The studies from the palbociclib development program providing evaluable safety data in support of the proposed usages are Studies 1023, 1003, 1008 (very limited), 1010, 1001 with additional information in a 90-day safety update from Study 1027 (not evaluable) and Study 1034 (very limited), but no CSR for either.
Palbociclib and fulvestrant combination
The CSR from Study 1023 with a safety update provided in the 90-day safety update; the latter updates the information from the CSR and therefore has been used by the evaluator.
Comment: As previously noted, the PI does not include information from this 90-day safety update and an updated PI needs to be provided by the sponsor .
Palbociclib and letrozole combination
This includes data from a Phase I/II randomised, open label, controlled study, and 1 Phase I/II study in Japanese patients:
‘top-line summary’ data for Study 1008;
Full CSR for Study 1003 (29 November 2013) with an update in the Module 2 SCS (2 January 2015), and then limited further update in the 90-day safety update (31 July 2015) of ‘Selected cumulative safety data summarised in this section include information on deaths and other SAEs along with patient disposition, geographic region, as well as baseline demographics and ECOG PS reported as of 31 July 2015 in Studies 1003, 1034, 1008, 1027’ (90-day safety update)
1010 Ph1P2 and Phase II, 1013, and IIR studies.
Study 1010 Phase I Part 2 (Ph1P2) and Phase II in Japanese patients
safety update and ‘other serious adverse events narratives’ draft CIOMS for Study 1027 which are blinded as to treatment allocation and therefore not evaluable;
Study 1034 (expanded access program), no CSR or discussion other than in safety update within 90-day safety update but limited presentation of CIOMS;
No Integrated safety summary for this proposed usage has been provided incorporating data from Studies 1003 and 1008;
Summary of Clinical Safety with a report date of 26 October 2015, cut-off date of 2 January 2015 for Study 1003 and 5 December 2014 for Study 1023, which integrates the safety from a range of studies including those using palbociclib monotherapy. This has largely been updated for Study 1023 and partly updated for Study 1003 by the 90-day safety update, and contains no safety information about Study 1008 and the top-line summary for Study 1008.
No integrated safety summary has been provided for the letrozole and palbociclib safety data (Clinical Question).
Other studies providing safety data
25 Phase I, II and III investigator-initiated research (IIR) studies in a range of solid tumours, in combination with a range of other treatments or as monotherapy; 17 of these are included in the Summary of Clinical Safety and 25 in the 90-day safety update.
Overall, safety data on deaths and other SAEs were summarised in the 90-day safety update for a total of 1028 patients participating in these 25 IIR studies, in which palbociclib was given either alone in patients with malignant solid tumours, including breast cancer, or as a component of combination therapy in patients with breast cancer.
Comment: The evaluator proposes to evaluate the randomised controlled data for each proposed indication separately, given the differing profiles of the co-administered treatments. Non-randomised data from the combination and from monotherapy will be evaluated for additional signals. Where palbociclib is used in combination with another treatment other than an aromatase inhibitor or fulvestrant (for example, chemotherapy) this will not be considered to provide relevant safety information regarding the proposed usage in this application. Similarly, where palbociclib is used as monotherapy in other solid tumours, any safety issues will be considered and interpreted in light of known adverse events associated with the underlying malignancy.
Adverse event reporting
Study protocols were available for Studies 1003, 1023 and 1008 and the reporting of the adverse events and abnormal test findings are based on these definitions:
An adverse event is:
‘An AE is any untoward medical occurrence in a clinical investigation subject administered a product or medical device; the event need not necessarily have a causal relationship with the treatment or usage.’
‘Abnormal test findings
The criteria for determining whether an abnormal objective test finding should be reported as an AE are as follows:
Test result is associated with accompanying symptoms, and/or
Test result requires additional diagnostic testing or medical/surgical intervention, and/or
Test result leads to a change in study dosing or discontinuation from the study, significant additional concomitant drug treatment, or other therapy, and/or
Test result is considered to be an AE by the investigator or sponsor.’
Merely repeating an abnormal test, in the absence of any of the above conditions, does not constitute an AE. Any abnormal test result that is determined to be an error does not require reporting as an AE.’
In the palbociclib development program, AEs were reported in accordance with FDA and ICH guidances. Uncertainty in AE data could still be present (as for any trial, despite following standard guidances), because, for example:
Reporting adverse event only if the patient is symptomatic or action is required could lead to underreporting of AEs, particularly for abnormal laboratory testing or diagnostic tests which may be significant but are frequently asymptomatic.
Relying on the investigator to nominate AEs by making attributions about a new medicine (that is, identifying the event or abnormal measurement as relevant, significant and/or related and therefore, recognizing that action is required) could lead to underreporting of both clinical and laboratory AEs.
There can be inter-investigator variability in reporting of outcomes, as well as potentially bias where a study is open label (Study 1003) and within blinded studies where distinct toxicity profiles could lead to assumptions about treatment allocation, effectively unblinding treatment(palbociclib causes significant neutropenia).
Pivotal studies that assessed safety as the sole primary outcome
No pivotal studies were provided that assessed safety as the sole primary outcome.
The Phase III Study 1008 is considered the pivotal study for the proposed palbociclib and letrozole combination, and the Phase I/II Study 1003 is considered supportive due to smaller numbers and significant methodological issues (see Efficacy section)
Study 1023 is the pivotal study for the proposed combination of palbociclib and fulvestrant. No other studies examined this combination for the proposed usage.
Phase I/II Study 1003 in patients with advanced breast cancer receiving palbociclib in combination with letrozole or letrozole alone.
Phase I/II Study 1010 in Japanese patients with advanced malignant solid tumours, including breast cancer, receiving palbociclib alone (phase 1 part 1) or in combination with letrozole (Phase I part 2 and Phase II).
Studies with evaluable safety data: dose finding and pharmacology
Phase I Study 1001 A Phase I Clinical, Pharmacokinetic, and Pharmacodynamic Evaluation of 2 Schedules of Oral PD 0332991, a Cyclin-Dependent Kinase Inhibitor, in Patients With Advanced Cancer
Most early dose finding studies are of palbociclib monotherapy (for example, Study 1001, Study 1010 Phase I Part 1) and offer limited safety information for the proposed usage in combination. Together with other monotherapy PK or PD studies, these will be evaluated for additional safety signals.
Studies evaluable for safety only
A small postmarketing study, Study 1034 provides limited safety information from the 90-day safety update and the limited narratives/CIOMS provided. No CSR is provided.
A safety update from Study 1027, a randomised Phase III study in Asian subjects, is included together with some CIOMS (marked as ‘draft’) which are blinded to treatment allocation; no CSR is provided. These data are as such, not evaluable.
Studies that assessed safety as the sole primary outcome
No studies were provided that assessed the safety as the sole primary outcome for the proposed usages.
Patient exposure (taken from 90-day safety update for Study 1023, Top-line summary for Study 1008)
Of 1661 patients with solid tumours identified in the 90-day safety update, 1160 women (69.8%) had advanced or metastatic breast cancer and received at least 1 dose of palbociclib 125 mg QD on Schedule 3/1 in combination with endocrine therapy (either letrozole, or fulvestrant +/- goserelin)
835 patients (70.8%) received palbociclib plus letrozole of whom
95 participated in completed open-label, randomised Study 1003;
12 of the 95 patients in Study 1003 initially received palbociclib alone on Schedule 2/1 in Cycle 1 of the Phase I part of the study then switched to the Schedule 3/1 in combination with letrozole taken continuously starting with Cycle 2;
48 participated/are participating in Study 1010 in Japanese patients
6 patients in the completed Ph1P2 portion
32 patients in the ongoing Phase II portion of the study);
444 (based on a 2:1 patient randomisation ratio for palbociclib versus placebo [666 × 2/3 = 444]) are participating in ongoing randomised, double-blind Phase III Study 1008;
20 patients in Study 1027, randomised 1:1, and still blinded as to whether they were receiving palbociclib or placebo in combination with continuous letrozole;
238 are participating in ongoing open-label Expanded Access Protocol 1034;
345 of the 1015 patients (34.0%) with advanced breast cancer received palbociclib plus fulvestrant with or without goserelin in completed randomised, double-blind Phase III Study 1023.
209/347 patients (60.2%) in the palbociclib plus fulvestrant arm had permanently discontinued
172/174 patients randomised to the placebo plus fulvestrant arm received at least one treatment;
138/174 patients (79.3%) in the placebo plus fulvestrant arm were permanently discontinued;
Hence, 136/347 patients (39.2%) in the palbociclib plus fulvestrant arm and
34/174 patients (19.5%) in the placebo plus fulvestrant arm were ongoing as of 31 July 2015.
Duration of exposure
The median duration of palbociclib treatment was more than 2-fold longer than that of placebo (330 [1 – 596] days and 137 [14 – 611] days, respectively). The median number of days on palbociclib (total number of days on which palbociclib was actually administered) was also more than 2-fold greater than that on placebo treatment (221 [1 – 436] days and 102 [14 – 460] days, respectively). The median relative dose intensity estimated for palbociclib was lower than that for placebo (89.8% [22% - 107%] and 99.5% [69% - 108%], respectively).
The duration of fulvestrant treatment and days on this treatment were greater in the palbociclib plus fulvestrant arm than in the placebo plus fulvestrant arm. In addition, the proportion of patients who had their fulvestrant dosing interrupted was also greater in the palbociclib plus fulvestrant arm than in the placebo plus fulvestrant arm. Of note, the study protocol did not allow for the fulvestrant dose to be reduced, but a single fulvestrant dose could be skipped or dosing delayed because of fulvestrant-related toxicity.
The median durations of fulvestrant treatment in the palbociclib plus fulvestrant arm (341 days) and the placebo plus fulvestrant arm (145 days) were slightly longer than those of palbociclib (330 days) and placebo (137 days) treatments in these arms.
From the main CSR (data cut-off 5 Dec 2014)
In the palbociclib plus fulvestrant arm, 71 pre-menopausal patients are included in the AT population. Only 70 pre-menopausal patients were treated with goserelin; 1 patient was randomised incorrectly to the pre-menopausal stratum in IMPALA while post-menopausal status was confirmed in the CRF.
In the palbociclib plus fulvestrant arm, 70 premenopausal patients received goserelin for a median duration of 183.0 days (range: 28 to 1254 days), and in the placebo plus fulvestrant arm, 36 premenopausal patients received goserelin for a median duration of 166.0 days (range: 28 to 1441 days).
Comment: Ovarian suppression may have commenced prior to and be continued beyond the duration of the study treatment, reflecting the prior and subsequent choices of therapy.
Comment: The increase in duration of treatment for both medicines reflects the longer PFS, while the longer median duration of fulvestrant especially compared with palbociclib in that combination arm most likely reflects fewer dose interruptions required for that medicine due to toxicities. The lower dose intensity reflected the need for treatment interruptions and dose reductions with palbociclib in combination with fulvestrant compared with placebo as would be expected due to placebo having no active ingredient. It does indicate that there is a reasonable level of dose-related toxicities with palbociclib but the duration of treatment reassures that this was manageable with strategies such as dose reduction, interruption and supportive measures.
Study 1008 – top-line data summary as of data cut-off 26 February 2016
As of this cut-off date:
199/444 (44.8%) patients were still receiving palbociclib and letrozole arm:
245/444 (55.2%) had discontinued permanently mostly due to objective progression or relapse (38.5%) but 12.6% were due to AE, global deterioration, or refusal
61/222 patients (27.5%) were still receiving the placebo and letrozole arm:
161/222 (72.5%) had discontinued permanently mostly due to objective progression or relapse (56.8%) with 12.7% were due to AE, global deterioration, or refusal
No data on median duration of exposure, dose intensity, and dose reductions were provided for either arm for this pivotal study.
Study 1003 – Phase I/II study as of data cut-off 31 July 2015
Phase I PK/safety
12 postmenopausal women with ER-positive, HER2-negative advanced breast cancer were assigned to palbociclib plus letrozole treatment of which all received at least 1 treatment.
10 patients were permanently discontinued from treatment, while 2 patients were ongoing as of that data cutoff date as of 31 July 2015;
as of 2 January 2015 cut-off (SCS) – no update in the 90-day safety update
median duration of palbociclib treatment in was approximately 12.3 months (373.5 days [range: 63 days - 2081 days]);
median relative dose intensity for palbociclib was 90.2% (range: 77.7% - 100.3%);
3 patients (25.0%) had their palbociclib dose reduced, 8 patients (66.7%) had their palbociclib dose interrupted, and 11 patients (91.7%) had their treatment cycle delayed.
165 women were randomised in this study:
83/84 postmenopausal women with ER-positive, HER2-negative advanced breast cancer were randomised to palbociclib plus letrozole treatment received at least 1 treatment;
76 patients (90.5%) receiving palbociclib plus letrozole were permanently discontinued from treatment, while 7 patients (8.3%) were ongoing as of July 31 2015;
77/ 81 postmenopausal women with ER-positive, HER2-negative advanced breast cancer were randomised to letrozole alone received at least 1 treatment;
75 patients (92.6%) were permanently discontinued from treatment, while 2 patients (2.5%) were ongoing as of July 31 2015.
As of 2 January 2015 (SCS) – no update in the 90-day safety update
median duration of palbociclib treatment was approximately 13.8 months (421.0 days [range: 7 days 1615 days]);
median relative dose intensity for palbociclib was 94.7% (range: 48.5% - 284.4%);
median duration of letrozole treatment duration in the palbociclib and letrozole arm was approximately 14.1 months (428 days [range: 7 days – 1615 days]).
In the letrozole arm:
median treatment duration in the letrozole alone arm was approximately 7.6 months (231 days [range: 28 days – 1241 days);
median relative dose intensity for letrozole was 100% for the letrozole alone arm (range: 81.5% – 100%).
Phase I part 2 portion
6 postmenopausal Japanese women with ER-positive, HER2-negative advanced breast cancer were assigned to palbociclib plus letrozole treatment of whom all received at least 1 treatment
median age was 62 years (range: 59 years – 76 years). 4 (66.7%) were younger than 65 years of age, 2 patients (33.3%) were 65 years of age or older at baseline.
50.0% had an ECOG PS of 0, 50% ECOG PS of 1.
2 patients were permanently discontinued from treatment, while 4 patients were ongoing as of 31 July 2015.
As of January 2 2015 (SCS) – no update in 90-day safety update
median duration of treatment was approximately 19 months (584.5 days [range: 28 days - 649 days]);
median relative dose intensity was 71.2% (range: 59.3% – 98.6%).
42/43 postmenopausal Japanese women with ER-positive, HER2-negative advanced breast cancer were assigned to palbociclib plus letrozole treatment of whom 42 received at least 1 treatment in the Phase II portion of Study 1010 as of 31 July 2015
median age was 62.5 years (range: 43 years – 84 years). 26 patients (61.9%) were younger than 65 years of age and 16 patients (38.1%) were 65 years of age or older at baseline.
All but 3 patients (92.9%) had an ECOG PS of 0 at baseline
3/43 (7%) permanently discontinued from treatment, while 39 (90.7%) were ongoing as of July 31 2015.
Comment: The 90-day safety update included 10 more patients than the SCS, a more recent summary of deaths and SAEs. No updated median duration of treatment was presented. It is not possible to make comparisons between the safety dataset summaries as new patients have joined, affecting the duration of treatment and also the denominator for adverse events. This study should be submitted with either more mature data or with a single, completely updated safety set to permit evaluation. The evaluator has evaluated the currently presented data for new safety signals but cannot comment further on the safety in this population.
All adverse events (irrespective of relationship to study treatment)
Integrated safety analyses
No integrated safety analyses are provided by the sponsor for studies 1003 and 1008, given the latter is a top line summary rather than a CSR. This is considered essential to provide accurate information for inclusion in the PI, particularly if the sponsor includes any further study updates when providing the CSR for Study 1008. See Clinical Questions.
Main/pivotal studies that assessed safety as the sole primary outcome
Pivotal and/or main efficacy studies
Combination of palbociclib and fulvestrant – Study 1023
Treatment-emergent adverse events (TEAEs) were common in both arms: 98.8% (62.3% Grade 3, 13% Grade 4, 1.2% Grade 5) in the palbociclib and fulvestrant arm compared with 90.7% (19.8% Grade 3, 2.3% Grade 4 and 1.7% Grade 5), in the placebo and fulvestrant arm. These are presented in order of decreasing frequency in the table below.
Comment: No updated table summarising the incidence of any AE by whether serious, Grade 3 or 4, grade 5, permanent discontinuations etc was provided for the cut-off July 31 2015. The table for the 5 December 2014 no longer adequately summarises what is known about the outcomes in this study.
The most commonly reported treatment emergent adverse events (TEAEs) in the palbociclib and fulvestrant arm were neutropenia (89%; ‘neutrophil count decreased’, ‘neutropenia’), fatigue (41.2%), nausea (33.9%), and vomiting (18.8%), bone marrow suppression as evidenced by neutropenia, white blood cell decreased (29.3%), anaemia (29%), leukopenia (25.8%), thrombocytopenia (23.1%)(captured in terms: ‘platelet count decreased’ and ‘thrombocytopenia’), headache (25.8%), diarrhoea (23.5%) and alopecia (18%), decreased appetite (15.9%). Gastrointestinal TEAEs of stomatitis (13% - see evaluator comment below) and oropharyngeal pain (12.5%), as well as rash (11 versus 5.2%) were reported more commonly than in the comparator arm.
Grade 3 or 4 events were very common in the palbociclib and fulvestrant arm (75.3%) and were largely made up of Grade 3 or 4 events of neutropenia (69.6%) with thrombocytopenia (0.6%) and WBC count decreased (0.6%). Dyspnoea accounted for the 0.6% and the remaining events accounting for the 3.5% of Grade 3 or 4 TEAEs are not presented (Clinical Questions). The 4 deaths were listed as being due to: disseminated intravascular coagulation (1), disease progression (1), hepatic failure (1), and general physical health deterioration (1).
Table 23: Study A5481023 Summary of all causality, treatment-emergent adverse events experienced by at least 10% of patients in either arm – all treated patients as at 31 July 2015 (Source 90-days safety update)
Of 240 patients experienced a Grade 3 or 4 episode described as neutropenia or neutrophil count decreased, 225 patients had a temporary discontinuation from treatment, 110 had a dose reduction and only 1 is reported as having to discontinue permanently (90-day safety update).
The most commonly reported TEAEs in the placebo and fulvestrant arm were fatigue (29.1%), nausea (27.9%), headache (19.8%), diarrhoea (19.2%), arthralgia (18%), constipation (15.7%), and pain in the extremity (15.1%). Alopecia rates were 6.4%.
Grade 3 or 4 events occurred in 22.1% with 3 deaths reported (1.7%). These adverse events were spread across 12 adverse events classifications with no clear pattern emerging (see Table below). Grade 4 TEAEs experienced by patients in the placebo plus fulvestrant arm were reported for 1 patient (0.6%) each and included white blood count decreased/febrile neutropenia, pathological fracture, cholecystitis, and hypoxia. The deaths from acute respiratory distress syndrome, cerebral haemorrhage and progressive disease, were attributed to the disease under study.
Rates of arthralgia, back pain, hot flushes, dyspnoea, pain, dizziness and injection site pain were similar between the arms.