The Preliminary Research On Computer-aided Detection Of Lung Nodule Based On Ct Images

Objective To research the methods of computer-aided analysis of CT image and computer-aided detection(or diagnosis) of lung nodule and their software realization.Materials and Methods In the experiments, we selected images of 20 patients underwent CT examination and were diagnosed as lung cancer patients to construct our experiment database. There are about 90-120 images per-patient and totally about 2,000 images. We only selected 4 images containing suspicious nodules per-patient to construct the experiment database. In order to test our algorithm, we processed 40 images totally containing 42 nodules with varies size.All the images were produced by SIMENS VOLUME ZOOM 4 spiral CT workstation with 512x512 pixel matrix size,16-bit allocation,12-bit depth and 4096 gray-level. The pixel size is 0.74 mmx0.74 mm with 5 mm or 7 mm slice thickness.Firstly, gray-level threshold and region-grow algorithm were applied on the image to segment lung fields and correction were applied if necessary, then multi-scale nodule enhancement filter and curved morphological analysis technologies were used to enhance, detect and display the lung nodules(ROI). In order to measure some indices of the segmented nodule, we also performed interactive region-growing technique and artificial neural networks (ANN) classifier on the nodule region. In order to classify nodules in 3-dimensional,3-dimensional reconstruction was used.Results In this paper, we proposed and realized a lung nodule computer-aided detection system based on multi-scale nodule enhancement, interactive region-growing segmentation, artificial neural network classifier and 3-dimensional technologies. By using VC++6.0 and OpenGL to develop the software, we realized the 2-Dimensional and 3-Dimensional detection and displaying of lung nodules. We processed 40 images containing 42 nodules with varies size. An experienced radiologist with 10 years experience was asked to detect all the nodules without CAD. And one month later the same radiologist was asked to review all the images to detect nodules.with CAD. RGC analysis was used to assess the performance of our CAD sytem, the area of ROC(AZ) is 0.96 with CAD, but without CAD,the area is 0.89.Conclusion We proposed and realized a lung nodule computer-aided detection system. Computer-aided diagnosis is an effective method for the improvements of lung nodule detection performance.