+ All Categories
Home > Documents > All GERCOR presentations will be available at canceronet

All GERCOR presentations will be available at canceronet

Date post: 07-Jan-2016
Category:
Upload: akando
View: 47 times
Download: 2 times
Share this document with a friend
Description:
Duration of Disease Control (DDC) or Time to Failure of Strategy (TFS) to evaluate a chemotherapy strategy in advanced colorectal cancer (ACC). - PowerPoint PPT Presentation
17
Duration of Disease Control (DDC) or Time to Failure of Strategy (TFS) to evaluate a chemotherapy strategy in advanced colorectal cancer (ACC). 1,2 B Chibaudel , 1,4 C Tournigand, 1 N Perez-Staub, 1 O Bourges, 1 F Maindrault-Goebel, 3,4 T André, 5 G Lledo, 1,4 C Louvet, 6 F Bonnetain, 1,4 A de Gramont (1) Hôpital Saint-Antoine, APHP, université Paris VI, Paris, France (2) GERCOR, Groupe Coopérateur Multidisciplinaire en Oncologie, Paris, France (3) Hôpital La Pitié-Salpétrière, APHP, université Paris VI, Paris, France (4) INSERM U8106 (5) Clinique Jean Mermoz, Lyon, France (6) Unité de biostatistique et de méthodologie FFCD, INSERM U866, Dijon, France ll GERCOR presentations will be available at http://www.canceronet.c
Transcript
Page 1: All GERCOR presentations will be available at canceronet

Duration of Disease Control (DDC)or Time to Failure of Strategy (TFS)

to evaluate a chemotherapy strategyin advanced colorectal cancer (ACC).

1,2B Chibaudel, 1,4C Tournigand, 1N Perez-Staub, 1O Bourges, 1F Maindrault-Goebel, 3,4T André, 5G Lledo, 1,4C Louvet, 6F Bonnetain, 1,4A de Gramont

(1) Hôpital Saint-Antoine, APHP, université Paris VI, Paris, France(2) GERCOR, Groupe Coopérateur Multidisciplinaire en Oncologie, Paris, France

(3) Hôpital La Pitié-Salpétrière, APHP, université Paris VI, Paris, France(4) INSERM U8106

(5) Clinique Jean Mermoz, Lyon, France(6) Unité de biostatistique et de méthodologie FFCD, INSERM U866, Dijon, France

All GERCOR presentations will be available at http://www.canceronet.com/

Page 2: All GERCOR presentations will be available at canceronet

Introduction• Therapeutic strategies are needed when several regimens, surgery,

maintenance or chemotherapy-free intervals are available: fixed sequence of two therapies (1), stop & go or alternated therapies (2,3)

• PFS is not an optimal endpoint to evaluate therapeutic strategies (4): Treatment effect on OS, based on the effect on PFS, is predicted extremely well when patients receive no effective second line therapy, but not in patients receiving subsequent lines of treatment→ PFS offers the direct measure of a single treatment course

• Composite endpoints have been proposed to evaluate efficacy of these strategies:– Duration of Disease Control (DDC): previously used as first

endpoint in several randomized studies (2,3)

– Time to Failure of Strategy (TFS) (5)

• Our goal was to compare these two composite endpoints and evaluate if they could be potential surrogate endpoints for OS

(4) Buyse et al, JCO 2007(5) Allegra et al, JCO 2007

(1) Tournigand et al, JCO 2004(2) Tournigand et al, JCO 2006(3) Maindrault et al, ASCO 2007

Page 3: All GERCOR presentations will be available at canceronet

Methods• Trials :

Pooled analysis of individual patient data from 3 randomized studies evaluating therapeutic strategies

• Groups :The OPTIMOX 1 study was divided in 3 groups according to the number of inclusions per centre, to finally obtain groups of about one hundred patients, like the arms of the other studies

• All endpoints were calculated using the Kaplan-Meier method

• Effect of treatment on PFS, DDC, TFS and OS was quantified through hazard ratios (HRs)

• Coefficient correlation was estimated through a linear regression model

Progression

N=311

N=309

N=104

N=109

N=111

OPTIMOX 1

OPTIMOX 2

C97-3

FOLFOX 4

FOLFOX 7 sLV5FU2

RFOLFOX 7

Reintroduction

R

mFOLFOX 7 sLV5FU2 mFOLFOX 7Reintroduction

mFOLFOX 7 Chemo-Free Interval mFOLFOX 7Reintroduction

R

FOLFIRI 1 FOLFOX 6

FOLFOX 6 FOLFIRI 1

N=98

N=311

N=309

N=104

N=109

N=111

OPTIMOX 1

OPTIMOX 2

C97-3

FOLFOX 4

FOLFOX 7 sLV5FU2

RFOLFOX 7

Reintroduction

R

mFOLFOX 7 sLV5FU2 mFOLFOX 7Reintroduction

mFOLFOX 7 Chemo-Free Interval mFOLFOX 7Reintroduction

R

FOLFIRI 1 FOLFOX 6

FOLFOX 6 FOLFIRI 1

N=98

Page 4: All GERCOR presentations will be available at canceronet

Definitions (1)Term Interval Delays/Holidays Censoring Rules

Sum of PFS

Except for PD at first evaluation after treatment reinitiation

Complete flexible intervals

1) End of study with no PD

2) Addition of new therapeutic agent with no PD

3) R0-R1 surgery not censored

Time to: (1)

1) addition of a new therapeutic agent (2)

2) PD on full therapy

3) PD during CT holiday and followed by no further therapy within 1 month

4) death

Flexible intervals

Except if treatment delay > 1 month after PD

1) End of study and:

- neither PD,

- nor new therapy intervention

2) R0-R1 surgery not censored

(1) Cessation of therapy as a result of toxicity is not considered as an endpoint(2) Agent not included in the original strategy ; endpoint measured at the first of either PD or initiation of new agent

TFS

DDC

Page 5: All GERCOR presentations will be available at canceronet

10 Must-have items

Item Definition

Original strategy

(data per course of treatment)

First cycle of the course Date

Last cycle of the course Date

Reason for cessation Progression

Toxicity

Maintenance or break

Other

Progression disease (PD) PD status No PD (0) / PD (1)

Progression Date

PD on full therapy ? No (0) / Yes (1)

New Therapy (NT)

(agent not included in the original strategy)

NT status No (0) / Yes (1)

First cycle Date

Follow-up Survival status Alive (0) / Dead (1)

Last news / Death Date

12

3

456

78

910

Page 6: All GERCOR presentations will be available at canceronet

General Definitions (2)

Treatmentcourse 1

Uncontrolleddisease

NewAgent

Progression

PFScourse1PFScourse2

Progression

Treatmentcourse 2

DDC = PFScourse1 + PFScourse2

TFS = PFScourse1 + Interval + PFScourse2

Interval

Page 7: All GERCOR presentations will be available at canceronet

Endpoints & Censoring Rules1. End of study with no PD

PD duringinterval

Course 1

TFS

>1month

Course 1

Censored subjects : 0Event : 1

DDC = PFScourse1 + PFScourse2

TFS = PFScourse1 + Interval + PFScourse2

DDC = PFScourse1 + PFScourse2

TFS = PFScourse1

DDC = PFScourse1

TFS = PFScourse1

No PD No PD

DDC / TFS

No PD No PDPD during

interval

No PD No PD

DDC

0

0

0

0

0

1

DDC / TFS

Course 2

Course 1 Course 2

Page 8: All GERCOR presentations will be available at canceronet

Endpoints & Censoring Rules2. End of study with PD

Course 2

PD duringinterval

Course 2>1month

TFS = PFScourse1 + Interval + PFScourse2

TFS = PFScourse1

No PD

DDC / TFS

No PDPD during

interval

PD at firstevaluation

DDC TFS

DDC = PFScourse1

PD at firstevaluation

DDC = PFScourse1

Censored subjects : 0Event : 1

1

1

1

1

Course 1

Course 1

Page 9: All GERCOR presentations will be available at canceronet

Results

• N = 1042 patients• Median potential follow-up : 39.8 months• Median age : 64 years [29-80]

Patients Characteristics

(%)

Page 10: All GERCOR presentations will be available at canceronet

Correlation between PFSand Overall Survival (OS)

PFS is not an optimal endpoint to evaluate therapeutic strategies

Correlation betweenmedian PFS and median OS

Correlation betweenLog HR PFS and Log HR OS

Page 11: All GERCOR presentations will be available at canceronet

Survivals

OPTIMOX 1 OPTIMOX 2 C97-3

FOLFOX4 Oxaliplatin Stop & Go

Oxaliplatin Stop & Go

Complete Stop & Go

FOLFIRIFOLFOX6

FOLFOX6FOLFIRI

N=312 N=308 N=98 N=104 N=109 N=111

PFS 8.9 m 9.0 m 8.7 m 6.4 m 8.5 m 8.0 m

P ; HR 0.65 ; 0.96 0.0015 ; 0.61 0.29 ; 1.17

DDC 9.3 m 9.9 m 12.5 m 9.0 m 11.4 m 9.9 m

P ; HR 0.47 ; 1.065 0.013 ; 0.67 0.92 ; 0.99

TFS 9.5 m 10.6 m 12.9 m 9.9 m 14.2 m 11.0 m

P ; HR 0.76 ; 1.03 0.003 ; 0.62 0.61 ; 0.93

OS 19.2 m 20.9 m 24.5 m 19.3 m 21.5 m 20.5 m

P ; HR 0.40 ; 1.08 0.32 ; 0.85 0.92 ; 0.97

Page 12: All GERCOR presentations will be available at canceronet

DDC vs. TFS

Complete Stop & Go

0 26 52 78 1040

20

40

60

80

100

Survival time (weeks)

FOLFIRI - FOLFOX6

0 26 52 78 1040

20

40

60

80

100

Survival time (weeks)

FOLFOX4

0 26 52 78 1040

20

40

60

80

100

Survival time (weeks)

Oxaliplatin Stop & Go

0 26 52 78 1040

20

40

60

80

100

Survival time (weeks)

OPTIMOX 1 study OPTIMOX 2 study C97-3 study

%

%

%

%

%

%

Oxaliplatin Stop & Go

0 26 52 78 1040

20

40

60

80

100

Survival time (weeks)

FOLFOX6 - FOLFIRI

0 26 52 78 1040

20

40

60

80

100

Survival time (weeks)

Page 13: All GERCOR presentations will be available at canceronet

Distribution

N (%)

DDC = TFS PFS = DDC = TFS

PFS < DDC = TFS

DDC = TFS < PFS

605

11

4

58.0

1.0

0.5

DDC < TFS PFS = DDC < TFS

PFS < DDC < TFS

206

199

20.0

19.0

TFS < DDC TFS < PFS = DDC 17 1.5

59.5%

39%

1.5%

Page 14: All GERCOR presentations will be available at canceronet

Advantages & Drawbacks

TFS

DDC

• Exclusion of uncontrolled disease interval

• Exclusion of inactive second course

→ direct measure of treatment effect

• Statistically significant correlation between median DDC and OS

• Median DDC shorter than median TFS

→ smaller sample size

• Based on pragmatic approach

• Uncontrolled disease interval included

• Active treatment delayed ≥1 month after minimal PD not included

• Extension of duration in case of no evaluation between the end of treatment and the following line

Page 15: All GERCOR presentations will be available at canceronet

Correlation between median endpointsand median Overall Survival (OS)

TFSDDC

Page 16: All GERCOR presentations will be available at canceronet

Correlation between Log Hazard Ratios Endpointsand Log Hazard Ratio Overall Survival (OS)

(logarithmic scale is used for both axes)

TFSDDC

Page 17: All GERCOR presentations will be available at canceronet

Conclusions• PFS is not an optimal endpoint when patients can

receive several lines ot treatment

• DDC and TFS achieve roughly the same results in the studies reviewed here.

• However, DDC could capture better the effect of a treatment in a stop and go strategy or a multiline strategy. It is adaptable to every strategies, requires a marginaly smaller sample size and in some cases will require less time to reach the endpoint

• Validation of these composite endpoints as surrogate for OS will require other statistical tests like the surrogate threshold effect


Recommended