Can a simple rule-out tool reliably select patients with solitary Bethesda IV nodules for conservative surgery?

Abstract

The study investigated the validity and reliability of an innovative diagnostic rule-out tool in selecting patients with solitary Bethesda IV nodules (suspicious for a follicular neoplasm) for thyroid lobectomy instead of a total thyroidectomy [1]. The authors described an innovative rule-out tool that was assembled by combining the negativity for three suspicious ultrasound features (irregular margins, microcalcification, and a taller-than-wide orientation), and mutational marker negativity (BRAF and NRAS). The tool was abbreviated as (US−/mutation−). The sonographic component of the tool was based on the meta-analysis by Remonti et al., which evaluated the diagnostic performance of US features for malignancy, based on studies that included nodules with a postoperative histopathological diagnosis [2]. Although they concluded that no US feature in isolation could predict the risk of malignancy with acceptable diagnostic accuracy, they demonstrated that irregular margins, a taller-than-wide orientation, and the presence of microcalcifications carried the highest risks of malignancy. The molecular component included the most common genetic alterations in differentiated thyroid cancers that are likely to be the outcome in a lesion suspicious for a follicular thyroid neoplasm [3, 4]. These included: RAS (mostly NRAS), and BRAF K601E (rarely V600E) mutations. Although a molecular component with a broader set of genetic mutations and/or alterations would be of additional value, they kept the tool basic, to fulfill the criteria of simplicity, availability, and cost-effectiveness. The tool was prospectively tested on a study cohort of 157 patients that fit the inclusion criteria set by the authors. It demonstrated an 89% negative predictive (NPV) value for malignancy and a 95% NPV for malignancy requiring total thyroidectomy according to the most recent clinical practice guidelines [5]. This means that the tool demonstrated a 95% preoperative diagnostic accuracy in refuting total thyroidectomy in patients with solitary Bethesda IV nodules. The two components of the tool (sonographic and molecular) demonstrated synergism as the NPV of the tool for malignancy and malignancy requiring total thyroidectomy decrease significantly when mutational markers were positive. Comparing (US−/mutation−) to (US−/mutation+): NPV (malignancy): 89% vs. 20%; p < 0.0001, and NPV (malignancy requiring TT): 95% vs. 40%; p < 0.0001. The strength of this tool is represented by its simplicity, the quick availability, high benefits in terms of cost-effectiveness. With the high sensitivity and high NPV, it holds a high capacity to affect patient’s management decisions. My belief is that this tool should be widely implemented in current practice. To confirm its potential and assess its real impact on pre-surgical diagnosis a multicentric randomized study is necessary. This will help to avoid selection biases and/or operator-dependent results (pathologist experience also does count). The actual limitation of this tool, as mentioned by the author, is that in its current form, it cannot reliably select patients for watchful waiting. It could be useful to also add the RET/PTC re-arrangement evaluation, to increase PPV. I encourage the author to check its feasibility.