Pixel-Level Classification of Five Histologic Patterns of Lung Adenocarcinoma

Shao, Dan3,4; Su, Fei4,5; Zou, Xueyu6; Lu, Jie4; Wu, Sitong4,5; Tian, Ruijun7; Ran, Dongmei2; Guo, Zhiyong1; Jin, Dayong4,5

2023-02-07

10.1021/acs.analchem.2c03020

ANALYTICAL CHEMISTRY   影响因子和分区

0003-2700

1520-6882

Lung adenocarcinoma is the most common histologic type of lung cancer. The pixel-level labeling of histologic patterns of lung adenocarcinoma can assist pathologists in determining tumor grading with more details than normal classification. We manually annotated a dataset containing a total of 1000 patches (200 patches for each pattern) of 512 x 512 pixels and 420 patches (contains test sets) of 1024 x 1024 pixels according to the morphological features of the five histologic patterns of lung adenocarcinoma (lepidic, acinar, papillary, micropapillary, and solid). To generate an even large amount of data patches, we developed a data stitching strategy as a data augmentation for classification in model training. Stitched patches improve the Dice similarity coefficient (DSC) scores by 24.06% on the whole-slide image (WSI) with the solid pattern. We propose a WSI analysis framework for lung adenocarcinoma pathology, intelligently labeling lung adenocarcinoma histologic patterns at the pixel level. Our framework contains five branches of deep neural networks for segmenting each histologic pattern. We test our framework with 200 unclassified patches. The DSC scores of our results outpace comparing networks (U-Net, LinkNet, and FPN) by up to 10.78%. We also perform results on four WSIs with an overall accuracy of 99.6%, demonstrating that our network framework exhibits better accuracy and robustness in most cases.

SCI ; EI

英语

NI论文

通讯

Shenzhen Science and Technology Program["KQTD20170810110913065","20200925174735005"]

Chemistry

Chemistry, Analytical

WOS:000929166900001

AMER CHEMICAL SOC

20230513493653

Biological organs ; Classification (of information) ; Deep neural networks ; Grading ; Image enhancement ; Image segmentation ; Statistical tests

Biological Materials and Tissue Engineering:461.2 ; Ergonomics and Human Factors Engineering:461.4 ; Information Theory and Signal Processing:716.1 ; Information Sources and Analysis:903.1 ; Mathematical Statistics:922.2

CHEMISTRY

Web of Science

1.Southern Univ Sci & Technol, UTS SUSTech Joint Res Ctr Biomed Mat & Devices, Dept Biomed Engn, Guangdong Prov Key Lab Adv Biomat, Shenzhen 518055, Peoples R China
2.Southern Univ Sci, Technol Hosp, Dept Pathol, Shenzhen 518055, Peoples R China
3.Yangtze Univ, Sch Elect & Informat, Jingzhou 434023, Peoples R China
4.Southern Univ Sci & Technol, UTS SUSTech Joint Res Ctr Biomed Mat & Devices, Dept Biomed Engn, Shenzhen 518055, Peoples R China
5.Univ Technol Sydney, Inst Biomed Mat & Devices IBMD, Fac Sci, Sydney, NSW 2007, Australia
6.Yangtze Univ, Sch Elect & Informat, Jingzhou 434023, Peoples R China
7.Southern Univ Sci & Technol, Dept Chem, Shenzhen 518055, Peoples R China

Shao, Dan,Su, Fei,Zou, Xueyu,et al. Pixel-Level Classification of Five Histologic Patterns of Lung Adenocarcinoma[J]. ANALYTICAL CHEMISTRY,2023,95(5):2664-2670.

Shao, Dan.,Su, Fei.,Zou, Xueyu.,Lu, Jie.,Wu, Sitong.,...&Jin, Dayong.(2023).Pixel-Level Classification of Five Histologic Patterns of Lung Adenocarcinoma.ANALYTICAL CHEMISTRY,95(5),2664-2670.

Shao, Dan,et al."Pixel-Level Classification of Five Histologic Patterns of Lung Adenocarcinoma".ANALYTICAL CHEMISTRY 95.5(2023):2664-2670.