KNEE INJURY DETECTION USING MRI WITH EFFICIENTLY LAYERED NETWORK - - PowerPoint PPT Presentation

▶

Nov 19, 2023 331 likes •464 views

KNEE INJURY DETECTION USING MRI WITH EFFICIENTLY LAYERED NETWORK (ELNET) Chen-Han Han Tsai, i, Nahum um Kiryati ati, , Eli i Konen, , Iris Eshed, , Arnald naldo o Mayer er PROBLEM MOTIVATION CONVEN ENTIONA IONAL L KNEE EE EXAMS

SLIDE 1

KNEE INJURY DETECTION USING MRI WITH EFFICIENTLY LAYERED NETWORK (ELNET)

Chen-Han Han Tsai, i, Nahum um Kiryati ati, , Eli i Konen, , Iris Eshed, , Arnald naldo

Mayer

er

SLIDE 2

PROBLEM MOTIVATION

MRI Acquisition Doctor’s Analysis Final Assessment Exam added to Queue Sorted by Exam Date 

MSK radiologists face a rising work demand each day



Triage improves efficiency by prioritization



Severe cases prioritized first

CONVEN ENTIONA IONAL L KNEE EE EXAMS MS

SLIDE 3

PROBLEM MOTIVATION



MSK radiologists face a rising work demand each day



Triage improves efficiency by prioritization



Severe cases prioritized first

Sorted by Level of Severity

TRIAGE GED D KNEE EE EXAMIN INATIONS IONS

MRI Acquisition Doctor’s Analysis Final Assessment

SLIDE 4

ELNET ARCHITECTURE

Fig-1: Illustration and configuration of ELNet.

SLIDE 5

ELNET CORE COMPONENTS

Fig-3: Multi-slice Normalization for 3D Inputs Fig-2: Block with 2 repeats Fig-4: BlurPool Down-sampling

SLIDE 6

EVALUATION DATASETS

 MRNet

et Datase taset1

 1370 knee MRI exams*  Labels : ACL tear / Meniscus tear / Abnormalities  Axial, coronal, and sagittal scans provided

 KneeM

eMRI I Datas aset et2

 917 knee MRI exams**  Labels: ACL Injured  Sagittal scan provided

1Bien et al, Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet, PLOS Medicine (2018) 2Štajduhar et al, Semi-automated detection of anterior cruciate ligament injury from MRI, Computer Methods and Programs in Biomedicine (2017)

Axial Plane Coronal Plane Sagittal Plane

SLIDE 7

ELNET SETUP

Detect ection ion Object ectiv ive Multi-Sl Slic ice e Norm Image e Modalit lity K Number er of Paramet eter ers Meniscus Tear Contrast Norm Coronal 4 ~ 0.2 M (850 kB) ACL Tear Layer Norm Axial 4

Abnormalities

Layer Norm Axial 4

ACL Tear (KneeMRI)

Contrast Norm Sagittal 2 ~ 0.05 M (438 kB)

 ELNet is trained from scratch  Previous SOTA MRNet ~183M parameters for each objective

SLIDE 8

MRNET EVALUATION

Fig-5: Evaluation of ELNet and MRNet performance on the MRNet Dataset

SLIDE 9

KNEEMRI EVALUATION

Fig-6: Comparison of ELNet performance across all 5 folds on the KneeMRI dataset

SLIDE 10

KNEEMRI EVALUATION

Fig-7: ROC’s of ELNet of KneeMRI Dataset across 5 folds

SLIDE 11

MODEL INTERPRETATION

Fig-8: Full-Grad visualization highlighting the tear locations in the knee

SLIDE 12

SUMMARY

 ELNet features

 Lightweight  Adequate performance  Easily trained from scratch

 May be applied to other pathologies involving 3D images (MRI, CT, etc.)