• Bethesda IV
• Indeterminate
• solitary thyroid nodule
• thyroid cancer
• total thyroidectomy

Background. Despite the many advances in the field of nodular thyroid disease and cytopathology reporting, fine needle aspiration cytology fails to establish a reliable diagnosis in 20-30% of cases. These nodules are labeled as cytologically indeterminate for malignancy. The term indeterminate collectively applies to nodules reported as Bethesda categories III, IV, and V. Such labeling is an eyesore to health care providers as it places their patients at risk of sub-optimal therapeutic decisions. This particularly applies to Bethesda category IV whose implied risk of malignancy falls in the gray zone between those of Bethesda categories III and V. Therefore, it can neither be considered low enough for a thyroid lobectomy to become custom, nor high enough to warrant a total thyroidectomy. An additional concern is related to benignity being the most common outcome in this subset of patients. This implies that these patients are at risk of being subjected to an unnecessary and potentially morbid thyroid lobectomy. This last concern is particularly prominent in centers around the world that customarily select patients with Bethesda IV nodules for a thyroid lobectomy.
At a national level, in centers around Italy, things are quite different. Most patients with Bethesda IV nodules (equivalent to TIR 3b according to the Società Italiana di Anatomia Patologica e Citopatologia Diagnostica/ International Academy of Pathology, Italian Division/ SIAPEC-IAP) are selected for a total thyroidectomy. A procedure that could be correctly labeled as “overly radical” for this subset of patients.
This study aims to tackle this nation-wide issue: the near routine selection of patients with TIR 3b/ Bethesda IV nodules for a total thyroidectomy. This is best done by developing, validating and popularizing a tool that could accurately prevent unbeneficial total thyroidectomies and at the same time be readily available, easy to assemble, and cost-effective.
This rule-out tool was first developed and published by the author and his colleagues in a previous retrospective study. The tool was assembled by combining negativity for suspicious gray-scale US features, and negativity for the genetic mutations commonly encountered in differentiated thyroid cancer (namely BRAF and NRAS). The gray-scale US features selected included: irregular margins, a taller-than-wide orientation, and the presence of microcalcifications, and was based on a recently published meta-analysis in the literature. The rule-out tool, abbreviated by the authors as: (US-/ Mutation-), demonstrated high predictivity for lesions that do not require a total thyroidectomy. Its predictivity of lesions for which a thyroid lobectomy is considered sufficient therapy was 94%.
The current study intends on prospectively validating the efficacy of the rule-out tool in the preoperative setting in triaging patients with Bethesda IV nodules to a lobectomy instead of a total thyroidectomy. Furthermore, demonstrate the synergism between the two components used to assemble it.
Methods. Between Jan. 2016 and Jan. 2018, 200 consecutive patients presented to an academic tertiary referral center with solitary thyroid nodules lacking all suspicious sonographic features set by the authors, and labeled as suspicious for a follicular neoplasm (Bethesda category IV) following FNAC. According to the authors’ published experience three grey-scale ultrasound (US) features in single or in combination are of sufficient clinical significance to label a thyroid nodule as suspicious for malignancy. These include: irregular margins, the presence of microcalcification, and a taller-than-wide configuration. The clinical significance of these US features has also been confirmed in a recent meta-analysis in the literature. Total thyroidectomy was justifiable in 33 out of the 200 patients for one the following reasons: hypothyroidism/ a background of Hashimoto’s thyroiditis (n=30), a positive family history for thyroid cancer (n=2), or a history of radiation exposure (n=1). These cases were excluded from the study, and the remaining 167 cytology smears were analyzed for NRAS and BRAF. Only 10 were positive for a mutational marker: BRAF V600E (n=1), BRAF K601E (n=1), and NRAS (n=8). Out of these mutation-positive lesions 8 were malignant, and according to the American Thyroid Association (ATA) risk stratification, six of these were high-risk and required a total thyroidectomy as a minimum surgical treatment. Ultimately 157 patients defined the authors’ study cohort and were all subjected to a thyroid lobectomy.
The rule-out tool put to test in this study consisted of two components: 1. Negativity for suspicious US features and 2. Negativity for mutational markers. It was abbreviated by the authors as: (US-/ Mutation-). Its diagnostic accuracy was assessed by calculating its negative predictive value (NPV) for both malignancy and malignancy requiring a total thyroidectomy. In other words, it was evaluated for its ability to preclude total thyroidectomy as the therapeutic modality required. Whether or not mutational marker negativity imparted an additional clinical benefit (i.e. a synergistic impact) as part of this “rule-out tool” was also evaluated. This was done by calculating the NPV of US negativity but mutational positivity (US-/mutation+) for both malignancy and malignancy requiring a total thyroidectomy and comparing it to those of (US-/mutation-).
Results. The 157-patient study cohort included 27 males and 130 females, with a male to female ratio of 1:5. The average age was 44 years (range: 14-75 years), and the mean nodule size was 34 mm (20-66 mm). Following
lobectomy, permanent pathology revealed 140 benign lesions and 17 malignant ones. Malignancies included: the FVPTC (n=12), classical PTC (n=2), FTC (n= 2), and tall-cell PTC (n=1). Out of the 17 malignant lesions, 8 (47%) demonstrated one or more ATA high-risk features that warranted a completion thyroidectomy.
From the results obtained, the NPV of (US-/mutation-) for malignancy was 89% (140/157). This increased to 95% (149/157) for malignancy requiring a total thyroidectomy. This implies a 95% diagnostic accuracy in refuting a total thyroidectomy in this subset of patients.
The synergistic impact that mutational marker negativity imparted as an essential component of the tool was assessed by calculating the NPV of US negativity but mutational positivity (US-/mutation+) for both malignancy and malignancy requiring a total thyroidectomy and comparing it to those of (US-/mutation-). The NPV of (US-/mutation+) for malignancy was 20% (2/10), and 40% (4/10) for malignancy requiring total thyroidectomy. The differences were statistically significant: [NPV for malignancy: 89% vs. 20%; p < 0.0001, and NPV for malignancy requiring total thyroidectomy: 95% s. 40%; p < 0.0001].
Conclusion. The combination of (US-/mutation-) is a valid and reliable rule-out tool with sufficient pre-operative diagnostic accuracy to spare patients with Bethesda IV nodules an overly radical total thyroidectomy.