Flat neoplasia of the colon: CT colonography with CAD

Taylor, Stuart A.; Suzuki, Noriko; Beddoe, Gareth; Halligan, Steve

doi:10.1007/s00261-008-9425-3

Cited by 11 publications

(3 citation statements)

References 68 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although current CAD schemes could be useful for detection of polypoid polyps, the detection of flat lesions is a major challenge [106], because existing CAD schemes have focused on the detection of pedunculated and sessile polyps; thus, they are designed for detecting the common polypoid shape. Existing CAD schemes use geometric, morphologic, and textural characteristics to distinguish polyps from normal structures in the colon (e.g., haustral folds, stool, the air/liquid boundary, the ileocecal valve, and a rectal catheter).…”

Section: Limitations Of Current Cad Schemes For Flat-lesion Detectionmentioning

confidence: 99%

Pixel-based Machine Learning in Computer-Aided Diagnosis of Lung and Colon Cancer

Suzuki

2013

Intelligent Systems Reference Library

View full text Add to dashboard Cite

Computer-aided diagnosis (CAD) for detection of lesions in medical images has been an active area of research. Machine learning plays an essential role in CAD, because representing lesions and organs requires a complex model that has a number of parameters to determine; thus, medical pattern recognition essentially requires ''learning from examples'' to determine the parameters of the model. Machine learning has been used to classify lesions into certain classes (e.g., abnormal or normal, lesions or non-lesions, and malignant or benign) in CAD. Recently, as available computational power increased dramatically, pixel/voxel-based machine learning (PML) has emerged in medical image processing/analysis, which uses pixel/voxel values in local regions (or patches) in images instead of features calculated from segmented regions as input information; thus, feature calculation or segmentation is not required. Because PML can avoid errors caused by inaccurate feature calculation and segmentation, the performance of PML can potentially be higher than that of common classifiers. In this chapter, MTANNs (a class of PML) in CAD schemes for detection of lung nodules in CT and for detection of polyps in CTC are presented.

show abstract

Section: Limitations Of Current Cad Schemes For Flat-lesion Detectionmentioning

confidence: 99%

Pixel-based Machine Learning in Computer-Aided Diagnosis of Lung and Colon Cancer

Suzuki

2013

Intelligent Systems Reference Library

View full text Add to dashboard Cite

show abstract

“…Many experts believe that flat neoplasms may be the precursor lesions accounting for a significant proportion of these cancers [8]. Suzuki et al [9] reported that flat cancers represent approximately 70% of T1 lesions and up to 10% of all cancers (in 1,026 patients).…”

Section: Flat Lesionsmentioning

confidence: 99%

“…The majority of CRCs develop from adenomatous polyps following the adenoma-carcinoma sequence; the remaining do not follow this developing pattern [6][7][8]. Many experts believe that flat neoplasms may be the precursor lesions accounting for a significant proportion of these cancers [8].…”

Section: Flat Lesionsmentioning

confidence: 99%

Flat lesions missed at conventional colonoscopy (CC) and visualized by CT colonography (CTC): a pictorial essay

Coppola

Regge²,

Flor

et al. 2013

Abdom Imaging

View full text Add to dashboard Cite

The purpose of this study is to describe our experience with cases of false negative findings at conventional colonoscopy (CC) that were identified by CT colonography (CTC). Conventional colonoscopy (CC) is the universally accepted gold-standard technique for the diagnosis of colonic polyps and cancers, however occasionally this method can generate false negative findings. We present examples of false negatives at CC, correctly identified by CT colonography (CTC), and later confirmed at a second endoscopy, describing the reasons of false negative, when possible.CT colonography is a highly sensitive technique for polyps and colorectal cancer (CRC) detection [1] across a broad spectrum of accepted indications and protocols (preparation, fecal tagging, CO 2 automatic insufflation, hypotonization). Its sensitivity is high especially for detecting polyps greater than 10 mm; however, CTC is less accurate than CC for small and diminutive polyps [1][2][3][4].However, when CC is performed by inexperienced operators and non-gastroenterologists or when lesions have atypical morphologies disappointing low sensitivities may occasionally be reported [5]. The commonalities of typical lesions missed at CC usually concern their location and shape: false negative results can be generated when incomplete colonoscopies are performed (cecum not reached), when lesions are hidden behind haustral folds or when they are flat or present atypical morphologies. On the other side, typical lesions missed at CTC are usually located in the rectosigmoid colon and in the splenic flexure, due to a minor effect of the insufflation in this segments [1]. Right colon is usually easier to distend and consequently evaluate at CTC, while the increased distance in CC makes right colon assessment more difficult. For the above reported reasons CT colonography and CC should be considered equivalent in terms of sensitivity in CRC detection and complementary [1].

show abstract

Computer-Aided Diagnosis in Computed Tomographic Colonography

Suzuki

Dachman²

2010

Atlas of Virtual Colonoscopy

View full text Add to dashboard Cite

Flat neoplasia of the colon: CT colonography with CAD

Cited by 11 publications

References 68 publications

Pixel-based Machine Learning in Computer-Aided Diagnosis of Lung and Colon Cancer

Pixel-based Machine Learning in Computer-Aided Diagnosis of Lung and Colon Cancer

Flat lesions missed at conventional colonoscopy (CC) and visualized by CT colonography (CTC): a pictorial essay

Computer-Aided Diagnosis in Computed Tomographic Colonography

Contact Info

Product

Resources

About