A Novel Karyoplasmic Ratio-Based Automatic Recognition Method for Identifying Glioma Circulating Tumor Cells.

Abstract

Background:

Detection of circulating tumor cells (CTCs) is a promising technology in tumor management; however, the slow development of CTC identification methods hinders their clinical utility. Moreover, CTC detection is currently challenging owing to major issues such as isolation and correct identification. To improve the identification efficiency of glioma CTCs, we developed a karyoplasmic ratio (KR)-based identification method and constructed an automatic recognition algorithm. We also intended to determine the correlation between high-KR CTC and patients' clinical characteristics.

Methods:

CTCs were isolated from the peripheral blood samples of 68 glioma patients and analyzed using DNA-seq and immunofluorescence staining. Subsequently, the clinical information of both glioma patients and matched individuals was collected for analyses. ROC curve was performed to evaluate the efficiency of the KR-based identification method. Finally, CTC images were captured and used for developing a CTC recognition algorithm.

Results:

KR was a better parameter than cell size for identifying glioma CTCs. We demonstrated that low CTC counts were independently associated with isocitrate dehydrogenase (IDH) mutations (p = 0.024) and 1p19q co-deletion status (p = 0.05), highlighting its utility in predicting oligodendroglioma (area under the curve = 0.770). The accuracy, sensitivity, and specificity of our algorithm were 93.4%, 81.0%, and 97.4%, respectively, whereas the precision and F1 score were 90.9% and 85.7%, respectively.

Conclusion:

Our findings remarkably increased the efficiency of detecting glioma CTCs and revealed a correlation between CTC counts and patients' clinical characteristics. This will allow researchers to further investigate the clinical utility of CTCs. Moreover, our automatic recognition algorithm can maintain high precision in the CTC identification process, shorten the time and cost, and significantly reduce the burden on clinicians.

• Authors: Xinyi Zhu, Shen Wen, Shuhang Deng, Gao Wu, Ruyong Tian, Ping Hu, Liguo Ye, Qian Sun, Yang Xu, Gang Deng, Dong Zhang, Shuang Yang, Yangzhi Qi, Qianxue Chen
• Journal: Frontiers in oncology (Front Oncol) Vol. 12 Pg. 893769 ( 2022) ISSN: 2234-943X [Print] Switzerland
• PMID: 35646680 (Publication Type: Journal Article)
• Copyright: Copyright © 2022 Zhu, Wen, Deng, Wu, Tian, Hu, Ye, Sun, Xu, Deng, Zhang, Yang, Qi and Chen.