Preoperative Identification of Papillary Thyroid Carcinoma Subtypes and Lymph Node Metastasis via Deep...

Area of Science:

• Biomedical Engineering
• Oncology
• Spectroscopy

Background:

• Accurate preoperative diagnosis of papillary thyroid carcinoma (PTC) histological subtypes and lymph node metastasis is crucial for personalized treatment.
• Current methods like cytology and ultrasound have limitations in reliability and accuracy for these diagnoses.

Purpose of the Study:

• To develop a deep learning-assisted surface-enhanced Raman scattering (SERS) chip for preoperative diagnosis of PTC histological subtypes and lymph node metastasis.
• To improve diagnostic accuracy and aid in personalized treatment strategies for PTC.

Main Methods:

• Utilized a deep learning-assisted SERS chip with fine-needle aspiration (FNA) samples.
• Employed a convolutional neural network algorithm to analyze Raman spectral fingerprints.
• Evaluated the platform's accuracy in distinguishing PTC subtypes and detecting lymph node metastasis.

Main Results:

• The SERS chip successfully distinguished PTC subtypes and lymph node metastasis with an accuracy of 95.83%.
• The platform achieved 100% accuracy in identifying central cervical lymph node metastasis.
• Demonstrated high performance in preoperative diagnosis using FNA samples.

Conclusions:

• The deep learning-assisted SERS platform offers a promising approach for accurate preoperative diagnosis of PTC.
• This technology supports personalized medicine by enabling individualized treatment strategies.
• Potential to reduce overtreatment and mitigate recurrence risk in PTC patients.