The present study explored the application of 3D-OmniView, a novel 3D-US technology, in evaluating the ETE of thyroid nodules. The results revealed that 3D-OmniView was more precise than 2D-US in predicting ETE of DTC nodules. The sensitivity, NPV, AUC and accuracy were improved by 3D-OmniView. ETE is easier to detect by ultrasound for nodules > 1 cm than for nodules ≤ 1 cm. 3D-OmniView could help more patients with nodules ≤ 1 cm.
The precise preoperative diagnosis of ETE of thyroid nodules is highly important, especially for papillary thyroid microcarcinoma(PTMC). Nodules with ETE are classified as high risk by ultrasound according to the 2015 American Thyroid Association guidelines[21]. The thyroid imaging reporting and data system of the American Radiological Society assigns a score of 3 points to ETE[22]. ETE also affects the choice of active surveillance, ablation or surgery for thyroid cancer. The American Thyroid Association guidelines recommend total or near-total thyroidectomy for DTCs with ETE and postsurgical radioactive iodine therapy for thyroid nodules of any size with gross ETE [21]. Moreover, numerous studies have identified gross ETE risk factors for recurrence, metastasis and survival. Although the effect of minimal ETE on PTMC clinical outcomes is controversial[23], some new studies have shown that minimal ETE is an independent predictor of persistent/recurrent disease and is associated with lymph node metastasis and lower disease-free survival rate[24, 25].
Ultrasound is the most popular imaging method for thyroid nodules. However, the accuracy of conventional 2D-US for the diagnosis of ETE is controversial[26]. First, the diagnostic criteria for ETE on ultrasound vary among different studies, from thyroid capsule contact, capsule contact > 25%, 25 ~ 50% or 50%, and disruption of the capsule to invasion of surrounding tissues. The sensitivity of 2D-US in predicting ETE decreases from 94.1–6.8%, and the specificity increases from 18.6–100%[3–12]. In this study, similar to the study of Ramundo (sensitivity, 43.2%; specificity, 81.9%)[7], we used a restrictive ultrasound definition of ETE (nodules abut the thyroid capsule with signs of disruption or disruption of the capsule and invade surrounding tissues). The sensitivity of 2D-US was 51.7%, and the specificity of 2D-US was 81.0%. We increased the sensitivity to 79.7% without decreasing the specificity of 3D-US. Second, 2D-US is a real-time dynamic imaging technique that is more dependent on the operator and more easily misses some information. Moreover, due to the lateral echo loss effect and acoustic shadow, exploration of the vascular and tracheal lateral boundaries of some nodules is limited. Therefore, we attempted to fill those gaps with 3D ultrasound.
In our study, 3D-OminiView showed higher sensitivity and accuracy for identifying ETE. 3D-US imaging provides volume data, thus reducing operator dependence and decreasing observer differences. 3D-US revealed a coronal plane that could not be visualized by 2D-US, reduced the influence of acoustic shadows and lateral wall echo drop-out, and showed a significantly clearer relationship between the thyroid nodules and the thyroid capsule along the tracheal and vessel sides. In addition, the focus of the observation could be localized to one point, which was simultaneously shown in three planes (transverse, longitudinal and coronal). When the thyroid nodules were adjacent to the thyroid capsule, 3D-US focused on these areas and provided dynamic observations in three planes at the same time. Moreover, 3D-OmniView can turn the disrupted line of the thyroid capsule into a plane of the thyroid capsule with protruding nodules, enabling clearer visualization of the relationship between the thyroid nodules and the capsule. As a result, 3D-OmniView identified ETE more readily and precisely than 2D-US.
Compared with other studies on the efficiency of 2D-US and 3D-US in diagnosing ETE, the present study maintained a relatively high specificity and improved the sensitivity (Table 3). This may be related to the use of 3D-OmniView technology and a restrictive US definition of ETE. In clinical practice, we found that 2D-US is sensitive enough to detect capsule contact of thyroid nodules, while the advantage of 3D-OmniView is reflected in the ability to distinguish capsule disruption and the relationship between nodules and the surrounding structure.
Table 3
Comparison of the efficiency of 2D-US and 3D-US in diagnosing ETE between different studies
Study | Criteria for predicting ETE | Sensitivity | Specificity |
2D-US | 3D-US | P | 2D-US | 3D-US | P |
Yi et al.[12] | contact capsule | 94.1% | 94.1% | 1.000 | 41.5% | 45.3% | 0.754 |
Gweon et al.[10] | contact capsule >25% | 78.2% | 86.5% | 0.14 | 27.2% | 27.2% | > 0.99 |
Kim et al.[11] | contact capsule >50% | 46.4% | 66.7% | 0.03 | 74.8% | 78.4% | 1.00 |
This study | capsule disruption or capsule disruption and invades surrounding tissues | 51.7% | 79.7% | <0.001 | 81.0% | 82.8% | 0.776 |
There are several limitations in this study. First, to focus on the ETE diagnosis, the nodules we enrolled were all adjacent to the thyroid capsule. This may have resulted in selection bias. Second, this was a single-center study with a small number of patients. Because only 12 DTC nodules presented gross ETE, we did not compare the diagnostic value of gross and minimal ETE by 2D-US and 3D-Omniview. A large, multicenter, prospective and comprehensive study is needed in the future to confirm these results.