Clinical Validation of Metabolic Markers Detected by Mass Spectrometry Imaging for Diagnosis of Thyroid Fine Needle Aspiration Biopsies

Accurate diagnosis of thyroid nodules by fine needle aspirate (FNA) biopsy is essential to modern day best practice care in patients who are at risk of thyroid cancer. Current diagnosis of suspicious thyroid nodules relies on the interpretation of cytology findings by cytopathology. Unfortunately, difficulties in FNA diagnosis due to overlapping cytological features, inadequate sample size, or lack of clear pattern result in an indeterminate diagnosis in ~ 20% of cases. Clinical guidelines recommend that patients with indeterminate FNA undergo further testing including repeat biopsy (painful, may yield same indeterminate result), genomic analysis (expensive, not always available), or diagnostic thyroid surgery (very expensive, painful, invasive, with many life altering complications). Shockingly, 70-90% of patients that undergo diagnostic surgery are found to present benign nodules by surgical pathology, meaning that surgery was completely unnecessary. Unnecessary surgeries have major negative consequences. For patients, diagnostic surgery hypothyroidism results in decreased quality of life and lifelong need for hormone replacement therapy. For the healthcare system, the cost from unnecessary surgeries is enormous. Despite best efforts in genomic analysis and improved cytologic classification, there still remains a large diagnostic gap and need for improved technology for preoperative diagnosis of thyroid cancers. To address this critical clinical need, we have combined our expertise in thyroid cancer/surgery (Dr. James Suliburk, Department of Surgery, Baylor College of Medicine, BCM), mass spectrometry imaging (Dr. Livia S. Eberlin, Department of Chemistry, The University of Texas at Austin), statistical analysis (Dr. Rob Tibshirani, Department of Biomedical Data Science, Stanford University), clinical chemistry (Dr. Rongrong Huang, Scientific Director of Clinical Chemistry, BCM), and clinical pathology (Dr. Thomas Wheeler, Department of Pathology, BCM), and developed an assay using mass spectrometry imaging and machine learning to diagnose FNA biopsies based on the detection of a profile of hundreds of metabolic markers directly from clinical specimens. Now, we propose to conduct critical analytical and clinical validation studies with FNA biopsies prospectively collected from patients undergoing treatment at BCM to rigorously validate the method for clinical implementation. During the UH2 research phase, we will establish key analytical performance metrics, quality control measures, and method standardization procedures to evaluate the performance of our assay and metabolic markers within its clinical context of use. During the UH3 research phase, we will validate the clinical and diagnostic performance for FNA diagnosis in comparison to gold standard pathologic evaluation. Our premise is that the rigorous studies proposed will complete the analytical and clinical tasks needed to validate our assay and predictive markers for thyroid FNA diagnosis, thus demonstrating its effectiveness as a diagnostic assay. With support from commercial partners, our ultimate objective is to develop this innovative metabolic test into a robust technology for high-throughput and accurate diagnosis of thyroid FNA material.