Summary of AVEO PHARMACEUTICALS INC

Item 2. Management's Discussion and Analysis of Financial Condition and Results of Operations.

Forward-Looking Information

The following discussion of our financial condition and results of operations should be read with our unaudited condensed consolidated financial statements and notes included in Part I. Item 1 of this Quarterly Report on Form 10-Q for the three and six months ended June 30, 2012, as well as the audited consolidated financial statements and notes and Management's Discussion and Analysis of Financial Condition and Results of Operations, included in our Annual Report on Form 10-K for the fiscal year ended December 31, 2011, filed with the Securities and Exchange Commission, or SEC, on March 30, 2012. This Management's Discussion and Analysis of Financial Condition and Results of Operations contains forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. These forward-looking statements are based on current expectations, estimates, forecasts, and projections and the beliefs and assumptions of our management and include, without limitation, statements with respect to our expectations regarding our research, development and commercialization plans and prospects, results of operations, general and administrative expenses, research and development expenses, and the sufficiency of our cash for future operations. Words such as "anticipate," "target," "project," "believe," "goals," "estimate," "potential," "predict," "may," "expect," "might," "could," "intend," variations of these terms or the negative of those terms and similar expressions are intended to identify these forward-looking statements. Readers are cautioned that these forward-looking statements are predictions and are subject to risks, uncertainties, and assumptions that are difficult to predict. Therefore, actual results may differ materially and adversely from those expressed in any forward-looking statements. Among the important factors that could cause actual results to differ materially from those indicated by our forward-looking statements are those discussed under the heading "Risk Factors" in Item 1A of Part II and elsewhere in this report.

Overview

We are a cancer therapeutics company, which does business as AVEO Oncology, committed to discovering, developing and commercializing targeted cancer therapies to impact patients' lives. Our product candidates are directed against important mechanisms, or targets, known or believed to be involved in cancer. Our proprietary Human Response Platform™, a novel method of building preclinical models of human cancer, provides us with unique insights into cancer biology and is being leveraged in the discovery and clinical development of our cancer therapeutics.

Tivozanib, our lead product candidate, the development of which is part of our 2011 partnership with Astellas Pharma Inc., or Astellas, is a potent, selective, long half-life inhibitor of all three vascular endothelial growth factor, or VEGF, receptors that is designed to optimize VEGF blockade while minimizing off-target toxicities. Our clinical trials of tivozanib to date have demonstrated a favorable safety and efficacy profile for tivozanib. In May 2012, we announced detailed data from our global, phase 3 clinical trial comparing the efficacy and safety of tivozanib with Nexavar ® (sorafenib), an approved therapy, for first-line treatment in advanced renal cell carcinoma, or RCC, which we refer to as the TIVO-1 study. The TIVO-1 study is being conducted in patients with advanced clear cell RCC who have undergone a prior nephrectomy (kidney removal) and who have not received any prior VEGF- and mTOR-targeted therapy. In this trial, we measured, among other things, each patient's progression-free survival, or PFS, which refers to the period of time that began when a patient entered the clinical trial and ended when either the patient died or the patient's cancer had grown by a specified percentage or spread to a new location in the body. PFS is the primary endpoint in the TIVO-1 study. Secondary endpoints include safety and overall survival. Key data from the TIVO-1 study include:

• Tivozanib demonstrated a statistically significant improvement in PFS over Nexavar with a median PFS of 11.9 months for tivozanib compared to a median PFS of 9.1 months for Nexavar in the overall study population.

• Tivozanib also demonstrated a statistically significant improvement in PFS with a median PFS of 12.7 months compared to a median PFS of 9.1 months for Nexavar in the pre-specified subpopulation of patients who received no prior systemic anti-cancer therapy for metastatic disease-a subpopulation that comprised approximately 70% of the total study population.

• Tivozanib demonstrated a well-tolerated safety profile as evidenced by a low rate of dose interruptions and reductions. The most commonly reported side effect for tivozanib was hypertension (44% for tivozanib vs. 34% for Nexavar), and for Nexavar was hand-foot syndrome (13% for tivozanib vs. 54% for Nexavar). Additionally, the incidence of other side effects that are commonly associated with other VEGF receptor inhibitors was relatively low, including diarrhea (22% for tivozanib vs. 32% for Nexavar), fatigue (18% for tivozanib vs. 16% for Nexavar), and neutropenia, a condition of having a lower than normal number of white blood cells (10% for tivozanib vs. 9% for Nexavar).

• Overall survival data from the TIVO-1 study are not yet mature and the median has not been reached in either the tivozanib or the Nexavar arm of the study. Based on early, interim analysis, 81% of patients receiving Nexavar achieved one year overall survival and 77% of patients receiving tivozanib achieved one year overall survival. In the TIVO-1 study, 53% of all patients randomized to the Nexavar arm of the study went on to receive a second drug after their disease progressed, nearly all of whom received tivozanib after Nexavar. In comparison, only 17% of all patients randomized to tivozanib went on to receive a subsequent therapy.

Table of Contents

In addition to the TIVO-1 study, we are evaluating tivozanib in multiple clinical trials including our BATON (Biomarker Assessment of Tivozanib in ONcology) program, a series of clinical trials assessing tivozanib biomarkers in solid tumors. The BATON trials include BATON-RCC, a phase 2 exploratory biomarker study evaluating first-line treatment with tivozanib in patients with advanced RCC, which completed enrollment in early 2012. The second BATON study underway is BATON-CRC, a phase 2 clinical trial being conducted by Astellas evaluating tivozanib in combination with modified FOLFOX6 (mFOLFOX6) compared to Avastin® (bevacizumab) in combination with mFOLFOX6 as first-line therapy in patients with advanced metastatic colorectal cancer, or CRC. We plan to initiate a new clinical study, TAURUS (TivozAnib Use veRsUs Sutent in advanced RCC:
Patient Preference), to demonstrate patient preference of tivozanib compared to sunitinib as first-line therapy in patients with advanced RCC. Additionally, we expect to initiate further clinical evaluation of tivozanib in breast cancer in the second half of 2012 as part of our BATON program.

We expect that the results of all of our phase 1 and phase 2 trials will help to inform our clinical development plans for tivozanib as a monotherapy and in combination with other anti-cancer therapies in multiple cancer indications.

We acquired exclusive rights to develop and commercialize tivozanib worldwide outside of Asia pursuant to a license agreement we entered into with Kirin Brewery Co. Ltd. (now Kyowa Hakko Kirin), or KHK, in 2006. Under the license agreement, we obtained an exclusive license to research, develop, manufacture and commercialize tivozanib, pharmaceutical compositions thereof and associated biomarkers for the diagnosis, prevention and treatment of any and all human diseases and conditions outside of Asia. KHK has retained all rights to tivozanib in Asia. We have obligations to make milestone and royalty payments to KHK. The royalty rates range from the low to mid-teens as a percentage of our net sales of tivozanib. We are also obligated to pay a specified percentage of certain amounts we receive from any third-party sublicensees, including Astellas. As discussed below under the heading "Strategic Partnerships," we entered into a strategic collaboration with Astellas in which we have agreed to share responsibility, including all profits and losses, with Astellas for continued development and commercialization of tivozanib in North America and Europe. Throughout the rest of the world, outside of North America, Europe and Asia, we granted Astellas an exclusive, royalty-bearing license to develop and commercialize tivozanib.

In addition to tivozanib, we have a pipeline of monoclonal antibodies derived from our proprietary Human Response Platform. Ficlatuzumab, our next most advanced product candidate, is an antibody which binds to hepatocyte growth factor, or HGF, thereby blocking its function. We have completed two phase 1 clinical trials of ficlatuzumab demonstrating the ability to combine ficlatuzumab with epidermal growth factor receptor, or EGFR, inhibitors Tarceva® (erlotinib) and Iressa® (gefitinib). In May 2012, we announced preliminary data from our phase 2 study comparing the combination of ficlatuzumab and gefitinib to gefitinib monotherapy in previously untreated Asian subjects with non-small cell lung cancer, a population with a high prevalence of EGFR sensitizing mutations. Preliminary results in the intent-to-treat population showed a trend favoring the ficlatuzumab/gefitinib combination. However, study results did not reach statistical significance.

Our third clinical candidate, AV-203, is a monoclonal antibody that targets the ErbB3 receptor, which we have partnered with Biogen Idec, Inc. In May 2012, we announced the initiation of a phase 1 clinical trial examining the safety, tolerability and preliminary efficacy of AV-203 along with exploratory biomarkers in patients with metastatic or advanced solid tumors. A core component of the study will be the exploration of predictive biomarkers for AV-203 response identified using our Human Response Platform.

We have also identified a number of other promising targets for the development of novel cancer therapeutics using our Human Response Platform. We have preclinical and discovery antibody programs underway focused on targets that appear to be important drivers of tumor growth, including programs directed towards the Notch receptors and Fibroblast Growth Factor receptors. We also developed a discovery antibody program targeting Recepteur d'Origine Nantais, or RON, which we licensed to Centocor Ortho Biotech Inc., or Centocor in May 2011, and have recently initiated a program focusing on cancer cachexia.

Our proprietary Human Response Platform was designed to overcome many of the limitations of traditional approaches to modeling human cancer. The traditional method of modeling human cancer uses a model referred to as a xenograft. A xenograft model is created by adapting cells from a human tumor to grow in a petri dish, and then injecting these cells in a mouse, where they grow into tumors. However, the resulting tumors differ from the original tumor in important respects, and, accordingly, xenograft models are often poor predictors of the success of cancer drugs in human clinical trials. In our Human Response Platform, we use patented genetic engineering techniques to grow populations of spontaneous tumors in animals containing human-relevant, cancer-causing mutations and tumor variation akin to what is seen in populations of human tumors. Because we believe that these populations of tumors better replicate what is seen in human cancer, we believe that our Human Response Platform provides us with unique insights into cancer biology and mechanisms of drug response and resistance, and represents a significant improvement over traditional approaches. We are utilizing this Human Response Platform alone and with our strategic partners to (i) identify and validate target genes which drive tumor growth, (ii) evaluate drugs which can block the function of these targets and (iii) identify biomarkers, which are indicators of drug response and resistance in patients, in an effort to evaluate which patients are most likely to respond favorably to

Table of Contents

treatment with such drugs. The potential identification and development of relevant biomarkers through our Human Response Platform is a core component of our oncology drug development efforts. Our goal is to utilize the biomarker data from BATON-CRC and other BATON clinical trials to further inform our clinical development strategy.

We have devoted substantially all of our resources to our drug discovery efforts comprising research and development, conducting clinical trials for our product candidates, protecting our intellectual property and supporting the general and administrative functions of these operations. We have generated no revenue from product sales through June 30, 2012, and have principally funded our operations through:

• $320.7 million of non-dilutive capital in the form of license fees, milestone payments and research and development funding received from our strategic partners;

• $15.0 million of gross proceeds in connection with the license agreement with Centocor executed in May 2011, which includes $7.5 million of gross proceeds from the sale of common stock to Johnson & Johnson Development Corporation, an affiliate of Centocor (recorded at $8.0 million, the fair market value on the date of issuance), and $7.5 million of non-dilutive capital in the form of license fees (recorded net of the discount on sale of common stock as $7.0 million of revenue);

• $169.6 million of funding from the sale of convertible preferred stock to investors prior to our initial public offering, including $77.5 million of equity sales to our strategic partners;

• $89.7 million of gross proceeds from the sale of common stock in connection with the completion of our initial public offering;

• $26.5 million of loan proceeds in connection with our loan agreement with Hercules Technology II, L.P. and Hercules Technology III, L.P.;

• $60.8 million of gross proceeds from the private placement of our unregistered common stock in November 2010; and

• $111.2 million of gross proceeds from the sale of common stock in connection with a follow-on public offering of our common stock in June 2011.

We do not have a history of being profitable and, as of June 30, 2012, we had an accumulated deficit of $268.7 million. We anticipate that we will continue to incur significant operating costs over the next several years as we advance our plan to expand our discovery, research, development and commercialization activities, including additional clinical development and planned commercialization of our lead product candidate, tivozanib, as well as the continued clinical development of our phase 2 product candidate, ficlatuzumab, and our phase 1 product candidate, AV-203. We will need additional financing to support our operating activities.

Strategic Partnerships

Kyowa Hakko Kirin

In December 2006, we entered into a license agreement with KHK under which we obtained an exclusive license, with the right to grant sublicenses subject to certain restrictions, to research, develop, manufacture and commercialize tivozanib, pharmaceutical compositions thereof and associated biomarkers. Our exclusive license covers all territories in the world, except for Asia. KHK has retained rights to tivozanib in Asia. Under the license agreement, we obtained exclusive rights in our territory under certain KHK patents, patent applications and know-how related to tivozanib, to research, develop, make, have made, use, import, offer for sale, and sell tivozanib for the diagnosis, prevention and treatment of any and all human diseases and conditions. Upon entering into the license agreement with KHK, we made a one-time cash payment in the amount of $5.0 million.

Under our license agreement with KHK, we may be required to

• make up to an aggregate of $50.0 million in future milestone payments upon the achievement of specified regulatory milestones;

• pay tiered royalty payments on net sales we make of tivozanib in our territory ranging from the low to mid-teens as a percentage of our net sales of tivozanib; and

• pay 30% of certain amounts we receive under our collaboration and license agreement with Astellas, which we describe below, in connection with Astellas' development and commercialization activities outside of North America and Asia related to tivozanib, other than amounts we receive in respect of research and development funding or equity investments, subject to certain limitations.

In March 2010, we made a $10.0 million milestone payment to KHK in connection with the dosing of the first patient in our phase 3 clinical trial of tivozanib.

Table of Contents

We also recorded $22.5 million of research and development expense during the six months ended June 30, 2011 associated with the payment made to KHK related to the up-front license payment received under the collaboration and license agreement with Astellas which we entered into in February 2011.

Astellas Pharma

In February 2011, we entered into a collaboration and license agreement with Astellas and certain of its indirect wholly-owned subsidiaries in connection with which we and Astellas will develop and commercialize tivozanib for the treatment of a broad range of cancers, including RCC, and breast and colorectal cancers. Under the terms of the collaboration agreement, we and Astellas will share responsibility for continued development and commercialization of tivozanib in the United States, Mexico and Canada, or North America, and in Europe under the joint development plan and joint commercialization plan, respectively. Throughout the rest of the world (which excludes North America, Europe and Asia), which we refer to as the royalty territory, Astellas has an exclusive, royalty-bearing license to develop and commercialize tivozanib. Our plan to commercialize tivozanib in collaboration with Astellas, as described herein, is subject to our and Astellas' receipt of necessary regulatory approvals from the U.S. Food and Drug Administration, or FDA, and foreign regulatory authorities based upon favorable results in clinical trials. There can be no assurance that such approvals will be obtained.

Assuming successful approvals of tivozanib by applicable regulatory agencies, we will hold all marketing authorizations in North America, including any new drug application in the United States, and Astellas will hold all marketing authorizations in the rest of the world, other than Asia.

Assuming successful approvals of tivozanib by applicable regulatory agencies, we, as the lead commercialization party in North America, will have lead responsibility for formulating the commercialization strategy for North America under the joint commercialization plan, with each of us and Astellas responsible for conducting fifty percent (50%) of the sales efforts and medical affairs activities in North America. Astellas will have lead responsibility for commercialization activities in Europe under the joint commercialization plan, and we will be responsible for conducting fifty percent (50%) of the medical affairs activities in the major European countries. All costs associated with each party's conduct of development and commercialization activities (including clinical manufacturing and commercial manufacturing costs, if any) in North America (including any regulatory milestones and royalties associated with tivozanib in North America which may become payable by us to KHK under our license agreement with KHK), and any resulting profits or losses, will be shared equally between the parties. All costs associated with each party's conduct of development and commercialization activities (including clinical manufacturing and commercial manufacturing costs, if any) in Europe, and any resulting profits or losses, will be shared equally between the parties. As between the parties, we will remain responsible for complying with our sublicense revenue sharing obligations, if any, to KHK under our license agreement with KHK in connection with the development and commercialization of tivozanib outside of North America.

We are responsible for manufacturing, through our third-party manufacturer, all of Astellas' requirements for tivozanib pursuant to a clinical supply agreement which we have entered into with Astellas, and a commercial supply agreement which the parties are currently negotiating.

Each party is obligated to use commercially reasonable efforts to develop and commercialize tivozanib in each of the United States, Mexico and Canada, and to develop and commercialize tivozanib in each European country specified in the agreement. Astellas is also obligated to use commercially reasonable efforts to develop and commercialize tivozanib in each country in the royalty territory.

During the term of the agreement, neither party nor its controlled affiliates may commercialize anywhere in North America, Europe or the royalty territory any product that has a specified mechanism of action (as further defined in the collaboration agreement) for any oncology indication, except that Astellas may commercialize specified compounds for hematological cancer. Astellas may also commercialize products (other than tivozanib) in the royalty territory, on a country-by-country basis, upon expiration of the applicable royalty term, and in North America and Europe upon expiration of all valid claims under the licensed patents.

In connection with the agreement, we received an initial cash payment of $125 million, comprised of a $75 million license fee and $50 million in research and development funding, both of which are non-creditable and non-refundable against any amounts due under the collaboration agreement. We retained net proceeds of approximately $97.6 million of the initial cash payment from Astellas, after payments to KHK and strategic, legal and financial advisors. We are also eligible to receive an aggregate of approximately $1.3 billion in potential milestone payments, comprised of (i) up to $85 million in substantive milestone payments upon achievement of specified clinical and development milestone events, (ii) up to $490 million in substantive milestone payments upon achievement of specified regulatory milestone events, including up to $90 million in milestone payments in connection with specified regulatory filings, and receipt of marketing approvals, for tivozanib to treat RCC in the United States and Europe, and (iii) up to approximately $780 million in milestone payments upon the achievement of specified commercial sales events. The first anticipated clinical and development milestone is due to us upon initiation of our next phase 3 clinical trial in RCC in combination with another therapeutic, or in breast cancer, colorectal cancer or another indication. The timing of this milestone is uncertain, as we have not finalized plans for our future trials. We have elected to recognize all milestone payments as revenue once the milestones have been triggered if the

Table of Contents

milestone is deemed to be substantive. Significant potential near-term regulatory milestones include acceptance by the FDA of the first filing of a New Drug Application, or NDA, for RCC (monotherapy) ($15 million) and acceptance by the European Medicines Agency, or EMA, of the first filing of a Marketing Authorization Application ($15 million). In addition, if tivozanib is successfully developed and launched in the royalty territory, Astellas will be required to pay to us tiered, double digit royalties on net sales of tivozanib in the royalty territory, if any, subject to offsets under certain circumstances. We are required to pay KHK low to mid-teen royalties on our net sales in North America, and 30% of certain amounts we may receive from Astellas in connection with Astellas' development and commercialization activities in Europe and the royalty territory, including up-front license fees, milestone payments and royalties.

Unless terminated earlier in accordance with its terms, the collaboration agreement with Astellas expires (a) with respect to the royalty territory, on a country-by-country basis, upon the latest to occur of: (i) the expiration of the last-to-expire valid claim of an AVEO patent or joint patent covering the composition of tivozanib, (ii) the expiration of the last-to-expire valid claim of an AVEO patent or joint patent covering the use of tivozanib, but only for so long as no generic competition exists in such country, and (iii) twelve years from first commercial sale of tivozanib in such country, and (b) with respect to North America and Europe as a whole, upon the expiration of all payment obligations between the parties related to development and commercialization of tivozanib in North America and Europe. After the second anniversary of the effective date of the collaboration agreement, Astellas has the right to terminate the collaboration agreement, in its entirety or solely with respect to the royalty territory, at any time upon 180 days prior written notice to us. Either party may terminate the collaboration agreement with respect to a specified territory or country as set forth in the collaboration agreement, if the other party fails to cure a material breach related to such territory or country, as applicable. We may also terminate the collaboration agreement in its entirety upon a patent-related challenge by Astellas, its affiliates or sublicensees, if such patent-related challenge is not withdrawn within 30 days following our notice to Astellas of such termination.

We are accounting for the joint development and commercialization activities in North America and Europe as a joint risk-sharing collaboration in accordance with Accounting Standards Codification, or ASC, 808 Collaborative Arrangements. In addition, these activities were not deemed to be separate deliverables under the agreement with Astellas.

Payments from Astellas with respect to Astellas' share of research and development costs incurred by us are recorded as a reduction to expense due to the joint risk-sharing provisions of the agreement in North America and Europe. As a result of the cost-sharing provisions in our agreement with Astellas, we reduced research and development expense by $7.4 million and $6.1 million during the three months ended June 30, 2012 and 2011, respectively, and by $15.3 million and $12.3 million during the six months ended June 30, 2012 and 2011, respectively. We also reduced general and administrative expense by $0.9 million and $0.2 million during the three months ended June 30, 2012 and 2011, respectively, and by $1.5 million and $0.4 million during the six months ended June 30, 2012 and 2011, respectively as a result of the cost-sharing provisions in our agreement with Astellas. The net amount due to us from Astellas pursuant to the cost-sharing provisions is $8.3 million at June 30, 2012.

Activities under the agreement with Astellas outside of the joint development and commercialization activities in North America and Europe were evaluated under ASC 605-25 Revenue Recognition - Multiple Element Arrangements, or ASC 605-25, to determine if they represented a multiple element revenue arrangement. The agreement with Astellas includes the following deliverables outside of the joint development and commercialization activities in North America and Europe:
a co-exclusive license to develop and commercialize tivozanib in North America and Europe, a royalty-bearing license to develop and commercialize tivozanib in the royalty-bearing territory, which includes our obligation to provide access to clinical and regulatory information resulting from the activities in North America and Europe to Astellas for its development and commercialization of . . .