Articles
`
`Lancet Oncol 2014; 15: 689–99
`Published Online
`May 2, 2014
`http://dx.doi.org/10.1016/
`S1470-2045(14)70178-0
`See Comment page 668
`See Online for an audio
`interview with Ian Krop
`*Collaborators are listed in the
`appendix
`Dana-Farber Cancer Institute,
`Harvard University School of
`Medicine, Boston, MA, USA
`(I E Krop MD); Asan Medical
`Center, University of Ulsan
`College of Medicine,
`Songpa-gu, Seoul, South Korea
`(Prof S-B Kim MD);
`MD Anderson Cancer Center,
`Madrid, Spain
`(A González-Martín MD);
`Karmanos Cancer Institute,
`Wayne State University,
`Detroit, MI, USA
`(Prof P M LoRusso DO);
`Department of Medical
`Oncology, Centre Antoine
`Lacassagne, Nice, France
`(Prof J-M Ferrero MD);
`Genentech, South San
`Francisco, CA, USA
`(M Smitt MD, R Yu PhD,
`A C F Leung MD); and
`Department of General Medical
`Oncology, University Hospitals
`Leuven, Leuven, Belgium
`(Prof H Wildiers MD)
`Correspondence to:
`Dr Ian E Krop, Dana-Farber
`Cancer Institute, Harvard
`University School of Medicine,
`450 Brookline Avenue, Boston,
`MA 02215, USA
`ikrop@partners.org
`
`See Online for appendix
`
`Trastuzumab emtansine versus treatment of physician’s
`choice for pretreated HER2-positive advanced breast cancer
`(TH3RESA): a randomised, open-label, phase 3 trial
`
`Ian E Krop, Sung-Bae Kim, Antonio González-Martín, Patricia M LoRusso, Jean-Marc Ferrero, Melanie Smitt, Ron Yu, Abraham C F Leung,
`Hans Wildiers, on behalf of the TH3RESA study collaborators*
`
`Summary
`Background Patients with progressive disease after two or more HER2-directed regimens for recurrent or metastatic
`breast cancer have few eff ective therapeutic options. We aimed to compare trastuzumab emtansine, an antibody–drug
`conjugate comprising the cytotoxic agent DM1 linked to trastuzumab, with treatment of physician’s choice in this
`population of patients.
`
`Methods This randomised, open-label, phase 3 trial took place in medical centres in 22 countries across Europe,
`North America, South America, and Asia-Pacifi c. Eligible patients (≥18 years, left ventricular ejection fraction ≥50%,
`Eastern Cooperative Oncology Group performance status 0–2) with progressive HER2-positive advanced breast
`cancer who had received two or more HER2-directed regimens in the advanced setting, including trastuzumab and
`lapatinib, and previous taxane therapy in any setting, were randomly assigned (in a 2:1 ratio) to trastuzumab emtansine
`(3·6 mg/kg intravenously every 21 days) or physician’s choice using a permuted block randomisation scheme by an
`interactive voice and web response system. Patients were stratifi ed according to world region (USA vs western Europe
`vs other), number of previous regimens (excluding single-agent hormonal therapy) for the treatment of advanced
`disease (two to three vs more than three), and presence of visceral disease (any vs none). Coprimary endpoints were
`investigator-assessed progression-free survival (PFS) and overall survival in the intention-to-treat population. We
`report the fi nal PFS analysis and the fi rst interim overall survival analysis. This study is registered with ClinicalTrials.
`gov, number NCT01419197.
`
`Findings From Sept 14, 2011, to Nov 19, 2012, 602 patients were randomly assigned (404 to trastuzumab emtansine
`and 198 to physician’s choice). At data cutoff (Feb 11, 2013), 44 patients assigned to physician’s choice had crossed over
`to trastuzumab emtansine. After a median follow-up of 7·2 months (IQR 5·0–10·1 months) in the trastuzumab
`emtansine group and 6·5 months (IQR 4·1–9·7) in the physician’s choice group, 219 (54%) patients in the
`trastuzumab emtansine group and 129 (65%) of patients in the physician’s choice group had PFS events. PFS was
`signifi cantly improved with trastuzumab emtansine compared with physician’s choice (median 6·2 months [95% CI
`5·59–6·87] vs 3·3 months [2·89–4·14]; stratifi ed hazard ratio [HR] 0·528 [0·422–0·661]; p<0·0001). Interim overall
`survival analysis showed a trend favouring trastuzumab emtansine (stratifi ed HR 0·552 [95% CI 0·369–0·826];
`p=0·0034), but the stopping boundary was not crossed. A lower incidence of grade 3 or worse adverse events was
`reported with trastuzumab emtansine than with physician’s choice (130 events [32%] in 403 patients vs 80 events [43%]
`in 184 patients). Neutropenia (ten [2%] vs 29 [16%]), diarrhoea (three [<1%] vs eight [4%]), and febrile neutropenia (one
`[<1%] vs seven [4%]) were grade 3 or worse adverse events that were more common in the physician’s choice group
`than in the trastuzumab emtansine group. Thrombocytopenia (19 [5%] vs three [2%]) was the grade 3 or worse adverse
`event that was more common in the trastuzumab emtansine group. 74 (18%) patients in the trastuzumab emtansine
`group and 38 (21%) in the physician’s choice group reported a serious adverse event.
`
`Interpretation Trastuzumab emtansine should be considered as a new standard for patients with HER2-positive
`advanced breast cancer who have previously received trastuzumab and lapatinib.
`
`Funding Genentech.
`
`Introduction
`HER2 is overexpressed in about 15–20% of invasive
`breast cancers and is associated with poor clinical
`outcome in the absence of systemic therapy.1 The addition
`of trastuzumab to standard chemotherapy improves
`survival in patients with HER2-positive metastatic breast
`cancer.2,3 However, despite the effi cacy of trastuzumab,
`most patients develop progressive disease during or after
`
`trastuzumab treatment, and additional intervention is
`required. In view of the evidence that HER2 over-
`expression persists and remains relevant beyond
`progression,4 –6 strategies to overcome insensitivity to
`treat ment have involved changing the HER2-directed
`agent or switching chemotherapies in subsequent lines
`of treatment.7,8 Moreover, combination treatment with
`trastuzumab plus
`lapatinib, another HER2-targeted
`
`www.thelancet.com/oncology Vol 15 June 2014
`
`689
`
`IMMUNOGEN 2012, pg. 1
`Phigenix v. Immuogen
`IPR2014-00676
`
`

`
`Articles
`
`therapy, has been shown to improve overall survival
`compared with lapatinib alone in patients with heavily
`pretreated metastatic breast cancer.6 However, few
`clinical studies have been done
`in patients with
`progressive disease who have already received both
`trastuzumab and
`lapatinib, and re-treatment with
`trastuzumab-containing regimens seems to have only
`moderate activity in this population.9,10
`Antibody–drug conjugates, comprising a potent
`cytotoxic molecule linked to a target-specifi c antibody, are
`a class of therapeutic agents that potentially reduce
`systemic toxicities and enhance antitumour activity by
`specifi cally directing cytotoxic compounds to tumours.
`Trastuzumab emtansine is an antibody–drug conjugate
`that
`delivers
`the microtubule-inhibitory
`agent
`DM1 directly to HER2-expressing tumour cells, where it
`is internalised by lysosomes and promotes apoptosis
`upon
`intracellular
`release.11
`In binding HER2,
`trastuzumab emtansine, like trastuzumab, inhibits cell
`signalling through the PI3K/AKT pathway, inhibits
`HER2 shedding, and induces antibody-dependent cellular
`cytotoxicity.12 Trastuzumab emtansine was recently
`approved in several countries and regions, including the
`USA and the European Union, as a single-agent treatment
`for patients with HER2-positive metastatic breast cancer
`who have previously received
`trastuzumab and a
`concurrent or sequential taxane in any setting, on the
`basis of results from the phase 3 EMILIA trial.13 In
`EMILIA, use of trastuzumab emtansine was associated
`with signifi cant reductions in both the risk of disease
`progression (hazard ratio [HR] 0·65, 95% CI 0·55–0·77)
`and death (HR 0·68, 0·55–0·85), with lower grade 3 or
`worse toxicity when compared with lapatinib plus
`capecitabine.13
`Although all patients in EMILIA had previously
`received
`trastuzumab, previous
`lapatinib was an
`exclusion criterion. Data from phase 2 clinical trials have
`shown the single-agent activity of trastuzumab emtansine
`in heavily pretreated patients with previous exposure to
`trastuzumab and lapatinib,14,15 but there are no defi nitive
`studies in this population and no clear standard of care
`exists for these patients.7 Therefore, new treatment
`options are needed. TH3RESA is the second phase 3 study
`of trastuzumab emtansine done in the metastatic breast
`cancer population and was designed
`to compare
`trastuzumab emtansine with treatment of physician’s
`choice in a population with progressive disease who had
`received both trastuzumab-containing and lapatinib-
`containing regimens for advanced breast cancer.
`
`Methods
`Study design and patients
`The TH3RESA study is a randomised, multicentre, open-
`label, phase 3 trial with enrolment in 22 countries across
`Europe, North America, South America, and Asia-Pacifi c.
`Eligible patients had HER2-positive, unresectable locally
`advanced or recurrent breast cancer or metastatic breast
`
`cancer (hereafter termed advanced breast cancer), had
`previously received both trastuzumab and lapatinib in
`the advanced setting and a taxane in any setting, and had
`documented
`investigator-assessed progression after
`treatment with two or more HER2-directed regimens for
`advanced breast cancer. Disease progression had to have
`occurred during both
`trastuzumab-containing and
`lapatinib-containing regimens, with at least 6 weeks of
`exposure to each agent, except when intolerance to
`lapatinib was identifi ed. HER2-positive status of tumour
`tissue, defi ned as in-situ hybridisation positivity or 3+ by
`immunohistochemical
`analysis, was prospectively
`confi rmed by a central laboratory. Patients with non-
`measurable or measurable disease according to Response
`Evaluation Criteria
`in Solid Tumors
`(RECIST)
`version 1.1 were enrolled.16 Additional eligibility criteria
`included age of 18 years or older, a left ventricular ejection
`fraction (LVEF) of 50% or higher as measured by
`echocardiography
`or multiple-gated
`acquisition
`scanning, an Eastern Cooperative Oncology Group
`performance status of 0–2, adequate organ function
`(including platelet count >100 000 cells per μL and
`aspartate aminotransferase and alanine aminotransferase
`≤2·5 × upper limit of normal), and provision of written
`informed consent.
`Major exclusion criteria were previous enrolment in a
`clinical trial of trastuzumab emtansine, grade 3 or worse
`peripheral neuropathy according to the National Cancer
`Institute Common Terminology Criteria for Adverse
`Events (NCI CTCAE) version 4.0,17 symptomatic or
`untreated CNS metastases or treatment for such
`metastases within 1 month of randomisation, a history of
`symptomatic congestive heart failure, and a history of
`myocardial infarction or unstable angina within 6 months
`of enrolment.
`The trial protocol was approved by the relevant
`institutional review boards of each study centre, and the
`trial was done in accordance with the Declaration of
`Helsinki, Good Clinical Practice guidelines, and
`applicable local laws. All patients provided written
`informed consent.
`
`Randomisation and masking
`Study
`investigators enrolled patients, who were
`randomised to the trial with use of an interactive voice
`and web
`response system. A permuted block
`randomisation scheme, with a block size of six, was used
`to ensure an approximate 2:1 allocation of patients to
`receive
`trastuzumab emtansine or
`treatment of
`physician’s choice, respectively, with stratifi cation
`according to world region (USA, western Europe, or
`other), number of previous regimens (excluding single-
`agent hormonal therapy) for advanced breast cancer
`(two to three vs more than three), and presence of
`visceral disease (any vs none). Neither patients nor
`investigators were masked to treatment assignment in
`this open-label trial.
`
`690
`
`www.thelancet.com/oncology Vol 15 June 2014
`
`IMMUNOGEN 2012, pg. 2
`Phigenix v. Immuogen
`IPR2014-00676
`
`

`
`Articles
`
`Procedures
`Patients randomly assigned to trastuzumab emtansine
`received a dose of 3·6 mg/kg intravenously every 21 days.
`If a patient needed a dose reduction, the dose was
`reduced fi rst from 3·6 mg/kg to 3·0 mg/kg and then
`from 3·0 mg/kg to 2·4 mg/kg. Patients given trastuzumab
`emtansine 2·4 mg/kg who developed an adverse event
`necessitating dose reductions were withdrawn from the
`study. Dose interruptions for up to 42 days from the last
`treatment dose were permitted
`for
`trastuzumab-
`emtansine-related
`thrombocytopenia, hepatotoxicity,
`neuro toxicity, cardiotoxicity, infusion-related reactions or
`hypersensitivity, pulmonary toxicity, or any other clinically
`signifi cant treatment-related toxicity that did not recover
`to grade 1 or baseline. The requirements for trastuzumab
`emtansine dose delays, reductions, and discontinuations
`owing to toxicities were protocol-defi ned and in keeping
`with current prescribing information.18
`Patients randomly assigned to treatment of physician’s
`choice were given an approved systemic
`therapy
`administered as per local practice at the investigator’s
`discretion and according to the needs of each patient.
`Treatment options were restricted to chemotherapy (any
`single agent), hormonal therapy for hormone-receptor-
`positive disease (single-agent or dual therapy), or
`HER2-directed
`therapy
`(single-agent, dual HER2-
`targeted therapy, or combination with either single-agent
`chemotherapy or single-agent hormonal therapy). Best
`supportive care alone, including palliative radiotherapy
`in the absence of systemic therapy, was not permitted.
`Treatment with trastuzumab emtansine or physician’s
`choice was continued until progressive disease or
`unmanageable toxicity. From September, 2012, onwards,
`after EMILIA data were reported,13 patients who had
`progressive disease while receiving
`treatment of
`physician’s choice were eligible to cross over to
`trastuzumab emtansine treatment, starting at 3·6 mg/kg.
`Tumour assessments were done every 6 weeks for the
`fi rst 54 weeks and every 12 weeks thereafter, irrespective
`of dose delays or interruptions, until investigator-assessed
`progressive disease or death. LVEF was measured by
`means of echocardiography (preferred method) or
`multiple-gated acquisition scanning at screening, week 6
`(ie, end of cycle 2), every 12 weeks thereafter until study
`discontinuation, and 30 days after the last treatment dose.
`Local laboratory assessments were done at baseline, on
`day 1 of each treatment cycle, and 30 days after the last
`treatment dose. Patients were continuously monitored
`for adverse events, which were graded using NCI CTCAE
`(version 4.0). A serious adverse event was any adverse
`event that was fatal, life threatening, led to inpatient
`hospital admission (or an extended hospital stay), resulted
`in persistent or clinically signifi cant disability or
`incapacity, resulted in a congenital anomaly or birth
`defect in a neonate or infant born to a mother exposed to
`the investigational product, or was considered to be a
`clinically signifi cant medical event by the investigator.
`
`Data for patient-reported outcomes were obtained
`using the European Organisation for Research and
`Treatment of Cancer Quality of Life Questionnaire
`Core-30 (EORTC QLQ-C30). Patients completed the
`EORTC QLQ-C30 on day 1 of each treatment cycle until
`study treatment discontinuation or investigator-assessed
`disease progression (whichever occurred later).
`
`Outcomes
`The coprimary endpoints were investigator-assessed
`progression-free survival (PFS) and overall survival in the
`intention-to-treat population, which consisted of all
`randomly assigned patients, irrespective of whether they
`received study treatment. PFS was defi ned as the interval
`from randomisation
`to fi rst documented disease
`progression according to RECIST or death from any
`cause, whichever occurred fi rst. Overall survival was
`defi ned as the interval from randomisation to death from
`any cause. Secondary endpoints were investigator-
`assessed objective response according
`to RECIST,
`duration of objective response, 6-month survival, 1-year
`
`972 patients assessed for eligibility
`
`370 excluded*
` 107 with ineligible brain metastases
` 101 without centrally confirmed
` HER2-positive disease
` 90 for other reasons
` 54 due to inadequate organ function
` 18 due to absence of previous treatment
`
` with protocol-required therapies
`
`602 randomly assigned to study
` treatment
`
`198 allocated to treatment of
` physician’s choice (ITT population)
` 13 withdrew before treatment
` 185 received treatment†
`
`404 allocated to trastuzumab
` emtansine (ITT population)
` 2 withdrew before treatment
` 402 received treatment†
`
`144 discontinued study
` treatment by data cutoff
` 108 PD
` 8 adverse events
` 3 death
` 17 patient’s decision
` 4 physician’s decision
` 3 protocol violations or
` non-compliance
` 1 for other reasons
`
`234 discontinued study
` treatment by data cutoff
` 193 PD
` 23 adverse events
` 3 death
` 4 patient decision
` 7 physician’s decision
` 2 protocol violations
` 2 for other reasons
`
`41 on study treatment at data cutoff
`
`168 on study treatment at data cutoff
`
`Figure 1: Trial profi le
`ITT=intention-to-treat. PD=progressive disease. *Reasons for patient ineligibility comprising 5% or more of the
`total number of screen failures are presented; all other reasons for ineligibility have been grouped under “other”,
`with each individual reason representing 2% or less of the total number of screen failures. †One patient
`randomised to the physician’s choice group received two cycles of trastuzumab emtansine by mistake; this patient
`was included in the trastuzumab emtansine group for safety analyses.
`
`www.thelancet.com/oncology Vol 15 June 2014
`
`691
`
`IMMUNOGEN 2012, pg. 3
`Phigenix v. Immuogen
`IPR2014-00676
`
`

`
`Articles
`
`survival, and safety. The safety population included all
`randomly assigned patients who
`received study
`treatment. Further secondary endpoints were general
`health status or quality of life and health-related quality
`of life, symptom severity and interference, and pain
`ratings as assessed by EORTC QLQ-C30.
`
`Statistical analysis
`The overall 5% type I error rate was split asymmetrically
`between the coprimary endpoints, with 0·5% allocated to
`PFS and 4·5% allocated to overall survival. We calculated
`
`Age (years)
`<65
`65–74
`≥75
`World region
`USA
`Western Europe
`Other
`Race
`White
`Asian
`Other*
`ECOG PS†
`0
`1
`2
`Hormone receptor status‡
`ER positive and/or PR positive
`ER negative and PR negative
`Unknown
`Visceral disease involvement
`Disease extent
`Metastatic
`Unresectable locally advanced or recurrent
`Measurable disease
`Number of previous regimens for advanced breast cancer§¶
`≤3
`4–5
`>5
`Previous exposure to HER2-directed therapy
`Trastuzumab
`Duration (months)
`Lapatinib
`Duration (months)
`Previously treated asymptomatic brain metastasis
`
`Physician’s choice
`(n=198)
`
`Trastuzumab emtansine
`(n=404)
`
`54 (28–85)
`164 (83%)
`28 (14%)
`6 (3%)
`
`53 (27–89)
`345 (85%)
`46 (11%)
`13 (3%)
`
`48 (24%)
`85 (43%)
`65 (33%)
`
`161 (81%)
`24 (12%)
`13 (7%)
`
`82 (41%)
`101 (51%)
`15 (8%)
`
`103 (52%)
`85 (43%)
`10 (5%)
`150 (76%)
`
`187 (94%)
`11 (6%)
`163 (82%)
`4 (1–19)
`78 (39%)
`65 (33%)
`55 (28%)
`
`99 (25%)
`171 (42%)
`134 (33%)
`
`325 (80%)
`57 (14%)
`22 (5%)
`
`180 (45%)
`200 (50%)
`22 (5%)
`
`208 (51%)
`185 (46%)
`11 (3%)
`302 (75%)
`
`391 (97%)
`13 (3%)
`345 (85%)
`4 (1–14)
`131 (33%)
`149 (37%)
`122 (30%)
`
`198 (100%)
`23·7 (0·7–508·8)
`198 (100%)
`7·62 (0·1–48·0)
`27 (14%)
`
`404 (100%)
`24·3 (1·4–140·5)
`404 (100%)
`7·98 (0·1–71·2)
`40 (10%)
`
`that a sample size of about 600 patients would provide 80%
`power to detect an HR of 0·65 for PFS (a 54% improvement
`in median PFS from 4 months in the physician’s choice
`group to 6·15 months in the trastuzumab emtansine
`group) at a two-sided signifi cance level of 0·5% and an
`HR of 0·76 for overall survival (a 32% improvement in
`median overall survival from 12 months in the physician’s
`choice group to 15·8 months in the trastuzumab
`emtansine group) at a two-sided signifi cance
`level
`of 4·5%.15,19 The primary PFS analysis was to be done when
`about 324 PFS events had occurred and only after all
`patients had enrolled and had the opportunity for at least
`one post-baseline tumour assessment. We planned two
`formal interim overall survival analyses (to be done at the
`time of the primary PFS analysis and at about 330 deaths,
`respectively) and one fi nal overall survival analysis (at
`about 492 deaths). The overall type I error was to be
`controlled at 0·045 for the formal overall survival interim
`analyses and fi nal overall survival analysis using the Lan-
`DeMets alpha-spending function with an O’Brien-
`Fleming boundary. The boundaries used at each interim
`and fi nal overall survival analysis depend on the actual
`number of observed deaths at each analysis. If either PFS
`or overall survival were statistically signifi cant at any
`analysis, the secondary endpoints were to be tested in a
`prespecifi ed order.
`For PFS and duration of objective response, we
`censored patients who had neither disease progression
`nor death at the date of the last tumour assessment in
`which an overall response other than unknown or
`unevaluable was recorded on or before the cutoff date.
`
`Treatment category
`Single-agent trastuzumab emtansine
`Combination with HER2-directed agent
`Trastuzumab plus chemotherapy
`Trastuzumab plus lapatinib
`Trastuzumab plus hormonal therapy
`Lapatinib plus chemotherapy
`Single-agent chemotherapy
`Chemotherapy agents†
`Vinorelbine
`Gemcitabine
`Eribulin
`Paclitaxel
`Docetaxel
`Other
`
`Physician’s choice
`(n=185)
`
`1 (<1%)*
`153 (83%)
`126 (68%)
`19 (10%)
`3 (2%)
`5 (3%)
`31 (17%)
`
`59 (32%)
`29 (16%)
`16 (9%)
`16 (9%)
`10 (5%)
`32 (17%)
`
`Data are median (range) or number (%). ECOG PS=Eastern Cooperative Oncology Group performance status.
`ER=oestrogen receptor. PR=progesterone receptor. *Includes multiracial patients. †Two patients in the trastuzumab
`emtansine group had missing ECOG PS scores; proportions are calculated out of a population of 402 patients. ‡At
`initial diagnosis of breast cancer. §Excluding hormonal treatment. ¶Two patients in the trastuzumab emtansine group
`had missing information; proportions are calculated out of a population of 402 patients.
`
`Table 1: Demographic and disease characteristics at baseline
`
`Data are number (%). Further details can be found in the appendix. *One patient
`randomised to the physician’s choice group (whose planned treatment was
`trastuzumab plus gemcitabine) received two cycles of trastuzumab emtansine by
`mistake. †With or without HER2-directed therapy.
`
`Table 2: Type of treatment in patients who received treatment of
`physician’s choice
`
`692
`
`www.thelancet.com/oncology Vol 15 June 2014
`
`IMMUNOGEN 2012, pg. 4
`Phigenix v. Immuogen
`IPR2014-00676
`
`

`
`Articles
`
`Physician’s choice
`(n=198)
`
` 3·3 (2·89–4·14)
`
`Trastuzumab
`emtansine
`(n=404)
`6·2 (5·59–6·87)
`
`129
`
`219
`
`Median PFS
`(95% CI), months
`Events
`
`Stratified HR 0·528 (95% CI 0·422–0·661); p<0·0001
`Unstratified HR* 0·521 (95% CI 0·418–0·648); p<0·0001
`
`Physician’s choice
`Trastuzumab emtasine
`
`A
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`Progression-free survival (%)
`
`0 0
`
`120
`
`334
`
`62
`
`241
`
`28
`
`114
`
`13
`
`66
`
`6
`
`27
`
`1
`
`12
`
`Number at risk
`Physician’s
`choice
`Trastuzumab
`emtansine
`
`198
`
`404
`
`Physician’s choice
`(containing
`trastuzumab)
`(n=149)
`
`Trastuzumab
`emtansine
`(n=404)
`
`
`
` 3·2 (2·83–4·11)
`
`6·2 (5·59–6·87)
`
`101
`
`219
`
`Median PFS
`(95% CI), months
`Events
`
`Stratified HR 0·558 (95% CI 0·437–0·711); p<0·0001
`Unstratified HR* 0·544 (95% CI 0·429–0·690); p<0·0001
`
`12
`
`5
`
`1
`
`0
`
`B
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`Progression-free survival (%)
`
`For the analysis of overall survival, we censored patients
`who were alive at the time of data cutoff at the last date
`they were known to be alive on or before the cutoff date.
`Patients with no post-baseline information were censored
`at the date of randomisation plus 1 day. For the analysis
`of overall response, we regarded patients with measurable
`disease at baseline who had no post-baseline record of
`tumour assessment as non-responders.
`We estimated median time-to-event outcomes and
`corresponding 95% CIs for each treatment group using
`Kaplan-Meier methods. We used the two-sided log-rank
`test, stratifi ed by the protocol-defi ned randomisation
`factors, to compare time-to-event outcomes between
`treatment groups. The unstratifi ed log-rank test was done
`as a sensitivity analysis. HRs and cor responding 95% CIs
`were estimated using Cox proportional hazards models,
`stratifi ed by the protocol-defi ned randomisation factors.
`Overall response data were compared between treatment
`groups using a stratifi ed Mantel-Haenszel χ² test.
`Statistical analyses were done with SAS (version 9.2).
`This study is registered with ClinicalTrials.gov, number
`NCT01419197.
`
`Number at risk
`Physician’s
`choice
`(containing
`trastuzumab)
`Trastuzumab
`emtansine
`
`149
`
`99
`
`50
`
`20
`
`404
`
`334
`
`241
`
`114
`
`66
`
`27
`
`12
`
`0
`
`Physician’s choice
`(no trastuzumab)
`(n=49)
`
` 3·4 (1·61–4·63)
`
`Trastuzumab
`emtansine
`(n=404)
`6·2 (5·59–6·87)
`
`28
`
`219
`
`Median PFS
`(95% CI), months
`Events
`
`Stratified HR 0·428 (95% CI 0·283–0·648); p<0·0001
`Unstratified HR* 0·438 (95% CI 0·295–0·651); p<0·0001
`
`C
`100
`
`80
`
`60
`
`40
`
`20
`
`Progression-free survival (%)
`
`Role of funding source
`The TH3RESA study was designed by the funder,
`Genentech, in collaboration with the study steering
`committee. Two non-Roche
`steering committee
`members and authors of this Article, IEK and PML,
`reviewed and approved the statistical analysis plan
`before fi nalisation of the original protocol. All steering
`committee members discussed and agreed to any
`protocol amendments,
`including changes
`to
`the
`statistical analysis after data from the EMILIA study
`became available. Employees of the funder managed
`the data and did the statistical analyses. Steering
`committee members reviewed the tables, listings, and
`graphs during the development of this manuscript and
`could have had access to the primary database, if
`requested. The Article’s senior author, HW, provided
`his sign-off on the clinical study report. Moreover, the
`study’s independent data monitoring committee, which
`provided external oversight, had access to all primary
`data throughout the course of the trial. All authors were
`involved in data analysis and interpretation, manuscript
`writing, and fi nal approval of the manuscript. The
`manuscript was also reviewed by the funder. The
`corresponding author had full access to all the data and
`had fi nal responsibility for the decision to submit for
`publication.
`
`14
`
`0 0
`
`693
`
`Figure 2: Kaplan-Meier curves of progression-free survival
`Probability of progression-free survival in all randomised patients in each
`treatment group (A), in all randomised patients in the trastuzumab emtansine
`group and those in the physician’s choice group who received a trastuzumab-
`containing regimen as study medication (B), and in all randomised patients in the
`trastuzumab emtansine group and those in the physician’s choice group who
`received a study medication regimen that did not contain trastuzumab (C).
`HR=hazard ratio. PFS=progression-free survival. *Provided as a sensitivity analysis.
`
`0
`
`0
`
`49
`
`404
`
`Number at risk
`Physician’s
`choice (no
`trastuzumab)
`Trastuzumab
`emtansine
`
`2
`
`21
`
`4
`
`12
`
`334
`
`241
`
`6
`8
`Months since randomisation
`
`8
`
`114
`
`1
`
`66
`
`10
`
`1
`
`27
`
`12
`
`0
`
`12
`
`www.thelancet.com/oncology Vol 15 June 2014
`
`IMMUNOGEN 2012, pg. 5
`Phigenix v. Immuogen
`IPR2014-00676
`
`

`
`Articles
`
`Results
`From Sept 14, 2011, to Nov 19, 2012, 602 patients were
`enrolled from 146 centres in 22 countries (appendix).
`A data cutoff date of Feb 11, 2013, was used for all reported
`analyses. HER2 status was established centrally using
`primary tumour samples in 420 patients and metastatic
`tumour samples in 88 patients. In 94 patients, the
`pathologist was not able to establish whether the tissue
`was derived from the primary tumour or a metastasis. Of
`the enrolled patients, 404 were randomly assigned to
`treatment with trastuzumab emtansine and 198 to
`physician’s choice; these patients make up the intention-
`to-treat population for effi cacy analyses (fi gure 1).
`Baseline demographic and disease characteristics were
`well balanced between treatment groups (table 1).
`Patients had received a median of four previous regimens
`(excluding single-agent hormonal therapy) for advanced
`breast cancer, three-quarters of patients had visceral
`
`disease, and just over half had hormone-receptor-positive
`breast cancer (table 1). One patient in the physician’s
`choice group received
`trastuzumab emtansine by
`mistake, and was therefore included in the trastuzumab
`emtansine group for safety analyses; the safety population
`consisted of 403 patients in the trastuzumab emtansine
`group and 184 patients in the physician’s choice group.
`Treatment with physician’s choice consisted of
`combination regimens with one or more HER2-directed
`agents in 153 (83%) of 184 patients, and single-agent
`chemotherapy in 31 (17%) patients (table 2; appendix).
`HER2-directed
`regimens
`typically contained
`tras-
`tuzumab, which was administered
`to 148
`(80%)
`of 184 patients who received their physician’s choice of
`treatment. Only 36 (9%) of 403 patients required dose
`reductions of trastuzumab emtansine, and a reduction
`by one dose level (to 3·0 mg/kg) was suffi cient for most
`(30 [83%]) of these patients. The median number of
`
`All patients, N
`
`Physician’s choice
`
`Trastuzumab emtansine
`
`HR* (95% CI)
`
`602
`
`509
`74
`19
`
`147
`256
`199
`
`488
`81
`35
`
`262
`301
`37
`
`n
`
`Event Median PFS,
`months
`
`n
`
`Event Median PFS,
`months
`
`198
`
`129
`
`3·3
`
`404
`
`219
`
`6·2
`
`0·52 (0·42–0·65)
`
`164
`28
`6
`
`108
`17
`4
`
`48
`85
`65
`
`161
`24
`13
`
`82
`101
`15
`
`103
`85
`10
`
`150
`48
`
`78
`120
`
`27
`171
`
`24
`61
`44
`
`104
`17
`8
`
`48
`68
`13
`
`66
`58
`5
`
`95
`34
`
`49
`80
`
`16
`113
`
`3·4
`3·2
`3·0
`
`4·1
`3·2
`3·1
`
`3·4
`2·8
`3·3
`
`3·6
`3·1
`1·6
`
`3·9
`2·9
`3·9
`
`3·4
`3·1
`
`3·3
`3·4
`
`2·9
`3·6
`
`345
`46
`13
`
`99
`171
`134
`
`325
`57
`22
`
`180
`200
`22
`
`208
`185
`11
`
`302
`102
`
`131
`271
`
`40
`364
`
`191
`25
`3
`
`58
`91
`70
`
`177
`30
`12
`
`84
`120
`13
`
`109
`105
`5
`
`168
`51
`
`60
`158
`
`24
`195
`
`5·8
`6·9
`NE
`
`5·8
`6·9
`5·8
`
`6·3
`5·4
`6·6
`
`7·0
`5·4
`6·9
`
`5·9
`6·0
`8·3
`
`6·2
`6·7
`
`6·9
`5·8
`
`5·8
`6·2
`
`0·55 (0·44–0·70)
`0·42 (0·22–0·80)
`0·14 (0·02–0·79)
`
`0·71 (0·44–1·14)
`0·44 (0·32–0·61)
`0·53 (0·36–0·78)
`
`0·50 (0·39–0·64)
`0·63 (0·35–1·14)
`0·57 (0·23–1·41)
`
`0·44 (0·31–0·64)
`0·63 (0·47–0·85)
`0·41 (0·19–0·92)
`
`0·56 (0·41–0·76)
`0·51 (0·37–0·71)
`0·17 (0·03–0·93)
`
`0·56 (0·44–0·72)
`0·41 (0·26–0·64)
`
`0·48 (0·32–0·70)
`0·55 (0·42–0·72)
`
`0·47 (0·24–0·89)
`0·53 (0·42–0·66)
`
`All patients
`Age group
`<65 years
`65–74 years
`≥75 years
`World region
`USA
`Western Europe
`Other
`Race
`White
`Asian
`Other
`ECOG PS
`
`0 1 2 E
`
`311
`270
`21
`
`R and PR status†
`ER+ and/or PR+
`ER– and PR–
`Unknown
`Disease involvement
`452
`Visceral
`150
`Non-visceral
`Number of previous regimens for advanced disease‡
`209
`≤3
`391
`>3
`Previously treated asymptomatic brain metastasis
`Yes
`67
`No
`535
`
`Figure 3: Subgroup analyses of progression-free survival
`The subgroups are based on data obtained at baseline. ECOG PS=Eastern Cooperative Oncology Group performance status. ER=oestrogen receptor. HR=hazard ratio.
`NE=not estimable. PR=progesterone receptor. *Unstratifi ed HR. †Measured at initial diagnosis of breast cancer. ‡Excluding hormonal treatment.
`
`0·2
`
`0·5
`
`1
`
`2
`
`Favours trastuzumab emtansine
`
`Favours physician’s choice
`
`694
`
`www.thelancet.com/oncology Vol 15 June 2014
`
`IMMUNOGEN 2012, pg. 6
`Phigenix v. Immuogen
`IPR2014-00676
`
`

`
`Articles
`
`than 0·0000016. Due to several planned assessments of
`overall survival over time, a p value of this stringency is
`necessary to control the type I error rate. Estimated 6-month
`survival (90·9% [95% CI 87·8–94·0] in the trastuzumab
`emtansine group vs 78·3% [71·5–85·2] in the physician’s
`choice
`group)
`and
`1-year
`survival
`(68·6%
`[59·9–77·3] vs 56·9% [42·2–71·6]) were numerically
`higher in the trastuzumab emtansine group than in the
`physician’s choice group.
`In patients with measurable disease at baseline,
`108 (31%) of 345 patients in the trastuzumab emtansine
`group achieved an objective response compared
`with 14 (9%) of 163 in the physician’s choice group
`(diff erence 22·7% [95% CI 16·2–29·2]; p<0·0001). The
`median duration of
`response was 9·7 months
`(95% CI 6·6–10·5) in the trastuzumab emtansine group,
`but it had not been reached at the data cutoff in
`the 14 patients with an objective response in the
`physician’s choice group.
`
`Physician’s choice
`Trastuzumab emtansine
`
`
`Physician’s
`Trastuzumab
`choice
`emtansine
`
`(