Back to GetFilings.com

Use these links to rapidly review the document
TABLE OF CONTENTS

UNITED STATES
SECURITIES AND EXCHANGE COMMISSION
Washington, D.C. 20549

FORM 10-K

For the fiscal year ended: December 31, 2003

Commission File Number: 000-32179

EXACT SCIENCES CORPORATION

(Exact Name of registrant as specified in its charter)

Registrant's telephone number, including area code: (508) 683-1200

Securities registered pursuant to Section 12(b) of the Act:

None

Securities registered pursuant to Section 12(g) of the Act:

Common Stock, $.01 Par Value

        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 report(s), and (2) has been subject to such filing requirements for the past 90 days. Yes ý    No o

        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.    o

        Indicate by checkmark whether or the registrant is an accelerated filer (as defined in the Exchange Act Rule 12B-2). Yes ý    No o

        The aggregate market value of the voting and non-voting common equity held by non-affiliates of the Registrant, as of the last business day of the Registrant's most recently completed second fiscal quarter was approximately $166,862,000 (based on the closing price of the Registrant's Common Stock on June 30, 2003 of $10.95 per share).

        The number of shares outstanding of the Registrant's $.01 par value Common Stock as of January 28, 2003 was 19,249,252.

DOCUMENT INCORPORATED BY REFERENCE

        The registrant intends to file a definitive proxy statement pursuant to Regulation 14A within 120 days of the end of the fiscal year ended December 31, 2003. Portions of such proxy statement are incorporated by reference into Part III of this Form 10-K.

EXACT SCIENCES CORPORATION

ANNUAL REPORT ON FORM 10-K

YEAR ENDED DECEMBER 31, 2003

TABLE OF CONTENTS

i

PART I

Item 1. Business

        This Business section and other parts of this Form 10-K contain forward-looking statements that involve risk and uncertainties. Our actual results may differ significantly from the results discussed in the forward-looking statements. Factors that might cause such a difference include, but are not limited to, those set forth in "Management's Discussion and Analysis of Financial Condition and Results of Operations—Factors That May Affect Future Results" and elsewhere in this Form 10-K.

Overview

        We are an applied genomics company that develops and commercializes proprietary DNA-based tests for the early detection of cancer. Our first commercial test, PreGen-Plus™, is used for screening colorectal cancer, the second leading cause of cancer death in the U.S. and the leading cause of cancer death among non-smokers. Each year, more than 50 million Americans over the age of 50 who should be screened annually for colorectal cancer fail to follow the American Cancer Society's screening guidelines. Of those people for whom screening is recommended, many reject the option of colonoscopy, which, while accurate as a means of detecting colorectal cancer, is invasive, requires unpleasant bowel preparation and involves risks of damaging the colon. Until the commercial launch of PreGen-Plus, the only non-invasive option for colorectal cancer detection had been fecal occult blood testing, or FOBT. FOBT, however, suffers from relatively low sensitivity, particularly in detecting the early stage, most curable cancers, and requires dietary modifications, unpleasant stool sampling and stool manipulation by the patient. With the U.S. launch in August 2003 of PreGen-Plus, our first commercially-available DNA-based cancer screening test for the average risk population, these patients now have a more accurate, non-invasive screening option for colorectal cancer. PreGen-Plus has been clinically shown to be four times more sensitive in detecting colorectal cancer than the most commonly used FOBT screening test on the market today.

Colorectal Cancer

Background

        Colorectal cancer is the most deadly cancer in the U.S. among non-smokers and the second most deadly cancer overall. The only cancer that kills more people each year is lung cancer. According to the American Cancer Society, each year, nearly 150,000 people are diagnosed with the illness and almost 60,000 people die from it.

        Medical practitioners commonly classify colorectal cancer into four stages at the time of diagnosis as shown in the following table:

1

        Detection of pre-cancerous adenomas and cancer in its earliest stages increases the likelihood of survival and reduces the cost of treatment and care. As a result, the American Cancer Society recommends that the 80 million Americans age 50 and above undergo regular colorectal cancer screening.

Colorectal Cancer and Genomics

        Genomics, broadly defined, is the study of the genome and, we believe, serves as the scientific discipline best suited for the early detection of colorectal cancer. Initial efforts in human genomics centered on identifying, mapping, sequencing and analyzing the definitive sequence of every gene in the human genome. Scientists are now focusing on applying that knowledge to the development of novel technologies used for the detection and management of disease, as well as the development of improved therapeutics.

        Cancer begins to develop when the DNA in a single normal cell mutates or changes in such a way that ultimately results in unregulated cell growth. In a ground-breaking paper published in the New England Journal of Medicine in 1988, Dr. Bert Vogelstein, one of our scientific collaborators, and his colleagues described a multi-step model of colorectal cancer development. In 1990, Dr. Eric Fearon, a former member of our scientific advisory board, and Dr. Vogelstein published a diagram depicting the development of colorectal cancer. An updated version of this diagram showing many of the genomic events involved in the development of colorectal cancer is shown below:

        The diagram illustrates that cancer develops in steps that results from alterations in multiple genes in an individual cell, and occurs frequently with chromosome loss. The diagram shows that these alterations lead to pathologic changes in the colon from normal epithelium—the tissue that lines the surface of the colon—through early and late adenomas, which are a form of pre-cancerous growth, to early cancer and late cancer. These alterations, shown in the above diagram, usually accumulate over many years, and are typically due to:

•

mutations in individual genes, such as the Apc, K-ras and p53 genes;



•

larger scale effects in which large parts of a chromosome or even entire chromosome and chromosome arms, such as 5q, 18q, 17p and 8p, are deleted; or



•

inactivation of the mismatch repair genes (the genes responsible for correcting misincorporated bases of DNA after DNA replication) that manifest themselves as deletions in polynucleotide DNA regions such as BAT-26.

        The multi-step process provides an abundance of genomic targets that may be used for the early detection of cancer when the disease can be treated at its most curable stage.

Our Solution

        It is widely accepted in the medical community that colorectal cancer screening is strongly recommended and that colorectal cancer is highly curable if detected early. However, according to the American Cancer Society, each year, nearly 150,000 people are diagnosed with the illness and almost 60,000 people die from it. Many of these people die because they are not screened for colorectal cancer or they use ineffective screening methods that either fail to detect the cancer or detect it at a later stage, when the five-year survival rate falls below 50%. Moreover, the number of people who die

2

annually from the disease has remained relatively unchanged over the last 20 years, despite the availability of multiple colorectal cancer screening options, all of which we believe fail to meet the collective needs of patients, doctors and payors.

        Since our founding in 1995 we have worked to apply the scientific discoveries about the human genome to address the significant unmet clinical need for an accurate, non-invasive colorectal cancer screening test that could reduce mortality through early detection and increased patient compliance. With the knowledge that survival rates approach 90% for patients whose colorectal cancers are detected in their earliest stages, but with too few patients getting adequately screened, we targeted the development of a safe, simple, non-invasive test that could save more lives. Our goal was to design a test that would be both easy to use and demonstrably more effective than other options such as FOBT in detecting early-stage cancers in an average-risk, asymptomatic population.

        These development efforts led to the creation of PreGen-Plus. Our test includes proprietary and patented technologies that isolate and analyze the trace amounts of human DNA that are shed into stool every day from the exfoliation of cells that line the colon. When colorectal cancer is present, a minute portion of the total isolated human DNA will represent DNA shed from cancerous or pre-cancerous lesions. Once the human DNA in the sample is isolated, PreGen-Plus identifies specific mutations and other abnormalities in that DNA associated with colorectal cancer and pre-cancerous lesions.

        We believe that no traditional screening method on the market today (or non-traditional screening methods, such as virtual colonoscopy and immunochemical FOBT) allows for the early and accurate detection of colorectal cancer in a manner that is acceptable to patients, medical practitioners and payors. We believe that because PreGen-Plus looks for the threshold indications of colorectal cancer at the molecular level (e.g., genetic changes in DNA) rather than the more traditional clinical manifestations of colorectal cancer (e.g., blood in stool, viewable polyps or identifiable lesions), it is a powerful screening tool for the detection of colorectal cancer at its earliest stages. We have conducted several clinical studies supporting the performance of PreGen-Plus, including a recent 5,500 patient multi-center study that showed the ability of the test to detect colorectal cancer in 57% of the cases that were in the earliest stages, more than four times the detection rate of the leading FOBT on the market today, to which it was compared. Given its ease of use when compared to more traditional colorectal cancer screening methods, we believe that, based on data collected from a subset of over 3,500 patients from our multi-center study, more people will use PreGen-Plus as their screening option and, as a result, patient compliance with screening will improve. Further, we believe that PreGen-Plus, over time, can help to substantially reduce colorectal cancer mortality, just as cervical cancer deaths have been substantially reduced through regular Pap smear testing.

        In August 2003, we commercialized PreGen-Plus in the United States through our exclusive licensee and strategic partner, Laboratory Corporation of America® Holdings (LabCorp®). We chose LabCorp as our strategic commercial partner for two important reasons. The first was our shared strategic vision about the influence the molecular diagnostics industry is expected to have on the healthcare system, as well as LabCorp's stated strategic focus on novel genomics-based products that could drive critical organic growth for its business. The second reason was a function of LabCorp's national breadth and distribution capability. LabCorp is the second largest commercial laboratory in the country and processes over 300,000 patient specimens daily through its system of 36 primary laboratories and over 1,000 patient service centers across the U.S. Additionally, LabCorp employs an 800-person primary care-focused sales force that has been trained extensively to sell PreGen-Plus. We expect that this will allow us to broaden our distribution reach in North America and maximize our commercial opportunity. In an effort to increase physician orders and third-party payor reimbursement, we and LabCorp are working with the physician, payor and patient communities to demonstrate the practical advantages of PreGen-Plus, including its cost-effectiveness. Given that PreGen-Plus is a safe, simple, and non-invasive test that has demonstrated a superior ability to detect early stage colorectal cancer when compared with the current non-invasive standard, we believe that physicians will increasingly order the test for their average-risk patients over the age of 50.

3

        PreGen-Plus is DNA-based and, therefore, its performance is not limited by the biology of cancer. Accordingly, we are able to continuously improve the performance characteristics of the test. Our applied research and development efforts are currently focused on increasing the ability of the PreGen-Plus test to detect colorectal cancers as well as pre-cancerous lesions. For example, in the recently launched commercial version of the test, we incorporated a new sample preparation technology called Effipure™. Effipure increases the yield of DNA that can be isolated from a stool sample, resulting in improved sensitivity compared to the performance of earlier versions of the test. As PreGen-Plus is not currently subject to the lengthy approval process of the U.S. Food and Drug Administration, or FDA, improvements to the test generally can be introduced to the market through LabCorp as they are developed. This advantage provides us the flexibility to commercialize the most advanced version of the test nearly as quickly as we can develop and validate it.

        PreGen-Plus has several advantages that we believe will lead to increased patient compliance and decreased mortality. These advantages include:

High sensitivity.    We believe that PreGen-Plus can lead to increased detection of colorectal cancer, including early stage cancers and pre-cancerous lesions. Based on our multi- center study data, PreGen-Plus demonstrated a sensitivity four times greater than the leading FOBT, currently the most common non-invasive screening method for colorectal cancer, and more than four times as effective as the leading FOBT in detecting cancer at its early stages, when survival rates approach 90%.

Non-invasive, painless and convenient testing.    Unlike current invasive screening and diagnostic methods, PreGen-Plus requires no pre-examination preparation, invasive procedures or anesthesia, and a sample can be collected in the privacy of one's home.

Simplicity.    PreGen-Plus requires no special bowel preparation, dietary restrictions, changes in medications, or manipulation of stool by the patient.

DNA-based test allows for continual and efficient improvements.    Unlike FOBT, PreGen-Plus is a DNA-based test and therefore its performance is not limited by the biology of cancer. Accordingly, this allows our scientific team to continue to improve the performance characteristics of PreGen-Plus through future technical innovations.

Our Testing Process

        Diagnostic tests typically require sample collection and preparation procedures as well as detection methods. We have developed a three-step sample collection and preparation process and five detection methods that apply genomics discoveries to the early detection of colorectal cancer.

        Specimen Collection and Transportation.    Our technologies for colorectal cancer are based on collecting a single whole stool sample in an easy, non-invasive manner. Utilizing our specially designed sample container, samples can be either brought by the patient to the laboratory or patient service center performing the colorectal cancer-screening test or sent directly from the patient's home using one of the many national couriers.

        Representative Sampling.    Before we developed our technologies, no one had been able to reproducibly extract human DNA and consistently find mutations in DNA in stool. We believe that this was due, in part, to the non-uniform distribution of abnormal DNA in stool. We have invented proprietary homogenization methods designed to ensure that the portion of stool sample that is processed at the laboratory will contain uniformly distributed DNA throughout the portion of the

4

sample being tested, and that the stool sample is, therefore, representative of the entire stool and colon. Based upon our data to date, we believe these methods lead to increased sensitivity and reproducible results.

        DNA Extraction, Purification and Amplification.    The isolation and amplification of human DNA found in stool is technically challenging because over 99% of DNA in stool is not human DNA, but is actually DNA from bacteria normally found in the colon. In addition, there are substances in stool that make the isolation and amplification of human DNA a difficult task. Our proprietary technologies allow for the reproducible isolation and amplification of human DNA found in stool.

        Cancer Detection Methods.    We have designed proprietary methods for detecting and identifying genomic markers associated with colorectal cancer that can be performed on existing instruments commonly available in clinical laboratories conducting molecular testing.

Commercial Strategy

        On June 26, 2002, we entered into a license agreement and long-term strategic alliance with LabCorp to commercialize PreGen-Plus. Since then we have been actively working with LabCorp to improve the performance characteristics of PreGen-Plus and its market acceptance through the incorporation of technical changes such as Effipure. In addition, the August 2003 commercial launch of PreGen-Plus enabled us and LabCorp to refocus our efforts on new sales and marketing initiatives to help stimulate demand for the test. We and LabCorp amended this license agreement on January 19, 2004 to, among other things, restructure certain product development milestones and increase the level of our collaboration on sales initiatives and test enhancements.

        Pursuant to the license agreement, as amended, we agreed to license to LabCorp all U.S. and Canadian patents and patent applications owned or exclusively licensed by us relating to PreGen-Plus. The license with LabCorp is exclusive in the U.S. and Canada for a five-year period after the commercial launch of PreGen-Plus followed by a non-exclusive license for the life of the patents. In return for the license, LabCorp has agreed to pay us certain up-front and milestone payments, and a per-test royalty fee based on the reimbursed amount of each test ordered by a physician and processed in LabCorp's facilities. These per-test royalty fees are subject to a minimum dollar amount per test. LabCorp made an initial payment of $15 million to us upon the signing of the agreement in June 2002, and a second payment of $15 million in August 2003 upon the commercial launch of PreGen-Plus. In addition, pursuant to the amended agreement, we may be eligible for additional milestone payments from LabCorp totaling up to $45 million, of which a total of up to $15 million relates to certain clinical guideline acceptance and policy-level reimbursement approvals and a total of up to $30 million relates to the achievement of significant LabCorp revenue thresholds. As part of the agreement, in June 2002 we issued to LabCorp a warrant to purchase 1,000,000 shares of our common stock, exercisable for cash over a three-year period, at an exercise price of $16.09 per share.

        In connection with the commercialization of PreGen-Plus, we have been developing and implementing a marketing and reimbursement strategy. We have built a strategic sales team of 11 highly skilled and experienced individuals to help strategically guide and support the 800-person LabCorp sales force on PreGen-Plus initiatives. Our reimbursement strategy consists primarily of educating large managed care organizations, large self-insured employers and large physician groups about the clinical benefits and cost-effectiveness of using PreGen-Plus. We believe that both the anticipated publication of our multi-center study results in a peer-reviewed journal and our cost-effectiveness study results that were presented at the Digestive Disease Week conference in May 2003 will aid in our efforts to gain reimbursement for the test. Between commercial launch and December 31, 2003, LabCorp received over 500 patient samples for testing from physicians across the country, billed insurers and received payment from numerous third-party payors. Furthermore, payors representing approximately 10 million covered lives have approved reimbursement of PreGen-Plus for their appropriate patients, including two large employer groups who have agreed to pay for PreGen-Plus for their employees.

5

Clinical Studies

        Our DNA-based technologies, including PreGen-Plus, have been the subject of extensive research and clinical studies. In numerous studies to date, the performance of PreGen-Plus has been examined in thousands of tissue and stool samples. In addition to several smaller clinical studies designed to measure the sensitivity and specificity of PreGen-Plus in detecting colorectal cancer, the performance of PreGen-Plus was compared to FOBT in a multi-center clinical study that enrolled approximately 5,500 average-risk, asymptomatic patients from more than 80 sites across the United States. The study was designed to determine whether PreGen-Plus was clinically superior to Hemoccult II®, an FOBT that is currently the most widely used non-invasive colorectal cancer screening option. The primary endpoint of the clinical study demonstrated strong statistical significance, with a p-value of less than 0.001. Results from the study, which were presented in October 2003 at the American College of Gastroenterology's Annual Conference, indicated that PreGen-Plus was four times more sensitive than this FOBT in detecting colorectal cancer (52% for PreGen-Plus versus 13% for FOBT), and more than four times more sensitive in detecting colorectal cancer in its earliest, most curable stages (57% for PreGen-Plus versus 13% for FOBT). There was no difference in specificity between PreGen-Plus and this FOBT, with both tests demonstrating a specificity of approximately 95%.

        Scientifically and clinically, study results provide validation for the technology and its use in clinical practice. Commercially, published clinical study results provide the information necessary for thought leaders to evaluate PreGen-Plus for inclusion into colorectal cancer screening guidelines. Guideline inclusion is important both to physicians and to payors, who frequently follow such guidelines in evaluating new technologies.

        The first colorectal cancer screening guidelines promulgated in 1997 by the GI Consortium, which includes physicians from the American College of Gastroenterology and the American Gastroenterological Association, stated that future studies of new technologies did not themselves have to encompass a mortality endpoint, but instead should be compared to currently available technologies that had already proven such a benefit. We therefore designed our multi-center study with this in mind, believing that demonstration of superiority with statistical significance would satisfy the directive from the GI Consortium, and thus increase the likelihood that the PreGen-Plus test would be included as an option in colorectal cancer screening guidelines.

        Results from our clinical studies that have been published are summarized in the table below. The results of these studies may not be directly comparable as these studies were conducted across a variety of patient populations and clinical settings and employed varying sample collection protocols. The multi-center study referenced above, as well as all of our published clinical studies to date, reflect the performance of our original, bead-based version of PreGen-Plus. The commercial test currently available includes several technological improvements, including Effipure, which we believe enhances the overall performance of the test.

*

Specificity can only be derived in studies that include a certain number of individuals without cancer. The studies in the table without a specificity figure did not contain the requisite number of disease-free individuals.

6

        In October 2001, Mayo Clinic initiated a clinical study of PreGen-Plus, which is ultimately intended to include approximately 4,000 patients and is designed to compare the results of PreGen-Plus with those of FOBT. Our role in this study is limited to sample processing. Based on our information to date, we expect Mayo Clinic to complete enrollment in this clinical study in 2005.

        In addition, we have had numerous abstracts accepted for presentation at industry and scientific meetings and have published articles in peer-reviewed journals, including Gastroenterology, The New England Journal of Medicine and the Journal of the National Cancer Institute.

        We expect that virtually all validations of PreGen-Plus technology improvements, including sensitivity improvements, will be based on internal research studies that take advantage of past empirical data and research results. With the results of our existing body of clinical data to date, we do not believe that additional, large, multi-year clinical studies will be necessary to achieve validations of our technology improvements in the future.

Research and Development

        Our research and development efforts focus on developing multiple, DNA-based methodologies for the early detection of cancer and pre-cancerous lesions. Specifically, we are working on developing methods to automate and simplify the collection, preparation and analysis of samples to produce cost-effective commercial tests. Our research and development expense, including stock-based compensation, for fiscal 2001, 2002, and 2003 was $14.2 million, $20.5 million and $17.3 million, respectively. Our research and development efforts for the near-term will focus almost exclusively on PreGen-Plus in the following areas:

        Technical performance improvement.    We continue to focus our research and development efforts on improving the sensitivity of PreGen-Plus for both invasive cancer and pre-cancerous lesions. We have demonstrated that increasing the yield and purity of human DNA extracted from a stool sample will result in an increase in the sensitivity of the test. The commercial version of PreGen-Plus that was launched in August 2003 incorporates Effipure, our new sample preparation technology that results in a higher yield of DNA as compared to our first generation, bead-based test that was used in all of our published studies to date. We intend to continue development work to improve human DNA yield and purity from a sample, increase the sensitivity of the test using its current configuration, and develop new configurations of the test to optimize performance.

        While our research efforts to date have focused on the detection of colorectal cancer, some of the new technologies that we are investigating may enable us to better detect pre-cancerous lesions, especially those that are most likely to progress to invasive colorectal cancer. As part of this effort, we have developed and are evaluating a new method for scanning regions of DNA at sites often associated with pre-cancerous lesion development.

        Process improvement.    We are undertaking efforts to automate and reduce the cost of the PreGen-Plus testing process by seeking to eliminate certain manual steps, reduce the use of expensive reagents and increase processing throughput. These efforts are intended to enable us to continue to offer LabCorp and future strategic partners the most sensitive, robust and low-cost genomics-based tests possible.

        Extensions to other cancers.    We believe our proprietary DNA Integrity Assay, or DIA®, may potentially be applicable to the detection of other cancers in addition to colorectal cancer. DIA is a non-gene-specific marker for the presence of cancer, as indicated by longer, less degraded strands of DNA. The presence of these longer strands of DNA is believed to be associated with escape from apoptosis (natural cell death), which itself is a hallmark of cancer. We have validated the DIA theory through a collaboration with a bioinformatics company using a virtual model of cancer, and we are now working with our collaborators on a pre-clinical model. In addition, several independent papers were recently published that support our observations around DIA. We intend to investigate the potential of DIA in other applications, including:

7

•

early detection of other common cancers among average-risk individuals;



•

individual monitoring of transitions from benign proliferative disorders, such as polyps, cysts and warts, to malignant tumors;



•

intra-individual therapeutic monitoring; and



•

post-therapy screening for disease recurrence.

Sales and Marketing

        The current primary focus of our sales and marketing organization is the commercialization of PreGen-Plus for colorectal cancer. Since the August 2003 commercial launch of PreGen-Plus, we have been working with LabCorp on various sales and marketing initiatives to help stimulate demand for the test. We have built a strategic sales team of 11 highly skilled and experienced individuals to help strategically guide and support the 800-person LabCorp sales force on PreGen-Plus initiatives.

        Our PreGen-Plus commercialization strategy being executed with LabCorp is designed to address the needs of four major constituencies:

Primary Care Physicians (including family practice, generalists, internists, and obstetricians and gynecologists, together "PCPs").    PCPs are principal targets of our promotional activities as we believe that they drive most colorectal cancer screening activities.

Gastroenterology Thought Leaders.    Gastroenterologists are highly vocal in advocating colorectal cancer screening, and perform the vast majority of the reference standard diagnostic procedure, colonoscopy. Because they are key to establishing new tests as standard of care and are highly influential with local primary care physicians, we are working closely with gastroenterology thought leaders.

Consumers.    Consumers are important promotional targets as we believe they can be very influential in the screening process.

Third-Party Payors.    We believe that all promotional targets, PCPs, gastroenterologists and consumers, will bring important pressure on the fourth major constituency, third party payors, such as Medicare, major national and regional managed care organizations and insurance carriers, and self-insured employer groups with the goal being payment for PreGen-Plus and, eventually, formal inclusion in plan reimbursement policies.

        To address these four important constituencies, we have engaged in the following broad sales and marketing activities:

Direct Sales To Physicians.    Sales initiatives to date have included direct detailing of medical professionals at numerous conventions and in their individual offices, resulting in widespread awareness of the product.

Sales Force Training.    We have completed a robust, training program designed to educate of LabCorp's sales representatives on PreGen-Plus.

Medical Education Programs.    We have and will continue to execute on numerous educational initiatives directed at luminaries in the field, as well as local PCPs, to promote the potential value of PreGen-Plus in their practices. These include continuing medical education ("CME") and non-CME symposia, publications, and speaker's bureau programming. The goal of these efforts is to increase awareness of PreGen-Plus and its potential role in reducing colorectal cancer mortality as well as to increase the likelihood of PreGen-Plus being included in formal clinical practice guidelines.

Advocacy Development.    We continue to work with influential advocacy groups to promote their awareness of PreGen-Plus, its performance characteristics, and its potential value in clinical practice toward the goal of reducing mortality from colorectal cancer. We intend to continue to

8

build on growing public awareness of colorectal cancer through our activities with these advocacy groups. Our efforts to date have led to inclusion of PreGen-Plus in various well-circulated brochures, and radio and television broadcasts.

Consumer Marketing Initiatives.    Because PreGen-Plus promises to be a more consumer-friendly screening option, patients who are aware of PreGen-Plus are more likely to ask their doctor for PreGen-Plus which, in turn, should help drive test volumes.

Managed Care Activities.    We continue to educate Medicare, major national and regional managed care organizations and insurance carriers, and self-insured employer groups about the need and clinical rationale for PreGen-Plus. Along with LabCorp, we are having discussions with key decision makers at most of the major payors, with the goal of shortening the review time and gaining approval for the inclusion of PreGen-Plus in formal practice guidelines within each payor's plan. In addition, we also continue to address reimbursement for PreGen-Plus from government payors, primarily the Centers for Medicare and Medicaid Services ("CMS", formerly known as the Health Care Financing Administration) by educating their senior staff about the need and clinical rationale for PreGen-Plus (See "Reimbursement").

Reimbursement

        We are currently working to obtain national coverage and reimbursement approval for tests using our technologies from Medicare as well as major national and regional managed care organizations and insurance carriers, and self-insured employer groups. In connection with the commercialization of PreGen-Plus, we have been developing and implementing a reimbursement strategy, consisting primarily of educating large managed care organizations, large self-insured employers and large physician groups about the clinical benefits and cost-effectiveness of using PreGen-Plus. We believe that both the anticipated publication of our multi-center study results in a peer-reviewed journal and our cost-effectiveness study results that were presented at the Digestive Disease Week conference in May 2003 will aid in our efforts to gain reimbursement for the test. Between commercial launch and December 31, 2003, LabCorp received over 500 patient samples for testing from physicians across the country, billed insurers and received payment from numerous third-party payors. Payors representing approximately 10 million covered lives have approved reimbursement of PreGen-Plus for their appropriate patients, including two large employer groups who have agreed to pay for PreGen-Plus for their employees.

        Medicare and other third-party payors will independently evaluate our technologies by, among other things, reviewing the published literature with respect to the results obtained from our clinical studies. We intend to assist these organizations in evaluating our technologies by providing scientific and clinical data in support of our assertions regarding the superiority and appropriateness of our technologies. In addition, data analysis has been presented showing the benefits of early disease detection and the resulting cost-effectiveness of our technologies. Current molecular diagnostic procedural terminology ("CPT") codes are available which will allow our technologies to be billed following completion of a test prescribed (ordered) by a physician for a patient. We believe that the existence of current CPT codes with applicability to our screening test will help facilitate Medicare's reimbursement process.

        The Federal Balanced Budget Act of 1997 required Medicare to reimburse for colorectal cancer screening for average-risk patients beginning on January 1, 1998 and mandated Medicare coverage for FOBT preformed by the guaiac method and flexible sigmiodoscopy. Congress amended the Budget Act of 1997 to include coverage for double contrast barium enema, a radiographic imaging test used to detect colorectal cancer in areas beyond the reach of flexible sigmiodoscopy. This was further expanded to include a screening colonoscopy every 10 years as an available option effective July 2001. We believe these actions provide evidence of the public interest in colorectal cancer screening methods and the federal government's willingness to fund these methods.

9

        Most importantly, the Federal Balanced Budget Act of 1997 allows new technologies to be included as colorectal cancer screening tests by action of the Secretary of Health and Human Services without the need for additional Congressional action. In the spring of 1999, we met with senior staff members of CMS to apprise them of our progress and to determine the steps we would need to take prior to a reimbursement determination. Following that meeting, we successfully petitioned the CMS staff to cover all medical expenses of a patient participating in our clinical studies who tests positive for colorectal cancer, which we believe was a favorable departure from prior CMS policy of not reimbursing for these costs.

        In October of 2002, we met with CMS to discuss the reimbursement process. Subsequent to that meeting, CMS published its approach to expanding the colorectal cancer screening benefit to include new technologies by use of a national coverage decision process, thereby avoiding the time-consuming notice and comment procedures otherwise applicable.

        In addition, we continue to work on building support in Congress and have met with several members of Congressional staffs and national organizations with an interest in colorectal cancer. In October 1999, we testified before the Subcommittee on Health of the House Ways and Means Committee in support of the Eliminate Colorectal Cancer Act of 1999. The Eliminate Colorectal Cancer Act of 1999 requires private insurers to cover colorectal cancer screening tests deemed appropriate by physicians and patients to the same extent as the Federal Balanced Budget Act of 1997 covers for Medicare.

        We believe that colorectal cancer screening tests based on our technologies will add a potentially lifesaving and cost-effective alternative to currently available colorectal cancer screening methods. We believe that reimbursement for FOBT tests ranges from $5 to $30, but, as stated earlier, FOBT sensitivity is relatively low, and is most effective in detecting later stage cancers when survival rates are low and treatment costs are high. We believe that reimbursement for flexible sigmiodoscopy ranges from $80 to $500, but at best, can directly detect no more than half of all colorectal cancers and adenomas since it only reaches the first third of the colon, where approximately 50% of lesions develop. Medicare and some private insurers currently reimburse for colonoscopy for cancer screening once every 10 years in average risk individuals. We believe that the cost of this procedure ranges from $700 to $2,000, and while colonoscopy is sensitive, the use of colonoscopy as a screening test to date has been limited due to low patient compliance and capacity constraints which result in generally long scheduling lead times for the procedure.

Government Regulation

General

        Certain of our activities are, or have the potential to be, subject to regulatory oversight by the Food and Drug Administration, or FDA, under provisions of the Federal Food, Drug and Cosmetic Act and regulations thereunder, including regulations governing the development, marketing, labeling, promotion, manufacturing and export of our products. Failure to comply with applicable requirements can lead to sanctions, including withdrawal of products from the market, recalls, refusal to authorize government contracts, product seizures, civil money penalties, injunctions and criminal prosecution.

        Generally, certain categories of medical devices, a category that may be deemed to include products based upon our technologies, require FDA pre-market approval or clearance before they may be marketed and placed into commercial distribution. The FDA has not, however, actively regulated in-house laboratory tests that have been developed and validated by the laboratory providing the tests. Additionally, the FDA has demonstrated prior enforcement discretion and is currently undergoing internal review on its legal authority for regulating these products. Pre-market clearance or approval is not currently required for this category of products. The FDA does regulate the sale of certain reagents, including some of our reagents, used in laboratory tests. The FDA refers to the reagents used in these tests as analyte specific reagents. Analyte specific reagents react with a biological substance including those intended to identify a specific DNA sequence or protein. These reagents generally do

10

not require FDA pre-market approval or clearance if they are (i) sold to clinical laboratories certified by the government to perform high complexity testing and (ii) labeled in accordance with FDA requirements, including a statement that their analytical and performance characteristics have not been established. A similar statement would also be required on all advertising and promotional materials relating to analyte specific reagents such as those used in our test. Laboratories also are subject to restrictions on the labeling and marketing of tests that have been developed using analyte specific reagents. We believe that in-house testing based upon our technologies, and any analyte specific reagents that we intend to sell to leading clinical reference laboratories currently do not require FDA approval or clearance. We cannot be sure, however, that the FDA will not change its policy in a manner that would result in tests based upon our technologies, or a combination of reagents, to require pre-market approval or clearance. In addition, we cannot be sure that the FDA will not change its position in ways that could negatively affect our operations either through regulation or new enforcement initiatives.

        Regardless of whether a medical device requires FDA approval or clearance, a number of other FDA requirements apply to its manufacturer and to those who distribute it. Device manufacturers must be registered and their products listed with the FDA, and certain adverse events, correction and removals must be reported to the FDA. The FDA also regulates the product labeling, promotion, and in some cases, advertising, of medical devices. Manufacturers must comply with the FDA's Quality System Regulation which establishes extensive requirements for design, quality control, validation and manufacturing. Thus, manufacturers and distributors must continue to spend time, money and effort to maintain compliance, and failure to comply can lead to enforcement action. The FDA periodically inspects facilities to ascertain compliance with these and other requirements.

        We and our strategic partner, LabCorp, are also subject to U.S. and state laws and regulations regarding the operation of clinical laboratories. The federal Clinical Laboratory Improvement Amendments of 1988 ("CLIA") and laws of certain other states impose certification requirements for clinical laboratories, and establish standards for quality assurance and quality control, among other things. Clinical laboratories are subject to inspection by regulators, and the possible sanctions for failing to comply with applicable requirements. Sanctions available under CLIA include prohibiting a laboratory from running tests, requiring a laboratory to implement a corrective plan, and imposing civil monetary penalties. If we or LabCorp fail to meet any applicable requirements of CLIA or state law, it could interrupt the commercial sale of PreGen-Plus and otherwise cause us to incur significant expense.

Diagnostic Kits

        Any diagnostic test kits that we, or our partners, may sell would require FDA clearance or approval before they could be placed into commercial distribution. There are two regulatory review procedures by which a product may receive such approval or clearance. Some products may qualify for clearance under a pre-market notification, or 510(k) process. Under such a process, the manufacturer provides to the FDA a pre-market notification that it intends to begin marketing the product, and demonstrates to the FDA's satisfaction, through appropriate studies, that the product is substantially equivalent to a comparative product that has been legally marketed and is currently in commercial distribution. Clearance of a 510(k) means that the product has the equivalent intended use, is as safe and effective as, and does not raise significant questions of safety and effectiveness than a legally marketed device. A 510(k) submission for an in vitro diagnostic device generally must include labeling information, performance data, and in some cases, it must include data from human clinical studies. Marketing may commence under a 510(k) submission when the FDA issues a clearance letter determining the product to be substantially equivalent to a comparative device.

        If a medical device does not qualify for the 510(k) submission process by not being substantially equivalent or raising new issues of safety and effectiveness, the FDA may require submission of a pre-market approval application, or PMA, before marketing can begin. PMA applications must demonstrate, among other matters, that the medical device is safe and effective. A PMA application is a more comprehensive submission than a 510(k) submission, resulting in longer review and approval

11

timeframes and usually includes the results of extensive pre-clinical and clinical studies and detailed information on the product, design and manufacturing system. Before the FDA will approve an original PMA, the manufacturer must undergo and pass a pre-approval inspection that assesses its compliance with the requirements of the FDA's Quality System Regulations.

        We believe that if our products are sold in FDA approved diagnostic test kit form; they would likely require PMA approval. As compared to the 510(k) process, the PMA process is traditionally more lengthy and costly, and we cannot be sure that the FDA will approve PMAs for our products in a timely fashion, or at all. Additionally, FDA requests for additional studies during the review period are not uncommon, and can significantly delay approvals. Even if we were able to gain approval of a product for one indication, significant changes to the product, its indication for use, its labeling or manufacturing and quality assurance would likely require additional approvals in the form of a PMA Supplement.

Specimen Container

        Once a physician orders a test, the patient will need to receive a specimen container to collect the patient's stool. Although specimen transport and storage containers are also medical devices regulated by the FDA, such containers generally have been exempted by regulation from the FDA's pre-market clearance or approval requirement and much of the Quality System Regulation. We believe that our specimen container falls within an applicable exemption, but we cannot be sure that the FDA will not assert that our container is not exempt and seek to impose a pre-market clearance or approval requirement.

Intellectual Property

        In order to protect our proprietary technologies, we rely on combinations of patent, trademark, and copyright protection, and other contractual restrictions to protect our proprietary technologies, as well as confidentiality agreements with employees, consultants, and third parties.

        We have pursued a patent strategy designed to maximize our patent position with respect to third parties. Generally, we have filed patents and patent applications that cover the methods we have designed to detect colorectal cancer as well as other cancers. We have also filed patent applications covering the preparation of stool samples and the extraction of DNA from heterogeneous stool samples. As part of our strategy, we seek patent coverage in the United States and in foreign countries on aspects of our technologies that we believe will be significant to our market strategy or that we believe provide barriers to entry for our competition.

        As of December 31, 2003, we had 30 patents issued and 29 pending patent applications in the United States and, in foreign jurisdictions, 33 patents issued and 98 pending applications. Our success depends to a significant degree upon our ability to develop proprietary products and technologies and to obtain patent coverage for such products and technologies. We intend to continue to file patent applications covering newly-developed products or technologies.

        Each of our patents generally has a term of 20 years from its respective priority filing dates. Consequently, our first patents are set to expire in 2018. We have filed terminal disclaimers in certain later-filed patents, which means that such later-filed patents will expire earlier than the twentieth anniversary of their priority filing dates.

        A third-party institution has co-inventorship rights with respect to one of our issued patents relating to use of our e-LOH detection method on pooled samples from groups of patients. Our current cancer screening detection methods do not include pooled samples. If any third party asserts co-inventorship rights with respect to any of our patents and is successful in challenging our inventorship determination, such patent may become unenforceable or we may be required to add that third party inventor to the applicable patent, resulting in co-ownership of such patent with the third

12

party. Co-ownership of a patent allows the co-owner to exercise all rights of ownership, including the right to use, transfer and license the rights protected by the applicable patent.

        We and a third-party institution have filed a joint patent application under the Patent Cooperation Treaty that will be co-owned by us and the third-party institution relating to the use of various DNA markers, including the DNA Integrity Assay, to detect cancers of the lung, pancreas, esophagus, stomach, small intestine, bile duct, naso-pharyngeal, liver and gall bladder in stool. This patent application does not relate to the detection of colorectal cancer and designates the United States, Japan, Europe and Canada as the territories in which rights are sought.

        We license on a non-exclusive basis certain polymerase chain reaction ("PCR") technology from Roche Molecular Systems, Inc. This license relates to a gene amplification process used in almost all genetic testing, and the patent that we utilize expires in mid-2004. Roche may terminate this license upon notice if we fail to pay royalties, fail to submit reports or breach a material term of the license agreement. In exchange for the license, we have agreed to pay Roche a royalty based on net revenues we receive from commercial tests that we perform at our facility. Our strategic relationship with LabCorp, however, contemplates commercial tests being performed in LabCorp's facilities, rather than in our facility, and therefore, LabCorp is now maintaining a license and paying royalties to Roche for the use of PCR in connection with performing the PreGen-Plus test at LabCorp.

        We license on a non-exclusive basis technology from Genzyme Corporation, a licensee of patents owned by Johns Hopkins University and of which Dr. Vogelstein is an inventor. This license relates to the use of the Apc and p53 genes (the "Genes") and methodologies related thereto in connection with our products and services and lasts through 2013, the life of the patent term of the last-licensed Genzyme patent. In exchange for the license, we have agreed to pay Genzyme a royalty based on net revenues we receive from commercial tests we perform at our facility and the sale of analyte specific diagnostic test kits, as well as certain milestone payments and maintenance fees. In addition, we must use reasonable efforts to make products and services based on these patents available to the public. Genzyme may terminate this license upon notice if we fail to pay milestone payments and royalties, achieve a stated level of sales or submit reports. In addition, if we fail to request FDA clearance for a diagnostic test as required by the agreement, Genzyme may terminate the license. As noted previously, our strategic relationship with LabCorp contemplates commercial tests being performed in LabCorp's facilities, rather than our facility, and therefore, LabCorp is now maintaining a license and paying royalties to Genzyme for the use of the Genes in connection with performing the PreGen-Plus test at LabCorp. However, pursuant to our agreement with LabCorp, we may be obligated to bear a portion of certain payments by LabCorp to Genzyme if LabCorp does not achieve certain PreGen-Plus sales thresholds.

        We license on an exclusive basis, in the field of stool-based colorectal cancer screening, from Matrix Technologies Corporation, d/b/a Apogent Discoveries, certain patents owned by Apogent relating to its Acrydite™ technologies. The license provides us and our sublicensees, with the ability to manufacture and use the Acrydite technology in the PreGen-Plus test. The Acrydite technology is useful in connection with our proprietary electrophoretic DNA gel capture technology used in the isolation of nucleic acids and the diagnosis of disease that we purchased from MT Technologies.

        We license on an exclusive basis from Johns Hopkins University certain patents owned by JHU that relate to digital amplification of DNA. We believe that this license will allow us and our partners to develop and commercialize novel detection technologies to enhance the performance of our current technologies. In exchange for the license, we have agreed to pay JHU certain royalties on revenues received by us relating to our or our sublicensees' sales of products and service.

        We license on a non-exclusive basis from Beckman Coulter certain patents owned by Beckman Coulter that relate to its Single Based Extension ("SBE") technology. The license provides us and our sublicensee, LabCorp, with the ability to use SBE in the PreGen-Plus test.

13

        We and LabCorp are currently negotiating additional third-party technology license and supply agreements that are necessary for the PreGen-Plus test.

Competition

        To our knowledge, none of the large genomics or diagnostics companies is developing tests to conduct stool-based DNA testing. However, these companies may be working on similar tests that have not yet been announced. In addition, other companies may succeed in developing novel or improving existing technologies and marketing products and services that are more effective or commercially attractive than ours. Some of these companies may be larger than we are and can commit significantly greater financial and other resources to all aspects of their business, including research and development, marketing, sales and distribution.

        The market for colorectal cancer screening is large, approximating 80 million Americans age 50 and above, of which over 50 million fail to follow the American Cancer Society's screening guidelines. As a result, the colorectal screening market has attracted competitors, some of which have significantly greater resources than we have.

        Currently, we face competition from procedure-based detection technologies such as flexible sigmoidoscopy, colonoscopy and virtual colonoscopy, a new procedure being performed in which a radiologist views the inside of the colon through a scanner, as well as existing and possibly improved traditional screening tests such as immunochemical FOBT.

        In addition, some competitors are developing serum-based tests, or screening tests based on the detection of proteins or nucleic acids produced by colon cancer in the blood. We believe serum-based testing is not able to detect disease at the earliest stages of cancer at levels of sensitivity and specificity comparable to that of stool-based testing.

        We believe the principal competitive factors in the cancer screening market include:

•

high sensitivity;



•

high specificity;



•

non-invasiveness;



•

acceptance by the medical community, especially primary care medical practitioners;



•

adequate reimbursement from Medicare and other third-party payors;



•

cost-effectiveness; and



•

patent protection.

Employees

        As of December 31, 2003, we had eighty-two employees, seven of whom have Ph.D.s and two of whom have M.D.s. Forty-nine employees are engaged in research and development, sixteen employees in sales and marketing and seventeen employees in general and administration. None of our employees is represented by a labor union. We consider our relationship with our employees to be good.

Available Information

        We were incorporated in the State of Delaware on February 10, 1995. Our executive offices are located at 100 Campus Drive, Marlborough, Massachusetts 01752. Our telephone number is 508-683-1200. Our Internet website address is http://www.exactsciences.com. Our Annual Report on Form 10-K, Quarterly Reports on Form 10-Q, Current Reports on Form 8-K and amendments to those reports filed or furnished pursuant to Section 13(a) or 15(d) of the Securities Exchange Act of 1934 are available through the investor relations page of our internet website as soon as reasonably practicable after we electronically file such material with, or furnish it to, the Securities and Exchange Commission

14

(SEC). Our Internet website and the information contained therein or connected thereto are not intended to be incorporated into this Annual Report on Form 10-K.

Item 2. Properties

        We lease approximately 56,000 square feet of space in our headquarters located in Marlborough, Massachusetts under a seven-year term. We also lease approximately 4,500 square feet in Maynard, Massachusetts that expires on August 31, 2006. We believe that these facilities will be adequate to meet our space requirements for the foreseeable future.

Item 3. Legal Proceedings

        From time to time we are a party to various legal proceedings arising in the ordinary course of our business. The outcome of litigation cannot be predicted with certainty and some lawsuits, claims or proceedings may be disposed of unfavorably to us. Intellectual property disputes often have a risk of injunctive relief which, if imposed against us, could materially and adversely affect our financial condition, or results of operations. From time to time, third parties have asserted and may in the future assert intellectual property rights to technologies that are important to our business and have demanded and may in the future demand that we license their technology. We are not currently a party to any material legal proceedings.

Item 4. Submission of Matters to a Vote of Security Holders

        No matters were submitted to a vote of security holders during the fourth quarter of fiscal 2003.

PART II

Item 5. Market for Registrant's Common Equity and Related Stockholder Matters

        Our common stock is listed on The Nasdaq National Market under the symbol "EXAS." The following table provides, for the periods indicated, the high and low sales prices per share as reported by The Nasdaq National Market.

        On January 28, 2004, the last sale price reported on The Nasdaq National Market for our common stock was $7.56 per share. As of December 31, 2003, there were approximately 19,245,977 shares of our common stock outstanding held by approximately 100 holders of record.

        We have never paid any cash dividends on our capital stock and do not plan to pay any cash dividends in the foreseeable future. Our current policy is to retain all of our earnings to finance future growth.

15

Item 6. Selected Financial Data

        The selected historical financial data set forth below as of December 31, 2002 and 2003 and for the years ended December 31, 2001, 2002 and 2003, are derived from our financial statements, which have been audited by Ernst & Young LLP, independent auditors, as of December 31, 2002 and 2003 and for the years then ended, and by Arthur Andersen LLP, our former independent public accountants, for the year ended December 31, 2001, and which are included elsewhere in this Form 10-K. The selected historical financial data as of December 31, 1999, 2000 and 2001 and for the years ended December 31, 1999 and 2000 are derived from our audited financial statements, which have been audited by Arthur Andersen LLP, our former independent public accountants and which are not included elsewhere in this Form 10-K.

        The selected historical financial data should be read in conjunction with, and are qualified by reference to "Management's Discussion and Analysis of Financial Condition and Results of Operations," our financial statements and notes thereto and the report of independent public auditors included elsewhere in this Form 10-K.