About 7-13% thyroid cancer patients have superior mediastinal lymph node metastasis, but its clinical identification is insufficient. Effective and individualized identification strategy need to be studied. Multiphoton imaging can accurately detect extracellular matrix collagen fibers by femtosecond laser. Quantitative features such as morphology and texture features extracted from the multiphoton images, can be used to predict lymph node metastasis of gastrointestinal cancer. Therefore, the investigators speculate that the characteristics of tumor microenvironment (TME) collagen based on multiphoton imaging may be used to predict superior mediastinal lymph node metastasis in thyroid cancer. The aim of this study is to investigate whether the TME collagen signature could predict superior mediastinal lymph nodes metastasis in thyroid cancer. The exposure of this observational cohort study is high TME collagen signature. The main measurement was the area under the receiver operating characteristic (AUROC) curves of collagen signature for predicting superior mediastinal lymph node metastasis.

Retrospective cohorts were enrolled from the Sir Run Run Shaw Hospital of Zhejiang University School of Medicine, The First Affiliated Hospital of Zhengzhou University and Nanfang Hospital from January 2020 to December 2023. Perspective cohort was obtained from all the three centers from January 2024 to July 2024.

Inclusion Criteria:

• Pathologically diagnosed as thyroid cancer
• Radical surgery of tumor and simultaneous resection of superior mediastinal lymph nodes

Exclusion Criteria:

• Reoperation
• Under the age of 18
• Combined with thoracopathy, lymphoma or other diseases that may cause mediastinal lymph node enlargement
• The superior mediastinal lymph nodes metastasis status was not separately reported
• Insufficient pathology section quality
• Pregnancy

Patients with thyroid cancer