DDMoRe Model Repository

Overview
Files
History
Model Definition
Estimation Steps

Short description:

Drug-disease model of tumor size and overall survival in metastatic colorectal cancer patients treated with cetuximab administered weekly or every second week. The drug-disease model was built in order to evaluate the effect of cetuximab on changes in tumor size. This model is a modified version of the Claret Tumor Growth inhibition model: it uses the K-PD approach with exposure expressed as Area Under the Curve (AUC). Also, KRAS mutation is introduced as covariate for resistance development.

Format:	PharmML (0.6.1)
Related Publication:	Drug-disease model of tumor size and overall survival in metastatic colorectal cancer patients treated with cetuximab administered weekly or every second week. P.Girard, T. Brodowicz, A. Kovar, B. Brockhaus, M. Zühlsdorf, M. Schlichting, A. Munafo, R. Esser, E. Van Cutsem, J. Tabernero Euroean Journal of Cancer, 6/2016, Volume 49 Affiliation: Merck Institute for Pharmacometrics, Lausanne (Switzerland) Medical university of Vienna, Vienna (Austria) Merck KGaA, Darmstadt (Germany) University hospital Gathuisberg, Leuven (Belgium) Vall d'Hebron University Hospital, Barcelona (Spain) Abstract: Background: Cetuximab, administered in a standard weekly regimen (q1w) with first-line chemotherapy (CT) can improve tumor response and overall survival (OS) in metastatic colorectal cancer (mCRC) pts. Cetuximab (400 or 500 mg/m2) every second week (q2w) was safely administered with reported activity in mCRC pts. Tumor size (response) and OS were further studied in a drug-disease model. Material and Methods: A pharmacokinetic (PK)-disease model of tumor size was constructed (using NONMEM7.2) in a pooled analysis of mCRC pts from the 502 (EVEREST, second-line CT + cetuximab q1w) study, and 045 and CECOG/CORE.1.2.002 (CORE) studies (both first-line CT with either cetuximab q1w or q2w). The model estimated changes in tumor size from baseline for early tumor shrinkage (ETS, % change in tumor size at wk 8) and time to tumor (re)growth (TTG, time to smallest tumor size). Covariates were examined including the effects of cetuximab q1w and q2w regimens. ETS and TTG were tested as predictors of OS using Cox models. Results: 369 pts from the 3 studies provided 3821 PK, 2053 tumor size and 233 death observations. Tumor size model fit was very good. Excluding study 502 (irinotecan-refractory pts), covariate analysis identified study (baseline tumor size was 32% lower in the CORE study; p < 0.01) and KRAS status (treatment was 43% less effective in KRAS mutated tumors; p < 0.01) as significant. TTG (p = 0.12) and ETS (p = 0.28) were not markedly different between the q2w and q1w regimens, were positively correlated (Spearman’s rho=0.64), and were identified as individual predictors of OS. For ETS and TTG, pts were grouped by low (L), intermediate (I), and high (H) terciles. In pts with KRAS wild-type tumors, OS was longer in the H-TTG (27 to 118 wk, median OS 31.4 mth) vs L-TTG group (0 to 11 wk, median OS 15.1 mth), or H-ETS (?66 to ?22%, median OS 25.9 mth) vs L-ETS group (?9 to +2%, median OS 12.8 mth). In a Cox model excluding study 502 (n = 215, death=93), ECOG PS grouped as fully active or restrictive activity (RA, p = 0.038), TTG (p < 0.001) and baseline tumor size (p < 0.001) were independent predictors of OS. In a model including ETS, ECOG PS RA (p = 0.0044), ETS (p = 0.013) and baseline tumor size (p = 0.0055) were independent predictors of OS. Treatment regimen was not significant. Conclusions: In this analysis changes in tumor size from baseline and OS were not significantly modified by the cetuximab dosing regimen in mCRC pts. The model identified TTG and ETS as predictors of OS
Contributors:	Nadia Terranova, Kheizurane_ElMekki

Context of model development:	Disease Progression model;
Long technical model description:	'- Cetuximab was administered either in a standard weekly (q1w) or every second week (q2w) regimen. - The drug-desease model was built in order to evaluate the effects of cetuximab regimens on changes in tumor size. ;
Model compliance with original publication:	Yes;
Model implementation requiring submitter’s additional knowledge:	No;
Modelling context description:	Drug-disease model of tumor size in metastatic colorectal cancer patieents from two clinical studies;
Modelling task in scope:	estimation;
Nature of research:	Early clinical development (Phases I and II);
Therapeutic/disease area:	Oncology;

Validation Status:	Annotations are correct.
Certification Comment:	This model is not certified.

Mandatory file
- Executable_Girard_2013_TGI.xml

Additional Files

Model owner: Nadia Terranova
Submitted: Dec 11, 2015 4:57:30 PM
Last Modified: Jul 22, 2016 5:32:49 PM

Revisions

Version: 9
- Submitted on: Jul 22, 2016 5:32:49 PM
- Submitted by: Nadia Terranova
- With comment: Updated model annotations.
Version: 7
- Submitted on: Jul 22, 2016 5:02:05 PM
- Submitted by: Nadia Terranova
- With comment: Updated model annotations.
Version: 4
- Submitted on: Dec 11, 2015 4:57:30 PM
- Submitted by: Kheizurane_ElMekki
- With comment: Edited model metadata online.

Independent variable T

Function Definitions

combinedError2 (additive, proportional, f) = \sqrt{({proportional}^{2} + ({additive}^{2} \times f^{2}))}

Structural Model sm

Variable definitions

\frac{dCETUX}{dT} = (- KE \times CETUX)

RESISTANCY = exp ((- KR \times T))

\frac{dTS}{dT} = ((KS \times TS) - (((KD \times RESISTANCY) \times CETUX) \times TS))

Initial conditions

CETUX = 0

TS = TS0

Variability Model

Level	Type
DV	residualError
ID	parameterVariability

Covariate Model

Continuous covariate KRAS

Continuous covariate STDY

Parameter Model

Parameters

POP_TS0

;

POP_KS

;

POP_KD

;

POP_KE

;

POP_KR

;

RUV_ADD

;

RUV_PROP

;

BETA_TS0_STDY

;

BETA_KD_KRAS

;

OMEGA_TS0

;

OMEGA_KD

;

OMEGA_KE

;

OMEGA_KR

;

OMEGA_KS

;

ETA_TS0 \sim N (0.0, OMEGA_TS0)

— ID

ETA_KD \sim N (0.0, OMEGA_KD)

— ID

ETA_KE \sim N (0.0, OMEGA_KE)

— ID

ETA_KR \sim N (0.0, OMEGA_KR)

— ID

ETA_KS \sim N (0.0, OMEGA_KS)

— ID

EPS_Y \sim N (0.0, 1.0)

— DV

log (TS0) = (log (POP_TS0) + ((STDY \times BETA_TS0_STDY) + ETA_TS0))

log (KD) = (log (POP_KD) + ((KRAS \times BETA_KD_KRAS) + ETA_KD))

log (KE) = (log (POP_KE) + ETA_KE)

log (KR) = (log (POP_KR) + ETA_KR)

KS = (POP_KS + ETA_KS)

Covariance matrix for level ID and random effects: ETA_KE, ETA_KD

(\begin{matrix} 1 & 1.72 \\ 1.72 & 1 \end{matrix})

Observation Model

Observation Y
Continuous / Residual Data

Parameters

Y = (TS + (combinedError2 (RUV_ADD, RUV_PROP, TS) \times EPS_Y))

Estimation Steps

Estimation Step estimStep_1

Estimation parameters

Fixed parameters

OMEGA_KS = 0

Initial estimates for non-fixed parameters

$POP_TS0 = 11$
$POP_KS = 0.0049$
$POP_KD = 5.0E-4$
$POP_KE = 1.46$
$POP_KR = 0.121$
$RUV_ADD = 0.463$
$RUV_PROP = 0.16$
$BETA_TS0_STDY = - 0.404$
$BETA_KD_KRAS = - 1.08$
$OMEGA_TS0 = 0.4$
$OMEGA_KD = 1.22$
$OMEGA_KE = 3.3$
$OMEGA_KR = 0.944$

Estimation operations

1) Estimate the population parameters

Algorithm SAEM

Step Dependencies

estimStep_1

DDMODEL00000107: Girard_2013_oncology_cetuximab_TGI

Revisions

Function Definitions

Structural Model sm

Variability Model

Covariate Model

Parameter Model

Observation Model

Observation YContinuous / Residual Data

Estimation Steps

Estimation Step estimStep_1

Estimation parameters

Estimation operations

Step Dependencies

Observation Y
Continuous / Residual Data