Methods of Intraoperative Diagnosis in Puncture Minimally Invasive Surgery of Liver Cancer

Introduction. Liver cancer is a leading cause of death in oncology. The final diagnosis is determined by a pathomorphological analysis of tissue specimens obtained during percutaneous puncture biopsy. Despite its obvious advantages, this method is associated with the possibility of obtaining nondiagnostic specimens and the need for long wait times. Therefore, additional diagnostic methods should be developed to improve the quality of surgical care for oncology patients. Optical methods are a sensitive tool for determining the metabolic characteristics of biotissues. Such methods may improve the efficacy of conventional puncture procedures by developing approaches for rapid diagnosis of liver neoplasm types.Aim. Development of intraoperative diagnostic methods for in vivo minimally invasive liver cancer surgery that allow differentiation between liver parenchyma and tumors, as well as classification of neoplasm types (primary malignant, metastatic, and benign) based on optical spectroscopy and machine learning.Materials and methods. The methods of clinical research, descriptive mathematical statistics, and machine learning were used.Results. Spectroscopic methods of intraoperative diagnostics, tested in clinic settings, are proposed. These methods demonstrated high diagnostic accuracy during percutaneous puncture biopsy of liver neoplasms. Application of the developed classifiers enables detection of pathological changes with a sensitivity and specificity of 0.90 and 0.95, respectively. When a tumor tissue is detected, differentiation of neoplasm type is possible with a sensitivity and specificity reaching 0.80 and 0.95, respectively.Conclusion. Recent advances in optics have enabled the implementation of optical technologies in minimally invasive surgery, particularly the integration of fiber optic probes into standard puncture needles. The methods described in this paper facilitate preliminary conclusion about the tumor type with automated processing of optical spectroscopy data during puncture interventions. The application of these methods in clinical practice will increase the accuracy and reliability of puncture biopsy, which is essential in determining a personalized treatment strategy.