Localization of parathyroid adenomas using ¹¹C-methionine pet after prior inconclusive imaging

Primary hyperparathyroidism (pHPT) is a common endocrine disorder, with the highest incidence in elderly women [1]. It occurs sporadically, but is also associated with hereditary syndromes such as multiple endocrine neoplasia (MEN) type 1 and 2. pHPT is characterized by hypercalcemia in the presence of high concentrations of PTH, which can lead to abdominal complaints, osteoporosis, kidney stones, muscle weakness, pain, depression and behavioral changes.

Surgery is the only curative and recommended treatment in patients with pHPT usually by means of a minimally invasive parathyroidectomy (MIP). In MIP, surgeons remove the adenoma via a unilateral approach with a minimal invasive incision of 1–2 cm. In 80 to 90% of the pHPT cases, only a single parathyroid adenoma is present, making this surgical strategy usually successful [2]. However, to be able to perform a unilateral MIP, accurate preoperative imaging is essential.

Worldwide, the current primary preoperative localization imaging standard consists of cervical ultrasonography (cUS) combined with ^99mTc-methoxyisobutylisonitrile single-photon emission computed tomography/computed tomography (MIBI-SPECT/CT) [3, 4]. Planar MIBI scintigraphy has the lowest sensitivity, around 70% [5, 6], performing better when combined with SPECT/CT [5].

However, although better, even MIBI-SPECT/CT alone is not optimal with a sensitivity of 85% [5, 7‐13]. But when the MIBI-SPECT/CT is combined with cUS (sensitivity from 22 to 82%) [2, 9‐11] a sensitivity of 80–90% can be achieved [14‐16]. This means that using these two modalities still in 10–20% of the cases, the surgeon will not be able to schedule the patient for a focused MIP operation.

¹¹C-methionine positron emission tomography/CT (¹¹C-MET PET/CT) is a nuclear imaging technique that can be used as a next step for imaging after prior negative localization. C-methionine accumulates in the parathyroid adenoma and is involved in the synthesis of the precursor of PTH [17‐19]. It improves the detection performance of parathyroid tissue due to a better spatial resolution of PET/CT [20]. When ¹¹C-MET PET/CT became available in 2006 at our institute, we used this nuclear imaging modality as a step up approach after inconclusive imaging.

The aim of this study was to determine the diagnostic performance of ¹¹C-MET PET/CT after prior negative localization in patients with pHPT.

This is a retrospective single center cohort study of patients with biochemically proven pHPT who underwent parathyroid surgery after localization by means of a ¹¹C-MET PET/CT in a teaching and tertiary referral hospital.

The medical charts of all patients who underwent ¹¹C-MET PET/CT between January 2006 and December 2014 were reviewed. To be eligible for inclusion, patients had to be older than 18 years and had to have a biochemically confirmed pHPT, for which parathyroid surgery was planned. A MIBI-SPECT/CT and/or cUS had to have been performed, however with negative or inconclusive result, after which patients underwent a ¹¹C-MET PET/CT. Patients were excluded if they were known to have a germline mutation predisposing for multiple gland disease or if an alternative diagnosis (e.g., parathyroid carcinoma) was known before surgery.

The medical charts were reviewed to determine the outcome of the imaging tests (negative = no localization stated in the original report, inconclusive = an original report describing a presumed adenoma with doubt), or positive = original reports describing the location of the presumed adenoma without any doubt). Also, data on gender, age, preoperative PTH, corrected calcium, intraoperative quick PTH (IOPTH), previous parathyroid surgery, the length of surgery and pathology outcome were collected. Corrected calcium was calculated using the following formula: Ca + ((40 – Alb) × 0.02). “Ca” is the serum calcium (mmol/l) and “Alb” is the serum albumin (g/l). The diagnosis pHPT was made by experienced endocrinologists from our center and all the patients were discussed in a multidisciplinary endocrine board.

Data obtained from patient records were anonymously stored using study-specific patient codes in a password protected database. The local ethical board evaluated the study and according to Dutch law, no additional review board approval was required.

cUS was performed in a number of different hospitals on various ultrasound systems. All patients who underwent cUS were examined in a supine position with a hyperextended neck using a high-frequency linear transducer, as is common practice. The neck was always examined from the level above the thyroid to the clavicle caudally. Findings suggestive for parathyroid adenomas were documented in two planes with special regard to size and anatomic correlation to adjacent structures.

Between 2006 and 2014, parathyroid imaging was performed with various protocols due to changes in gamma cameras and radiotracers. Also, some procedures were performed in other hospitals according to slightly different imaging protocols. However, all protocols adhered to the international guidelines [21]. At the University Medical Center Groningen (UMCG) until 2010, images were performed using a Multispect 2 gamma camera (Siemens), on which only SPECT images could be made. Afterwards, patients were scanned on a Symbia T16 gamma camera with CT (Siemens), resulting in SPECT/CT images. Furthermore, always ^99mTechnetium (^99mTc)-sestamibi (MIBI) was used for preoperative localization as “dual phase” technique. This technique was always combined with a “dual tracer” subtraction technique for thyroid only visualization, although a switch in tracer was made in 2012, and ¹²³I was replaced by ^99mTc-pertechnetate. Thus, currently the MIBI-SPECT/CT imaging protocol includes early and late planar MIBI images combined with ^99mTc-pertechnetate planar subtraction images, and late MIBI SPECT/CT 3D images.

In the current study, two PET cameras were used. Patients were either scanned on a Ecat EXACT HR + PET only system or a PET/CT (Biograph mCT, 64 slice CT) camera (in use since October 2009) (both Siemens) and had to fast for 6 h while drinking 1 l of water prior to the PET procedure. PET images were taken 20 min after injection of 7 MBq/kg ¹¹C–methionine which was produced on site as described by Phan et al. [22]. The head and neck area was scanned in two bed positions (2D imaging, 7 min per bed position and 2 min transmission scan) using the HR + camera, while it was scanned in three bed positions using the mCT camera (3 min per bed position, prior low dose CT for attenuation correction; 100 kV, 30 Quel Ref mAs, 1.5 pitch). PET images were iteratively reconstructed using four iterations, 16 subsets with a 5 mm Gausian filter for HR+, and using three iterations, 21 subsets with 4 mm Gausian filter for mCT. Low dose CT images were reconstructed with 2 mm slice thickness.

Data were analyzed using descriptive statistics on a patient based level. Mean (± SD) or median with range were calculated when appropriate. Differences between duration in surgery in different groups were calculated using a Mann-Whitney U test. Categorical variables were expressed in proportions. SPSS version 22 statistical software was used. A p value of <0.05 was considered significant.

This retrospective single center cohort study evaluates the diagnostic performance of ¹¹C-MET PET/CT after prior non-conclusive localization via MIBI-SPECT/CT and/or cUS in patients operated for biochemically confirmed pHPT. In this challenging clinical setting, ¹¹C-MET PET/CT was able to adequately localize the adenoma correctly in nearly two-third of the patients. Thus, in these patients the surgeon was subsequently able to perform a successful focused approach.

We found that ¹¹C-MET PET/CT in this series localized the adenoma in 64% at the correct side of the neck. In 3 patients, the ¹¹C-MET PET/CT was false positive, and the surgeon could not locate an adenoma. Reasons for false positivity might be benign and malignant thyroid lesions, as was the case in one of our patients [22].

Although few studies describe the diagnostic performance of ¹¹C-MET PET/CT, our sensitivity of 72% is comparable with earlier studies. Beggs et al. who included 51 patients, found a sensitivity of 83% [23]. In Beggs et al., also patients in whom other imaging techniques such as MIBI scanning, CT, or ultrasound had earlier failed to localize the adenoma were selected. In a recent smaller study with 18 patients, Braeuning et al. found a sensitivity per patient of 91.7% and a sensitivity per lesion of 73.3% after a prior negative ^99mTc-MIBI-SPECT/CT [24]. Another smaller study by Chun et al. who included 16 patients found a sensitivity of 91.7%, but did not select patients with prior negative imaging [20]. Traub-Weidinger et al. included 15 patients with negative ^99mTc-MIBI SPECT/CT and earlier neck surgery because of pHPT and/or thyroid disorder and found a sensitivity of 40% [25]. With a sensitivity of 40% for ¹¹C-MET PET/CT, these results differ from other studies. The varying surgical history of the patients in this series may be an explanation [25]. Martinez-Rodriguez et al. included 14 patients in a prospective study and showed a sensitivity of 76.9%, but only 2 patients had previous neck surgery [26]. Additionally, no patients with prior negative imaging were selected. So, with 28 included patients (10 with previous neck surgery) this is one of the largest studies reflecting the performance of ¹¹C-MET PET/CT in patients with pHPT after prior negative imaging.

We realize that our retrospective study has limitations. In this study, only patients were included who had undergone ¹¹C-MET PET/CT and were subsequently surgically treated. Correspondingly, of the 15 patients not referred for surgery, ¹¹C-MET PET/CT was negative in 11 patients. This referral bias may have led to overestimation of the sensitivity of the ¹¹C-MET PET/CT. However, our study group also included patients with negative results of ¹¹C-MET PET/CT that underwent surgery. Also, performance of MIBI-SPECT/CT highly depends on the protocol used. If all patients would have received a dual phase MIBI-SPECT/CT, more adenomas could have been localized and in the remaining patients, adenomas might have been harder to localize. This may also have led to some overestimation of the performance of ¹¹C-MET PET/CT.

In this study, the results of surgery were seen as the gold standard, which could be a drawback of the current analysis. Although a surgical neck exploration is technically protocolled, heterogeneity in surgical techniques remains inevitable. Also, in this study imaging procedures changed slightly during the years, which might have influenced our results.

The strong point of our study is that it describes and reflects the strategy of preoperative ¹¹C-MET PET/CT after earlier negative imaging. The success of a focused approach strongly depends on how well the adenoma is localized by imaging preoperatively.

In this series duration of surgery is relatively long, since anesthesiology and waiting time for IOPTH was included and the data were gathered in a tertiary referral hospital which adds to more extended surgery times because of a patient selection for more difficult cases. Furthermore, we conduct a training program for residents and therefore also provide surgical training during operations which required time because of learning curve. Most importantly however, if ¹¹C-MET PET/CT correctly located the adenoma, the duration of surgery was significantly shorter.

In general, preoperative localization saves operation time because of a double effect. Firstly, operating via keyhole surgery is a shorter and more effective procedure compared to an open bilateral exploration and secondly the surgeon knows where the adenoma should be located. Furthermore, if the surgeon finds the adenoma where it was expected to be, often less perioperative adjuncts such as IOPTH are needed, also shortening the procedure. Finally, when the patient is informed about the small risk of having a 5% chance of a second adenoma, a second contralateral procedure in the future will still be in a surgical virgin territory and therefore is an acceptable calculated risk. We speculate that eventually in the long run, this might lead to a decrease in healthcare costs. More research on this topic is mandatory.

Globally, cUS and MIBI-SPECT/CT are used first in line to localize parathyroid adenomas being relatively cheap and widely available imaging techniques [3, 4]. Replacing cUS and MIBI-SPECT/CT with ¹¹C-MET PET/CT is not a realistic option. ¹¹C-MET PET/CT has a more complicated radiochemistry tracer production, includes the need for a cyclotron on site, and has significantly higher costs. When available, it is therefore more realistic to use ¹¹C-MET PET/CT as a step up approach only after earlier negative localization by MIBI-SPECT/CT and/or cUS.

Other various imaging protocols in the detection of a parathyroid adenoma are also feasible. For example, four-dimensional computed tomography (4DCT) is another reported strategy as a next step imaging procedure, because of its high sensitivity (88%) [27, 28]. Choline PET, known from its use in prostate cancer diagnostics, can either be performed with ¹¹C–choline or ¹⁸F–choline and may be a good imaging alternative for ¹¹C-MET PET/CT. First results in literature look promising [9, 29, 30], however, results comparing the two tracers directly are not available. Also, in this area more research is warranted. Furthermore, a cost-effectiveness analysis on the various anatomical and nuclear imaging techniques and the order in which to use them in the setting of pHPT could also be performed.

In conclusion, this retrospective analysis shows ¹¹C-METPET/CT to be a sensitive method for the localization of parathyroid adenomas in patients with clinically suspected pHPT (sensitivity of 72%). ¹¹C-MET PET/CTcorrectly detects parathyroid adenomas in 64% of operated patients after prior negative MIBI-SPECT/CT and/or cUS. Since ¹¹C-MET PET/CT is able to additionally localize adenomas, resulting in more patients who can be operated via MIP, the duration of surgery and thus healthcare costs potentially decrease.