Quantitative Dose-Response Modelling of GnRH-Antagonist Therapy: A Pharmacodynamic Framework for Dose Selection

Objective: Endocrine therapy for endometriosis is dosed empirically. We built and tested a continuous pharmacodynamic model relating residual endocrine drive to lesion behaviour. Methods: Lesion burden was expressed as an exponential function of a composite endocrine drive, scaled by an inherited susceptibility coefficient, with an activation threshold and a regression term. Structural identifiability was examined analytically. Susceptibility was modelled as lognormal, its dispersion anchored to a published odds ratio. One parameter was fitted to the pooled untreated progression rate from a meta-analysis of 285 untreated and 730 hormonally suppressed patients; progression under suppression, the treated-versus-untreated odds ratio and mean volume change were withheld for validation. Parameter influence was quantified by variance-based global sensitivity analysis. Results: Only lumped products of the rate parameters are identifiable from burden trajectories. At a residual drive of 0.813 the model predicted progression in 10.0% of lesions against 12.4% (95% CI 9.0–16.1) reported, an odds ratio of 0.41 against 0.52 (95% CI 0.41–0.66), and a 12-month volume change of −26.0% against −28.5%. Interactions accounted for 47% of output variance. Conclusions: The model reproduces independent clinical observations not used in its calibration and yields a closed-form criterion for the suppression required to arrest lesion expansion.