Aidence Enhancing Radiology with Artificial Intelligence - - PowerPoint PPT Presentation

▶
aidence
SMART_READER_LITE
LIVE PREVIEW

Aidence Enhancing Radiology with Artificial Intelligence - - PowerPoint PPT Presentation

Nov 28, 2022 357 likes •556 views

GTC EUROPE 2017 Aidence Enhancing Radiology with Artificial Intelligence Localization in 3D Biomedical Image Data using Deep Learning Mark-Jan Harte, CEO GTC EUROPE 2017 About Aidence Founded in 2015, based in Amsterdam Deep learning

slide-1
SLIDE 1

GTC EUROPE 2017

Aidence

Enhancing Radiology with Artificial Intelligence Localization in 3D Biomedical Image Data using Deep Learning Mark-Jan Harte, CEO

slide-2
SLIDE 2

GTC EUROPE 2017

About Aidence

  • Founded in 2015, based in

Amsterdam

  • Deep learning for automatic

medical image analysis

  • 3rd place in the Kaggle Data

Science Bowl 2017

2
slide-3
SLIDE 3

GTC EUROPE 2017

Challenges for AI in Radiology

Technical ○ Sample size ○ Class imbalances due to mostly healthy/background tissue present ○ Accurate labeling is a pain ○ Validation dataset for regulatory approval required

3

Common resolution ImageNet: 256x256x3 = 196.608 Common resolution CT scan: 300x512x512x1 = 78.643.200

slide-4
SLIDE 4

GTC EUROPE 2017

CT Chest

  • Early detection of lung nodules

leads to 20% mortality reduction

  • Human sensitivity ~80%

Nodules are:

  • Small
  • Anywhere
  • Highly variable in number

Lung Nodules

4
slide-5
SLIDE 5

GTC EUROPE 2017

Detection

Training

  • Input is 128x128x7 voxels
  • Target is 58x58 rectangle mask
  • Loss: Normalized cross entropy
  • Output: segmentation (probability map)
  • Keep fine grained spatial details
  • Network size not too big
  • Second network to filter out false positives

○ Larger, less restrictions

Lung Nodules

5
slide-6
SLIDE 6

GTC EUROPE 2017

Network architecture

  • Fully 3D convolutional

○ Efficient inference on big CT scans ○ No same padding ○ No pooling ○ No strides

  • Dilated convolutions

○ Reduce resolution more quickly ○ Keep network size in check (180K params)

  • Weight normalization [Salimans & Kingma, 2016]

○ Easier (than batch norm) to distribute over multiple GPUs

Lung Nodules

6
slide-7
SLIDE 7

GTC EUROPE 2017

MR Lumbar Spine

  • Foramen is the passage where a

nerve exits the spine

  • Foraminal stenosis is a common

cause of leg pain

○ Time-consuming to find on scoliotic Spines

  • Task: locate and classify all of them

Lumbar Foramina

7
slide-8
SLIDE 8

GTC EUROPE 2017

Localization

  • Foramina are big
  • Foramina are located at a certain

location

  • There are 10 lumbar foramina per spine
  • Output a segmentation (probability map)
  • Requires less fine grained details

Lumbar Foramina

8
slide-9
SLIDE 9

GTC EUROPE 2017

Localization architecture

  • 3D U-Net architecture

○ Easier to modify and/or extend ○ Less need for efficient inference ○ Input: 481x481x3

  • Binary segmentation (probability map)
  • Loss: (Fuzzy) Dice score

○ Trains faster and requires less data for our network than the normalized cross entropy loss

Lumbar Foramina

9
slide-10
SLIDE 10

GTC EUROPE 2017

Lumbar Foramina

10
slide-11
SLIDE 11

GTC EUROPE 2017

Lumbar Foramina

11
slide-12
SLIDE 12

GTC EUROPE 2017

Combating over-generalization

  • Reduce the resolution for more context
  • A segmentation per foramen level
  • Bias towards the lower and bigger foramina

Lumbar Foramina

12

Recall L1 L2 L3 L4 L5 Base 0.97 0.96

slide-13
SLIDE 13

GTC EUROPE 2017

Combating over-generalization

  • Reduce the resolution for more context
  • A segmentation per foramen level
  • Bias towards the lower and bigger foramina
  • Normalize the Dice score

Lumbar Foramina

13

Recall L1 L2 L3 L4 L5 Base 0.97 0.96 Normalized 0.87 0.96 0.97 0.98 0.95

slide-14
SLIDE 14

GTC EUROPE 2017

Lumbar Foramina

14
slide-15
SLIDE 15

GTC EUROPE 2017

Lumbar Foramina

15
slide-16
SLIDE 16

GTC EUROPE 2017

A Comparison

  • Lung nodules:

○ Challenge: ■ Detect small nodules in a vast volume ■ Requires fine grained spatial details ○ Solution: ■ Fully convolutional for efficient inference ■ Dilated convolutions to keep network size in check ■ Trains on 10,000s of samples

  • Lumbar foramina

○ Challenge: ■ Detect big foramina in a (relatively) small volume ■ Distinguish lumbar foramina from thoracic ones ○ Solution: ■ Use 3D U-Net with Dice score ■ Normalize the Dice score and reduce the resolution for increased performance ■ Trains on 100s of samples

16
slide-17
SLIDE 17

GTC EUROPE 2017

In practice

  • FDA and CE clearance necessary for diagnostic impact

○ Certify the training pipeline, inference pipeline, annotation tooling, deployment… ○ Concept of independent test set matches very well ○ Discussions on continuous learning (FDA, ACR)

  • Aidence lung nodule detection submitted for CE

○ Feedback received; clearance expected soon

17
slide-18
SLIDE 18

GTC EUROPE 2017

“Your software has, in this short time, detected a patient with a nodule that has clearly grown in 3 years, is probably malignant and missed by 4 consecutive radiologists.”

  • W.M, radiologist
18

markjan@aidence.com