Integration of Pre-Clinical Safety Assessments in Early Phase Oncology Programs

Gregory Friberg, MD1 and John S. Hill, PhD1,2

1 Early Development Oncology 2 Early Development Clinical Pharmacology,

Medical Sciences, Amgen Inc.

Outline

• Introduction

• Background • Cardiovascular safety considerations in the Oncology setting

• Goals of early phase oncology studies

• Incorporation of cardiovascular safety monitoring

• Clinical examples

2

Drug Development for Grievous Illness is Challenging, and Success Rates Are Low

For Internal Use Only. Amgen Confidential.

Therapeutic Class Number of Projects Probability of Success Oncology & Immunology 6,566 1.8%

Neuroscience 3,817 2.9% Benign Hematology 822 3.8%

GI & Metabolism 2,046 4.5% Pulmonary 1,165 4.8%

Cardiovascular 2,139 4.9% Dermatologic 859 6.6%

Data from >28,000 R&D projects (i.e. patents filed) between 1990 and 2004 (Pharmaceutical Industry Database, IMT Institute)

(Pammolli F, et. al., Nat Rev Drug Disc, 2011)

Costly Late-Phase Program Failures are Rising

For Internal Use Only. Amgen Confidential.

Rigorous, early-phase assessment and prioritization is essential in the setting of finite resources

(patient and financial)

(Pammolli F, et. al., Nat Rev Drug Disc, 2011)

Clinical and Non-clinical Safety Signals in Drug Development

Cardiovascular and liver toxicities are the main safety reason for: • Drug development discontinuations (all stages of discovery and development).

• Serious adverse events and adverse drug reactions (clinical development and post-marketing).

• Drug withdrawals (post-marketing).

5

Cardiovascular Safety Publication Trends (Keywords “QT”, “hERG”, “Torsades de Pointes”)

6 J-P Valentin, Br. J. Pharmacol. (2010), 159, 5 - 11

Publications referring to “QT”, “hERG” and “Torsades de Pointes” have grown significantly over the past decade

Incidence of Potential Drug-Induced Arrhythmia Adverse Events

7

Accumulated QT reports

Accumulated QRS reports

Accumulated PR reports

from: J-P Valentin, Br. J. Pharmacol. (2010), 159, 5 – 11.

Adverse Events Reporting System (AERS) •Potential drug-induced events •Voluntarily entered (physicians & patients) •Mandatorily entered data (manufacturers)

AERS records show consistent reporting of drug-induced arrhythmias over the past decade

8

Implementation of Guidance Documents for Cardiovascular Safety Assessment

There has been an increase in number and scope of regulatory guidelines across all therapeutic areas (increased surveillance)

from: J-P Valentin, Br. J. Pharmacol. (2010), 159, 5 - 11

Oncology Chemotherapeutics Historically Associated with Cardiovascular Toxicities

9

• Anthracyclines • Dose-dependent contractile dysfunction and heart failure due to

myocardial cell toxicity / death

• 5-Flurouracil & Capecitabine • Cardiovascular ischemia due to increased vasoconstriction

• Cyclophosphamide • Association with acute/sub-acute congestive heart failure and

cardiovacular ischemia due to cardiac necrosis

• Vinca alkaloids • Association with cardiac ischema

• Mitoxantrone • Association with cardiomyopathy (> with prior doxorubicin)

Targeted Oncology Drugs Can Also Mediate a Spectrum of Cardiovascular Events

Drug Event Frequency Mechanism

Bevacizumab Cardiovascular ischemia ≤ 3.8% Thromboembolism

Contractile dysfunction 1-3%

Sunitinib Cardiovascular ischemia ≤ 3% Thromboembolism

Contractile dysfunction / heart failure

8-15% Mitochondrial dysfunction

Arterial hypertension -

Sorafenib Cardiovascular ischemia ≤ 3% Thromboembolism

Arterial hypertension

-

Trastuzumab Contractile dysfunction / heart failure

3-18%; 0.4-3.6% Myofbrilar disorganization

Lapatinib Contractile dysfunction / heart failure

1.4%; 0.2%

10

Adapted from: Stortecky and Suter, Curr. Opinion in Oncol., (2010), 22, 312 - 317

Oncology Drugs Reported to Impact Cardiac Re-polarization (additional examples)

Class Drug HDAC Inhibitors Vorinostat

Depsipeptide (FK228) Panobinostat

MTKIs Sunitinib Vandetinib

Vascular Targeting Agents

Combretastatin A4 Ph. 5,6-Dimethylxanthenone AA

Farnesyl Transferase Inhibs.

Lonafarnib

Src / Abl Kis Dasatanib Nilotinib

11

Class Drug Topoisomerase Inhibs. Aclarubicin

ErbB-1 / -2 Rec. Inhibs.

Lapatanib

mTOR Inhibs. Temsirolimus

Microtubule Targeting Agents

Ixabepilone

Alkylating Agents Bisulfex

Proteosome Inhibs. Bortezomib Others Arsenic trioxide

Acodazole

from: Morganroth et al, Clin. Pharm. Ther. (2009), 87(2), 166-174.

The Goals of Early Phase Oncology Studies

• Describe the pharmacology of the drug • Pharmacokinetics • Pharmacodynamics

• Establish Phase 2 dose and schedule (e.g. maximum tolerated) • Observe anti-tumor activity (efficacy)

• Identify patient selection and response biomarker(s)

• Establish enough data on safety and tolerability to inform the next phase of development • Always have the limitation of small patient numbers • Identify key safety signals (e.g. cardiovascular) requiring further

investigation

Are there any “road-blocks” to the further development?

12

What Information is Typically Available to Enable Planning of Early Phase Oncology Studies?

• Pre-clinical data • hERG ion channel (and possibly hERG trafficking) and ECG

data in some species (usually canine) • General viewed as predictive when positive, but absence of a signal

does not mean a clinical risk can be ruled out. • Toxicology data from relevant species • PK data and human dose projections

• Guidance • S9 Nonclinical evaluation for anticancer pharmaceuticals • E14: Clinical evaluation of QT/QTc interval prolongation and pro-

arrhythmic potential for non-antiarrhythmic drugs

13

Early Patient Selection: Who enrols in Phase I Oncology Trials?

• Advanced solid tumor patients • “Limited” expected lifespan (per S9 guidance) • Unresponsive to standard therapy, limited options • All-comers vs. restriction to certain populations

• Well enough to receive therapy • Life expectancy >3 months • Adequate organ function (hematologic, renal, hepatic)

• Not high-risk for particular safety concerns • No uncontrolled medical disorders (e.g. brain mets) • No medications with potential drug-drug interactions

Key Considerations in Planning ECG Assessment in First in Human Oncology Studies

• Co-morbidities can impact cardiac assessments: • nausea, vomiting, diarrhea, mucositis • polypharmacy

• Some concomitant medications may be critical to patient support:

• 5-HT3 antagonists • anti-epileptics • antibiotics and antifungals

• Typical phase 1 exclusion criteria: • subjects with baseline QTcF >470 msec • myocardial infarct within prior 6 months • symptomatic congestive heart failure or unstable angina • uncontrolled hypertension • cardiac arrhythmia requiring medication

15

Other Considerations in Planning ECG Assessment in First in Human Oncology Studies

• A key element of the ICH E14 Guidance is the description of the “Thorough QT/QTc Study” • Most oncologic drugs cannot be studied in healthy volunteers (at

therapeutic doses) • Use of placebos and controls as suggested in ICH E14 is

ethically controversial in cancer patients

16

If “thorough QT studies” are impractical, it is essential to gain as much information as possible

from the available clinical studies

What ECG Surveillance/Precautions Should be Incorporated in FIH Oncology Studies?

• Concomitant-Med Restrictions: should be avoided if possible

• ECG Machines: ideally standardized, centrally calibrated and maintained ECG machines

• Multiple Baseline Measures: 3 sets of triplicate ECGs

• Collection Methods: triplicates separated by ≥30 sec • Timing taking into account the modality (large vs small

molecule) • Multiple timepoints (initial dosing, steady state, and study exit) • Should be gathered close to the Cmax / Tmax

• PK Associations: “Concentration – Effect” should be determined at these multiple time points

17

What ECG Surveillance Should be Incorporated in FIH Studies? (cont.)

• Patient electrolytes monitored and be proactively treated

• QT and QTc safety parameters should be prospectively determined • Accepted that QTcF (Fridericia correction) is optimal

• Analysis may be by “site read”, however a central core laboratory for standardized assessment and reporting may be warranted.

The intensity of ECG assessments in subsequent studies will be determined by data from FIH study, consideration of the properties of

the drug, and aspects of the patient population

18

Incorporation of Other Cardiovascular Assessments in Early Oncology Studies

• Will depend on composition and MOA of the drug and any pre-clinical signals. • LVEF: echocardiograms, MUGA • Ischemia: transdermal doppler, serum analytes (e.g.,

troponin)

• Combined Modalities (e.g. Antibody Drug Conjugates) will need to consider the safety of the component elements (eg, Ab – Linker – Drug and the final released “warhead”).

19

20

Example of ECG Collection Methods: Drug X

Phase 1 Phase 2 Patients 53 38 Standardization (machine)

None None

Baseline ECGs 4 2

Replicates Triplicates (Baseline, Cmax

x2)

Single

Collection Analog, Site Analog, Site

Reader Machine Machine

21

ECG Analysis Methods

• Population PK Analysis • PK vs. QTcF change

• Outlier analysis- QTcF Data Tables • QTcF > 480ms or 500ms at any time (Shift) • Max change in QTcF > 30ms or 60ms at any

time

• Individual Case Evaluations • Focus on cases identified above

22

Drug X Concentration vs. Change in QTcF (n=53 patients)

23

QTcF Shift Table

Cohort 1 (Dose “Y”)

Cohort 2 (Dose “3Y”)

Cohort 3 (Dose “10Y”)

Cohort 4 (Dose “12Y”)

Cohort 5 (Dose “20Y”)

24

Maximum QTcF change from baseline

Cohort 1 Cohort 2 Cohort 3 Cohort4 Cohort 5

Published Experience of ECG Analysis in Phase I Anti-Cancer Drug Development

• 525 patients enrolled in 22 early phase anticancer studies at MD Anderson (55% FIH studies) • Retrospective chart review (ECGs) • Multiple tumor types • Multiple Drugs (e.g. PI3K/AKT/mTOR, Hsp90, cytotoxic agents, etc.) • 8518 ECGs (~16 ECG/patient, mostly in first 4 wks)

• All subjects had a baseline ECG 1 week prior to entry • Post-dose ECGs were at 1, 6, 12 and 24 hrs post first dose in first

treatment cycle • Subsequent ECGs during first 4 weeks of treatment and end of study.

25

1. A Naing et al., Ann. Oncol., (2012), published online June 27, 2012

Key Findings of Naing et al (2012) • 33 patients (6%) had prolonged QTc at baseline.

• 7 patients (21%) had normalization of QTc after dosing.

• 74 patients (14%) with normal QTc at baseline had elevated QTc above ULN at any time on study. • All were considered clinically insignificant.

• 2 patients (0.4%) experienced serious cardiac events (1 myocardial infarct and 1 unstable atrial flutter). • Neither event detected by ECG

• Authors’ conclusions: • Frequent ECG monitoring provided no clinically significant information in this

small highly selected patient series. • Clinical evaluation continues to be important for patient safety and “more

modest ECG monitoring could be considered in early phase oncology trials.”

26

Late Attrition is Typically Due to Lack of Efficacy, not Safety

$

Preclinical and early-phase safety assessments help prevent late stage surprises

(Arrowsmith J, Nat Drug Disc; 2011)

Acknowledgements

• Mike Engwall, Hugo Vargas

28