Development and Validation of a Deep Learning Model for Diagnosing Lymph Node Metastasis in Nasopharyngeal Carcinoma Using Histologic Whole Slide Images and Time-dependent Magnetic Resonance Images

What this Trial Is About

Researchers are developing and validating an AI model to help diagnose lymph node metastasis in patients with nasopharyngeal carcinoma (NPC). The AI uses MRI images and pathology data to analyze single lymph nodes and surrounding areas, aiming to improve diagnosis before and after chemotherapy. This study also evaluates the AI's potential to correct past diagnoses and guide radiotherapy treatment plans, as well as its economic benefits in managing NPC. The AI model includes automatic identification and segmentation of lymph nodes and primary tumors using semi-supervised learning. It integrates MRI scans from different time points and pathological features from tumor samples to predict metastasis likelihood. The study involves retrospective analysis, validation with biopsy results in head and neck cancer patients, and prospective clinical trials to assess safety and effectiveness in guiding radiation doses. Participants will undergo MRI and PET/CT scans at diagnosis, with some having lymph node biopsies if imaging results differ. The study measures outcomes like the accuracy of diagnosis using area under the curve (AUC), sensitivity, and specificity over an average of two years. Researchers will monitor how well the AI model detects metastasis, its impact on treatment decisions, and its economic value, with participant involvement lasting throughout the study period.

A Deep Learning Model for Diagnosing Lymph Node Metastasis in Nasopharyngeal Carcinoma(NPC)