<title>Image processing for computer-aided diagnosis of lung cancer screening system by CT (LSCT)</title>

Abstract

in this paper, we report the image processing technique for computer-aided diagnosis of lung cancer screening system by CT (LSCT). LSCT is the newly developed mobile-type CT scanner for the mass screening of lung cancer by our project team. in this new LSCT system, one essential problem is the increase of image information to be diagnosed by a doctor to about 30 slices per patient from I X-ray film. To solve this difficult problem, we are trying to reduce the image information drastically to be displayed for the doctor by image processing techniques. We propose a new method named Variable New-Quoit filter for the automatic recognition of the pathological shadow candidates. Our computer aided diagnosis (CAD) system can satisfactoiily reduce the number of CT cross sections by this method, containing the abnormal shadow candidates.

Computer-assisted detection of pulmonary nodules: evaluation of diagnostic performance using an expert knowledge-based detection system with variable reconstruction slice thickness settings

“…Due to the inherently high contrast between soft tissues and lung tissues, it represents a particularly promising tool for optimized automated detection of pulmonary nodules from MSCT datasets. Several CAD approaches are currently undergoing clinical evaluation with preliminary evidence that CAD may be suited to guide the radiologist to suspicious lesions [20,[25][26][27][28][29][30][31]. Mathematical models for computer-aided detection of pulmonary nodules can be broadly divided into two categories: density-based approaches using the high density interval between the nodule and the pulmonary parenchyma employ techniques such as multiple thresholding [7,25,26], region-growing [20], locally adaptive thresholding in combination with region-growing [27] and fuzzy clustering [28] for nodule identification.…”

“…For this purpose, approaches such as "N-Quoit filter" [29], template-matching [30], object-based deformation [31] and the anatomy-based generic model [32] have been developed. In addition analysis of curved surface morphology [33,34] and subtraction of bronchovascular structures from the chest CT images [35] have been used for nodule detection.…”

Computer-assisted detection of pulmonary nodules: performance evaluation of an expert knowledge-based detection system in consensus reading with experienced and inexperienced chest radiologists

“…Various CAD tools using the inherently high contrast between the nodular and lung tissues as platform for nodule recognition within density-based or model-based algorithms have been developed for analysis of thin section CT datasets [1,5,[12][13][14][15]38]. However, the currently published data is scarce, with only few studies having been performed to apply CAD to more than 50 nodules.…”

“…Consensus of reader 1+CAD significantly outperformed all other readings, demonstrating a benefit in using CAD as an inexperienced reader replacement. It is questionable whether inexperienced readers can be regarded as adequate for interpretation of pulmonary nodules in consensus with CAD, replacing an experienced radiologist.Computer-assisted detection of pulmonary nodules: performance evaluation of an expert knowledge-based detection system in consensus reading with experienced and inexperienced chest radiologists mance and may enhance the detection of suspicious lesions [1,[10][11][12][13][14][15][16]. However, CAD systems for automated lung nodule detection employing various density-based or model-based recognition algorithms have been predominantly tested on small numbers of artificial or in vivo pulmonary nodules without standardized databases that impede the comparison of detection performances between different authors.…”

Recent progress in computer-aided diagnosis of lung nodules on thin-section CT

“…Similarly, CAD schemes for nodule detection in thick-section CT images have been developed by many investigators [41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56]. The typical performance of current CAD schemes in thicksection CT is an 80-90% sensitivity with 1-2 false positives per section, which is translated into tens of false positives per CT scan.…”