| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
CLINICAL STUDY |
1 Departments of Endocrinology, 2 Radiology, 3 Nuclear Medicine, 4 Surgery and 5 Pathology, St Bartholomew’s Hospital, QMUL, London EC1A 7BE, UK
(Correspondence should be addressed to J P Monson; Email: Johnmonson{at}aol.com)
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
Design: We report a retrospective 10-year experience (n = 27) of the use of PVS in complicated primary hyperparathyroidism and contrast the use of PVS with neck ultrasound, magnetic resonance imaging (MRI), computed tomography (CT) and sestamibi imaging modalities.
Results: In 14 out of 25 patients who underwent surgery PVS results were completely concordant with surgical and histological findings and 88% of patients achieved post-operative cure. Out of 13 patients referred after previous failed surgery, 12 underwent further surgery which was curative in 9. In total PVS yielded useful positive (n = 13) and/or negative information (n = 6) in 19 out of 25 patients undergoing surgery. Using histology as the gold standard, 59% of PVS studies were entirely consistent with histology, as compared with 39% of ultrasound scans, 36% of sestamibi scans and 17% of MRI/CT scans.
Conclusions: PVS is a valuable adjunct to MRI/CT and sestamibi scanning in selected patients with complicated hyperparathyroidism when performed in an experienced unit.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
All catheters were performed by one out of two experienced interventional radiologists (M Matson, J Dacie), a single radiologist performed cross-sectional imaging (RH Rezneck) and surgery was undertaken by a single surgeon (R Carpenter).
A diagnosis of MEN1 was based on the presence of pituitary and/or pancreatic islet cell tumours in addition to hyperparathyroidism, and/or a family history of typical endocrinopathy in a patient with hyperparathyroidism, and/or a defined mutation in the menin gene.
Familial hypocalcuric hypercalcaemia was excluded due to the measurement of urine calcium clearance to creatinine clearance and this was >0.03 in all patients.
The pathology of the parathyroids was examined by an endocrine histopathologist. All glands were weighed on receipt and classified as adenomatous, hyperplastic or carcinoma based on macroscopic and histologic criteria. The diagnosis of adenoma was made on glands with an encapsulated tumour with a compressed rim of normal parathyroid. Any other excised glands had to be of normal or atrophic appearance. The adenomas were a mixture of chief cells and oxyphil cells, with the former predominating. Parathyroid hyperplasia was identified when multiple large glands with a multinodular enlargement were seen. The hyperplasia was usually made up of chief cells with occasional oxyphil cell clusters, no compressed rim of normal parathyroid was identified. The diagnosis of carcinoma was made on a constellation of features, including gland weight, the presence of fibrous bands, spindled tumour cells, mitotic figures, capsular invasion and vascular invasion. No cases of double adenoma or water clear hyperplasia were identified in this series.
A positive catheter was defined as a serum parathyroid hormone gradient of two times the peripheral measurement, as previously described (3). Peripheral serum parathyroid hormone concentrations were determined by taking the mean of measurements in the subclavian veins. This measure of positivity has been used in our audit. However, as is the nature of a retrospective review; in some cases the primary physician (WM Drake, PJ Jenkins, SL Chew and JP Monson) has used other definitions of positivity (e.g. 1.5 times the peripheral measurement) where this has occurred we have commented on this.
A combination of surgical and histological findings was used as the gold standard for analysis. Complete consistency had to be demonstrated for the localisation method to be analysed as positive, i.e. the localisation method had to demonstrate – right upper neck disease not just – right-sided disease.
Details of methodology are as follows:
Radiological technique
Venous access is obtained via the right femoral vein and seven French sheath inserted. Catheters are advanced under fluoroscopic screening into the relevant veins. Six French-shaped catheters with one or two side holes at the distal end are used to facilitate sampling. A multi-purpose catheter is often able to reach many of the desired sites; a headhunter catheter is sometimes useful to enter the left internal jugular vein, where a valve can sometimes be obstructive. A cobra catheter is useful for accessing the superior and middle thyroid veins and the left superior intercostal vein and thymic mediastinal vein. This last vein is often the most technically challenging with a small ostium opening into the left innominate vein. Even when entered, it is sometimes impossible to aspirate blood. In these cases in which the catheter obstructs the ostium of a small vein, we have found it useful to allow blood to slowly drip from the end of the catheter or to use a gentle siphoning technique with tubing attached to the end of the catheter placed below the level of the patient. Using these techniques we have not found the need to routinely use coaxial microcatheters.
Sestamibi
Approximately 20 MBq 123I were injected intravenously followed by an injection of 600 MBq 99mTc-Ses-taMIBI 60 min later. A zoomed 20-min dual isotope dynamic acquisition was performed using a 256 x 256 matrix and asymmetric energy windows of 126–147 and 151–174 keV for 99mTc and 123I respectively. The dynamic images were corrected for motion when necessary. Dynamic image frames were then summed to create static images. The static images were then compared using a statistical technique whereby the 123I image was used as a template of the thyroid and any difference between corresponding pairs of pixels in the 99mTc-SestaMIBI and 123I images were assessed as to whether they constituted a significant difference in relation to the global regression between all the corresponding pixel pairs under consideration. Such differences were assigned p-values and displayed as images. This method is analogous to image subtraction but designed to be of a more objective nature was used to compare 123I and 99mTc-SestaMIBI static images. An additional static acquisition approximately 2-h post 99mTc-SestaMIBI injection was also carried out as a washout image in order to assess any areas of retained tracer activity.
Ultrasound
Ultrasound scans were performed using an Acuson Sequoia 512 ultrasound system (Siemens, Mountain View, CA, USA). All scans were performed with a 15 L8W probe with a 52 mm footprint. Scans were performed routinely both in the longitudinal and in the transverse planes, where a parathyroid mass was identified, it was measured using electronic calipers. The normal parathyroids, where possible, were also measured.
CT
Scans were performed using a GE Lightspeed Ultra 8-slice multidetector scanner or a GE High Speed ZX/i (General Electric, Milwaukee, WI, USA). Scans were performed at 2.5 mm space intervals, from the base of the skull to the carina, both before and after intravenous injection of contrast medium. Contrast medium of 100 ml was injected at 3 ml/s and the scans commenced at 40 s following the onset of contrast medium injection.
MRI
MRI was performed using a 1.5 T superconductive magnet (Signa, General Electric Medical Systems, 1996). Sequences included an axial T1-weighted from the base of the skull to the carina (TR 500–540, TE 8–9, slice width 4 mm, interslice gap 0–2 mm, number of acquisitions 2, matrix 256 x 256) and T2-weighted axial (TR 3333–7508, TE 80–90 EF, ETL 8, slice width 3–4 mm with an interslice gap of 0–1 mm, number of acquisitions 2, matrix 256 x 256–512 x 512). This was supplemented by fat-saturated T2-weighted sequences together with coronal Short Tau Inversion Recovery sequences. Intravenous contrast medium was not administered.
PTH assay
Our parathyroid hormone (PTH) assay is intact PTH (DPC IMMULITE 2500). It is specific for intact PTH 1–84 but has 40% cross-reactivity with fragment 7–84. Precision across the range is approximately 8%.
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
. Although PVS is more invasive than anatomical scanning, no complications of the procedure were observed.
|
A total of 27 patients underwent PVS. Subsequent surgical cure was defined as persisting normocalcaemia or requirement for 1
-hydroxycholecalciferol therapy (i.e. hypoparathyroid state) during a mean and a median of three patient years of follow-up. Thirteen patients were referred having undergone previous neck surgery; of these, nine were cured after repeat surgery subsequent to localising investigations, including PVS. In the 25 patients who underwent operation, 14 PVS (59%) were completely consistent with surgical and histological findings. However, in total 19 (76%) PVS were considered by the supervising physician to contribute valuable positive (n = 13) or negative (n = 6) information. Surgical cure was achieved in 22 out of 25 patients who underwent surgery giving a cure rate of 88%.
Using histological findings as the gold standard, PVS had a positive predictive value (PPV) of 81% (using a diagnostic threshold of two times the peripheral samples). The positive predictive values for ultrasound scanning, sestamibi scanning and MRI/CT were 73, 72 and 57% respectively. It should be recognised that as one of the criteria for performing PVS was equivocal or non-conclusive diagnostic studies, by definition the positive predictive value of these other studies would be expected to be lower than in an unselected population of patients with hyperparathyroidism. Results were further analysed to determine the optimum diagnostic gradient to use to determine a positive PVS. At ratios of 1.5, 2.5 and 3x peripheral sample, the PPV is 72%; for 2.5x is 75% and 3x is 83% respectively.
Eight patients (35%) had a diagnosis of MEN1 made prior to surgery and one patient had familial hyperparathyroidism. PVS was performed to determine the presence of ectopic parathyroid tissue and this was evident in five out of eight MEN1 patients who underwent surgery. The routine surgical approach in familial hyperparathyroid disease is total neck exploration and either total or 3.5 gland clearance. Without the information derived from PVS in these patients 63% would not have been cured with the first procedure as their mediastinal disease would not have been identified.
Eight PVS were negative, four because of technical failure to cannulate the thymic mediastinal vein. Therefore, four gave a false negative result. Three PVS gave a false positive result; patients 9, 11 and 20. Two of these patients (9 and 20) were not rendered normocalcaemic by surgery and hence it is possible that adenomatous tissue in these areas was not found at surgery. Patient 11 was falsely positive in the superior vena cava, which could indicate aberrant venous drainage from abnormal parathyroid tissue.
The magnitude of the ratio between the peripheral samples and the highest sample measured can predict the accuracy of PVS findings. If this ratio is >3, then all of our PVS findings were consistent with findings at surgery. A total of 53 patient years of follow-up has been achieved on those subjects who underwent operation.
Details of individual patients
These are described in Tables 1
and 2. In patient 1 PVS results were highly suggestive; serum parathyroid hormone in the left innominate vein sample was 102 pmol/l, the next highest sample was 28 pmol/l from the right atrium. Given the high likelihood of recurrence of parathyroid disease in patients with MEN1 (8) hypercalcaemia may recur.
|
Patient 3 was referred from another centre with the clinical problem of a mediastinal mass, a skull base lesion and primary hyperparathyroidism. The skull base mass was diagnosed as an invasive prolactinoma. PVS was positive in multiple areas. Initial neck exploration revealed a right-sided parathyroid mass with histological confirmation of parathyroid carcinoma; neck clearance was performed but the patient remained hypercalcaemic. Review of the catheter data demonstrated that the thymic mediastinal vein and superior vena cava (SVC) had a gradient of 1.5 over peripheral concentrations. Due to dilution effect from venous drainage these areas should normally demonstrate parathyroid hormone concentrations similar to periphery. Therefore, surgical exploration of the mediastinum was undertaken and revealed a retro-oesophageal parathyroid lesion with histological characteristics of carcinoma. This patient remains normocalcaemic on 1-hydroxycholecalciferol therapy 2 years later.
In patient 4 multiple gland involvement was suggested by both sestamibi and ultrasound scanning. PVS failed to catheterise the mediastinal vein. In the absence of imaging evidence of mediastinal disease the surgeon did not explore the mediastinum.
In patients 5, 6 and 8, the presence of multinodular goitre meant that all other scans were equivocal. This is a recognised limitation of sestamibi scanning (9).
In patient 7 sestamibi and neck ultrasound both showed left-sided disease. However, sestamibi suggested both upper and lower parathyroid disease. As this was the patient’s fourth neck operation (two for parathyroid disease; one removing an adenoma and one re-exploration negative; and one for papillary thyroid carcinoma) and the patient had previous neck radiotherapy, surgery was limited by necessity and all possible information was considered helpful. PVS also localised disease to the left side and two parathyroid adenomas were removed. The patient remains normocalcaemic on 1-hydroxycholecalciferol 8 years later.
Patient 9 had undergone two previous neck operations, at which normal parathyroid glands were removed, and remained hypercalcaemic. At PVS the left superior thyroid, middle thyroid, right low thyroid and the thymic mediastinal veins could not be cannulated. However, PVS was positive in the left inferior thyroid and the left innominate veins. Unfortunately surgical exploration was negative. The patient suffered a myocardial infarction perioperatively and was unfit for further surgery.
Patient 10 had MEN1 and had undergone non-curative surgery in 1994. Ultrasound suggested two left-sided parathyroid lesions; sestamibi scanning located one lesion. PVS also suggested two left-sided lesions and absence of mediastinal disease.
In patient 11 PVS ruled out the presence of mediastinal disease and therefore was contributory, but was categorised as unhelpful as it was only positive in the SVC and not consistent with surgical findings (a false positive PVS study).
Patient 12 required a second operation. Sestamibi scan prior to the first operation was negative for mediastinal disease. However, it demonstrated multiple gland involvement and therefore, by our criteria, this patient would have been a candidate for PVS. PVS performed prior to the second operation showed mediastinal and right-sided secretion of PTH. A second sestamibi at this stage also indicated mediastinal disease.
Patient 13 had neurofibromatosis, making cross-sectional imaging and surgical findings difficult to interpret and hence functional localisation was indicated. PVS was convincing of right-sided disease (Fig. 1) and in view of concordance with the sestamibi scan the surgeon did not explore the left side of the neck.
|
Patient 15 had a biochemical diagnosis of hyperparathyroidism and was symptomatic of renal calculi. Localisation of a single parathyroid lesion was evident on ultrasound. However, sestamibi scan was negative. PVS was indicative of a single parathyroid lesion and this was confirmed at surgery.
Patient 16 had MEN1 and four gland hyperplasia on histology. All scanning and PVS gave a negative result, but PVS excluded the presence of mediastinal disease and this patient remains normocalcaemic after 2 years.
Patient 17 had MEN1 and required a repeat operation (three previous neck operations). Ultrasound, sestamibi and MRI scan all suggested right-sided lower disease but PVS was negative. A hyperplastic gland was removed from the right side.
Patient 18 had undergone three previous operations. Sestamibi scan suggested two right-sided parathyroid lesions and these were confirmed at surgery. Ultrasound and PVS had indicated a single right-sided lesion.
Patient 19 has familial hyperparathyroidism and PVS failed to cannulate the mediastinal vein or show a gradient. Surgical findings demonstrate four gland hyperplasia.
Patient 20 had undergone one previous operation. Sestamibi scan and MRI suggested mediastinal disease and PVS was positive in the right atrium. At surgery a normal parathyroid gland was removed from the mediastinum; exploration of left side of the neck was negative. The patient declined further surgery.
Patient 21 had two previous adenomas removed. PVS was the only investigation that demonstrated mediastinal disease and the patient remains normocalcaemic on 1-hydroxycholecalciferol 4 years later.
Patient 22 had end stage renal failure and tertiary hyperparathyroidism. PVS was negative by our criteria but the left innominate vein sample (203 pmol/l) was clearly elevated over the next lowest sample (143 pmol/l) and this was consistent with surgical localisation of disease.
Patient 23 required reoperation and PVS was the only study that was negative; all other investigations pointed to the left-sided adenoma that was removed at surgery.
PVS in patient 24 was performed because of equivocal ultrasound and sestamibi scans and negative MRI. PVS and sestamibi both indicate mediastinal disease, and surgery is planned.
Patient 25 had a right lower gland adenoma localised by PVS.
In patient 26, analysis of PVS using a gradient of 2x the peripheral parathyroid hormone measurements demonstrated involvement of both the right upper and lower parathyroids. When a gradient of 1.5x peripheral measurements was used, the PVS was the only investigation to conclusively demonstrate four gland disease.
In patient 27 sestamibi and PVS supported surgical findings.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
Repeat operations
Patients who have persisting hypercalcaemia post-operatively have a much higher likelihood of having multiple adenomas, hyperplasia or carcinoma (13). Reasons for persisting hypercalcaemia post-operatively include missed ectopic disease (3) and either no or inaccurate preoperative imaging. Traditional imaging has been shown to have limitations when overactive parathyroid glands are small, there is multiple parathyroid gland involvement or when thyroid disease is coexistent (14). Ultrasound is operator-dependent and is limited in detecting tissue behind bone. Previously, MRI has been the investigation of choice for mediastinal disease (14). However, the information from MRI is limited to anatomical considerations. Undescended adenomas have been reported as reasons for persisting hypercalcaemia post-operatively (5); the incidence of such lesions is reported to be 7% and none were found in our series. However, PVS found 80% of these adenomas in this single study (5). Clearly, it is desirable that the first operation is curative in the context of both patient safety and longer term health, and PVS offers further opportunity for diagnostic information.
Familial hyperparathyroidism
Familial hyperparathyroid syndromes include MEN1 and 2A, hyperparathyroidism-jaw tumour syndrome and familial isolated hyperparathyroidism. The diagnosis of any of these syndromes will alter patient management.
Approximately 20% of patients with multiple gland involvement will have a hereditary syndrome, such as MEN1 (10). Thirty percent of MEN1 patients will have an ectopic parathyroid gland (8); in our study this has reached 60%. Full neck exploration is necessary in these patients; mediastinal exploration would only be undertaken, should preoperative investigations suggest mediastinal disease and preoperative PVS is an effective way of determining the presence of such disease. Other reviews have called for MEN1 screening only in patients who have a family history or are symptomatic from other endocrinopathies (15). However, given that 30% of our patients had MEN1 we believe it is important to screen patients who present with multiple parathyroid gland involvement for MEN1. PVS has been previously described as initial work up for MEN1 patients by the NIH group (16). Bilateral disease but no mediastinal disease was found in four out of five subjects investigated using PVS 63% of our subjects investigated with PVS had mediastinal disease.
A prospective review of preoperative localisation (17) demonstrated a sensitivity of 95% for sestamibi scanning in cases of single adenoma but sensitivity decreased to 25% in multiple gland disease. A further retrospective review described a sensitivity of 62% using sestamibi in patients with multiple gland disease (18). Intraoperative PTH monitoring may be misleading in over 50% of patients with multiple gland involvement (17). The presence of mediastinal disease is an important clinical question and investigations other than PVS are not always reliable in this area.
PVS
In our series, thirteen of the PVS studies yielded information not available from any of the other modalities of localisation. None of our patients experienced any complications from the PVS study. Previously PVS has only been used in the setting of persistent hyperparathyroidism. However, when there is suggestion of multiple gland disease it is appropriate to perform this procedure preoperatively to increase the likelihood of achieving cure by a single operation.
|
Conclusion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
|
Acknowledgements |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
