Abstract and Introduction
Abstract
Objective: Ultrasound (US) has gained a critical role in thyroid cancer treatment planning, yet it is limited by its user-dependent nature. The aim of this study was to compare the impact of US performed by radiologists specializing in thyroid imaging (hqUS) and US performed by radiographers in the community (cUS) on treatment plans of patients diagnosed with well-differentiated thyroid malignancies.
Design: Retrospective single-centre case series with chart review.
Patients: Patients diagnosed with thyroid cancer during 2017–2019 that had cUS followed by hqUS pre-operative counselling were included in this retrospective analysis.
Measurements: The main outcome was management alternations based on one of two sonographic measures: (1) extrathyroid extension (ETE); (2) The presence of central or lateral lymph nodes suspicious for metastases (LNM), which were compared with the final pathology.
Results: Among those with non-recurrent tumour (n = 76), ETE was reported 22 times more by hqUS compared with cUS (28.9% vs 1.3%, P < .001). Central and lateral LNM were reported approximately 6.5 and 1.5 times more by hqUS, respectively (25.0% vs 3.9%, P < .001 and 15.8% vs., 9.2%, P = .227, respectively). Overall, hqUS altered the initial treatment plan of 35.5% of patients. In 27.6% of patients, hqUS and its subsequent surgery resulted in a change to the patients’ 2015 ATA risk stratification system. In 40% of patients with microcarcinomas, hqUS findings mandated surgery according to findings that were not reported by cUS. False-positive rate was 5.2%.
Conclusions: Community US may under-diagnose important features such as ETE and LNM, leading to potential under-treatment in many patients. High-quality US of the neck should be considered in patients with differentiated thyroid carcinoma before making any treatment decisions.
Introduction
In recent years, ultrasound (US) has been accepted as the single most important imaging modality for initial risk assessment, decision-making and follow-up of thyroid nodules.[1] The initial US evaluation determines whether a nodule should be aspirated.[1,2] Other features may determine the future course of action in patients diagnosed with a suspected malignancy.
Following pivotal studies on more conservative approaches for the management of thyroid cancer,[3–5] the latest American Thyroid Association (ATA) guidelines have allowed for two major changes in initial decision-making for low-risk thyroid cancer:[1] recognizing hemithyroidectomy as adequate for tumours up to 4 cm without adverse features, and allowing active surveillance of microcarcinomas (tumours < 1cm). Yet, to allow for both options, aggressive features, namely, extrathyroid extension (ETE) or lymph nodes metastasis (LNM) should not be present. Therefore, the ability of the physician to offer the optimal treatment option completely depends upon the precision and reliability of neck US.
Despite its many advantages, US is not without limitations, mostly its user-dependent nature. Several studies have reported moderate-substantial inter-observer variability relating to assessments of various US features.[6–10] Among these, assessment of margins and capsular invasion of the cancerous process showed the greatest variability.[6,8,10] In addition, while the majority of common knowledge on US sensitivity and predictive values comes from high-quality neck US (hqUS) performed by experienced neck sonographers using high-quality transducers in referral centres, there are few reports in the English literature on the quality of baseline studies performed by radiographers in the community (cUS), although it is more commonly practiced than hqUS. Carneiro-Pla and Amin,[11] who evaluated the differences between cUS and surgeon-performed US, reported that cUS reports failed to mention lymph node status in nearly three-quarters of patients, a contralateral thyroid nodule in nearly a quarter of patients, and missed all cases of ETE. Another study, which evaluated the consistency of cUS reports regarding the US features of nodules, found that only 14% of reports related to 4 or more features that are needed for an initial evaluation, thus exposing major quality differences between cUS and hqUS.[12]
When considering the low sensitivity (30%) reported for central LNM—even by hqUS[13] or surgeon-performed US,[14] it is possible that determining an N0 stage based on cUS may result in under-treatment of patients as well as a missed indication for therapeutic neck dissection. Similarly, a missed ETE, which have been reported in up to 40% of pre-operative hqUS, may also lead to under-treatment.[15]
The study’s hypothesis was that performing an additional hqUS following a pre-operative cUS would have a significant impact on two aspects of decision-making: (1) whether to recommend surgery to patients who were eligible for active surveillance based on cUS reports and (2) planning the extent of surgery in all other patients diagnosed with thyroid malignancy (ie changing from lobectomy to total thyroidectomy or performing any type of neck dissection). For this purpose, we retrospectively examined the records of all patients who had both cUS and hqUS prior to undergoing surgery with a subsequent diagnosis of thyroid carcinoma. We believe that the results of the study are of importance to thyroid clinicians facing patients who had cUS and were referred to their clinic for treatment planning.