J.G. Coen van Hasselt(1,2), Jan H.M. Schellens(1,2,3), Melvin R. Mac Gillavry(4), Jos H. Beijnen(1,2,3), Alwin D.R. Huitema(1,2)
1. Department of Clinical Pharmacology, The Netherlands Cancer Institute, Amsterdam, The Netherlands. 2. Department of Pharmacy & Pharmacology, Slotervaart Hospital/Netherlands Cancer Institute, Amsterdam, The Netherlands. 3. Department of Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands. 4. Department of Cardiology, Slotervaart Hospital, Amsterdam, The Netherlands
IntroductionTrastuzumab treatment is associated with occurrence of cardiac toxicity, for which monitoring of the left ventricular ejection fraction (LVEF) is indicated. The performance of the currently used monitoring protocol as defined in the summary of product characteristics (SPC) is however unknown.
ObjectivesThe objectives of this analysis were to: i) quantitatively evaluate the currently used cardiac monitoring strategy, as defined in the SPC; ii) suggest further improvements in cardiac monitoring strategies by inclusion of adaptive properties to the current monitoring protocol; iii) demonstrate how a model-based framework can be applied to evaluate repeated measurement (cardiac) monitoring protocols.
MethodsA model-based framework was developed comprising a PK-PD-model for trastuzumab associated changes in LVEF, and a protocol-execution model, describing treatment interventions based on the observed response. Three monitoring protocol scenarios were considered: S1) the SPC-based protocol which represents the currently used monitoring protocol, S2) an adaptive protocol is based on S1, but which includes exclusion of low-risk patients, and S3) a protocol considers a scoring-based algorithm for determining if an intervention is necessary. Evaluation metrics were defined allowing evaluation of diagnostic and therapeutic properties of monitoring strategies. These included the observed area above the LVEF<45% curve (AUC45), dose intensity (efficacy measure), success of dose reduction, and number of LVEF observations.
ResultsFor scenario S3 it was found that patients not experiencing cardiac toxicity receive a marginally higher dose intensity then patients in S1/S2. The success of a protocol-defined dose reduction was improved from 40.2% for the SPC-based protocol, to 75.8% for a scoring-based protocol, thereby decreasing the observed severity of cardiotoxicity. The mean number of LVEF measurements was reduced by 17%, when the monitoring protocol for low-risk patients is adjusted.
ConclusionThis model-based evaluation approach enabled successful evaluation and optimization of cardiac monitoring protocols. The proposed framework and associated evaluation metrics could be applied for evaluation of drugs requiring longitudinal monitoring of (cardiac) toxicity.
Reference: PAGE 21 (2012) Abstr 2467 [www.page-meeting.org/?abstract=2467]
Poster: Oncology