Rafal Dziadziuszko, MD, PhD

University of Colorado Cancer Center, Aurora, CO, USA

Medical University of Gdansk, Poland

Pharmacogenomic markersin EGFR-targeted therapy

of lung cancer

EMEA Workshop on Biomarkers, 15 December 2006

Boyle et al., Ann Oncol 2005

Cancer mortalityin the European Union; 2004

LUNG 20%

COLON & RECTUM 12%

STOMACH 8%

BREAST 8%PROSTATE 5%

LYMPHOMAS 4%

LEUKAEMIAS 3%

OTHER 40%

• Standard chemotherapy providesmodest survival benefit at the expenseof significant toxicity and costs

• Survival rates from lung cancer almostunchanged for decades

• Significant improvement from targeted therapiesin other solid tumors (breast cancer, renalcancer, GIST) and haematologic malignancies

Rationale for targeted therapyof lung cancer

• Orally available EGFR tyrosine kinase inhibitors(TKIs: gefitinib, erlotinib, lapatinib, canertinib, HKI 272)

• Anti-EGFR monoclonal antibodies(cetuximab, panitumumab, matuzumab, pertuzumab)

Classes of EGFR inhibitorsunder clinical development

• Phase I studies: relatively good tolerance; dose limiting toxicities: skin rash and diarrhea

• Phase II monotherapy studies in non-small celllung cancer (NSCLC): ~10-20% response ratesand ~40% disease control rates in pretreatedpatients

Gefitinib and erlotinib: findings from early clinical studies

• No advantage of EGFR TKIs combined withchemotherapy in unselected NSCLC patientsin the first-line treatment (four phase III studies; >4.000 patients)

• Significant survival benefit (HR=0.70) with erlotinibmonotherapy vs placebo in unselected patients relapsedafter one or two lines of chemotherapy (BR.21)

• Insignificant survival benefit (HR=0.89) with gefitinibmonotherapy in a similar setting (ISEL)

Gefitinib and erlotinib: findings from phase III studies

BR.21: survival

Shepherd et al., NEJM, 2005

HR=0.70 (0.58–0.85) Stratified log-rank p<0.001

100

80

60

40

20

0

Perc

enta

ge

0 6 12 18 24 30

ErlotinibPlacebo

At riskErlotinib 488 255 145 23 4 0 Placebo 243 107 50 9 0 0

Time (months)

• Never-smokers (RRs ~ 20-30%)

• Asian ethnicity (RRs ~ 30%)

• Female gender (RRs ~ 15-20%)

• Adenocarcinoma (RRs ~ 10-20%)

Clinical markers of increasedresponsiveness to EGFR TKIs

BR.21: Forest plot of survival by subsets

Erlotinib:placeboPS 0–1PS 2–3

MaleFemale

<65 years≥65 years

AdenocarcinomaSquamous-cell carcinoma

Other histologyPrior weight loss <5%

Prior weight loss 5–10%Prior weight loss >10%

Never-smokerCurrent/ex-smoker

1 prior regimen2+ prior regimens

0 1 2 3 4HR Tsao et al., NEJM, 2005

Biologic selection to EGFR TKIs

GGCGGGCCAAACTGCTG

EGFR gene copy numberby FISH

EGFR gene mutations

EGFR protein expressionby IHC

EGFR FISH

13.8%Gene Amplification

17.0%

27.3%

2.2%

24.1%

15.7%

EGFR (%)

High Polysomy

Low Polysomy

High Trisomy

Low Trisomy

Disomy

PATTERN

ISEL STUDY

Hirsch et al., J Clin Oncol 2006

16% vs. 3%

20% vs. 2%

26% vs. 11%

36% vs. 3%

RR FISH+ vs. FISH-

31%

45%

32%

32%

% FISH Positive

0.50*(0.25-0.97)

Gefitinib500 mg/d

82Hirsch et al.SWOG 0126

0.44**(0.23-0.82)

Erlotinib150 mg/d

125Tsao et al.BR.21

0.61**(0.36-1.03)

Gefitinib250 mg/d

370Hirsch et al.ISEL

Gefitinib250 mg/d

Drug

0.44* (0.23-0.82)

102Cappuzzo et al.

HR (95% CI)

NAuthor

EGFR TKIs studies: impact of gene copy number by FISH

*HR for FISH+ vs. FISH- subsets; all patients treated with gefitinib

**HR for EGFR TKI vs. placebo in FISH+ patients

Survival according to EGFR genecopy number – BR.21 and ISEL

MONTHS MONTHS

HR=0.44 (0.23, 0.82)P=.008

ISEL FISH + BR.21 FISH +

HR=0.61 (0.36, 1.04)P=.07

20

40

60

80

100

0 4 8 12 16MONTHS MONTHS

20

40

60

80

100

0 6 12 18 3024

HR=0.85 (0.48, 1.51)P=.59

BR.21 FISH -ISEL FISH -

HR=1.16 (0.81, 1.64)P=.42

20

40

60

80

100

0 4 8 12 16

20

40

60

80

100

0 6 12 18 3024

Surv

ival

, %

Surv

ival

, %

ISEL FISH interaction test P=.04 • BR.21 FISH interaction test P=.10

Gefitinib Placebo

GefitinibPlacebo

Erlotinib Placebo

Erlotinib Placebo

Tsao et al, NEJM 2005; Hirsch et al., J Clin Oncol 2006

IHC and EGFR status: scoring systemScore=0

Score=300 Score=400

Score=200

EGFR POSITIVE: 62/100 pts=62%

N= 166 5

(3.0%)

N=80 3

(3.8%)

N=17 1

(5.6%)

N=69 1

(1.5%)EGFR -

N=348 38

(10.9%)

N=106 12

(11.3%)

N=84 13

(13.4%)

N=158 13

(8.2%)EGFR +

ORR (%)ORR (%)ORR (%)ORR (%)

TOTALBR.21IDEALISELEGFR Status

Response according to EGFR protein expression (IHC)

BR.21: Survival accordingto EGFR protein expression

Interaction P = 0.25

100

80

60

40

20

0

Perc

enta

ge

0 6 12 18 24 30

At riskErlotinib117 71 43 5 5 0 Placebo 67 23 12 5 0 0

100

80

60

40

20

0

Perc

enta

ge0 6 12 18 24 30

At riskErlotinib 93 42 22 8 3 0 Placebo 48 24 14 3 0 0

Months Months

ErlotinibPlacebo

Log-rank: p=0.02HR=0.68 (0.49, 0.95)

ErlotinibPlacebo

Log-rank: p=0.70HR=0.93 (0.63, 1.36)

EGFR+ EGFR–

Tsao et al., NEJM 2005

EGFR gene mutations

747-750

L858G719

TM K DFG Y Y Y Y

Autophosphorylationdomain

Tyrosine kinaseLigand binding domain

K R H DFGGXGXXG L L Y

718 745 776 835 858 861 869 964

18 19 20 21 22 23 24

757-750

Exon:

Paez:

Lynch:

Pao:

Mutacje punktoweDelecje

719 858

Pao et al., PNAS 2004

NS54% vs. 5%17%Gefitinib250 mg/d

89Cappuzzoet al.

0.16*(0.05-0.52)

64.7% vs. 13.7%

18.9%Gefitinib250 mg/d

90Han et al.

60% vs. 8.8%

82% vs. 11%

83% vs. 10%

RR Mut+ vs. Mut-

12%

59%

56%

% Mut+

0.27*(0.13-0.53)

Gefitinib250 mg/d

66Takano et al.

0.32*(0.12-0.91)

Gefitinib250 mg/d

83Cortes-Funeset al.

Gefitinib250 mg/d

Drug

0.34*(0.12-0.99)

59Mitsudomi et al.

HR (95% CI)

NAuthor

*Mut+ vs. mut- subsetsNS - non significant

Retrospective studies: impact of EGFR mutations

Prospective studies: impact of EGFR mutations

NR1.77

(0.25-0.97)

46% vs. 10%72% vs. 55%

18%10%

Gefitinib250 and500 mg/d

79312

Bell et al.IDEALINTACT

53% vs. 18%

37.5% vs. 2.6%

16% vs. 7%

RR Mut+ vs. Mut-

12.7%

12%

22.6%

% Mut+

NRGefitinib250 mg/d

215Hirsch et al.ISEL

NR (NS)Erlotinib150 mg/d

228Eberhardt et al.TRIBUTE

Erlotinib150 mg/d

Drug

0.77 (0.40-1.50)

197Tsao et al.BR.21

HR (95% CI)

NAuthor

NR – not reported; NS – non significant

BR.21: Survival accordingto EGFR mutations

NErlotinib 21 11 5 1 1 0 Placebo 19 10 5 1 0 0

Log-rank: p=0.13HR=0.73 (0.49, 1.10)

Interaction test, P= 0.97

NErlotinib 93 59 34 9 1 0Placebo 44 18 11 6 0 0

Wild-type EGFR100

80

60

40

20

00 6 12 18 24 30

MONTHS

ErlotinibPlacebo

Mutant EGFR100

80

60

40

20

0

MONTHS

ErlotinibPlacebo

Log-rank: p=0.45HR=0.77 (0.40, 1.50)

0 6 12 18 24 30

SUR

VIVA

L PR

OB

AB

ILIT

Y

Tsao et al., NEJM 2005

Prognostic value of EGFR mutationsin advanced NSCLC

I

II I

I I

I I I II

I I II I I

II IIIIIIIIIIIIIIII IIIIIIIIII IIIIIIIIII III I I IIII II IIIIIIIII I I I III I I I

I

IIII I I II I I II I I II II I I I

1839IL/0014 and 1839IL/0017FIGURE FS5.EGFR MUTATION SURVIVAL: KAPLAN MEIER PLOT

POPULATION : INTENTION-TO-TREAT

TICK MARKS INDICATE CENSORED OBSERVATIONS

GROUP IRESSA & EGFR MUT. + PLACEBO & EGFR MUT. +IRESSA & EGFR MUT. - PLACEBO & EGFR MUT. -

PROPORTION EVENT FREE

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

SURVIVAL TIME (MONTHS)0 4 8 12 16 20 24

Prop

ortio

n Ev

ent F

ree

EGFR mutation-positive (chemotherapy & gefitinib)

EGFR mutation-negative (chemotherapy & gefitinib)

EGFR mutation-positive (chemotherapy & placebo)

EGFR mutation-negative (chemotherapy & placebo)

Overall Survival (months)

EGFR Mutation Status and Overall Survival INTACT

Bell et al., Clin Cancer Res, 2006

Survival vs. EGFR mutation type

Jackman et al., Clin Cancer Res, 2006

• Several biomarkers identified (gene copynumber, EGFR protein expression, EGFR mutations, serum proteomics)

• None routinely used for patient selection

• Clinical trials in selected patient populationsor stratified for these markers ongoing

Current status of biomarkers for selectionof NSCLC patients to EGFR TKIs

• Poor translational components of clinical studies(none prospectively enriched or stratifiedfor biomarkers)

• Neglecting differences in biology accordingto demographic and clinical characteristics(i.e. smoking history, ethnicity)

• Poor standarization and validationof technologies for biomarker assesment

What went wrong with biomarkers in clinicaldevelopment of EGFR TKIs in NSCLC?

EGFR TKI preclinical studies in Colorado

Sensitive Resistant

ResistantSensitive

Clinical trial design issues

Prognostic markerAssociates with maineffect regardlessof treatment

May be used for risk-stratified treatment

Not suitable for targeted-therapy trialdesigns

Predictive markerInteraction withtreatment

Appropriate for targeted-therapytrial designs

Crowley J., Taormina IASLC Meeting, 2006

Targeted therapy clinical trial designs

• All-comers design: Randomize everyone, measure marker / stratify by marker

• Targeted design: Randomize positive patients only

• Strategy design: Randomize to strategy based on marker

Register Measure marker Randomize

Register Measure marker Randomize M+

Register Measure marker Randomize

AB

BA

Tx basedon marker

Tx not basedon marker

A or B

A or B

M+ M-

M+

M+

Crowley J., Taormina IASLC Meeting, 2006

• Incorporation of biomarker studies early in preclinical and clinical development

• Understanding of biomarker significance for disease biology (prognostic vs. predictive)

• Better standarization and validationof technologies for biomarker assesment

Future directions