The invasion depth measurement of bladder cancer using T2-weighted magnetic resonance imaging
- PDF / 1,410,218 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 72 Downloads / 154 Views
RESEARCH
BioMedical Engineering OnLine Open Access
The invasion depth measurement of bladder cancer using T2‑weighted magnetic resonance imaging Yang Liu1† , Haojie Zheng2†, Xiaopan Xu1†, Xi Zhang1, Peng Du1, Jimin Liang2* and Hongbing Lu1* *Correspondence: [email protected]; [email protected] † Yang Liu, Haojie Zheng and Xiaopan Xu contributed equally to this work, and they are co-first authors. 1 School of Biomedical Engineering, Air Force Medical University, No. 169 Changle West Road, Xi’an, Shaanxi 710032, PR China 2 School of Life Sciences and Technology, Xidian University, 266 Xinglong Section of Xifeng Road, Xi’an, Shaanxi 710126, PR China
Abstract Background: Invasion depth is an important index for staging and clinical treatment strategy of bladder cancer (BCa). The aim of this study was to investigate the feasibility of segmenting the BCa region from bladder wall region on MRI, and quantitatively measuring the invasion depth of the tumor mass in bladder lumen for further clinical decision-making. This retrospective study involved 20 eligible patients with postoperatively pathologically confirmed BCa. It was conducted in the following steps: (1) a total of 1159 features were extracted from each voxel of both the certain cancerous and wall tissues with the T2-weighted (T2W) MRI data; (2) the support vector machine (SVM)based recursive feature elimination (RFE) method was implemented to first select an optimal feature subset, and then develop the classification model for the precise separation of the cancerous regions; (3) after excluding the cancerous region from the bladder wall, the three-dimensional bladder wall thickness (BWT) was calculated using Laplacian method, and the invasion depth of BCa was eventually defined by the subtraction of the mean BWT excluding the cancerous region and the minimum BWT of the cancerous region. Results: The segmented results showed a promising accuracy, with the mean Dice similarity coefficient of 0.921. The “soft boundary” defined by the voxels with the probabilities between 0.1 and 0.9 could demonstrate the overlapped region of cancerous and wall tissues. The invasion depth calculated from proposed segmentation method was compared with that from manual segmentation, with a mean difference of 0.277 mm. Conclusion: The proposed strategy could accurately segment the BCa region, and, as the first attempt, realize the quantitative measurement of BCa invasion depth. Keywords: Bladder cancer, Support vector machine, Feature selection, Segmentation, Invasion depth
Background Bladder cancer (BCa) is the sixth-most common cancer in male worldwide [1–4]. It is estimated that 549, 000 new cases and 200,000 deaths occurred every year, with threequarters of them occurring in men [1, 5]. Based on National Comprehensive Cancer
© The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit
Data Loading...
