Diagnostic performance of the second-generation molecular tests in the assessment of indeterminate thyroid nodules: A systematic review and meta-analysis

• Published in: American journal of otolaryngology 2022-05, Vol.43 (3), p.103394-103394, Article 103394
• Main Authors: Lee, Esther, Terhaar, Samantha, McDaniel, Lea, Gorelik, Daniel, Gerhard, Eleanor, Chen, Chen, Ma, Yan, Joshi, Arjun S., Goodman, Joseph F., Thakkar, Punam G.
• Format: Article
• Language: English
• Subjects: Bias; Biopsy; Bivariate analysis; Cellular biology; Confidence intervals; Cytology; Diagnostic systems; GSC; Malignancy; Mathematical models; Medical diagnosis; Meta-analysis; Mutation; Nodules; Performance evaluation; Search engines; Sensitivity; Statistical analysis; Surgery; Systematic review; ThyGeNext; Thyroid; Thyroid cancer; Thyroid nodule; Thyroseq v3; Tumors
• ISSN: 0196-0709; 1532-818X
• DOI: 10.1016/j.amjoto.2022.103394

The objective of this systematic review and meta-analysis was to evaluate the diagnostic performance of the second-generation molecular tests in the diagnosis of thyroid nodules with indeterminate fine-needle aspiration biopsy results. We searched PubMed, Google Scholar, Scopus, and Cochrane Library for studies published between January 2017 and March 2021. Inclusion criteria were indeterminate thyroid results from fine-needle aspiration (FNA) that included Bethesda categories III and IV, use of Afirma GSC, Thyroseq v3, and ThyGeNext as an index test, and conclusive histopathological results. Studies with no post-surgical diagnoses were excluded. For each included study, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were obtained. Sensitivity and specificity were pooled jointly using a bivariate binomial random-effects model. Statistical significance was indicated at p-value less than 0.05. Our search yielded 431 non-duplicate articles, of which 15 were included in the study (7 GSC, 6 Thyroseq v3, and 2 ThyGeNext). ThyGeNext studies were excluded from the meta-analysis due to the small sample size. Pooled data for GSC studies on 472 thyroid nodules showed a sensitivity of 96.6 (95% confidence interval: 89.7–98.9%), specificity of 52.9% (23.4–80.5%), PPV of 63% (51–74%), and NPV of 96% (94–98%). Pooled data for ThyroSeq studies on 530 thyroid nodules showed a sensitivity of 95.1% (91.1–97.4%), specificity of 49.6% (29.3–70.1%), PPV of 70% (55–83%), and NPV of 92% (86–97%). There was no statistically significant difference in diagnostic performances of the two tests (p-values for sensitivity = 0.89, specificity = 0.82, PPV = 0.43, NPV = 0.17). High sensitivity and high NPV in GSC and Thyroseq v3 have potential to help rule out malignancy among thyroid nodules with indeterminate cytology results. There was no difference in diagnostic performances between the two molecular tests indicating that either test is appropriate to determine the malignancy of thyroid nodules. Further long-term outcome data are warranted to make a clear recommendation.