Focal Malignant Liver Lesions: Diagnosis by Dynamic Incremental CT, Early Phase

The purpose of our study was to evaluate the accuracy of dynamic incremental bolus-enhanced conventional CT (DICT) with intravenous contrast administration, early phase, in the diagnosis of malignancy of focal liver lesions. A total of 122 lesions were selected in 74 patients considering the following criteria: lesion diameter 10 mm or more, number of lesions less than six per study, except in multiple angiomatosis and the existence of a valid criteria of definitive diagnosis. Lesions were categorized into seven levels of diagnostic confidence of malignancy compared with the definitive diagnosis for acquisition of a receiver-operator-characteristic (ROC) curve analysis and to determine the sensitivity and specificity of the technique. Forty-six and 70 lesions were correctly diagnosed as malignant and benign, respectively; there were 2 false-positive and 4 false-negative diagnoses of malignancy and the sensitivity and specificity obtained were 92 and 97%. The DICT early phase was confirmed as a highly accurate method in the characterization and diagnosis of malignancy of focal liver lesions, requiring an optimal technical performance and judicious analysis of existing semiological data.

Automatic Small Bowel Tumor Diagnosis by Using Multi-Scale Wavelet-Based Analysis in Wireless Capsule Endoscopy Images

BACKGROUND: Wireless capsule endoscopy has been introduced as an innovative, non-invasive diagnostic technique for evaluation of the gastrointestinal tract, reaching places where conventional endoscopy is unable to. However, the output of this technique is an 8 hours video, whose analysis by the expert physician is very time consuming. Thus, a computer assisted diagnosis tool to help the physicians to evaluate CE exams faster and more accurately is an important technical challenge and an excellent economical opportunity. METHOD: The set of features proposed in this paper to code textural information is based on statistical modeling of second order textural measures extracted from co-occurrence matrices. To cope with both joint and marginal non-Gaussianity of second order textural measures, higher order moments are used. These statistical moments are taken from the two-dimensional color-scale feature space, where two different scales are considered. Second and higher order moments of textural measures are computed from the co-occurrence matrices computed from images synthesized by the inverse wavelet transform of the wavelet transform containing only the selected scales for the three color channels. The dimensionality of the data is reduced by using Principal Component Analysis. RESULTS: The proposed textural features are then used as the input of a classifier based on artificial neural networks. Classification performances of 93.1% specificity and 93.9% sensitivity are achieved on real data. These promising results open the path towards a deeper study regarding the applicability of this algorithm in computer aided diagnosis systems to assist physicians in their clinical practice.