Deep transfer learning for characterizing chondrocyte patterns in phase contrast X-Ray computed tomography images of the human patellar cartilage

• Published in: Computers in biology and medicine Vol. 95; pp. 24 - 33
• Main Authors: Abidin, Anas Z., Deng, Botao, DSouza, Adora M., Nagarajan, Mahesh B., Coan, Paola, Wismüller, Axel
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.04.2018; Elsevier Limited
• Subjects: Algorithms; Arthritis; Artificial neural networks; Biocompatibility; Biomedical materials; Bus interconnections; Cartilage; Cartilage - diagnostic imaging; Cartilage diseases; Chondrocytes; Classification; Computation; Computed tomography; Convolutional neural network; Datasets; Deep transfer learning; Diabetic retinopathy; Embedding; Feature extraction; Humans; Image contrast; Image retrieval; International conferences; Machine Learning; Medical imaging; Neural networks; Neural Networks, Computer; Osteoarthritis; Osteoarthritis - diagnostic imaging; Patella - diagnostic imaging; Patellar cartilage; Phase contrast; Phase contrast imaging; Stochasticity; Tomography, X-Ray Computed - methods; Transfer learning; Visualization; Patellar cartilage; Phase contrast imaging; Convolutional neural network; Deep transfer learning
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2018.01.008

Phase contrast X-ray computed tomography (PCI-CT) has been demonstrated to be effective for visualization of the human cartilage matrix at micrometer resolution, thereby capturing osteoarthritis induced changes to chondrocyte organization. This study aims to systematically assess the efficacy of deep transfer learning methods for classifying between healthy and diseased tissue patterns. We extracted features from two different convolutional neural network architectures, CaffeNet and Inception-v3 for characterizing such patterns. These features were quantitatively evaluated in a classification task measured by the area (AUC) under the Receiver Operating Characteristic (ROC) curve as well as qualitative visualization through a dimension reduction approach t-Distributed Stochastic Neighbor Embedding (t-SNE). The best classification performance, for CaffeNet, was observed when using features from the last convolutional layer and the last fully connected layer (AUCs >0.91). Meanwhile, off-the-shelf features from Inception-v3 produced similar classification performance (AUC >0.95). Visualization of features from these layers further confirmed adequate characterization of chondrocyte patterns for reliably distinguishing between healthy and osteoarthritic tissue classes. Such techniques, can be potentially used for detecting the presence of osteoarthritis related changes in the human patellar cartilage. •Phase Contrast Imaging allows visualization of osteoarthritic changes in the patellar cartilage.•Features from pre-trained CNNs can be used characterize healthy and diseased patterns.•Features from Inception-v3 perform better than CaffeNet and GLCM at the classification task.•Fine-tuning with a small dataset can improve classifier performance.•Visualization techniques can help further substantiate the characterization obtained.