Apparent Survival Benefit of Postoperative Radioactive Iodine in Lymph Node-Positive Papillary Thyroid Carcinoma: A Negative-Control Analysis for Unmeasured Confounding.

Abstract

BackgroundObservational studies report an overall-survival advantage for postoperative radioactive iodine (RAI) in lymph node-positive papillary thyroid carcinoma (PTC), but such estimates are vulnerable to confounding by indication and to bias from immortal time and the healthy-user effect. We assessed the credibility of this apparent benefit rather than RAI efficacy itself.MethodsUsing SEER (2010-2015), we assembled a cohort of N1b, M0, totally thyroidectomized PTC patients and applied a 3-month landmark analysis (sensitivity analyses at 1 and 6 months) with stabilized inverse-probability-of-treatment weighting. We estimated RAI effects on overall survival, cancer-specific survival, and relative survival, and on a negative-control outcome-non-thyroid-cancer death-that RAI cannot plausibly affect. Residual confounding was quantified with E-values.ResultsAmong 4275 patients, RAI showed an apparent survival benefit in direction across all three endpoints. However, the negative-control outcome likewise showed an apparent benefit (IPTW hazard ratio 0.713, 95% CI 0.562-0.906), indicating residual confounding not removed by weighting. The apparent survival benefits therefore could not be distinguished from unmeasured confounding.ConclusionsIn registry data lacking key confounders, a negative control identifies the apparent RAI survival benefit as a candidate artifact of unmeasured confounding. The findings neither support nor refute RAI efficacy.