Comparing early and delayed [^99mTc]Tc-MIBI SPECT/CT parathyroid scans: agreement, confidence levels, and clinical predictive factors

Abstract

Background

Parathyroid scan is an important imaging modality for localizing hyperfunctioning parathyroid tissue in patients with hyperparathyroidism. Unfortunately, whether early or delayed timing is the optimal protocol for [^99mTc]Tc-MIBI SPECT/CT parathyroid remains under debate. This study aimed to evaluate the agreement and compare the confidence levels of physicians when interpreting early and delayed [^99mTc]Tc-MIBI SPECT/CT parathyroid scans. Additionally, it sought to identify clinical factors that related to positive scan result. We conducted a prospective study where the early and delayed [^99mTc]Tc-MIBI SPECT/CT was separately interpreted as either positive or negative. Furthermore, these interpretations were categorized based on whether they fell within more or less confidence levels of the readers and were correlated with clinical information.

Results

We enrolled 39 patients with hyperparathyroidism with 158 possible locations of parathyroid glands. The per-location agreement between the early and delayed scans was moderate (concordant rate: 80.3%, Kappa = 0.558), and the per-patient agreement was slight (concordant rate: 71.8%, Kappa = 0.093). The confidence of interpretation was significantly higher for the delayed scans. Calcium supplementation, low serum parathyroid hormone levels, and low serum phosphate levels were associated with positive early scans. High calcium level and high parathyroid hormone levels were associated with positive delayed scans.

Conclusions

Our study highlights the impact of the timing of SPECT/CT in [^99mTc]Tc-MIBI parathyroid scans. The different confidence levels between early and delayed scans, along with clinical factors, imply that various factors affect parathyroid scan interpretation, and individualized scanning protocols adjusted for specific settings may be needed to optimize the successful localization of hyperfunctioning parathyroid tissue.

Background

Hyperparathyroidism is an endocrine disease that can adversely affect multiple organs. The disease is characterized by the excessive secretion of parathyroid hormone (PTH), which increases serum calcium levels from bone resorption, as well as increased calcium resorption and the production of 1,25-dihydroxyvitamin D in the kidneys. Hyperparathyroidism can be classified as primary, secondary, and tertiary. Primary hyperparathyroidism results from an intrinsic abnormality of one or more parathyroid glands, such as parathyroid adenoma. Secondary hyperparathyroidism occurs as a result from physiologic PTH secretion stimulated by hypocalcemia, which is often caused by vitamin D deficiency and chronic renal disease. In some patients, long-standing secondary hyperparathyroidism can lead to the autonomous secretion of PTH and hypercalcemia, resulting in a condition known as tertiary hyperparathyroidism. Hyperparathyroidism patients may present with a variety of symptoms and signs, such as hypercalcemia, osteoporosis, and kidney or urinary tract stones [1].

One curative treatment of hyperparathyroidism is the surgical removal of hyperfunctioning parathyroid glands, and precise preoperative localization of the hyperfunctioning glands is required for a successful operation. Several imaging modalities are for the localization of parathyroid glands, including ultrasonography, computed tomography (CT), magnetic resonance imaging, and radionuclide parathyroid scans. Among these modalities, radionuclide parathyroid scans possess the advantages of being able to localize hyperfunctioning ectopic parathyroid gland, and can be safely performed in patients with renal failure [2, 3].

Radionuclide parathyroid scans can be performed using various techniques, with the dual phase technique being one of the widely used. This technique utilizes the radiopharmaceutical technetium-99m sestamibi ([^99mTc]Tc-MIBI) and leverages its prolonged retention in parathyroid tissue to localize parathyroid glands and differentiate them from the adjacent thyroid tissue. Although delayed scanning after the washout of [^99mTc]Tc-MIBI from the thyroid tissue was indispensable in the era of planar gamma cameras, the advent of single-photon emission computed tomography (SPECT)/CT allows for differentiation between parathyroid tissue from thyroid tissue on the early scans, reducing the total time needed to performed the parathyroid scans and enabling the localization of parathyroid gland with rapid [^99mTc]Tc-MIBI washout [4].

Despite the potential benefits of early SPECT/CT in parathyroid scans, the evidence on this issue remains inconclusive. Some studies suggest that substituting a delayed SPECT/CT with an early SPECT/CT in parathyroid scans may result in a deterioration of scan sensitivity and detection rate [5], while others suggest that early SPECT/CT can achieve superior accuracy and detection sensitivity compared to delayed SPECT/CT [2, 6,7,8].

Given the absence of universal consensus regarding the optimal timing for SPECT/CT in parathyroid scans, this study was conducted to evaluate the agreement between interpretations of parathyroid scans using early SPECT/CT and those using delayed SPECT/CT. Additionally, we compared the confidence levels in the interpretation of early scans to those of delayed scans and analyzed clinical information predictors related to positive early and delayed scans.

Methods

This study is a single-center prospective study, in which we recruited patients sent for [^99mTc]Tc-MIBI parathyroid scans at (The name of the institution is intentionally omitted to ensure blind review process) from December 2022 to September 2023. The patients had to be 18 years or older and had serum parathyroid hormones and blood chemistry test measured within 6 months prior to the parathyroid scans. Patients with history of neck surgery and those who could not complete the image acquisition and reconstruction process were excluded (Fig. 1). Demographic data, concurrent medications, serum intact PTH levels, serum calcium, serum phosphate, serum albumin, serum vitamin D, and serum creatinine, were retrieved from medical records.

Fig. 1

A diagram showing patient recruitment and enrollment process

Once recruited, the patients underwent parathyroid scans with the standard protocol of our institution with an additional early SPECT for the purpose of this study. The standard protocol consisted of [^99mTc]pertechnetate planar scans, early [^99mTc]Tc-MIBI planar scans, delayed [^99mTc]Tc-MIBI planar scans, and delayed [^99mTc]Tc-MIBI SPECT/CT. To elaborate, after the pertechnetate images were obtained, a single dose of oral perchlorate was given to wash out thyroid pertechnetate activity, and 740 MBq of [^99mTc]Tc-MIBI was injected. The early SPECT, which was acquired for the purpose of this study and was not performed in routine clinical service, was obtained 20 min after the injection of [^99mTc]Tc-MIBI and 70 min before the delayed SPECT. The two SPECTs and a low-dose CT were acquired using 64 projections of 128 × 128 matrix covering 360 degree angle using either a GE Discovery NM/CT 670 (GE Healthcare, USA) or a Siemens Symbia Intevo Bold (Siemens, USA) imaging system. The patients were in supine position throughout the entire process. Both the early SPECT and the delayed SPECT were co-registered with the CT to form SPECT/CT images using the Syngo.via workstation software (Siemens, Germany).

The early and the delayed SPECT/CT scans were interpreted separately by two readers. The readers were physicians with two and five years of experience in nuclear medicine. During the interpretation, the readers could review CT images, SPECT images, fused SPECT/CT images, and maximum intensity projection (MIP) images. Planar images were not used in the interpretation, and clinical information, including laboratory results, was not disclosed to the readers throughout the interpretation process. If there were disagreements on the interpretation, the two readers reached a consensus through discussion. The SPECT/CT scans were examined for the potential locations of parathyroid glands. Each scan was classified as either positive or negative for hyperfunctioning parathyroid tissue, both on a per-patient and per-location basis. Additionally, the readers classified the confidence levels of each interpretation into more confident or less confident levels.

SPSS and R software packages were used for statistical analysis. Data were presented as numbers with percentages, means with standard deviations, or medians with interquartile ranges, as appropriate. The agreement between the early and the delayed SPECT/CT was described as concordance rates with Cohen’s kappa [9]. The confidence levels were evaluated by comparing the proportions of scans the readers classified as belonging the more confident level using McNemar’s Chi-square test for paired proportions. The agreement and the confidence levels were analyzed both on a per-location and per-patient basis. Categorical predictors of positive scan were identified using Pearson’s Chi-square test. Numeric predictors of positive scans were identified using receiver operating characteristic (ROC) area under the curve (AUC) analysis considering AUC > 0.600 as a satisfactory predictors [10]. P values < 0.05 were regarded as statistically significant.

Results

This study included a total of 39 patients, with 15 (38.5%) being males and 24 (61.5%) being females, all of whom had completed the research scan protocols. The average age of the patient was 55.0 ± 14.0 years. Other baseline characteristics of the patients can be found in Table 1. In addition to the four normal locations of the parathyroid glands in each patient, two patients were suspected to have ectopic parathyroid glands, resulting in a total of 158 potential locations of parathyroid glands.

Table 1 Patient characteristics

Interpretation agreement is demonstrated in Table 2. To summarize, the location-based analysis showed that the early scans were positive in 53 (33.5%) locations and the delayed scans were positive in 52 (32.9%) locations. The early and the delayed scans were concordant in 97 locations (Fig. 2) and discordant in 61 locations (Fig. 3), equating to a concordance rate of 80.3% (Kappa = 0.558, P value < 0.001), a moderate agreement based on the kappa value. Forty-eight (30.4%) locations in the early scans and 97 (61.4%) locations in the delayed scans were considered belonging to the more confidence level. This indicates a significantly greater proportion of locations in the delayed scan falling into the category of higher confidence (McNemar’s χ2(1) = 34.0, P value < 0.001).

Table 2 Interpretation agreement on scan positivity between the early and the delayed scans

Fig. 2

A case with concordant scans. A 42-year-old female patient presented with hypercalcemia and osteoporosis. Her laboratory investigation showed serum calcium of 11.4 mg/dL (normal range: 8.5–10.5 mg/dL), serum phosphate of 2.4 mg/dL (normal range of 2.3–4.7 mg/dL), and serum intact parathyroid hormone of 86.0 pg/mL (reference range: 15–65 pg/mL). She was diagnosed by an endocrinologist as primary hyperparathyroidism and was sent to perform a parathyroid scan to localize hyperfunctioning parathyroid tissue. Her early [^99mTc]Tc-MIBI SPECT/CT revealed an avid hypodense nodule inferior to the left lower pole of the thyroid gland, which was suspicious for enlarged left lower parathyroid gland that was evident in both the early and the delayed scans. The patient proceeded to surgery, in which the nodule was resected and the pathological diagnosis revealed left lower parathyroid gland with water clear cell hyperplasia. Subfigures: CT (a), early SPECT/CT (b), delayed SPECT/CT (c), early MIP (d), and delayed MIP (e)

Fig. 3

A case of discordant scans. A 47-year-old female patient presented with hypercalcemia during a check-up. Her blood test showed serum calcium of 11.5 mg/dL (normal range: 8.5–10.5 mg/dL), serum phosphate of 3.9 mg/dL (normal range: 2.3–4.7 mg/dL), and serum intact parathyroid hormone of 133 pg/mL (reference range: 15–65 pg/mL). Her [^99mTc]Tc-MIBI SPECT/CT demonstrated two lesions that were discordant between the early and delayed scans. The lesion inferior to the right lower pole of the thyroid gland showed MIBI avidity only in the delayed scan, while the other lesion posterior to the left lower pole of the thyroid gland showed MIBI avidity only in the early scan. An operation was later performed, and the histopathologic diagnosis revealed the right and the left lesion to be lymph nodes and thyroid tissue with lymphocytic thyroiditis, respectively. Subfigures: CT of the right lower lesion (a), early SPECT/CT of the right lower lesion (b), delayed SPECT/CT of the right lower lesion (c), CT of the left lower lesion (d), early SPECT/CT of the left lower lesion (e), delayed SPECT/CT of the left lower lesion (f), early MIP (g), delayed MIP with an arrow indicating the right lower lesion (h)

In the patient-based analysis, the early scans were positive in 32 (82.1%) patients and the delayed scans were positive in 31 (79.5%) patients. The early and the delayed scans were concordant in 28 patients, resulting in a concordance rate of 71.8% (Table 2, Kappa = 0.093, P value = 0.560), a slight agreement based on the kappa value. The confidence levels of 21 (53.8%) patients in the early scans and 32 (61.4%) patients were in the more confident level, indicating a significantly larger proportion of patients in the delayed scans belonging to the more confidence levels (McNemar’s χ2(1) = 6.25, P value < 0.012).

Regarding clinical predictors related to positive scans in a patients-based analysis, a history of calcium supplementation was significantly associated with positive early scans (P value = 0.003). There was no demonstrable significant association between sex (P value = 0.792), the use of calcium channel blockers (P value = 0.139), the use of calcimimetics (P value = 0.780), or vitamin D supplementation, and positive early scans. No significant association was observed between sex (P value = 0.950), the use of calcium channel blockers (P value = 0.139), the use of calcimimetics (P value = 0.780), calcium supplementation (P value = 0.077) or vitamin D supplementation (P value = 0.820), and positive delayed scans.

Additionally, the ROC analysis revealed that low serum intact PTH (Fig. 4a-b), low serum phosphate, and high serum vitamin D were good predictors for positive early scans, while high serum intact PTH (Fig. 4c-d), high serum calcium, and low serum vitamin D were good predictors for positive delayed scans. The details of the ROC analysis, including AUC values, as well as the cutoffs, sensitivity, and specificity for predictors with ROC > 0.600, are illustrated in Table 3.

Fig. 4

The ROC analysis of serum intact parathyroid hormones levels as predictors for positive scans, illustrated as ROC curves with box and whisker plots. Low serum PTH levels were associated with positive early scans (a, b), while high serum PTH levels were associated with positive delayed scans (c, d)

Table 3 Numeric clinical predictors of positive scans. Cutoffs, sensitivity, and specificity were calculated for predictors with AUC > 0.600

Discussion

No single scanning protocol for parathyroid scans has emerged as the universally accepted most optimal for patients with hyperparathyroidism. Various protocols utilize different radiopharmaceuticals and timings of image acquisition. In this study, our focus was on the timing of SPECT/CT acquisition in [^99mTc]Tc-MIBI parathyroid scans. We measured the agreement of interpretation between the early and the delayed scans, compared the confidence levels of the readers in interpreting the scans at each image acquisition time, and tried to identify clinical factors associated with positive scans [11].

Regarding the agreement, we found that the two time points of SPECT/CT provide a moderate per-location agreement and a slight per-patient agreement. In terms of the confidence of the readers, the readers felt more confident in interpreting the delayed scans. Concerning the clinical factors, our analysis suggested that positive early scans were associated with calcium supplementation, low serum intact PTH, low serum phosphate, and high serum vitamin D levels, while positive delayed scans were associated with high serum intact PTH levels, high serum calcium, and low serum vitamin D levels.

As compared with prior studies, our concordance rate and the kappa value were lower than the concordance rate of 92.0% and the kappa value of 0.945 of the previous similar studies. This could be attributed to the difference in the scanning protocols, where our scanning protocols the injection of [^99mTc]pertechnetate before the early [^99mTc]Tc-MIBI SPECT/CT scan, which may interfere with the interpretation, particularly in the early scans [2].

The significantly higher confidence levels of the investigators in interpreting delayed scans could be attributed to several factors: the interference of [^99mTc]Tc-MIBI activity in the thyroid tissue adjacent to the parathyroid gland location, the longer time between the early SPECT acquisition and the CT scan, which might cause a higher degree of misregistration on the early SPECT/CT scans. This observation supports the use of delayed scan, which has also been shown to be more sensitive than the early SPECT/CT in identifying parathyroid adenomas, raising the detection rate from 38.9% to 58.3% in one cohort study [5].

The association between positive early scans and low serum phosphate, as well as the association between positive delayed scans and high serum calcium and high serum intact PTH levels is concordant with previous studies [12]. These abnormalities in blood chemistry are generally accepted as an indicator for severe disease and larger size of hyperfunctioning parathyroid tissue producing more excess PTH and leading to more severe blood chemistry alterations. Larger size of the hyperfunctioning parathyroid tissue is important in scan positivity since the [^99mTc]Tc-MIBI SPECT/CT has limited resolution of approximately 6 mm [13].

Interestingly, we found that the use of calcium supplements was associated with a positive result on the early scan but not on the delayed scan. The mechanism in which calcium supplements affect the sensitivity of [^99mTc]Tc-MIBI parathyroid scans is unclear. It is possible that calcium supplementation is often recommended for patients with hypocalcemia, and prolonged hypocalcemia can lead to hyperplasia of parathyroid tissue. This condition is associated with the rapid washout of [^99mTc]Tc-MIBI, resulting in a negative delayed scan [14]. Alternatively, calcium supplement might interact directly or indirectly with the anion efflux pumps, since the administration of the calcium channel blockers was shown to decrease the sensitivity of [^99mTc]Tc-MIBI parathyroid scans [15].

There are few limitations that should be considered in this study. First, we have yet to get the pathological diagnosis to confirm or compare the true sensitivity and specificity between the early and the delayed scans. Second, the interpretation of the early and the delayed SPECT may be biased by the familiarity toward the delayed SPECT. Lastly, the number of subjects in this study may not be very large and the study was performed in a single center, which could affect the generalizability of this study.

Despite its limitations, this study has some potentially important implications: First, the different predictors of positive predictors between the early and the delayed scans may imply individualized scan protocols selection for individual patients, for example, center with routine delayed SPECT may consider adding early SPECT for patients with calcium supplements, low PTH, or low phosphate levels. It also implies that the optimal protocols for parathyroid scintigraphy should remain under constant investigation to maximize the accuracy of the outcomes. We also advocated for future investigation to assess the accuracy of the early and the delayed parathyroid scintigraphy compared to the gold standard of pathological diagnosis.

Conclusions

The early and the delayed [^99mTc]Tc-MIBI SPECT/CT scans had moderate agreement in scan positivity on a per-location basis and slight agreement on a per-patient basis. However, confidence levels in interpreting delayed scans were higher compared to early scans. Positive early scans were correlated to a history of calcium use, low serum phosphate levels, and low serum PTH levels, while positive delayed scans were associated with high serum PTH levels and high serum calcium levels. Tailoring scanning protocols based on individual clinical information could potentially enhance scan sensitivity and detection rates.