Intelligent SERS navigation system to guide lung cancer surgery through intraoperative metabolic acidosis.

e20061 Background: Lobectomy combined with systematic lymph node dissection is the standard treatment for early stage lung cancer. However, partial resection, such as sublobar resection, is gaining attention because it can protect normal lung tissue to the greatest extent. Several studies have shown that partial resection has a good effect in improving prognosis. However, excessive lymph node dissection may negatively affect the immune system, thereby impairing the efficacy of postoperative adjuvant immunotherapy. Therefore, more precise surgical strategies are needed to avoid unnecessary lymph node resection. The acidic microenvironment of tumor is one of its malignant characteristics, and its application in surgical navigation has shown great potential in recent years. Methods: Based on the acidic tumor microenvironment, we established an innovative pH-based surface-enhanced Raman scattering (SERS) chip to accurately identify the acidic margin and metastatic lymph nodes of lung cancer. A total of 56 lung cancer patients were enrolled in this study. The pH characteristics of tumor, normal tissues and benign and malignant lymph nodes were measured by SERS chip, and the tumor resection margin and related lymph nodes were rapidly located. In order to further improve the diagnostic efficiency, a multi-classification model of pH and a classification model of benign and malignant pH were developed by combining deep learning technology. The data set was divided into training set, validation set and test set according to the ratio of 7:2:1 for model training and validation. Results: Through SERS chip, we successfully achieved accurate recognition of tumor resection margin and metastatic lymph nodes. The accuracy of the deep learning model on the training set, validation set and test set were 0.92, 0.972 and 0.860, respectively, indicating that the system could efficiently and accurately identify tumor resection margin and malignant lymph nodes. Conclusions: The pH-responsive SERS navigation system proposed in this study provides a novel technical scheme with clinical application potential for precise surgical treatment of invasive solid tumors. This system can improve the accuracy and efficiency of surgery, and is expected to accelerate its clinical translation and application in tumor surgery. Future studies will further optimize the performance of this technique and promote its wide application in various types of solid tumor surgery through large-scale clinical validation.