Summary Basis of Decision for Xalkori ™

The Canadian regulatory decision on the clinical pharmacology, efficacy, and safety of Xalkori was based on a critical assessment of the Canadian application/submission. The foreign reviews completed by the United States FDA (March 30, 2011) were used as an added reference in the assessment Xalkori.

In parallel with the review of this submission, the Medical Device Bureau, Health Canada assessed and provided a positive regulatory decision of a companion diagnostic kit to identify patients with ALK-positive disease. This is reflected in the indication wording within the Product Monograph.

The clinical efficacy claim for Xalkori as monotherapy to treat patients with ALK-positive advanced NSCLC with or without brain metastases is supported based on a preliminary report of data from the pivotal expanded Phase I Study A8081001 (herein referred to as Study A1001) and supportive Phase II Study A8081005 (herein referred to as Study A1005). Safety data were also included from a supportive Phase III Study A8081007 (herein referred to as A1007). This study is ongoing. Studies A1001 and A1005 also provided pharmacodynamic, pharmacokinetic, and safety data.

Studies A1001 and A1005 are ongoing, open-label, single-arm, multicentre studies with 116 and 136 enrolled patients, respectively. Prior to enrolling in Study A1001, patients were required to have ALK-positive tumours identified using a number of local clinical trial assays. The median duration of treatment for patients in Study A1001 was 32 weeks. Patients in Study A1005 were required to have received at least one prior treatment regimen and harbour ALK-positive tumours prior to entering the study. ALK-positive NSCLC was identified using a Health Canada-approved Vysis ALK break-apart fluorescence in situ hybridization (FISH) assay. The median duration of treatment for Study A1005 was 22 weeks. For both studies, all patients received prior systemic therapy, with the exception of 16 patients (13%) in Study A1001 who had no prior systemic treatment for locally advanced or metastatic disease. Each patient received the recommended dose of 250 mg Xalkori taken orally twice per day.

Study A1007 is an ongoing Phase III, open-label, two-arm, randomized, multicentre, multinational study comparing crizotinib to standard of care chemotherapy (permetrexed or docetaxel) in patients with previously treated ALK-positive advanced NSCLS. In this study ALK-positive NSCLC patients were identified using a validated break-apart FISH assay kit. At the time of this review, 71 patients with ALK-positive NSCLC had received crizotinib treatment.

3.3.1 Pharmacodynamics

At the recommended therapeutic dose of 250 mg twice daily, crizotinib had a pronounced negative chronotropic effect in patients with ALK-positive NSCLC. In Study A1001, there was a clinically and statistically significant mean heart-rate reduction at steady-state of -10 to -14 beats per minute (bpm). In Study A1005, the mean heart rate reduction at steady-state was -15 to -16 bpm. A mechanistic explanation for this effect could be the antagonistic effects of the drug on L-type calcium channels. A negative chronotropic effect raises concerns about bradycardia and bradyarrhythmias.

At a dose of 250 mg twice daily, crizotinib caused QTc interval prolongation according to the Fridericia cube root heart rate correction. In Study A1005, the magnitude of the estimated effect was mean 10.3 msec at 6 h post-dosing on Day 22. QTc prolongation with crizotinib is consistent with its activity as an inhibitor of human Ether-à-go-go Related Gene (hERG) channels. QTc prolongation creates an electrophysiological environment that favours the occurrence of Torsade de Pointes, which can progress to ventricular fibrillation and sudden cardiac death. Concerns regarding Torsade de Pointes are increased when QTc prolongation occurs in association with a reduction in heart rate. Crizotinib also appeared to cause some prolongation of the QRS duration, with a mean 2.5 msec increase from baseline at 6 h post-dosing on Day 22 in Study A1005.

The electrocardiogram (ECG) data from studies A1001 and A1005 were read using an automated reading method. Automated ECG reading methods are considered to be potentially unreliable when performing measurements of low amplitude waveforms, such as T waves and P waves. Erroneous automated readings are of particular concern in a patient population in which many subjects may have abnormal ECG waveforms because of the underlying disease state, co-morbidities, and concomitant medications. It is therefore necessary to consider the possibility that the analyses provided in these studies under-estimated the magnitude of the QTc or PR prolongation produced by crizotinib.

Two unexplained deaths occurred for which an arrhythmic cause cannot be excluded.

Both outlier analyses and analyses of central tendency raise concerns about an association of crizotinib with decreased serum calcium levels. Hypocalcaemia is a risk factor for Torsade de Pointes during treatment with a QTc prolonging drug.

Crizotinib decreases systolic and diastolic blood pressure, effects that seem likely to be related to its blockade of L-type calcium channels and alpha-1 adrenergic receptors. This PD effect raises concerns about orthostatic hypotension and syncope in patients with low blood pressure at baseline. Patients may need adjustment of concomitant antihypertensive therapies when receiving crizotinib.

Many tyrosine kinase inhibitors are associated with cardiomyopathy and decreased ventricular performance which can result in congestive heart failure. The administration of crizotinib 250 mg twice daily was associated with adverse events (AEs) such as edema, dyspnoea, cough, and wheezing that raise concerns about cardiotoxicity. An alternative explanation for the respiratory events would be pulmonary toxicity, as crizotinib has been associated with pneumonitis. In subsequent clinical trials, prospective echocardiography or multi-gated acquisition scans (MUGAs) should be performed for monitoring of ejection fraction.

Finally, treatment-emergent thrombotic events occurred in Studies A1001 and A1005 for which a causal relationship to crizotinib should be considered.

3.3.2 Pharmacokinetics

The single-dose PKs of Xalkori (crizotinib) were evaluated in healthy subjects and in patients with advanced tumours. The multiple-dose PKs of crizotinib were evaluated in patients with advanced solid tumours.

A single-dose, four-period crossover comparative bioavailability study (Study A8081011) in healthy subjects demonstrated that the Xalkori (crizotinib) commercial image 250 mg capsules met the recommended standards for bioequivalence in comparison to the clinical study formulations (immediate release tablets and powder in capsule formulations). In addition, the data from this study support the proposed labelling of the effect of food on crizotinib PKs. A high-fat meal decreased the rate and extent of absorption of crizotinib from the Xalkori capsules by approximately 14% in comparison to fasting conditions.

The maximum tolerated dose was 250 mg twice daily based on dose-limiting toxicities. The lowest crizotinib plasma concentrations (trough concentrations) achieved at steady-state following the 250 mg twice-daily dose exceeded the target efficacious concentrations predicted for ALK inhibition (111 ng/mL total drug).

Absorption

Following repeated 250 mg twice-daily dosing (recommended dose) in patients with advanced solid tumours, crizotinib plasma concentration reached steady-state within 15 days. In general, the systemic exposure appeared to increase in a greater than dose-proportional manner over the dose range of 200 to 300 mg twice daily.

Distribution

Crizotinib showed extensive distribution into tissues from the plasma. Binding of crizotinib to human plasma proteins in vitro was 91% and appeared to be independent of drug concentrations.

Metabolism

In vitro studies demonstrated that CYP3A4 and CYP3A5 were the major enzymes involved in the metabolic clearance of crizotinib. The primary metabolic pathways in humans were oxidation and O-dealkylation. In vitro studies in human microsomes demonstrated that crizotinib was a time-dependent inhibitor of CYP3A.

Following a single dose of 250 mg [¹⁴C]-labeled crizotinib in healthy volunteers, crizotinib accounted for 33% of the total recovered radioactivity in pooled plasma; crizotinib lactam metabolite was the predominant metabolite in plasma and accounted for 10% of the total recovered radioactivity.

Crizotinib lactam was approximately 2.5- and 7.7-fold less potent than crizotinib in inhibiting ALK and c-Met tyrosine kinases, respectively, in vitro. The O-desalkyl crizotinib and O-desalkyl crizotinib lactam were inactive against ALK and c-Met.

Elimination

Following a single 250 mg oral dose in patients with advanced solid tumours, the terminal half-life of crizotinib was 42 hours and the mean apparent clearance was 100 L/hr. At steady state, the clearance appeared to be lower (65 L/h) possibly due to auto-inhibition of CYP3A by crizotinib following repeated dosing.

Following the administration of a single 250 mg radiolabeled crizotinib dose to healthy subjects, 63% and 22% of the administered dose was recovered in faeces and urine, respectively. Unchanged crizotinib represented approximately 53% and 2.3% of the administered dose in faeces and urine, respectively.

Drug-Drug Interactions

CYP3A Inhibitors

Crizotinib is predominantly metabolized by CYP3A. Co-administration of a single 150 mg oral dose of crizotinib in the presence of ketoconazole (200 mg twice daily), a strong CYP3A4 inhibitor, resulted in a 3.2-fold increase in crizotinib systemic exposure compared to that when crizotinib was administered alone. The Product Monograph recommends avoiding concurrent use of strong CYP3A inhibitors. Caution should be exercised when moderate CYP3A inhibitors are co-administered.

CYP3A Inducers

Co-administration of a single 250 mg crizotinib dose with rifampin, a strong CYP3A inducer, resulted in an 82% decrease in crizotinib exposure compared to the administration of crizotinib alone. The Product Monograph recommends avoiding the concurrent use of strong CYP3A inducers.

CYP3A Substrates

Crizotinib is a moderate inhibitor of CYP3A and may increase plasma concentrations of co-administered CYP3A substrates. The Product Monograph recommends that caution be exercised in administering crizotinib in combination with drugs that are predominantly metabolized by CYP3A, particularly those that have narrow therapeutic indices and that are associated with life-threatening arrhythmias.

P-glycoprotein (P-gp) Substrates

Crizotinib is an inhibitor of P-gp in vitro. Consequently, crizotinib may increase plasma concentrations of co-administered drugs that are substrates of P-gp.This is indicated in the Product Monograph.

Gastric Protectant Medications

The aqueous solubility of crizotinib is pH dependent, with high pH (less acidic) resulting in lower solubility. A formal drug-interaction study with an antacid has not been conducted. The use of antacids was permitted during crizotinib treatment in clinical trials.

Other Drug-Drug Interactions

The concomitant use of Xalkori with QT interval-prolonging drugs or with drugs that can disrupt electrolyte levels should be avoided to the extent possible. For further information, please consult the Product Monograph.

Drug-Food Interactions

A high-fat meal reduced crizotinib exposure by approximately 14% when a 250 mg dose was given to healthy volunteers. The Product Monograph recommends that crizotinib be administered with or without food.

Foods and herbal products that may modify CYP3A activity should be avoided. These include grapefruit, pomegranate, Seville oranges, and St. John's wort.

Special Populations and Conditions

Hepatic Impairment

Xalkori has not been studied in patients with hepatic impairment. The clinical studies conducted excluded patients with alanine aminotransferase (ALT) or aspartate aminotransferase (AST) levels greater than 2.5-times the upper limit of normal (ULN) or, if due to underlying malignancy, greater than 5.0-times the ULN, or with total bilirubin greater than 1.5-times the ULN.

Renal Impairment

Xalkori has not been studied in patients with severe renal impairment. No starting dose adjustment is recommended for patients with mild or moderate renal impairment, although data are limited. Clinical data are not sufficient to determine if there is the potential need for starting dose adjustment in patients with severe renal impairment. No data are available for patients with end-stage renal disease.

Ethnicity

After 250 mg twice daily dosing, steady-state crizotinib maximum concentration (C_max) and the area under the curve for the dosing interval (AUC_t) in Asian patients were 1.57- and 1.50-fold those seen in non-Asian patients, respectively. There was a higher incidence of Grade 3 or 4 adverse events (AEs) in non-Asians (17%) than Asians (10%).

The identified risks and precautions related to PK findings have been properly labelled in the PM to facilitate the safe use of the drug.

3.3.3 Clinical Efficacy

The efficacy of Xalkori as monotherapy to treat patients with ALK-positive advanced NSCLC with or without brain metastases has been assessed in the ongoing studies A1001 and A1005. For a description of these studies please refer to the beginning of section 3.3 Clinical Basis for Decision.

In both studies combined, 255 previously treated or untreated ALK-positive advanced or metastatic NSCLC patients received the recommended dose of 250 mg crizotinib twice daily. The primary efficacy endpoint for both studies was Objective Response Rate (ORR) according to Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.0 in Study A1001 and RECIST Version 1.1 in Study A1005. The ORR was 61% in Study A1001 (with two complete responses) and 51% in Study A1005 (with one complete response). These values, although mostly comprised of partial responses, were substantial. Notably, more than 90% of patients in these studies showed some degree of target lesion shrinkage.

The secondary endpoints for both studies included Duration of Response (DR), Time to Tumour Response (TTR), Disease Control Rate (DCR), Progression-Free Survival (PFS), and Overall Survival (OS). The median time to response in both studies was between 6 and 7.7 weeks with a median duration of response of 42 to 48 weeks. There were 31 patients in study A1001 who had stable disease for a DCR at 8 weeks of 79%. Fifty-five percent (55%) of objective tumour responses in this study were achieved during the first 8 weeks of treatment. There were 45 patients in study A1005 who had stable disease for a DCR at 6 weeks of 85%. Seventy-nine percent (79%) of objective tumour responses in this study were achieved during the first 8 weeks of treatment. To date, there are no data demonstrating an OS or PFS benefit in this patient population.

Although these preliminary data are based on single-arm studies with no control arm, the ORR is substantial and may in fact underestimate the true value which is expected from randomized controlled study data. The indication for Xalkori is granted on the basis of the ORR data from the ongoing studies A1001 and A1005. Confirmation of clinical benefit will be assessed based on the two Phase III randomized controlled studies (Study A1007 and A1014). The final Clinical Study Reports (CSRs) from Study A1007 and A1014 are the confirmatory studies listed as post-approval commitments. In addition, the sponsor will also provide the final complete CSRs from Study A1001 and A1005 to Health Canada.

3.3.4 Clinical Safety

The clinical safety of Xalkori was evaluated based on the data provided from the studies described at the beginning of section 3.3 Clinical Basis for Decision.

The major AEs reported included gastrointestinal and visual disturbance predominantly mild to moderate in severity. Serious adverse events (SAEs) included elevated transaminase levels, QT prolongation, interstitial lung disease, neutropaenia, and lymphopaenia. These SAEs were typically manageable through dose modification. Nevertheless, deaths did occur including those from pneumonitis, interstitial lung disease, cardiac arrest, and hepatic failure. Hy's Law cases were reported in patients treated with Xalkori. The proportion of patients with AEs differed slightly with age with more AEs noted for patients over the age of 65. As well, more AEs were reported in females versus (vs.) males and in Asian vs. non-Asian patients. In general, results to date suggest that crizotinib 250 mg, twice daily, is generally well tolerated and has a manageable safety and laboratory profile.

Although the number of patients in the safety database is adequate, the proportion of patients with ALK-positive NSCLC is small. In addition, the safety update for Study A1001 did not provide information on Grade 1-4 AEs in these patients, but only provided information on SAEs and deaths. Data from patients who were not ALK-positive and who received Xalkori are useful in helping to further evaluate the safety profile of Xalkori. Of note is that many of these patients, including those in the Dose Escalation Cohort and those with Xalkori-refractory tumours, had a short duration of exposure to Xalkori. Data with regard to these patients may not sufficiently reflect the AEs associated with crizotinib. In addition, the absence of randomized controlled trials in this submission makes it difficult to distinguish AEs due to disease progression and those due to the use of Xalkori. Two Phase III randomized studies (Study A1007 and Study A1014) are expected to allow for a comparison of safety and efficacy between Xalkori-treated and comparator-drug treated patient groups. Health Canada has requested the final CSRs for these Phase III studies as post-approval commitments.

Three major safety issues were identified in this submission:

• QT interval prolongation and bradycardia;
• Hepatotoxicity, including fatal outcomes; and
• Pneumonitis, including fatal outcomes.

These issues have been addressed through appropriate labelling (a Serious Warnings and Precautions box) in the Product Monograph. Other important safety risks reported in studies with Xalkori are managed through strict labelling. The labelling also highlights the importance of the requirement to utilize laboratories with demonstrated proficiency in using a validated diagnostic assay to assess ALK fusion, to avoid inappropriate treatment in ALK-negative patients for whom the benefit of Xalkori is not established. For more information about QT interval prolongation and bradycardia, see section 3.3.1 Pharmacodynamics.

No data exist in patients with hepatic or severe renal impairment. There is an ongoing multiple-dose study to determine the appropriate dose of Xalkori in patients with various levels of severity of hepatic impairment. The anticipated study completion date is July 2013 and the final CSR is expected by January 2014. Although not part of the post-approval commitments, the Sponsor has agreed to submit the final CSR to the Health Canada to evaluate the risk of Xalkori treatment in patients with hepatic impairment.

There is also an ongoing clinical trial to determine the appropriate dose of Xalkori in patients with severe renal impairment. The anticipated study completion date is April 2012 and the final CSR is expected by October 2012. Although not part of the post-approval commitments, the Sponsor agreed to submit the final CSR to the Bureau to evaluate the risk of Xalkori treatment in patients with severe renal impairment.

The sponsor has agreed to provide final CSRs to further evaluate the possible serious safety risks associated with the use of Xalkori. A list of these reports is found in section 3.3.5 Additional Issues.

3.3.5 Additional Issues

In keeping with the provisions outlined in the Notice of Compliance with Conditions (NOC/c) Guidance, the sponsor is required to submit the results of the following confirmatory studies:

• Study A1007 is an ongoing Phase III, randomized, open-label study of the clinical efficacy and safety of crizotinib vs. standard of care (pemetrexed or docetaxel) in patients with advanced NSCLC harbouring a translocation or inversion event involving the ALK gene locus. The anticipated completion date of the study is December 2013. Submission of the final CSR is expected by June 2014.
• Study A1014 is an ongoing Phase III, randomized, open-label study of the efficacy and safety of crizotinib vs. pemetrexed/cisplatin or pemetrexed/carboplatin in previously untreated patients with non-squamous carcinoma of the lung harbouring a translocation or inversion event involving the ALK gene locus. To evaluate the clinical benefit of crizotinib as a first-line therapy, the sponsor will submit the final CSR from this study. The anticipated completion date of the study is December 2015. Submission of the final CSR is expected by June 2016.

In addition, to evaluate the efficacy and safety of the use of crizotinib, the sponsor will submit the final exposure-response analysis for PFS, response rate, OS, and safety endpoints, utilizing data from each study. These data are important given that the efficacy endpoints, PFS, response rate, and OS are outcomes that Health Canada deems to be sufficient to grant a Notice of Compliance (NOC) for the proposed indication, along with a continued favourable benefit/risk profile for crizotinib.

The sponsor has also agreed to Health Canada's request to submit the following:

• The final CSR from Study A1001 and Study A1005;
• Final study reports to further evaluate the serious risk of drug interactions (including CYP2B and CYP2C enzymes, CYP3A inhibitors, and CYP3A inducers);
• A final CSR to evaluate the serious risk of visual impairment;
• A final CSR to evaluate the serious risk of QTc interval prolongation;
• Pending results for echocardiography/MUGA. The high incidence of oedema and dyspnea raise concerns about decreased left ventricular ejection fraction (LVEF) and heart failure. These data will be provided in the final CSRs for Study A1001, A1005, and A1007. The sponsor should incorporate troponin and B-type natriuretic peptide as clinical laboratory biomarkers in subsequent clinical studies, in addition to MUGA or echocardiography monitoring for the measurement of ejection fraction;
• A final CSR to evaluate the risk of crizotinib treatment in patients co-administered pH elevating agents;
• A final CSR (as part of the CSR for Study A1001) to evaluate the efficacy and safety of the use of crizotinib in patients with ALK-negative tumours; and
• As part of future Periodic Safety Update Reports (PSURs) for products authorized under the Notice of Compliance with Conditions Guidance (PSUR/Cs), data with regard to crizotinib resistance.

These items are not listed as post-approval commitments.

In addition, as part of the market authorization, Health Canada has requested a Risk Management Plan (RMP) to monitor whether the benefits continue to outweigh any risk. Health Canada has requested the following post-marketing information from the sponsor (and agreed to by the sponsor) which addresses the following issues raised by Health Canada and which are not included as post-approval commitments in the Letter of Undertaking. In future RMP submissions, the sponsor should address the following issues:

• Information pertaining to the conception, specificity, sensitivity, and availability of companion tests;
• A discussion as to what extent a companion test or other available tests could generate false negative/positive results and the strategy adopted to minimize the impact of these false negative/positive results;
• The sponsor should provide an estimation of mean treatment duration;
• Data from future clinical trials and/or post marketed surveillance activities should be collected and analysed to evaluate the effect of crizotinib on the elderly population;
• The rationale supporting the decision to include vision disorders and QT prolongation as "potential" risks rather than "identified" risks associated with the use of crizotinib should be more explicitly developed;
• The sponsor should update the RMP with more details from non-clinical studies related to cardiotoxicity and QTc prolongation in order to better appraise the cardiotoxicity associated with crizotinib. Those results should identify whether crizotinib's metabolites conserved their inhibitory effect on hERG and Ca²⁺ channels;
• The sponsor should provide a more in depth discussion of the inhibitory effects of crizotinib on 5-HT4 receptors (along with the other 5-HT receptors) and on the dopamine active transporter (DAT) and its possible role in bradycardia observed in patients using crizotinib;
• The sponsor should discuss the implication of the inhibitory effects of crizotinib on 5-HT4 receptors and DAT on the observed neuropathy associated with the use of crizotinib. The sponsor should also provide a more in depth discussion of the data ascertaining that crizotinib does not cross the blood-brain barrier;
• Considering that several clinical studies are still ongoing and that long-term effects of crizotinib on patients are still unknown, Health Canada recommends that the sponsor submit PSURs annually for the next 3 years (which Health Canada will review) following the Notice of Compliance in order to monitor adverse events associated with the cardiac, respiratory and hepatic functions; and
• Updates on the ongoing clinical studies can be summarized in the PSUR, or presented as a separate report if warranted.

All the identified and potential risks addressed in the RMP should be discussed separately in each PSUR report.