Detection of parathyroid adenomas with 4DCT: Towards a true four-dimensional technique


 Background. 4DCT is a commonly performed examination in the management of primary hyperparathyroidism, combining three-dimensional imaging with enhancement over time as the fourth dimension. We propose a novel technique consisting of 16 different contrast phases, instead of three or four different phases. The main aim of this study was to see if this protocol allows for the detection of parathyroid adenomas within dose limits. Our secondary aim was examining the enhancement of parathyroid lesions over time.Methods. For this prospective study, we included 15 patients with primary hyperparathyroidism prior to surgery. We obtain a 4DCT with 16 different phases: an unenhanced phase followed by 11 consecutive arterial phases and 4 venous phases. Centered on the thyroid, continuous axial scanning is performed over a fixed 8cm or 16cm coverage volume after start of contrast administration.Results. In all patients an enlarged parathyroid can be demonstrated, mean lesion size is 13.6mm. Mean peak arterial peak enhancement for parathyroid lesions is 384 HU compared to 333 HU for the normal thyroid. No statistical difference could be found. Time to peak (TTP) is significantly earlier for parathyroid adenomas compared to normal thyroid tissue: 30.8s versus 32.3s (p value 0.008). Mean Slope of Increase (MSI) of the enhancement curve is significantly steeper compared to normal thyroid tissue: 29.8% versus 22.2% (p value 0.012). Mean dose length product was 890.7 mGy.cm with a calculated effective dose of 6.7 mSv.Conclusion. We propose a feasible 4DCT scanning-protocol for the detection of parathyroid adenomas. We manage to obtain a multitude of phases, allowing for a dynamic evaluation within an acceptable exposure range when compared to classic helical 4DCT. Our 4DCT protocol may allow for a better visualization of the pattern of enhancement of parathyroid lesions, as enhancement over time curves can be drawn. This way wash-in and wash-out of contrast in suspected lesions can be readily demonstrated. Motion artifacts are less problematic as multiple phases are available.

. Taking duplicates into account, we identi ed 38 unique scanning protocols.
Timing of the different scanning phases in the literature varies: with the exception of four studies, all cited authors choose to obtain an arterial phase. For 25 different studies the arterial images are obtained 25 seconds after contrast-administration, in 13 studies authors propose to scan after 30 seconds. Five other authors come to an obfuscate consensus, stating that arterial images are obtained by scanning after 25-30 seconds. 42 authors then do report to perform arterial scanning after 25 up to 30 seconds, identical protocols included.
The main aim of this study was to see if we could create a 4DCT protocol allowing for the detection of parathyroid adenomas using multiple added contrast phases and this within dose limits. Our secondary aim was to examine the enhancement of parathyroid lesions over time, determining the maximum peak enhancement of parathyroid lesions.

Methods
For this prospective study, we included 15 patients with primary hyperparathyroidism (i.e., an elevated serum level of calcium and raised levels of parathyroid hormone) and a positive ultrasound prior to surgery. This study was approved by the medical ethics committee of our hospital. All patients were informed about the nature of the procedure as well as the risks involved (radiation, administration of iodinated contrast). All patients signed an approved informed consent form prior to the examination.
Patients under legal age (18y) were excluded. Patients who had undergone prior surgery of the thyroid or parathyroid were also excluded. All patients had undergone prior US examination, as this is the standard of care for the detection of parathyroid adenomas in our hospital.
Scanning is performed on a Revolution CT, GE Healthcare. Patients receive a venous catheter placed in a cubital vein that is checked for patency. Arms are placed in a neutral position alongside the body. The patient's head is xated in a head cradle. The patient is explained the sensation of contrast administration and instructed not to move and avoid swallowing. Centered on the thyroid, continuous axial scanning is performed over a xed 8 cm coverage volume (100 kVp, SmartmA 10-480 mA, thickness 0.625 mm, 0.5s rotation scanning time). Wide beam axial scanning was chosen over helical scanning in order to limit the dose (65). For the last three patients we widened the scanning range from an 8cm volume to a 16cm volume, covering a larger view of the upper mediastinum.
First, we obtain a non-enhanced scan (NECT). At the same time contrast administration is initiated: a bolus of 90 mL Xenetix 350 mg I/mL is injected at 6 mL/s followed by a 50mL saline ush (6mL/s). After a delay of 20 seconds 11 subsequent phases with a 2-second interphase delay are obtained. These are the arterial phases. With a 10-second interphase delay 4 more phases are obtained. These are the venous phases. A schematic overview of this protocol is provided in Figure 1.
Images were reviewed on a GE workstation. The different scan phases were deformably registered after which the registered image data can be analyzed on a voxel-by-voxel basis, thereby retaining spatial information for the analysis. The data were analyzed and interpreted by two senior members of staff with experience in perfusion imaging.
The following parameters were considered in the normal thyroid, parathyroid and lymph nodes. Maximum peak enhancement (HUmax) is de ned as the maximum concentration of contrast agent over time, measured in the region of interest. It is expressed in Houns eld units (HU). Time to peak (TTP) is de ned as the time at which HUmax is reached. It is expressed in seconds (s). Mean slope of increase (MSI) corresponds to the steepness of the enhancement curve of a given region of interest. It is expressed in percent (%). An overview of these different values is provided in Table 2.
Statistical analyses were performed using SPSS software ver. 23.0 (IBM, Armonk, NY, USA). For this small dataset we assumed the absence of normality and symmetry. The Wilcoxon Signed Ranks test was used to evaluate differences between values. Statistical signi cance was set at p < 0.050.
For the evaluation of the radiation dose, we considered the dose length product (DLP) and the effective dose (E). The DLP is a technical dose descriptor expressed in mGy.cm that represents the total radiation output of the scan. The effective dose, expressed in mSv, represents the total body dose and was calculated following current ICRP-103 guidelines (66) by using a CT patient dosimetry software (CT-Expo v1.7.1, G Stamm and H Nagel, Hann-over) that considers all technical acquisition parameters of the individual scans, including the scan range. A conversion factor of 0.0075 was used.

Results
The mean patient age of patients included in our study was 61 years. 6 patients were male, 9 patients were female. In all patients, a single enlarged parathyroid could be detected, coinciding with ultrasound ndings and surgical localization. The mean lesion size was 13.6mm. Mean peak arterial peak enhancement for parathyroid lesions was 384 HU compared to 333 HU for the normal thyroid. No statistical difference could be found between these values.
The different phases were reviewed side-by-side, as demonstrated in Figure 2. Here we suspected a parathyroid adenoma posterior from the left thyroid lobe in close relation to the common carotid artery. The different phases can also be reviewed in absolute relation to their timing. This allows us to plot the various phases on a curve depicting enhancement over time, as demonstrated in Figure 3. In this example arterial wash-in and venous wash-out of a suspected parathyroid adenoma can be easily demonstrated. Motion artifacts due to swallowing are present on a single phase in the example shown, affecting the curve of the small parathyroid lesion (34s after administration of contrast). 11 out of 15 detected adenomas demonstrate a higher peak enhancement in the arterial phase compared to normal thyroid tissue. This is called the arterial wash-in effect. 4 out of 15 adenomas have a slightly lower peak enhancement in the arterial phase compared to normal thyroid tissue. These 4 adenomas do show a lower enhancement compared to the thyroid on the later venous phases. This is the venous wash-out effect. The enhancement pattern of lymph nodes differs from parathyroid and thyroid tissue with lower maximum enhancement (mean HUmax 102 HU). These lesions do not demonstrate a signi cant wash-in or wash-out of contrast.
Mean dose length product was 890.7 mGy.cm with a calculated effective dose of 6.7 mSv. Mean dose length product for the scans with an 8 cm volume was 728.3 mGy.cm with a calculated effective dose of 5.5 mSv. Mean dose length product for the three scans with a 16 cm volume was 1540.2 mGy.cm with a calculated effective dose of 11.6 mSv.

Discussion
In all patients, a single enlarged parathyroid could be detected, coinciding with ultrasound ndings and surgical localization. By drawing enhancement over time curves, we can easily evaluate the pattern of enhancement of a region of interest over time. This way wash-in of contrast and wash-out of contrast can be readily demonstrated. Motion artifacts (due to swallowing for example) can affect the measurement in the ROI of small lesions, as we demonstrated in the example (Figure 3). If this effect would have been present on a single arterial phase, this might have risked missing the lesion as the difference in enhancement between the lesion and the normal thyroid did become smaller. Since we do possess other phases for this case, this problem can then at least partially be overcome.
We found a statistically signi cant difference in TTP and MSI between parathyroid and thyroid tissue, with parathyroid adenomas showing a steeper curve and enhancing on average 1.5 seconds earlier. Since the interval for TTP for parathyroid adenomas in our study varies between 26 and 42 seconds after contrast-administration, arterial enhancement could be less conspicuous with a single arterial phase 25s after contrast administration as this is the most common protocol found in the literature.
Cervical lymph nodes are a known mimicker of parathyroid adenomas. Based on their slow and continuous enhancement, these structures can be readily differentiated from parathyroid adenomas. Another know mimicker is caused by ectopic thyroid tissue. It can be suggested that this ectopic thyroid tissue would behave in a similar way to normal thyroid tissue and thus differ from parathyroid adenomas. As we did not come across a relevant case, this remains speculation.
The effective dose associated with our proposed 4DCT protocol has a mean value of 6.7 mSv and can be as low as 1.4 mSv. The effective dose of 4DCT protocols in the literature is situated between 10.4 mSv and 13.8 mSv. The effective dose of scintigraphy is estimated at 7.8 mSv (99mTc-Sestamibi-SPECT) or 18.4 mSv (hybrid Sestamibi-SPECT) (39,58). The last three patients were scanned with a 16cm volume. The calculated effective dose for these three cases was calculated at 11.5 mSv, still within dose limits for traditional helical 4DCT as referenced by the literature (58). We then managed to de ne parameters for a dynamic 4DCT examination within an acceptable exposure range.
This study has several limitations. First there is the small sample size of the study. It is however di cult to include patients when an existing and approved standard of care technique like ultrasound produces excellent results. Our study is also prone to selection bias since we only included patients with positive ultrasound ndings that were eligible for surgery. Lesions in patients, when surgery is not considered, may be smaller and thus more di cult to detect. Comparing the suspected parathyroid lesion to thyroid tissue may also prove problematic in cases of thyroid disease like thyroiditis with associated diffuse in ammation of the thyroid gland. Lastly, we admit our initial scanning protocol may not have been ideal for the detection of ectopic glands, as an 8cm volume is small. A 16cm volume is still within the possibilities of axial scanning and can then include the upper mediastinal structures, however the dose will be higher as stated above.

Conclusion
We propose a feasible and directly implementable scanning-protocol for the detection of parathyroid adenomas. We include 16 different contrast phases, allowing for a dynamic evaluation within an acceptable exposure range when compared to classic helical 4DCT. Enhancement over time curves can be drawn. This way wash-in and wash-out of contrast in suspected lesions can be readily demonstrated. Since the interval for TTP for parathyroid adenomas in our study varies between 26 and 42 seconds after contrastadministration, arterial enhancement of the lesion could be less conspicuous when scanning 25 seconds after contrast administration, the most common practice in the literature. Since we obtain multiple phases, motion artifacts that could hinder the detection of small lesions can be overcome at least partially. Time to peak is found to be signi cantly earlier for parathyroid adenomas compared to normal thyroid tissue, MSI is signi cantly steeper. Lymph nodes can be readily differentiated from parathyroid adenomas due to their slow and continuous enhancement. Availability of data and materials The data analyzed during the current study is comprised within the article. The different sets of CT images are not publicly available due to medical con dentiality.

Competing interests
The authors declare that they have no competing interests  Tables   Table 1. Overview of the timing of the different scanning phases as obtained in 4DCT protocols for primary hyperparathyroidism in the literature