Artificial Intelligence Based Autonomous Socket Proposal Program: Socket Design Experiences

The aim of this study is to develop an artificial intelligence-based autonomous socket recommendation program that will provide a more comfortable and easier test socket production with high time-cost efficiency and to share experiences about socket designs in these processes.

Study Overview

• Status: Completed
• Conditions: Artificial Intelligence; Prosthetic
• Intervention / Treatment: Other: the stumps of all patients were scanned with the Artec Eva Lite brand 3D scanner.

Detailed Description

For the artificial intelligence-based software planned to be created, the stumps of all patients were scanned with the Artec Eva Lite brand 3D scanner. The scanned patterns were saved as point clouds. The socket parts of the prostheses used by the same patients were also scanned with the same scanner device and recorded.

The point dataset consisting of stump-socket matches obtained from the patients was used for the software.

In order to train the artificial intelligence model, a working environment has been created in which artificial intelligence libraries and tools can be used on the computer. For this purpose, first Anaconda data science platform was established. Thereupon, Python programming language and Tensorflow deep learning library were installed, other libraries required for the training of the artificial intelligence model were added, and the working environment was made ready. A deep learning algorithm was used in the artificial intelligence model developed for training the data. The purpose of using deep learning, which is one of the most up-to-date and popular artificial intelligence algorithms, is to achieve more accurate results by increasing the performance and accuracy rate. First, the dataset is 90% reserved for training and 10% for testing. Then, a deep learning model was created with the Sequantial() model selected from the Keras library. In the model, a total of 7 layers are used, the first of which is the input layer and the last is the output layer. While "relu" is used as the activation function for the input layer and intermediate layers, the "linear" function is used for the output layer. While creating the model, "Adam" was chosen as the optimizer. In the model trained with a total of 500 "repetitions", "batch size" is assigned as 5. The trained model was then tested with the test data and a success rate of 61% was achieved. Afterwards, the model and weights were recorded. After the model training was completed, a new Python program was developed. The previously developed models and weights were loaded while the program was running and were used to propose a socket for the new die data to be given. When the program is run, the stump name for which a socket is requested is asked.

Thus, the program proposes a new socket after receiving the stubby data set from the user and testing it in the trained model. This 3D socket model is shown to the user via the Python Plotly Graphics Library.

• Study Type: Observational
• Enrollment (Actual): 101

Contacts and Locations

Study Locations

• Turkey
  • Şahinbey
    • Gaziantep, Şahinbey, Turkey, 27000
      • Hasan Kalyoncu University

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 14 years to 61 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

Patients aged between 18-45 years with amputation who came to Hasan Kalyoncu University and who met the inclusion criteria of the study.

Description

Inclusion Criteria:

- Conscious patients >18 years old having undergone amputation surgery

Exclusion Criteria:

• • Severe visual and perception impairment

  • Surgical intervention with functional sequelae in the extremities
  • Pain that does not allow tests to be done
  • Patients with diseases with neurological dysfunction (stroke, multiple sclerosis, etc.)

Study Plan

How is the study designed?

Design Details

• Observational Models: Case-Control
• Time Perspectives: Retrospective

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Model of the stump scanned with a 3d scanner; For the artificial intelligence-based software planned to be created, the stumps of all patients were scanned with the Artec Eva Lite brand 3D scanner. The scanned patterns were saved as point clouds	Other: the stumps of all patients were scanned with the Artec Eva Lite brand 3D scanner.; the stumps of all patients were scanned with the Artec Eva Lite brand 3D scanner.
Socket matched to stump; The socket parts of the prostheses used by the same patients (with other group) were also scanned with the same scanner device and recorded.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Software ( Artificial Intelligence Based Autonomous Socket Proposal Program)	The foresight of the software to be developed will be evaluated. It will be evaluated how suitable a socket design can be suggested for the stump dimensions entered into the system. Thanks to the software, the time taken for socket design will be compared with the time taken for sockets produced with classical methods. The time/cost effectiveness of the software will be evaluated.	2 years

Collaborators and Investigators

• Sponsor: Hasan Kalyoncu University
• Investigators: Principal Investigator: Murat ÇINAR, Doctor, Hasan Kalyoncu University

Publications and helpful links

General Publications

• Ten Kate J, Smit G, Breedveld P. 3D-printed upper limb prostheses: a review. Disabil Rehabil Assist Technol. 2017 Apr;12(3):300-314. doi: 10.1080/17483107.2016.1253117. Epub 2017 Feb 2. Review.
• O'Brien L, Cho E, Khara A, Lavranos J, Lommerse L, Chen C. 3D-printed custom-designed prostheses for partial hand amputation: Mechanical challenges still exist. J Hand Ther. 2021 Oct-Dec;34(4):539-542. doi: 10.1016/j.jht.2020.04.005. Epub 2020 Jun 19.
• Vujaklija I, Farina D. 3D printed upper limb prosthetics. Expert Rev Med Devices. 2018 Jul;15(7):505-512. doi: 10.1080/17434440.2018.1494568. Epub 2018 Jul 5. Review.
• Abbady HEMA, Klinkenberg ETM, de Moel L, Nicolai N, van der Stelt M, Verhulst AC, Maal TJJ, Brouwers L. 3D-printed prostheses in developing countries: A systematic review. Prosthet Orthot Int. 2022 Feb 1;46(1):19-30. doi: 10.1097/PXR.0000000000000057.

Study record dates

Study Major Dates

• Study Start (Actual): January 1, 2020
• Primary Completion (Actual): June 10, 2021
• Study Completion (Actual): March 1, 2022

Study Registration Dates

• First Submitted: April 10, 2022
• First Submitted That Met QC Criteria: April 17, 2022
• First Posted (Actual): April 22, 2022

Study Record Updates

• Last Update Posted (Actual): April 22, 2022
• Last Update Submitted That Met QC Criteria: April 17, 2022
• Last Verified: April 1, 2022

More Information

Terms related to this study

• Other Study ID Numbers: MAC2022

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: No

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No