Diagnostic Yield of Thyroid Incidentaloma Evaluation and the Impact of Guideline Variability: Insights from a Retrospective Cohort Study

Abstract

Introduction:

Thyroid incidentalomas are lesions detected incidentally on imaging performed for unrelated indications. Determining which lesions warrant further evaluation to exclude malignancy introduces a dilemma: intensive diagnostic strategies maximize malignancy detection but increase diagnostic workload, patient burden, and overdiagnosis of indolent cancers, whereas restrictive approaches reduce these harms but risk missing malignancies. Guidelines balance these opposing considerations differently, resulting in substantial variation in recommendations for evaluating conventional imaging-detected thyroid incidentalomas (ultrasound, computed tomography, magnetic resonance imaging). In this study, we aimed to investigate the diagnostic yield of thyroid incidentaloma evaluation and the impact of guideline variability on malignancy detection and diagnostic workload.

Methods:

We retrospectively assessed all patients referred to Zuyd Thyroid Center (2018–2023) for thyroid nodule evaluation. Data on diagnostic outcomes, treatments, and complications were collected. A guideline-based simulation analysis examined how applying the European Thyroid Association (ETA; most intensive), the American Thyroid Association (ATA), and the Dutch Federation of Medical Specialists (most restrictive) guidelines would affect malignancy detection and diagnostic workload.

Results:

Of 1825 referred patients, 630 (34.5%) had one or more incidentalomas (median age 66 years [interquartile range 54–74]; 70.6% female). Malignancy rate was significantly higher for nuclear imaging-detected incidentalomas than for conventional imaging (21.1% vs. 1.8%, p < 0.001). In the guideline-based simulation analysis, ETA recommended evaluating all cases, while ATA (without clinical warning signs) evaluated 85.3%, both identifying all malignancies. The Dutch guideline avoided evaluation in 96.9% of patients, missing 7 malignancies among 528 unevaluated cases (1.3%), all subtypes with good prognosis even if left undetected.

Conclusions:

Conventional imaging-detected incidentalomas carry a low malignancy risk, resulting in a high diagnostic workload to detect few malignancies. More restrictive evaluation criteria can substantially reduce diagnostic burden but may miss a small number of low-risk malignancies. These findings highlight the need for improved selection criteria that minimize unnecessary procedures while ensuring detection of clinically relevant malignancies.