Critical Methological Issues in Recent Randomised Trials Paolo Bruzzi Epidemiologia Clinica IRCCS...
-
Upload
crescenzo-tortora -
Category
Documents
-
view
214 -
download
0
Transcript of Critical Methological Issues in Recent Randomised Trials Paolo Bruzzi Epidemiologia Clinica IRCCS...
Critical Methological Issues in Recent Randomised Trials
Paolo Bruzzi
Epidemiologia Clinica
IRCCS AUO San Martino IST
MSO – ROMEa November 23, 2012
Topics
1. Changes in the methodology of the Phases of cancer trials
2. Efficacy: Do we still need (Large) Randomised Controlled Trials?
3. Perspectives ?
Topics
1. Changes in the methodology of the Phases of cancer trials
2. Efficacy: Do we still need (Large) Randomised Controlled Trials?
3. Perspectives ?
Conventional Methodology
- Phase I: dose increases -> MTD
- Phase II: Uncontrolled -> % Response in metastatic pts with a specific cancer
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Conventional Methodology
- Phase I: dose increases -> MTD
The More the Better?
DOSE –RESPONSE!
PHASE I-II TRIALS
Conventional Methodology
- Phase II: Uncontrolled -> % Response in metastatic pts with a specific cancer
Conventional Methodology
- Phase II: Uncontrolled -> % Response in metastatic pts with a specific cancer
- Response? Direct anticancer effect?
- Metastatic pts? - Activity in pts with less disease burden- Need of repeated biopsies
Modern Methodology
- Phase II trials:
- Biological Endpoints
- Window-of-opportunity studies- Neoadjuvant trials – Locally advanced dis.
- Randomised Controls
Conventional Methodology
- Phase I: dose increases -> MTD
- Phase II: Uncontrolled -> % Response in metastatic pts with a specific cancer
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Topics
1. Changes in the methodology of the Phases of cancer trials
2. Efficacy: Do we still need (Large) Randomised Controlled Trials?
3. Perspectives ?
Conventional Methodology
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Conventional Methodology
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Conventional Methodology
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Conventional Methodology
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Conventional Methodology
- Phase III: Large RCT’s in relatively heterogeneous pts: OS (or EFS)
Reasons for the need of Randomised Control groups
a) Inability to predict individual outcome
b) Inability to predict group outcome
c) Inability to predict effect of treatments based on mechanisms
Reasons for the need of Randomised Control groups
a) Inability to predict individual outcome
b) Inability to predict group outcome
c) Inability to predict effect of treatments based on mechanisms
STILL TRUE!
On the other hand ...
if a new drug,
- with a well-identified molecular target
- which is present in subgroups of cancers in different sites
- produces a strong benefit in one of these cancers,
…it may become ETHICALLY unacceptable to run a randomised trial in other cancers with the same target
Example
Mortality
Tumor X Nil vs A 15% vs 12.5%
N=12000 P = 0.0007
H0 Rejected: A is effective in X
Example
Mortality
Tumor X Nil vs A 15% vs 12.5%
N=12000 P = 0.0007
Tumor Y Nil vs A 15% vs 7.5%
N= 240 P=0.066
H0 not rejected: A not shown effective in y
Prior Information:
Mortality
Tumor X Nil vs A 15% vs 12.5%
N=12000 P = 0.0007
Tumor Y Nil vs A 15% vs 7.5%
N= 240 P=0.066
Prior Information: tumors X and Y are BRAF+
Mortality
Tumor X Nil vs A 15% vs 12.5%
N=12000 P = 0.0007
Tumor Y Nil vs A 15% vs 7.5%
N= 240 P=0.066
Prior Information: tumors X and Y are BRAF+A = Anti BRAF Mortality
Tumor X Nil vs A 15% vs 12.5%
N=12000 P = 0.0007
Tumor Y Nil vs A 15% vs 7.5%
N= 240 P=0.066
Interpretation?
Prior Information: tumors X and Y are BRAF+A = Anti BRAF
I have to plan a trial in the rare tumor Z, which is usually BRAF+, and for which there is no effective treatment
Do I need a randomised control group?
Is it ethically acceptable?
GLEEVEC
CML -> Large RCT
GIST -> Large uncontrolled trial
Chordomas -> Case Series
New paths to drug use Large RCT in a frequent cancer with the
target - Proof of principle – Toxicity
Uncontrolled (but formal) trial(s) in other cancers with the target
Off label use in individual cases with the target
New paths to drug useLarge RCT in a frequent cancer with the
target - Proof of principle – Toxicity
Uncontrolled (but formal) trial(s) in other cancers with the target
Off label use in individual cases with the target
Acceptable? Methodology?
Topics
1. Changes in the methodology of the Phases of cancer trials
2. Efficacy: Do we still need (Large) Randomised Controlled Trials?
3. Perspectives?
3. Perspectives
a) Adaptive trials
b) ‘Personalised’ Medicine
c) New Statistical Approaches
3. Perspectives
Adaptive trials
‘Personalised’ Medicine
New Statistical Approaches
3. Perspectives
a) Adaptive trials
b) ‘Personalised’ Medicine
c) New Statistical Approaches
Radical changes in the way cancer trials are designed and analysed
Three modern revolutions
1950 Randomized Clinical Trial
1985 Evidence Based Medicine
1990 Molecular Medicine
Three modern revolutions
1950 Randomized Clinical Trial
1985 Evidence Based Medicine
1990 Molecular Medicine
Empirical Approach
Preclinical work + Clinical observations
Clinical rationale
Empirical Approach
Preclinical work + Clinical observations
Clinical rationale
CLINICAL STUDIES
INTERPRETATION
Fondamenti della sperimentazione clinica randomizzata
• Protezione da risultati falsamente positivi:– Randomizzazione– Protocolli rigidi/Piano statistico predeterminato– Intention to treat– (doppio cieco)
• Protezione da risultati falsamente negativi– Dimensioni
Fondamenti della sperimentazione clinica randomizzata
• Protezione da risultati falsamente positivi:– Randomizzazione– Protocolli rigidi/Piano statistico predeterminato– Intention to treat– (doppio cieco)
• Protezione da risultati falsamente negativi– Dimensioni
RCT -> EBM in Oncology Golden Age
• Rigid protocols – Drugs – Doses – Cycles– Modifications for toxicity or
progression/relapse
• Generic Selection Criteria– Site (e.g. Stomach)– Histology (ADK vs Lymphoma)– Stage (early vs late)
RCT -> EBM in Oncology Golden Age
• Large and Simple Clinical Trials - Systematic Reviews – Meta-analyses
• Clinical Guidelines/Recommendations – Generic
• Flexibility in pt management not considered
Evidence Based Medicine
≈
Cookbook Medicine?
New Century
• Technological advances – Discoveries– Cellular functions
– Molecular mechanisms/pathways
– Genes/Mutations /Gene Functions
– Genomics/Proteomics/(Metabolomics)
– Targeted Drugs
Patients Heterogeneity
New Approaches!
3. Perspectives
a) Adaptive trials
b) ‘Personalised’ Medicine
c) New Statistical Approaches
a) Adaptive design
FDA’s draft guidance for industry on adaptive design clinical trials
(http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM201790.pdf).
Adaptive design clinical trial
FDA’s Definition:
“… a study that includes a prospectively planned opportunity for modification of one or more specified aspects of the study design and hypotheses based on analysis of data (usually interim data) from subjects in the study”
Why adaptive designs are so attractive?
• Early Responses
• Effects in Subgroups of patients
• Toxicity
It is possible to obtain, during the trial,
crucial information to improve some of its design features
Most conventional trials have an adaptive component
Stopping rules based on interim analyses:
• Toxicity
• Rejection of null-hypothesis
• Futility
Conventional Trials
Few, if any, planned interim analyses
1. Change selection criteria ?
2. Change treatment protocol ?
3. NO Change of the Endpoint
4. NO Change of the Statistical Plan
Adaptive designs
Interim analyses critical in study design
1. Change selection criteria
2. Change treatment protocol
3. Change Endpoint
4. Change Statistical Plan
Adaptive designs
Interim analyses
1. Change selection criteria (target)
2. Change treatment protocol
3. Change Endpoint
4. Change Statistical Plan
3. Perspectives
a) Adaptive trials
b) ‘Personalised’ Medicine
c) New Statistical Approaches
Adaptive designs
Interim analyses
1. Change selection criteria (target)
2. Change treatment protocol
3. Change Endpoint
4. Change Statistical Plan
3. Perspectives
a) Adaptive trials
b) ‘Personalised’ Medicine
c) New Statistical Approaches
New Statistical approaches
• Bayesian statistics (rare tumors and subgroups)
• Changing the null hypothesis (Sobrero & Bruzzi, 2009-2012)
• Statistical methods for uncontrolled trials
New Statistical approaches
• Bayesian statistics (rare tumors and subgroups)
• Changing the null hypothesis (Sobrero & Bruzzi, 2009-2012)
• Statistical methods for uncontrolled trials
Metodologia dei trials clinici: Elementi critici
• Primary Aim(s)
• Endpoint(s)
• Selection of patients
• Study Design
• Randomization
• Assessment of outcome
• Statistics (Statistical plan, ITT analysis)
Metodologia dei trials clinici: Prospettive
• Validita’ Interna (Statistica) = Assenza di bias
• Validita’ Esterna= Implicazioni- Generalizzazioni (Proof of Principle)- Applicazioni (pratica clinica, sanita’, ricerca,…)
Metodologia dei trials clinici: Elementi critici
• Primary Aim(s)• Endpoint(s)• Selection of patients• Study design• Randomization• Assessment of outcome• Statistics (Statistical plan, ITT analysis)
Validita’ Interna (Assenza di Bias)
Industry-Sponsored Trials
• Randomizzazione
• Valutazione outcome
• Statistica
I moderni trial dell’industria, nella stragrande maggioranza, sono studi di altissima qualita’, privi di bias rilevanti
} Tecnicamente perfetti !
Metodologia dei trials clinici: Elementi critici
1. Primary Aim(s)
2. Endpoint(s)
3. Selection of patients
4. Study Design
5. Randomization
6. Assessment of outcome
7. Statistics (Statistical plan, ITT analysis)
Metodologia dei trials clinici: Elementi critici
1. Scopo/i Primario/i
2. Endpoint(s)
3. Selezione dei pazienti
4. Disegno di studio
‘Validita’ Esterna’ =
= Rilevanza Clinica e di Sanita’ Pubblica
1. Scopo
Possibili scopi di un Trial Clinico
Valutare:
- (Attivita’ – meccanismi -> Fase II)
- Efficacy (proof of principle)
- Effectiveness (benefici clinici concreti)
Industry-Sponsored Trials
• Scopo Primario:
Rifiutare l’ipotesi nulla H0: P<0.05
H0: Trattamento Sperimentale (con/senza treatment (with/without standard) identico allo Standard
Significativita’ Statistica
• P = Probabilita’ di osservare, per caso, un differenza grande quanto quella osservata o piu’ grande se i due trattamenti sono identici (H0)
Nota: Nei trials di fase III, H0 spesso non plausibile
(precedenti trials di fase II, in altre malattie/stadi)
Significativita’ Statistica
• P = Probabilita’ di osservare, per caso, un differenza grande quanto quella osservata o piu’ grande se i due trattamenti sono identici (H0)
• Rilevanza Clinica: Beneficio per i pazienti sufficiente per far adottare il trattamento sperimentale come standard (considerando costi, tossicita’, rischi)
Trials Sponsorizzati dalle industrie farmaceutiche
1. Scopo: Trovare una differenza statisticamente significativa
2. Endpoint
3. Criteri di Selezione
4. Disegno di Studio a) Sample Size
b) Collocazione temporale delle analisi
2. Endpoints
Endpoints naturali nei trials di efficacia:
= Scopi del trattamento:
Aumentare
Quantita’
e/o di vita
Qualita’
2. Endpoints
Endpoints naturali nei trials di efficacia:
• Overall Survival
• Quality of Life
• (Qualy)
2. Endpoints
Endpoints spesso usati nei trials di efficacia:
• Endpoints di attivita’
• Endpoints Surrogati non sempre validati
Endpoints Surrogati in oncologia
• Relapse-Free Survival, Disease-Free Survival nella malattia operata
• Progression-Free Survival nella malattia metastatica
Perche’ i trials dell’industria si basano cosi’ fortemente su
endpoints surrogati ?
Perche’ i trials dell’industria si basano cosi’ fortemente su
endpoints surrogati ?
- Per abbreviare il tempo alle analisi ad interim e finale (piu’ eventi) (Basterebbe aspettare)
- Effetto piu’ forte! soprattutto nel periodo iniziale del follow-up - Maggiore potenza e ‘rilevanza clinica’
3. Popolazione in studio
• Pazienti selezionati (es. Eta’)– Compliance– Suscettibili agli effetti del trattamento– Meno sensibili alla tossicita’ – Massimizzare gli effetti del trattamento
Possibilita’ di extrapolare a popolazioni di pazienti differenti?
(Trials adiuvanti nel BC - eta’ mediana : 50years)
4. Disegno di Studio
a) Sample Size
b) Timing delle analisi ad interim e finale
4. Study Design
a) Sample Size : Aumentando le dimensioni dello studio,
differenze Clinicamente poco importanti -> Statisticamente
significative
Es. Molti farmaci ‘targeted’ con effetti sul PFS<3 mesi in vari tumori solidi e p<0.001
HR 0.6 0.7 0.8 0.9 1.0 1.1
MCWE, H1
Conventional Trial
No Difference, H0
HR =0.8Power= 80%635 events
HR=0.86 P=0.05
HR 0.6 0.7 0.8 0.9 1.0 1.1
MCWE, H1
Conventional Trial
No Difference, H0
HR =0.8Power= 95%1050 events
HR=0.89 P=0.05
HR 0.6 0.7 0.8 0.9 1.0 1.1
MCWE, H1
Conventional Trials
No Difference, H0
HR =0.8Power= 95%1050 events HR=0.89
Confidence Limits: From No Effect to MCWE
Treatment effects in trials of targeted drugs in advanced solid tumors
All p-values <0.001 (except one)
0100200300400500600700800900
1000
0 1 2 3 4 5 6 7
Increase in median PFS (months)
N. o
f P
atie
nts
P<0.03
4. Study design
b) Timing delle analisi: Interim Analisi precoci– Effetti precoci molto plausibili (Attivita’)– Effetti maggiori e minore focalizzazione sui
fallimenti a lungo termine (tumori avanzati)– Minore focalizzazione sulla tossicita’ a lungo
termine (malattia precoce)
Typical treatment effect in advanced solid tumors
Interim Analysis
Interim analyses• I metodi di correzione della p usati nella analisi ad interim servono a preservare solo l’errore alfa
• Per vari motivi statistici, gli studi interrotti per interim analisi positive tendono a fornire SOVRASTIME dell’efficacia del trattamento sperimentale
• Queste stime quindi non dovrebbero essere usate per valutazioni cliniche e costi/benefici
Case-study: Aromatase Inhibitors in early breast cancer
Three original papers• Anastrozole alone or in combination with tamoxifen
versus tamoxifen alone for adjuvant treatment of postmenopausal women with early breast cancer: first results of the ATAC randomised trial - Lancet 2002
• A Randomized Trial of Letrozole in Postmenopausal Women after Five Years of Tamoxifen Therapy for Early-Stage Breast Cancer – NEJM Nov. 2003
• A Randomized Trial of Exemestane after two to three years of Tamoxifen therapy in postmenopausal women with primary Breast Cancer – NEJM March 2004
Trials results (1st analysis)
Trial Stage Sample Size
F-up Events DFS red. %
Tam vs Anastrovs A+T
Postmenop 61% N-
84% HR+
6200
(2/3 arms)
2.7 years 696 -17%
(-14%)
Letro vs Placebo
5-yrs Tam
Postm ER+
50% N-
5187 2.4 years 207 (73) - 43%
Tam vs
Exem.
2-3yrs Tam
Postm/ER+? 51% N-
4742 2.5 years 449 (199)
-32%
Effetti - a lungo termine?- su mortalita’?
• Cross-over programmato
• Altre terapie
Diluzione effetto?
• Tecniche statistiche non ITT
Ultima Generazione di trials sponsorizzati dall’industria
• Trials sovradimensionati per garantire P significativa ad effetti moderati
• Analisi concentrate su endpoints surrogati e/o nelle analisi ad interim
• Effetti precoci spesso plausibili, ma benefici e tossicita’ a lungo termine?
• Endpoints Surrogati + Interim Analyses: Sovrastima del beneficio?
Trials sponsorizzati dall’ Industria
• I moderni trials: virtualmente privi di difetti dal punto di vista metodologico e statistico
• Disegnati per massimizzare la probabilita’ di osservare un effetto ‘statisticamente signficativo’ del farmaco sperimentale, a prescindere dalla rilevanza clinica dell’effetto ‘reale’
Trials sponsorizzati dall’ Industria (2)
• Questi trials forniscono una prospettiva distorta sulla reale efficacia dei nuovi farmaci
• Ciononostante, sulla base di questi risultati, spesso le agenzie regolatorie approvano questi farmaci, scaricando il problema degli alti costi x benefici limitati sui sistemi sanitari locali
Soluzioni?
• Trials Indipendenti (?)
• Contrattazione ‘forte’ tra agenzie sanitarie-comunita’ scientifica internazionale e Big Pharma
• Nuovi requisiti per l’approvazione dei farmaci
Proposals of new statistical designs for phase III cancer trials
A. Sobrero, P. Bruzzi 2009-2011
Sintesi della proposta
- Trials di efficacia
- Scopo : rifiutare
H0: Delta < MCWE
(Minimal Clinically Worthwhile Effect)
Example
• Metastatic Colorectal Cancer
• Expected Overall Survival: 24 months
• Experimental treatment: Limited toxicity and costs
• MCWE: increase in OS = 3 months (HR= 0.9)
HR 0.5 0.6 0.7 0.8 0.9 1.0
MCWE, H0
New proposal
APPROVE
Discard
LIMBO
H0: No effect
Issues to discuss
1. Choice of the MCWE
2. P value computation
3. Power/required sample size
4. Analysis and interpretation of the results
5. Adaptive designs/Interim analyses
Conclusione
• La metodologia delle sperimentazioni cliniche e dell’intero processo di sviluppo delle terapie oncologiche e’ destinata a modificarsi radicalmente nei prossimi anni per rispondere alle esigenze che derivano dai progressi nella quantita’ e nel tipo delle conoscenze disponibili
Processo di sviluppo delle nuove metodologie
Esperti + Stakeholders
Cost-efficacy
Bevacizumab (BEV) plus chemotherapy (CT) continued beyond first progression in patients with metastatic
colorectal cancer (mCRC) previously treated with BEV + CT: Results of a randomised phase III intergroup study –
TML (ML18147)
D Arnold1, T Andre2, J Bennouna3, J Sastre4, P Österlund5, R Greil6 E Van Cutsem7, R von Moos8, I Reyes-Rivera9, B Bendahmane10, S Kubicka11
on behalf of the AIO, GERCOR, FFCD, UNICANCER GI, TTD, GEMCAD and AGMT groups
1Hamburg, Germany; 2Paris, France; 3Nantes, France; 4Madrid, Spain5Helsinki, Finland; 6Salzburg, Austria; 7Leuven, Belgium; 8Chur, Switzerland 9South San
Francisco, USA; 10Basel, Switzerland; 11Reutlingen, Germany
Aims and objectives
• The efficacy and safety of BEV continued after first PD has not been investigated in a randomised clinical trial
• TML (ML18147) is the first randomised phase III study to evaluate BEV continued with standard second-line CT in patients with mCRC who progressed after BEV plus standard first-line CT
Statistical considerations• Study initiated as AIO KRK 0504 then transferred to Roche
(after enrolment of 261 patients)
– Primary endpoint changed from PFS to OS
– Sample size increased from 572 to 810 patients
• Designed to detect 30% (HR 0.77) improvement in median OS (90% power, 2-sided 5%) 613 events required for analysis
• OS curves estimated using Kaplan–Meier method, differences assessed using unstratified log-rank tests
– Unstratified Cox regression model used to estimate HR for OS
– Stratified log-rank tests and Cox regression analyses used as supportive analyses
Main eligibility criteria
Inclusion
• Patients ≥18 years with histologically confirmed diagnosis of mCRC
• Eastern Cooperative Oncology Group (ECOG) PS 0–2
• PD (≥1 measurable lesion according to RECIST v1 assessed by investigator, documented by CT or MRI), ≤4 weeks prior to start of study treatment
• Previously treated with BEV plus standard first-line CT, not candidates for primary metastasectomy
Exclusion
• Diagnosis of PD >3 months after last BEV administration
• First-line patients with PFS in first-line of <3 months
• Patients receiving <3 consecutive months of BEV in first-line
CharacteristicCT
(n=411)BEV + CT(n=409)
Male, % 63 65
Age, median years 63 63
ECOG performance status, %012
43525
44515
First-line PFS, %≤9 months>9 months
5644
5446
First-line CT, %Irinotecan-basedOxaliplatin-based
5842
5941
Demographic and baseline characteristics: Randomised patients
Patients were accrued between February 2006 and June 2010
Demographic and baseline characteristics: Randomised patients (cont’d)
aPatient population refers to sequential enrolment of patients in original AIO study and subsequent enrolment in ML18147 when study was transferred to Roche
Characteristic CT
(n=411) BEV + CT
(n=409)
Duration from last BEV dose to randomisation, %
≤42 days >42 days
7723
7723
Patient populationa, %AIOML18147
3268
3268
Liver metastasis only, %NoYes
7129
7327
No. of organs with metastasis, %1>1
3961
3664
Second-line chemotherapy during study: Randomised patients
Second-line CT regimen, % CT
(n=407) BEV + CT
(n=407)
Irinotecan-based CT 43 42
FOLFIRI 14 16
LV5FU2 + CPT11 (Douillard regimen1) 7 7
XELIRI 12 12
Other regimens 10 7
Oxaliplatin-based CT 57 58
FOLFOX4 / mFOLFOX4 18 19
FOLFOX6 13 16
FUFOX 9 6
XELOX 11 14
Other regimens 6 4
1. Douillard et al. Lancet 2000;355:1041–7
Primary endpoint:overall survival
OS: ITT populationO
S e
stim
ate
Time (months)
1.0
0.8
0.6
0.4
0.2
00 6 12 18 24 30 36 42 48
No. at riskCT 410 293 162 51 24 7 3 2
0BEV + CT 409 328 188 64 29 13 4 1
0
CT (n=410)BEV + CT (n=409)
9.8 mo 11.2 mo
Unstratifieda HR: 0.81 (95% CI: 0.69–0.94)
p=0.0062 (log-rank test)
Stratifiedb HR: 0.83 (95% CI: 0.71–0.97)
p=0.0211 (log-rank test)
aPrimary analysis method; bStratified by first-line CT (oxaliplatin-based, irinotecan-based), first-line PFS (≤9 months, >9 months), time from last dose of BEV (≤42 days, >42 days), ECOG performance status at baseline (0, ≥1)
Median follow-up: CT, 9.6 months (range 0–45.5); BEV + CT, 11.1 months (range 0.3–44.0)
Subsequent anti-cancer therapies: Safety population
Subsequent therapy, %CT
(n=409)BEV + CT(n=401)
Patients who received ≥1 subsequent anti-cancer therapy
67.7 68.6
Subsequent anti-cancer therapies
BEV 12.2 11.5
Anti-EGFR 41.3 39.2
Other 50.4 54.9
EGFR: epidermal growth factor receptor
Subgroup analysis of OS: ITT population
aPatient population refers to sequential enrolment of patients in original AIO and subsequent enrolment in ML18147 when study was transferred to Roche. All patients listed under AIO were included in primary analysis
Category Subgroup n HR (95% CI)
All All 819 0.81 (0.69–0.94)
Patient populationa AIO 260 0.86 (0.67–1.11)
ML18147 559 0.78 (0.64–0.94)
Gender Female 294 0.99 (0.77–1.28)
Male 525 0.73 (0.60–0.88)
Age <65 years 458 0.79 (0.65–0.98)
≥65 years 361 0.83 (0.66–1.04)
ECOG performance status 0 357 0.74 (0.59–0.94)
≥1 458 0.87 (0.71–1.06)
First-line PFS ≤9 months 449 0.89 (0.73–1.09)
>9 months 369 0.73 (0.58–0.92)
First-line CT Oxaliplatin-based 343 0.79 (0.62–1.00)
Irinotecan-based 476 0.82 (0.67–1.00)
Time from last BEV ≤42 days 630 0.82 (0.69–0.97)
>42 days 189 0.76 (0.55–1.06)
Liver metastasis only No 592 0.81 (0.67–0.97)
Yes 226 0.79 (0.59–1.05)
No. of organswith metastasis
1 307 0.83 (0.64–1.08)
>1 511 0.77 (0.64–0.94)
HR 0 1 2
Subgroup analysis of PFS (ITT population)
aPatient population refers to sequential enrolment of patients in original AIO and subsequent enrolment in ML18147 when study transferred to Roche. All patients listed under AIO were included in primary analysis
Category Subgroup n HR (95% CI)
All All 819 0.68 (0.59–0.78)
Patient populationa AIO 260 0.65 (0.51–0.84)
ML18147 559 0.69 (0.58–0.82)
Gender Female 294 0.85 (0.67–1.07)
Male 525 0.60 (0.50–0.72)
Age <65 years 458 0.66 (0.55–0.80)
≥65 years 361 0.71 (0.57–0.87)
ECOG PS 0 357 0.59 (0.48–0.74)
≥1 458 0.76 (0.63–0.92)
First-line PFS ≤9 months 449 0.75 (0.62–0.90)
>9 months 369 0.58 (0.47–0.72)
First-line CT Oxaliplatin-based 343 0.68 (0.55–0.85)
Irinotecan-based 476 0.67 (0.56–0.81)
Time from last BEV ≤42 days 630 0.72 (0.61–0.85)
>42 days 189 0.56 (0.41–0.75)
Liver metastasis only No 592 0.68 (0.57–0.80)
Yes 226 0.68 (0.52–0.89)
No. of organswith metastasis
1 307 0.74 (0.59–0.94)
>1 511 0.64 (0.53–0.77)
HR 0 1 2
Summary• BEV + standard second-line CT, crossed over from BEV +
standard first-line CT, significantly prolongs OS and PFS– OS
• Median: BEV + CT 11.2 months, CT 9.8 months• HR: 0.81 (95% CI: 0.69–0.94), p=0.0062a
– PFS• Median: BEV + CT 5.7 months, CT 4.1 months• HR: 0.68 (95% CI: 0.59–0.78), p≤0.0001a
• Findings from subgroup analyses for OS generally consistent with overall population– Treatment effect according to gender appeared to be different; however,
treatment-gender interaction test was not statistically significant
• Differences in best overall response rate not statistically significant; low response rate in both treatment groups
• AEs not increased when continuing BEV beyond PD; AE profile consistent with previous findings
Conclusions
• First randomised clinical trial that prospectively investigated the impact of continued VEGF inhibition with BEV beyond first progression
• Study confirms that continuing BEV beyond first progression while modifying CT is beneficial for patients with mCRC and leads to a significant improvement in OS and PFS
• This provides a new second-line treatment option for patients who have been treated with BEV + standard CT in first line while maintaining an acceptable safety profile
• Findings indicate a potential new model for treatment approaches through multiple lines and this is currently being investigated in other tumour types
OS: ITT populationO
S e
stim
ate
Time (months)
1.0
0.8
0.6
0.4
0.2
00 6 12 18 24 30 36 42 48
No. at riskCT 410 293 162 51 24 7 3 2
0BEV + CT 409 328 188 64 29 13 4 1
0
CT (n=410)BEV + CT (n=409)
9.8 mo 11.2 mo
Unstratifieda HR: 0.81 (95% CI: 0.69–0.94)
p=0.0062 (log-rank test)
Stratifiedb HR: 0.83 (95% CI: 0.71–0.97)
p=0.0211 (log-rank test)
aPrimary analysis method; bStratified by first-line CT (oxaliplatin-based, irinotecan-based), first-line PFS (≤9 months, >9 months), time from last dose of BEV (≤42 days, >42 days), ECOG performance status at baseline (0, ≥1)
Median follow-up: CT, 9.6 months (range 0–45.5); BEV + CT, 11.1 months (range 0.3–44.0)
Long term benefit?
Long term benefit?
Long term benefit?
2 mos
2-3 months 3.5 months
Clinically Worthwhile Effect?
• Median follow-up: 3.8 vs 2.3 months
• HR = 0.37 (0.26-55): Meaningless!
• Long term benefit: None (Mature data?)
• Increase in median survival: ≈ 2 months
• Average benefit:< 2 months– 30% of the pts: 3 months– 30% of the pts: 2 months– 40% of the pts: 0
} Clinically worthwhile?
Conclusions• The medical research community should rethink the
terms of cooperation with industry in clinical trials, taking into account a wider clinical and public health perspective.
Conclusions • The medical research community should rethink the
terms of cooperation with industry in clinical trials, taking into account a wider clinical and public health perspective.
• Our health systems risk bankruptcy for the
skyrocketing costs of drugs that were developed on their own patients, using strategies that ignore the patients’ needs and priorities.
Conclusions • The medical research community should rethink the
terms of cooperation with industry in clinical trials, taking into account a wider clinical and public health perspective.
• Our health systems risk bankruptcy for the skyrocketing
costs of drugs that were developed on their own patients, using strategies that ignore the patients’ needs and priorities.
• Governments, health systems, and regulatory agencies must identify new paths for drug development and set new standards for drug approval