo tablets. Less than one percent of the patients who received Salagen Tablets withdrew from these studies due to lack of efficacy. Sweating was the most commonly reported side effect; however, less than one percent of patients taking the approved dosing regimen withdrew from the study due to sweating.

Sjögren’s syndrome

Sjögren’s syndrome is a chronic autoimmune disorder in which the body’s own immune system attacks the moisture-producing glands, including the salivary glands, causing them to lose their ability to produce adequate moisture. Symptoms of Sjögren’s syndrome vary in degree and type, but the common component is chronic dryness. Patients can exhibit dry mouth, swollen glands, dry eyes, vaginal dryness and fatigue. The syndrome can be manifested alone (primary Sjögren’s) or in combination with other autoimmune disorders (secondary Sjögren’s). We believe that approximately 50,000 of the 200,000 Sjögren’s syndrome patients could benefit from Salagen Tablets.

In patients with Sjögren’s syndrome-related dry mouth, Salagen Tablets can help to relieve the symptoms of oral dryness. Salagen Tablets can stimulate the salivary glands to increase production of saliva, which is essential to maintaining good oral health. For Sjögren’s syndrome patients, previously available therapies included tear and saliva substitutes. These types of products provide transient relief at best and often fail to prevent complications.

In two 12-week trials that were the primary basis for the supplemental approval of Salagen Tablets for Sjögren’s syndrome in 1998, a total of 629 primary or secondary Sjögren’s syndrome patients were assessed for the ability of Salagen Tablets, versus placebo, to relieve the symptoms of dry mouth and to stimulate saliva production. Patients receiving Salagen Tablets four times a day reported a significant improvement in their symptoms associated with oral dryness, and they also demonstrated a significant increase in saliva for the full 12 weeks. Similar to the Salagen Tablet trials in head and neck cancer patients, less than one percent of the patients

8

receiving Salagen Tablets withdrew from the trial due to lack of efficacy. The most common side effect was mild to moderate sweating. Less than four percent of patients taking the approved dosing regimen withdrew from the study due to sweating.

The FDA granted us orphan drug status for Salagen Tablets in 1994 as a treatment for the symptoms of xerostomia induced by radiation therapy in head and neck cancer patients and in 1998 for the symptoms of dry mouth associated with Sjögren’s syndrome. Our orphan drug protection for Salagen Tablets for the treatment of symptoms of radiation-induced xerostomia in head and neck cancer patients expired in March 2001 and our orphan drug protection for Sjögren’s syndrome will expire in 2005. Expiration of our orphan drug protection for Salagen Tablets may result in competition from manufacturers of generic versions of Salagen Tablets.

Hexalen Capsules for Ovarian Cancer

In November 2000, we purchased worldwide rights to Hexalen (altretamine) capsules from MedImmune Oncology, Inc. Hexalen capsules are an orally administered chemotherapy that is approved as a second-line treatment of ovarian cancer. Hexalen capsules are approved for the treatment of ovarian cancer in 21 countries including the United States. Sales of Hexalen capsules in the United States were $2.8 million in 2003.

In the two trials that were the primary basis for its approval in the United States, Hexalen capsules were administered as a single agent for 14 or 21 days of a 28-day cycle. In the 51 patients with measurable or evaluable disease, there were seven complete responses and two partial responses for an overall response rate of 18 percent. The duration of these responses ranged from two months in a patient with a palpable pelvic mass to 36 months in a patient who achieved a complete response. In some patients, tumor regression was associated with improvement in symptoms and performance status. Side effects of Hexalen capsules are comparable to those seen with other approved chemotherapies and include mild to moderate bone marrow suppression, nausea and vomiting, and peripheral sensations of touch.

A more recent trial in 97 ovarian cancer patients, which was published in Gynecologic Oncology (Vol. 82 pages 317-322, 2001), investigated the ability of six months of treatment with Hexalen capsules to extend survival following achievement of a complete response with front-line therapy. At two years, following completion of front-line therapy, patients in this trial demonstrated a higher survival rate compared to that seen in earlier trials.

9

Products Under Development

The following table summarizes the status of ongoing development of Aloxi products, irofulven, other acylfulvene analogs, MG98 and DNA methyltransferase inhibitors:

Post Operative Nausea and Vomiting (PONV) Overview

Post operative nausea and vomiting is a common consequence of anesthetic and surgical procedures. Patients undergoing abdominal, gynecological, ear, nose and throat, cardiovascular, and eye surgery are at the highest risk for PONV. If not prevented, PONV can cause hospital re-admissions and increase healthcare costs for patients who undergo surgery. In the United States, more than 40 million surgical procedures are performed annually, and the incidence of PONV is estimated at 25% to 30%. If approved for the prevention of PONV, Aloxi injection would compete in an approximately $400 million PONV market in the United States.

Aloxi (palonosetron hydrochloride) injection is a potent, highly selective serotonin subtype 3, or 5-HT₃, receptor antagonist differentiated by its strong receptor binding affinity and extended half-life and has already been approved for the prevention of chemotherapy-induced nausea and vomiting. We are working with Helsinn Healthcare, the licensor of palonosetron hydrochloride, to design phase 3 programs with a goal to seek FDA approval of Aloxi injection for the prevention of PONV and to develop an oral Aloxi formulation. Phase 3 trials are expected to begin during 2004 for each program.

10

Irofulven for Chemotherapy Treatment

Irofulven is our lead cancer therapy product candidate under development and is part of our family of proprietary cancer therapy compounds called acylfulvenes. Acylfulvenes, including irofulven, are semi-synthetic derivatives of the natural product illudin S obtained from the Omphalotus olearius mushroom. We licensed rights to the entire class of acylfulvene agents, including irofulven, from the Regents of the University of California in 1993.

Irofulven:

Mechanism of action studies have indicated that irofulven is rapidly taken up by sensitive tumor cell types where it reacts with tumor cell DNA and protein targets in a novel manner, producing rapid inhibition of DNA synthesis and DNA lesions that are difficult for the tumor cell to repair. The initiation of DNA damage by irofulven begins tumor selective apoptosis that ultimately causes cell death. Clinical trials will determine whether the differential effect of irofulven on tumor cells compared to healthy cells translates into important clinical benefit. We further believe that irofulven could be the first of a series of acylfulvene analogs that merit development as cancer therapies.

We believe irofulven has a unique mechanism of action that makes it well-suited for study in refractory patient populations and in combination with other cancer therapies. Preclinical data and early clinical data demonstrate irofulven’s activity against tumors that are known to be resistant to other therapies. Preclinical studies have also demonstrated the additive or synergistic effect of irofulven with a number of marketed cancer therapies. To date, irofulven has demonstrated a side effect profile relative to certain marketed cancer therapies, which we believe enhances the prospect of irofulven combining well with these cancer therapies. Therefore, we believe that pursuing multiple development paths in different refractory cancers and in combination with other cancer therapies is warranted. Preclinical toxicology studies in rats and dogs and initial clinical data from the dosing of over 1,000 cancer patients with irofulven have demonstrated that irofulven is adequately tolerated as a chemotherapeutic compound and its side effects are reversible. The primary dose-limiting side effect has been bone marrow suppression, usually observed as decreased blood platelet or white cell counts. Other drug-related side effects have been nausea, vomiting, visual disturbances and fatigue. Bone marrow suppression has been controlled through dose adjustments or treatment delays to allow recovery of platelet or white cell counts. Nausea and vomiting are prophylactically controlled with standard, currently available treatments. Visual disturbances and fatigue are managed by dose adjustments and are reversible after discontinuation of treatment.

One method of identifying cancer therapy targets for further human trials is to conduct preclinical tests on mice that have been implanted with human, solid tumors. Testing of irofulven in these disease models has demonstrated dose-related anti-cancer activity, including an increase in survival or tumor regressions in the following types of cancers:

11

Irofulven Single Agent Trials

An investigational new drug application for irofulven was submitted to the FDA in September 1995 and phase 1 human safety testing was initiated in December 1995. Phase 1 clinical trials are conducted in small patient populations and are generally not designed to measure efficacy. In October 1997, we initiated a second phase 1 trial to evaluate the effect of longer infusion times. Both initial phase 1 trials completed enrollment in 1998 after establishing a maximum tolerated dose level and a recommended dosing regimen for the initial phase 2 trials.

The investigators participating in our initial phase 1 trials observed anti-cancer activity. Seven out of a total of 32 patients who received daily treatments at a dose between 8 and 18 mg/m² in these trials demonstrated stabilization of disease or tumor shrinkage as a response to irofulven.

In late 1999, we initiated a phase 1 dose optimization trial using intermittent or weekly dosing in patients with malignant solid tumors that were refractory to anti-cancer treatment or for which no standard treatment exists. In this challenging patient population, greatly improved tolerance to acute effects of treatment were observed compared to the five consecutive day dosing schedule used to initiate our phase 2 trials program. In addition, comparable dose intensity was achieved and evidence of anti-cancer activity was observed in this trial. Based on these results and consultation with our panel of outside oncology experts, weekly or bi-weekly dosing schedules will generally be used in ongoing and future trials.

We have tested irofulven in a variety of phase 2 solid tumor trials. We chose to investigate pancreas, ovary, prostate and liver because:

Anti-cancer activity of irofulven has been observed in each of these trials, including objective tumor shrinkage.

To obtain marketing approval for irofulven in the United States, pivotal registration trials will need to be successfully completed and submitted to the FDA. Phase 3 trials typically use survival as the primary endpoint, compared to a randomly determined control group and have a higher number of enrolled patients than in earlier phases.

Irofulven for Pancreatic Cancer

Our initial pivotal registration trial of irofulven was for the treatment of refractory pancreatic cancer. Pancreatic cancer is the fifth most common cause of cancer-related death in the United States, and is an aggressive disease that has few effective treatment options. In February 2001, we began a pivotal phase 3 trial of irofulven for treating refractory pancreatic cancer patients. Initiation of this trial was based on phase 2 results in pancreatic cancer patients, improved acute tolerance with every-other-week dosing of irofulven seen in the dose optimization trial, and discussions with the FDA and our panel of outside oncology experts.

The phase 3 trial was a randomized, multi-center, international trial in advanced-stage pancreatic cancer patients whose disease progressed after treatment with gemcitabine, the current standard-of-care treatment. The primary endpoint was overall survival times following treatment with irofulven compared to continuous infusion 5-fluorouracil (5-FU). Two patients were randomized into the irofulven treatment arm for every patient enrolled in the 5-FU control arm. In April 2002, we stopped enrollment in this phase 3 clinical trial of irofulven for gemcitabine-refractory pancreatic cancer patients. Despite evidence of irofulven activity, preliminary analysis of

12

the phase 3 data by an independent Data and Safety Monitoring Board, or DSMB, indicated that the comparator agent 5-FU demonstrated a greater than expected survival benefit, making it statistically improbable that the final study results could achieve our planned objectives for the trial. For this reason, we will no longer pursue this specific indication for irofulven monotherapy in gemcitabine-refractory pancreatic cancer. Following closure of the trial to enrollment, we continued to make irofulven available to the remaining enrolled pancreatic cancer patients in this trial who had experienced clinical benefit.

Irofulven Combination Trials

We believe that an advantage of combination therapy is the potential to achieve enhanced anti-cancer benefit with an acceptable side effect profile compared to either agent used alone. Because irofulven has a unique mechanism of action, retains activity against tumors that are known to be resistant to other cancer therapies and in preclinical trials demonstrated additive or synergistic effects in combination with a number of marketed chemotherapies, we are conducting drug combination trials with irofulven. The first clinical step in exploring combination therapy is to conduct phase 1, dose-ranging trials to determine the maximum tolerated dose of both drugs together. The anti-tumor activity in a completed phase 1 trial of irofulven in combination with irinotecan led to the initiation of a phase 2 trial of irofulven with irinotecan in patients with advanced gastrointestinal tumors. We are currently conducting four dose-ranging trials of irofulven in combination with gemcitabine, capecitabine, docetaxel, and oxaliplatin. Data from these trials and possible follow-up phase 2 trials in specific tumor types will be evaluated to determine whether further drug combination development with these marketed therapies is warranted.

Irofulven for Hormone Refractory Prostate Cancer

Throughout phase 1 and 2 monotherapy and combination therapy trials, consistent evidence of clinical activity in hormone refractory prostate cancer, or HRPC, has been reported, including activity in patients previously exposed to chemotherapy. Phase 2 data presented at the 39^th Annual Meeting of the American Society of Clinical Oncology (ASCO) in 2003 indicated that irofulven in combination with prednisone, a steroid, induced decreases of more than 50 percent in prostate specific antigen (PSA) levels in 25 percent of evaluable patients and disease stabilization in 86 percent of patients evaluable for objective tumor response. PSA is a blood marker used by clinicians for treatment decisions. Activity has also been shown in prostate cancer patients treated with irofulven plus the marketed chemotherapies, capecitabine or cisplatin, in phase 1 trials. Following comprehensive analysis of our available data, we are preparing to initiate a phase 2 combination program for irofulven in HRPC patients previously treated with docetaxel. Most HRPC patients in the United States who are receiving chemotherapy are treated with docetaxel, and we believe significant unmet need exists among prostate cancer patients who have failed hormone therapy and who have progressed on docetaxel.

NCI Clinical Trials

Under a 1996 Clinical Trials Agreement, the National Cancer Institute, or NCI, is sponsoring and overseeing, at its own expense, a clinical trials program using irofulven in its network of designated cancer centers and other institutions. We have agreed to provide drug product for these trials and will have access to any resulting data. We intend to utilize the results from NCI-sponsored trials to further characterize the safety of irofulven as well as evaluate the potential of irofulven in additional refractory solid tumor types. The NCI has sponsored 14 trials of irofulven in a range of solid tumor cancers and leukemias and two trials are ongoing.

Other Acylfulvene Analogs

In addition to irofulven, we have obtained license rights to acylfulvene analogs. The synthesis and the initial biological testing of over 100 of these analogs has been performed at the University of California, San Diego, or UCSD. At UCSD, both in vitro and in vivo anti-tumor activity has been demonstrated for a significant number of the acylfulvene analogs. Further testing by the NCI of some of these analogs has confirmed broad spectrum anti-

13

tumor activity. The broad-spectrum anti-tumor activity of the acylfulvene analogs suggests that this class of compounds has the potential to produce additional clinical development candidates. We are currently evaluating these analogs for further preclinical development.

MG98 and Small Molecule DNA Methyltransferase Inhibitor Programs

In August 2000, as part of our strategy to expand our portfolio of marketed and development stage anti-cancer products, we entered into an exclusive license, research and development agreement with MethylGene Inc. for North America rights to its proprietary anti-cancer product candidate MG98 and its DNA methyltransferase small molecule inhibitor program. Included within our license rights is a United States patent on a method for reversing the tumor-causing state of a cell by administering an agent that corrects an aberrant methylation pattern in the DNA of the cell. MethylGene is a chemistry-driven, rational drug design and development company focused on the inhibition of enzyme targets that are associated with disease. It pursues two approaches to enzyme inhibition: rationally designed mRNA inhibitors that block the production of enzymes and rationally designed small molecule inhibitors that block the activity of enzymes.

In a December 2003 amendment to the License Agreement, MethylGene acknowledged full satisfaction of our payment obligations and suspended further payment obligations by us pending a planned June 2004 data review of the small molecule inhibitor program and following completion by MethylGene of a planned clinical trial with MG98. If we resume development, our financial responsibilities under the License Agreement would also resume.

Sales and Marketing

We currently promote our products in the United States directly to medical oncologists, hematologic oncologists, radiation oncologists, rheumatologists, internal medicine physicians, and other physician specialists using our 104-person oncology-focused sales organization. In 2003, we expanded the sales organization in preparation for the commercial launch of Aloxi injection. Our sales organization will continue to be used to promote our current and future oncology products in the United States.

In addition, we have an in-house marketing, manufacturing and commercial management staff of 38 persons who support our commercial activities. We use a variety of marketing programs to reach our targeted audiences, including distribution of product-specific brochures in face-to-face meetings and direct mailings, exhibits at select medical meetings and journal advertising. Our sales and marketing organizations are highly trained and experienced.

We use international collaborations to commercialize our products outside the United States. We currently have agreements with Novartis Pharma AG (formerly CIBA Vision AG), Kissei Pharmaceutical Co., Ltd. and Pfizer Inc. to develop and commercialize Salagen Tablets for the European, Japanese and Canadian markets, respectively. Under these agreements, we receive payments based on development milestones and/or product sales in the pertinent territory. We also have distribution agreements for Salagen Tablets in Israel, Korea, Singapore, Taiwan, Hong Kong, Colombia, Thailand, and Malaysia. We receive payments based upon product sales to these distributors.

Research and Development

We maintain active drug development programs for our new drug candidates and commercialized products. Current drug development efforts are primarily focused on Aloxi products and irofulven. Licensing development milestone payments for Aloxi products are recognized as research and development expense. We also participate in post-marketing studies to support the continued commercialization of Aloxi injection, Salagen Tablets and Hexalen capsules. We seek product candidate acquisitions in order to expand our development pipeline.

14

We have incurred significant research and development costs in the past and believe that substantial capital resources will be required to support current and future development programs. We spent approximately $36.1 million in 2001, $32.2 million in 2002, and $50.1 million in 2003 on research and development. In recent years, 74 to 80 percent of our research and development expense was attributable to contracts with third parties. Approximately 80 percent of our research and development expense in 2003 was attributable to third party services, license fees or FDA user fees. Funding for research and development is expected to come from internally generated funds, joint ventures, strategic alliances or other sources of capital, including equity or debt offerings.

Successful drug development requires a broad spectrum of scientific, clinical and product development expertise. As part of our strategy, we do not directly conduct basic research or traditional drug discovery activities because we primarily intend to acquire rights to product candidates that are in the human clinical stage of development. This approach substantially reduces our research risk due to product failure and eliminates the need for direct investment in discovery research laboratories and personnel.

We manage our human clinical development of product candidates by selectively outsourcing certain activities. We have in-house medical communications capabilities and regulatory affairs expertise which allow us to maintain support systems, monitor adverse drug experience reporting, and file new drug applications with the FDA. We outsource other development activities, such as the conduct of preclinical studies and clinical trials, product formulation and production of clinical supplies.

We expect to continue to contract with third parties until it is necessary and economical to add these capabilities internally. As of December 31, 2003, we employed 55 persons in research and development, regulatory affairs and product formulation.

International Alliances

Dainippon Pharmaceutical Co., Ltd.

During 1995, we entered into a cooperative development and commercialization agreement with Dainippon, whereby we granted Dainippon an exclusive license to develop and commercialize acylfulvenes, including irofulven, in Japan. Dainippon granted us an irrevocable, exclusive, royalty-free license allowing us to use any technology or data developed by Dainippon relating to the acylfulvenes. Dainippon and MGI agreed in the first quarter of 2003 to terminate this agreement effective August 2003. Under the agreement, Dainippon paid initial and continuing quarterly milestone payments totaling $11.1 million through April 2000. From April 2000 through January 2002, $4.3 million in deposit payments were received from Dainippon, which we repaid to Dainippon in August 2003.

Kissei Pharmaceutical Co., Ltd.

In December 1994, we entered into a license agreement with Kissei under which we granted to Kissei an exclusive, royalty-bearing license to develop and commercialize Salagen Tablets in Japan. Kissei granted back to us an irrevocable, non-exclusive, royalty-free license allowing us to use any technology or data developed by Kissei relating to Salagen Tablets. Kissei paid us an initial license fee upon execution of the agreement and agreed to pay us additional milestone payments and we have received all $2.5 million of the required license fees and milestone payments. In addition, Kissei agreed to pay us royalties equal to a certain percentage of net sales revenues, subject to annual minimum requirements. In May 2003 Kissei filed a submission to the Japanese regulatory authorities for the use of Salagen Tablets in treating the symptoms of radiation-induced xerostomia in head and neck cancer patients. The product is currently undergoing regulatory review for marketing approval in Japan. Unless earlier terminated by the parties for cause or by mutual agreement, the term of the agreement is for ten years from the date Salagen Tablets are first launched in Japan. Thereafter, the agreement automatically renews for additional one-year periods.

15

Novartis Pharma AG (formerly CIBA Vision AG)

In April 2000, we entered into a license agreement with Novartis under which we granted Novartis an exclusive, royalty-bearing license to develop and commercialize Salagen Tablets in Europe, Russia and certain other countries. Novartis granted to us an irrevocable, non-exclusive, royalty-free license allowing us to use any technology or data developed by Novartis relating to Salagen Tablets. Simultaneous with this agreement, the previous agreements with Chiron B.V. for Salagen Tablet rights in Europe were terminated. Novartis paid us $1.5 million of net license fees upon completion of transfer activities. Additional milestone payments may be received if specified sales targets are achieved. Further, Novartis agreed to pay us royalties equal to a percentage of net sales revenues. Unless earlier terminated by the parties, the term of the agreement is for 12 years and may be extended for additional two-year periods.

In addition, we simultaneously entered into a supply agreement with Novartis under which we agree to supply Novartis’ requirement of Salagen Tablets until the termination of the license agreement with Novartis.

Pfizer Inc. (formerly Pharmacia Corporation)

In November 1994, we entered into a license agreement with Pfizer Inc, or Pfizer, under which we granted to Pfizer an exclusive, royalty-bearing license to develop and commercialize Salagen Tablets in Canada. Pfizer granted to us an irrevocable, non-exclusive, royalty-free license allowing us to use any technology or data developed by Pfizer relating to Salagen Tablets. Pfizer paid us an initial license fee of $75,000 upon execution of the agreement and agreed to pay us royalties equal to a percentage of net sales revenues, subject to annual minimum requirements. In addition, we agreed to pay Pfizer royalties if we promote Salagen Tablets in Canada in the first or second year following termination of the agreement. Either party may terminate the agreement upon one-year prior written notice. Thereafter, for an additional two-year period we would pay a portion of Salagen Tablet net sales in Canada to Pfizer. In addition, we simultaneously entered into a supply agreement with