Comparative evaluation of conventional and deep learning methods for semi-automated segmentation of pulmonary nodules on CT

Abstract

Background: Accurate segmentation of pulmonary nodules on computed tomography (CT) scans plays a crucial role in the evaluation and management of patients with suspicion of lung cancer (LC). When performed manually, not only the process requires highly skilled operators, but is also tiresome and timeconsuming. To assist the physician in this task several automated and semi-automated methods have been proposed in the literature. In recent years, in particular, the appearance of deep learning has brought about major advances in the field.Methods: Twenty-four (12 conventional and 12 based on deep learning) semi-automated-'one-click'methods for segmenting pulmonary nodules on CT were evaluated in this study. The experiments were carried out on two datasets: a proprietary one (383 images from a cohort of 111 patients) and a public one (259 images from a cohort of 100). All the patients had a positive transcript for suspect pulmonary nodules. Results:The methods based on deep learning clearly outperformed the conventional ones. The best performance [Sørensen-Dice coefficient (DSC)] in the two datasets was, respectively, 0.853 and 0.763 for the deep learning methods, and 0.761 and 0.704 for the traditional ones.Conclusions: Deep learning is a viable approach for semi-automated segmentation of pulmonary nodules on CT scans.

CircFOXP1: A novel serum diagnostic biomarker for non-small cell lung cancer

“…28 Bianconi et al discovered that deep learning is a viable approach for semi-automated segmentation of pulmonary nodules on CT scans. 26 However, these approaches presented some challenges. For semi-automated quantification methods, much larger datasets were needed to complete the calculation.…”

“…24 Semi-automated quantification methods with radiomics could reduce the false positive results significantly. [25][26][27] Spanu et al found that TF SPECT was a highly accurate diagnostic method in the detection of intrathoracic malignant lesions. 28 Bianconi et al discovered that deep learning is a viable approach for semi-automated segmentation of pulmonary nodules on CT scans.…”

Impact of Lesion Delineation and Intensity Quantisation on the Stability of Texture Features from Lung Nodules on CT: A Reproducible Study

“…However, for some features simple mathematical transformations could be applied to make the features independent of the number of quantisation levels (see for instance Appendix A ). In order to avoid or reduce the inter-observer bias related to manual lesion delineation, automated and semi-automated methods offer great promises in terms of speed, accuracy and repeatability [ 65 ]. Previous studies have shown that semi-automated segmenters can improve on manual delineation and generate more reproducible radiomics features [ 37 , 38 ].…”

How AI Can Help in the Diagnostic Dilemma of Pulmonary Nodules

“…Accurate nodule volumetry requires good nodule segmentation. Manual segmentation of lung nodules is time consuming and is far less accurate in comparison to deep learning semiautomated methods [ 120 ]. Most of the available algorithms concerned with pulmonary nodule detection rely on growing edge method where a predefined threshold acts as a seed that connects all nearby voxels of higher density [ 121 ].…”