Learning Patient-Specific Lumped Models for Interactive Coronary - - PowerPoint PPT Presentation

▶
learning patient specific lumped models for
SMART_READER_LITE
LIVE PREVIEW

Learning Patient-Specific Lumped Models for Interactive Coronary - - PowerPoint PPT Presentation

Apr 06, 2024 236 likes •408 views

Learning Patient-Specific Lumped Models for Interactive Coronary Blood Flow Simulations Paper #168, Poster F-5 Hannes Nickisch, Yechiel Lamash, Sven Prevrhal, Moti Freiman, Mani Vembar, Liran Goshen and Holger Schmitt October 08, 2015 Learning

slide-1
SLIDE 1

1 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Learning Patient-Specific Lumped Models for Interactive Coronary Blood Flow Simulations

Paper #168, Poster F-5

Hannes Nickisch, Yechiel Lamash, Sven Prevrhal, Moti Freiman, Mani Vembar, Liran Goshen and Holger Schmitt

October 08, 2015

slide-2
SLIDE 2

2 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Overview: FFR and FFR-CT

  • Significance of a coronary stenosis assessed by FFR
  • FFR = Fractional Flow Reserve = Pd/Pa
  • FFR: measured through invasive cathlab procedure through a pressure wire
  • FFR-CT: simulated from a routine cardiac CTA scan using a biophysical model
  • Our contribution: Interactive simulation using lumped models

FFR

FFR-CT

slide-3
SLIDE 3

3 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Visualisation

slide-4
SLIDE 4

4 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Workflow

(a) Standard coronary CTA scan (b) Automatic cardiac segmentation (c) Coronary lumen segmentation (d) Coronary tree (centerline+cross sections) (e) FFR simulation

slide-5
SLIDE 5

5 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Simulation: Input and Output

Vessel Segmentation Boundary Conditions FFR value

slide-6
SLIDE 6

6 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Simulation Pipeline: Finite Elements (FE)

Vessel Segmentation Surface Meshing Volume Meshing Boundary Conditions FE Simulation FFR value

slide-7
SLIDE 7

7 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Simulation Pipeline: Lumped Model (LM)

Vessel Segmentation Surface Meshing Volume Meshing Boundary Conditions FE Simulation LM Simulation FFR value

slide-8
SLIDE 8

8 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Lumped Models

  • Fast:

extremely quick to evaluate

  • Simple:

no surface or volume meshing required

  • Already there: most FE boundary conditions are lumped
  • Similar to electrical circuits via hydraulic analogy
  • voltage ≡ pressure, current ≡ volumetric flow rate
  • wire ≡ pipe, resistance ≡ constricted pipe,
  • Non-linear 0d models aka reduced order approximation to 3d FE models
  • Fit using statistical machine learning and hydraulic features
slide-9
SLIDE 9

9 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

  • Segment transfer function = sum of local (cross sectional) transfer functions

Lumped Model from Geometry

radius length f

f: volumetric flow rate p: pressure

slide-10
SLIDE 10

10 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

  • Segment transfer function = sum of local (cross sectional) transfer functions
  • Local transfer function = weighted sum of effect transfer functions
  • Symmetric polynomial from literature
  • Poiseuille, Bernoulli, ovality, expansion, constriction, bifurcation, curvature, ..

Lumped Model from Geometry

radius length f

f: volumetric flow rate p: pressure

slide-11
SLIDE 11

11 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Lumped Model Learning

  • Adjust effect weights so that lumped model matches CFD simulation
  • Simulate a training data base of cases using OpenFOAM
  • Lumped Model estimate
  • Nonnegative least squares fit
slide-12
SLIDE 12

12 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Learning Validation

  • E=8 coefficients , 35 coronary trees, 10 flow levels, 20fold resampling
  • MAE 2.76±0.56mmHg, runtime ≤2min
slide-13
SLIDE 13

13 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Ground Truth Validation

  • 41 patients, 59 invasive FFR measurements (GT)
  • Clinical FFR threshold 0.8 → AUC=0.77
  • Green: correct, red: incorrect, gray: intrinsic error margin
slide-14
SLIDE 14

14 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Interactive Prediction

slide-15
SLIDE 15

15 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015

Summary and Discussion

  • Simple and fast surrogate model to FE simulations of fluid flow
  • Trained using machine learning
  • Evaluated on FFR prediction tasks
  • Allow for interactive computation of hemodynamic parameters
  • Applicable to general networks of elongated structures
  • Hamburg (Research)

Hannes Nickisch, Sven Prevrhal, Holger Schmitt

  • Haifa (Advanced Development)

Yechiel Lamash, Moti Freiman, Liran Goshen

  • Cleveland (Clinical Science)

Mani Vembar

slide-16
SLIDE 16

16 Learning Lumped Coronary Blood Flow Models: Nickisch et al., October 08, 2015