Development of an Artificial Neural Network Algorithm to Assess Cartilage Healthiness

Author(s)

Jonathan YeFollow

School Name

Academic Magnet High School

Grade Level

12th Grade

Presentation Topic

Physiology and Health

Presentation Type

Mentored

Oral Presentation Award

4th Place

Abstract

Osteoarthritis, otherwise known as degenerative joint disease, is an incredibly common disease that affects over 10% of men and 18% of women over 60 years of age. Though the exact cause of osteoarthritis is unknown, the loss of cartilage and chronic joint pain are common features of OA. Currently, disease modifiable treatments are not available. One hurdle of developing an effective treatment for OA is that physicians are unable to assess the cartilage healthiness of patients with high spatial resolution. Recent studies suggest that two-photon laser scanning auto-fluorescence microscopy have the ability to image chondrocyte viability of articular cartilage without dye labeling. As the chondrocyte is the only cell type in articular cartilage to maintain the tissue, the chondrocyte viability is incredibly important in monitoring disease progression and the effects of treatment. However, the current two-photon imaging method relies on manual cell classification and counting, which is time consuming and not practical for real-time measurement. To solve this problem, in this study, a machine learning algorithm, one type of artificial intelligence algorithms, is proposed to automate the process of live cell identification for measuring the chondrocyte viability through the use of the neural network toolbox provided by Matlab (MathWorks). After training the algorithm with 150 cell images, it had a success rate of over 80% in recognizing live/dead cells, which can be further improved with a larger amount of training images. This result has demonstrated that, combined with the development of the imaging tool, the machine learning is a very promising approach towards the assessment of cartilage health in vivo.

Location

Founders Hall 142 A

Start Date

3-30-2019 10:45 AM

Presentation Format

Oral and Written

Group Project

No

COinS
 
Mar 30th, 10:45 AM

Development of an Artificial Neural Network Algorithm to Assess Cartilage Healthiness

Founders Hall 142 A

Osteoarthritis, otherwise known as degenerative joint disease, is an incredibly common disease that affects over 10% of men and 18% of women over 60 years of age. Though the exact cause of osteoarthritis is unknown, the loss of cartilage and chronic joint pain are common features of OA. Currently, disease modifiable treatments are not available. One hurdle of developing an effective treatment for OA is that physicians are unable to assess the cartilage healthiness of patients with high spatial resolution. Recent studies suggest that two-photon laser scanning auto-fluorescence microscopy have the ability to image chondrocyte viability of articular cartilage without dye labeling. As the chondrocyte is the only cell type in articular cartilage to maintain the tissue, the chondrocyte viability is incredibly important in monitoring disease progression and the effects of treatment. However, the current two-photon imaging method relies on manual cell classification and counting, which is time consuming and not practical for real-time measurement. To solve this problem, in this study, a machine learning algorithm, one type of artificial intelligence algorithms, is proposed to automate the process of live cell identification for measuring the chondrocyte viability through the use of the neural network toolbox provided by Matlab (MathWorks). After training the algorithm with 150 cell images, it had a success rate of over 80% in recognizing live/dead cells, which can be further improved with a larger amount of training images. This result has demonstrated that, combined with the development of the imaging tool, the machine learning is a very promising approach towards the assessment of cartilage health in vivo.