Bowen Song1,2, Guopeng Zhang3, Hongbing Lu3, Huafeng Wang1, Wei Zhu2, Perry J Pickhardt4, Zhengrong Liang5. 1. Department of Radiology, Stony Brook University, Stony Brook, NY , 11790, USA. 2. Department of Applied Mathematics and Statistics, Stony Brook University, Stony Brook, NY , 11790, USA. 3. Department of Biomedical Engineering, Fourth Military Medical University, Xi'an , 710032, Shaanxi, China. 4. Department of Radiology, University of Wisconsin School of Medicine and Public Health, Madison, WI , 53792, USA. 5. Department of Radiology, Stony Brook University, Stony Brook, NY , 11790, USA. jerome@mil.sunysb.edu.
Abstract
PURPOSE: Differentiation of colon lesions according to underlying pathology, e.g., neoplastic and non-neoplastic lesions, is of fundamental importance for patient management. Image intensity-based textural features have been recognized as useful biomarker for the differentiation task. In this paper, we introduce texture features from higher-order images, i.e., gradient and curvature images, beyond the intensity image, for that task. METHODS: Based on the Haralick texture analysis method, we introduce a virtual pathological model to explore the utility of texture features from high-order differentiations, i.e., gradient and curvature, of the image intensity distribution. The texture features were validated on a database consisting of 148 colon lesions, of which 35 are non-neoplastic lesions, using the support vector machine classifier and the merit of area under the curve (AUC) of the receiver operating characteristics. RESULTS: The AUC of classification was improved from 0.74 (using the image intensity alone) to 0.85 (by also considering the gradient and curvature images) in differentiating the neoplastic lesions from non-neoplastic ones, e.g., hyperplastic polyps from tubular adenomas, tubulovillous adenomas and adenocarcinomas. CONCLUSIONS: The experimental results demonstrated that texture features from higher-order images can significantly improve the classification accuracy in pathological differentiation of colorectal lesions. The gain in differentiation capability shall increase the potential of computed tomography colonography for colorectal cancer screening by not only detecting polyps but also classifying them for optimal polyp management for the best outcome in personalized medicine.
PURPOSE: Differentiation of colon lesions according to underlying pathology, e.g., neoplastic and non-neoplastic lesions, is of fundamental importance for patient management. Image intensity-based textural features have been recognized as useful biomarker for the differentiation task. In this paper, we introduce texture features from higher-order images, i.e., gradient and curvature images, beyond the intensity image, for that task. METHODS: Based on the Haralick texture analysis method, we introduce a virtual pathological model to explore the utility of texture features from high-order differentiations, i.e., gradient and curvature, of the image intensity distribution. The texture features were validated on a database consisting of 148 colon lesions, of which 35 are non-neoplastic lesions, using the support vector machine classifier and the merit of area under the curve (AUC) of the receiver operating characteristics. RESULTS: The AUC of classification was improved from 0.74 (using the image intensity alone) to 0.85 (by also considering the gradient and curvature images) in differentiating the neoplastic lesions from non-neoplastic ones, e.g., hyperplastic polyps from tubular adenomas, tubulovillous adenomas and adenocarcinomas. CONCLUSIONS: The experimental results demonstrated that texture features from higher-order images can significantly improve the classification accuracy in pathological differentiation of colorectal lesions. The gain in differentiation capability shall increase the potential of computed tomography colonography for colorectal cancer screening by not only detecting polyps but also classifying them for optimal polyp management for the best outcome in personalized medicine.
Authors: Perry J Pickhardt; J Richard Choi; Inku Hwang; James A Butler; Michael L Puckett; Hans A Hildebrandt; Roy K Wong; Pamela A Nugent; Pauline A Mysliwiec; William R Schindler Journal: N Engl J Med Date: 2003-12-01 Impact factor: 91.245
Authors: Perry J Pickhardt; David H Kim; B Dustin Pooler; J Louis Hinshaw; Duncan Barlow; Don Jensen; Mark Reichelderfer; Brooks D Cash Journal: Lancet Oncol Date: 2013-06-07 Impact factor: 41.316
Authors: Thomas M Gluecker; C Daniel Johnson; William S Harmsen; Kenneth P Offord; Ann M Harris; Lynn A Wilson; David A Ahlquist Journal: Radiology Date: 2003-05 Impact factor: 11.105
Authors: Perry J Pickhardt; Bryan Dustin Pooler; David H Kim; Cesare Hassan; Kristina A Matkowskyj; Richard B Halberg Journal: Gastroenterol Clin North Am Date: 2018-06-29 Impact factor: 3.806
Authors: Weiguo Cao; Zhengrong Liang; Marc J Pomeroy; Kenneth Ng; Shu Zhang; Yongfeng Gao; Perry J Pickhardt; Matthew A Barish; Almas F Abbasi; Hongbing Lu Journal: J Med Imaging (Bellingham) Date: 2019-12-27
Authors: Yifan Hu; Zhengrong Liang; Bowen Song; Hao Han; Perry J Pickhardt; Wei Zhu; Chaijie Duan; Hao Zhang; Matthew A Barish; Chris E Lascarides Journal: IEEE Trans Med Imaging Date: 2016-01-18 Impact factor: 10.048