Nicolas PENEL Andrew KRAMAR Centre Oscar Lambret , Lille, France

Activity endpoints in soft tissue sarcoma phase II trials

Quality and correlations with overall survival

Nicolas PENEL Andrew KRAMAR

Centre Oscar Lambret, Lille, France

Primary endpoints

Critical choiceFew promising drugsPromising drugs failed to improve overall survival

Q1: What is the quality of reported primary endpoints ?Q2: What are the correlation between activity endpoints

and overall survival ? Q3: What are the distribution of activity endpoints in

positive and negative trials ?

Method - GeneralCriteria of selection of trials:

Phase II trials Chemotherapy (single agents or combination) or

moleculary targeted agents After failure/intolerance to doxorubicin Full reports issued between January 1999 and August

2011 English-written reports

Systematic analysis of 53 trials 77 strata

Q 1: What is the quality of reported primary endpoints ?

Nature of primary endpointPrimary endpoint StudiesAbsence of defined primary endpoint 12/53 (22%)Defined primary endpoint

Best objective response 21/53 (39%)Progression-free rate at 6 months

7/53 (13%)

Progression-free rate at 3 months

4/53 (6%)

Progression-free rate at 4 months

2/53 (4%)

Time to progression 2/53 (2%)Progression-free survival 1/53 (2%)Rate of “remission” 1/53 (2%)

Precise definition of primary endpoint

Primary endpoint Studies

Absence of defined primary endpoint 12/53 (22%)

Defined primary endpoint

Not precisely defined 5/53 (10%)

Precisely defined 36/53 (68%)

Design/Methodology Key-issues Categories StudiesDesign Stratification 6/53 (11%)

Randomization 3/53 (5%)Central radiological review

Yes 2/53 (3%)No 51/53 (97%)

Stastical hypothesis

Yes 41/53 (77%)No 12/53 (33%)

interpretation of the results Results StudiesPromising drug 7/77 (10%)Inactive drug 38/77 (50%)Ininterpretable results because of absence of statistical hypothesis

12/77 (15%)

Ininterpretable results because of absence of reported data

20/77 (25%)

Q 2: What are the correlation between activity endpoints and overall survival ?

0

2

4

6

8

10

12

14

16

18

20

22

24

Med

ian

Ove

rall

Sur

viva

l

0 .1 .2 .3 .4 .5 .6Best Objective Response Rate

Trial Point L-Fit

Poor correlation between mOS and BORR: p=0.058

Good correlation between 6-month PFR and OS (p=0.005)

Endpoints strata R pBest objective response 48 0.276 0.058Best tumour control rate 48 0.276 0.2573-month progression-free rate

39 0.466 0.002

6-month progression-free rate

41 0.430 0.005

Median progression-free survival

45 0.402 0.006

Q 3: Distributions of endpoints in cases of active or inactive drugs?

Categories Inactive drug Active drug p EORTC STBSG Definition

3-month PFR <39% Or 6-month PFR <14%

3-month PFR ≥39% & 6-month PFR ≥14%

-

Strata 33 26 3-month PFR (%) Median (range)

26.0 (0.0-42.0) 48.0 (40.0-75.0) (0.0001)

6-month PFR (%) Median (range)

9.0 (0.0-39.0) 30.0 (15.0-55.0) (0.0001)

BORR (%) Median (range)

0.0 (0.0-19.0) 10.0 (0.0-53.0) 0.0001

BTCR (%) Median (range)

29.0 (6.0-50.0 43.0 (14.0-77.0) 0.00001

Median PFS (months) Median (range)

1.9 (0.2-3.03) 3.35 (1.8-12.0) 0.0001

Median OS (months) Median (range)

10.3 (4.9-22.8) 11.8 (4.9-22.4) 0.463

Q1: Key-findingsNumerous (7 ≠) and not suitable primary endpoints (BORR)Poorly defined endpoint (32% of the studies)Absence of central radiological review (98% of the studies)Absence of statistical hypothesis (33%)Ininterpretable results (40%)

Q2: Endpoints possibly correlated with OS

3-month progression free rate6-month progression free rate Median progression-free survival

Q3: Current definition of active drugs

Using current definitions of active/inactive drugs:

All primary endpoints are statistically higher with « active drugs »

But

OS was not statistically different in active compared to inactive drugs

ConclusionBetter definition the primary endpointRole of the central radiological reviewStatistical hypothesis based on primary endpoint Endpoints correlated with OS (PFR3 , PFR6 and PFS)

But Current definitions of active drug failed to identify

drugs able to improve the OSWe have to refine the thresholds of PFR3 and PFR6

defining active drugs

Thank your for your attention