Selective venous sampling supports localization of adenoma in primary hyperparathyroidism.

BACKGROUND: Selective venous sampling (SVS) is an invasive localization study for persistent or recurrent hyperparathyroidism.
PURPOSE: To assess the role of SVS in addition to non-invasive imaging for primary hyperparathyroidism (pHPT).
MATERIAL AND METHODS: This study was approved by the institutional review board and included 14 patients who underwent SVS and subsequent parathyroidectomy between January 2014 and April 2017 following a clinical diagnosis of pHPT. All patients underwent pre-SVS non-invasive imaging, including ultrasound, computed tomography (CT), and 99mTc-MIBI scintigraphy, and sensitivity was assessed using the operative and pathological findings.
RESULTS: In all but one case, a single parathyroid adenoma was responsible for the pHPT; the remaining case exhibited a chemical response following surgical removal of parathyroid tissue. The sensitivity (%) for ultrasound, CT, 99mTc-MIBI scintigraphy, and SVS was 76.9, 84.6, 69.2, and 76.9, respectively. SVS yielded positive results in four patients with discordant results and one patient with non-detectable results on imaging. In seven patients, a significant increase in the intact parathyroid hormone level was recognized only in the thyroid veins. The procedure time was in the range of 52-183 min (median = 89.5 min).
CONCLUSION: The addition of SVS to a non-invasive imaging study would be helpful to locate the responsible lesion of pHPT with discordant or non-detectable results on imaging for initial surgical treatment as well.

Introduction

Primary hyperparathyroidism (pHPT) is a relatively common disease (1–3). Surgery is currently the curative treatment for pHPT, although some patients do not require surgery because they are asymptomatic (4). Currently, minimally invasive parathyroidectomy (MIP) is preferred and appears to cause fewer cases of hypocalcemia and recurrent laryngeal nerve injury as compared with conventional bilateral neck exploration (5–7).

In the era of MIP, surgeons require more precise preoperative localization studies to identify the location of the functioning parathyroid gland and its surrounding structures. These studies usually involve non-invasive imaging examinations such as ultrasound (US) and 99mTc-Sestamibi (MIBI). Although both methods provide good detectability, some cases remain complicated by non-localizing or discordant results. Selective venous sampling (SVS) is an invasive method of localization (1). The precise localization afforded by SVS should benefit surgeons especially if non-invasive imaging is inconclusive. Several studies have demonstrated the feasibility of SVS (1,8–15); however, to our knowledge, few have evaluated its ability to localize the hyperfunctioning gland in patients indicated for initial surgery.

This study aimed to show the role of SVS in addition to non-invasive imaging for pHPT.

Material and Methods

Patients

Institutional review board approval was obtained for this study and the requirement for informed consent was waived due to the retrospective design. Our hospital is a tertiary center with referrals. Twenty-six patients underwent SVS between January 2014 and April 2017 following a clinical diagnosis of pHPT. Twelve asymptomatic patients had been followed up because they were not satisfied with the surgical criteria (16). In total, 13 of 26 patients subsequently underwent parathyroidectomy as initial surgical treatment. Furthermore, one patient previously had undergone a left parathyroidectomy for pHPT; however, the parathyroid adenoma was not identified and the clinical findings did not improve (Tables 1 and 2, Case 3). Consequently, this patient was indicated for repeat surgical treatment for persistent HPT. A total of 14 patients (6 men, 8 women; age range = 17–86 years; median age = 66.5 years) were included in this study.

Table 1.

Individual patient details illustrating findings, evaluation of imaging results, location of culprit gland, and calcium/iPTH value before and after surgery.

Case	1	2	3	4	5	6	7	8	9	10	11	12	13	14
Age	66	67	71	79	63	74	64	72	79	17	50	35	86	57
Sex	M	F	F	F	M	M	M	F	F	F	F	M	F	M
US	×	○LtL	○RtL	○RtL	○RtU	○RtL, LtL	○RtL	×	○LtL	○RtL	○RtL	×	○Rt	○Rt
CT	○RtL	○LtL	○RtL	○RtL	○RtU	○RtL	○RtL	×	○LtL	○RtL	○RtL	×	○Rt	○Rt
MIBI	○RtL	×	×	○RtL	×	○RtL	○RtL	×	○LtL	○RtL	×	○LtL	○Rt	○Rt
Site of parathyroidectomy	RtL	LtL	RtU	RtL	RtU	RtL, LtL, LtU	RtL	RtU	LtL	RtL	RtL, RtL, LtL	LtL	RtU	RtU
Location	RtL	LtL	RtU	RtL	RtU	RtL	RtL	RtU	LtL	RtL	Unspecified	LtL	RtU	RtU
Ca before surgery (mg/dL)	11.6	12	10.6	10.7	11.4	11.7	11.7	14.5	12.8	9.8	9.8	11.9	11.4	10.9
Ca after surgery (mg/dL)	9.2	9.3	8.9	9.4	9.4	9	9.1	9.2	9.4	9.1	9.1	10.2	9.4	9.2
iPTH before surgery (pg/mL)	93	80	80	189	127	91	260	332	78	442	97	128	581	144
iPTH after surgery (pg/mL)	16	42	None	30	51	10	47	55	33	56	63	46	63	21

, location was suggested; ×, non-detectable (or undetected); RtL, right lower; RtU, right upper; LtU, left upper; LtL, left lower; Rt, right side; Ca, serum calcium.

Table 2.

Results of iPTH value (pg/mL) at each sampling point, predictive site on SVS, location of culprit gland, and procedure time.

Case		1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. Rt. jugular (distal site of STV)		115	128	87	131	115	108	221	138	121	1100	56	117	591	138
2. Rt. jugular (joining site of STV)		104	123	97	137	1550*	106	274	133	110	9800*	63	118	695	128
3. Rt. jugular (proximal site of STV)		103	124	82	143	371*	121	238	150	123	3340*	87	120	1070	125
4. Rt. jugular (joining site of MTV)		412*	106	88	140	423*	203	233	158	144	3040*	60	121	676	131
5. Rt. jugular (proximal site of MTV)		431*	121	1910*	142	803*	231*	215	142	136	2690*	60	110	649	153
6. Rt. brachiocephalic		146	112	173	129	235*	177	268	151	159	2210*	61	111	660	237
7. Rt. ITV						NA	484*	15,200*	6960*	Impossible	1150*	Impossible	151	4370*	2170*
8. Lt. ITV	8. d Lt. ITV (distal)	69	14,200*	94	Impossible	NA	NA	294	166	Impossible	1450*	1880*	3270*	3620*	129
	8. p Lt. ITV (proximal)	94	2950*	85	Impossible
9. Lt. jugular (distal site of STV)		104	127	94	126	107	105	233	146	121	623	63	130	554	125
10. Lt. jugular (joining site of STV)		102	140	98	135	122	121	229	171	124	573	98	108	584	138
11. Lt. jugular (proximal site of STV)		107	156	94	138	123	142	241	163	119	617	70	120	630	147
12. Lt. jugular (joining site of MTV)		102	126	83	137	119	118	237	146	134	630	74	121	641	140
13. Lt. jugular (proximal site of MTV)		100	153	87	116	115	111	220	141	124	641	74	121	640	138
14. Lt. brachiocephalic		100	130	88	158	111	99	232	149	175	620	68	115	644	145
15. SVC		107	156	87	101	102	108	230	230	112	1110	84	173	898	112
16. Rt. STV						16,100*	131	301	NS	NS	1100	2420*	99	2710*	7630*
17. Rt. MTV						6020*	Impossible	NA	NA	Impossible	2600*	3850*	Impossible	NA	8610*
18. Lt. STV						NS	107	236	NS	NS	834	971*	217	661	151
19. Lt. MTV						NS	210	NA	NA	Impossible	640	NA	181	NA	Impossible
Predictive site		○RtL	○LtL	○RtL	×	○RtU	○RtL	○RtL	○RtL	×	ΔRt	×RtU, Lt	○LtL	×Rt, LtL	ΔRt
Location		RtL	LtL	RtU	RtL	RtU	RtL	RtL	RtL	LtL	RtL	unspecified	LtL	RtU	RtU
Procedure time (min)		83	183	80	113	81	139	77	86	52	92	87	124	108	149

Significant value of iPTH.

Rt, right; Lt, left; STV, superior thyroid vein; MTV, middle thyroid vein; ITV, inferior thyroid vein; SVC, superior vena cava; NS, not selected; NA, not applicable; ○, probable; Δ, possible; ×, non-detectable; RtL, right lower; RtU, right upper; LtL, left lower; Impossible, the branch could not be catheterized.

Preoperative non-invasive localizing examination

All 14 patients underwent non-invasive localizing examinations, including US, computed tomography (CT), and MIBI scintigraphy, and single photon emission computed tomography (SPECT). For the purposes of this study, we used the findings reported in our institution’s radiology reporting system.

Ultrasound

Each patient underwent an ultrasonographic scan (Aplio ultrasound system; Toshiba Medical System, Kawasaki, Japan) performed by an experienced US technologist using a linear probe with a frequency of 8 MHz. The technologist evaluated suspicious nodules at and/or adjacent to the thyroid gland and recorded the findings in the radiological reporting system.

MIBI scintigraphy and SPECT

Dual-phase scintigraphy was performed 20 min following radiotracer administration (600–740 MBq of MIBI) and at 2 h post injection. Planar and SPECT images were obtained using a two-head scanner (Symbia Evo Excel, e.cam; Siemens Medical Solutions, Malvern, PA, USA). A nodule exhibiting retained uptake in the delayed phase was considered a suspicious lesion.

CT scanning technique and image analysis

All contrast-enhanced CT images from the skull base to the lower edge of the aortic arch were obtained on a multidetector-row CT scanner (Aquilion 64; Toshiba Medical Systems, Kawasaki, Japan). Iodinated contrast material (300 mg I/mL) at a dosage of 2 mL/kg was injected at a rate of 1.5 mL/s. At 90 s post injection, the CT images were obtained using the following parameters: collimation = 32 × 0.5 mm; tube voltage =120 kVp; tube current = 200 mA; gantry rotation time = 0.75 s; and beam pitch = 0.844 (helical pitch = 27). The scan data were reconstructed to axial images with slice thicknesses of 1 mm and coronal images with slice thicknesses of 2–3 mm without section interval overlaps. The imaging field of view was 240 × 240 mm and the pixel size was 0.468 × 0.468 mm.

The radiologists performing the SVS evaluated the suspicious nodule before venous sampling with US and MIBI examination data. Thyroid vein anatomy was recorded on CT scans in the manner described in a previous report on preparatory practices for SVS (17). First, the superior and middle thyroid veins and the sites at which these veins joined the internal jugular veins were identified. Second, the inferior veins and the sites at which they joined the brachiocephalic veins were identified using transverse and coronal images in accordance with a previous study reporting the various branching patterns of inferior veins (18).

Selective venous sampling

After puncturing the right femoral vein, a 4-Fr multipurpose, Berenstein catheter (Heartcath, 100 cm or 120 cm; Terumo, Tokyo, Japan) or 4.2-Fr H1H catheter (IR catheter, 100 cm; Hanako, Saitama, Japan) was used to obtain blood samples at the bilateral jugular and brachiocephalic veins. The initial sampling points at the jugular vein were set by the initial protocol until December 2014; however, beginning in January 2015, these sites were modified by the joining sites of the thyroid veins as determined by CT scans. Of the thyroid veins, only the left inferior thyroid vein was sampled until December 2014 (Fig. 1a). Subsequently, attempts were made to catheterize all thyroid veins (Fig. 1b).

Fig. 1.

Sampling points for selective venous sampling (SVS). (a) Initial protocol sampling points until December 2014. Only the left inferior thyroid vein was sampled at the distal and proximal sites. (b) Sampling points of the modified protocol since January 2015. An attempt was made to catheterize every thyroid vein. 1, 9, the upper superior thyroid vein; 2, 10, the joining site of the superior thyroid vein; 3, 11, between the superior and middle thyroid veins; 4, 12, the joining site of the middle thyroid vein; 5, 13, the joining site of the internal jugular and subclavian veins; 6, the right brachiocephalic vein; 7, the right inferior thyroid vein; 8, the left inferior thyroid vein; 14, the left brachiocephalic vein; 15, the superior vena cava; 16, the right superior thyroid vein; 17, the right middle thyroid vein; 18, the left superior thyroid vein; 19, the left middle thyroid vein.

When the left interior thyroid vein was catheterized, retrograde venogram was performed to opacify the other thyroid veins through their connections (Figs. 2 and 3), which enabled us to determine the sampling point in the catheterized thyroid vein and confirm the existence of a particular thyroid vein that was not depicted on CT scans. Next, a 2.3-Fr microcatheter with a side hole (Gold Crest; HI-LEX Co., Hyogo, Japan) was used in a coaxial manner. We introduced the microcatheter into both the right and left thyroid veins where they formed a common trunk and collected more blood samples from these veins when possible. The samples were not collected if we could not visualize a certain thyroid vein on CT scan or retrograde venogram. The intact PTH (iPTH) level was measured in all blood samples and the SVS procedure time was recorded from the medical charts.

Fig. 2.

A 67-year-old female patient with primary hyperparathyroidism (Case 11). A retrograde venogram from the common trunk of the inferior thyroid veins (7 and 8) shows the superior (16 and 18) and right middle (17) thyroid veins through the venous network.

Fig. 3.

A 64-year-old male patient with primary hyperparathyroidism (Case 7). A retrograde venogram from the left superior thyroid vein is shown. The bilateral middle thyroid veins are not identified. Other thyroid veins are opacified through the venous network. A significant increase in the intact PTH was recognized only in the right inferior thyroid vein (7), and the SVS localization diagnosis was the right lower quadrant, which was considered probable.

Interpretation of the SVS results

We defined a significant increase in iPTH concentration as a twofold increase relative to the level in the superior vena cava. The responsible lesion sites were divided into four quadrants (right upper, right lower, left upper, and left lower) assuming there were four parathyroid glands. The result was defined as probable if the responsible quadrant could be clearly specified, as possible if we could not localize the quadrant but could localize the approximate region (e.g. left or right side), and as non-detectable if we could not localize the approximate region or failed to observe a significant increase in iPTH. Probable and possible were both considered positive results.

Correlations with operative findings and postoperative biochemical responses

The suspicious lesions predicted by US, CT, and MIBI were compared with the sites of responsible parathyroid lesions according to the operative and pathological findings; the SVS results were also correlated with the operative and pathological findings. Sensitivity was calculated for non-invasive studies and SVS, respectively.

The post-parathyroidectomy iPTH values were obtained from the patients’ medical charts. Treatment was considered successful if the iPTH value improved to within the normal range.

Results

Operative findings and postoperative biochemical responses

Table 1 shows the results of non-invasive imaging studies, the operative locations of the responsible glands, and the biochemical responses. Pathologically, a single parathyroid adenoma was the responsible lesion in all but one case. In one patient (Case 11), surgical treatment was performed for three parathyroid glands; however, an adenoma was not detected pathologically, thus the responsible lesion was unspecified. Therefore, 13 patients in whom the responsible lesion was specified were the subjects of the sensitivity calculation.

The serum calcium and iPTH values decreased postoperatively in all 14 patients, including Case 11 (Table 1). Accordingly, all treatments were considered successful in terms of the biochemical response.

Noninvasive localization studies

Of the 13 adenomas (13 patients), US depicted ten, CT scans depicted 11, and MIBI scintigraphy depicted nine; the sensitivity of US, CT, and MIBI scintigraphy was 76.9%, 84.6%, and 69.2%, respectively. The cross-sectional imaging and MIBI scintigraphy findings were consistent in eight patients in whom the responsible adenoma quadrants were consistent with those identified in imaging studies. However, the two types of studies yielded discordant results in four of the 13 patients; in the remaining patient, none of the non-invasive imaging studies were able to depict the suspicious lesion.

Table 2 presents the iPTH values at all SVS points. No patients developed complications related to SVS. The procedure time was 52–183 min (median = 89.5 min).

SVS detected the quadrant suspected of being responsible for the culprit gland in eight of 13 patients and the responsible side in two of 13 patients. In the other three patients, the thyroid veins could not be sampled and no significant increase in iPTH was observed in two patients, and it could not be even localized approximate region in one patient; accordingly, these cases were defined as non-detectable. The sensitivity of SVS was 76.9%.

Notably, in the patients with a single adenoma, SVS yielded positive results in four patients with discordant results and one patient with non-detectable results on non-invasive imaging studies (probable in all patients). Conversely, three patients in whom SVS was non-diagnostic had concordant findings on non-invasive localization.

Significant increases in iPTH levels were identified in the jugular veins of five of 14 patients and the greatest iPTH gradient was noted in the thyroid veins in two patients (Fig. 4). In seven patients, a significant increase was recognized only in the thyroid veins and four of these patients had a steep increase in iPTH concentration in the inferior thyroid vein; however, it did not increase in the brachiocephalic vein near the joining site.

Fig. 4.

A 74-year-old male patient with primary hyperparathyroidism (Case 6) who was pathologically diagnosed with parathyroid adenoma in the right lower gland. (a) US showed the low echoic nodule caudal to both thyroid lobes, which could not be distinguished from the lymph nodes. (b) CT demonstrated a small nodule caudal to the right thyroid lobe. (c) MIBI in the delayed phase showed slightly increased uptake in the right thyroid lobe (arrowhead), which was ambiguous as a significant feature. (d) A retrograde venogram from the right superior thyroid vein and left middle thyroid vein are shown, and the values of intact PTH (iPTH) were noted at the veins. The most significant increase in iPTH was recognized in the right inferior thyroid vein (7) and a significant increase in iPTH was also revealed in the right jugular vein where it joins with the middle thyroid vein (5). The SVS localization diagnosis was probably the right lower quadrant.

Discussion

There are two important clinical findings regarding SVS. First, SVS is feasible in patients with discordant or non-detectable non-invasive imaging study results. Second, SVS of blood from the thyroid veins was beneficial.

Although diagnostic non-invasive imaging studies are able to localize lesions (2,19–24), discordant or equivocal results are often obtained from multiple examinations. Our study revealed that SVS yielded positive (i.e. probable or possible) results in all patients with discordant or non-detectable findings and has been reported to yield a high sensitivity (1,11–15). Two previous studies found that SVS enabled localization of responsible lesions in affected patients with non-localizing or discordant findings (25,26). Therefore, the performance of SVS in addition to non-invasive imaging studies could have a role in improving the confidence of predictions as well as the accuracy of localization diagnoses.

We found that it was beneficial to collect samples selectively from the thyroid vein whenever possible; however, few studies have discussed the importance of the sampling point in diagnostic ability. In the present study, we observed the greatest iPTH gradient in the thyroid veins, although two patients also exhibited elevated iPTH levels in the jugular veins. In addition, seven patients had significant increases in iPTH levels only in the thyroid veins. Additionally, the iPTH level decreased rapidly as the sampling point became incrementally distant from the lesion area. Gimm et al. suggested that SVS of the thyroid veins would provide detailed spatial resolution (1); therefore, if the thyroid veins are sampled selectively to the extent possible, we can expect a higher sensitivity. Conversely, catheterization of multiple thyroid veins is time consuming and can be technically challenging. SVS complexity seems to increase in accordance with the variability in the structure of the thyroid veins. The bilateral superior, middle, and inferior thyroid veins have been depicted in the literature as typical scenarios (8,9); however, multidetector-row CT scans have revealed variations in the structure of the thyroid veins (18). Several patterns of inferior veins featuring different numbers and joining sites to the brachiocephalic vein have been observed and the middle veins are frequently absent. Therefore, familiarity with the thyroid vein anatomy should improve the success of thyroid vein catheterization (17). Retrograde venous angiography can depict the venous network in the thyroid region and can thus assist clinicians in determining sampling points in the thyroid veins.

This study had several limitations. First, the blood sampling method changed during the study period, i.e. the procedure was not consistent. Second, we could not selectively obtain samples from every thyroid vein in all cases. Third, all pathologically proven responsible lesions in 13 patients were identified as single adenomas. In the remaining patient, multiple quadrants were identified on SVS as suspicious, and even though the resected tissue included parathyroid tissue, an adenoma was not pathologically confirmed. Although the patient experienced symptom relief and exhibited a biochemical response, the feasibility of SVS for multinodular lesions remains unclear. Fourth, recent reports have described improved sensitivity rates for four-dimensional CT and SPECT/CT (27–30); however, these modalities were not used in the present study, and their inclusion might have altered the significance of SVS.

In conclusion, the use of SVS in addition to non-invasive imaging studies could have a role in improving confidence regarding the localization of a responsible lesion. We considered it beneficial to collect samples selectively from the thyroid veins whenever possible after anatomical imaging to determine the site that would yield the greatest iPTH gradient and thus the best SVS spatial resolution.