Evaluating the Predictive Capability of Transcriptomic Profiling for Identifying the Primary Site of Metastatic Tumors

Overview

This study will enroll patients with metastatic malignancies. Tumor samples (fresh or formalin-fixed paraffin-embedded tissue specimens) will undergo RNA extraction and next-generation sequencing (RNA-seq). Once the raw data is obtained, the system will analyze the transcriptomic feature values (cancer-specific RNA transcripts and tissue-specific RNA transcripts) expressed in the tumor tissue samples to further predict tissue origin using a machine learning model. The output includes probabilities and confidence intervals for tissue origin.

This is a non-interventional, observational study. Through a single-center, prospective clinical trial, the study aims to utilize the transcriptomic profiling for tumor tissue origin identification to predict the tissue origin of primary sites in metastatic tumors and evaluate the accuracy and specificity of this prediction solution. Primary Endpoint: The accuracy of the transcriptomic profiling for tumor tissue origin identification in predicting the primary site of metastatic tumors (expressed as overall accuracy with its 95% confidence interval). Secondary Endpoints: 1. The specificity and sensitivity of the transcriptomic profiling for tumor tissue origin identification in predicting the primary site of metastatic tumors. 2. Exploratory analysis of characteristic molecular markers expressed in metastatic lesions from different primary sites.

Conditions

Eligibility