Introduction For women diagnosed with advanced-stage epithelial ovarian cancer, complete cytoreductive surgery (CRS) is the most powerful independent parameter for prolonged survival. An intraoperative imaging technique to detect tumor deposits could help achieve complete CRS. Hyperspectral imaging (HSI) provides information on tissue composition, including tissue oxygenation, hemoglobin-, and tissue water and fat indices. In an earlier ex-vivo study it was shown that HSI can be used for ex-vivo tumor detection. Currently, we are evaluating whether HSI can be used in-vivo to distinguish tumor from healthy tissue.
Methods HSI data of healthy and tumorous peritoneum, omentum, ovary, and mesentery were obtained in-vivo and preprocessed by image calibration and glare removal. The data were correlated to histopathology and used to train classifiers. The ability to delineate tumorous from healthy tissues was determined using leave-one-out cross-validation.
Results A total of 18 images from 12 patients were included. In total 302.258 data points were extracted based on the knowledge of the surgeons and histopathological information. Our data shows that different organs that are affected by ovarian cancer yield different spectra. Additionally, we observe a difference in the spetra of tumor and non-tumor tissue.
Conclusion/Implications HSI enables the classification of various tissue types, including tumor and non-tumor tissue. To improve classification outcomes, it is crucial to obtain more data and to make separate groups of healthy and tumorous tissues for each of these tissue types. HSI is a promising technique to differentiate between healthy tissue and ovarian cancer lesions and eventually help surgeons to achieve complete CRS.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.