ASN's Mission

ASN leads the fight to prevent, treat, and cure kidney diseases throughout the world by educating health professionals and scientists, advancing research and innovation, communicating new knowledge, and advocating for the highest quality care for patients.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on Twitter

Kidney Week

Abstract: PO0336

Using a Quantitative Systems Pharmacology Model of CKD-MBD to Guide Therapy Minimizing Calcium Flux from Bone and into Vasculature

Session Information

Category: Bone and Mineral Metabolism

  • 402 Bone and Mineral Metabolism: Clinical

Authors

  • Gaweda, Adam E., University of Louisville, Louisville, Kentucky, United States
  • Brier, Michael E., University of Louisville, Louisville, Kentucky, United States
  • Lederer, Eleanor D., University of Texas Southwestern Department of Internal Medicine, Dallas, Texas, United States
Background

CKD-MBD is characterized by bone loss and vascular calcification. Pharmacologic treatment of CKD-MBD involves dosing of three agents to minimize these complications through optimal balance of Calcium (Ca), Phosphorus (P), and PTH. Having developed a Quantitative Systems Pharmacology (QSP) model of CKD-MBD, we test the hypothesis that an AI method called Deep QLearning (DQL) in conjunction with our model can be used to determine the impact of precision therapy on the mineralization defect in patients with End Stage Renal Disease (ESRD).

Methods

Applying a quantitative systems pharmacology (QSP) model of CKD-MBD to mimic disease progression, we trained a Deep Neural Network (virtual physician) to minimize the Ca bone efflux and the Ca vascular tissue influx regardless of achieved serum Ca, P, and PTH predicted by the model. The virtual physician observed Ca, P, PTH and adjusted the doses of P binder, vitamin D, and a calcimimetic. We evaluated a trained virtual physician through simulation of CKD-MBD treatment over 18 months on a population of 100 virtual ESRD patients with varying baseline Ca, P, PTH levels, P intake, and Ca sensing receptor sensitivity.

Results

Simulations produced an average 30% decrease in bone Ca efflux and a 20% decrease in Ca influx to vascular tissue over baseline values. Average P decreased from 7.4 to 5.1 mg/dL, average Ca increased from 8.5 to 9.2 mg/dL, median PTH decreased from 1650 to 315 pg/mL.

Conclusion

Using a QSP model of CKD-MBD, we trained an AI agent to minimize the Ca fluxes from the bone and into the vascular tissue by prioritizing optimization of Ca fluxes over the achievement of specific Ca, P, and PTH levels. Our approach demonstrates beneficial synergy of Systems Biology and AI in modeling complex biologic processes.

Change in Ca flux distributions over time in the simulated patient cohort.

Funding

  • Veterans Affairs Support