Prevalence of the pyramidal lobe
The PL is known as an ancillary lobe of the thyroid gland, correlating with the caudal end of the embryologic thyroglossal duct. Its prevalence in literature ranges from 35.7 to 60.0% in cadaver studies and from 12.3 to 61.0% when assessed during thyroid surgery [
1‐
10]. In the present study, we observed a prevalence of 80%, which is markedly higher than what was previously reported. We believe that the methods applied in this study contributed to this finding and raise the possibility that PLs might have been overlooked beforehand. A crucial weakness in the design of previous studies regarding the PL prevalence was the definition of negative controls. Whereas other investigators used CT, anatomical dissection, or intraoperative visual identification, we histopathologically analyzed complete tissue blocks of the anterior neck region.
In contrast to the previous cadaveric studies conducted by Braun et al. [
1], Ranade et al. [
2], Ozgur et al. [
3], Prakash et al. [
4] and Milojevic et al. [
5], we used fresh, non-embalmed cadavers. As fixation with formalin leads to decolorization and alters the tissue properties, the usage of fresh, non-embalmed cadavers enables a better differentiation of thyroid tissue from its surroundings. Additionally, the presence and location of the PL was indicated by ultrasound conducted prior to dissection. The microscopic examination of the tissue blocks in all cases in which the PL was not observed using ultrasound or on dissection further minimized the possibility of overlooking the PL. In the present study, one PL was not found during dissection and would have gone unnoticed without microscopic examination.
In comparison to Geraci et al., Zivic et al., Kim et al., Ryu et al. and Irawati et al., who reported the PL’s prevalence based on findings during thyroid surgery, we had the advantage of not being confined to a regular surgical area. To improve the exposure of the anterior cervical structures, the post-mortem nature of our study allowed us to perform larger skin incisions than normally required for hemi-/thyroidectomy in vivo. Furthermore, the patients included in these studies underwent thyroid surgery for various pathologies of the thyroid gland, potentially creating a selection bias [
6‐
10]. While we cannot ensure the exact representation of the general Austrian population with our study sample, we aimed to minimize any sort of bias by randomly selecting body donors at our institution.
All individuals examined in this study were born before the compulsory iodization of salt (10 mg KI/kg salt) by law in Austria in 1963, which used to be an area of iodine deficiency and endemic goiter. Due to persistent deficiency the initial supplement was increased to 20 mg KI/kg in 1990 [
28]. Considering the mean age of our study population at 81.3 ± 10.2 years, these body donors had already reached adult age by this time. As reported by Heinisch et al., goiter prevalence was expected to remain relatively unchanged in the population that had suffered from iodine deficiency prior to supplementation [
30]. However, neither thyroid volume, nor thyroid parenchyma echogenicity, nor age interfered with the accuracy of PL detection according to regression analysis.
Despite the smaller sample size (
n = 50) yielding lower statistical power in comparison to the aforementioned studies [
1,
2,
4‐
10] (except Ozgur et al. [
3], Table
1), the 95% confidence interval [68.9%; 91.1%] suggests a higher prevalence than formerly observed and strongly supports the opinion that the PL has to be considered a regular part of the thyroid gland [
1,
10].
Sonographic detection of the pyramidal lobe
It was previously shown that ultrasound offers good diagnostic value in the detection of the PL [
8,
9]. However, these prior studies were conducted with an in-hospital population presenting with thyroid or neck pathology (e.g., known malignancy, neck mass, lymphadenopathy), potentially generating selection bias or afflicting anatomy. Additionally, surgical findings or the prevalence of the PL on CT was used as reference [
8,
9,
23,
24]. To our knowledge, this is the first study investigating the sonographic detection of the PL in a post-mortem setting, confirmed by histopathological confirmation. With the methods applied, we were able to reliably exclude false-negative results.
Similarly to Kim et al. [
24], who compared the detection rates of the PL of multiple investigators using ultrasound and CT, we investigated the impact of the examiner’s experience level. Examiner I is a radiologist specialized in small parts ultrasound with more than 15 years of experience. In contrast, examiner II had acquired basic head and neck ultrasound skills during medical school. Diagnostic performance for examiner I was sensitivity of 85.0%, specificity of 50.0%, PPV of 87.2%, NPV of 45.5% and accuracy of 78.0%. At a very basic level of training, examiner II achieved sensitivity of 42.5%, specificity of 60.0%, PPV of 81.0%, NPV of 20.7% and accuracy of 46.0%, not yielding a better discrimination than by chance. These widely differing results highlight the importance of a proficient examiner. Expert assessment, optimal adjustment of settings and adept handling of the ultrasound probe enable the differentiation of delicate structures such as the PL.
Examiner I`s diagnostic performance was comparable to preceding studies in terms of sensitivity (72.6–85.2%), PPV (85.3–93.3%) and accuracy (80.3–87.5%), (Table
7) [
8,
9,
23,
24]. Regarding specificity (79.2–95.3%) and NPP (73.7–78.6%), however, we achieved considerably lower values at 50.0% and 45.5%, respectively [
8,
9,
23,
24]. We believe the superior resolution of this ultrasound system to be in part responsible for the relatively high false-positive rate of 50% (5/10). As noted by Ryu et al. and Kim et al., longitudinally arranged muscular fibers or fibro-fatty tissue can mimic the presence of a PL on ultrasound and cause false-positive results [
9,
24]. High-resolution imaging by modern ultrasound systems may enhance the discrimination of such longitudinally arranged structures and cause a false-positive result. Therefore, close attention should be paid to the echogenicity of the tissue in question.
Table 7
Comparison of diagnostic indices with previous studies
Prevalence (%) | 80.0 | 59.3 | 59.8 | 60.0 | 47.6 |
Sensitivity (%) | 85.0 | 81.0 | 82.3 | 85.2 | 72.6 |
Specificity (%) | 50.0 | 79.2 | 95.3 | 81.5 | 91.5 |
PPV (%) | 87.2 | 85.3 | 93.3 | 87.3 | 89.3 |
NPV (%) | 45.5 | 73.7 | 78.2 | 78.6 | 77.3 |
Accuracy (%) | 78.0 | 80.3 | 87.5 | 83.7 | 82.1 |
Another potential reason for the discrepancy in specificity is that Kim et al. and Kim et al. used the prevalence of the PL reported by CT as gold standard, but not surgical or anatomical confirmation [
23,
24]. Although CT was previously shown to have favorable diagnostic value (sensitivity 91.4% and specificity 94.4%) when compared to surgical prevalence as gold standard, an uncertain number of false-positive and false-negative cases remain [
9].
Despite the poor specificity observed in the present study, we advocate the screening and assessment of the PL in routine thyroid ultrasound as it can be detected with satisfactory sensitivity and provides additional information to the treating physician.
For clinical implication, a PL located along the former thyroglossal duct with separation from the main thyroid gland is of special interest. This was reported by Braun et al. in 1.7%, by Kim et al. in 3.8% and Ryu et al. in 5.2%, respectively [
1,
8,
9]. Accounting for 4% (2/50) in the present study, these remnants of thyroid tissue pose a high risk of not being detected during surgery as they lack communication to the thyroid isthmus. In the two cases presented in this study, the separated PL was located ventrally on the thyrohyoid membrane and would have most likely been overlooked by a surgeon screening the isthmus for a protruding PL. Both cases were correctly identified by examiner I using ultrasound. In the case of a separated PL, imaging prior to surgery could yield decisive information.