OBJECTIVES

4-7% of adult US population have palpable thyroid nodule. Only 1 of 20 clinically identified nodules is malignant. This corresponds to approximately 2-4 per 100,000 people/year^a. Genomic classifiers; multigene assays designed to predict malignancy and avoid unnecessary surgeries, can return indeterminate posing possibility of under/over performing surgeries with the consequence of either lowering chances of early stage treatment, or incurring unnecessary resources that can impact patient quality of life. The study used claims data to assess impact of genomic classifiers on underdiagnosing thyroid malignancies versus unnecessary surgeries.

^aNEnglJMed.2004Oct21;351(17):1764-71.,Clinical practice.The thyroid nodule.

METHODS

Study analyzed 18+ patients diagnosed with thyroid benign neoplasms or nodular diseases as identified using respective ICD-9/10-CM codes [01Jan2013 through 31Dec2014] defining index event. Patients must have evidence of thyroid/Molecular panel within 3 months suggesting treating physician suspected malignancy. Surgery within 6 months post-diagnosis assigned patient to “Surgery arm” and lack thereof assigned patient to “No Surgery arm”. Patients must have no evidence of malignancy 12 months post-diagnosis, and if they had surgery, 12 months post-surgery, suggesting test returned negative/indeterminate. Patients were tracked 4 years to identify malignancy. Baseline and demographic characteristics were summarized using descriptive analysis. Hazard Ratio to develop malignancy was assessed using Cox Proportional Hazards (CoxPH) method.

RESULTS

2,982 patients qualified for study; 83% females. 89% had no surgery. 3% of patients with no surgery developed malignancy 30 months from diagnosis on average (median:28). 6% of those with surgery developed malignancy in 20.5 months on average (median:17). CoxPH Regression model showed HR of 0.53 (0.32-0.87) PValue: 0.0128.

CONCLUSIONS

Overall, not performing surgery showed to not increase chances of developing malignancy (HR 0.53 (0.32-0.87) PValue: 0.0128). In the study, patients with no surgery had lower chance of developing malignancy proving non-inferiority to surgery at 95% CI level. Probability of developing malignancy with/without surgery varied by age, gender and diagnosis.