Related Experiment Video
Updated: Jun 30, 2026
07:42
Assessment of Resistance to Tyrosine Kinase Inhibitors by an Interrogation of Signal Transduction Pathways by Antibody Arrays
Published on: September 19, 2018
Assessment of Papillary Thyroid Carcinoma by Profiling Multiple Matrix Metalloproteinase Activities Using a Machine
Zhenshengnan Li1, Xudong Sun2,3, Bowen Chen1
1Department of Thyroid Surgery, General Surgery Center, The First Hospital of Jilin University, Changchun, Jilin 130021, China.
Analytical Chemistry
|June 29, 2026
Summary
A new machine learning-assisted peptide microarray platform accurately diagnoses papillary thyroid carcinoma (PTC) using matrix metalloproteinase (MMP) activity. This tool aids in distinguishing benign from malignant thyroid nodules and predicting metastasis.
Area of Science:
- Biomarker Discovery
- Medical Diagnostics
- Machine Learning Applications
Background:
- Diagnosing papillary thyroid carcinoma (PTC) with ambiguous ultrasound features remains challenging.
- Matrix metalloproteinases (MMPs) are crucial in tumor invasion and metastasis, showing promise as PTC biomarkers.
Purpose of the Study:
- To develop a machine learning-assisted peptide microarray sensing platform (MLPM) for accurate PTC diagnosis.
- To combine high-throughput peptide microarray analysis with gradient boosting (GB) algorithm for enhanced predictive performance.
Main Methods:
- Developed a peptide microarray for simultaneous detection of seven MMP activities with high sensitivity and wide dynamic range.
- Utilized a gradient boosting (GB) algorithm to analyze MMP activity profiles for diagnostic prediction.
- Validated the MLPM platform on diverse sample types including cells, tissues, and plasma.
Main Results:
- The peptide microarray achieved low limits of detection (pg mL-1 level) and wide dynamic ranges (>3 orders of magnitude).
- The GB model achieved an AUC of 0.99 for discriminating benign from malignant nodules and 0.97 for predicting lymph node metastasis.
- Independent blind validation demonstrated high accuracy of 96.0% and 94.1% for these diagnostic tasks.
Conclusions:
- The MLPM platform offers a robust and generalizable approach for accurate PTC diagnosis.
- This integrated system effectively utilizes MMP activity profiles for critical clinical decision-making in thyroid nodule assessment.
- The study highlights the potential of combining advanced sensing technologies with machine learning for improved cancer diagnostics.
