Main Logo
Logo

Society for Pediatric Radiology – Poster Archive

  319
  0
  0
 
 


Final ID: Paper #: 021

Assessment of Classic Metaphyseal Lesions on Bone Surveys for Non-Accidental Trauma with Deep Learning Algorithms

Purpose or Case Report: To assess whether a convolutional neural network (CNN) can be trained via transfer learning to accurately diagnose metaphyseal corner fractures on long bone radiographs by comparing its performance to that of a board-certified pediatric radiologist. Secondary endpoints will include the ability of the CNN to detect the type of long bone presented on the radiograph as well as identify the chronicity of the fracture.
Methods & Materials: In this IRB approved retrospective study, we reviewed records from our children hospital’s radiology database. Consecutive patients, 12 months in age or younger, from January 2009 to September 2019 were screened for bone surveys in the evaluation of suspected non-accidental trauma performed at our institution. Patients with inadequate or incomplete radiographic examinations due to poor patient positioning or overlying artifact were excluded.

A total of 132 frontal radiographs of the femur were retrospectively utilized for the training and testing of a CNN binary classifier. Radiographs were obtained from 61 patients who were found to have femoral metaphyseal corner fractures and from 71 patients with normal femoral metaphyses and no metaphyseal corner fractures. Twenty cases were randomly allocated from the total for validation and an additional 20 for testing. We utilized transfer learning (VGG16) and data augmentation in our algorithm due to the low sample size. The model was assessed using the test dataset with area under the Receiver Operating Characteristic curve and accuracy.
Results: Using a small training dataset of 82 images, we achieved an accuracy of 95% with both validation (n = 20) and test (n = 20) datasets. Our model achieved an area under the Receiver Operating Characteristic curve of 0.94.
Conclusions: A CNN trained on a small set of images with data augmentation is able to detect metaphyseal corner fractures on long bone radiographs with relatively high accuracy. We predict that further analysis with segmentation using a larger dataset will further improve accuracy.
  • Xue, Christine  ( Loma Linda University , Loma Linda , California , United States )
  • Nowrangi, Rajeev  ( Loma Linda University , Loma Linda , California , United States )
  • Smith, John  ( Loma Linda University , Loma Linda , California , United States )
  • Acharya, Patricia  ( Loma Linda University , Loma Linda , California , United States )
Session Info:

Scientific Session I-C: Informatics/AI

Informatics, Education, QI, or Healthcare Policy

SPR Scientific Papers

More abstracts on this topic:
Assistive Management for Independent Gastrojejunal Enteral Tube Clinic

Desai Sudhen, Steve Mccaulley, Vaidya Vinay

How Well do Hospitals Comply with ACR-SPR Guidelines when Performing Skeletal Surveys for Child Abuse?

Marine Megan, Smith Jessica, Wanner Matthew, Jennings Samuel, Hibbard Roberta, Karmazyn Boaz

More abstracts from these authors:
Pediatric Radiology Advocacy and Research - The Crossover We Didn't Know We Needed

Yaya Carlos, Miranda Schaeubinger Monica, Morales-tisnés Tatiana, Otero Hansel, Annam Aparna, Acharya Patricia, Joshi Aparna

Assessment of Germinal Matrix Hemorrhage on Cranial Ultrasound with Deep Learning Algorithms

Kim Kevin, Nowrangi Rajeev, Mcgehee Arianna, Joshi Neil, Acharya Patricia

Due to circumstances surrounding the coronavirus pandemic, this final ePoster exhibit was not submitted.
You have to be authorized to contact abstract author. Please, Login or Signup.

Please note that this is a separate login, not connected with your credentials used for the SPR main website.

Not Available

Comments

We encourage you to join the discussion by posting your comments and questions below.

Presenters will be notified of your post so that they can respond as appropriate.

This discussion platform is provided to foster engagement, and stimulate conversation and knowledge sharing.

Please click here to review the full terms and conditions for engaging in the discussion, including refraining from product promotion and non-constructive feedback.

 

You have to be authorized to post a comment. Please, Login or Signup.

Please note that this is a separate login, not connected with your credentials used for the SPR main website.


   Rate this abstract  (Maximum characters: 500)