The evolution of biosimilars in oncology, with a focus on trastuzumab

Cancer therapy has evolved significantly with increased adoption of biologic agents (“biologics”). That evolution is especially true for her2 (human epidermal growth factor receptor-2)–positive breast cancer with the introduction of trastuzumab, a monoclonal antibody against the her2 receptor, which, in combination with chemotherapy, significantly improves survival in both metastatic and early disease. Although the efficacy of biologics is undeniable, their expense is a significant contributor to the increasing cost of cancer care. Across disease sites and indications, biosimilar agents are rapidly being developed with the goal of offering cost-effective alternatives to biologics. Biosimilars are pharmaceuticals whose molecular shape, efficacy, and safety are similar, but not identical, to those of the original product. Although these agents hold the potential to improve patient access, complexities in their production, evaluation, cost, and clinical application have raised questions among experts. Here, we review the landscape of biosimilar agents in oncology, with a focus on trastuzumab biosimilars. We discuss important considerations that must be made as these agents are introduced into routine cancer care.


INTRODUCTION
In recent years, cancer treatment has been revolutionized by the introduction of biologic agents ("biologics"). That revolution is especially evident for her2 (human epidermal growth factor receptor 2)-positive breast cancer with the introduction of trastuzumab, a monoclonal antibody against her2 that significantly improves survival in both metastatic and early disease.
Although their efficacy is undeniable, expensive biologics, including trastuzumab, are a significant contributor to the increasing cost of cancer care. In developing countries, fewer than 10% of patients have access to her2targeted therapies, largely because of the cost 1 . In 2012, the World Health Organization released a proposal for the inclusion of trastuzumab in its list of essential medicines. Trastuzumab was subsequently added to the list in 2015, with the caveat that the petition for inclusion was based on the possibility of obtaining a lower-cost biosimilar product.
Biosimilars are pharmaceuticals whose molecular shape, efficacy, and safety are similar, but not identical, to the original product. Across disease sites and indications, biosimilar agents are rapidly being developed with the goal to create price competition and to provide cost-effective alternatives to biologics. Although these agents have the potential to improve patient access, complexities in their production, evaluation, and clinical application have raised questions among experts.

HISTORY OF BIOSIMILAR AGENTS
Compared with small-molecule drugs, biologics are larger and more complex-and therefore more susceptible to production differences. Unlike traditional small-molecule drugs for which generic versions can be produced, biologics cannot be identically copied. Biologic agents are manufactured using cell lines and processes exclusive to the manufacturer. They require multiple steps for cloning, selecting, and expanding the cell line, and then isolating and purifying the product 2 . At multiple points during that process, clinically significant alterations can potentially occur. A different cell line, for example, might result in a difference in post-translational protein modification that can affect immunogenicity and alter a drug's pharmacokinetics and dynamics.
The European Medicines Agency (ema) was the first regulatory body to develop, in 2003, guidelines for biosimilars. In 2010, Canada adopted the guidance document Information and Submission Requirements for Biosimilar Biologic Drugs 3 , which was updated in 2016. Also in 2010, then U.S. president Barack Obama signed into law the Affordable Care Act, which amended the Public Health Service Act to create an abbreviated pathway to licensure for biosimilar agents. For biosimilars to be approved, Health Canada, the ema, and the U.S. Food and Drug Administration (fda), require that the quality, activity, safety and efficacy of the new agent be comparable to those of the original agent (Table i).
Since 2006, 25 biosimilar agents have been approved by regulatory agencies in Europe and North America (Table ii). The first biosimilars were the somatropin analogs, introduced in Europe. Erythropoietin biosimilars followed in 2007, and agents similar to filgrastim, in 2008. It was not until more recently that agents similar to the monoclonal antibodies, which are larger and more complex biologics, were introduced. The first biosimilar was the anti-tumour necrosis factor α antibody infliximab, introduced in Europe in 2013. With the patents on several monoclonal antibodies now approaching or past expiry (including trastuzumab, bevacizumab, cetuximab, and rituximab), development programs for similar therapeutics are under way.

CURRENT LANDSCAPE AND CONSIDERATIONS FOR TRASTUZUMAB BIOSIMILARS
Trastuzumab (Herceptin) is a fully humanized monoclonal antibody to the extracellular domain iv of her2, developed by Genentech (San Francisco, CA, U.S.A.). In the landmark trial by Slamon and colleagues 4 , trastuzumab added to standard chemotherapy for metastatic disease significantly improved overall survival (os) by approximately 5 months. Trastuzumab has since been shown to confer a survival benefit in multiple settings, including adjuvant and neoadjuvant treatment of her2-positive (her2+) breast cancer, and her2+ metastatic gastric cancer [5][6][7][8] . Herceptin is now well-established in guidelines as a standard of care, but it remains costly: estimates place the cost at more than CA$4,500 per month or CA$54,000 for a full 1-year course of treatment 9,10 .
The patent on Herceptin expired in 2014 in Europe, opening the door for biosimilar agents to enter the market and to lower the price by creating competition. In July 2017, a trastuzumab biosimilar agent, MYL-1410 (Mylan, Canonsburg, PA, U.S.A.), received unanimous recommendation for approval from the fda's Oncologic Drugs Advisory Committee, and on 1 December 2017, it was approved. Despite that rapid progress, important regulatory and clinical factors have to be considered. Those factors include the sensitivity of the endpoints used to determine equivalence and safety, and extrapolation of indications, interchangeability, post-market surveillance, and naming.

TRASTUZUMAB BIOSIMILARS IN DEVELOPMENT Early Disease
At the 2017 meeting of American Society of Clinical Oncology in Chicago, two phase iii trials in early-stage breast cancer with trastuzumab biosimilars were presented (Table iii). A phase iii study involving 549 patients with stages i-iiia her2+ breast cancer evaluated CT-P6, the Celltrion Healthcare version of an anti-her2 monoclonal antibody. The CT-P6 product was found to be noninferior in the rate of pathologic complete response (pcr), defined as the absence of invasive disease in the breast, and in the absence of invasive and in situ disease in the breast and axilla (ypT0/isypN0), with a risk ratio of 0.92 12 . Those findings met the predefined margin for risk. A second randomized phase iii study with SB3, a trastuzumab biosimilar from Samsung Bioepis, demonstrated equivalence in neoadjuvant treatment, with a primary endpoint of pcr in the breast. The rate was 51.7% for SB3 compared with 42.0% for trastuzumab, for an adjusted ratio of 1.259 (95% confidence interval: 1.112 to 1.426) 15 .
At the 2017 meeting of the European Society for Medical Oncology, studies on two additional agents in the neoadjuvant setting were presented. Pfizer's PF-05280014 was examined in the neoadjuvant setting in combination with docetaxel; pharmacokinetics was the primary endpoint, and pcr was a secondary endpoint. In this case, the biosimilar agent was noninferior, with a pcr of 47% compared with 50% for trastuzumab (see NCT02187744 at http://ClinicalTrials.gov). Amgen's ABP 980 also resulted in a noninferior pcr when combined with docetaxel in the neoadjuvant setting 14 .

Metastatic Disease
An evaluation of CT-P6 was also conducted in metastatic her2+ breast cancer (Table iii). In a pooled analysis of data from phase i/iib and iii studies comparing CT-P6-paclitaxel with trastuzumab-paclitaxel for metastatic disease, the overall response rates (orrs) during the first 8 cycles of treatment were 57% and 62% respectively (5% difference; 95% confidence interval: -0.14 to 0.04). That finding met the criteria for equivalence. Median time to progression was 11.07 months with CT-P6 and 12.52 months with trastuzumab, with serious adverse event rates of 13.5% and 12.1% respectively 11 . The full trial has yet to be published; however, in January 2014, CT-P6 was granted approval in South Korea for metastatic breast cancer (mbca), early breast cancer (ebc), and gastric cancer. In the population with metastatic disease, Biocad's BCD-022 also demonstrated a noninferior orr (primary endpoint) of 53.6% compared with 53.70% in the group receiving Herceptin-paclitaxel. Complete responses, partial responses, stable disease, and progression rates were also similar ( Table iii).
The Heritage study, a phase iii study evaluating the biosimilar MYL-1401O combined with docetaxel in her2+ mbca, showed a noninferior orr of 69.6% compared with 64% with Herceptin-docetaxel, meeting the endpoint of equivalence (hazard ratio: 1.09; 95% confidence interval: 0.954 to 1.237). Progression-free survival (pfs) events were similar in the two groups, at 21.1% and 17.8% respectively. The pfs results were updated at the 2017 international congress of the European Society for Medical Oncology, showing a hazard ratio for pfs of 0.96. The os at 48 weeks was 89.1% (MYL-1401O) compared with 85.1% (Herceptin). Those results led to a recommendation from the Oncologic Drugs Advisory Committee to approve the agent in mbca and ebc in addition to advanced gastric cancer. Also at the 2017 congress, PF-05280014 in combination with paclitaxel Should take into account any safety or effectiveness concerns; should have mechanism to differentiate between events associated with the product and those with reference product (four-letter identification suffix known as "biologic modifier") n Adverse drug reaction reports and periodic safety update reports required n The authority to suspend an authorization is outlined in the Food and Drug Regulations n Products must be labelled indicating that the product is a "subsequent entry biologic" n There should be no claims that the biosimilar is better Labelling n Summary of product characteristics must be derived from those of the reference product n Labels require "biosimilarity statement" describing the biosimilar product's relationship to its reference product n Comparative data demonstrating biosimilarity should not be included on the label n Statement indicating that the product is a biosimilar and that similarity between the drugs has been established n Comparative data generated by the biosimilar for which the decision for market authorization was made summarized in tabular format n Relevant safety and efficacy information from the biologic drug authorized in Canada to which a reference is made n There should be no claims for bioequivalence or clinical equivalence in the first line for mbca was shown, in a randomized double-blind study, to produce a similar response, with a risk ratio for orr of 0.940 [95% confidence interval: 0.842 to 1.049 (within the pre-specified equivalence margin of 0.8 to 1.25)] 13 . The 1-year pfs was 56% for PF-05280014 and 52% for trastuzumab. The 1-year survival was similar (88.84% vs. 87.96%). Safety was also similar in the two groups (Table iii).

REGULATORY CONSIDERATIONS IN BIOSIMILAR DEVELOPMENT Comparability Trial Endpoint
When new cancer therapies are evaluated, os has been considered the "gold standard," especially by regulatory authorities. In reality, survival endpoints might not be sensitive enough when considering comparability (such as in the case of biosimilar agents). It might also mean that the sample size required for adequate statistical inference would be prohibitively large.
In 2015, Jackisch et al. 17 evaluated the sensitivity of endpoints for both mbca and ebc in similarity studies of trastuzumab and biosimilar agents. They used orr and pfs data for mbca and total pcr and event-free survival (efs) in the neoadjuvant setting reported in a meta-analysis of data from trastuzumab clinical trials. Despite prior findings suggesting tumour response as a potential surrogate for pfs in mbca 18 , they found that using the shorter-term endpoint of orr to measure equivalence could lead to substantial differences in long-term pfs. At an equivalence margin of 10% for orr, a difference in pfs of close to 3.2 months could be observed. For equivalence margins of 15%, the difference in pfs could be more than 4 months. For a stricter margin of 5% for orr, the sample size required to correlate with pfs would be close to 4000 patients. As with orr and pfs, total pcr and efs have been shown to correlate, at least in the her2+ subgroup 19 . Again, based on a meta-analysis of clinical trials, Jackrisch et al. 17 calculated that a 10% equivalence margin in total pcr corresponded with a difference of 3.8% in 3-year efs; for a 15% equivalence margin, the predicted loss in 3-year efs was 6.8%.

Extrapolation of Indications
Based on the results of phase iii clinical trials, trastuzumab is currently indicated in the treatment of her2+ mbca, ebc, and metastatic gastric cancer. "Extrapolation of indications" allows an agent to be used for the treatment of certain conditions or in populations in which it has not been directly studied, based on evidence of similarity in another condition or population. For example, if an agent were to be shown to be similar to trastuzumab in mbca, extrapolation of indications would allow that agent to be used in ebc or gastric cancer.
Regulatory bodies have made recommendations about indication extrapolation for biosimilars (Table i). A common requirement is that relevant mechanisms of action should be the same for the reference drug and for the biosimilar agent. For trastuzumab, multiple mechanisms of action are proposed as being important for effectiveness, including her2 degradation, antibody-dependent cellular cytotoxicity, and interference in downstream signalling 20 . The relative contribution of each mechanism to the treatment of various populations or various cancers is unknown, and each might be disproportionately affected in biosimilars by production differences and post-translational modifications.
Beyond mechanistic uncertainty, other factors might limit clinical similarity in various indications. Minor differences in drug products can affect immunogenicity, and those differences are essentially impossible to exclude without clinical trials. To support indication extrapolation, it is therefore recommended by the fda that immunogenicity be investigated in the patient population that carries the highest risk of an immune response and immune-related adverse events. Similarly, the ema suggests that clinical trials be carried out in a sufficiently sensitive and homogeneous population. Most mbca patients have already been exposed to treatment with chemotherapy, radiotherapy, or both, which have known immunosuppressive effects 21 . As a result, compared with women affected by ebc, women affected by mbca are more likely to have some degree of immunologic impairment, making mbca patients a less-sensitive population for comparisons of biosimilars in clinical trials. Moreover, patients with metastatic disease are a heterogeneous group, with variation in prior treatment exposure, line of therapy, disease burden, comorbidities, and location of metastases. Given those issues, it would be reasonable to consider extrapolation from the neoadjuvant or adjuvant setting to metastatic disease; the reverse, however, would be uncertain.
Extrapolation of indications can also lead to extrapolation to various combinations of drugs in which the biosimilar agent has not been directly tested. The MYL-1401O study (Table iii) looked at the biosimilar agent in combination with docetaxel in the first line for metastatic disease, because that combination was the standard of care established by Slamon et al. 16 in 2001. Since that time, however, the combination of taxane chemotherapy and trastuzumab with pertuzumab, a second her2-targeted antibody, has been associated with increased os and is now established as the standard of care in the first line for mbca. Whether the equivalence demonstrated in the Heritage study is representative of the combination with pertuzumab is uncertain.

Interchangeability and Exchange
"Interchangeability" implies that two medical treatments have an identical therapeutic effect, such that patients can be safely switched from one to the other. Despite trials to ensure that biosimilars are therapeutically equivalent, it might not be safe for patients already maintained on a biologic agent to be switched over to a biosimilar, and vice versa. Minor differences in structure have the potential to result in a serious immunologic effect. For that reason, data showing the safety and efficacy of switching from the reference drug to the biosimilar agent must be presented before the two agents are considered interchangeable.
In the United States, the fda has two pathways for licensing biosimilar products. One is relatively simple and is used for agents that will not be deemed interchangeable; interchangeability requires the other, more complex path. In Europe, the interchangeability designation is determined by each individual country; currently, no consistent definition has been established. Without adequate interchangeability evidence, biosimilars are usually prescribed only to patients who have not previously been treated with the reference drug. That limitation on prescribing can restrict the degree to which the biosimilar can competitively affect the price of the reference drug.

Pharmacovigilance and Safety Monitoring
As is the case with most biologics, clinical testing before the approval of a biosimilar might not identify all possible associated adverse events. Evaluation of clinical safety for biosimilars should therefore continue after marketing begins 22 . All manufacturers of biologics, including those manufacturing biosimilars, must submit pharmacovigilance plans as part of the marketing authorization application 22 . The plan includes pre-and post-authorization immunogenicity testing, a risk management plan based on safety issues identified during the clinical trials, and post-marketing safety commitments such as targeted questionnaires, phase iv studies, and specialized follow-up for long-term use 23,24 .
The goal of a pharmacovigilance plan is to identify and understand the frequency and nature of product-associated adverse events that might not have been observed during clinical testing, and to provide a framework to rapidly report and manage such incidences 24,25 . Given that the full scope of trastuzumab's mechanisms of action remains largely unknown, post-authorization immunogenicity testing is of particular importance for the pharmacovigilance plan of trastuzumab biosimilars 26 . Subsequent immunogenicity events can range from inconsequential non-neutralizing antibodies to more severe toxicities, including loss of efficacy and build-up of true resistance to the reference product. The potential effects on both efficacy and safety render the immunogenicity of a trastuzumab biosimilar a critical feature that must be carefully monitored in post-marketing settings 26 . The ema, fda, and Health Canada have guidelines for post-marketing surveillance of biosimilars (Table i).
Central to all pharmacovigilance plans is the need to be able to accurately trace the medicine given to a patient, which makes labelling of the product is important 27 . In the United States in 2014, the fda issued updated regulations for labelling, including detailed recommendations for the labelling of biosimilar products. Health Canada also updated their guidelines to include recommendations on labelling. Health Canada recommends including a table containing information about the comparability testing, but the fda intentionally suggests excluding such a table on the grounds that it might be confusing or potentially misleading ( Table i). The ema has taken a different approach, recommending that, as with generic drugs, biosimilars should derive the summary of product characteristics for the label from the reference product. That recommendation has been criticized for not highlighting the differences between biosimilars and reference products.

THE VALUE OF BIOSIMILARS
Much of the value proposition for biosimilars consists of the potential for a substantial cost reduction, based on previous experience with generic small-molecule compound drugs 28 . However, given the considerations discussed here, the manufacturing of biosimilars clearly requires a more extensive and lengthier clinical testing program 29 .
The research and development costs of biosimilars are many times those for developing and manufacturing small-molecule generics 30 . For example, in the United States, estimates suggest that bringing a biosimilar to market will cost between $10 million and $40 million and will take 6-9 years; for generics, the expected cost is $1-$2 million, with a 3-year timeline 31 . Biosimilars are therefore likely to be marketed only at a 20%-30% discount compared with the cost of the original products; generics are generally sold at about a 75% discount 31 . Preferential formulary placement for the biosimilar would require a discount of 20%-50% compared with the originator biologic 32 .
In addition, few drug manufacturers have the complex research and development capabilities to advance a biosimilar to market. It is therefore unlikely that the competition dynamics for biosimilars will echo those of the small-molecule drug market 33 . In the absence of a significant discount, preference in the short-term might be given to the reference biologic 32 . Although it is not the intention of similarity exercises, a biosimilar agent might prove to be more effective than the reference drug-for example, when the Samsung Bioepis and Mylan products demonstrated higher pcr rates in comparisons with trastuzumab (Table i). Such a situation could open the door to further increased pricing.
Because branded biologic agents are generally associated with high cost, the value of biosimilars to the health care system might be inversely associated with the accessibility and affordability of branded biologic agents 32 . In countries in which branded biologic agents are traditionally available, patients and physicians alike might question the justification of administering a biosimilar solely to lower costs 32 . The lack of education for many stakeholders involved in the biosimilar decision-making process could further limit the perceived value of biosimilars and their integration into clinical practice 34 . Many key parties in countries in which branded biologic agents are available and accessible might be unfamiliar with biosimilars: how they differ from generics and interchangeable biologic drugs, whether the differences are or are not important, and how to manage any perceived risk that might be associated with biosimilars 34 .
In the long term, predictability and clarity in the regulatory requirements and market for biosimilars could encourage greater participation of diverse stakeholders worldwide in biosimilar development and could support more price competition 32 . By 2020, more than US$67 billion of total global sales of biologic therapies will be coming off-patent in the United States and the European Union 28 . Even with a discount as low as 20% from biosimilars, the projected savings in the United States are substantial, ranging from US$3 billion to US$4.5 billion annually, and up to US$378 billion over the next two decades 31 . An annual savings of €1.6 billion has been predicted for Europe if biosimilars to 5 patent-expired biologic drugs are successfully developed 30 . Because of escalating health care costs, biosimilars show great promise for enhancing access to necessary medications for larger audiences, while containing payer costs in various disease states 35 .

CONCLUSIONS
Biosimilar agents are being developed at a rapid pace and will play an important role in cancer treatment. The aim of clinical trials with biosimilars is to demonstrate clinical equivalence. Regulatory guidelines for biosimilar antibodies exist, but questions about how best to assess equivalence and to integrate those agents into clinical practice remain. Introducing them into the standard of care will be a dynamic process involving multiple stakeholders, but could potentially help to control the cost of cancer care. Whether biosimilar agents will be the sole option available, or whether the choice of the reference biologic will be available, remains to be seen. Education for physicians about these complex issues will be imperative.