Predicting specific outcome in dosimetry-based treatment for Graves' disease

Predicting Specific Outcome in Dosimetry-Based Treatment for Graves' Disease Background In ¹³¹I-NaI (radioiodine, RAI) therapy for the treatment of hyperthyroidism there is a lack of consensus between professional societies on the desired outcome of the therapy (euthyroidism or hypothyroidism) and on the optimal methodology to achieve the outcome (from empirical activities to personalised absorbed dose calculation). Analysis was performed on data from a prospective study of 284 patients with Graves' disease who were administered radioiodine calculated to deliver an absorbed dose of 60Gy to the thyroid. Multivariate logistic regression was carried out to determine prognostic covariates and produce a model to give probability of outcome. Outcomes investigated were cure of hyperthyroidism (including euthyroid and hypothyroidism), euthyroidism, and hypothyroidism. Outcome was assessed at 12 months post administration. Receiver operating characteristics (ROC) curves were produced and model performance was assessed using area under the curve (AUC). The AUC of the predictive model exceeded 70% for each outcome. Thyroid volume and percentage uptake in a ¹³¹I-NaI tracer study were identified as being predictive of cure of hyperthyroidism. Volume and effective half-life of the tracer were predictive of euthyroidism, and effective half-life and tracer uptake were predictive of hypothyroidism. This work allows us to determine which patient specific factors are significant predictors of response when dose is constrained. We produce models that allow for patient specific predictions of euthyroid and hypothyroid outcome following RAI therapy at 60 Gy. (author)