Three-dimensional lung nodule segmentation and shape variance analysis to detect lung cancer with reduced false positives

Abstract

The three-dimensional analysis on lung computed tomography scan was carried out in this study to detect the malignant lung nodules. An automatic three-dimensional segmentation algorithm proposed here efficiently segmented the tissue clusters (nodules) inside the lung. However, an automatic morphological region-grow segmentation algorithm that was implemented to segment the well-circumscribed nodules present inside the lung did not segment the juxta-pleural nodule present on the inner surface of wall of the lung. A novel edge bridge and fill technique is proposed in this article to segment the juxta-pleural and pleural-tail nodules accurately. The centroid shift of each candidate nodule was computed. The nodules with more centroid shift in the consecutive slices were eliminated since malignant nodule's resultant position did not usually deviate. The three-dimensional shape variation and edge sharp analyses were performed to reduce the false positives and to classify the malignant nodules. The change in area and equivalent diameter was more for malignant nodules in the consecutive slices and the malignant nodules showed a sharp edge. Segmentation was followed by three-dimensional centroid, shape and edge analysis which was carried out on a lung computed tomography database of 20 patient with 25 malignant nodules. The algorithms proposed in this article precisely detected 22 malignant nodules and failed to detect 3 with a sensitivity of 88%. Furthermore, this algorithm correctly eliminated 216 tissue clusters that were initially segmented as nodules; however, 41 non-malignant tissue clusters were detected as malignant nodules. Therefore, the false positive of this algorithm was 2.05 per patient.

Automated delineation of non‐small cell lung cancer: A step toward quantitative reasoning in medical decision science

“…The Laplace operator, which uses second derivatives, can localize the direction of the pixels along the edges, but also suffers from sensitivity to local irregularities. Canny detector, on the other hand, provides better results compared to the gradient and Laplacian operator, since it combines the operation of filtering, enhancing, and detecting the target object . High false negatives due to noise sensitivity are the major limitation of an edge‐based model, and various morphological operators may be required to further post‐process the segmented object.…”

Automated delineation of non-small cell lung cancer

“…By invading the surrounding structures, such as blood vessels, chest and mediastinal walls, the NSCLCs may occlude the existent boundaries of the lungs, which increases the difficulty of accurately segmenting the lung thoracic areas on the chest CT images and hinders the subsequent detection and segmentation of lung cancers [17]. The existing algorithms for delineating lung nodules are also insufficiently efficient to segment the relatively larger NSCLC masses [3][4][5][6][7]. To rectify this problem, an algorithm was proposed that acknowledges the radio-morphological condition of a NSCLC.…”

“…NSCLCs are often large-sized with an irregular shape and invasion of the surrounding structures in comparison with lung nodules. These features present challenges and cause the failure of automatic segmentation and detection algorithms developed for small lung nodules [3][4][5][6][7], particularly for the following reasons: (1) NSCLCs are often attached to the pleural or mediastinal wall and are easily excluded during thoracic segmentation. (2) Existing algorithms always assume a small, solid, and round lung nodule, in contrast to the characteristic features of NSCLC.…”

Lungs Nodule Detection Using Semantic Segmentation and Classification with Optimal Features

“…The morphological operations have been applied to remove the vessel cavities. Then, juxta pleural nodule segmentation was done using edge bridge, fill technique, 3D shape, and edge analysis to differentiate malignant nodules from benign nodules [20]. Clusters of highlighted voxels > 3mm was treated as nodule candidate [16].…”

“…Over the past decade, lung cancer cases declined by 2% per year in men and almost remains stable in women [30]. However, the survival rate can be increased if the cancer is detected in its early stages [20]. For the early detection of lung cancer, several modalities are used such as Computed Tomography (CT) scan, X-rays and the Magnetic Resonance Imaging (MRI).…”