Chrisoula Tahtsidou 1,2, Fenja Klima 2,3, Jonas Huhn 1, Noureldin Sharaf 1, Nora Isberner 4, Juliane Staudinger 5, Georg Hempel 5, Ulrich Jaehde 6, Charlotte Kloft 2,3, Oliver Scherf-Clavel 1
1 Clinical Pharmacy and Pharmacotherapy, Department of Pharmacy, Ludwig-Maximilians-Universitaet Muenchen (Munich, Germany), 2 Graduate Research Training Program PharMetrX (Berlin/Potsdam, Germany), 3 Department of Clinical Pharmacy and Biochemistry, Institute of Pharmacy, Freie Universitaet Berlin (Berlin, Germany), 4 Division of Infectious Diseases and Tropical Medicine, Department of Internal Medicine II, University Hospital Wuerzburg (Wuerzburg, Germany), 5 Institute of Pharmaceutical and Medical Chemistry, Department of Clinical Pharmacy, University of Muenster (Muenster, Germany), 6 Department of Clinical Pharmacy, Institute of Pharmaceutical Sciences, University of Bonn (Bonn, Germany)
Objective:Substantial interindividual pharmacokinetic variability results in up to 45% of patients receiving oral anticancer drugs being under- or overexposed[1]. Therapeutic drug monitoring (TDM) has the potential to identify suboptimal concentrations and thereby improve efficacy while reducing toxicity. However, TDM is traditionally based on venous blood sampling, which is time consuming, and often inconvenient for patients. Volumetric absorptive microsampling (VAMS) offers a minimally invasive alternative enabling at home sampling and may facilitate routine TDM[2]. For cabozantinib, a tyrosine kinase inhibitor, accumulating evidence demonstrates exposure-toxicity and -efficacy relationships, but therapeutic targets are defined only in serum or plasma[3, 4], while reliable translation of capillary VAMS concentrations into serum‑based exposure metrics is lacking. Thus, this project investigated the feasibility of capillary microsampling for cabozantinib TDM and aimed to develop a model informed VAMS to serum conversion strategy using nonlinear mixed effects (NLME) modeling.
Methods:Data were obtained from two prospective, non‑interventional studies enrolling patients with advanced/metastatic renal cell carcinoma (RCC) treated with cabozantinib: the multicenter ON‑TARGET study[5] and a single-center observational study at the University Hospital of Wuerzburg. During outpatient visits, time-paired venous samples and VAMS capillary samples (MITRA®) were collected. Additionally, the Wuerzburg cohort performed at‑home VAMS sampling.
Four published serum population pharmacokinetic (popPK) models including RCC patients[6-9] were evaluated using NONMEM®7.5.1 via PsN 5.6.0 and Pirana 24.9.2. The best‑performing serum model was extended to incorporate VAMS concentrations. Two conversion strategies were investigated: a hematocrit (Hct)‑based conversion according to Iacuzzi et al., applying individual or typical sex‑specific Hct values[10], and an empirical conversion factor with interindividual variability (IIV), estimated from the consolidated data. For at-home samples, the last known Hct was applied. In patients without any Hct information, sex-specific population mean values were used. Predictive performance was assessed using goodness‑of‑fit (GOF) diagnostics, prediction‑corrected visual predictive checks (pcVPCs), median individual prediction error (MDIPE) and median absolute individual prediction error (MDAIPE). Individual predictions were derived to compare serum‑based and VAMS‑based exposure. Finally, serum minimal concentrations (Cmin) were predicted using (1)serum measurements within the dosing interval, and (2)only capillary concentrations to evaluate the feasibility of VAMS‑based TDM and the potential to identify concentrations above the proposed toxicity threshold[3]. As the model included IIV only and age and body weight showed minimal within-patient variation, Cmin predictions were averaged per patient and dose to account for unequal sample numbers across patients.
Results:From ON‑TARGET, 230 venous samples of 25 patients and 87 capillary samples from a subset of 12 patients were included in the analysis. The Würzburg cohort contributed 41 venous and 108 capillary samples from 18 patients. Across all evaluated serum popPK models, our population consistently showed a higher apparent clearance (CL/F). The modified model by Tran et al.[8] provided the best fit with an estimated CL/F of 3.44L/h (8.2% RSE) and good MDIPE and MDAIPE of -0.42% and 19.4%, respectively, and was selected as the basis for the VAMS‑to‑serum conversion framework.
Both empirical and Hct-based conversion approaches demonstrated good predictive performance (MDIPE <5%, MDAIPE <20%), supported by GOF diagnostics and pcVPCs. By empirical conversion, capillary concentrations were reduced by 44.8% (5.1% RSE) compared to serum samples with IIV of 53.8%CV (17.5% RSE). Individual Hct approximations derived from hemoglobin measurements was available for 93% of patients from outpatient visits but did not meaningfully improve predictions compared with typical Hct values. Thus, in exploratory analyses using only capillary concentrations, serum Cmin predictions were derived from models applying the typical-Hct-based and empirical conversion.
The Hct-based VAMS model resulted in 76.1% of dose-stratified individual mean Cmin predictions being within ±20% of corresponding serum-derived Cmin estimates, compared with only 50% when applying the empirical conversion approach without covariate testing. In first exploratory analyses, the Hct-model showed good concordance with serum-derived toxicity classifications, identifying 21.7% of patients above the threshold versus 28.3% based on serum concentrations.
Conclusion:This study shows that capillary microsampling could be a reliable alternative to venous sampling for cabozantinib TDM. The NLME-based conversion models adequately translated capillary concentrations into serum‑equivalent exposure. Still, larger cohorts and extended covariate analyses are needed to clarify serum-VAMS discrepancies and to establish a clinically robust conversion strategy. This approach could allow reliable estimation of serum exposure using solely VAMS samples, potentially reducing the burden of clinic visits while enabling identification of patients with overexposure.
References:
[1] Groenland, S.L., et al., Individualized dosing of oral targeted therapies in oncology is crucial in the era of precision medicine. European Journal of Clinical Pharmacology, 2019. 75(9): p. 1309-1318.
[2] Meertens, M., et al., Clinical Application of Volumetric Absorptive Microsampling for Therapeutic Drug Monitoring of Oral Targeted Anticancer Drugs. Therapeutic Drug Monitoring, 2025. 47(5): p. 625-634.
[3] Cerbone, L., et al., Association of cabozantinib pharmacokinetics, progression and toxicity in metastatic renal cell carcinoma patients: results from a pharmacokinetics/pharmacodynamics study. ESMO Open, 2021. 6(6): p. 100312.
[4] Castellano, D., et al., Exposure-response modeling of cabozantinib in patients with renal cell carcinoma: Implications for patient care. Cancer Treatment Reviews, 2020. 89.
[5] Mc Laughlin, A.M., et al., Developing a Nationwide Infrastructure for Therapeutic Drug Monitoring of Targeted Oral Anticancer Drugs: The ON-TARGET Study Protocol. Cancers, 2021. 13(24): p. 6281.
[6] Lacy, S., et al., A population pharmacokinetic model of cabozantinib in healthy volunteers and patients with various cancer types. Cancer Chemotherapy and Pharmacology, 2018. 81(6): p. 1071-1082.
[7] Nguyen, L., et al., Updated Population Pharmacokinetic Model of Cabozantinib Integrating Various Cancer Types Including Hepatocellular Carcinoma. Journal of Clinical Pharmacology, 2019. 59(11): p. 1551-1561.
[8] Tran, B.D., et al., Cabozantinib exposure–response analysis for the phase 3 CheckMate 9ER trial of nivolumab plus cabozantinib versus sunitinib in first-line advanced renal cell carcinoma. Cancer Chemotherapy and Pharmacology, 2023. 91(2): p. 179-189.
[9] Clinical Pharmacology and Biopharmaceutics Review for Cabometyx. 2016, U.S. Food and Drug Administration.
[10] Iacuzzi, V., et al., Dried Blood Spot Technique Applied in Therapeutic Drug Monitoring of Anticancer Drugs: a Review on Conversion Methods to Correlate Plasma and Dried Blood Spot Concentrations. Pharmaceutical Research, 2021. 38(5): p. 759-778.
Reference: PAGE 34 (2026) Abstr 11950 [www.page-meeting.org/?abstract=11950]
Poster: Drug/Disease Modelling - Oncology