Article Text
Abstract
Introduction Cerebral arteriovenous malformation (AVM) rupture risk is thought to be approximately 3% peryear, but varies with high risk anatomic features. Recent studies have raised the question about whether treatment is indicated for unruptured AVMs. The important decisions about treatment versus conservative management is largely based on a collection of outcome studies that may not match a patient’s particular AVM. The ultimate purpose of this research is to create a predictive computational model for a patient’s specific AVM anatomy to give them a lifetime risk of rupture profile. Our hypothesis is that the flow and pressure patterns within the nidus of the AVM can be determined by creating a 3D printed model of a simple AVM with internal imbedded pressure and flow sensors. This information can then be incorporated into a fluid structure interaction model of a simple arteriovenous malformation.
Methods In an IRB approved study, unruptured Spetzler Martin Grade I and II AVMS were evaluated to select an AVM that demonstrated a relatively de-compact nidus and simple inflow and outflow vessels with cerebral angiography imaging with at least 4 frames/second anterior-posterior and lateral runs. This AVM was then processed in Matlab as well as manually segmented to create a 3D printing stereolithography file. The AVM was then printed in the Cleveland Clinic 3D printing laboratory with hollow lumens within the vessels and was fitted with inline flow probes within the inflow, outflow and nidus vessels and pressure transducers. This AVM was then inserted into a flow circuit connected to a cardiac cycle pump. The pump was then run with fluid that mirrors the viscosity of blood and cycled for 10 repetitions to verify stability of the pressure and flow readings at each 10 mmHg increment of output pressure between 100 mmHg to 260 mmHg. The model was then imaged in a cerebral angiography machine and connected to a pressure injector to compare transit time between the flow model and the previously obtained cerebral angiogram data from the patient.
Results We successfully built and measured flow and pressure from a 3D printed replica of a simple brain AVM incorporated into a cardiac cycle fluid circuit. This data is currently being compared to cerebral angiography transit rates to validate the data and fit the projected flow and pressure measurements to clinically obtained data. The data additionally is being incorporated into inflow and outflow parameters in the ANSYS fluid structure interaction model of the same simple AVM to compare predicted pressure and flow through the simulation compared to the 3D printed AVM measurements.
Conclusions We have made progress in the multistep task of building a fluid structure interaction model of a simple AVM by measuring flow rate and pressure measurements in an anatomically correct 3D printed model of a simple AVM. This data allows for more accurate simulation and for further evaluation of whether flow and pressure characteristics can be simplified for a compact AVM nidus with the goal of developing a predictive model of the natural history of AVM rupture.
Disclosures N. Moore: 1; C; Joe Niekro/SNIS Research Grant. R. Klatte: None. A. Erdemir: None. T. Masaryk: None. N. Parkar: None. M. Bain: None. M. Hussain: None.