By Will Boggs MD
NEW YORK (Reuters Health) - Mass-spectrometry imaging of fine-needle aspiration (FNA) biopsies can accurately identify the malignant potential of thyroid nodules, researchers report.
"We feel DESI-MS (desorption electrospray ionization mass spectrometry) will be an outstanding complementary test to the cytologist in the evaluation of thyroid nodules," said Dr. James Suliburk from Baylor College of Medicine, in Houston, Texas.
"As the technique becomes more refined, we may be able to obviate the need for cytologic review of FNA in some cases," he told Reuters Health by email. "One powerful feature of the technology is the ability to perform analysis on cells and in real time obtain results without need for extensive processing/preparation of the FNA smear."
Ultrasound-guided FNA biopsy is a standard-of-care technique for preoperative diagnosis of suspicious thyroid lesions, but discrimination between malignant and benign thyroid nodules from FNA cytology can be challenging.
Dr. Suliburk and colleagues used DESI-MS imaging to analyze 178 tissue samples to determine molecular signatures of normal, benign follicular adenoma (FTA), malignant follicular carcinoma (FTC) and papillary carcinoma (PTC). They tested the classifiers on 69 prospectively collected clinical FNA smears from 57 patients.
The first predictive model discriminated PTC from benign thyroid with 92% sensitivity, 94% specificity and 93% per-sample agreement with pathology, the researchers report in the Proceedings of the National Academy of Sciences, online October 7.
The second predictive model differentiated FTA from FTC with 80% accuracy, 81% sensitivity and 80% specificity on a per-patient basis, with only three of 16 FTC samples misclassified as benign in the validation set.
The FNA test set consisted of 69 samples from 57 patients, including 21 benign, 19 FTA, 28 PTC and one FTC smears from 34 benign and 24 malignant nodules. The benign-versus-PTC model yielded 96% sensitivity and 91% specificity on a per-nodule basis, misclassifying four of 57 FNA nodules.
The FTC-versus-benign-thyroid model correctly classified 30 of 34 benign samples, with overall sensitivity of 100%, specificity of 88% and overall accuracy of 89% for the FNA samples.
"We are excited/surprised with the quality of the results," Dr. Suliburk said. "In particular, of the clinical samples that were analyzed using current commercially available state of the art genomic testing, our mass spec correctly predicted benign behavior in 6 of the 6 specimens, whereas the commercially available assay predicted 'suspicious for malignancy' in 5 of the 6. In this limited clinical sample set, we could have prevented the need for surgery to exclude/diagnose malignancy in all of these patients."
"While sample size is very limited here, we are cautiously optimistic that we have developed a very valuable technique to evaluate for malignancy in indeterminate thyroid nodules," he said. "Anytime we can offer a faster, more precise diagnosis, the patient wins."
Dr. Michael S. Landau of the University of Pittsburgh Medical Center, in Pennsylvania, who has reviewed different thyroid-FNA practices for cytology and molecular testing, told Reuters Health by email, "This is an interesting method of analysis of thyroid nodules but far at this time from clinical use."
"It would be interesting to know if their method can distinguish benign from malignant Huerthle cell nodules and also benign follicular nodules from follicular carcinoma, non-Huerthle cell type," he said. "These are scenarios with which current methods still struggle, and there is a need for better biomarkers. However, classic type of papillary carcinoma, which comprises the overwhelming majority of their malignant cases, is usually easy to diagnose inexpensively by microscopic examination and/or analysis of nucleic acids."
Dr. Suliburk and several of his coauthors are inventors on a provisional patent application related to the use of mass spectrometry to diagnose thyroid cancer.
Proc Natl Acad Sci U S A 2019.