Importance:
Thyroid nodules classified as atypia of undetermined significance have a significant risk of malignancy and can be managed by surveillance or diagnostic surgery. There are currently no publications evaluating surveillance compliance.

Objective:
To determine whether thyroid nodule surveillance compliance is influenced by patient demographics or plan type, and whether patient or nodule characteristics influence triage to surgery.

Design: Retrospective case series from 2010-2018.

Setting: United States Military Health System.

Participants: 481 patients with a thyroid nodule fine-needle aspiration classified as atypia of undetermined significance for whom treatment and follow-up information was available.

Main Outcomes and Measures: Compliance with nodule surveillance and the influence of patient demographics, sonographic nodule characteristics and surveillance plan on compliance or triage to surgery.    

Results: A total of 289 nodules were surveilled and 192 diagnostic lobectomies were performed. An initial surveillance plan was documented in 93% (268/289) and 86% (231/268) complied. The most common plans were repeat biopsy in 78% (210/268) or ultrasound in 20% (53/268). A second plan was documented in 88% (204/231) of those who complied. The most common plans were ultrasound in 87% (178/204) or repeat biopsy in 8% (17/204). Compliance with the second plan was 64% (130/204), significantly lower than with the first (OR 3.6, 95% CI: [2.3, 5.6], P < 0.0001). Only 45% (130/289) were surveilled twice. Age and gender did not significantly affect compliance rates. Compliance with primary care ultrasound surveillance was 40% (21/52), significantly lower than with a specialist (77% [137/179]; OR 4.8, 95% CI: [2.5, 9.3, P < 0.0001). In the surgical cohort, age was lower (47 vs 52, P = 0.0002), nodule size was greater (2.6 vs 2.0 cm, P < 0.0001), and internal vascularity was more common (80% vs 66%, P = 0.003). The risk of malignancy after diagnostic lobectomy was 38% (73/192).

Conclusions and Relevance: Compliance with surveillance of thyroid nodules classified as atypia of undetermined significance was poor in this military cohort. Ultrasound surveillance by a specialist may be more reliable than with primary care. Age, nodule size, and internal vascularity may have influenced providers towards recommending diagnostic lobectomy.