A neural pathomics framework for classifying colorectal cancer histopathology images based on wavelet multi-scale texture analysis

Abstract

Colorectal cancer (CRC) constitutes the third most commonly diagnosed cancer in males and the second in females. Precise histopathological classification of CRC tissue pathology is the cornerstone not only for diagnosis but also for patients’ management decision making. An automated system able to accurately classify different CRC tissue regions may increase diagnostic precision and alleviate clinical workload. However, tissue classification is a challenging task due to the variability in morphological and textural characteristics present in histopathology images. In this study, an artificial neural network was trained to classify between eight classes of CRC tissue image patches derived from a public dataset with 5000 CRC histopathology image tiles. A total of 532 multi-level pathomics features examined at different scales were extracted by visual descriptors such as local binary patterns, wavelet transforms and Gabor filters. An exhaustive evaluation involving a variety of wavelet families and parameters was performed in order to shed light on the impact of scale on pathomics based CRC tissue differentiation. Our model achieved a performance accuracy of 95.3% with tenfold cross validation demonstrating superior performance compared to 87.4% reported in recent studies. Furthermore, we experimentally showed that the first and the second levels of the wavelet approximations can be used without compromising classification performance.

Deep Radiotranscriptomics of Non-Small Cell Lung Carcinoma for Assessing Molecular and Histology Subtypes with a Data-Driven Analysis

“…Furthermore, another agreement with the findings ( Table S6 ) of the proposed analysis was observed for discriminative features of histology subtypes with studies reporting cluster shade [ 62 ], first-order, GLCM, GLSZM [ 63 ] and a combination of high level emphasis and small area emphasis [ 64 ] as important features. It is worth noting that the majority of the identified radiomics were wavelet-based features ( Table S6 ), indicating that a significant part of the differentiating information exists only in specific frequency bands and can be deciphered through scale-space wavelet analysis [ 65 ].…”

AI-Based Pipeline for Classifying Pediatric Medulloblastoma Using Histopathological and Textural Images

“…Afterwards, every covariance matrix is normalized by the total number of its components to determine the covariance relative frequency among the gray levels of mutual pixels [61]. In this study, we have used the traditional GLCM textural features approach, although other techniques such as the doughnut GLCM [62], GLCM based on Haralick features [58], WPT-GLCM, WPT-LBP-GLCM, and WPT-Gabor-GLCM (WPT: wavelet packet transform, LBP: local binary patterns) [63] have been deployed by other studies.…”

Deep Learning With Radiomics for Disease Diagnosis and Treatment: Challenges and Potential

“…Evidence supports that the classification and grading of many tumors, such as breast, colorectal, prostate, glioma, and lung cancer, are possible through histopathological images (21,71,88,146,147). Specifically, Sharma and Mehra (146) evaluated the discriminative power of handcrafted and baseline pathology and depth features in a breast cancer multi-classification problem, with linear SVM and VGG16 networks exhibiting excellent predictive performance.…”

“…Specifically, Sharma and Mehra (146) evaluated the discriminative power of handcrafted and baseline pathology and depth features in a breast cancer multi-classification problem, with linear SVM and VGG16 networks exhibiting excellent predictive performance. Trivizakis et al (21) proposed a multiscale texture analysis framework for CRC classification. They obtained an accuracy of 95.3% in the recognition task of eight types of CRC tissue image patches, which is better than the 87.4% obtained in recent studies.…”

“…Deep learning-based radiomics-driven multiclassification methods for breast cancer typically employ supervised models based on transfer learning (146). Multiscale texture characteristics of WSIs have demonstrated the best performance in recent studies in terms of the CRC tissue region differentiation (21). Khan et al (121) used an extreme learning machine model combining transfer learning and feature fusion to classify voxels from multimodality MRI of patients with brain tumors automatically.…”