F18-choline PET/CT guided surgery in primary hyperparathyroidism when ultrasound and MIBI SPECT/CT are negative or inconclusive: the APACH1 study

Patients with a biologically proven PHPT (hypercalcemia and elevated or inappropriately normal parathyroid hormone (PTH) levels) and negative or inconclusive ultrasound and MIBI SPECT/CT results prior to surgery were prospectively included. Exclusion criteria were severe kidney failure (creatinine clearance <30 ml/min), profound Vitamin D deficiency or the Multiple Endocrine Neoplasia-1 (MEN-1) syndrome. The local ethics committee approved the study protocol (Ref. 2014–41, Comité de protection des personnes Nord-Ouest III). All patients gave written informed consent. This trial is registered as EUDRACT 2014–003852-30, Clinical trial NCT02432599.

All patients underwent cervical ultrasound and MIBI SPECT/CT prior to inclusion. Ultrasound was performed by local or external radiologists. Results were classified as negative or inconclusive according to the written report. The presence or absence of concomitant thyroid nodules was recorded.

MIBI SPECT/CT was performed identically in the two participating centers on SymbiaT2 systems (Siemens Medical Solutions). After intravenous injection of 740 MBq of MIBI, an early pinhole acquisition of the anterior lower neck was performed 10 min post-injection, followed by a late SPECT/CT acquisition of the neck and upper chest 90 min post-injection. A detailed description of the MIBI SPECT/CT protocol can be found elsewhere [6].

Images were interpreted by an experienced nuclear medicine physician (EQ, NA). A negative MIBI SPECT/CT was defined as the absence of focal uptake on the early and delayed images. An inconclusive MIBI SPECT/CT result was defined as faint uptake compared to the surrounding background without CT substrate or uptake most likely related to a thyroid nodule.

Upon inclusion, serum values of calcium, parathyroid hormone (PTH), albumin, phosphorus, vitamin D and creatinine were measured for all patients, and the creatinine clearance was calculated according to the MDRD formula [16].

Sixty minutes after intravenous injection of 1,5 MBq/kg of FCH, a low-dose CT was performed (CAREdose ref. mAs 100, 130 kV, slice 3 mm, pitch 1.0), followed by a one bed position PET acquisition of 10 min covering the neck and upper chest in 3D list-mode on a Biograph 6 TrueV PET/CT system (Siemens Medical Solutions). The injected activity and the exact delay between injection and the start of the acquisition were recorded. The effective dose due to the low-dose CT was calculated by multiplying the dose-length product with the conversion factor 0.0059 mSv/mGy/cm [17], and the effective dose due to the FCH administration by multiplying the injected dose in MBq by the conversion factor 0.019 mSv/MBq [18]. The raw PET data were iteratively reconstructed with 3 iterations, 21 subsets, point-spread-function (PSF) modeling (TrueX), matrix size 256*256 and zoom 1.0. The list mode data were reconstructed with an increment of 2 min (2 min, 4 min, 6 min, 8 min and 10 min). Scatter and attenuation corrections were applied. No post-reconstruction filter was used. Image analysis was performed on Leonardo workstations (Siemens Medical Solutions). Only the 10 min reconstruction was used for the clinical FCH PET/CT report.

The FCH PET/CT was considered positive in the case of clear focal uptake(s) in a predisposing area. The exact location of each focus was noted (the side and upper or lower position, or the ectopic position), and when measurable the maximum transverse CT diameter. The FCH PET/CT was considered inconclusive in the presence of faint focal uptake superior to the surrounding background without CT substrate and negative in the absence of focal uptake. A second blind read of all exams was performed 9 months after the last inclusion by the same readers in order to estimate the inter observer agreement.

The semi-quantitative analysis was performed with MIM-software (version 5.6, MIM Software Inc., Cleveland, OH). The PTA SUVmax measurement was performed by placing a VOI on the parathyroid uptake with the PET Edge tool. Background SUVmean was measured by placing a spherical VOI with a 1 cm diameter on the contralateral thyroid lobe, if present, and the contralateral sternocleidomastoid (SCM) muscle. The PTA-to-thyroid and PTA-to-SCM ratios were calculated for all (2 min, 4 min, 6 min, 8 min and 10 min) attenuation corrected reconstructions.

All patients underwent surgery within four weeks following the FCH PET/CT by a dedicated head and neck or endocrine surgeon in one of the participating centers. The surgeon had access to the clinical FCH PET/CT report. In case of a positive FCH PET/CT, an outpatient MIP was performed. The surgical procedure was adapted in the case of suspected multiple or ectopic PTAs. In case of an inconclusive FCH PET/CT, surgery was performed on the site of the dubious focus. In case of a negative FCH PET/CT, a conventional inpatient BCE was performed.

The exact location was noted for each resected specimen, as were the total surgery time and surgical complications if any.

During surgery, an intra-operative frozen section was performed to confirm the presence of parathyroid tissue. Final analysis was performed on paraffin-wax embedded sections stained with hematoxylin and eosin. When necessary, immunohistochemistry with anti-PTH antibody was performed. Parathyroid adenoma and parathyroid hyperplasia were considered true positive.

In the days following surgery, the serum calcium level was repeated. Patients were considered cured in case of histological proof of PTA or parathyroid hyperplasia and a normalization of the serum calcium level after surgery.

The planned sample size for the present single-stage design was based on the estimated sensitivity of FCH PET/CT for PTA detection [12] that should be superior to 0.60 to be sufficient. With unilateral alpha of 0.10, an anticipating sensitivity of 0.90 and a power = 0.80 we determined that 20 patients should be included to detect 12 PTAs. The sample size was increased to 24 to correct for drop-out. Quantitative variables were described with mean and standard deviations, whereas qualitative variables were described with numbers and percentages. The sensitivity and positive predictive value were calculated on a lesion and patient level respectively. The Wilcoxon rank-sum test was used for continuous variables. The kappa statistic according to Fleiss-Cuzick was used to determine the inter-observer agreement, with 95% confidence intervals using an inverted modified Wald test approach, as recommended by Zou and Donner [19]. Kappa values were interpreted as follows: <0 poor agreement, 0.0–0.20 slight agreement, 0.21–0.40 fair agreement, 0.41–0.60 moderate agreement, 0.61–0.80 substantial agreement, 0.81–1.00 almost perfect agreement. The nonparametric Friedman test was used to compare the ratios between SUV_max of the adenomas to the SUV_mean in background (either muscle or thyroid) amongst the different reconstructions. A post hoc test was performed with the Dunn test for multiple comparisons. For all tests, a two-tailed P value of 0.05 or less was considered statistically significant.

Analyses were performed with STATA, version 12 software (Stata Corp, College Station, TX) and Prism (GraphPad Software, La Jolla, CA).

From March 2015 through February 2017, 25 patients were included and 24 patients completed the study protocol. The patient characteristics, blood test results and ultrasound and MIBI SPECT/CT results can be found in Table 1.

For the FCH PET/CT acquisition, the injected FCH activity was 113 ± 33 MBq, and the delay between injection and acquisition was 63 ± 8 min. Effective doses due to the FCH injection and the low-dose CT were 2.67 ± 2.50 mSv and 1.14 ± 0.35 mSv, respectively.

FCH PET/CT showed 21 positive foci and 3 inconclusive foci in 22 patients, with a maximum transverse CT diameter of 13.1 ± 8.6 mm. The FCH PET/CT results and the flow of patients through the study are depicted in Fig. 1. The inter observer agreement was considered almost perfect when the inconclusive results were classified either as positive (κ = 0.92 (95%CI: 0.73–0.99)) or negative (κ = 0.88 (95%CI: 0.68–0.97)), and considered substantial when all three groups were taken into account (κ = 0.79 (95%CI: 0.63–0.90)).

SUV_max of PTAs on the 10 min reconstruction was 6.73 ± 3.27. SUV_mean in the thyroid background and in surrounding muscle background were 2.43 ± 0.63 and 1.36 ± 0.49, respectively. No statistical significance was observed between PTA/muscle ratios for the different reconstructions. With regards to the PTA/thyroid ratio, a trend towards higher ratio for the 2 min reconstruction (3.51 ± 2.18) versus 8 min (3.08 ± 2.03) and 10 min (3.09 ± 2.15) reconstructions was observed, but statistical significance was not reached. The results of the semi-quantitative analysis of the list-mode PET data are shown in Fig. 2.

Surgery was performed 50 ± 49 days after the FCH PET/CT. One patient did not undergo surgery. The FCH PET/CT result guided surgery in 22 patients (88%). The different surgical procedures and the adverse events can be found in Fig. 1. The mean duration of the surgical procedure was 44 ± 24 min for the MIP and 100 ± 45 min for the other surgical procedures (p = 0.0096).

The results of the histopathological analysis of the resected specimens can be found in Fig. 1.

When dichotomizing the FCH PET/CT results, thereby classifying the inconclusive FCH PET/CT results as positive, 21 foci were considered true positive (TP), 0 true negative (TN), 3 false positive (FP) and 2 false negative (FN), resulting in per lesion and per patient sensitivities of 91.3% (95%CI: 72.0–98.9) and 90.5% (95%CI: 69.6–98.8), respectively, of FCH PET/CT for PTA detection. The corresponding per lesion and per patient positive predictive values were 87.5% (95%CI: 67.6–97.3) and 86.4% (95%CI: 65.1–97.1), respectively.

In the FCH PET/CT positive group, the FP result occurred in a patient with a previous history of thyroidectomy. The FCH-PET/CT showed a single PTA in the right superior position that was not found at surgery. PHPT and a right superior FCH positive focus persisted after surgery.

The FN results occurred in two patients with negative FCH PET/CT and a 15 mm left superior and 16 mm right superior PTA at BCE.

Three patients had an inconclusive FCH PET/CT result, two with faint left inferior uptake, corresponding to one normal parathyroid and one PTA, and one with faint right inferior uptake corresponding to a lymph node.

All site classifications were concordant, except for one patient in whom the FCH PET/CT position was classified as right inferior, whereas the surgical classification was right superior.

Four patients had a previous history of thyroid or parathyroid surgery. One of the two patients with previous thyroid surgery corresponded to the patient with the FP result described above, and the other patient had a previous history of a right hemithyroidectomy for vesicular adenoma. FCH PET/CT showed a TP right inferior PTA, which was removed by MIP, leading to cure. One of the two patients with previous parathyroid surgery was a patient with familial hyperparathyroidism (HRPT2 gene mutation) and a previous upper left PTA resected by MIP, with persistence of PHPT after surgery. FCH PET/CT showed a TP left inferior PTA of 9 mm, which was resected by MIP leading to cure. The other patient had a previous history of parathyroidectomy by BCE. Conventional imaging and FCH PET/CT results were all negative, and surgery was not performed.

Overall, 21 (88%) of the patients were considered cured after surgery. Two cases are illustrated in Figs. 3 and 4. The serum calcium level after surgery in the patients considered cured was 2.36 ± 0.17 mmol/l, versus 2.72 ± 0.06 mmol/l in the patients with persistent PHPT. BCE could be avoided in 18 (75%) patients (17 MIPs and 1 ectopic parathyroidectomy).