Comparison of duplex ultrasound and hemodynamic assessment with computed tomography angiography in patients with arterial thoracic outlet syndrome

Abstract

Thoracic outlet syndrome (TOS) presents with a variety of neurovascular symptoms, and its diagnosis cannot be established purely on the basis of clinical assessments. Computed tomography angiography (CTA) is currently the most useful investigative modality for patients with suspected vascular TOS. However, CTA facilities are limited, and CTA itself is an expensive and a resource-intensive technique associated with risks such as radiation exposure and contrast toxicity. Therefore, a screening test to identify the need for CTA may facilitate clinical management of patients with suspected TOS. Data for patients with suspected arterial TOS who underwent duplex ultrasound with arterial hemodynamic assessment (HDA) (pulse-volume recording and Doppler arterial pressure measurement) at King Saud University Medical City Vascular Lab between 2009 and 2018 were collected. The sensitivity, specificity, positive and negative predictive values (NPV), and area under the curve for duplex ultrasound and arm arterial HDA with CTA were reviewed. The data for 49 patients (mean age, 31 ± 14 years) were reviewed, of which 71% were female. The sensitivity, specificity, positive predictive value, and NPV of duplex ultrasound were 86.7%, 49.3%, 26.5%, and 94.6%, respectively. For arm arterial HDA, these values were 73.3%, 78.9%, 42.3%, and 93.3%, respectively. The combination of arm arterial HDA with duplex ultrasound scores yielded sensitivity, specificity, positive predictive value, and NPV of 93.3%, 42.3%, 25.5%, and 96.8%, respectively. The combination of duplex ultrasound with arm arterial HDA showed higher sensitivity and NPV than either test alone. The specificity of arm arterial HDA was significantly higher than that of the other measurements. When suspected, arterial TOS could be ruled out using duplex ultrasound and arm arterial HDA. These 2 investigations may help determine the need for CTA.

1. Introduction

The term thoracic outlet syndrome (TOS) was first used by Peet and colleagues in 1956 to describe a spectrum of conditions resulting from compression of the neurovascular bundle exiting the thoracic outlet.^[1,2] Compression of the brachial plexus, subclavian artery (SCA), or subclavian vein (SCV) results in neurogenic, arterial, or venous TOS, respectively.^[3] Neurogenic TOS is the most common form and constitutes over 95% of all TOS cases; venous TOS is found in 2 to 3% of TOS patients; and arterial TOS is a less common form found in <1% of all TOS patients.^[4] TOS is also more commonly diagnosed in females, and patients are usually aged 20 to 50 years at diagnosis.^[5]

The underlying causes of TOS can be multifactorial and may be related to congenital factors such as a cervical rib, traumatic events such as whiplash or a fall, or functional factors such as repetitive use injury in athletes and musicians.^[6] TOS is characterized by the presence of distinct symptoms, including upper-extremity pallor, paresthesia, weakness, muscle atrophy, and pain.^[7] Although TOS is suspected on the basis of clinical examination, the scope for differential diagnosis is wide, necessitating accurate diagnostic evaluation. Direct imaging techniques are often used to elucidate the underlying location and structure of compression. However, such diagnostic testing methods often yield equivocal or negative findings in cases of neurogenic TOS due to the lack of apparent or radiologically identifiable structural causes, making neurogenic TOS a diagnosis of exclusion.^[8,9] Assessment for vascular TOS more often results in direct identification of stenosis or occlusion of the SCA or SCV.^[10–12] The standards for reporting vascular TOS proposed by the Society for Vascular Surgery indicate that imaging results should be considered for the diagnosis of either arterial or venous TOS.^[13,14] In either case, early diagnosis is critical for timely intervention, since long-term compression can result in disability and permanent structural defects.

Several modalities are frequently utilized for diagnosis of patients who exhibit symptoms of TOS, including plain radiography, nerve-conduction studies/electromyography, ultrasonography, computed tomography angiography (CTA), dynamic CTA, magnetic resonance angiography (MRA), and digital subtraction angiography.^[8,13] Plain chest radiography is the most useful technique for an initial assessment and to identify or rule out bony anatomical abnormalities or defects, such as the presence of a cervical rib, first-rib anomalies, fracture calluses, congenital osseous malformations, and focal bone lesions that may compress the thoracic outlet.^[15] Electrophysiological tests, such as nerve-conduction studies or electromyography can help establish a diagnosis of neurogenic TOS.^[8] Newer imaging techniques, such as MR neurography and diffusion tensor imaging, are also emerging as useful diagnostic tools for neurogenic TOS. MR neurography facilitates non-invasive visualization of nerve morphology and signaling in peripheral nerves, while diffusion tensor imaging assesses central nervous system abnormalities.^[16,17]

For vascular TOS, duplex ultrasound is typically the initial imaging test. CTA and MRA can provide anatomical details in unclear cases,^[18] while digital subtraction angiography allows accurate evaluation of the vasculature around the thoracic outlet. However, digital subtraction angiography is invasive and associated with the risk of contrast toxicity^[19]; therefore, it is usually performed when therapeutic intervention is planned. The American College of Radiology Appropriateness Criteria state that CTA and MRA are appropriate tools for diagnosis of vascular TOS.^[14] CTA provides superior analysis of the vascular anatomy with respect to other structures, for example, the course of the nerves cannot directly be seen on a CT scan and has to be inferred from other surrounding structures.^[20] In contrast, MRA is more effective for the assessment of soft tissue structures, particularly the presence of accessory muscles (scalenus minimus, subclavius posticus, duplicated omohyoid inferior belly, or pectoralis minimus muscle), muscle hypertrophy (omohyoid inferior belly, pectoralis minor, scalene, or subclavius), and fibrous bands.^[18,20,21] To this end, CTA and MRA can be performed concomitant with postural maneuvers to facilitate diagnosis in patients with dynamically acquired compression.^{[7,20,22–24]} However, MRA cannot always distinguish between physiological and pathological compression of vascular structures. Additionally, MRA results do not always correlate with patient symptoms and are associated with a high false-positive rate of venous compression in asymptomatic populations.^[25,26] Encouraging results from CTA have identified this imaging modality as the most appropriate technique for identification of arterial or venous TOS, as evidenced by the correlation between significant SCA stenosis on dynamic CTA and thoracic outlet symptomatology.^[27]

Although the field of imaging-based diagnostics is rapidly evolving, these diagnostic imaging tests remain resource-intensive. CTA is costly, resource-intensive, and not available in all settings. Additionally, the technique is associated with radiation exposure and contrast-related risks. Finally, CTA evaluations only include the patient’s anatomy, precluding dynamic and functional assessments of blood flow through the thoracic outlet. Thus, other less invasive and more readily available diagnostic techniques may be useful to rule out a diagnosis of arterial TOS or strengthen the need for further diagnostic assessment.

Duplex ultrasound is a readily available and frequently employed technique in cases of suspected vascular TOS, and this imaging modality is considered an excellent initial assessment tool. Duplex ultrasound is often used for assessment of the presence of thrombosis, compression, stenosis, and aneurysmal changes with remarkably high sensitivity and specificity.^[28,29] However, this technique is operator-dependent, requires experience, and challenged by patient’s body habitus, and even healthy individuals may show a certain degree of arterial compression during the exaggerated positions commonly employed in ultrasound testing, with almost 30% of the healthy population demonstrating significant changes in Doppler waveforms in stress positions when compared with those in a neutral position.^[30] Hemodynamic assessment is also useful since arterial TOS can cause significant differences (>20 mm Hg) in arterial blood-pressure readings between both arms or during stress-position assessments,^[8] and identification of these differences can bolster the evidence for arterial TOS. However, since arterial TOS does not necessarily cause pressure differences in some cases, the rate of false-results in arterial TOS diagnoses using this method requires further evaluation; nevertheless, both duplex ultrasound and hemodynamic assessments are readily available, non-invasive, safe, and relatively inexpensive.

Considering the limitations and challenges inherent in the adjuvant imaging modalities currently used for diagnosis of arterial TOS and the challenges in identifying the index tests with the highest diagnostic accuracy,^[26] a better understanding of the utility of the more readily available diagnostic methods and their combinations for accurate assessment of arterial TOS is important to ensure timely diagnosis. Thus, the accuracy of duplex ultrasound and Doppler arterial HDA alone or in combination was compared with the accuracy of CTA in suggesting the diagnosis of arterial TOS.

2. Methods

2.1. Patients

Data from patients with symptoms suggestive of arterial TOS (claudication, hand pain, pallor, coldness, paresthesia, and/or digital ischemia) at King Saud University Medical City between 2009 and 2018 were reviewed. Duplex ultrasound assessment of blood flow in the SCA and SCV, arm arterial HDA, and contrast-enhanced helical CTA were performed for each patient. The tests were performed on the left and right upper limbs for each patient. The study was approved by the institutional review board for Health Sciences Colleges Research on Human Subjects, King Saud University College of Medicine (approval number: E-21-5911). The need for informed consent was waived due to the retrospective nature of the study.

2.2. Duplex ultrasound imaging

Duplex ultrasound imaging to examine the blood flow through the arteries and veins was performed using an ultrasound Phillips Epic 7-Color Doppler (Philips Medical Systems, Best, Netherlands), a 12-5-MHz linear transducer, and an 8-5-MHz curved transducer. The instrument utilized a constant angle of 60°, sample volume size of 1.5 mm, DR-56, wall filter of 60 Hz, and frame rate on medium mode. These ultrasound scans were performed with the patient in a sitting position in B-mode and pulsed wave Doppler with the 12-5-MHz linear transducer and 8-5-MHz curved transducer.

Duplex ultrasound evaluations were performed with the patient’s head rotated toward the contralateral side and the arms in a neutral position. In this position, the proximal SCA was examined anteriorly via the supraclavicular region, and the distal portion of the SCA was examined via the infraclavicular region. Subsequently, the ultrasound examination was repeated with the arm in a hyperabduction stress position (180°). The SCA was examined for thrombosis, stenosis, aneurysm, and post-stenotic dilatation. In addition, the artery diameter and peak systolic velocities were recorded, and the waveform was analyzed. The findings obtained on both sides were compared. The SCV was similarly examined for evidence of thrombosis.

The duplex ultrasound scan suggested a diagnosis of arterial TOS if there is a presence of an aneurysm, thrombus, symptomatic positional compression, stenosis, or post-stenotic dilatation was identified. During the stress position scan, a reduction of more than 2 mm or 30% in the SCA diameter was considered suggestive of arterial TOS. Additionally, a 2-fold increase in the peak systolic velocity from the neutral position, flow reduction/occlusion, or the presence of SCV thrombosis also suggested the diagnosis of arterial or venous TOS. Additionally, a difference of more than 20% in the peak systolic velocity between both upper limbs indicated a higher possibility of a positive arterial TOS finding.

2.3. Arterial HDA

Arm arterial HDAs were performed using calibrated instruments. Pulse-volume recordings and Doppler arterial pressure measurements were conducted using the Nicolet Vasoguard Pressure machine (VIASYS Healthcare, Conshohocken, PA) with a 7.5-MHz transducer. The pressure-cuff size was 10 cm for arm measurements. Bilateral upper-limb brachial artery pressure was measured in patients in the neutral position by using the pulse waveform. An initial difference of more than 20 mm Hg between both sides was suggestive of arterial TOS. These measurements were then acquired in a hyperabduction stress position (180°) with the head rotated to the contralateral side. The pressure recording and the waveform were noted. A drop of more than 20 mm Hg in the brachial artery pressure (especially with flattening of the waveform amplitude) suggested a positive diagnosis of arterial TOS.

2.4. Computed tomography

Contrast-enhanced helical CT was performed using a 64-row multidetector CT scanner (LightSpeed VCT; GE Healthcare, Chicago, IL) with a slice thickness of 0.625 mm, pitch of 1.375, and field of view of 50 cm. These scans included both shoulders and yielded images encompassing the C6 level to the mid-chest region. A total of 120 mL (60 mL per position) of iodinated non-ionic contrast material was utilized. Iodixanol iodine (320 mg/mL) was injected in the lower limb vein at a flow rate of 4 mL/s, and the scan timing after injection was based on a preliminary test bolus.

CTA was performed with the patient in a neutral position (arms at sides) and in a dynamic stress position (arms hyperabducted to 180° and in external rotation). The images were evaluated for the degree of arterial compression at the thoracic outlet following the hyperabduction stress maneuver in comparison with the arterial dimensions in the neutral position. The degree of stenosis was determined according to the SCA diameter reduction. A diameter reduction of more than 2 mm or 30% was considered suggestive of arterial TOS. In addition, the presence of notation of stenosis, aneurysmal changes, thrombus, or post-stenotic dilatation was considered suggestive of arterial TOS; the presence of SCV thrombosis suggested the presence of venous TOS.

2.5. Statistical analysis

All analyses were performed using IBM SPSS Statistics for Windows, Version 26 (Chicago, IL). CTA scan results were used as a reference; that is, positive CTA scan results were indicative of a true-positive result and negative CTA scan results were indicative of a true-negative finding. Duplex ultrasound and arterial HDA results were considered false-positive if the values indicated a positive result while the CTA scan result was negative. A false-negative result was indicated when the test results were negative while the CTA scan result was positive. On the basis of these measurements, the sensitivity, specificity, negative predictive value (NPV), positive predictive value, and false-negative rate were calculated. Missing data were treated with multiple imputation under the missing-at-random assumption. McNemar’s test was performed to compare the sensitivity and specificity of different investigation modalities, and all reported P values were 2-tailed. The cutoff P value was 05. Cochran’s formula was used for sample-size calculation based on a confidence level of 95%, ±5% precision, and the incidence of arterial TOS. Lastly, the distribution of continuous variables was assessed using D’Agostino’s K-squared test. Variables that followed a normal distribution were assessed using parametric tests, whereas non-normally distributed data were assessed using non-parametric tests.

3. Results

A total of 49 patients were enrolled. The patients’ mean age was 30.8 ± 13.8 (range 6–65) years, and 35 (71.4%) patients were female. The resultant dataset included the 98 limbs of 49 patients who underwent duplex ultrasound and CTA. Arm arterial HDA was performed in 45 patients (86 limbs). Table 1 shows the frequencies and percentages of positive and negative results for each of the test categories and their respective subcategories.

Table 1.

Positive and negative results of the different diagnostic tests.

Test	Subcategory	Positive, n (%)	Negative, n (%)
Duplex ultrasound (n = 49)	Thrombosis	0 (0)	98 (100)
	Stenosis	15 (15.3)	83 (84.7)
	Aneurysm	1 (1)	97 (99)
	Significant change in size >2 mm or 30%	11 (15.3)	61 (84.7)
	Significant change in PSV (2× increase or flow reduction/occlusion)	46 (52.9)	41 (47.1)
Change in arm arterial pressure (>20 mm Hg) (n = 45)		26 (30.2)	60 (69.8)
CTA scan (n = 49)	Thrombosis	1 (1)	97 (99)
	Stenosis	6 (6.1)	92 (93.9)
	Aneurysm	9 (9.2)	89 (90.8)
	Change in size	14 (17.1)	68 (82.9)

CTA = computed tomography angiography, PSV = peak systolic velocity.

On the basis of these results, aggregate scores for each of the test categories (duplex ultrasound, arm arterial HDA, and CTA) were computed (Fig. 1). For these scores, a positive result for any of the subcategories indicated a general positive result for the test. The duplex ultrasound test yielded 53 positive limb results, while CTA indicated only 19 positive results. Furthermore, arm arterial HDA showed 26 positive results.

Figure 1.

Number of positive and negative results for limbs tested for arterial thoracic outlet syndrome with computed tomography angiography, duplex ultrasound, and arm arterial hemodynamic assessment. CTA = computed tomography angiography, HAD = hemodynamic assessment, TOS = thoracic outlet syndrome.

After computing these aggregate scores, the sensitivity, specificity, positive predictive value, and NPV were computed by comparing the result of each vascular laboratory test and their combination with the CTA scan results. Therefore, the results of the duplex ultrasound and arm arterial HDA as well as their combination in comparison with CTA were considered (Table 2). A total of 86 limbs were assessed by all 3 tests and were included in the analysis.

Table 2.

Sensitivity, specificity, PPV, and NPV for duplex ultrasound, arm arterial HDA, and their combination in comparison with CTA.

Test/combination	TP, n	TN, n	FP, n	FN, n	Sensitivity, %	Specificity, %	PPV, %	NPV, %
Duplex ultrasound	13	35	36	2	86.7	49.3	26.5	94.6
Arm arterial HDA	11	56	15	4	73.3	78.9	42.3	93.3
Duplex ultrasound + arm arterial HDA	14	30	41	1	93.3	42.3	25.5	96.8

CTA = computed tomography angiography, FN = false-negative, FP = false-positive, HDA = hemodynamic assessment, NPV = negative predictive value, PPV = positive predictive value, TN = true-negative, TP = true-positive.

The combination of duplex ultrasound and arm arterial HDA exhibited better sensitivity and NPV than either test alone. The area under the curve for the combination was 0.74, suggesting that this combined approach had acceptable discriminative prediction ability (Fig. 2). Although no significant differences were identified in terms of sensitivity, the specificity of the combined approach was significantly lower than those of the 2 tests individually. Moreover, the specificity of arm arterial HDA showed higher statistical significance than the specificities of duplex ultrasound alone or the combination of duplex ultrasound and arm arterial HDA (Table 3).

Figure 2.

Receiver operating characteristic (ROC) curve for duplex ultrasound with arm arterial hemodynamic assessment score versus computed tomography angiography.

Table 3.

Comparison of the sensitivity and specificity of different diagnostic techniques by using McNemar’s test.

Test/combination	P value for the difference in sensitivity	P value for the difference in specificity
Duplex ultrasound vs arm arterial HDA	.317	<.001
Duplex ultrasound vs duplex ultrasound + arm arterial HDA	.317	.025
Arm arterial HDA vs duplex ultrasound + arm arterial HDA	.083	<.001

HDA = hemodynamic assessment.

Finally, the test results for limbs that showed positive results on CTA (n = 19) were assessed. Of these, 19 limbs underwent duplex ultrasound, while only 15 underwent arm arterial HDA. The combination of duplex ultrasound and arm arterial HDA exhibited the highest sensitivity of 93.3% (Fig. 3). Furthermore, the combination of duplex ultrasound and arm arterial HDA showed a lower false-negative rate than the results of either of the 2 tests individually (Table 4).

Figure 3.

Sensitivity and negative predictive value for each test and their combination in comparison with computed tomography angiography results (n = 86). HAD = hemodynamic assessment, NPV = negative predictive value, TOS = thoracic outlet syndrome.

Table 4.

Sensitivity and FNR for duplex ultrasound and arm arterial HDA of limbs showing a positive CTA result for arterial TOS (n = 19).

Test/combination	TP, n	FN, n	Sensitivity, %	FNR, %
Duplex ultrasound (n = 19)	16	3	84.2	15.8
Arm arterial HDA (n = 15)	11	4	73.3	26.7
Duplex ultrasound + arm arterial HDA (n = 15)	14	1	93.3	6.7

CTA = computed tomography angiography, FN = false-negative, FNR = false-negative rate, HDA = hemodynamic assessment, TOS = thoracic outlet syndrome, TP = true-positive.

As regards limbs showing positive results for arterial TOS on vascular lab studies and/or CT angiography with no other test (e.g., Raynaud’s tests, MRI) suggestive of another diagnosis (e.g., Raynaud’s Phenomenon, neurogenic TOS), surgical supraclavicular decompression was performed for 15 limbs with surgical findings suggestive of SCA compression and surgical decompression, including scalenectomy and cervical rib or first rib resection. Follow-up showed clinical improvement.

4. Discussion

The combination of duplex ultrasound and arm arterial HDA results showed a higher sensitivity and NPV than either test alone and was most successful in identifying true-positive cases in comparison with the CTA results. This analysis indicates that duplex ultrasound may be used in combination with arm arterial HDA to reliably rule out the need for CTA in patients with symptoms of arterial TOS. Although no differences in sensitivity were observed among duplex ultrasound, arm arterial HDA, and their combination, the specificity of arm arterial HDA was the highest, and the specificity of the combined approach was significantly lower than that of each individual test.

Although the diagnosis of arterial TOS is initially based on clinical findings, imaging studies are required to eliminate the possibility of other conditions with similar presentations and to specify the site and etiology of the compression, its grade, and any predisposing anatomical factors. While CTA is often the investigation of choice for such diagnoses, it is not as safe and convenient as HDA and duplex ultrasound since CTA exposes patients to radiation and potential contrast toxicity and is more expensive.

This analysis of patients who showed positive findings suggestive of arterial TOS on CTA revealed that the combination of duplex ultrasound and arm arterial HDA was the most sensitive (93.3%) and yielded a lower false-negative rate than either of the 2 tests individually, although the differences were not statistically significant. The frequencies of true-positive (14 of 15) and false-negative (1 of 15) cases as well as the sensitivity (93.3%) and false-negative rate (6.7%) of the combined approach support our conclusion that this combination may be a preferable investigational line for the evaluation of the presence of arterial TOS in suspected cases. This combination of tests is particularly warranted in cases where CTA is not available or for screening cases that require further investigation by CTA.

Duplex ultrasound is considered an excellent initial imaging test for the diagnosis of suspected vascular TOS, and its use is supported in the reporting standards issued by the Society for Vascular Surgery.^[13,14] Duplex ultrasound also has the advantage of allowing assessments of dynamic blood flow.^[21] Additionally, a recent report indicated that identification of compression of the SCV or SCA via duplex ultrasonography can also facilitate the diagnosis of neurogenic TOS,^[31] further supporting the use of this imaging modality. However, the high false-positive rates obtained in duplex ultrasound dynamic analysis indicate the need for updated ultrasound diagnostic protocols.^[30] On the other hand, significant (>20 mm Hg) differences in arterial pressure during stress positioning can also support the diagnosis of arterial TOS, and this relatively quick and simple method for assessment can provide weight to a suspected diagnosis. The effectiveness of the combination of these 2 non-invasive methods in comparison with CTA is particularly clinically relevant due to the possible established reliability of this combination and their safety and relatively lower cost in comparison with CTA.

This study had several limitations. First, the study was retrospective in nature; it only included 49 patients (98 limbs) and only recruited patients from a single tertiary center. Nevertheless, this study demonstrates that the combination of duplex ultrasound with arm arterial HDA could be utilized to determine the need for further evaluation by CTA in patients with suspected arterial TOS. Duplex ultrasound and arm arterial HDA are relatively simple and readily available investigative modalities that warrant further investigation though prospective, controlled studies with larger sample sizes.

In conclusion, this study demonstrated that the combination of duplex ultrasound and arm arterial HDA may be useful as a screening examination to warrant the need for further confirmatory CTA in cases of suspected arterial TOS. These methods and their combination proved to be sensitive, suggesting that the use of the safe, readily available, relatively inexpensive, non-invasive, and reliable duplex ultrasound and arm arterial HDA should be further examined for their relevance in the assessment of patients who present with symptoms suggestive of arterial TOS.

Acknowledgments

Dr. Hussam Anas, MD – Data collection; and Ms. Małgorzata Jakubowska, PhD – Statistical analysis.

Author contributions

Conceptualization: Talal A. Altuwaijri.

Data curation: Talal A. Altuwaijri.

Formal analysis: Talal A. Altuwaijri.

Methodology: Talal A. Altuwaijri.

Project administration: Talal A. Altuwaijri.

Validation: Talal A. Altuwaijri.

Writing – original draft: Talal A. Altuwaijri.

Writing – review & editing: Talal A. Altuwaijri.