Accessing the influence of ^99mTc-Sesta-MIBI-positive thyroid nodules on preoperative localisation studies in patients with primary hyperparathyroidism

All patients with the biochemical diagnosis of PHPT who underwent surgical parathyroidectomy with/without thyroid surgery between 2009 and 2015 were included in this analysis.

Inclusion criteria were sporadic PHPT with available data regarding clinical symptomatology, preoperative US and MIBI scan, method of PTX, method of thyroid surgery, final histology and follow-up data.

All patients preoperatively underwent a standard double-phase MIBI scintigraphy with SPECT and high-resolution US with colour Doppler of the cervical region to localise one or more enlarged (presumed hyperfunctioning) parathyroid gland(s). US was also used to evaluate the morphology of the thyroid gland, in accordance to previously published standard protocols [22, 23]. All MIBI and US images and reports were retrospectively reviewed.

True positive (TP): Removal of parathyroid gland with adequate decline of PTH and postoperative normocalcemia = correct prediction of complete resection.

False positive (FP): Removal of parathyroid gland with adequate decline of PTH and persistent hypercalcemia = incorrect prediction of complete resection.

True negative (TN): Removal of parathyroid gland with inadequate decline of PTH, thus requiring additional gland resection or operative failure and persistent hypercalcemia = correct prediction of incomplete excision.

False negative (FN): Removal of parathyroid gland with inadequate decline of PTH but postoperative normocalcemia = incorrect prediction of incomplete excision.

The results of preoperative localisation determined surgical strategy, in accordance with the ESES consensus statement [22, 24]. Minimal invasive parathyroidectomy (MIP) (skin incision = 20–25 mm) was performed on the side of the predicted enlarged/hyperfunctioning single gland disease by MIBI, irrespective of US.

Categorical variables are reported as absolute numbers and percentages, while continuous variables are given as mean with standard deviation. Sensitivity and positive predictive value (PPV) were calculated using 2 × 2 tables. Descriptive statistics analysed patient characteristics. Chi-quadrate tests and Fischer’s exact test were used to determine statistical significance. Analysis of data was performed using SPSS version 23 for Windows (Chicago, IL, USA).

Four hundred ninety-seven patients (female: n = 341 (69.8%); male: n = 156 (31.2%)) were included in this analysis. The average age was 60.2 (± 13.1) years.

In total, 153/497 (30.8%) patients were asymptomatic. Conversely, 228/497 (45.9%) patients showed symptoms of arterial hypertension, hypercalcemia, osteopenia, bone pain and depression. Typical symptoms, such as bone manifestations and renal manifestations, were observed in 69.2% and 20.3%, respectively. Depression was seen in 45.9% of patients, 0.6% presented with hypercalcaemic crisis and 0.2% had a brown tumour.

In total, 318/497 (64.0%) patients showed ≥ 1 thyroid nodule in the preoperative US. Overall, 198/318 (62.3%) of these patients had concomitant thyroid surgery: 103/198 (52.0%) thyroidectomies and 95/198 (48.0%) hemi-thyroidectomies.

True positive US results were seen in 132/299 (44.1%) patients and 65/198 (32.8%) patients that underwent sole PTX and PTX with concomitant thyroid surgery, respectively. False positive results were observed in 113/299 (37.8%) and 84/198 (42.4%) patients that underwent sole PTX and PTX plus thyroid surgery, respectively. False negative results were seen in 54/299 (18.1%) patients with sole PTX and 49/198 (24.7%) patients with PTX and thyroid surgery. Statistical analysis showed a significant difference in the ability of preoperative US to more often correctly identify parathyroid adenomas in patients without thyroid nodules (p = 0.029; see Table 1).

In patients who underwent sole PTX, 170/299 (56.9%) patients had true positive MIBI results whereas 101/198 (51.0%) patients with concomitant thyroid surgery had true positive results. False positive results were seen in 68/299 (22.7%) and 58/198 (29.3%) patients that underwent sole PTX and PTX plus thyroid surgery, respectively. False negative results were observed in 61/299 (20.4%) patients with sole PTX and 39/198 (19.7%) patients with PTX and thyroid surgery. There was no significant statistical difference in the MIBI localisation of parathyroid adenoma in patients with concomitant thyroid nodules (p = 0.248; see Table 2).

Furthermore, in patients with concomitant thyroid nodules, MIBI result was true positive in 30/101 (29.7%) patients with thyroid carcinoma (27 patients had papillary thyroid microcarcinoma ranging from 1 to 10 mm) and in 71/101 (70.3%) patients with benign thyroid nodules. In terms of false positive results, 13/58 (22.4%) patients had thyroid carcinoma (eight patients had papillary microcarcinoma and one patient medullary microcarcinoma, all ranging between 0.3 and 10 mm) and 45/58 (77.6%) benign nodules. Lastly, 14/39 (35.9%) patients had thyroid carcinoma (11 patients had papillary microcarcinoma ranging from 0.4 to 2.5 mm), but false negative MIBI result, and 25/39 (64.1%) patients had false negative MIBI results, but benign thyroid nodule in the final histology. This difference was not statistically significant (p = 0.341; see Table 3).

The most common surgical procedure conducted was MIP in 210/497 (42.3%) patients followed by BNE in 177/497 (35.6%) and UNE in 110/497 (22.1%) patients. Patients underwent concurrent thyroid surgery when suspicious unilateral or bilateral multinodular goitre(s) or solitary nodule(s) ≥ 5 mm demonstrated US characteristics typical for malignancy (described above in “Materials and methods”). This allowed for the exclusion and sufficient treatment of thyroid carcinoma. In total, 299/497 (60.2%) patients underwent sole PTX. The remaining 198/497 (39.8%) patients had PTX with concomitant thyroid surgery. Of these 198 patients, 95/198 (48.0%) underwent a hemithyroidectomy and 103/198 (52.0%) total thyroidectomy. In 37/198 (18.7%) patients, a conversion from UNE to BNE was required based on inadequate IOPTH decline and false preoperative localisation, whereas a conversion in patients with sole PTX was only observed in 35/299 patients (11.7%; see Table 4).

The majority of patients (69.8%) suffering from PHPT had single adenomas or hyperplasia in the final histology, and are female and over the age of 60 years, similar to the literature [25, 26]. This study observed a rate of papillary thyroid (micro)carcinoma and PHPT of 10.3%, lying within the literature rates of 0.9–18.2% [27]. One must consider that the majority of the carcinomas in this study were papillary microcarcinomas, which has an incidence rate of up to 12% [28] and that Austria is an endemic goitre region.

In the literature, the rate of osteoporosis lies between 39 and 62.9% [29, 30]. The higher prevalence observed in this study could be influenced by bias as the diagnosis of osteoporosis may have led to the screening for PHPT. The frequency of depression observed in the study lies within the rate cited in the literature (30–62%). However, one must keep in mind that neuropsychiatric symptoms can be missed in the diagnosis of PHPT, especially in elderly patients [31, 32]. Kidney stones are typically seen in 10–20% of patients [33]. The slightly higher rate seen in this study might be related to the higher prevalence of osteopenia/osteoporosis, given that the calcium status of the bones influences the production of kidney stones [34].

According to the consensus statement of ESES, preoperative imaging should include MIBI scan as the initial test, but US by an experienced investigator can also be conducted [24]. In our institution, all patients undergo preoperative US and MIBI scan. In a study by Prager et al. [2], the combined sensitivity of US and MIBI scan was 98%.

In this study, 64.0% of patients had at least one thyroid nodule in preoperative US. While this rate is higher than in previous studies [35], in endemic goitre regions, the rate of PHPT with thyroid nodules ranges between 15 and 84% [2‐6]. Therefore, preoperative US is so important in patients with PHPT [2, 3]. In this analysis, sensitivity and PPV of US lie within the literature range of 51–96% [1, 2]. A high rate of false negative patients contributed to the low sensitivity. Additionally, the low PPV of this analysis was due to the high rate of false positive results. While US is safe, less expensive, non-invasive and quick, its sensitivity can be decreased in cases of multiple parathyroid adenomas or parathyroid hyperplasia, as well as concomitant thyroid nodules [36, 37]. In addition, US is operator-dependent, and it can be difficult to discriminate between parathyroid adenoma, small thyroid nodules and lymph nodes [38].

In patients with concomitant thyroid nodules, the sensitivity and PPV of US were both very low with 57.0% and 43.6%, respectively, which has been observed in other studies [2, 39]. Statistical analysis showed a significant difference in preoperative US diagnostics between patients with and without thyroid nodules (p = 0.029). While this study showed no statistically significant difference between benign nodules and thyroid carcinoma in the preoperative US, sensitivity was decreased in patients with benign nodules versus those with thyroid carcinoma. Studies have investigated US and MIBI in PHPT patients with thyroid nodules. While Mazzeo et al. [40] showed no difference in the sensitivity between the two, several studies have demonstrated that the sensitivity of both US and MIBI was decreased in patients with concomitant thyroid nodules [11, 39], as was observed in this study.

In MIBI imaging, thyroid nodules can result in a higher uptake without washout, leading to false positive and even false negative results [10]. The overall sensitivity of MIBI lies between 61 and 80% [41, 42]. The sensitivity in this analysis was 73.6%, which lies within the reported range. MIBI also had a high rate of false negative patients, which contributed to the relatively low sensitivity. PPV of MIBI was higher than with US. Size of the parathyroid adenoma and concomitant thyroid nodules are the most important factors influencing the PPV of MIBI [42]. Given that in an endemic goitre region more thyroid nodules are observed, this can explain the relatively low PPV in this analysis.

In patients with concomitant thyroid nodules, sensitivity and PPV were slightly lower than in patients without nodules. Statistical analysis showed no significant difference. A negative influence of thyroid lesions on MIBI sensitivity has also been observed in previous studies [20, 39]. Hyper-functional thyroid cells in benign lesions can cause false positive MIBI results [14]. The overall sensitivity and PPV of MIBI in benign thyroid nodules were 73.9% and 61.2%, respectively. In this study, 9.0% of patients had false positive MIBI results with benign thyroid nodules, 8.6% of patients nodular goitre in the final histology, 0.2% of patients Graves’ disease and 0.2% of patients thyroiditis. As with the study by Prager et al. [2], this study did not show a significant difference in the sensitivity and PPV of MIBI in patients with and without thyroid nodules.

Thyroid carcinoma is found in 1–36% of patients with PHPT [7, 14, 35]. In this study, 11.3% of patients had thyroid carcinoma in the final histology, lying within the reported literature range. Papillary microcarcinoma was the most common in 82.1% of patients, followed by papillary thyroid carcinoma in 8.9% of patients, 3.6% medullary microcarcinoma, 1.8% medullary thyroid carcinoma, 1.8% insular thyroid carcinoma and 1.8% follicular thyroid carcinoma.

The sensitivity of MIBI in patients with thyroid carcinoma observed in this study lies within the reported literature range of 55–91% [5, 14, 20, 35]. In this analysis, PPV was slightly higher than sensitivity. Both sensitivity and PPV did not show statistically significant differences between thyroid carcinoma and benign thyroid nodules. A meta-analysis by Treglia et al. [14] determined that MIBI is a sensitive diagnostic tool in the workup analysis of malignant thyroid nodules. However, it seems that MIBI uptake is not specific for thyroid carcinomas [14]. This analysis also found no correlation between histological type of thyroid carcinoma and MIBI uptake. Furthermore, there seems to be no correlation between MIBI uptake and thyroid nodule size [43]. In terms of false positive MIBI results, 22.4% of patients had thyroid carcinoma in the final histological report. Upon further retrospective analysis of the MIBI scans, five patients were MIBI positive, but the uptake was falsely identified as a parathyroid adenoma and not thyroid carcinoma. In an additional five patients, MIBI scan was positive for the same side as the thyroid carcinoma, but in a different location, and in two patients, thyroid carcinoma was on the opposite side of the positive MIBI scan. In one patient, multi-glandular disease could not be ruled out, but the location of the positive uptake correlated with the thyroid carcinoma. This study could also not determine a correlation between the histology of the thyroid carcinoma and its influence on MIBI uptake, similar to the meta-analysis by Treglia et al. [14]. In total, 15% of patients with microcarcinomas had false positive MIBI scans. However, in all of these patients, the correct side was diagnosed in the MIBI scan. Therefore, the presence of the small cancers did not interfere with the imaging to the same degree as the larger nodules (> 10 mm). Nevertheless, preoperative detection of occult thyroid malignancy in MIBI scan is important in patients initially only being considered for MIP.

There are several factors that can affect the sensitivity of an imaging technique, such as the level of expertise and variations in the patient population, but also if the institution is a low- or high-volume centre [44]. By using more than one imaging modality, a more accurate preoperative location can usually be diagnosed [42]. While the new techniques, such as PET/CT, are promising, it is associated with higher costs and is not available in low-volume centres. One must take into consideration that higher initial imaging costs may correlate with higher rates of intraoperative parathyroid localisation, thus subsequently lowering the overall costs of parathyroid treatment.

The majority of operations conducted were MIP in 42.3% of patients, followed by BNE in 35.6% of patients and UNE in the remaining 22.1% of patients. Concomitant thyroid surgery was conducted in 39.8% of patients. Given that Austria is an endemic goitre region, this high rate in additional thyroid operations is not surprising. A survey conducted by the American Association of Endocrine Surgeons stated that 14–30% of parathyroid surgeons would plan for concomitant thyroid surgery, regardless of nodule size [45]. However, even in endemic goitre regions, the extent of resection should be individualized due to the higher morbidity associated with total thyroidectomy [46]. In the literature, overall morbidity and mortality rates of concomitant (hemi)-thyroidectomy are rather low [47]. Recurrent laryngeal nerve palsy occurs in 2.1–4.1% (permanent in 0.3–1.2%) [48‐50] and hypoparathyroidism in 6.6–24% (permanent in 0–4.4%) [48, 51, 52]. On the other hand, a carcinoma rate of 6–39% has been reported in PHPT patients with concomitant thyroid nodules [4, 53, 54]. The main indication for concomitant thyroid surgery is exclusion and appropriate treatment of malignancy and national guidelines should be adhered to [55, 56]. While patients who receive concomitant thyroid operations have a higher risk for temporary recurrent laryngeal nerve palsy and hypoparathyroidism, the high rate of incidentally detected thyroid carcinoma in endemic goitre regions indicates thyroid surgery when thyroid nodules are detected in preoperative ultrasound [57].