Mortality Benefit of Ultrasound for Incidental Thyroid Nodules Identified With PET Imaging: A Non-Inferiority Emulated Target Trial

The investigators hypothesize that all-cause mortality in patients with an incidental thyroid nodule on PET-CT who did not have thyroid ultrasound (the exposure) within 3 months of the PET-CT is non-inferior within a 5% margin to those who have thyroid ultrasound at 7-years. That is, among patients with an incidental thyroid nodule on PET-CT, mortality is no more than 5% larger (in absolute difference) for those who do not have thyroid ultrasound compared to those who do. The investigators will also report mortality differences at landmark timeframes of 1-year, 3-years, 5-years, and 10-years. To estimate group differences in mortality, the investigators will conduct a non-inferiority emulated target trial utilizing clone-censor weighting to address potential immortal time bias introduced by the 3-month grace period. The investigators will adjust for demographic, potential confounder, and mortality risk adjustor factors. The investigators will stratify analyses based on baseline disease severity (estimated 5-year relative survival risk) and disease status (progression, lymph node involvement, other sites of metastases). All subjects will be accrued from the Mass General Brigham healthcare system, which includes two academic medical centers, a specialty head and neck hospital, and multiple community hospitals and numerous community clinics.