Case-based lung image categorization and retrieval for interstitial lung diseases: clinical workflows

Abstract

The proposed approach is in accordance with the classical workflow of clinicians searching for similar cases in textbooks and personal collections. The developed system enables objective and customizable inter-case similarity assessment, and the performance measures obtained with a leave-one-patient-out cross-validation (LOPO CV) are representative of a clinical usage of the system.

Automated 3D Ιnterstitial Lung Disease Εxtent Quantification: Performance Evaluation and Correlation to PFTs

“…Automated image analysis tools, also referred as computer-aided diagnosis (CAD) schemes, focused on the quantification of ILD extent have been proposed [15][16][17][18][19][20][21][22] to improve efficiency, accuracy, and reproducibility in disease extent assessment. Furthermore, the contribution of CAD schemes in facilitating disease assessment in case of volumetric chest datasets originating from multi-detector CT (MDCT) has been recently highlighted [23][24][25].…”

“…CAD schemes proposed for ILD extent quantification are based on gray level thresholding techniques [15][16][17] and texture-based voxel/pixel classification schemes of lung parenchyma [18][19][20][21][22]. Texture-based image analysis approaches have also been exploited for the differentiation of automatically [26,27] or manually [28][29][30] defined lung parenchyma regions/voxels of interest (VOIs) into various ILD patterns, allowing for a coarse identification of disease patterns and not for a detailed disease extent quantification scheme.…”

“…CAD schemes for ILD extent quantification are differentiated with respect to the 2D or 3D disease extent output. Most CAD schemes exploit high-resolution CT (HRCT) image acquisition protocols [18,19,21,22] which limit the coverage of the entire lung parenchyma volume and lack anatomic comparability in case of follow-up studies [25,26], while only a few CAD schemes are applicable on MDCT image data thus providing volumetric quantification of disease extent [15][16][17]20].…”

“…The importance of estimating the contribution of each specific pattern in disease progression has been recently highlighted [9], rendering the development of CAD schemes capable of identifying and quantifying distinct reticular and ground glass patterns crucial. Nevertheless, only a few CAD schemes allow for the quantification of distinct reticular and ground glass patterns [18,20,22]. The development of CAD schemes that provide 3D quantification of ILD extent and are further capable of quantifying distinct reticular and ground glass patterns accounts for an ongoing research issue [20].…”

Sistematização da percepção médica na construção de sistemas para recuperação de imagens por conteúdo

“…Em imagens médicas, tecidos de mama (Pereira-Jr. et al 2007), osso (Rubin et al 2003), encéfalo (Balan et al 2005) e pulmão (Depeursinge et al 2012) (Wang et al 2006) já apresentam trabalhos de reconhecimento de texturas com um bom grau de evolução.…”