Radiological differentiation of phaeochromocytoma from other malignant adrenal masses: importance of wash-in characteristics on multiphase CECT.

Rationale and introduction: To evaluate the computerised tomography (CT) characteristics of phaeochromocytoma (PCC) that differentiate them from other non-benign adrenal masses such as adrenocortical carcinoma (ACC), primary adrenal lymphoma (PAL) and adrenal metastases (AM).
Methods: This retrospective study was conducted at a tertiary health care institute from Western India. Patients presented between January 2013 and August 2016 with histological diagnosis of PCC or other non-benign adrenal mass having adequate reviewable imaging data comprising all four CECT phases were included.
Results: The study cohort consisted of 72 adrenal masses from 66 patients (33 PCC, 22 ACC, 4 PAL, 13 AM). Unlike other masses, majority of PCC (25/33) showed peak enhancement in early arterial phase (EAP). PCC had significantly higher attenuation in EAP and early venous phase (EVP), and higher calculated percentage arterial enhancement (PAE) and percentage venous enhancement (PVE) than other adrenal masses (P < 0.001). For diagnosis of PCC with 100% specificity, PAE value ≥100% and EAP attenuation ≥100 HU had 78.8 and 63.6% sensitivity respectively. ACC were significantly larger in size as compared to PCC and metastasis. The adreniform shape was exclusively found in PAL (two out of four) and AM (4 out of 13). None of the enhancement, wash-in or washout characteristics were discriminatory among ACC, PAL and AM.
Conclusion: Peak enhancement in EAP, PAE value ≥100% and EAP attenuation ≥100 HU differentiate PCC from other malignant adrenal masses with high specificity.

Introduction

There is a recent rise in detection of adrenal incidentalomas (3–10%) with increasing use of radiological investigations. Benign adenomas are the most common lesions, accounting for 80% of adrenal incidentalomas, while adrenocortical carcinoma (ACC, 8%), phaeochromocytoma (PCC, 7%) and adrenal metastasis (AM, 5%) are other less common masses (1). Most crucial step in management of an adrenal incidentaloma is to differentiate a benign ‘leave-alone’ lesion, from the one that warrants further management. Hence, radiological differentiation of benign adenomas from other non-adenoma adrenal masses has been extensively studied (2). Phaeochromocytoma, a tumour that arises from chromaffin cells of adrenal medulla, constitutes an important non-adenoma adrenal mass. It is usually a benign tumour but can rarely be malignant (<10%). Although biochemical evaluation for catecholamine excess state can diagnose most of the PCCs, identification of specific and distinguishing radiological characteristics may provide useful clues to the clinicians. Because of obvious clinical implications, radiological differentiation of PCC from benign adenomas has been evaluated in several computed tomography (CT)-based studies (2). However, studies on radiological differentiation of PCC from other common malignant adrenal masses such as ACC, AM and primary adrenal lymphoma (PAL) are limited. Certain magnetic resonance imaging (MRI)- and MR spectroscopy-based characteristics have been shown to distinguish PCC from other malignant adrenal masses (3). In this regard, data on utility of CT-based characteristics, a more commonly used modality for abdominal imaging, are limited.

Characteristics on triple phase contrast-enhanced CT (unenhanced (UP), 60 s early venous (EVP) and 10–15 min delayed venous phase (DVP)) have not been found useful to differentiate PCC from other non-adenoma masses (4). Although hypervascular nature with avid enhancement during early arterial phase (EAP) is a known feature of PCC, systematic evaluation of arterial phase parameters to distinguish PCC from other malignant adrenal masses has not been reported (5). Hence, we aimed to study the wash-in characteristics that can distinguish PCC from other malignant adrenal masses in a retrospective cohort of patients with adrenal masses who had undergone multiphase CECT of adrenals.

Materials and methods

Patients

The study was conducted at a tertiary health care institute from Western India. The study was approved by the Seth GSMC and KEM Hospital Ethics Committee and considering retrospective nature of the study, waiver for patients’ consent was granted. We did a retrospective chart analysis of patients presented to our department between January 2013 and August 2016 with adrenal mass. Inclusion criteria were patients with available histological diagnosis of PCC or other non-benign adrenal mass having adequate reviewable imaging data comprising all four CECT phases that is UP, EAP, EVP and DVP. Final study cohort consisted of 72 adrenal masses in 66 patients (six patients had bilateral adrenal masses). These included 33 PCC, 22 ACC, 4 PAL and 13 AM (Fig. 1).

Figure 1

Flow chart showing assessment of adrenal masses for inclusion in the study.

Baseline demographic features like age and pre-operative hormonal profile were recorded for each patient. PCC were classified as metanephrine-secreting (increased plasma free metanephrine with/without elevation of plasma free normetanephrine), normetanephrine-secreting (elevated plasma free normetanephrine but normal plasma free metanephrine) or non-secretory (normal plasma free metanephrines). ACC was considered as cortisol secreting, if overnight dexamethasone supressed serum cortisol was ≥1.8 µg/dL. Diagnosis of patients with AM was based on primary tumour histopathology and/or adrenal biopsy.

CT protocol

Imaging of the abdomen was performed on a 64-slice multidetector CT system (Brilliance 64, Philips Healthcare) using standardised protocol. Patients were placed in supine position, with arms pulled caudally. The scanning protocol consisted of four identical helical scans obtained in an automated, predetermined and timed sequence. Scanning parameters were 120 kVp, with automatic exposure control (range, 140–220 mA), 0.75 seconds rotation time, pitch of 0.797 and a 0.625-mm detector configuration, with beam width of 40 mm. First phase was obtained at baseline (UP). 100 mL of iodinated contrast material (Iohexol: Omnipaque 300, GE Healthcare) was injected in the antecubital vein through a preplaced 18-gauge cannula, at the rate of 3 mL/s and was followed by a saline flush. The second phase (EAP) was obtained after 20 s of contrast administration. Third phase (EVP) and fourth phase (DVP) were obtained at 1 min and 15 min respectively. All scans obtained were stored on mass storage device (Seagate, Cupertino, CA, USA) and retrieved by attaching mass storage devices to picture archiving and communication system (PACS) as direct storage device. Images were retrieved by transferring them back to the PACS. Images were reconstructed in standard radiology workstation consisting of trans-axial, sagittal and coronal projections.

Imaging analysis

Two radiologists (with 21 years and 7 years of experience) who were blinded to histopathological diagnosis reported the CT images on workstation. Morphological parameters like size (maximum diameter in axial images), shape (spherical, adreniform or irregular) and presence of necrosis, haemorrhage and calcification were noted. For each phase, attenuation in HU was measured by placing elliptic region of interest (ROI) of minimum 1 cm2, at three separate sections of the mass, and mean of these three values was calculated. ROI was placed on maximally enhancing areas in the EAP and replicated in identical locations for other (UP, EVP and DP) phases. For ROI placement, due care was taken to cover the maximum possible area, after excluding areas of calcification, necrosis, haemorrhage and surrounding fat. Additionally, aortic HU (at the level of renal artery) was measured in EAP. Relative washout percentage (RWP = (EVP HU − DP HU) × 100/EVP HU) and absolute washout percentage (AWP = (EVP HU − DP HU) × 100/(EVP HU − UP HU)) were calculated. ‘Peak enhancement phase’ was defined as the phase in which enhancement was maximum with attenuation value greater than other phases by >5 HU. Percentage arterial enhancement (PAE) was calculated as a measure of enhancement in EAP as compared to that in UP and was calculated as PAE = (EAP HU − UP HU) × 100/UP HU. Percentage venous enhancement (PVE) was calculated as a measure of enhancement in EVP as compared to UP and was calculated as PVE = (EVP HU − UP HU) × 100/UP HU.

Statistical analysis was done using SPSS version 23, IBM. Six patients in our cohort had bilateral masses. The statistical analysis was done with ‘mass’ (rather than patient) as a unit. Any missed data (demographic features and hormonal profile) were addressed using multiple imputation method. Categorical variables (shape, presence/absence of necrosis, calcification and haemorrhage, peak enhancement phase) were expressed in actual numbers or percentages and compared by Kruskal–Wallis test. Continuous variables (age, size, HU in different phases, AWP, RWP, PAE, PVE and aortic HU) were expressed as mean ± s.d. and were compared using one-way ANOVA. Post hoc test (LSD) was done to study intergroup comparisons.

Inter-rater reliability for shape, presence of necrosis and calcification was assessed by calculating Cohen’s Kappa (K) value. K value more than 0.6 was considered as good level of agreement. Inter-observer variation in measurement of attenuation values was assessed by inter-class correlation coefficient (ICC), and a value more than 0.8 was considered as good correlation. Receiver-operating characteristic (ROC) curves were used to determine the cut-off value(s) for different CECT parameters that can best differentiate PCC from other adrenal masses with maximum specificity. P value <0.05 was considered as statistically significant.

Results

The study cohort consisted of 72 adrenal masses from 66 patients (6 patients with bilateral adrenal masses). Mean age of the study population was 41.4 ± 16.8 years. Cohen’s kappa (k) value for inter-rater reliability among the two radiologists for shape, presence of necrosis and calcification was 0.978 (P < 0.001, 95% CI 0.95–0.98). ICC for HU measurements between the two radiologists was 0.98 (P < 0.001, 95% CI 0.97–0.98). Comparison of morphological, contrast enhancement and washout characteristics of various adrenal masses is shown in Table 1. Age of presentation was not significantly different among the different groups of adrenal masses.

Table 1

Comparison of morphological, contrast enhancement and washout characteristics of various types of adrenal masses.

	PCC (n = 33)	ACC (n = 22)	Lymphoma (n = 4)	Metastasis (n = 13)	P value
Age (years)	38.8 ± 18.6	41.5 ± 12.3	36.4 ± 18.2	56.6 ± 5.9	0.145
Morphological characteristics
Shape
Spherical	23	13	2	7	<0.001a
Adreniform	0	0	2	4
Irregular	10	9	0	2
Size (cm)	5.5 ± 2.4	9.1 ± 3.9	7.1 ± 1.4	4.5 ± 2.7	<0.001b
Necrosis	32	22	4	12	0.42
Calcification	2	7	0	0	0.194
Haemorrhage	0	0	0	0
Enhancement patterns
Unenhanced phase enhancement (HU)	37.1 ± 6.1	38.4 ± 6.02	37.1 ± 1.3	32.6 ± 13.6	0.239
Arterial phase enhancement (HU)	117.6 ± 41.1	54.8 ± 11.9	48.2 ± 5.2	48.8 ± 17.6	<0.001c
Early venous phase enhancement (HU)	102.7 ± 29.4	70.1 ± 12.1	58.1 ± 1.5	60.5 ± 19.7	<0.001d
Delayed venous phase enhancement (HU)	76.5 ± 42.3	55.9 ± 12.5	53.0 ± 4.2	41.5 ± 24.2	0.003e
Absolute percentage washout (%)	34.1 ± 20.6	44.2 ± 23	24.2 ± 24.1	29.9 ± 6.7	0.275
Relative percentage washout (%)	21.2 ± 20.5	19.6 ± 12.7	10.7 ± 2.3	36.3 ± 3.1	0.70
Percentage arterial enhancement (%)	227 ± 129	43.4 ± 28.1	29.7 ± 9.8	29.7 ± 9.8	<0.001c
Percentage venous enhancement (%)	186.6 ± 102.4	84.1 ± 31.5	56.6 ± 5.9	137.3 ± 16	0.009
Aortic enhancement (HU)	285 ± 65	256.9 ± 85.2	257 ± 86	289.2 ± 33	0.768

Post hoc test (LSD).

aP value significant between lymphoma vs PCC (<0.001), ACC (<0.001) and metastases (0.001); bP value significant between PCC vs ACC (P < 0.001) and ACC vs metastasis (P < 0.001); cP value significant between PCC vs ACC (P < 0.001), lymphoma (<0.001) and metastasis (P < 0.001); dP value significant between PCC vs ACC (P < 0.001), lymphoma (0.001) and metastasis (P < 0.001); eP value significant between PCC vs ACC (P = 0.012) and metastasis (P = 0.001).

ACC, adrenal cortical carcinoma; PCC, phaeochromocytoma.

PCC

Thirty-three PCCs were included in the analysis (three patients had bilateral PCCs). Twenty-two PCCs were normetanephrine-secreting, nine were metanephrine-secreting and two were non-secretory. In UP, the attenuation of PCC was not significantly different from those of other subgroups (Figs 2, 3 and Table 1). Majority of PCCs (25/33) showed peak enhancement in EAP, whereas six PCCs had peak enhancement in EVP and remaining two PCC were equally enhancing (peak) in EAP and EVP (difference being <5 HU). None of the masses from other subgroups had peak enhancement in EAP. PCCs had significantly higher attenuation in EAP and EVP, and higher calculated PAE and PVE than other adrenal masses (P < 0.001). Among these, arterial phase characteristics had the best AUC in ROC curve analysis to differentiate PCC from other masses (Fig. 4 and Table 2). For diagnosis of PCC with 100% specificity, PAE value ≥100% and EAP HU ≥100 had 78.8% and 63.6% sensitivity respectively.

Figure 2

(A, B and C) Axial CECT images with attenuation value mentioned in different phases. (A) Right adrenal phaeochromocytoma (nor-metanephrine secreting) showing peak enhancement in early arterial phase. (B) Right adrenal phaeochromocytoma (metanephrine secreting) showing peak enhancement in early venous phase. (C) Left adrenocortical carcinoma.

Figure 3

(A and B) Axial images of CECT with attenuation value mentioned in different phases. (A) Left adrenal metastases. (B) Left primary adrenal lymphoma.

Figure 4

The receiver-operating curve (ROC) analysis for different parameters for discriminating phaeochromocytoma from other malignant adrenal masses.

Table 2

Diagnostic performance of various imaging parameters represented by area under the curve calculated using receptor-operating characteristic curves.

Parameter	AUC	95% CI
Percentage arterial enhancement (%)	0.93	0.87–0.99
Early arterial phase enhancement (HU)	0.91	0.86–0.99
Early venous phase enhancement (HU)	0.86	0.77–0.95
Percentage venous enhancement (%)	0.80	0.69–0.91
Delayed phase enhancement (HU)	0.81	0.71–0.91
Relative percentage washout (%)	0.49	0.41–0.68
Absolute percentage washout (%)	0.67	0.55–0.80

Out of seven PCCs which had PAE <100%, five were metanephrine secreting. Similarly, out of 12 PCCs with EAP HU <100, eight were metanephrine secreting. Thus, based on the observation that higher proportion of PCC with PAE <100% and EAP HU <100 were metanephrine-secreting, sub-group analysis was done to compare CT characteristics of PCC with these secretory subtypes (Table 3). Metanephrine-secreting PCC had significantly lower attenuation in EAP and EVP, lower PAE and PVE as compared to nor-metanephrine-secreting PCC. Additionally, proportion of the PCC with peak enhancement in EAP or EAP HU ≥100 were significantly lower in the metanephrine-secreting PCC group.

Table 3

Comparison of computed tomography characteristics of metanephrine-secreting and normetanephrine-secreting phaeochromocytoma.

	Normetanephrine secreting (n = 22)	Metanephrine secreting (n = 9)	P value
Number of tumours with maximum enhancement in arterial phase (%)	20 (90.9%)	5 (55.5%)	0.047
Number of tumours with arterial phase enhancement >100 HU (%)	18 (81.8%)	1 (11%)	<0.001
Size (cm)	5.6 ± 2.2	5.8 ± 3.2	0.83
Unenhanced phase enhancement (HU)	37.2 ± 6.6	37.5 ± 5.6	0.907
Arterial phase enhancement (HU)	136.6 ± 33.6	78.4 ± 24.8	<0.001
Early venous phase enhancement (HU)	115 ± 26.4	75.6 ± 18.2	<0.001
Delayed venous phase enhancement (HU)	86.6 ± 48.3	56 ± 14.1	0.07
Percentage arterial enhancement (%)	279.4 ± 114.8	117.1 ± 90	0.001

HU, Hounsfield units.

ACC, AM and PAL

Twenty-two ACCs (ten were cortisol secreting), 4 PAL and 13 AM (three patients with bilateral metastases) were included in our study. ACCs were significantly larger in size as compared to PCC and metastasis. The adreniform shape was exclusively found in PAL (two out of four) and AM (4 out of 13). While necrosis was prevalent in almost all the groups of adrenal masses, calcifications were seen in a minority of PCC and ACC. Adrenal masses in each of these three subgroups had peak enhancement in EVP and none of the enhancement, wash-in or washout characteristics were discriminatory among these three subgroups (Table 1).

Discussion

Up to 14% of adrenal incidentalomas are reported to be PCCs (1). Especially amidst the non-adenoma adrenal masses, PCC form a sizeable proportion of incidentalomas (~35%), making its radiological characterisation a pertinent area of investigation. While CECT-based differentiation of PCC from benign adenomas has been extensively studied, there are limited studies on radiological differentiation of PCC from other malignant masses. In the current study, we report multiphase CECT characteristics that differentiate PCC from other malignant adrenal masses like ACC, AM and PAL. Previously, Szolar et al. (4) have reported CT attenuation characteristics on unenhanced, 1min and 10 min post contrast scans of 73 adrenal masses which included ACC, PCC, AM and adenomas. They found no significant difference in either mean attenuation HU values in any of the three phases or washout percentages among PCC, ACC and AM. In contrast, we found that mean attenuation HU values in EVP were significantly higher in PCC than in ACC and AM. Lesser number of PCC (n = 17) and presence of few (n = 2) hypervascular metastases in their cohort might have accounted for this difference in findings (4).

In our study, we found that addition of wash-in EAP parameters (PAE value ≥100%, peak enhancement in EAP and EAP attenuation ≥100 HU) provided the best diagnostic accuracy to differentiate PCC from other malignant adrenal masses. Although wash-in characteristics of adrenal masses on CECT have been studied previously (6, 7, 8, 9), studies reporting systematic evaluation of diagnostic accuracy of various EAP parameters to differentiate PCC from other adrenal masses have been limited. In a study by Szolar et al., wash-in characteristics of 74 adenomas and 61 nonadenomas were studied (6). They found the adenomas to have significantly higher attenuation at 30 s and higher PAE (referred in their paper as ‘percentage increase in attenuation’) than nonadenomas. However, their non-adenoma masses mostly consisted of metastases (n = 47) and few PCC (n = 6, 30 s scan available for 2 PCC only) and no comparison was made amidst the various non-adenoma masses. Use of PVE as a wash-in parameter was reported by Foti et al. in their study of 107 adrenal masses (56 adenomas and 51 metastasis) (9). While they found no significant difference in PVE (referred in their study as ‘relative percentage wash-in ratio from the unenhanced to the portal venous phase’) between adenomas and metastasis, there were no PCCs in their cohort. ‘Wash-in attenuation’ (calculated as the difference between 35 s enhanced attenuation and unenhanced attenuation) and ‘phase of achieving peak enhancement’ are the other wash-in parameters reported in studies mainly comparing adenomas with nonadenomas (7, 8). However, we believe that our results are not directly comparable with any of these previous studies reporting wash-in characteristics of adrenal masses, due to basic differences in composition of the cohorts (with no adenomas in our cohort and few PCC in previous studies).

We found interesting differences in attenuation parameters between biochemical sub-types of PCC, with normetanephrine-secreting PCC having higher attenuation profile than metanephrine-secreting PCCs. To the best of our knowledge, this has not been reported previously. This is most probably due to implication of hypoxia pathway defects (Cluster 1 genes: VHL and SDHx) in the pathogenesis of former than the latter (Cluster 2: RET, NF1, TMEM127) (10). Hypervascular nature of PCC is also the most probable reason for their higher arterial enhancement than other malignant masses.

Although not an aim of our study, differentiating PCC from benign adenomas is important from the clinical standpoint. Since both adenomas and PCC can have good as well as poor contrast washout patterns (2, 11), it would be informative to test the performance of wash-in EAP parameters in differentiating PCC from adenomas. In this regard, a study by Northcutt et al. compared EAP (25 s) and EVP (1 min) enhancement of 12 PCC with 41 adenomas (7). They also found EAP parameters (absolute arterial enhancement >110 HU and peak enhancement in EAP) as specific for PCC, distinguishing them from adenomas. While PAE and PVE of adenomas vs PCC have not been studied adequately, EAP attenuation values of adenomas have been consistently shown to be significantly lower than that of PCC and can provide useful clues (7, 12, 13).

Our findings can have some implications in the evaluation of incidental adrenal masses. Arterial phase wash-in characteristics can be readily obtained from the routinely performed triple-phase (EAP, EVP and 3-min DVP) CECT abdomen scans, the scan in which an adrenal incidentaloma is often diagnosed first. In case of an incidentaloma, documenting hypervascular nature using wash-in characteristics differentiates PCC from ACC and other malignant masses, as demonstrated in our study, as well as from adrenal adenoma as demonstrated by previous studies (7, 12, 13). This may reduce the need for a repeat adrenal focussed multiphase CECT imaging and may also help to minimise the laboratory evaluation by choosing plasma-free metanephrines, particularly plasma free normetanephrine as the most pertinent initial biochemical test.

Technical factors related to contrast administration, cardiovascular status of patients, and true biological tumoural vascular characteristics are the few important factors that can influence arterial enhancement pattern of a tissue (14, 15). Since this is a single-centre study performed over a relatively shorter duration, the technical aspects like type, amount and rate of contrast agent injection have largely remained uniform for patients included in the study. The fact that the aortic HU was similar among different tumour groups rules out the confounding effect of cardiovascular status on enhancement patterns (Table 1). Hence, it is likely that the difference in enhancement patterns in our study reflects the biological difference in underlying tumoural vascular characteristics. Considering robust reference of histopathology for diagnosis and reporting of images in a blinded fashion (to minimise the observer error) were other strengths of our study.

By attempting to differentiate PCC from other malignant adrenal masses and to radiologically characterise biochemical sub-types of PCC, we have studied an area less addressed in previous CECT studies on PCCs. However, exclusion of benign masses limits the wide clinical applicability of our results. We have included consecutive patients satisfying the inclusion criteria (to minimise the selection bias); however, due to the retrospective nature of the study, a large proportion of patients with adrenal masses were excluded due to lack of four phase adrenal imaging and/or histopathology (Fig. 1) which might have added significant selection bias. Small sample size is another major limitation of our study. Additionally, few hypervascular adrenal masses such as metastasis from renal-cell carcinoma or hepatocellular carcinoma with potential to yield higher PAE values were not part of our cohort. Hence, we propose the need of larger, multi-centric, prospective studies with wider spectrum of benign as well as malignant masses which helps to confirm and generalise the study findings.

To conclude, we present a report on multiphase CECT-based characterisation of PCC and found that certain arterial phase-based wash-in parameters like PAE value ≥100%, peak enhancement in EAP and EAP attenuation ≥100 HU can differentiate PCC from other malignant adrenal masses with high specificity.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This work did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.