Anatomic variants of the subscapular-thoracodorsal arterial system: A radiologic analysis of 200 arterial systems

Abstract

Background:

Composite flaps based on the subscapular arterial system are excellent choices for complex defects, including those of the head and neck, though rates of anatomic variants are not well described.

Objective:

Characterize subscapular-thoracodorsal arterial system in a large cohort of patients using CT chest angiography.

Methods:

CTA chest studies from 100 adult patients were analyzed to characterize the bilateral subscapular-thoracodorsal arterial systems.

Results:

Out of 200 arterial systems, 25 (12.5%) were lacking a subscapular artery, with the thoracodoral and circumflex scapular arteries arising independently off of the axillary (or other nearby vessels). Strikingly, the subscapular artery was absent bilaterally in 5 patients and absent unilaterally in 15 patients, meaning that one in five patients harbored abnormal anatomy on at least one side. There was no radiographic evidence of atherosclerosis in the studied vessels in any patient, including smokers and patients with atherosclerosis in other vessels.

Conclusion:

Variations in the subscapular-thoracodorsal arterial system appear more frequent than previously described. For select patients requiring complex reconstruction using the scapular system, CTA chest may aide surgical planning.

Introduction

Reconstruction of large, composite defects, including those of the scalp, midface, and oral cavity remain a formidable challenge for the head and neck microvascular surgeon. While advances in technique have allowed for increased usage of double free flap reconstructions, these options are associated with increases in operative times and patient morbidity.¹ The subscapular-thoracodorsal system can be used for mega-flap free tissue transfers, providing osseous, soft tissue, and cutaneous components based on a single vascular pedicle that can be flexibly arranged into different three-dimensional configurations.²

Variations of this anatomy, however, are usually discovered intra-operatively. Absence of a common SSA trunk converts a one-vessel, mega-flap reconstruction into a double free flap, requiring two separate anastomoses, thereby altering the risk profile of the surgery, and disrupting the original reconstructive plan. In addition, in the vessel depleted neck (e.g. previously operated or radiated patients), a limited number of recipient arteries and veins may be available, making such a finding intra-operatively a challenging situation.

Most commonly, the subscapular artery (SSA) originates as a branch from the third part of the axillary artery (AA), with the circumflex scapular artery (CSA) and thoracodorsal artery (TDA) arising as the two main divisions of the SSA. The periosteal branch of the CSA supplies the lateral border of the scapula, while the transverse and descending cutaneous branches supply the scapular and parascapular cutaneous flaps, respectively. The TDA supplies the scapular tip through the angular branch and latissimus dorsi and overlying skin through perforators arising from its two terminal branches. Classically, the TDA is estimated to arise directly from the AA in 5% of cases where there is absence of a common subscapular arterial trunk, though this estimate is primarily based on a series of 58 cadaver dissections (100 axillary arteries).³ Subsequent, cadaver-based anatomic studies have estimated this variant anatomy to exist in 3.3% (of n = 60 arteries)⁴ and 19% (of n = 16 arteries) of cases.⁵

While cadaveric dissections allow for the description of branching patterns of smaller vessels (e.g. cutaneous perforators), CT angiography (CTA) allows for 3D reconstruction of the SSA, CSA, and TDA branching patterns with sufficient resolution to describe the rates of variants in a larger patient cohort. The purpose of this study is to characterize rates and patterns of anatomic variations of the subscapular-thoracodorsal arterial system radiologically in a large series of patients.

Methods

The study received approval by the Institutional Review Board at the Washington University School of Medicine. One hundred adult patients who underwent a CTA chest study at BJC Healthcare between May 2020 and May 2021 were retrospectively identified and included in this study, ensuring that there was an even 50% split for each sex. A dedicated board certified radiologist (H.O.) reviewed the images for the following: presence of a SSA, CSA, TDA on each patient side, vessel of origin for CSA and TDA, and radiologic evidence of atherosclerosis. Depending on the course of the SSA, the vessel length was measured in the coronal, sagittal or oblique plane. The SSA and CSA vessel widths at their origin as well as the length from the CSA origin to the lateral margin of the scapula were measured on the axial images. Patient demographic and clinical variables were collected, including age, sex, race, height, weight, BMI, clinical (e.g. known coronary artery disease) and/or radiographic evidence (e.g. aortic atherosclerosis) of systemic atherosclerosis, and smoking history. Univariate linear regression was used to assess relationships between vessel length and width and patient height, weight, and BMI, stratified by sex. Wilcoxon rank-sum testing was used to compare vessel length and width between laterality and between sexes.

Results

Demographic and clinical features of the included 100 patients are summarized in Table 1. Of note, there was no radiographic evidence of atherosclerosis in the SSA, CSA, or TDA in any patient, even in patients with positive smoking histories (25% current, 39% former) and either clinical or radiographic evidence of atherosclerosis elsewhere (60.6%).

Table 1.

Demographic and clinical characteristics of included patients. sd = standard deviation

	Patients
	Overall (n = 100)	Male (n = 50)	Female (n = 50)
Features	mean (sd)	mean (sd)	mean (sd)
Age	57.13 (15.8)	57.6 (17.1)	57.6 (17.1)
BMI	30.31 (7.3)	29.3 (6.56)	31.3 (7.9)
Height	1.72 (0.12)	1.80 (0.09)	1.63 (0.08)
Weight	89.4 (22.9)	95.1 (22.1)	83.8 (22.4)
Race	n (%)	n (%)	n (%)
White	82 (82)	43 (86)	39 (78)
Black	18 (18)	7 (14)	11 (22)
Smoking	n (%)	n (%)	n (%)
Current	25 (25.2)	12 (24)	13 (26.5)
Former	39 (39.4)	18 (36)	21 (42.8)
Never	35 (35.3)	20 (40)	15 (30.1)
Systemic Atherosclerosis	n (%)	n (%)	n (%)
Yes	60 (60.6)	32 (64)	28 (57.1)

Table 2 summarizes the anatomic variation observed in the present study. In 25/200 analyzed axillary arterial systems, the SSA was absent (Table 2, Figure 1). In all of these cases, the TDA arose directly from the AA, and the CSA arose from the AA in 23 patient sides, from the posterior circumflex humeral artery in 1 case, and the origin could not be confidently identified as present or absent due to presence of contrast in the adjacent venous system. Interestingly, five of these patients had absent SSA bilaterally and 15 patients had an absent SSA unilaterally, meaning that 20% of our included patients had an absent SSA trunk on at least one side.

Table 2. Presence, length and caliber of the SA, CSA, and TDA.

Wilcoxon Rank Sum test for comparisons. Length in cm and width in mm.

	Length and Caliber of SSA, CSA, and TDA
	Overall	Right	Center	p	Male	Female	p
SSA
Present	175	85	90		44	41
Length	2.46 (1.11)*	2.65 (1.23)	2.28 (0.96)	0.079	2.88 (1.24)	2.03 (0.77)	<0.001
Width	3.27 (0.92)	3.38 (0.96)	3.16 (0.87)	0.122	3.54 (0.98)	2.98 (0.74)	<0.001
CSA
Present	197	98	99		48	50
Length**	2.92 (0.88	2.98 (0.91)	2.86 (0.84)	0.417	3.19 (0.90)	2.66 (0.77)	<0.001
Width	2.00 (0.63)	2.05 (0.66)	1.95 (0.61)	0.299	2.23 (0.67)	1.77 (0.49)	<0.001
TDA
Present	200	100	100		50	50
Width	1.45 (0.39)	1.46 (0.39)	1.44 (0.40)	0.66	1.53 (0.39)	1.37 (0.38)	<0.001

^*9 could not be reliably measured.

^**Distance to lateral margin of scapula. SSA = subscapular artery, CSA = circumflex scapular artery, TDA = thoracodorsal artery

Figure 1.

(A) Post-contrast 3D reconstructed and (B) post-contrast maximum intensity projection (MIP) CT angiographic images through the chest demonstrate separate origins of the right circumflex scapular artery (arrow) and thoracodorsal artery (arrow head). (C) Post-contrast 3D reconstructed and (B) post-contrast maximum intensity projection (MIP) CT angiographic images through the chest demonstrate a subscapular artery (arrow) arising from the axillary artery which gives rise to the circumflex scapular (dashed arrow) and thoracodorsal artery (arrow head)

CSA that arose from the AA were larger than those that arose from the SSA (mean width 2.45 mm vs 1.94 mm, p = 0.004). Similarly, TDA that arose from the AA were wider than TDA that arose from the SSA (mean width 1.70 mm vs 1.41 mm, p = 0.005). Mean vessel length and width were significantly greater in men versus women (Table 2, p <0.001), though vessel length of the TDA could not be consistently measured. While there were weak correlations between SSA and CSA width and height, weight and BMI, there was surprisingly little correlation between patient height and vessel length (Figure 2).

Figure 2.

Correlation of SSA (left) and CSA (right) vessel length (top) and width (bottom) with height, weight, and BMI, stratified by sex. Length is in cm; width is in mm.

Discussion

Our present study is the largest analysis of the anatomic variation of the subscapular-thoracodorsal arterial system present within and across patients. Among included patients, the TDA arose from the AA in 12.5% (25/200) of patient sides in our dataset. This rate sits within the wide range of previously reported rates from cadaveric studies (3-19%)^3-6, though more than double that of the classically described 5%. More remarkable, this absence of a subscapular trunk was observed in 5 patients bilaterally (10 patient sides), and 15 patients unilaterally (15 patient sides). Thus, 20 patients (5 bilateral + 15 unilateral) out of 100 patients included had at least one variant subscapular-thoracodorsal system without a clear predilection for laterality. These findings stand in contrast to one prior study that also used chest CTA images to characterize the subscapular arterial tree, in which they reported that the CSA branched directly from the AA in only 3.2% of patients (of n = 92 cases), and they did not report rates of intrapatient variations.⁷ Taken together, our findings suggest that preoperative CTA of the chest may aid surgical planning for patients who will undergo a mega-flap reconstruction incorporating both the CSA and TDA off of a common subscapular trunk. In particular, a CTA chest study could guide which side to use, whether to use a chimeric flap or not, and whether to use the scapula tips versus the scapular border. These findings would be especially useful for patients with a vessel depleted neck in whom performing two anastomoses may prove more challenging.

The mean radiographic vessel diameters of the SSA, CSA, and TDA in our study (3.3 mm, 2.0 mm, and 1.45 mm respectively) fell within the reported ranges from prior cadaveric studies though all were consistently below the previously reported means.^3-5 There are multiple possible explanations for the discrepancy. These results could suggest that CTA may slightly underestimate vessel caliber or possibly discrepancies between external vessel diameter (as measured in anatomic dissections) compared to intraluminal width (as detected on CTA). In our analysis we also found that the diameters for the CSA and TDA were significantly wider when these vessels arose from the AA compared to the SSA.

The length of the subscapular trunk can have important implications for recipient vessel selection. Measured SSA lengths on CTA from our cohort were consistent with reported figures from anatomic dissections (mean = 2.5 cm, range = 0.4-6.3 cm).^3,4,8 Lhuaire et al found that 9/80 dissections revealed a longer SSA that originated more proximally on the AA and was therefore significantly longer than the majority of SSAs, which originated from the third part of the AA as classically described.⁸ While they proposed that this vascular configuration should be considered a distinct sub-classification, our analysis suggest that SSA length has a wide, right-skewed distribution.

There were several limitations to our study. In our analysis, CTA proved to have insufficient resolution to capture the length and course of the TDA completely. Distally, this vessel becomes quite small in diameter and is very circuitous. The contrast diminishes in the distal diminutive portion of the vessel, and there is variation from patient to patient depending on the contrast bolus, so a confident assessment can be challenging. To measure the length, one would have to “straighten” out the vessel in silico, which would yield inconsistent results depending on the degree of opacification across the entire vessel in a given patient. Furthermore, we were unable to delineate the venous anatomy through the use of CTA due to the timing of the contrast bolus infusion during image acquisition. Additionally, this study did not evaluate patients with head and neck cancer who were candidates for a composite reconstruction, so the clinical utility of CTA for preoperative planning could not be directly assessed. While the goal of the present study was to characterize the SSA-TDA arterial anatomy in detail with CTA, it is possible that non-arterial phase contrast-enhanced CT imaging of the chest may be sensitive enough to detect the presence or absence of a subscapular trunk. An interesting future direction of study could evaluate how well this anatomy is delineated in patients with head and neck cancer who are undergoing a contrast-enhanced chest CT as part of a pre-operative staging workup. Additional future directions of study should include a retrospective analysis that quantifies the incidence of attempted mega-flap reconstructions that ultimately required two vessel anastomoses due to the unexpected discovery of a variant CSA-TDA system. However, given the routine use of preoperative CTA for fibula free flap surgery⁹ or perforator mapping in breast reconstruction¹⁰, preoperative chest CTA may aid preoperative planning in select patients with complex reconstructive needs, particularly in the management of patients in whom a single-vessel anastomosis is critical.

Conclusion

In our study of 100 patients, one in five patients had variant subscapular-thoracodorsal arterial anatomy and lacked a subscapular arterial trunk on at least one side. These findings suggest that routine CT angiography of the chest may aid in the surgical planning of complex head and neck cancer patients who may benefit from a composite free flap, particularly patients with a vessel depleted neck.

Highlights.

• CTA Chest effectively mapped vasculature of thoracodorsal-subscapular system

• One in five patients lacked a subscapular artery unilaterally or bilaterally

• CTA Chest may guide planning for patients with complex reconstructive needs

Abbreviations:

CSA

circumflex scapular artery

CTA

computed tomography angiography

SSA

subscapular artery

TDA

thoracodorsal artery