Combination of imageomics and histopathology for prognostic analysis of non-small cell lung cancer
Principal Investigator
Name
Liangxing Li
Degrees
M.M
Institution
Southern Medical University
Position Title
Laboratory Technician
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1092
Initial CDAS Request Approval
Jul 6, 2023
Title
Combination of imageomics and histopathology for prognostic analysis of non-small cell lung cancer
Summary
Lung cancer is one of the most common and deadly malignancies worldwide, with non-small cell lung cancer (NSCLC) accounting for 85 % of all primary lung cancers. The TNM staging system is an important tool for postoperative prognosis, which can guide patient treatment. Lung cancer can be classified into one to four stages based on the size, location, and presence of lymph node metastasis and distant metastasis of the tumor. However, among NSCLC patients who underwent complete tumor resection at the same stage of the disease, the survival time varies greatly, with some patients experiencing recurrence shortly after surgery and others achieving complete cure, indicating the need for new prognostic assessment methods.
Radiomics can achieve quantitative processing. It can extract tumor imaging features from images at high throughput: quantifying tumor size, shape, texture, and more. Therefore, imageomics has been widely applied in predicting cancer survival rates, tumor staging, and tumor histology.
With the widespread use of full slide tissue scanning, computer image processing technology can improve the efficiency, accuracy, and consistency of histopathological evaluation. Pathohistology is also born, which can capture various data from pathological images in high-throughput and analyze them to generate quantitative features that characterize different phenotypes of tissue samples. The types of image features include cell size, shape, pixel intensity distribution in cells and nuclei, and texture of cells and nuclei. For example, automatically identifying tissues such as tumors, fat, and blood cells in slices like images, so the machine learning-based histopathology method can be used for prognostic analysis of various cancers.
In this study, we try to integrate the radiological features of CT scan and the pathological features of histopathological images, and combine machine learning to establish a prognostic evaluation model for NSCLC.
Radiomics can achieve quantitative processing. It can extract tumor imaging features from images at high throughput: quantifying tumor size, shape, texture, and more. Therefore, imageomics has been widely applied in predicting cancer survival rates, tumor staging, and tumor histology.
With the widespread use of full slide tissue scanning, computer image processing technology can improve the efficiency, accuracy, and consistency of histopathological evaluation. Pathohistology is also born, which can capture various data from pathological images in high-throughput and analyze them to generate quantitative features that characterize different phenotypes of tissue samples. The types of image features include cell size, shape, pixel intensity distribution in cells and nuclei, and texture of cells and nuclei. For example, automatically identifying tissues such as tumors, fat, and blood cells in slices like images, so the machine learning-based histopathology method can be used for prognostic analysis of various cancers.
In this study, we try to integrate the radiological features of CT scan and the pathological features of histopathological images, and combine machine learning to establish a prognostic evaluation model for NSCLC.
Aims
Integrate the radiological features of CT scan and the pathological features of histopathological images, and combine machine learning to establish a prognostic evaluation model for NSCLC.
Collaborators
none