Convolutional Neural Networks For Modeling Temporal Biomarkers And - - PowerPoint PPT Presentation

▶
convolutional neural networks for modeling temporal
SMART_READER_LITE
LIVE PREVIEW

Convolutional Neural Networks For Modeling Temporal Biomarkers And - - PowerPoint PPT Presentation

Sep 02, 2022 181 likes •481 views

Convolutional Neural Networks For Modeling Temporal Biomarkers And Disease Predictions Narges Razavian New York University Langone Medical Center GTC 2017 In collaboration with: David Sontag PhD , Saul Blecker MD , Ann-Marie Schmidt MD ,

slide-1
SLIDE 1

Convolutional Neural Networks For Modeling Temporal Biomarkers And Disease Predictions

Narges Razavian

New York University Langone Medical Center GTC 2017

In collaboration with: David SontagPhD, Saul BleckerMD, Ann-Marie SchmidtMD, Enrico BertiniPhD, Rahul Krishnan, YD Choi, Josua Krause, Somesh Nigam, Aaron Smith-McLallen, Ravi Chawla

slide-2
SLIDE 2

Deep learning progress Healthcare world getting digital

Parallel Developments

EHR adoption by healthcare centers in the US Error rate on Image-Net object recognition challenge

slide-3
SLIDE 3

What is captured in the EHR?

Source: healthcare.gov

slide-4
SLIDE 4

Healthcare has joined the data-rich world

slide-5
SLIDE 5

Moving from Treatment to Prevention

Challenges: Each Individual has a different ‘healthy’ baseline.

  • Temporal Patterns/Trends are predictive

Each biomarker varies at a different speed in our bodies Measurements are sparse, asynchronous and correlated Many correlated outcomes are observed per patient

  • Can we leverage this correlation?
slide-6
SLIDE 6

Biomarkers and Outcomes

Biomarkers measurements

  • ver time
slide-7
SLIDE 7

Biomarkers and Outcomes

Biomarkers measurements

  • ver time

Phenotype (diseases)

  • ver time
slide-8
SLIDE 8

Biomarkers and Outcomes

Biomarkers measurements

  • ver time

Phenotype (diseases)

  • ver time
slide-9
SLIDE 9

Biomarkers and Outcomes

Biomarkers measurements

  • ver time

Phenotype (diseases)

  • ver time
slide-10
SLIDE 10

Step 1 Learn each biomarker from other biomarkers time-series

slide-11
SLIDE 11

Kernel Regression

Observations X (Measurement Time-Series) Time Not Observed Want to estimate

slide-12
SLIDE 12

Kernel Regression

Observations X (Measurement Time-Series) Time Not Observed Want to estimate

E[X(v)] = xP(x |

∫

t = v, Xtrain)dx

E[X | t = v, Xtrain] = x

∫

P(x,t = v | Xtrain) P(t = v | Xtrain) dx

E[X | t = v, Xtrain] = x

∫

K(x − xi,v −ti)

xi,ti

∑

K(v −ti)

ti

∑

dx

slide-13
SLIDE 13

Kernel Regression

Observations X (Measurement Time-Series) Time Not Observed Want to estimate

E[X(v)] = xP(x |

∫

t = v, Xtrain)dx

E[X | t = v, Xtrain] = x

∫

P(x,t = v | Xtrain) P(t = v | Xtrain) dx

E[X | t = v, Xtrain] = x

∫

K(x − xi,v −ti)

xi,ti

∑

K(v −ti)

ti

∑

dx E[X | t = v, Xtrain] = (K ⊗ Xtrain)(v) (K ⊗ I(Xtrain :Observed))(v)

slide-14
SLIDE 14

Use convolution framework to LEARN those kernels

slide-15
SLIDE 15

We can learn the kernel (No need for parametric forms and cross validations) Easily extendible to multivariate! Unsupervised: All needed is (asynchronous) sequence of

  • bservations.

Fast to train. Fast to apply.

Benefits

slide-16
SLIDE 16

Data: 30K Individuals from the original training set. Dataset split equally between train, test and validate set. Loss: MSE. Train and evaluate only on (lab, person) with more than 1 observaGon.

Mul$variate Kernels learned for each input dimension (total 18)

slide-17
SLIDE 17

More details in our ICLR paper

Narges Razavian, David Sontag Temporal Convolutional Neural Networks for Diagnosis from Lab Tests http://arxiv.org/abs/1511.07938 Open Source code available (torch/lua implementation): https://github.com/clinicalml/deepDiagnosis

slide-18
SLIDE 18

Step 2 Predict 200+ correlated outcomes using multi- resolution convolutional neural networks and multi-task learning

slide-19
SLIDE 19

Multi-Resolution Convolution Networks

The Architecture - model (1)

slide-20
SLIDE 20

Multi-Resolution Convolution Networks

The Architecture - model (2)

slide-21
SLIDE 21

Prediction AUCs on the held-out test set

slide-22
SLIDE 22

More details in our JMLR paper

Narges Razavian, Jake Marcus, David Sontag, Multi-task Prediction of Disease Onsets from Longitudinal Lab Tests JMLR, 2016 http://arxiv.org/abs/1608.00647 Open Source code available (torch/lua implementation): https://github.com/clinicalml/deepDiagnosis

slide-23
SLIDE 23

Following up in clinical world

  • Prediction models built and deployed for

– Nurse calls and home visits for 250,000+ NYUMC patients at high risk for a number of these outcomes – Improved documentation in EHR

  • Automation of mandatory visits/screening/follow-ups
  • Best practice alerts
  • Reimbursement for intense lifestyle management programs
  • Extending to broader outcomes and domains
slide-24
SLIDE 24

New York University (i2b2) Database

slide-25
SLIDE 25

New York University (i2b2) Database

Nuclear Medicine Procedures Magnetic Resonance Imaging

slide-26
SLIDE 26

Conclusions

  • Applications of deep learning in healthcare are unlimited
  • Unsupervised learning + back-propagation + deep learning

can recover biomarker models from asynchronous high- dimensional time-series data

  • Multi-task learning benefits prediction tasks with smaller

datasets.

slide-27
SLIDE 27

Thanks!

Questions/comments: Narges.Razavian@nyumc.org Open Source Package: https://github.com/clinicalml/deepDiagnosis

slide-28
SLIDE 28