High-resolution 0.25 mm Detector CT Has Limited Impact on Right Adrenal Vein Detectability in Preprocedural Contrast Enhanced CT for Adrenal Venous Sampling

Abstract

Objective:

This study aims to identify factors associated with the detectability of the right adrenal vein (RAV) on preoperative contrast-enhanced CT scans of adrenal venous sampling (AVS) in the era of high-resolution CT (HRCT).

Materials and Methods:

In this retrospective study, 36 patients (15 men and 21 women; mean age, 56 y) who underwent preoperative contrast-enhanced CT [11 patients in HRCT with 0.25 mm detector matrix (Cannon Medical Systems) and 25 patients in conventional multidetector CT with 0.5 mm matrix] were included. A contrast agent dose of 600 mgI/kg was injected, and CT images were acquired at a fixed scan delay of 50 and 80 seconds. Adrenal venography and venous sampling were performed for the diagnosis of suspected primary hyperaldosteronism. The qualitative detectability of RAV on preoperative CT was assessed with adrenal venography as a reference. Clinical and imaging factors associated with a good detectability of RAV were analyzed via regression analysis. Optimal acquisition timing was assessed by analyzing the time-intensity curve and contrast enhancement pattern of the inferior vena cava using CT data from a separate cohort (n=5).

Results:

The qualitative detectability of RAV was deemed good in 15 patients and poor in 21 patients. Regression analysis revealed that only heterogeneous enhancement of inferior vena cava with bolus high attenuation, corresponding to an optimal acquisition timing from time-intensity curve analysis, was associated with a good detectability of RAV (odds ratio, 5.06). The use of HRCT was not statistically significant.

Conclusions:

Optimal acquisition timing is a crucial factor for the detectability of RAV in preprocedural CT for AVS, while high-resolution 0.25 detector CT appears to have limited significance.

Primary aldosteronism is one of the major causes of secondary hypertension and results from the autonomous secretion of aldosterone from a unilateral adrenal adenoma or bilateral cortical hyperplasia (idiopathic hyperaldosteronism).^1,2 Differentiating the laterality of adrenal lesions is essential because surgical resection is recommended for a unilateral adrenal adenoma, while bilateral cortical hyperplasia should be treated with aldosterone antagonist drugs.³ Adrenal venous sampling (AVS) is the standard procedure for subtyping adrenal lesions and is recommended before surgical intervention.³

Catheterization of the right adrenal vein (RAV) during AVS can be difficult due to anatomic variant and small orifice of RAV,^4–6 often leading to prolonged examination times and increased radiation exposure. Preprocedural contrast-enhanced multidetector CT is a crucial imaging modality because it provides a detailed variation of RAV anatomy with a high detection rate.^6–8 However, the detection rate of RAV is lower than that of the left adrenal vein, as seen in a study using 320-row detector mapping CT (88% vs. 100%).⁹ Therefore, further efforts to improve RAV visualization on preprocedural CT are warranted.

The maximum spatial resolution, evaluated with a high-contrast object and low image noise, is largely dependent on the detector matrix size, which is 0.5 to 0.625 mm for multidetector CT for more than 20 years. A detector matrix size of 0.25 mm square has been implemented in a newly developed high-resolution CT (HRCT) (Cannon Medical Systems),¹⁰ which has improved the detectability of small, high attenuated arteries both in vitro and in vivo.^11,12 We hypothesized that this state-of-the-art HRCT could also enhance the visualization of the RAV compared with conventional CT. However, visualizing small veins, as opposed to arteries, may be challenging because small veins are less likely to be well-contrasted than arteries.

This study aimed to evaluate the potential effect of HRCT in improving the visualization of the RAV compared with conventional CT in patients scheduled for AVS for the diagnosis of primary aldosteronism.

MATERIALS AND METHODS

The institutional review board approved this single institutional retrospective study (reference number #2822), and the patient’s informed consent was waived.

Study Population

We searched for potentially eligible patients from our radiology archives. Contrast-enhanced abdominal CT has been performed for preoperative mapping of fluoroscopic AVS with an adrenal venous protocol in patients with a clinically suspected primary hyperaldosteronism. A 320-row multidetector CT with conventional resolution was used until November 2020, after which a 160-row multidetector CT with a high-resolution 0.25 mm detector matrix (HRCT) was implemented to improve adrenal vein visualization. Inclusion criteria were: (a) patients aged 18 years or older, (b) availability of both preoperative thin slice CT raw data and adrenal venography during AVS, (c) preoperative CT using either of the 2 CT scanners (HRCT or 320-row CT), and (d) a time interval >3 months between preoperative CT and adrenal venography (Fig. 1). Patients’ characteristics (age, sex, and body mass index) were recorded. There was no patient who had a significant cardiopulmonary or neurological dysfunction.

FIGURE 1.

Patients’ selection flowchart.

Acquisition Parameters of CT and Procedure of Adrenal Venous Sampling

Two CT scanners were used for preoperative examination; a 160-row detector CT scanner with a detector matrix size at the isocenter of 0.25×0.25 mm (Aquilion Precision; Cannon Medical Systems, Otawara, Japan) and a 320-row detector conventional CT with a matrix size of 0.5×0.5 mm (Aquilion ONE; Cannon Medical Systems, Otawara, Japan). The acquisition protocols were: tube voltage, 120 kVp; tube current, automatically adjusted (Volume EC); detector collimation, 0.25×0.25 mm (super high-resolution mode) for HRCT and 0.5×0.5 mm for conventional CT; gantry rotation time, 0.75 seconds/rotation; field of view, 400 mm; slice thickness, 0.5 mm; and a preset noise index of 15 HU. After acquiring non–contrast-enhanced images, an iodinated contrast agent (600 mgI/kg of body weight) was injected with a fixed 30-second injection duration. Images were subsequently acquired at 50 and 80 seconds of scan delay. Although the bolus tracking technique has been implemented in preprocedural mapping CT of AVS,^6,7 it was not used for mapping CT of AVS in our institution. Since the optimal scan delay for the adrenal vein has not been established, we selected a scan delay of 50 and 80 seconds based on previous clinical experience. Images were reconstructed with a slice thickness of 1.0 mm and a 1024×1024 matrix for HRCT (using an advanced intelligent clear-IQ engine) and a 512×512 matrix for conventional CT (using adaptive iterative dose reduction 3D) and a soft tissue kernel (FC03). CT dose index (mGy), as displayed on the CT console, was recorded.

Right adrenal venography was performed by manual injection of contrast material via a 2D or 3D curved 5F catheter canulated at the RAV orifice.⁴ Venous sampling of bilateral adrenal veins and inferior vena cava (IVC) was performed before and after the infusion of an adrenocorticotropic hormone analog, with careful attention to avoid contrast material contamination.

To assess whether the CT image timing matches the peak of RAV, we evaluated the relationship between the contrast enhancement pattern of the IVC and the time-intensity curve (TIC) using an additional 4-dimensional abdominal CT data from another cohort (n=5). Continuous low-dose upper abdominal volume CT data were acquired every 3 seconds between 20 and 80 seconds after contrast agent injection while patients were shallow breathing. TICs of the aorta, portal vein, and IVC at the right adrenal gland level were created by placing circular regions of interest.

Image Assessment

CT data acquired with a 50-second scan delay was used for image analysis. Data from the 80-second delay was not used, as the 50-second scan delay was close to the late arterial phase, which has been reported to achieve a high RAV detection rate.^6,7 Two radiologists (an abdominal radiologist and an interventional radiologist) independently assessed CT images.

(a) Detectability of the RAV on thin slice CT, the primary outcome, was evaluated with reference to sagittal and coronal reformatted images, using angiographic venography as a reference. The result was dichotomized into good or poor detection. Good detection was defined as a clear linear structure between the IVC and the right adrenal gland with high attenuation equal to or greater than the portal vein. Poor detectability indicated an equivocal or undetectable structure without strong enhancement (Fig. 2).

FIGURE 2.

Detectability assessment of the right adrenal vein on preoperative CT. The detectability of the right adrenal vein (RAV) on preoperative axial thin-slice images (A and D) and reformatted images (B and E) was assessed using right adrenal venography as a reference (C and F). In the upper row, the RAV is clearly depicted on preoperative CT images (white arrows in A and B) with high attenuation of the iodinated contrast agent, which was confirmed in right adrenal venography (C). In the lower row, the RAV could not be detected on preoperative CT images (D and E) even with reference to the right adrenal venography (F). In this case, an adrenal adenoma was noted in the right adrenal gland (gray arrow in E).

In addition, the following factors were assessed; (b) morphologic type of RAV, whether originating directly from IVC or from a common trunk with the hepatic accessory vein as previous reports^8,13; (c) contact between the right adrenal gland and IVC, assessed for fat tissue interposition or close contact without interposed fat (Fig. 3); (d) contrast enhancement pattern of IVC at the right adrenal gland level, assessed as either heterogenous enhancement with a high-attenuation bolus or homogenous enhancement (see Fig. 4 for the relationship between enhancement pattern of IVC and TIC).

FIGURE 3.

The distance between the right adrenal gland and inferior vena cava was assessed whether separate with fat tissue interposition (arrow in A) or close contact without fat tissue interposition (B).

FIGURE 4.

Relationship between time-intensity curve and contrast enhancement pattern of the inferior vena cava. The time-intensity curve of the aorta, portal vein (PV), and inferior vena cava (IVC) were generated by placing regions of interest (in the bottom row images) on continuously acquired CT images taken at 3-second intervals from 20 to 80 seconds after contrast agent injection. Symbols indicate mean values and vertical bars represent the SD of CT attenuation values (n=5). Peak enhancement of the IVC was observed at around 50 seconds, where heterogenous enhancement with high bolus attenuation was noticed (green arrow in C). After that, the contrast enhancement of the IVC gradually decreased, showing homogenous enhancement (green arrow in E).

Visceral fat and subcutaneous fat area (cm²) and body circumference (cm) at the umbilical level were automatically calculated using software (SYNAPS VINCENT; Fujifilm Medical, Tokyo, Japan). Image noise (HU) was measured as a mean value of SDs from 3 regions of interest placed on homogenous subcutaneous fat tissue. Quantitative analysis was performed by an abdominal radiologist.

Statistical Analysis

Inter-reader agreement on quantitative assessments (RVA detectability, morphologic type, right adrenal gland-IVC contact, and IVC enhancement pattern) between the 2 readers was evaluated using the Cohen kappa analysis, with results interpreted according to previous reports.¹⁴ Discordant categorical results were resolved by consensus and presented as raw date with percentages. Categorical variables were compared using a χ² or Fisher exact test, while continuous variables were compared using a Welch t test or Manny-Whitney U test, depending on its normal distribution.

The primary outcome of this study was RAV detectability on CT. Potential influencing factors included basic patients’ demographics, RAV morphology, right adrenal gland-IVC contact, IVC enhancement pattern (related to appropriate acquisition timing), CT scanner type, patient body habitus (measured on CT), and image noise. Univariate and multivariate logistic regression analyses were performed to determine independent factors associated with good RAV detectability (the dependent variable). Statistical analysis was performed using Prism 9 (GraphPad Software, San Diego, CA).

RESULTS

Of the 60 eligible patients, 24 were excluded, leaving 36 patients in the study cohort (see Fig. 1 for the patients’ selection flowchart). Among these 36 patients, preoperative abdominal CT was performed by using conventional CT for 25 patients and HRCT for 11 patients. The mean CTDIvol in the patient group undergoing HRCT was 14.4±1.5 mGy, compared with 17.0±3.0 mGy in the conventional CT group, with a significant difference between the 2 groups (P=0.005). The mean interval between the preoperative CT and AVS was 8 days. During AVS, both RAV cannulation and venous sampling were successfully performed in all patients. The inter-reader agreement for the RAV detectability was 0.50 (95% CIs, 0.22−0.78; moderate agreement), for the RAV morphology was 0.65 (95% CIs, 0.03-1.00; substantial agreement), for the right adrenal gland-IVC contact was 0.35 (95% CIs, 0.06-0.63; fair agreement), and for the IVC enhancement pattern was 0.60 (95% CIs, 0.34-0.86; moderate agreement).

The relationship between the enhancement pattern of IVC and TIC is shown in Figure 4, where peak enhancement occurred 50 seconds after the start of contrast injection. At this time, a bolus of contrast agent backed from the kidney entered the IVC and mixed with noncontrasted blood, creating a heterogenous enhancement pattern. According to this result, heterogenous enhancement with a high-attenuation bolus corresponds to optimal timing, while homogenous enhancement corresponds to suboptimal timing.

In the qualitative image analysis, RAV detectability was rated as good in 15 patients and poor in 21 patients. The results comparing clinical and imaging factors between patients with good and poor RAV detectability are presented in Table 1. Among these factors, only the IVC enhancement pattern showed a significant group difference (P=0.02).

TABLE 1.

Comparison of Clinical and Imaging Characteristics Between Good and Poor Detectability of Right Adrenal Vein

Variables	Patients with good detectability (n=15)	Patients with poor detectability (n=21)	P
Sex, man	7 (47)	8 (38)	0.61
Age, y	53.2±5.0	52.1+13.2	0.84
BMI	26.7±4.2	24.5±4.4	0.07
Visceral fat area, cm2	114±48	91±56	0.08
Subcutaneous fat area, cm2	194±96	167±82	0.49
Body circumference, cm	90±10.2	84±10.9	0.13
Type of CT scanner			0.95
Conventional CT	10 (67)	15 (71)
High-resolution CT	5 (33)	6 (29)
CT dose index, mGy	18.3 (13.9-34.2)	15.5 (9.2-30.6)	0.22
Image noise, HU	15.4±2.0	16.8±3.3	0.12
Distance between IVC and right adrenal gland			0.68
Separate	3 (20)	3 (14)
Contact	12 (80)	18 (86)
Morphologic pattern			0.91
Branching from IVC	13 (87)	18 (86)
Common trunk with AHV	2 (13)	3 (14)
Contrast enhancement pattern of IVC at 50 s			0.02
Heterogenous enhancement with bolus high attenuation	10 (66)	6 (29)
Homogenous enhancement	5 (33)	15 (71)

Data are presented as mean±SD or median (interquartile ranges) for continuous variables and raw date (percentages) for categorical variables.

AHV indicates accessory hepatic vein; BMI, body mass index; HU, Hounsfield’s unit; IVC, inferior vena cava.

In the univariate logistic regression analysis, only heterogenous enhancement with a high-attenuation bolus was a significant factor for good RAV detectability (odds ratios, 5.00; 95% CIs, 1.25-22.7) (Table 2). The use of HRCT did not significantly contribute to the detectability of RAV. In the multivariate logistic regression analysis including all variables, heterogenous enhancement with a high attenuation bolus remained a significant factor for good RAV detectability (odds ratios, 13.9; 95% CIs, 1.86-210), while other factors were not significant (Table 2).

TABLE 2.

Univariate and Multivariate Regression Analyses for Good Detectability of Right Adrenal Vein

	Odds ratio (95% CI)
Variables	Univariate analysis	Multivariate analysis
Sex (male)	1.42 (0.36-5.56)	0.22 (0.01-2.23)
Age (≥53 y)	1.03 (0.27-4.24)	0.68 (0.04-6.81)
BMI (≥25)	2.66 (0.69-11.3)	1.25 (0.05-52.9)
Visceral fat area (≥100 cm2)	3.75 (0.95-16.2)	9.70 (0.81-248)
Subcutaneous fat area (≥175 cm2)	1.26 (0.33-4.85)	0.59 (0.04-6.99)
Body circumference (≥87 cm)	2.00 (0.53-8.03)	0.47 (0.01-20.4)
Type of CT scanner (high-resolution)	1.25 (0.28-5.30)	1.54 (0.07-30.8)
CT dose index, mGy (≥17 mGy)	1.65 (0.43-6.56)	5.66 (0.21-249)
Image noise (≥16 HU)	0.45 (0.11-1.75)	0.36 (0.02-4.13)
Distance between IVC and right adrenal gland (separate)	0.65 (0.08-3.90)	0.45 (0.03-5.12)
Branching pattern (IVC)	1.08 (0.15-9.12)	0.52 (0.02-10.8)
Contrast enhancement pattern of IVC (heterogenous)	5.00 (1.25-22.7)	13.9 (1.86-210)

Continuous variables are dichotomized by the mean value. Odds ratios with 95% CIs for a good detectability of the right adrenal vein on preoperative CT are presented.

BMI indicates body mass index; IVC, inferior vena cava.

DISCUSSION

Our hypothesis for this study was that the ultra-high resolution detector matrix installed on HRCT could improve the subjective detectability of the RAV, short and small caliber vein (2 to 4 mm),¹³ on preprocedural mapping CT of AVS. However, in both univariate and multivariable regression analyses, the use of HRCT was not an independent predictor. Only heterogenous enhancement with a high attenuation bolus of the IVC corresponding to the peak enhancement was an independent predictor of good RAV detectability.

HRCT has demonstrated improved visualization of cranial lenticulostriate arteries,¹⁵ spinal Adamkiewicz artery,¹² coronary artery,¹⁶ and small visceral arteries.¹⁷ In phantom studies, HRCT also showed superior detectability of simulated small-caliber arteries as compared with conventional CT,¹¹ with small-caliber vessels filled with iodine contrast agents. While previous studies targeted arteries, which are easily enhanced, our study focused on small veins. Based on previous reports of HRCT’s superior detectability for high-contrast objects and our current results, UHRCT can be able to provide superior detectability only when the object shows high attenuation. Good detectability of HRCT requires adequate enhancement of the target vessels.

Adequate contrast enhancement of RAV can be achieved when the CT is acquired at an optimal time window that matches the peak enhancement of RAV. Although detailed TIC data of RAV has not been reported, it is reasonable to assume that an enhancement pattern of RAV is similar to that of the IVC at the adrenal gland levels. In our analysis of TIC of IVC, peak enhancement occurred 50 seconds after the start of contrast injection, when a bolus of contrast agent backed from the kidney entered the IVC and mixed with noncontrasted blood, creating a heterogenous enhancement pattern. Heterogenous enhancement pattern of the IVC was associated with good RAV detectability in our regression analysis, highlighting the importance of adequate enhancement for detecting small veins, even with HRCT.

The proportion of good detectability in our study, 42% (15/36), was lower than in previous reports,^6,7 where a bolus tracking technique was used to capture adequate enhancement of RAV, resulting in an excellent detection rate of RAV (78% to 93%) in the late arterial phase images. This discrepancy may be due to differences in qualitative evaluation criteria and imaging methods.

High-resolution can be also achieved with state-of-the-art photon counting CT, which is expected to lead to improvements in spatial resolution and image quality in the future. Our findings suggest that optimal acquisition timing is crucial for achieving improved visualization of small veins even in the era of photon counting CT.

We hypothesized that an anatomic variant of the RAV, a common trunk with the hepatic accessory vein, would contribute to the poor detectability of the RAV due to its small vessel diameter and complex anatomy. However, this factor did not show a significant impact, likely due to the limited number of patients with variant RAV in this study. In addition, the close contact between the right adrenal gland and the IVC without interposing fat tissue may also hinder RAV detectability. This is because the short distance between the right adrenal gland and the IVC, coupled with poor contrast against adipose tissue, makes the RAV harder to identify. While this factor may influence RAV detectability to some degree, optimal acquisition timing is a much more critical factor in ensuring the effective visualization of the RAV.

Our study has several limitations. First, this is a single-center retrospective study with a limited sample size, which introduces sampling bias and reduces statistical power. Further large-scale studies are needed. Second, our CT acquisition protocol using a fixed scan delay may miss the optimal scan timing. The use of a bolus tracking technique will be appropriate.^6,7 Although a subanalysis restricted to cases with adequate acquisition timing might provide more reliable insight, the small sample size of such cases (n=10 vs. n=6 of heterogenous enhancement pattern of the IVC) limits the robustness of the statistical analysis. Third, our dichotomized qualitative image assessment may either overestimate or underestimate subtle and ambiguous findings. Lastly, in addition to factors in our regression analysis, there may be other intrinsic and uncontrollable factors that could potentially affect RAV detectability.

In conclusion, although this was a retrospective small study, the simple use of high-resolution 0.25 mm detector matrix CT did not immediately improve the detection of RAV in a preprocedure CT for AVS. Optimal acquisition timing of the targeted venous structure is essential for achieving good detectability.