The CT Imaging Character And Differential Diagnosis About Solitary Pulmonary Nodules And Computer Aided Detection

Part ⅠThe CT imaging character and d ifferential diagnosis about solitary pulmonary nodulesObjective:To discuss the diagnosis value of multi-spire CT on the solitary pulmonary nodules.Methods:To collect the imaging data of156patients those had been approved benign and malignant in our hospital during July2005to March2011(all cases underwent plain and enhanced scans, and some nodules those have difficulties in diagnosis apply computer-assisted analysis), analysis morphological features, the enhanced value and enhanced form after enhanced about solitary pulmonary nodules. Above Image characteristics were compared with pathological results,in order to achieve the differential diagnosis between benign and malignant pulmonary nodules.Results:The nodules group with lobed,burr, spiculate bubble, vacuole sign, air bronchogram, and changes in the surrounding glassy ground, blood vessels and pleural indentation sign cluster was significantly higher than benign nodules group, the cavity was less than benign nodules group.there were significant differences (P <0.01). Enhanced scan to20HU as a study value,80cases of90cases of malignant nodules in peak enhancement between the20～60HU,8cases less than20HU,2cases higher than60HU;66cases of benign nodules in54patients with less than20HU, including28cases those been calcificated more obviously enhance less, the other10cases strengthen more than60HU.2patients in the enhanced value of between20～60HU.78cases of malignant nodules within the peak in lmin, and the form of parabola, benign nodules time-density curves of various shapes, the time to peak of varying lengths.Conclusions:1. Major CT signs of malignant nodules include:lobulation, spiculation, vacuole sign, air bronchogram, pleural indentation and vascular convergence sign sign, part of the malignant nodules were ground glass.2. Malignant nodules enhanced features:peak enhancement in the20～60HU, less than3cm and more uniform enhancement of lung cancer, a few can be heterogeneous enhancement greater than3cm of lung cancer is heterogeneous enhancement can also be manifested as necrosis of the center peripheral enhancement, different histological types of lung cancer is also enhanced form of difference. Malignant nodules time-density curves reach peak in1min, and the form of parabola.3. Benign and malignant nodules in the differential diagnosis:malignant nodules lobulation, spiculation, vacuole sign, air bronchogram, vascular convergence sign and pleural indentation were significantly higher than benign nodules, benign nodules in CT images calcified significantly more malignant nodules, benign nodules or enhancement value≤20HU、≤60HU, its time-density curves of various shapes, the time to peak of varying lengths. The computer-aided separation and diagnosis of difficult diagnosis pulmonary nodulesObjective:Using computer-aided segmentation to difficult diagnosis pulmonary nodules (the nodules connected with the pleura, blood vessels and ground-glass opacity nodules), and with pathologic findings were compared in general form, that achieves the purpose of diagnosis, evaluate the value of computer-aided segmentation in difficult diagnosis lung nodulesMethods:In this study,52cases of difficult diagnosis pulmonary nodules,14cases connected with the pleura,18cases associated with blood vessels,20patients with ground-glass opacity nodules use mathematical morphology and the random walk method assist segmentation, extract the pulmonary nodules, than the general form of pathological results were compared, the two were statistically analyzed.Results:52cases of difficult pulmonary nodules removed the blood vessels surrounding the pleura by above methods, the lung nodule extract, and ground-glass nodules have been good edge segmented., in which, the pathological results of43cases have no obvious divarication with the general statistically significant.Conclusions:Using computer-aided diagnosis of nodular segmentation, it Remove the influence of the vascular or pleural from the nodules and indicate the edge of the shape and size of the nodule size, it is beneficial for benign nodules, malignant identification. Film-reading in a doctor first before computer-aided detection, the image information to doctors nodule diagnosis, which will greatly reduce the number of doctors read the piece, to improve the efficiency of doctors read the piece, while overcoming the inertia and the human eye Gray is not sensitive to defects, improve the detection accuracy, thereby effectively improve the treatment success rate.