Case 1:18-cv-00924-CFC-SRF Document 563-12 Filed 11/09/20 Page 1 of 54 PageID
`#: 38028
`
`Articles
`
`Amgen.
`
`© 2018 Elsevier Ltd. All rights reserved.
`
`Trastuzumab is approved in many countries for the
`treatment of metastatic breast cancer, early breast cancer,
`and metastatic gastric cancer, 1
`' and it is the standard of
`care for patients with HER2-overexpressing breast
`cancers. 3
`' Trastuzumab is a monoclonal antibody that
`binds to the extracellular domain of HER2, blocking
`receptor activation and the subsequent proliferation of
`cells expressing HER2. It also induces the downstream
`effects of antibody-dependent cellular cytotoxicity in and
`cellular phagocytosis ofHER2-expressing cells.'
`Several trastuzumab biosimilars are in development.
`Guidelines for the development of biosimilars recom(cid:173)
`mend a totality of evidence approach with stepwise
`development
`to
`ensure comprehensive analytical
`characterisation. Studies should include structural and
`functional assessments followed by phase 1 pharmaco(cid:173)
`kinctic and, if feasible, pharmacodynamic studies to
`show similarity to the reference product. 78 At least one
`comparative clinical study in a representative population
`with sensitive endpoints (ie, are clinically relevant,
`readily assessible, and show a size of treatment effect
`that is large enough to detect differences between similar
`treatments if any exist) is also needed to confirm
`similarities in safety, efficacy, and immunogenicity.'
`
`Results from phase 3 studies have shown clinical
`similarity to trastuzumab reference product for CT-P6
`Incheon, South Korea), 10 MYL-14010
`(Celltrion,
`(Biocon, Bangalore, India, and Mylan, Canonsburg,
`PA, USA)," and SB3 (Samsung Bioepis, Incheon.
`South Korea and Merck, Kenilworth, NJ, USA). 12
`Two studies were done in the ncoadjuvant setting"n'
`and one in the metastatic setting. 11 No studies, however,
`have been designed to assess the effect of switching
`from
`the
`trastuzumab reference product
`to
`the
`biosimilar. The
`trastuzumab biosimilar ABP 980
`(Amgen Inc, Thousand Oaks, CA USA) is analytically
`similar to trastuzumab with respect to structure,
`function, and pharmacokinetic profile, 13 which suggests
`that
`there should be no clinically meaningful
`differences between these drugs in efficacy, safety, or
`irnmunogenicity.
`We assessed the clinical similarity of ABP 980 and
`trastuzumab in women with HERZ-positive early breast
`cancer in the neoadjuvant and adjuvant settings, based
`on the proportion of patients achieving a pathological
`complete response. We compared safety, tolerabil ity, and
`immunogenicity, including after switching treatment
`from trastuzumab to ABP 980 to generate data about
`clinical use.
`
`Evidence before this study
`We searched Pub Med on June 11, 2012, for papers on
`trastuzumab in the neoadjuvant treatment of early breast
`cancer with the search terms ("trastuzumab" AND
`"neoadjuvant" AND "breast"). We identified 220 papers that
`included clinical studies and reviews of trastuzumab and papers
`on other topics that discussed trastuzumab. We selected studies
`in which data were collected for HER2-positive patients;
`neoadjuvant treatment included at least epirubicin or
`doxorubicin and cyclophosphamide in combination with a
`taxane (docetaxel or paclitaxel) for at least 18 weeks;
`pertuzumab or lapatinib were not allowed; the definition of
`pathological complete response was consistent with that
`proposed for this study; and patients received neoadjuvant
`trastuzumab treatment for at least 18 weeks. We also included
`data from an abstract presented at the Annual Meeting of the
`American Society of Clinical Oncology, Chicago, IL, USA,
`June 3-7, 2011, and two studies that were unpublished at the
`time of the literature search but have since been published.
`Together, the studies showed thattrastuzumab was safe and
`effective for the neoadjuvant treatment of early breast cancer.
`
`Added value of this study
`In this randomised, double-blind, phase 3 comparative trial, we
`assessed ABP 980 (Amgen Inc, Thousand Oaks, CA, USA) as a
`
`potential biosi rnilar to trastuzu mab for the treatment of
`HER2-positive early breast cancer. We assessed safety based on
`pathological complete response in breast tissue and axillary
`lymph nodes. During the adjuvant phase some patients in the
`trastuzumab group switched to ABP 980, which allowed
`assessment of the clinical safety and immunogenicity. We were
`also able to assess the feasibility of central independent
`pathological review of response in a large multi centre study. To
`our knowledge, these are novel study design features. Our
`results add to the totality of evidence generated in analytical,
`functional, and pharmacokinetic studies and support clinical
`similarity of ABP 980 to thetrastuzumab reference product.
`
`Implications of all the available evidence
`All the data indicate that there are no clinically meaningful
`differences between ABP 980 and trastuzumab. Our findings
`add to the growing body of evidence supporting the potential
`clinical usefulness of ABP 980. Additionally, switching from
`trastuzumab to a biosimilar seems to be safe. The use of
`trastuzumab biosimilars could expand treatment options for
`clinicians, mitigate cost barriers for payers, and increase
`patients' access to important therapy.
`
`www.thelancet.com/oncology Published online June 4, 2018 http://dx.doi.org/10.1016/S1470-2045(18)30241-9
`
`AMGKAN02922077
`
`

`

`Case 1:18-cv-00924-CFC-SRF Document 563-12 Filed 11/09/20 Page 2 of 54 PageID
`#: 38029
`
`Articles
`
`seeoniimforappendix
`
`Study design and participants
`We designed a randomised, multicentre, double-blind,
`active-controlled, phase 3 equivalence trial to compare
`ABP 980 with trastuzumab in adult women with HER2-
`positive early breast cancer. Patients were recruited from
`97 study centres in 20 countries, mainly in Europe and
`South America (appendix pp 14-16).
`Eligible patients were women aged 18 years or older with
`histologica1Jy confirmed invasive breast cancer and an
`Eastern Cooperative Oncology Group performance status
`score of O or 1, who were planning to have surgical
`resection of their breast tumour with sentinel or axillary
`lymph node dissection and neoadjuvant chemotherapy.
`Inclusion criteria were HERZ-positive disease confirmed
`by a central laboratory before randomisation (defined as
`3+ overexpression on immunohistochemistry or HER2
`amplification on fluorescence
`in situ hybridisation),
`known oestrogen-receptor and progesterone-receptor
`status at study entry, measurable disease in the breast after
`diagnostic biopsy (defined as longest tumour diameter
`2:2-0 cm), and left ventricular ejection fraction (LVEF)
`of at least 55% on a two-dimensional echocardiogram.
`Exclusion criteria were presence of bilateral breast cancer
`or known distant metastases; previous treatment for
`primary breast cancer,
`including chemotherapy, a
`biological agent, radiotherapy, or surgery; concomitant
`active malignancy; and malignant disease
`in
`the
`previous 5 years, except treated basal-cell carcinoma of the
`skin or carcinoma in situ of the cervix.
`The protocol was reviewed and approved by the relevant
`independent ethics committees for each centre. All
`patients provided written informed consent. This study
`was done in accordance with the terms of the Declar;:ition
`of Helsinki, Good Clinical Practice guidelines, and all
`applicable regulatory requirements.
`
`Randomisation and masking
`All patients had to complete screening and ;:i 12-week run(cid:173)
`in period of chemotherapy to be eligible for randomisation.
`After run-in, patients were randomly assigned 1:1 to
`receive ABP 980 or trastuzumab. R;:indomis;:ition was
`stratified by T stage (<T4 vs T4), node status (yes vs no),
`hormone receptor status (positive for oestrogen receptor,
`progesterone receptor, or both vs negative for oestrogen
`receptor and progesterone receptor), planned paclitaxel
`dosing schedule (once weekly for 12 weeks vs every 3 weeks
`for four cycles), and geographical region (eastern Europe
`vs western Europe vs other). Sentinel lymph node
`assessment was not a stratification factor.
`We used a computer-generated randomisation schedule
`with a permuted block design (blocks of four) in each
`str;:itum, which w;:is prepared by PRA International
`(Paris, France) before the start of the study, to assign
`patients to treatment groups. At the start of screening,
`each patient received a unique identification number
`before undergoing any study procedures. This number
`
`was used for individual patient identification throughout
`the study, although it was not necessarily the same as the
`randomis;:ition number. Upon completion of run-in
`chemotherapy, researchers at the study sites used an
`interactive voice and web response system (IXRS, Almac,
`Souderton, PA, USA) to receive a centrally assigned
`unique randomis;:ition number th;:it was used for centr;:il
`randomisation of each patient to treatment group and
`treatment allocation. Patients were randomly assigned to
`receive ABP 980 throughout the study, trastuzumab
`throughout the study, or neoadjuvant trastuzumab
`followed by ;:idjuvant ABP 980.
`The pharmacists who prepared investigational products
`were aware of treatment allocation. Patients, physicians,
`the sponsor, investigators, ;ind study site staff were
`masked to treatment allocation until the final database
`was locked. 1he pathologists who assessed complete
`response at the local and centra I laboratories were also
`masked to treatment allocation.
`
`Procedures
`During the 28-day screening period, we took patients'
`medical histories, did physical examimtions, electro(cid:173)
`cardiograms,
`two-dimensional echocardiograms, and
`laboratory testing in blood samples, assessed vit;:il signs,
`serious adverse events, and disease progression or
`recurrence, and established Eastern Cooperative Oncology
`Group performance status score (assessed locally) and
`HER2 and hormone receptor statuses (assessed centrally).
`After screening, patients entered
`the 24-week
`neoadjuvant treatment phase. This phase began with a
`12-week run-in chemotherapy period (during which
`clinical response was not assessed) when patients
`received intravenous epirubicin 90 mg/mZ and cyclo(cid:173)
`phosphamide 600 mg/m2 every 3 weeks for four cycles.
`After run-in chemotherapy and surgery, patients with
`adequate cardiac function, assessed by left ventricular
`ejection fraction on two-dimensional echocardiograms,
`were randomly assigned to one of three treatment
`groups: ABP 980, trastuzumab, or neoadjuvant trastu(cid:173)
`zumab followed by adjuvant ABP 980. Neoadjuvant
`tre;:itment began with one cycle of 8 mg/kg investi(cid:173)
`gational product (ie, either ABP 890 or trastuzumab)
`given in an intravenous infusion over 90 min as a
`loading dose; administration as a push or bolus dose
`was not allowed. Trastuzumab and ABP 980 were
`received in 150 g vials of lysophilised sterile powders
`that were qualitatively and quantitatively the same. The
`containers, however, differed in appearance, and to
`achieve masking the products were reconstituted with
`7 · 2 mL sterilised water for injection, yielding 7 · 4 mL
`solutions containing approximately 21 mg/mL of either
`drug, and transferred to intravenous bags labelled with
`patients' randomisation numbers.
`If the loading dose was tolerated, patients received
`three cycles of trastuzumab or ABP 980 6 mg/kg given as
`30 min intravenous infusions once every 3 weeks. All
`
`www.thelancet.com/ oncology Published onli ne June 4, 2018 http://dx.doi.org/10.1016/51470-2045(18)30241-9
`
`3
`
`AMGKAN02922078
`
`

`

`Case 1:18-cv-00924-CFC-SRF Document 563-12 Filed 11/09/20 Page 3 of 54 PageID
`#: 38030
`
`Articles
`
`patients also received intravenous paclitaxel 175 mg/mZ
`with all doses of investigational product (or 80 mg/m2
`every week for 12 cycles if that was the local standard of
`care). Patients were observed to check for infusion(cid:173)
`related symptoms for at least 6 h after the start of the first
`infusion and for 2 h after the start of subsequent
`infusions. Interruption or slowing of the rate of the
`infusion was allowed if infusion-related symptoms
`occurred, and could be resumed at the 30 min infusion
`rate once symptoms abated.
`Patients underwent surgery (lumpectomy or mast(cid:173)
`ectomy with sentinel or axillary lymph node dissection)
`within 3-7 weeks of receiving the last dose of neoadj uvanl
`investigational product, then entered the adjuvant phase.
`During the adjuvant phase, patients either continued
`with ABP 980 or trastuzumab (dose 6 mg/kg) or switched
`from trastuzumab to ABP 980 6 mg/kg intravenous
`infusions given over 30 min every 3 weeks for up to
`1 year after the first dose of neoadjuvant treatment.
`Laboratory assessments were done during screening
`(visit 1), dming treatment (neoadjuvant phase visits
`2-9 and adjuvant phase visits 10-22), and at the end of
`the study, 30 days after the end of treatment (visit 23).
`TI1ese assessments were serum chemistry (visits 1, 2-9,
`14, 18, 22, and 23), haematology (visits 1, 2-9, 10-22,
`and 23), measurements of antibodies against the
`investigational product (immunogenicity; visits 1, 5, 9,
`10, 14, 18, 22, and 23); and pharmacokinetics (visits 5-9,
`10, 14, 18, 22, and 23).
`Patients could withdraw from the study at any time and
`for any reason. Safety concerns (eg, due to an adverse
`event, failure to use contraception. or protocol require(cid:173)
`ments) and disease progression or recurrence were
`clinically assessed at each visit as potential causes for
`withdrawing patients from the investigational product or
`procedural assessments per protocol.
`We did not allow investigational product dose adjust(cid:173)
`ments, but if LVEF decreased from the value seen on
`echocardiograms after chemotherapy run-in and before
`randomisation by 10 percentage points or more and to less
`than 50%, treatment was suspended and a repeat LVEF
`assessment was done within approximately 3 weeks. If
`LVEf had not improved or had declined further, the
`investigational product was discontinued. If symptomatic
`cardiac failure developed, it was treated according to local
`standard of care. Administration of an investigational
`product could be delayed or discontinued for decreases in
`LVEF, symptomatic cardiac failure, or other adverse events.
`Based on the known safety profile of trastuzumab, we
`prespecified cardiac
`failure. neutropenia,
`infusion
`reactions, pulmonary
`toxicity, hypersensitivity, and
`infections and infestations as events of interest. We used
`Standardized MedDRA Queries to retrieve relevant
`system organ classes and preferred terms in the Medical
`Dictionary for Regulatory Activities version 19.0, if
`available. If no standardised query was available for a
`given event of interest, we used a customised search
`
`strategy to identify relevant terms. Investigators graded
`adverse events according to Common Terminology
`Criteria for Adverse Events version 4.0. Previous and
`concomitant medications were coded with the WHO
`Drug Dictionary version 2015 D ECOL
`Adverse events and disease progression or recurrence
`were assessed at all visits during the neoadjuvant and the
`adjuvant phases. Two-dimensional echocardiography
`was done at screening, and at visits 5 and 9 of the
`neoadjuvant phase, and results were assessed before
`administration of the investigational product. During the
`adjuvant phase, we assessed patients for adverse events,
`concomitant medications, and disease progression or
`recurrence at all visits and did two-dimensional echo(cid:173)
`cardiograms at visits 14 and 18.
`The efficacy analysis was done after the last patient had
`had surgery and been assessed for pathological complete
`response or had withdrawn from the sh1dy. Here we
`present the pathological complete response efficacy
`analysis and the safety and immunogenicity data from the
`final database lock. All tumour samples were assessed by
`local pathologists. Representative tumour samples were
`laboratory for assessment by
`sent
`to
`the central
`two independent central pathologists who were unaware
`of each other's findings. The pathologists determined the
`samples as adequate or inadequate for evaluation based
`on the presence or absence of tumour bed and integrity or
`loss of nuclear detail. The central pathology findings were
`documented on worksheets specifically developed for the
`study and included the folio-wing items: adequate or
`inadequate specimen quality; presence or absence of
`tumour bed; presence or absence of invasive breast cancer;
`results differing from the local assessment for the number
`of blocks with invasive breast cancer present; results
`differing from the local assessment for the estimated
`percentage of viable residual tumour; presence or absence
`of ductal carcinoma in situ; presence or absence oflymph
`nodes; and presence or absence of lymph-node-invasive
`cancer. If the central results were concordant, those from
`first central pathologist were entered into the database
`and were deemed to be representative. If results were
`discordant,
`the worksheets were
`reviewed by an
`adjudicating pathologist who made a final independent
`interpretation, which was entered into the database.
`
`Assays validated according to FDA guidance1' were
`used to detect antibodies against the investigational
`products. All samples were first tested in an electrochemi(cid:173)
`lurninescence-based bridging immunoassay that used
`ABP 980 as antigen to detect binding antibodies. Samples
`were then tested to confirm specificity of response. Those
`that showed signal inhibition greater than the drug
`depletion cutoff point in the presence of excess soluble
`drug were reported as positive for binding antibodies
`against investigational products. Positive samples were
`tested in a non-cell-based, time-resolved, fluorescence(cid:173)
`based competitive target-binding assay to determine
`neutralising activity. A confirmatory assay was done on
`
`4
`
`www.thelancet.com/oncology Published online June 4, 2018 http://dx.doi.org/10.1016/S1470-2045(18)30241-9
`
`AMGKAN02922079
`
`

`

`Case 1:18-cv-00924-CFC-SRF Document 563-12 Filed 11/09/20 Page 4 of 54 PageID
`#: 38031
`
`Articles
`
`all samples to determine whether the inhibition of drug
`activity was due to neutralising antibodies to ABP 980. A
`post-treatment sample was defined as positive for
`neutralising antibodies if it was simultaneously positive
`for binding antibodies and neutralising activity.
`We recorded the numbers and percentages of patients
`in each treatment group who had pre-existing or
`developed binding and neutralising binding antibodies
`against investigational products. Pre-existing antibody
`incidence was defined as the number of patients with
`positive antibody results at the time of or before the first
`dose of investigational product divided by the number of
`patients with an immunoassay result on or before the
`first dose. We defined patients who developed antibodies
`as the number of patients with a negative antibody result
`or no result available at or before baseline and a positive
`antibody result at any time after the first dose of
`investigational product divided by the number of patients
`with at least one immunoassay result after baseline. A
`transient antibody result was defined as a positive result
`after baseline with a negative result at the patient's last
`time tested within the study period.
`
`Outcomes
`The co-primary efficacy endpoints were risk difference
`and risk ratio (RR) of pathological complE'te rE'sponse,
`defined as the absence of invasive tumour cells in the
`breast tissue and in axillary lymph nodes regardless of
`ductal carcinoma in situ (as defined by the FDA).'' The
`primary analysis was based on local laboratory findings
`in patients with assessable tumour samples. We did
`sensitivity analyses based on central pathology findings
`to reduce variability benveen pathologists at the local
`level. Efficacy results are reported for the neoadjuvant
`phase (ABP 980 and trastuzumab groups).
`Secondary efficacy endpoints were risk differences and
`RRs for pathological complete response in breast tissue
`(absence of invasive tumour cells, regardless of residual
`ductal carcinoma in situ); 1isk differences and RRs for
`pathological complete response in breast tissue and
`axillary lymph nodes in the absence of ductal carcinoma in
`situ (defined as the absence of invasive tumour cells in
`breast tissue and axillary lymph nodes and absence of
`ductal carcinoma). TI1ese results will be reported separately.
`Safety assessments reported in this Article are the
`incidence of treatment-emergent adverse events, changes
`in LVEF, exposure
`to
`investigational product and
`paclitaxel, and formation of antibodies against an
`investigational product (immunogenicity). Safety results
`are presented for the neoadjuvant phase (ABP 980 and
`trastuzumab groups) and adjuvant phase (ABP 980,
`trastuzumab, and switching groups). Other safety
`outcomes that will be reported elsewhere were on-study
`event-free survival, overall survival, pharmacokinetics,
`concomitant medications, laboratory tests (including
`serum chemistry and haematology), vital signs, and
`physical examination.
`
`Subgroup analyses done in prespecified groups for the
`neoadjuvant phase, adjuvant phase, and entire study.
`These included age group, race, T stage, axillary lymph
`node involvement, hormone receptor status, paclitaxel
`dosing schedule, and geographical region, and will be
`reported separately.
`
`"
`
`Statistical analysis
`The primary efficacy hypothesis was that ABP 980
`would be equivalent to trastuzumab when each was
`given in combination with standard-of-care neoadjuvant
`cancer treatment (paclitaxel). The planned sample size
`was 808 lo ensure that 768 patients (384 in each group)
`were randomly assigned treatment. We calculated that
`this number would achieve 90% power to show
`equivalence when assessed by RR for pathological
`complete response with 5% dropout during run-in
`chemotherapy phase. This sample size was also
`calculated to provide at least 90% power to show
`equivalence when assessed by risk difference between
`groups for pathological complete response with margins
`of -13% and 13% and a two-sided 0-05 significance
`level. We assumed that the proportion of patients who
`would achieve a pathological complete response would
`be approximately 42 · 5%
`in
`the ABP 980 and
`trastuzumab groups. 1
`We initially used a sequential testing method to test
`similarity between ABP 980 and trastuzumab by
`comparing the two-sided 90% CI for risk difference
`between the ABP 980 and trastuzumab groups with
`statistical margins of -13% and 13%. If the test on the
`risk difference was successful, similarity was then tested
`by RR of pathological complete response at a hvo-sided
`significance level of O -05 by comparing the two-sided
`90% CI between the ABP 980 and trastuzumab groups
`with statistical margins of O · 759 and 1 · 318.
`The population assessable for pathological complete
`response was defined as all randomised patients who
`received any amount of investigational product,
`underwent surgery, and had an available pathological
`complete response assessment from the local laboratory.
`The safety analysis population consisted of all patients
`who were randomised and received any amount of
`investigational product. We did sensitivity analyses in
`the intention-to-treat and per-protocol populations (data
`not shown). The intention-to-treat population included
`all patients randomly assigned to a study group,
`regardless of whether they received any investigational
`product. The per-protocol population
`included all
`patients who were randomised, had local laboratory
`pathological complete response results, and had no
`protocol deviations that prevented assessment of the
`primary objective.
`All statistical analyses were done with SAS version 9.1.3
`or later. This study is registered with ClinicalTrials.gov,
`number NCT01901146,
`and Eudra, number CT
`2012-004319-29.
`
`www.thelancet.com/ oncology Published onli ne June 4, 2018 http1// dx.doi.org/10.1016/51470-2045(18)30241-9
`
`5
`
`AMGKAN02922080
`
`

`

`Case 1:18-cv-00924-CFC-SRF Document 563-12 Filed 11/09/20 Page 5 of 54 PageID
`#: 38032
`
`Articles
`
`I 906 patients screened
`
`,,
`
`I 827 patients enrolled
`
`I 725 patients randomly assigned
`
`79 excluded
`13 did not sign consent
`20 missed screening visits or procedures
`6 no measurahle diseasP
`5 positive for HBsAG, HCV, or HIV
`35 other reasons
`
`102 excluded
`60 HER2 expression unconfirmed
`15 withdrew rnnsent
`9 physician decision
`6 disease progression or recurrence
`3 pmlocol violation
`9 other reasons
`
`I
`
`I
`
`I
`
`I
`
`I
`
`364 patients assigned to ABP 980
`
`6 did not compiete surgery
`2 withdrew consent
`2 disease progression or recurrence
`1 physician decision
`ldied
`
`'
`358 assigned to adjuvant phase
`
`9 did not start adjuvant treatment
`2 withdrew conser1t
`1 disease progression or recu rrence
`4 physician decision
`1 lostto follow-up
`1 needed alternative treatment
`
`I 361 patients assigned to trastuzumab
`
`I
`
`14 did not complete surgery
`5 withdrew consent
`3 d isease progression or recurrence
`2 physician decision
`2died
`1 lost to follow-up
`l needed alternative treatment
`
`H 9 manually assigned to trastuzumab
`
`group excluded from efficacy analysis*
`
`~~.~
`
`338 included in neoadjuvdnt efficacy
`analysis and assigned to adjuvant
`phase
`
`..
`I 347 assigned to acljuvant phase
`
`I
`
`5 did not start adjuvant treatment
`2 withdrew consent
`2 physician decision
`1 protocol violation
`
`I
`
`I
`
`I
`
`I
`I 349 continued to adjuvant ABP 90
`~~ - - - - - - - - - - - - - - - - -~
`
`j 171 continued to adjuvant trastuzumab
`
`I I 171 switched from trastuwmab to adjuvant ABP 90
`
`26 diswntinued adjuvant lreatrnent
`4 withdrew consent
`12 disease progression or recurrence
`9 physician decision
`1 other reason
`
`-+
`
`7 discontinued adjuvant treatment
`1 withdr·ew consent
`4 disease progression or recurrence
`2 physician decision
`
`14did no l star l adjuvant Lreatrnenl
`4 withdrew consent
`3 disc.isc progression or recurrence
`4 physician decision
`3died
`
`I 1164 completed both phases of the study
`I 323 completed both parts of the study
`~------------------~
`
`I 1157 completed both phases of the study
`
`Figure 1c Trial profile
`HBsAg=hepatitis B surface antigen. HCV=hepatitis C virus. *Nine patients were assigned to the trastuzumab group because of a delay in manufacturing of ABP 980 at the start of the study. These
`patients were excluded from the primary efficacy analysis but included in the final safety analysis.
`
`6
`
`www.thelancet.com/oncology Published online June 4, 2018 http://dx.doi.org/10.1016/S1470-2045(18)30241-9
`
`AMGKAN02922081
`
`

`

`Case 1:18-cv-00924-CFC-SRF Document 563-12 Filed 11/09/20 Page 6 of 54 PageID
`#: 38033
`
`Articles
`
`Role of the funding source
`The funder had a role in study design, data analysis, data
`interpretation, and writing of the report, and had access
`to the raw data, but had no role in data collection. The
`corresponding author had full access to all the data in the
`study and had final responsibility for the decision to
`submit for publication.
`
`We enrolled patients between April 29, 2013, and
`Sept 29, 2015. The data cutoff for the primary analysis was
`May 5, 2016, and the database lock for the final analysis
`was March 29, 2017. Of 906 patients screened, 79 were
`excluded (figure 1). 827 patients were enrolled and
`725 were randomised (figure 1). The treatment groups
`were well balanced in terms of baseline characteristics
`(table 1). The baseline distribution of sentinel lymph node
`biopsies was balanced the two groups (39 [11%] patients
`in the ABP 980 group and 29 [9%] in the trastuzumab
`group). Lymph node surgery was not done in 13 patients
`after neoadjuvant treatment because they had negative or
`only up to two positive sentinel nodes; these patients
`were equally distributed between the two treatment
`groups of the neoadjuvant phase (six [2%] of 358 patients
`in the ABP 980 group vs seven [2%] of 338 in the
`trastuzumab group) Patients' exposure to investigational
`products is shown in table 2. Exposure to paclitaxel
`during the neoadjuvant phase was similar in the
`ABP 980 and
`trastuzumab groups. Paclitaxel was
`administered only in the neoadjuvant phase. The mean
`cumulative dose for patients receiving paclitaxel every 3
`weeks was 686·0 (SD 65 ·2) mg/m' in the ABP 980 group
`and 679 · 0 (83 · 0) mg/m' in the trastuzumab group. For
`patients who received paclitaxel weekly. the mean cumu(cid:173)
`lative dose was 913·0 (SD 131·2) mg/m' in the ABP 980
`group and 906·0 (132-8) mg/m' in the trastuzumab
`group. Median follow-up was 12 months (IQR 1 ·04-1,08)
`in patients who only received ABP 980, 12 months
`(1 · 04-1 · 07) in those who only received trastuzumab, and
`12 months (1 · 04-1 · 08) in the patients who switched from
`trastuzumab to ABP 980 in the adjuvant phase.
`All patients who underwent surgery were assessable
`for the primary endpoint of pathological complete
`response (696 patients in total; 358 of whom received
`ABP 980 and 338 who received trastuzumab). 172 (48%,
`95% Cl 43-53) of 358 patients who received neoadjuvant
`ABP 980 and 137 (41%, 35-46) of 338 patients who
`received neoadjuvant trastuzurnab achieved a patho(cid:173)
`logical complete response in breast tissue and axillary
`nodes based on local laboratory assessments. The risk
`difference (ABP 980 minus trastuzumab) of pathological
`complete response was 7- 3% (90% Cl 1-2-13 ·4). The RR
`(ABP 980 vs trastuzumab) of pathological complete
`response was 1 · 188 (90% CI 1-033-1· 366). The primary
`endpoint, however, was not met, because the upper
`boundaries of the 90% Cls for risk difference and RR
`exceeded the predefined equivalence margins (figure 2).
`
`ABP 980 (n=364)
`
`Trastuzumab (11=190)
`
`Switched from
`adjuvant trastuzumab
`toABP980(n=171)
`
`53·0 (46·0-6CJ.CJ)
`
`53·0 (45 0-60,0)
`
`53 U (44·0-62,0)
`
`331 (91%,)
`
`10(3%)
`
`23(6%)
`
`175 (92%)
`
`2 (1%)
`
`13(7%)
`
`158 (92%)
`
`2(1%)
`
`11(6%)
`
`70-6 (61-60-8100)
`
`70 2 (62,00-79 00)
`
`733 (62 20-81-30)
`
`Age (years)
`
`Ethnicity
`
`White
`
`Rlack or African American
`
`Other
`
`Weight (kg)
`Geographical region
`Eastern Europe
`Western Europe
`
`Other
`
`271 (75%)
`
`43 (12%)
`so (14%)
`
`ECOG performance status score
`
`0
`
`1
`
`lUmourstage
`
`<T4
`
`T4
`
`298 (82%)
`
`66 (18%)
`
`282 (78%)
`
`82 (23%)
`
`Axi lla lymph node involvement
`
`Yes
`
`No
`
`277 (76%)
`
`87 (24%)
`
`Hormone receptor status
`
`Positive for ER, PR, or both
`
`265 (73%)
`
`Negative for ER and PR
`
`99 (27%)
`
`Histological grade
`
`141 (74%)
`
`24(13%)
`
`25 (13%)
`
`163 (86%)
`
`27 (14%)
`
`147 (77%)
`
`43 (23%)
`
`136 (72%)
`
`54 (28%)
`
`140 (/4%)
`
`50 (26%)
`
`1 (1%)
`
`93 (49%)
`
`67 (35%)
`
`29 (15%)
`
`132 (77%)
`
`22 (13%)
`
`17(10%)
`
`149 (87%)
`
`22(13%)
`
`134 (78%)
`
`37(22%)
`
`130(76%)
`
`41 (24%)
`
`128 (/5%)
`
`43(25%)
`
`0
`
`80 (47%)
`
`65 (38%)
`
`26(15%)
`
`8 (2%)
`
`174 (48%)
`
`120 (33%)
`
`62 (17%)
`
`2
`
`3
`Unknown
`
`Left ventricular ejection
`fraction(%)
`
`65 (610-68-0)
`
`65 (60 0-68·0)
`
`65 (60 0-68-0)
`
`Data are median (IQR) or n (%). Percentage values might not total 100% because of rounding. ECOG=Eastern
`Cooperative Oncology Group. ER=oestrogen receptor. PR::::cprogesterone receptor.
`
`Table 1: Baseline characteristics of safety population
`
`In the sensitivity analyses based on central pathology
`review of tumour samples, 162 (48%, 95% CI 42-53) of
`339 patients in the ABP 980 group and 138 (42%, 36-47)
`of 330 in the trastuzumab group showed pathological
`complete response in breast tissue and axillary nodes.
`The risk difference between groups and RR of
`ABP 980 versus trastuzumab were within the predefined
`equivalence margins (figure 2).
`the
`in
`The overall
`incidence of adverse events
`two treatment groups during both the neoadjuvant and
`adjuvant phases was similar (tables 3, 4, appendix pp 3-7).
`In the neoadjuvant phase. 19 (5%) of 364 patients in the
`ABP 980 group and 23 (6%) of 361 in the trastuzumab
`group had adverse events that led to dose delays of
`investigational products, three (1%) and two (1%),
`respectively, had events that led to discontinuation of
`treatment, and four (1%) and two (1%), respectively, had
`events that led to withdrawal from the study. In the
`adjuvant phase, 16
`(5%) of 349 patients
`in
`the
`
`www.tnelancet.com/ oncology Published onli ne June 4, 2018 http://dx.doi.org/10.1016/51470-2045(18)30241-9
`
`7
`
`AMGKAN02922082
`
`

`

`Case 1:18-cv-00924-CFC-SRF Doc