What’s the point of ENDPOINTS? Translating trial outcomes into …€¦ · These slides were presented at a Takeda Oncology sponsored symposium at the 18th British Thoracic Oncology

This symposium was organised and funded by Takeda UK Ltd. Data relating to Takeda medicines was used as illustrative examples in this symposium.Prescribing information is available at the end of the presentation.

What’s the point of ENDPOINTS? Translating trial outcomes into clinical practice

.Job code: UK/BRIG/1911/0094x(1)Date of preparation: April 2020

These slides were presented at a Takeda Oncology sponsored symposium at the 18th British Thoracic Oncology Group (BTOG) Annual Conference and are intended for Healthcare Professionals only. These have been made available to Healthcare Professionals who attended the meeting and have requested a copy of the slide deck

presented or have downloaded from www.alunbrig.co.uk/hcp. These slides are not for further sharing or use without Takeda’s permission.

Agenda

Welcome and introductions Professor Samreen Ahmed

Why focus on endpoints and interpretation of clinical trial outcomes? Professor Samreen Ahmed

An overview of common endpoints used in lung cancer clinical trials Professor Allan Hackshaw

Practical interpretation of lung cancer clinical trial endpoints Dr Tom Newsom-Davis

Emerging ideas in lung cancer clinical trials Professor Allan Hackshaw

Final remarks and close Professor Samreen Ahmed

Disclosures

Professor Samreen Ahmed

▪ Consultancies/Travel grants/Lectures fees: Pfizer, Roche, AZ, Takeda, Novartis, Eisai, MSD

Professor Allan Hackshaw

▪ Educational Activities on behalf of: Boehringer, Roche, Abbvie, Merck Serono, MSD, Daiichi Sankyo and UCB (to employees and external health professionals).

▪ Ownerships and possessions in health care companies: Shares in Illumina and ThermoFisher

▪ Advisory boards: Precision Xtract (for Abbvie), and Grail Inc (biotech company)

Dr Tom Newsom-Davis

▪ Support to attend educational conferences: AstraZenca, BMS, Boehringer Ingelheim, Lilly, MSD, Otsuka, Roche, Takeda

▪ Advisory and Speaker Bureaus: Amgen, AstraZeneca, BMS, Boehringer Ingelheim, Eli-Lilly, MSD, Novartis, Otsuka, Pfizer, Roche, Takeda

Professor Samreen Ahmed

University Hospitals of Leicester NHS Trust, Leicester, UK

Why focus on endpoints and interpretation of clinical trial outcomes?

This presentation utilises data from lung cancer clinical trials to illustrate learning points around endpoints and is not meant to be fully illustrative of the efficacy and safety profile of each product. For full information about each product, please refer to the prescribing information (available for brigatinib at the end of this presentation) or the Summaries of Product Characteristics.

How do we interpret clinical trials to inform treatment decisions?

AEs, adverse events; DOR, duration of response; iPFS, intracranial progression free survival; OS, overall survival; PFS, progression free

survival; QOL, quality of life; RWE, real-world evidence.

Trial design

Phase

Randomised

N=?Primary endpoint

Inclusion/exclusion

Cross-over

Comparator arm

Single-arm

Interventional/observational

Reported outcomes

PFS

QOL

OS

DORiPFS

AEs

Treatment decision

Clinical interpretation

Audience questionWhich measure of efficacy would you most commonly use in clinical practice?

1. Hazard ratio on the survival curve

2. Median PFS

3. Median OS

4. Intracranial PFS

5. Duration of response (DoR)

PFS, progression-free survival.

*Only targeted and immunotherapeutic agents have been summarized in this image1. Aggarwal C and Borghaei H. The Oncologist 2017;22:700–708.

AEs, adverse events; CNS, central nervous system; DOR, duration of response; EGFR, epidermal growth factor receptor; FDA, Food and Drug Administration; HRQOL, health-related quality of life; MTD, maximum tolerated

dose; ORR, objective response rate; OS, overall survival; PD-L1, programmed death-ligand 1; PK, pharmacokinetics; PFS, progression free survival; PROs, patient reported outcomes; QOL, quality of life; TTF,

time-to-failure; TTP, time-to-progression.

Which endpoints do we consider when making treatment decisions?

How do we evaluate these when making

everyday clinical decisions?

Endpoints reported in 18 clinical trials for FDA approved agents for NSCLC lung cancer trials*1

TimeToTumour

Response

How do we evaluate and interpret those endpoints when making treatment decisions?

What is the best endpoint to

assess intracranial

efficacy? – ORR, PFS, DOR?

Intracranial efficacy

How do you interpret quality of life outcomes?

Patient reported outcomes

…DOR, duration of response; HR, hazard ratio; INV, investigator; IRC, independent review committee;

ORR, objective response rate; PFS, progression-free survival; TKI, tyrosine kinase inhibitor.

Efficacy measures

When do we look at IRC vs INV

assessed outcomes?

How relevant is early separation

of survival curves?

How do you interpret hazard

ratios and the overall shape of

the Kaplan-Meier curve?

In addition to PFS and OS, how

important are other measures of efficacy such

as DoR?

Key questions when interpreting efficacy measures

CI, confidence interval; DOR, duration of response; HR, hazard ratio; INV, investigator; IRC, independent review committee; NSCLC, non-small cell lung cancer; PFS, progression free survival.

aLicensed dose for ALUNBRIG is 180 mg once daily with 7 days lead-in at 90mg once daily.The ALTA trial was not designed to statistically compare between arms.1. Huber RM, et al. J Thoracic Oncol 2019;S1556-0864(19)33645-7. 2. Alunbrig Summary of Product Characteristics.

ALTA trial: Brigatinib in crizotinib-refractory ALK+ advanced NSCLC1

When do we look at IRC vs INV

assessed outcomes?

In addition to the median PFS, how

often do you evaluate the overall

shape of the curve or the hazard ratio? How important is DOR as a

measure of efficacy?

Median duration of response in responders13.8 months (95% CI 10.2–19.3 months)

Investigator-Assessed90 mg-> 180 mg once-dailya

n=110

IRC-Assessed90 mg-> 180 mg once-dailya

n=110

Median PFS, months (95% CI) 15.6 (11.1–21.0) 16.7 (11.6–21.4)

1-year PFS probability, % (95% CI) 58 (47–67) 61 (49–70)

Arm A: mPFS: 9.2 months (95% CI 7.4–12.8)Arm B: mPFS: 16.7 months (95% CI 11.6–21.4)

IRC-assessed systemic PFS

Alunbrig (brigatinib) is indicated as monotherapy for the treatment of adult patients with anaplastic lymphoma kinase (ALK)-positive advanced non-small cell lung cancer (NSCLC) previously treated with crizotinib.2 The most common (in ≥10% of patients at a frequency of ≥30% at any grade) treatment-emergent adverse events are diarrhoea, nausea, cough and increased blood creatine phosphokinase.2

Key questions when interpreting intracranial efficacy measures

CI, confidence interval; HR, hazard ratio; i, intracranial; IRC, independent review committee; NSCLC, non-small cell lung cancer; ORR, intracranial overall response rate; mDOR, median

duration of response; mPFS, median progression-free survival; NR, not reached; ORR, objective response rate.

What is the best endpoint to

assess intracranial

efficacy? – iORR, iPFS, iDOR?

Intracranial mPFS 18.4 months (95% CI 12.6–23.9)

Intracranial ORR 67% (95% CI 41–87)

Intracranial mDOR 16.6 months (95% CI 3.7–NR)

IRC-assessed measures of intracranial efficacy for brigatinib in crizotinib-refractory ALK+ advanced NSCLC1,a

IRC-assessed intracranial PFS in patients with any brain metastases at baseline

a80 mg once daily with 7-day lead-in at 90 mg once daily.The ALTA trial was not designed to statistically compare between arms.1. Huber RM, et al. J Thoracic Oncol 2019;S1556-0864(19)33645-7.


Key questions when interpreting patient reported outcomes (PRO)

aLicensed dose for ALUNBRIG is 180 mg once daily with 7 days lead-in at 90mg once daily.** Includes one grade 5 event 1. Lenderking WR, et al. Future Oncology 2019;15:2841–2855; 2. Kim D-W, et al. J Clin Oncol 2017;35:2490–2498.

AE, adverse event; ALK, anaplastic lymphoma kinase; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BL, baseline; C, cycle; CPK, creatine

phosphokinase; D, day; GHS, global health status; QD, once daily; QoL, quality of life.

How do you interpret

quality of life outcomes?

Mean change in Global Health Status/Quality of Life subscale over treatment

The PRO measurement scale used in the study

was the European Organization for

Research and Treatment of Cancer Quality of Life

Questionnaire Core-30 (EORTC QLQ-C30 V3.0)

Overall mean change from baseline Arm A: 5.3 (95% CI: 2.2–8.4) Arm B: 6.2 (95% CI: 3.1–9.2)


Limited conclusions can be drawn from this quality of life data in this unblinded study with no placebo or active control arm

This presentation utilises data from lung cancer clinical trials to illustrate learning points around endpoints and is not meant to be fully illustrative of the efficacy and safety profile of each product.

For full information about each product, please refer to the prescribing information (available for brigatinib at the end of this presentation) or the Summaries of Product Characteristics.

Common endpoints and terminology in lung cancer trials


CRUK Cancer Trials Centre, University College London, UK


What makes a ‘good’ endpoint for treatment trials?

▪ Simple and easy to measure

▪ Reliable and reproducible

▪ Has clinical relevance to patients, health professionals and decision makers

▪ Is ‘sensitive’ to the experimental treatment: if the therapy really is more effective the endpoint can show this

QoL, quality of life.European network for Health Technology Assessment. 2013. Available at: www.eunethta.eu/wp-content/uploads/2018/01/Clinical-endpoints.pdf. Accessed January 2020.

Commonly used time to event endpoints: PFS and OS

Reaches progression first

Reaches death first without progression

Reaches both progression and death

Date of trial enrollment/randomisation

Date of progression

Date of death

▪ Any of these 3 eventsconstitute PFS

▪ Any patient alive andnot progressed is censored

▪ Only this event (death) constitutes overall survival

▪ Any patient alive is censored

Reaches death

OS, overall survival; PFS, progression-free survival.

Overall Survival (OS)

Progression FreeSurvival (PFS)

OR

OR

Hackshaw A. A concise guide to clinical trials. April 2009. DOI:10.1002/9781444311723.

PFS and OS: What are the strengths and limitations?

OS is considered the gold standard for oncology clinical trials1

PFS is a widely used surrogate endpoint1


…however it is……because it is…

✓ Easily and precisely measured

✓ Completely objective

✓ Assessed earlier than OS✓ Not influenced by crossover

and subsequent therapies✓ Have more events, requiring

smaller and shorter trials

ꭓ Affected by crossover and/or subsequent therapies

ꭓ Needs long enough follow-up to get enough events

ꭓ Includes non-cancer deaths*

*Not a limitation when caused by drug toxicity2

ꭓ Potentially influenced by assessment bias

ꭓ Depends on frequency of radiological

ꭓ May not correlate highly with survival

ꭓ Uncertainty over their clinical relevance/value

1. Genentech. Oncology endpoints in a changing landscape. January 2016. Available at: www.managedcaremag.com/sites/default/files/graphics/OncoEndpoints_MC.pdf. Accessed January 2020; 2. Hackshaw A. A concise guide to clinical trials. April 2009. DOI:10.1002/9781444311723.

Audience questionDuration of Response (DoR) takes into consideration…

1. Only patients who achieve Complete Response (CR) to treatment

2. Patients who achieve Complete Response (CR) or Partial Response (PR) to treatment

3. All patients included in the Intent-To-Treat (ITT) population (same as for PFS)


What is Duration of Response?

Reaches progression first


Reaches progression and death together

Date of trial enrollment/randomisation

Date of progression

Date of death

Reaches deathOverall Survival (OS)1

Progression FreeSurvival (PFS)1

together

OR

OR

Duration of Response (DOR)2

Documentation of tumourresponse

▪ Measured from start of response

▪ Any of these 3 eventsconstitute loss of response

▪ Any responder alive andnot progressed is censored

Reaches progression


Reaches both progression and death

firstOR

OR1. Hackshaw A. A concise guide to clinical trials. April 2009. DOI:10.1002/9781444311723; 2. Genentech. Oncology endpoints in a changing landscape. January 2016. Available at: https://tinyurl.com/w7wfwae. Accessed January 2020.

Is Duration of Response (DoR) a useful efficacy measure?

▪ Only based on a subset of patients, i.e. only those who have had a complete or partial response1

▪ Baseline characteristics might not be balanced anymore; potential for confounding/bias1

However…

▪ A long median DoR may reflect fast onset of efficacy (early response)

▪ Can highlight a plateau

▪ Useful for advanced cancers

1. Ellis S, et al. Contemp Clinl Trials 2008;29:456–465; 2. Kim DW, et al. Lancet Oncol 2016;17:452–463.

Patients with ALK-rearranged advanced NSCLC who responded to ceritinib 750 mg treatment in the Phase 1 ASCEND-1 trial.2

ALK, anaplastic lymphoma kinase; DoR, duration of response; NSCLC, non-small cell lung cancer.

▪ Considers the entire Kaplan-Meier curve1

▪ Tells us whether risk has gone up or down compared to a control therapy (but does not reflect the background risk)

▪ In this example:2

― Risk (hazard) of progressing/dying reduced by 51% compared to chemotherapy (at any time point)

― Risk is nearly halved

1. Hackshaw A. A concise guide to clinical trials. April 2009. DOI:10.1002/9781444311723; 2. Shaw AT, et al. N Engl J Med 2013;368:2385–2394.

Patients with ALK-rearranged advanced NSCLC treated with crizotinib (n=173) or chemotherapy (n=174)2

3 ways of examining treatment efficacy: What does hazard ratio tell us?

ALK, anaplastic lymphoma kinase; CI, confidence interval; PFS, progression-free survival NSCLC, non-small cell lung cancer.

▪ Tells us only about that point on the curve when 50% of patients have had an event (progression or death)1

▪ In this illustration:2

― Half the crizotinib patients live without progression for 7.7 months

― Half the chemotherapy patients live without progression for 3 months

― Median PFS among crizotinib patients is 4.7 months longer than chemotherapy

1. Hackshaw A. A concise guide to clinical trials. April 2009. DOI:10.1002/9781444311723; 2. Shaw AT, et al. N Engl J Med 2013;368:2385–2394.

Median PFS7.7

3.050

3 ways of examining treatment efficacy: What does median PFS tell us?



40%

20%

▪ You/investigator choose an appropriate time point1

▪ In this example:2

― 40% crizotinib patients live without progression to 10 months

― 20% chemotherapy patients live without progression to 10 months

― 40–20 = 20 more patients per 100 are alive without progression by 10 months when given crizotinib

▪ Called absolute risk difference

1. Morgan CJ. J Nucl Cardiol 2019;26:391–393; 2. Shaw AT, et al. N Engl J Med 2013;368:2385–2394.

3 ways of examining treatment efficacy: What does landmark PFS tell us?



The 3 effect sizes each tell us something different about the effect of a new therapy

Hazard Ratio

Median PFS

Landmark PFS(absolute risk difference)

*The risk of an event at a single time point is called a hazard; and the ratio of hazards should be the same at all time points (except the very start).1. Shaw AT, et al. N Engl J Med 2013;368:2385–2394; 2. Hackshaw A. A concise guide to clinical trials. April 2009. DOI:10.1002/9781444311723. PFS, progression-free survival.

Assessment as an efficacy measure2What did it tell us?1

▪ Risk (hazard) of progressing/dying reduced by 51% (at any time point)

▪ Crizotinib patients live without progression 4.7 months longer, on average

▪ Measure of relative effect▪ Uses the whole Kaplan-Meier curve▪ Scientifically the most robust effect size▪ Underlying assumption of ‘proportional hazards’*

▪ 20 more patients per 100 are alive without progression at 10 months when given crizotinib

▪ Measure of absolute effect▪ But only at one point in the curve; can be influenced by

unexpected blips▪ Uses time

▪ Measure of absolute effect▪ But only at one time point in the curve; can be influenced

by unexpected blips▪ Measure of ‘impact’

Audience questionWhich method for examining treatment efficacy do you prefer?

1. Hazard ratio

2. Median PFS

3. Landmark PFS


▪ ~30% patients with advanced NSCLC develop CNS metastases: but 50–60% for ALK-positive patients1,2

▪ Because patients are living longer, up to 90% may ultimately develop brain metastases during the course of their disease3

▪ CNS/brain mets can impact patient’s mental/cognitive function4

Incidence of CNS progression5

(all patients)

An event is:▪ New CNS lesion▪ Baseline CNS lesion grows

Intracranial PFS6

(patients with baseline brain mets)a

An event is:▪ Brain progression▪ Death, any cause

Intracranial response rate6

(patients with baseline brain mets)a

aLicensed dose for ALUNBRIG (brigatinib) is 180 mg once daily with 7 days lead-in at 90mg once daily.1. Remon J, Besse B. Front Oncol. 2018;8:88; 2. Griesinger F, et al. Oncotarget. 2018;9(80):35181–35194; 3. Rusthoven CG and Doebele RC. J Clin Oncol 2016;34:2814–2819; 4. Schagen SB, et al. EJC Suppl. 2014;12(1):29–40; 5. Peters S, et al. N Engl J Med 2017; 377:829-838; 6. Huber RM, et al. J Thorac Oncol 2019 DOI: https://doi.org/10.1016/j.jtho.2019.11.004.

ALK, anaplastic lymphoma kinase; CI, confidence interval; CNS, central nervous system; iCR, intracranial complete response; iDOR, intracranial duration of response; iORR,

intracranial objective response rate; iPR, intracranial partial response; NR, not reported; NSCLC, non-small cell lung cancer.

iDOR16.6 months

95% CI: 3.7–NR

Intracranial duration of response6

Time from iCR/iPR to progression or deatha

Importance

vs. vs.

vs.

iORR67% (12)

95% CI: 41–87%

Intracranial efficacy of brigatinib in patients with ALK+ advanced NSCLC post-crizotinib (ALTA trial)2

Intracranial efficacy endpoints

▪ Quantifies the patient’s perception of their mental and physical well-being

▪ Show specific symptoms and issues that matter for lung cancer patients

▪ For advanced cancers, just maintaining QoL is a ‘positive’ outcome▪ Can show improvements in QoL compared with the standard/control therapy

Be aware: QoL is not easy to interpret!

We have to know:▪ The scale of each measure

(here 0-100); EORTC-QLQC30 and

▪ Whether high score indicates ‘good’ or ‘bad’ QoL (here high=worse QoL) and

▪ What is a clinically relevant improvement (here ~5 unit change)

1. Roever, Evidence Based Medicine and Practice 2016;1:2; 2. Polanski J, et al. Onco Targets Ther 2016:9:1023–1028; 3. Damm et al. Health Econ Rev 2013;3:15; 4. Romero M, et al. SpringerPlus 2013;2:644; 5. Shaw AT, et al. N Engl J Med 2013;368:2385–2394.

ALK, anaplastic lymphoma kinase; EORTC-QLQC30, European organisation for research and treatment quality of life

questionnaire C30; OS, overall survival; PFS, progression-free survival; QoL, quality of life.

Health-related QoL is a useful supporting endpoint to OS and PFS

Importance1–4

Overall change from baseline in symptoms and global QoL in patients with ALK-rearranged advanced NSCLC treated with crizotinib or chemotherapy5

Another way of analysing the same quality of life data

▪ Very different approach to previous

▪ Time taken for patient’s score to increase by 10 units (i.e. QoL gets worse)

▪ Someone has to define what ‘worse’ means

▪ Risk of having QoL deteriorate by 10 units is halved among crizotinib patients compared to chemotherapy

▪ Median time taken for QoL to deteriorate

―~1.5 months chemotherapy

―~5 months crizotinib

Time to deterioration with respect to a compositeend point of three symptoms

(cough, dyspnoea, or chest pain)

Shaw AT, et al. N Engl J Med 2013;368:2385–2394. QoL, quality of life.

Summary – trial endpoints

▪ Always be clear about how each endpoint was measured

▪ Is it measured on all patients (intention-to-treat for overall survival) or a subset (duration of response)?

▪ Could the endpoint be significantly affected by bias or lack of trial treatment blinding?

▪ Has it been significantly affected by treatment crossovers?

▪ Is the treatment effect small, moderate or big? The answer to this always depends on the choice of effect size (hazard ratio, median, or risk difference)

Practical interpretation of lung cancer clinical trial endpoints

Dr Tom Newsom-Davis

Chelsea and Westminster Hospital, London, UK


Audience questionWhich of the following trials did not have a statistically significant improvement in median PFS in the experimental arm?

1. KEYNOTE-010

2. FLAURA

3. IMPOWER-150

4. ALEX

5. IMPOWER-133

Audience questionWhich of the following trials did not have a statistically significant improvement in median PFS in the experimental arm?

1. KEYNOTE-0101

2. FLAURA2

3. IMPOWER-1503

4. ALEX4

5. IMPOWER-1335

1. Herbst RS, et al. Lancet 2016;387:1540–1550; 2. Ramalingam S, et al. Ann Oncol 2016;28(suppl_5):v605–v649; 3. Socinski MA, et al. N Eng J Med 2018;378:2288–2301; 4. Peters S, et al. N Eng J Med 2017;377:829–838; 5. Liu S, et al. Presented at WCLC. 2018.

12-month PFS

5.4%

12.6%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24Months

100

90

80

70

60

50

40

30

20

10

0

PFS

(%

)

Atezolizumab + CP/ET

Placebo + CP/ET

HR=0.77(95% CI: 0.62–0.96)p=0.017

0 3 6 9 12 15 18 2421

Pro

bab

ility

of

PFS

0.0

0.2

0.4

0.6

0.8

1.0

17.5%

4.7%

Time from randomisation (months)

HR=0.78(95% CI: 0.645–0.936)p=0.047

12-month PFS

Investigator assessed PFS from the Impower-133 trial(Blinded)1

CASPIAN(Open Label)*,2

Blinded and open-label trialsmPFS in extensive stage SCLC

Durvalumab + chemo

Chemo

*Durvalumab in combination with platinum-based chemotherapy regimens with or without tremelimumab does not currently have a marketing authorisation in the UK for small-cell lung cancer3

1. Liu S, et al. Presented at WCLC. 2018; 2. Paz-Ares L. Presented at WCLC. 2019; 3. Durvalumab NICE HTA 2019.CI, confidence interval; CP, carboplatin; ET, etoposide; HR, hazard ratio; mPFS, median

progression-free survival; SCLC, small cell lung cancer.

The results from these studies are not intended to be used for cross-trial comparison but rather illustrate interpretation of endpoints.

IPASS (Iressa Pan-ASia Study)* Investigator-assessed PFS

IPASS (Iressa Pan-ASia Study)* BIRC-assessed PFS

*Phase III, multicentre, randomised, open-label, parallel-group IPASS study, which compared EGFRTKI gefitinib (IRESSATM) with carboplatin plus paclitaxel (carbo-platin/paclitaxel) doublet chemotherapy as first-line treatment for Asian patients with advanced NSCLC.1. Wu YL, et al. Lung Cancer 2017;104:119–125.

BIRC, blinded independent review committee; CI, confidence interval; HR, hazard ratio; INV, investigator; PFS, progression-free survival.

▪ Reflective of real-world clinical practice▪ Method usually not a problem for double-blind trials,

but potential for bias in open-label trials

▪ Blinded independent review committee (BIRC) minimises this bias

Investigator vs. independent review outcomes Is one better and are both needed?

Do landmarks mean more than medians?KEYNOTE-001: Pembrolizumab in advanced NSCLC

Garon EB, et al. J Clin Oncol 2019;37:2518–2527.CI, confidence interval; NSCLC, non-small cell lung cancer; OS, overall survival

PD-L1, programmed death ligand 1; TPS, tumour proportion score.

Kaplan-Meier estimates of 5-year overall survivalTreatment-naïve patients by PD-L1 TPS status

Kaplan-Meier estimates of 5-year overall survivalPreviously treated patients by PD-L1 TPS status

However, it is important to exercise care in interpretation…

▪ Authors concluded that atezolizumab had no effect on PFS

▪ But HR 0.95 is difficult to interpret; the proportional hazards assumption clearly fails

▪ Comparison of medians does does not reflect the shape of the curve over the course of time

Rittmeyer A, et al. Lancet 2017;389:255–266(Suppl).CI, confidence interval; HR, hazard ratio; mPFS, median

PFS; PFS, progression-free survival.

Atezolizumab

Docetaxel

Median 2.8 months

Median 4.0 months

Hazard ratio 0.9595% CI 0.82–1.10P=0.49

Progression-free survival for atezolizumab vs docetaxel in previously treated NSCLC (OAK)

Median PFS, but no median OS benefitPROFILE-1014: 1st line crizotinib in ALK+ advanced NSCLC

PFS HR ORR

Crizotinib 10.9m

0.45

74%

Chemo 7.0m 45%

OS HR Probability of 1-year survival

Crizotinib NR

0.82

84%

Chemo NR 79%

Solomon B et al. N Eng J Med 2014;371:2167–2177.CI, confidence interval; HR, hazard ratio; NR, not reached; NSCLC, non-small cell lung cancer; ORR, objective

response rate; OS, overall survival; PFS, progression free survival.

Progression-free survival in the intention-to-treat population (PROFILE-1014)

Overall survival in the intention-to-treat population (PROFILE-1014)

PROFILE-1014: Survival by subsequent treatment1L crizotinib in ALK+ advanced NSCLC

*Crossover from the chemotherapy arm to crizotinib and from the crizotinib arm to chemotherapy.Solomon BJ, et al. J Clin Oncol 2018;36:2251–2258. NSCLC, non-small cell lung cancer; OS, overall survival; TKI, tyrosine kinase inhibitor.

OS in the intention-to-treat population adjusted for crossover* (PROFILE-1014)

Median OS, but no median PFS benefitKEYNOTE-010: Pembrolizumab vs docetaxel in previously treated PDL1+

advanced NSCLC

PFS for total population OS for total population

Treatment Median (95% CI), mo

HR(95% CI)

P-value

Pembro 2mg/kg 3.9 (3.1–4.1) 0.88 (0.74–1.05) 0.07

Pembro 10mg/kg 4.0 (2.7–4.3) 0.79 (0.66–0.94) 0.004

Docetaxel 4.0 (3.1–4.2) – –

Treatment Median (95% CI), mo

HR(95% CI)

P-value

Pembro 2mg/kg 10.4 (9.4–11.9) 0.71 (0.58–0.88) 0.0008

Pembro 10mg/kg 12.7 (10.0–17.3) 0.61 (0.49–0.75) <0.0001

Docetaxel 8.5 (7.5–9.8) – –

Herbst RS, et al. Lancet 2016;387(10027):1540–1550.CI, confidence interval; HR, hazard ratio; NSCLC, non-small cell lung cancer; OS, overall survival;

PFS, progression-free survival; TPS, tumour proportion score.

Duration of response (DoR) versus PFS: IMPOWER150: Atezolizumab plus carboplatin and paclitaxel +/- bevacizumab in

untreated metastatic non-squamous NSCLC

Time (months)

Median, 8.3 mo(95% CI: 7.7, 9.8)

Median, 6.8 mo(95% CI: 6.0, 7.1)

HR, 0.59 (95% CI: 0.50, 0.70)

P < 0.0001Median follow-up: ~20 mo

0%

10%

20%

30%

40%

50%

60%

70%

80%

Arm A Arm B Arm C

Ob

ject

ive

resp

on

se r

ate

(%)

40%

56%

41%

Arm A: atezo + CPArm B: atezo + bev + CPArm C: bev + CP

PR CR

Median DOR (months)

8.3 11.5 6.0

ITT-WT population consisted of patients without ALK rearrangements or EGFR mutations.Socinski MA, et al. J Clin Oncol 2018;36:abstr 9002.

Atezo, atezolizumab; bev, bevacizumab; CI, confidence interval; CP, chemotherapy; CR, complete response; DOR, duration of response; ITT, intention to treat; NR, not reached; ORR, overall response rate;

PFS, progression free survival; PR, partial response; SD, stable disease; WT, wild type.

Progression-free survival (WT-ITT) Objective response (ITT)

Is significant always significant?SQUIRE: 1L platinum chemotherapy +/- necitumumab in patients with stage IV

squamous NSCLC

Overall survival

Progression-free survival

Thatcher N, et al. Lancet Oncol 2015;16:763–774. CI, confidence interval; NSCLC, non-small cell lung cancer; 1L, first-line.

Is insignificant always insignificant?

KEYNOTE-6041*

“…there was also an improvement in OS … however, these OS results did not meet statistical significance per the pre-specified statistical plan (HR=0.80 [95% CI, 0.64–0.98])”1

CASPIAN2

Overall survival (primary endpoint)

1. KEYNOTE-604 press release. 6 January 2020. Available at: tinyurl.com/w645foo. Accessed January 2020;2. Paz-Ares et al. Lancet 2019;394:1929–1939. CI, confidence interval; EP, platinum etoposide; HR, hazard ratio; OS, overall survival.

*Patients with newly diagnosed extensive stage SCLC (N=453) were randomised to:Pembrolizumab plus etoposide and investigator's choice of platinum chemotherapy (carboplatin or cisplatin), ORPlacebo plus etoposide and investigator's choice of platinum chemotherapy (carboplatin or cisplatin)1

Please note these trials used pembrolizumab and durvalumab in untreated extensive stage SCLC. The results from these studies are not intended to be used for cross-trial comparison but rather to illustrate interpretation of endpoints.

In summary: What are the considerations when interpreting trial data in clinical practice?

▪ What was the objective of the trial/analysis?

– Was it powered for statistical significance

– Was it open-label or double-blinded?

▪ Was it designed to show superiority or non-inferiority?

▪ Have you considered hazard ratio, landmark PFS and the overall shape of the curve in addition to the mPFS?

▪ Might you need to consider the Duration of Response (DoR) in addition to mPFS when evaluating efficacy?

▪ Does median PFS benefit translate into OS benefit and vice versa?

▪ What does the IRC-assessed outcome show in addition to the investigator-assessed outcome? How different are they?

▪ Are the results statistically significant? How relevant is that in your evaluation?

IRC, independent-review committee; mPFS, median progression-free survival.

Emerging ideas in lung cancer trials


CRUK Cancer Trials Centre, University College London, UK


Challenges for modern advanced cancer trials

HTA, health technology assessment; OS, overall survival; PFS, progression-free survival; QoL, quality of life.Wise J, et al. Drug Discov Today 2018;23:788–801; Ajithkumar TV, Gilbert DC. Clin Oncol (R Coll Radiol). 2017 Dec;29(12):767-769.

`

`````````````````````````````````````````````

Allow crossovers to encourage participation by clinicians and patients?

Satisfying various organisations (regulators, HTA/payers, guideline authors, clinicians, patients)

Bigger trials needed to demonstrate superiority of new therapies when compared against highly effective standards

Payers sometimes want expensive head-to-head comparisons between the same class of drugs to show ‘added value’

Multiple eligibility criteria & high adherence means trial patients are not representative of the real-world

Problem when finding treatment benefits for PFS but not OS; and if there is no clear benefit on QoL, what do you do?

Effect of treatment crossoversProfile 1007: Crizotinib vs standard chemotherapy in previously treated

ALK+ aNSCLC1

▪ Striking results on PFS

▪ Different HTA agencies had a different perspective on these results:

―E.g. NICE acknowledges this effect

― IQWiG (Germany) focused on OS (little mention of PFS)2,3

Median PFS 7.7 vs 3.0 months

1. Shaw AT, et al. N Engl J Med 2013;368:2385–2394; 2. IQWiG Press release 2013. Available at: https://tinyurl.com/v6fjge9. Access January 2020; 3. IQWiG reports. Commission No. A16-41. September 2016.

ALK, anaplastic lymphoma kinase; aNSCLC, advanced non-small cell lung cancer; CI, confidence interval; IQWiG, Institute for Quality and Efficiency in Health Care; HTA, health technology assessment; PFS,

progression-free survival.

What was the challenge?Profile 1007: Crizotinib vs standard chemotherapy in previously treated

ALK+ aNSCLC

ALK, anaplastic lymphoma kinase; aNSCLC, advanced non-small cell lung cancer; HR, hazard ratio; IQWiG, Institute for Quality and Efficiency in Health Care; NICE, National

Institute for Health and Care Excellence; OS, overall survival; PFS, progression-free survival.

IQWiG had a different view despite the large effect on PFS:3

▪ Did not accept the adjusted analyses (requires various assumptions)

▪ From their perspective “Looks like a single arm trial because so many patients ended up receiving crizotinib”

87% of chemotherapy patients switched to crizotinib after progression1

Standard chemotherapy

20.3 months2

Crizotinib

22.8 months2

HR, 1.02, p=0.542

Crossover adjusted analysis1

▪ HR: 0.79 (less mature data); HR: 0.38 (more mature data)▪ Accepted by NICE as supporting evidence but with difficulty

Overall survival

1. NICE. TA422. December 2016; 2. Shaw AT, et al. N Engl J Med 2013;368:2385–2394; 2. IQWiG Press release 2013. Available at: https://tinyurl.com/v6fjge9. Access January 2020.

Potential value of real world evidence (RWE) – large observational studies, electronic health data, disease registries

Two major ongoing initiatives to get greater agreement/standardisation on how regulators, payers and academic/pharma

investigators should collect and analyse RWE alongside randomised trials:

▪ USA FDA DUPLICATE

▪ European Commission Innovative Medicines Initiative

Wise J, et al. Drug Discov Today 2018;23:788–801. FDA, Food and Drug Administration; RWE, real-world evidence; QoL, quality of life.

Pharmaco-epidemiology Effects of drugs in large numbers of people, both efficacy and side effects

Precision/stratified medicine - Finding more effective treatments according to smaller patient subgroups defined by their characteristics, or disease features

Rare/uncommon disorders - Combining single-arm treatment trials with observational controlled data, where large RCTs are not feasible (to have some sort of comparison)

Patient reported outcomes - Using personal electronics, wearable gadgets, & social media, to look at or obtain large amounts of data on patients ‘real life’ use of drugs, including adherence & QoL

RWE when we only have single arm clinical trials: Uncommon disorderUsing electronic health records as a real-world comparator: An example

▪ In ALK+ advanced NSCLC, alectinib (2 single arm trials), to be compared with ceritinib (comparator)

▪ Aim: To compare 3 different statistical methods

– IPTW: Inverse probability of treatment weighting

– PSM: Propensity score matching

– GenMatch

▪ They try to ‘create’ a randomised comparison with balanced patient characteristics

Source of dataNo. of

patients

Alectinib 2 single arm trials 183

Ceritinib (comparator)

Flatiron electronic health records.

250 cancer clinics (~1.5M patients)

Diagnosed 2011–2014, followed to 2016

67

Davies I, et al. Value Health 2017;20(9):A735; Davies J, et al. J Comp Eff Res 2018;7(9):855–865. ALK, anaplastic lymphoma kinase; NSCLC, non-small cell lung cancer; RWE, real-world evidence.

▪ Two statistical methods produced good balance in baseline factors

▪ One method did not

▪ The 2 gave consistent treatment effects

▪ This is one of the problems for decision-makers: which method(s) are best?!

Davies J, et al. Value Health 2017;20(9):A735; Davies J, et al. J Comp Eff Res 2018;7(9):855–865.

Ceritinib

Ceritinib

Alectinib

Alectinib

CI, confidence interval; HR, hazard ratio; NR, not reached; OS, overall survival.

Summary of emerging concepts

▪ More careful consideration should be given to allowing crossovers

▪ Greater scope for using real-world data (RWD) as supporting evidence

▪ Needs more guidance for sponsors on when RWD is or is not appropriate:

– What defines ‘reliable’ RWD?

– How do we analyse RWD robustly?


ALUNBRIG®▼ (brigatinib) PRESCRIBING INFORMATION

Refer to the Summary of Product Characteristics (SmPC) before prescribing.

UK: Adverse events should be reported to the Medicines and Healthcare products Regulatory Agency. Reporting forms and information can be found

at www.mhra.gov.uk/yellowcard. Adverse events should also be reported and additional information on our products is available on request from

Takeda UK Ltd [email protected]

Ireland: Adverse Events should be reported to the Pharmacovigilance Unit at the Health Products Regulatory Authority ([email protected]). Information

about Adverse Event reporting can be found on the HPRA website (www.hpra.ie). Adverse events should also be reported and additional

information on our products is available on request from Takeda UK Ltd [email protected]

Presentation: Brigatinib 180 mg, 90 mg and 30 mg film-coated tablets.Indications: As monotherapy for the treatment of adult patients with anaplasticlymphoma kinase (ALK)-positive advanced non-small cell lung cancer (NSCLC)previously not treated with an ALK inhibitor. As monotherapy for the treatmentof adult patients with ALK-positive advanced NSCLC previously treated withcrizotinib. Dosage and administration: Recommended starting dose is 90 mgonce daily for the first 7 days, then 180 mg once daily. If ALUNBRIG is interruptedfor 14 days or longer for reasons other than adverse reactions, treatment shouldbe resumed at 90 mg once daily for 7 days before increasing to the previouslytolerated dose. If a dose is missed or vomiting occurs after taking a dose, anadditional dose should not be administered and the next dose should be takenat the scheduled time. Treatment should continue as long as clinical benefit isobserved. Dosing interruption and/or dose reduction may be required. Refer toSmPC for full dose adjustments. Paediatric population: No data are available.Elderly patients: Dose adjustment is not required. No available data on patientsaged > 85 years. Hepatic impairment: A reduced starting dose of 60 mg oncedaily for the first 7 days, then 120 mg once daily is recommended for patientswith severe hepatic impairment (Child-Pugh class C). Renal impairment: Areduced starting dose of 60 mg once daily for the first 7 days, then 90 mg oncedaily is recommended for patients with severe renal impairment (eGFR < 30mL/min). Patients with severe renal impairment should be closely monitored fornew or worsening respiratory symptoms (e.g., dyspnoea, cough, etc.) particularlyin the first week. Contraindications: Hypersensitivity to the active substance orto any of the excipients. Warnings and precautions: Refer to SmPC forrecommended dose modifications. Pulmonary adverse reactions: Most occurredwithin the first 7 days of treatment, but they can occur later in treatment.Patients should be monitored for new or worsening respiratory symptoms (e.g.,dyspnoea, cough, etc.), particularly in the first week of treatment. If interstitiallung disease/pneumonitis is suspected, the dose of ALUNBRIG should bewithheld, and the patient evaluated for other causes of symptoms (e.g.,pulmonary embolism, tumour progression, and infectious pneumonia). Dosemodification may be required. Hypertension: Heart rate and blood pressureshould be monitored regularly. Hypertension should be treated according tostandard guidelines. Heart rate should be monitored more frequently in patientsif concomitant use of a medicinal product known to cause bradycardia cannot beavoided. Withhold ALUNBRIG in patients with severe hypertension (≥ Grade 3)until hypertension has recovered to Grade 1 or to baseline. Modify doseaccordingly. Bradycardia: Caution should be exercised when administeringALUNBRIG in combination with other agents known to cause bradycardia. Ifsymptomatic bradycardia occurs, treatment with ALUNBRIG should be withheldand concomitant medicinal products known to cause bradycardia should beevaluated. Upon recovery, the dose should be modified according to SmPC. Incase of life-threatening bradycardia, if no contributing concomitant medicationis identified or in case of recurrence, treatment with ALUNBRIG

should be discontinued. Visual disturbance: Advise patients to report any visualsymptoms. Consider ophthalmologic evaluation/dose reduction for new orworsening severe symptoms. Creatine phosphokinase (CPK) elevation: Patientsshould be advised to report any unexplained muscle pain, tenderness, orweakness. CPK levels should be monitored regularly during ALUNBRIGtreatment. Based on the severity of the CPK elevation, and if associated withmuscle pain or weakness, treatment with ALUNBRIG should be withheld, and thedose modified. Elevations of pancreatic enzymes: Lipase and amylase should bemonitored regularly during treatment with ALUNBRIG. Based on the severity ofthe laboratory abnormalities, treatment with ALUNBRIG should be withheld, andthe dose modified. Hepatotoxicity: Liver function, including AST, ALT and totalbilirubin should be assessed prior to the initiation of ALUNBRIG and then every 2weeks during the first 3 months of treatment. Thereafter, monitoring should beperformed periodically. Based on the severity of the laboratory abnormalities,treatment should be withheld, and the dose modified. Hyperglycaemia: Fastingserum glucose should be assessed prior to initiation of ALUNBRIG and monitoredperiodically thereafter. Antihyperglycaemic medications should be initiated oroptimised as needed. If adequate hyperglycaemic control cannot be achievedwith optimal medical management, ALUNBRIG should be withheld untiladequate hyperglycaemic control is achieved; upon recovery, dose reduceALUNBRIG as per the SmPC or permanent discontinuation may be considered.Lactose: ALUNBRIG contains lactose monohydrate. Patients with rare hereditaryproblems of galactose intolerance, total lactase deficiency, or glucose-galactosemalabsorption should not take ALUNBRIG. Interactions: Avoid use with strongCYP3A inhibitors. Refer to SmPC for dosage modifications for ALUNBRIG ifconcomitant use with a strong CYP3A inhibitor cannot be avoided. Strong andmoderate CYP3A inducers should be avoided. Grapefruit or grapefruit juiceshould be avoided. Coadministration of ALUNBRIG with CYP3A substrates with anarrow therapeutic index should be avoided as ALUNBRIG may reduce theireffectiveness. Co-administration of ALUNBRIG with substrates of P-gp, BCRP,organic cation transporter 1, multidrug and toxin extrusion protein (MATE) 1 and2K may increase their plasma concentrations. Patients should be closelymonitored when ALUNBRIG is co-administered with substrates of thesetransporters with a narrow therapeutic index. Fertility, pregnancy and lactation:Women of reproductive potential should be advised not to become pregnantand to use effective non-hormonal contraception during treatment withALUNBRIG and for at least 4 months following the final dose. Men should beadvised not to father a child during ALUNBRIG treatment. Men with femalepartners of reproductive potential should be advised to use effectivecontraception during and for at least 3 months after the last ALUNBRIGtreatment. No clinical data on the use of ALUNBRIG in pregnant women.ALUNBRIG should not be used during pregnancy unless the clinical condition ofthe mother requires treatment. Breast-feeding should be stopped duringtreatment with ALUNBRIG.

No human data are available on the effect of ALUNBRIG on fertility. Undesirableeffects: Very common (≥1/10): Pneumonia, upper respiratory tract infection,anaemia, anaemia, lymphocyte count decreased, APTT increased, white bloodcell count decreased, neutrophil count decreased, hyperglycaemia,hyperinsulinaemia, hypophosphataemia, hypomagnesaemia, hypercalcaemia,hyponatraemia, hypokalaemia, decreased appetite, headache, peripheralneuropathy, dizziness, visual disturbance, hypertension, cough, dyspnoea, lipaseincreased, diarrhoea, amylase increased, nausea, vomiting, abdominal pain,constipation, stomatitis, AST increased, ALT increased, alkaline phosphataseincreased, rash, pruritus, blood CPK increased, myalgia, arthralgia, bloodcreatinine increased, fatigue, oedema, pyrexia. Common (≥1/100 to <1/10):Decreased platelet count, insomnia, memory impairment, dysgeusia,bradycardia, electrocardiogram QT prolonged, tachycardia, palpitations,pneumonitis, dry mouth, dyspepsia, flatulence, blood lactate dehydrogenaseincreased, hyperbilirubinaemia, dry skin, photosensitivity reaction,musculoskeletal chest pain, pain in extremity, musculoskeletal stiffness, non-cardiac chest pain, chest discomfort, pain, blood cholesterol increased, weightdecreased. Other serious undesirable effects: Pancreatitis. Refer to the SmPC fordetails on full side effect profile and interactions. UK basic NHS price: 30mg x56: £2,450, 30mg x 28: £1,225, 90mg tablets x 28: £3,675, 180mg tablets x 28:£4,900, ALUNBRIG treatment initiation pack: £4,900. Legal classification: POM.Marketing authorisation (MA) numbers: EU/1/18/1264/003,EU/1/18/1264/011, EU/1/18/1264/008, EU/1/18/1264/007, EU/1/18/1264/010,EU/1/18/1264/012. Name and address of MA holder: Takeda Pharma A/S,Dybendal Alle 10, 2630 Taastrup, Denmark. PI approval code:UK/BRIG/2004/0015. Date of Preparation: April 2020.

mailto:[email protected]

http://www.hpra.ie/

Documents

What’s the point of ENDPOINTS? Translating trial outcomes into …€¦ · These slides were presented at a Takeda Oncology sponsored symposium at the 18th British Thoracic Oncology