Segmentation of lung computed tomography images based on SegNet in the diagnosis of lung cancer

Background: The diagnosis of lung cancer (LC) may be difficult to make in the elderly. We report on the diagnostic elements available in life in an LC necropsy case series of asbestos-exposed workers and describe the frequency of non-neoplastic asbestos-related diseases as biological exposure indices. Methods: We reviewed pathologic and clinical records of an unselected series of autopsies (1997–2016) in patients with LC employed in the Monfalcone shipyards. We assessed the consistency with autopsy results of diagnoses based on, respectively, radiologic, cytologic, and histologic findings. Results: Data on 128 autopsy-confirmed LC cases were available; in life, 119 had been diagnosed as LC. Among these, 49 had histologic confirmation of diagnosis (17 with immunophenotyping); histology had been negative in 4. Cytology had been the main positive finding and the basis for diagnosis in 24 cases, but had been negative in 13. Chest computed tomography had been the basis for diagnosis in 45; in 18 cases, it had been negative. Nine patients had received a diagnosis different from LC, among whom 4 had been suspected to have malignant pleural mesothelioma by chest computed tomography. Pleural plaques were found in 124 and histologic asbestosis in 46 cases. Conclusions: Autopsies confirmed all LC diagnoses received in life, including 46 that would have been considered only possible LC based on clinical workup. The overall survival in this case series was poor. The high prevalence of pleural plaques and asbestosis suggest severity of asbestos exposures.

Download Full-text

Automatic Segmentation of the Gross Target Volume in Non-Small Cell Lung Cancer Using a Modified Version of ResNet

Technology in Cancer Research & Treatment ◽

10.1177/1533033820947484 ◽

2020 ◽

Vol 19 ◽

pp. 153303382094748

Author(s):

Fuli Zhang ◽

Qiusheng Wang ◽

Haipeng Li

Keyword(s):

Computed Tomography ◽

Lung Cancer ◽

Small Cell Lung Cancer ◽

Cell Lung Cancer ◽

Automatic Segmentation ◽

Target Volume ◽

Small Cell ◽

Small Cell Lung ◽

Computed Tomography Images ◽

Positive Rate

Radiotherapy plays an important role in the treatment of non-small cell lung cancer. Accurate segmentation of the gross target volume is very important for successful radiotherapy delivery. Deep learning techniques can obtain fast and accurate segmentation, which is independent of experts’ experience and saves time compared with manual delineation. In this paper, we introduce a modified version of ResNet and apply it to segment the gross target volume in computed tomography images of patients with non-small cell lung cancer. Normalization was applied to reduce the differences among images and data augmentation techniques were employed to further enrich the data of the training set. Two different residual convolutional blocks were used to efficiently extract the deep features of the computed tomography images, and the features from all levels of the ResNet were merged into a single output. This simple design achieved a fusion of deep semantic features and shallow appearance features to generate dense pixel outputs. The test loss tended to be stable after 50 training epochs, and the segmentation took 21 ms per computed tomography image. The average evaluation metrics were: Dice similarity coefficient, 0.73; Jaccard similarity coefficient, 0.68; true positive rate, 0.71; and false positive rate, 0.0012. Those results were better than those of U-Net, which was used as a benchmark. The modified ResNet directly extracted multi-scale context features from original input images. Thus, the proposed automatic segmentation method can quickly segment the gross target volume in non-small cell lung cancer cases and be applied to improve consistency in contouring.

Download Full-text