Comparison of three diagnostic imaging modalities for use in diabetic inferior arterial lesions

Abstract

BACKGROUND

Underdiagnosis of peripheral arterial disease results in inadequate treatment and more serious consequences. Hence, clinicians have focused on early diagnosis and treatment.

AIM

To investigate the effectiveness of the combination of doppler ultrasonography (DUS), three-dimensional dynamic contrast-enhanced magnetic resonance angiography (CE-MRA), and CT angiography (CTA) in assessing lower extremity arterial disease in diabetes mellitus (DM).

METHODS

This study retrospectively analyzed the imaging and clinical data of 116 patients diagnosed with DM complicated with lower extremity vascular diseases from January 2021 to June 2023. All patients underwent unilateral or bilateral DUS, CTA, and CE-MRA as well as invasive digital subtraction angiography (DSA). The application values of DUS, CE-MRA, and CTA were compared.

RESULTS

A total of 152 lower extremity arteries in the 116 patients were graded following the classification of vascular branches. The Kappa values between DUS and DSA were 0.780, 0.755, and 0.806 for diagnosing moderate stenosis and 0.484, 0.699, and 0.449 for severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries, respectively. The Kappa values between CE-MRA and DSA were 0.784, 0.814, and 0.835 for diagnosing moderate stenosis and 0.694, 0.748, and 0.606 for severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries, respectively. The Kappa values between CTA and DSA were 0.900, 0.858, and 0.878 for diagnosing moderate stenosis and 0.882, 0.823, and 0.756 for severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries, respectively.

CONCLUSION

DUS, CE-MRA, and CTA demonstrated comparable accuracy in diagnosing lower extremity arterial disease in DM, and the consistency between CTA and DSA diagnoses was higher than the other two imaging methods.

Key Words: Doppler ultrasonography; Three-dimensional dynamic contrast-enhanced magnetic resonance angiography; Computerized tomography angiography; Diabetes; Lower-extremity arterial disease

Core Tip: Diabetes mellitus is a major risk factor for lower extremity peripheral arterial disease, and lower extremity complications are prevalent and increasing globally. However, peripheral arterial disease underdiagnosis results in undertreatment and subsequent complications with more serious consequences. Therefore, in this study we analyzed the diagnostic value of color doppler ultrasonography, magnetic resonance multimodal imaging, and CT angiography in diabetes mellitus complicated with lower limb arterial disease using digital subtraction angiography as the gold standard for diagnosing this disease.

INTRODUCTION

Peripheral arterial disease (PAD) of the lower limbs is a local lesion and is a local manifestation of systemic atherosclerosis[1]. Partial or complete occlusion of one or more arteries of the lower extremities primarily causes PAD and is associated with prevalent cardiovascular risk factors[2]. Individuals may be asymptomatic; however, symptoms typically involve intermittent claudication and chronic limb-threatening ischemia (with resting pain, ulceration, or gangrene)[3]. PAD of the lower extremities is endemic globally due to an aging population and an increase in cardiovascular risk factors such as hypertension, dyslipidemia, smoking, and type 2 diabetes mellitus (DM)[4].

DM is a major risk factor for PAD; thus, it is more prevalent in patients with PAD than in those with no PAD[5]. Further, PAD is one of the major complications of DM, and symptomatic PAD adversely affects the quality of life of patients[6]. Lower extremity complications are prevalent and are increasing globally, affecting approximately 131 million people with an estimated global prevalence of 1.8%[7]. They significantly increase morbidity and mortality in DM, resulting in leg ulcers and amputations in some cases, and are characterized by physical disability, decreased productivity, and mood disturbances. Neuropathy is a crucial cause of leg ulcers and amputations, and PAD is equally important[8]. However, PAD underdiagnosis results in inadequate treatment and more serious consequences. Therefore, clinicians have focused on early diagnosis and treatment.

Digital subtraction angiography (DSA) is the gold standard for PAD diagnosis[9]. However, DSA is an invasive test that causes complications in approximately 1.5% of patients, with a high cost of testing, which limits its clinical use. Further, non-invasive imaging examination is the most important assessment method for lower extremity vascular diseases. Considering the continuous development of medical technology, imaging modalities mainly include doppler ultrasonography (DUS)[10], three-dimensional (3D) dynamic contrast-enhanced magnetic resonance angiography (CE-MRA)[11], and CT angiography (CTA)[12]. Each modality demonstrates its own merits, with no comprehensive assessment criterion for selecting a particular approach. Hence, the novelty of the study was to comprehensively evaluate non-invasive diagnostic methods in selected patients with DM complicated with lower extremity arterial disease (LEAD) and assessed the efficacy of DUS, CE-MRA, and CTA to help in the early diagnosis and treatment of LEAD in patients with DM in the future.

MATERIALS AND METHODS

Study population

This study retrospectively analyzed the imaging and clinical data of patients with DM complicated with lower extremity vascular diseases diagnosed at Chongqing Kang Hua Zhong Lian Cardiovascular Hospital from January 2021 to June 2023. Inclusion criteria were: (1) The clinical diagnosis of DM-induced LEAD, with the main types of lower extremity arterial stenosis and occlusion; (2) All patients with pale lower limb skin and local numbness and pain or obvious signs of dyskinesia before inspection; (3) All patients who underwent unilateral or bilateral DUS, CTA, and CE-MRA as well as invasive DSA; and (4) Complete clinical and imaging data. Exclusion criteria were: (1) Severe coagulation dysfunction; (2) Malignancies of other systems; (3) Autoimmune diseases; (4) Neurological dysfunction; (5) Acute or chronic infectious diseases; and (6) Missing clinical and imaging data. The control group included another 60 healthy individuals who concurrently underwent physical examinations. A total of 116 eligible patients were finally included, with 152 lower limb arteries. Figure 1 illustrates the flow chart of the study.

Figure 1 Flow chart of the study. DM: Diabetes mellitus; LEAD: Lower extremity vascular diseases; DUS: Doppler ultrasonography; CTA: CT angiography; CE-MRA: Contrast-enhanced magnetic resonance angiography; DSA: Digital subtraction angiography; FBG: Fasting blood glucose; 2hPG: 2-h postprandial blood glucose; HbA1c: Glycosylated hemoglobin.

Data collection

General data (age, gender, underlying diseases, disease course, etc.) of patients and controls were obtained by reviewing the electronic medical records. Biochemical indicators, including fasting blood glucose, 2-h postprandial blood glucose, and glycosylated hemoglobin, were acquired.

Inspection methods

DSA: The Philips Allura Xper FD20 angiography system was utilized. Seldinger’s technique was used to puncture the affected or contralateral radial artery or femoral artery, and retrograde or anterograde intubation was conducted until reaching the affected limb artery for routine DSA.

DUS: DUS was performed using the Philips IE22 ultrasound system. The lower limb arteries of the patients in a supine or prone position were observed through ultrasound images, including lumen stenosis, lumen occlusion, and plaques, and the smoothness of the vascular lumen–intima was analyzed.

CTA: Somatom Definition AS 128-slice 64-row CT scanner (Siemens, Germany) was employed. The patient was placed in a dorsal elevated and supine position for scanning from the abdomen to the dorsum pedis. The automatic bolus tracking technique was used for scanning after injecting 80 mL of iopromide (300 mg/mL), a contrast medium, into the anterior elbow vein of the patient with a high-pressure syringe at a flow rate of 4.0 mL/sec and 30 mL of normal saline administration at 4.0 mL/sec. After completing the scanning process, the images during the optimal diastolic and systolic phases were selected for automatic reconstruction.

CE-MRA: CE-MRA was performed with the German Siemens Skyra 3.0T magnetic resonance imaging machine. In the first step, with the patient in a supine or dorsal elevated position, 3D rapid gradient-echo imaging was performed. Gadobenate dimeglumine (20 mL) was pushed through the anterior cubital vein with a high-pressure syringe at a speed of 3.0 mL/sec. The second step was to push normal saline (30 mL) with the same injection rate. The area from the dorsal artery to the abdominal aorta was scanned using automatic table movement and segmentation techniques and time-resolved imaging of contrast kinetics. The image acquisition involved three segments, namely from the calf to the dorsal foot artery, from the abdominal aorta to the pelvic artery, and from the femoral artery to the popliteal artery.

Evaluation methods

To obtain the highest quality images, a team consisting of one senior and one junior physician in the Angiography Unit of the Department of Ultrasound, who had received specialized training in standardized examination of lower extremity arteries, performed all ultrasound or imaging assessments in this study. The examination procedures and points were consistent. Further, two radiologists with > 10 years of work experience conducted image assessments. Vascular branches involve three grades: Grade 1 indicates the superior geniculate artery, including the superficial femoral artery, external iliac artery, and common iliac artery; grade 2 denotes the inferior genicular artery, including the popliteal artery, posterior tibial artery, anterior tibial artery, and peroneal artery; grade 3 represents the dorsal aorta. Vascular lesions were classified as mild stenosis (< 50% stenosis), moderate stenosis (50%-75% stenosis), and severe stenosis or occlusion (> 75% stenosis).

Statistical analysis

Statistical Package for the Social Sciences version 25.0 statistical software was used to sort out and analyze data. Count data were presented as cases or percentages, and the χ² test was conducted for between-group comparisons. Continuous data were all expressed as mean ± SD, and the comparison of means between groups was performed using the t-test. Considering DSA diagnosis as the gold standard, the diagnostic efficiency of the three diagnostic modalities for LEAD was assessed, including sensitivity, accuracy, and specificity. The Kappa consistency test was conducted to assess the consistency of three diagnostic methods in diagnosing LEAD.

RESULTS

General data

The DM group consisted of 54 males and 62 females, with an average age of 53.62 ± 3.37 years and an average course of 7.35 ± 0.41 years. Patients with DM demonstrated statistically higher fasting blood glucose, 2-h postprandial blood glucose, and glycosylated hemoglobin levels than healthy individuals (P < 0.05; Table 1).

Table 1 Comparison of general information.

Characteristics	DM group (n = 116)	Control group (n = 60)	χ2/t	P value
Age (year)	53.62 ± 3.37	53.78 ± 3.09	0.307	0.759
Gender			2.158	0.142
Male	54	21
Female	62	39
Hyperlipidemia	41	19	0.238	0.626
Hypertension	54	23	1.085	0.297
Coronary heart disease	38	16	0.690	0.406
Course of DM (year)	7.35 ± 0.41
Fasting blood glucose (mmol/L)	8.17 ± 1.04	4.36 ± 0.59	26.253	< 0.0001
2-h postprandial blood glucose (mmol/L)	12.87 ± 1.48	7.01 ± 0.63	29.296	< 0.0001
Glycosylated hemoglobin (%)	9.81 ± 0.67	5.41 ± 0.52	44.399	< 0.0001

DM: Diabetes mellitus.

DSA findings

Table 2 shows the DSA diagnosis of 152 lower extremity arteries in 116 patients with DM based on arterial branch classification. A total of 78, 51, and 23 grade 1 arteries, 50, 53, and 49 grade 2 arteries, and 85, 47, and 20 grade 3 arteries demonstrated mild, moderate, and severe stenosis or occlusion, respectively. Figure 2 illustrates the imaging diagnosis results of different arterial branches in patients with DM complicated with LEAD.

Figure 2 Patient imaging findings. A: Doppler ultrasonography showed occlusion of the lateral femoral circumflex artery; B: Contrast-enhanced magnetic resonance angiography diagnosis results; C: Volume rendering after CT angiography examination. LCFA: Lateral circumflex femoral artery.

Table 2 Digital subtraction angiography diagnosis results of patients.

Arterial branch classification	Mild stenosis	Moderate stenosis	Severe stenosis or occlusion	Total
Grade 1	78	51	23	152
Grade 2	50	53	49	152
Grade 3	85	47	20	152

Diagnostic value of DUS in DM complicated with LEAD

The sensitivity of DUS in diagnosing moderate stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries were 86.3%, 84.9%, and 91.5%, with specificity of 92.1%, 90.9%, and 91.4% and accuracy of 90.1%, 88.8%, and 91.4%, respectively. The Kappa values between DUS and DSA for diagnosing moderate stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries were 0.780, 0.755, and 0.806, respectively.

The sensitivity of DUS in diagnosing severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries was 82.6%, 79.6%, and 70.0%, with specificity of 82.2%, 90.3%, and 86.4% and accuracy of 82.2%, 86.8%, and 84.2%, respectively. The Kappa values of DUS vs DSA in diagnosing severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries were 0.484, 0.699, and 0.449, respectively (Table 3).

Table 3 Diagnostic value of doppler ultrasonography for different arterial branches in diabetic lower extremity arterial disease.

Degree of stenosis	Arterial branch classification	DUS diagnosis	DSA diagnosis		Kappa	Sensitivity (%)	Specificity (%)	Accuracy (%)
			Moderate/severe	Non-moderate/non-severe
Moderate stenosis	Grade 1	Moderate	44	8	0.780	86.3	92.1	90.1
		Non-moderate	7	93
	Grade 2	Moderate	45	9	0.755	84.9	90.9	88.8
		Non-moderate	8	90
	Grade 3	Moderate	43	9	0.806	91.5	91.4	91.4
		Non-moderate	4	96
Severe stenosis or occlusion	Grade 1	Severe	18	23	0.484	82.6	82.2	82.2
		Non-severe	5	106
	Grade 2	Severe	39	10	0.699	79.6	90.3	86.8
		Non-severe	10	93
	Grade 3	Severe	14	18	0.449	70.0	86.4	84.2
		Non-severe	6	114

DUS: Doppler ultrasonography; DSA: Digital subtraction angiography.

Diagnostic value of CE-MRA in DM complicated with LEAD

The sensitivity of CE-MRA in diagnosing grade 1 arteries, grade 2 arteries, and grade 3 arteries with moderate stenosis were 90.2%, 90.6%, and 93.6%, with specificity of 90.1%, 91.9%, and 92.4% and accuracy of 90.1%, 91.4%, and 92.8%, respectively. The Kappa values between CE-MRA and DSA in diagnosing moderate stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries were 0.784, 0.814, and 0.835, respectively.

The sensitivity of CE-MRA in diagnosing severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries was 82.6%, 85.7%, and 80.0%, with specificity of 93.0%, 90.3%, and 90.9% and accuracy of 91.4%, 88.8%, and 89.5%, respectively. The Kappa values for CE-MRA diagnosis of severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries were 0.694, 0.748, and 0.606, respectively (Table 4).

Table 4 Value of contrast-enhanced magnetic resonance angiography in diagnosing diabetic lower extremity arterial disease of arterial branches of different grades.

Degree of stenosis	Arterial branch classification	CE-MRA diagnosis	DSA diagnosis		Kappa	Sensitivity (%)	Specificity (%)	Accuracy (%)
			Moderate/severe	Non-moderate/non-severe
Moderate stenosis	Grade 1	Moderate	46	10	0.784	90.2	90.1	90.1
		Non-moderate	5	91
	Grade 2	Moderate	48	8	0.814	90.6	91.9	91.4
		Non-moderate	5	91
	Grade 3	Moderate	44	8	0.835	93.6	92.4	92.8
		Non-moderate	3	97
Severe stenosis or occlusion	Grade 1	Severe	19	9	0.694	82.6	93.0	91.4
		Non-severe	4	120
	Grade 2	Severe	42	10	0.748	85.7	90.3	88.8
		Non-severe	7	93
	Grade 3	Severe	16	12	0.606	80.0	90.9	89.5
		Non-severe	4	120

CE-MRA: Contrast-enhanced magnetic resonance angiography; DSA: Digital subtraction angiography.

Diagnostic value of CTA in DM complicated with LEAD

The sensitivity of CTA in diagnosing moderate stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries was 100.0%, 94.3%, and 93.6%, with specificity of 93.1%, 92.9%, and 95.2% and accuracy of 95.4%, 93.4%, and 94.7%, respectively. The Kappa values between CTA and DSA for diagnosing moderate stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries were 0.900, 0.858, and 0.878, respectively.

The sensitivity of CTA in diagnosing severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries was 100.0%, 91.8%, and 85.0%, with specificity of 96.1%, 92.2%, and 95.5% and accuracy of 96.7%, 92.1%, and 94.1%, respectively. The Kappa values between CTA and DSA for diagnosing grade 1 arteries, grade 2 arteries, and grade 3 arteries with stenosis were 0.882, 0.823, and 0.756, respectively (Table 5).

Table 5 Diagnostic value of computerized tomography angiography for different arterial branches in diabetic lower-extremity arterial disease.

Degree of stenosis	Arterial branch classification	CTA diagnosis	DSA diagnosis		Kappa	Sensitivity (%)	Specificity (%)	Accuracy (%)
			Moderate/sever	Non-moderate/non-severe
Moderate stenosis	Grade 1	Moderate	51	7	0.900	100.0	93.1	95.4
		Non-moderate	0	94
	Grade 2	Moderate	50	7	0.858	94.3	92.9	93.4
		Non-moderate	3	92
	Grade 3	Moderate	44	5	0.878	93.6	95.2	94.7
		Non-moderate	3	100
Severe stenosis or occlusion	Grade 1	Severe	23	5	0.882	100.0	96.1	96.7
		Non-severe	0	124
	Grade 2	Severe	45	8	0.823	91.8	92.2	92.1
		Non-severe	4	95
	Grade 3	Severe	17	6	0.756	85.0	95.5	94.1
		Non-severe	3	126

CTA: Computerized tomography angiography; DSA: Digital subtraction angiography.

DISCUSSION

The incidence of DM is rapidly increasing. More than 700 million individuals globally are predicted to suffer from DM by 2045[13]. LEAD is a complication of DM and exhibits no obvious clinical symptoms in the early stage[14], making it an important public health problem. DSA is the gold standard for the clinical diagnosis of LEAD, but its invasiveness may cause vascular injury and other complications that are unacceptable to patients.

DUS, CE-MRA, and CTA, all non-invasive imaging modalities, have their merits and demerits in diagnosing LEAD comorbid with DM. In imaging arterial diseases are characterized by a lipid-rich necrotic core, intraplaque hemorrhage, calcification, etc.[15]. The proximal superficial femoral artery is the most predominant site of femoral artery stenosis (41.1%)[16], followed by the common femoral artery. The most common compression site of the popliteal artery is the gastrocnemius head[17]. The most predominantly seen site of anterior tibial artery damage or occlusion is the distal two-thirds, causing dorsalis pedis artery ischemia and blood redistribution through posterior tibial artery anastomosis[18].

In this study the sensitivity of DUS and CE-MRA for severe stenosis of grade 3 arteries was lower than that of grade 1 arteries and grade 2 arteries. CTA exhibited high diagnostic efficiency for both patients with moderate and severe stenosis. CE-MRA provides good visualization and grading technique for arterial stenosis. One of the advantages of MRA is its non-invasive nature and is less operator-dependent than DUS and develops a comprehensive image of the entire artery. Unlike DSA and CT, MRA does not involve radiation exposure. However, the contrast agent used in MRA causes kidney diseases. In addition, intravenous contrast utilization induces mild adverse effects in patients with renal failure.

DUS is a superior vascular mapping tool because it is cost-effective and eliminates complications associated with iodine contrast agents and ionizing radiation, especially for patients with renal diseases[19]. Further, DUS is an important tool for detecting lipid-rich plaques[20], lumen stenosis, and arterial wall thickness that is used to diagnose focal stent stenosis[21].

CTA is another excellent technique to provide high-resolution tissue contrast imaging in vascular diseases. It utilizes an iodinated contrast agent, like DSA, but is less invasive with a low risk of kidney disease caused by contrast agents and radiation[22]. Further, benefits that outweigh the risks can still be used in patients with kidney disease[23].

However, as previously noted, different approaches have their advantages and disadvantages. MRA is better at mapping calcified arteries and reducing artifacts compared with CTA[24]. In the popliteal artery, CTA can overcome the anatomical limitations of the popliteal fossa (gastrocnemius muscle) observed in DUS and MRA[25].

The results of this study indicated that the accuracy of CTA in diagnosing moderate and severe stenosis of grade 1 arteries, grade 2 arteries, and grade 3 arteries has a high consistency with the DSA diagnostic results. Thus, the results indicated that CTA has very high diagnostic efficiency, consistent with the research results of most scholars at home and abroad[26,27]. This is because CTA can display changes in lower limb varicose veins, especially in the venous equilibrium phase. Further, with higher resolution CTA effectively reflects the situation of vascular calcification. Therefore, to improve the diagnostic value of DM complicated with LEAD, appropriate examination methods should be selected in clinical practice. We believe that CTA should be utilized in the examination of patients with DM with severe stenosis of grade 3 lower extremity arteries.

Foreign studies indicated that DSA should only be utilized as a pretreatment diagnosis and non-invasive imaging as a routine assessment[28], indicating that non-invasive imaging examination has significant guiding value in clinical applications. Another study revealed that DUS demonstrated a diagnostic accuracy of up to 95% for moderate and severe anastomosed or occluded lower limb arteries[29]. In this study, the accuracy of DUS for moderate stenosis arteries and the accuracy for severe stenosis arteries were all < 95% reported in the aforementioned studies. This may be closely associated with the personal operation and experience of imaging physicians.

The cost-effectiveness of different imaging methods in clinical practice still needs to be considered. De Vries et al[30] conducted a randomized trial and revealed that the use of CE-MRA over standard DUS for the initial workup of PAD reduced the number of additional imaging procedures by 42%, with no difference in total cost. However, MRA is more expensive and less broadly available in contrast to carotid DUS and CTA[31]. Further, 3D approaches are more precise than 2-dimensional ultrasonography, and the costly expense of 3D imaging equipment and software to assess a 3D image may restrict its practical use[32]. In conclusion, in selecting a non-invasive diagnostic method that improves diagnostic accuracy, a combination of all three can be used, but only if the use of a single diagnostic method is not clearly identifiable and if economic conditions allow.

Limitations of this study should be acknowledged. The included participants were limited and exclusively inpatients; thus, the single-center small sample size may affect the generalizability of the results. Further, some important characteristics of patients may not be recorded due to the nature of the retrospective study. In the future, a large sample size prospective multicenter study is warranted to further investigate the application of DUS, CE-MRA, and CTA in DM complicated with LEAD.

CONCLUSION

DUS, CE-MRA, and CTA all have high accuracy in diagnosing LEAD, and the diagnostic efficacy of CTA was superior over the other two imaging methods.