Evaluation the venous pathology of the lower extremities with triggered angiography non-contrast-enhanced sequence magnetic resonance imaging (TRANCE-MRI)

Background To explore the diagnostic performance of non-contrast-enhanced magnetic resonance imaging using triggered angiography non-contrast-enhanced sequence (TRANCE-MRI) in evaluation of venous pathology of the lower extremity. Methods This is a single-center prospective cohort study of 25 patients with suspected venous disease in their lower extremities. Each patient received a Doppler ultrasound exam before the scheduled TRANCE-MRI on a 1.5T MR scanner (Philips Ingenia, Philips Healthcare, Best, The Netherlands). The following lymphography and CTA were arranged according to the diagnostic indications. Results The venous scenarios of the 25 patient were divided as follows: 11 had deep venous thrombosis (DVT), seven had a static ulcer, three had symptomatic varicose veins (VV), two had recurrent VV after surgery, and two had lymphedema. TRANCE-MRI unexpectedly found that 4 patients (16%) had occult peripheral arterial occlusive disease. Of the 11 patients with scenario of DVT, 4 patients (36.4%) did not actually have DVT on TRANCE-MRI, and the symptoms were due to malignancy, external compression, and congenital anomalies. One patient (4%) with radiation-related lymphedema was falsely diagnosed as external iliac vein compression. Interrater agreement for DVT in the thigh between the ultrasonography and TRANCE-MRI was substantial agreement (Cohen's kappa κ, 0.72). The sensitivity, specicity and accuracy of TRANCE-MRI were 85.7%, 88/9% and 88%, respectively. Venous thrombi and collateral veins could be clearly outlined by TRANCE-MRI, including in middle femoral veins that might be dicult to detect by ultrasound. Conclusion TRANCE-MRI provided not only vascular image of the lower extremity but also information about the pelvis and abdomen. However, false positive results may occur in iliac vessels. TRANCE-MRI could outline venous thrombi and collateral veins from the abdomen to both calves, and thus, could be a powerful tool in the treatment of venous pathology in the lower extremities. 3D=three-dimensional, CT=computed tomography, CTA=computed tomography angiography, DVT=deep venous thrombosis, FOV=eld of view, GSV=great saphenous vein, IR=inversion recovery, IRB=institutional review board, MRI=magnetic resonance imaging, MRV=magnetic resonance venography, NSF=nephrogenic systemic brosis, Q-Flow=quantitative ow scan, STIR=short tau inversion recovery, TE=echo time, TOF=time-of-ight, TR=repetition time, TRANCE-MRI=triggered angiography non-contrast Enhanced MRI, TSE=turbo spin-echo, US=ultrasonography, VV=varicose vein, SU=static ulcer,


Background
Venous pathology in the lower extremities is a critical public health problem with economic and social consequences. It includes scenarios from minor varicose veins (VV) and annoying venous static ulcers (SU) to potentially deadly deep vein thrombosis (DVT). [1][2][3] Ultrasonography (US) is used for routine initial evaluation of the venous system in the lower extremities. Previously, conventional venography had been considered the gold standard for detection of DVT in patients with VV. [4,5] However, this procedure is time-consuming, invasive, and requires the use of ionizing radiation. Although computed tomographic (CT) venography is less invasive than conventional venography, it still requires contrast media plus radiation exposure, and the quality of the image is not as high as CT arteriography. [5,6] Magnetic resonance venography (MRV) with use of contrast media has been revealed as highly sensitive for detecting pathology in a variety of blood vessels when compared with conventional angiography. Magnetic resonance imaging (MRI) does not involve radiation exposure but the magnetic resonance contrast agents still have undesirable effects. Nephrogenic systemic brosis (NSF) is a common complication of using gadolinium-based contrast agents in patients with pre-existing impairment of kidney function. [7][8][9] TRiggered Angiography Non-Contrast Enhanced (TRANCE) is a technique, which exploits differences of vascular signal intensity during the cardiac cycle for subsequent image subtraction, providing not only venogram but also arteriogram without use of contrast agents.
TRANCE-MRI has been used in patients with renal insu ciency and vascular diseases; however, few applications of this technique in the venous pathology of the lower extremities can be found. [10][11][12] Methods Subjects The Institutional Review Board (IRB) of Chang Gung Memorial Hospital approved this study (IRB number: 201700389B0). We prospectively collected information on consecutive patients who had been evaluated by TRANCE-MRI for venous pathology in their lower extremities at a vascular wound care center of a tertiary hospital between April 2017 and March 2018. Patients were eligible for inclusion in the study if they had a clinical indication for computed tomography angiography (CTA) of the pelvic and leg vessels. Each patient was suspected to have venous pathology in their lower extremities. Exclusion criteria were non-MRI-compatible ferromagnetic devices and pregnancy. In addition, patients with poor compliance and patients with multiple comorbidities that prevented them from lying down for the 1-hour protocol of the TRANCE-MRI were excluded. There are 30 patients were evaluated initially. One patient was excluded due to possible pregnancy and the other was unable laydown owing to complicated spine disease. In 28 patients who scheduled of TRANCE-MRI examination, two patients were morbidly obese which could not go through the MRI scanner. One female patient failed to complete TRANCE-MRI owing to her restless legs. There were 16.7% patients (5/30) excluded from this TRANCE-MRI study. All patients received a noninvasive color Doppler exam for venous status of their lower extremities before the scheduled TRANCE-MRI. The Doppler exams performed in supine position. The femoral veins, great saphenous vein, popliteal veins and perforating vein in calves were checked. Pelvic veins were not included in the Doppler exams.
Lymphoscintigraphy byTc-99m phytate and CTA were arranged according to the diagnostic indications.
We used ultrasonography and TRANCE-MRI to assess DVT in the thigh. Cohen's kappa coe cient was used to measure interrater agreement between ultrasonography and TRANCE-MRI. To evaluate the sensitivity, speci city, and accuracy of TRANCE-MRI, ultrasound was determined as true condition because it already exists in the gold standard.

MRI acquisition
MR imaging was acquired with a 1.5T MR scanner (Philips Ingenia, Philips Healthcare, Best, The Netherlands). Patients underwent imaging in a supine position and with Peripheral Pulse Unit (PPU) trigger. All images of the arterial systems were evaluated by three-dimensional (3D) turbo spin-echo (TSE) at systole and diastolic period. In the imaging with the TSE TRANCE, the following ranges of parameters were used: repetition time (TR), 1 beats; echo time (TE), shortest; ip angle, 90°; voxel size, 1.7 x 1.7 x 3mm eld of view (FOV), 350 x 420. In systole, arterial blood is owing fast. This causes dephasing of the signal and leads to ow voids; thus, the arteries were black with systolic triggering. In diastole, blood ow in the arteries is slow. The signal does not dephase; thus, the arteries were bright on the diastolic scans. Subtraction of the two phased scans will make up a 3D data set with only arteries. Another one images of the venous systems were evaluated by 3D TSE Short tau inversion recovery (STIR) at systole period. In the imaging with the TSE STIR TRANCE, the following ranges of parameters were used: TR, 1 beats; TE, 85; inversion recovery (IR) delay time, 160; voxel size, 1.7 x 1.7 x 4mm; FOV, 360 x 320. STIR provides extra background suppression because fat and bones are also suppressed. With systolic triggering, the arteries were black. The result was a dataset using TRANCE-MRI in the venous system in which no subtraction was required. The Quantitative Flow (Q-Flow) scan was routinely performed to determine the appropriate trigger delay times for systolic and diastolic triggering. The Figure 1 summarizes the principle of TRANCE-MRI technique. All the images were acquired without use of gadolinium contrast medium ( Figure 2). The TRANCE-MRI protocol requires 60 minutes for imaging acquisition, 25 minutes for MRV, and 35 minutes for MRA.

Results
Between April 2017 and March 2018, 25 patients were enrolled in this study and evaluated using TRANCE-MRI for venous pathology in their lower extremities at a vascular wound center in a tertiary hospital. These 25 patients were classi ed into four different venous scenarios: deep vein thrombosis (DVT), static ulcer (SU), symptomatic varicose veins (symptomatic VV) and recurrent varicose veins after venous surgery (recurrent VV). For example, those patients present with typical clinical gures (sudden swelling legs, history of recent trauma and surgery) and physical examination (dark skin color, pitting edema, no cracking / aking change on the skin) . Their venous scenarios were classi ed as DVT whilst enrolled into this study. The descriptive characteristics of this population are listed in Table 1. The venous scenarios from the vascular wound center were DVT in 11 patients (44%; Figure 3), static ulcers in seven patients (28%), symptomatic VV in three patients (12%), recurrent VV after surgery in two patients (8%), and possible lymphedema in two patients (8%). Nine patients had undergone relevant surgeries, such as stripping of the great saphenous vein (GSV), truncal ablation of the GSV, axillo-femoral arterial bypass, hip replacement, ap reconstruction for severe crushing injury, and hysterectomy for cervical cancer, before coming to the vascular wound center. All 25 patients were evaluated by Doppler scan for leg veins and TRANCE-MRI for both venous and arterial systems. Interrater agreement for DVT in the thigh between the ultrasonography and TRANCE-MRI was substantial agreement (Cohen's kappa κ, 0.72). The sensitivity, speci city and accuracy of TRANCE-MRI were 85.7%, 88/9% and 88%, respectively. A patient was diagnosed partial occlusion of right femoral vein on ultrasonography, but negative nding on TRANCE-MRI. Two patients had negative nding on ultrasonography, but were noticed small thrombi on TRANCE-MRI. Three patients had a CT scan with contrast media injection. One patient had suspected radiation-related lymphedema and received radionuclide lymphoscintigraphy and a venogram to con rm the diagnosis. Notably, in four patients, occult peripheral arterial occlusive diseases were found accidently by TRANCE-MRI of the arterial system. Details of the age, sex, imaging ndings of TNACE-MRI and other studies were summarized in Table 2. The patients who presented with DVT were older and may have received warfarin at the referral visit. The case 24 classi ed as having lymphedema had an asymmetric swollen thigh with hardened skin for years. Ten patients were proven to have DVT by TRANCE-MRI, including two cases of iliac vein thrombosis. Notably, TRANCE-MRI accidentally identi ed two cases of PAOD. In the remaining four patients, cases 8 through 11, venous congestions in the lower extremities were attributed to external compression and congenital anomaly. One patient, case 8, had external iliac veins that were compressed by enlarged lymph nodes and was then diagnosed with advanced prostate cancer. In cases 9 and 10, venous congestions were caused by external compression from hip prosthesis with osteomyelitis and joint effusion. The venous congestion of the lower extremities may have been caused by a double inferior vena cava in case 11 (Additional le 1).
Of the seven patients with SU, ve of them, cases 12 through 16, were related to venous pathology, including severe VV and chronic DVT. SU were caused by poor lymph drainage after complicated reconstructed surgery in case 17. In case 18, the unhealed wound was attributed to subacute occlusion of the left axillo-femoral arterial bypass. Three patients with symptomatic VV, including claudication, repeat cellulitis, and bleeding, were evaluated by TRANCE-MRI, and two of them were treated by truncal ablation of the GSV plus sclerotherapy. Cases 22 and 23 had truncal ablation and stripping of varicose GSVs at other institutions and were evaluated by TRANCE-MRI for recurrent venous claudication. Both had diffuse collateral veins in both lower limbs with occluded GSVs. Two patients were evaluated by TRANCE-MRI for lymphedema in their lower extremities. One patient, case 24, had typical radiation-related lymphedema after treatment for cervical cancer and had a venous system evaluation by TRANCE-MRI for a scheduled lymph node transfer by the plastic surgeon. TRANCE-MRI of the venous system revealed equivocal interruption of the left common iliac vein. Conventional venography and testing balloon angioplasty was performed with a 14-mm balloon. High-grade lesions were not present over the left common iliac vein and the symptom of morbid limb did not improve after angioplasty. The left thigh and calf shrank signi cantly after laparoscopic lymph node harvest and lymph node transplantation in the ankle.

Discussion
Venous pathology in the lower extremities may be suspected in patients with engorged calf veins, unhealed shadowing wounds, and asymmetrically swollen legs. [1,3,13] US is operator-dependent, timeconsuming, and inadequate at providing information about the pelvic and abdominal areas. Conventional venography has been considered the gold standard for detection of DVT in patients with VV; however, it is invasive, time-consuming, and requires the use of radiation plus contrast media. [5,14] . Although CT venography is useful for exclusion of pulmonary embolism in patients with signs of thrombosis in the legs, it still cannot replace US as the rst line image for detecting DVT. [15][16][17][18] MRI with contrast media, or time-of-ight (TOF) MRV, has been revealed as highly sensitive for detecting pathology in a variety of blood vessels when compared with conventional angiography. In TOF-MRV, blood ow is used as the intrinsic contrast agent and signal is based on an in-ow effect. The signal in the vessel depends on the ow up to a threshold speed, which is calculated by the slice thickness, in mm, divided by repetition time, in ms. However, the vessels can be observed most clearly when they are orthogonal to the twodimensional plane, because in-plane vessels sometimes experience loss of signal. [19][20][21] In summary, TOF-MRV is less invasive than conventional venography and CT venography, avoids the side effects of iodinated contrast material such as renal damage, and is less operator-dependent than US. The main disadvantage of TOF-MRV is that the FOV is small for each image acquisition and need much time to obtain a whole image of the lower extremity. MRI with gadolinium-based contrast media is an alternative and relatively rapid method for imaging lower extremity. Although MRI does not involve radiation exposure, the non-iodinated contrast agents involved still have undesirable effects. NSF is a common complication of gadolinium-based contrast agents in patients with pre-existing impairments of kidney function. [7,8] The TRANCE technique in MRI, which exploits differences in vascular signal intensity during the cardiac cycle with subsequent image subtraction, was rst described by Wedeen in 1985. [22] TRANCE-MRI has been applied in cranial neurologic diseases and arterial diseases; however, few applications of this technique in venous pathology in the lower extremities can be found. [19,[23][24][25][26] The principle of TRANCE-MRI technique is that different blood ow velocities will have different signal intensities on TSE sequence. High signal intensity, as well as bright color, re ects slow velocity, such as venous blood ow and diastolic arterial blood ow. The high velocity of systolic arterial blood ow will result in a ow void effect and is dark, as well as low signal intensity. TRANCE MRI can present high resolution and isolated vascular structures, such as arteries or veins. Presenting only the venous structure without accompanying the arterial structure is di cult to be achieved on MRI or CT with use of contrast medium, because the proper acquisition time is short and variable. Therefore, TRANCE MRI is a useful tool for venous pathology of the lower extremities because it provides additional pelvic information, no contrast agents, and no radiation toxicity. This study highlight that TRANCE MRI may be a safe and useful tool for lower extremity imaging, especially venous pathology. TRANCE-MRI may be preferred in patients with chronic renal insu ciency, history of abdominal/pelvic/orthopedic surgery and allergy to contrast media.
In this study, we do not speci cally describe how to distinguish between acute and chronic thrombosis. Distinguishing acute from chronic DVT is a potential advantage of MRI, with irregular wall thickening in the presence of collaterals and diminutive lumen suggestive of chronic DVT. Our MRI protocol provide coronal and axial images, as well as 3D MRA and MRV images. We will use the original unremoved background image to examine possible tumors or other causes of compression for all vascular lesions. (Figure 4). TRANCE-MRV showed that many subjects had equivocal interruption of the left common iliac vein, but no venous thrombosis, collateral vessels or related symptoms. This may be because the left is located between the right common iliac artery and the spine, which is an anatomically relatively narrow location ( Figure 5).
Several advantages of TRANCE-MRI application in venous pathology in the lower extremities exist. First, TRANCE-MRI provides not only images of the arteries and veins in the lower extremities but also information on the pelvis and abdomen, which is valuable in patients with a venous scenario of DVT. DVT may be mistaken as external compression of the pelvic vessels. Moreover, it is notorious as a sign of occult malignancies. Among the 11 patients with a venous scenario of DVT, four of them (36.4%) had no DVT and the symptoms were attributed to malignancy, external compression by degenerated hip prosthesis, external compression by knee effusion, and congenital anomaly. Second, the thrombi and collateral veins can be clearly outlined, including middle femoral veins that might be di cult to detect by US. This may be helpful in catheter-based thrombolytic therapy and rescue therapy in recurrent VV after truncal ablations of GSV. Finally, because TRANCE-MRI has no radiation and does not use contrast media, it is safe for patients with impaired renal function.
Compared with TRANCE-MRI, ultrasonography played a relatively small role in assessing varicose veins of the lower extremities and deep veins of the pelvis and abdomen. We still consider that ultrasound should be used preferentially when assessing venous lesions in the lower extremities because it is noninvasive and cost-effective. TRANCE-MRI is a non-invasive examination without use of contrast medium, which provide not only images of the arteries and veins in the lower extremities but also information about the pelvis and abdomen. If a patient has a pelvic vein problem or complicated varicose veins before surgery, we recommend an MRI.
We did learn of some drawbacks to TRANCE-MRI from this study. First, TRANCE-MRI of the venous system may cause a false positive in the left iliac vessels, which could be attributed to the complex anatomy and overlapping of the vessels with different blood ow directions. Other observations, such as increasing diameter and number of collateral veins, constant lling defect, and application of intravascular ultrasound, may decrease the risk of incorrect diagnosis. Second, the TRANCE-MRI protocol requires 60 minutes for imaging acquisition, 25 minutes for MRV, and 35 minutes for MRA. Thus, it is not suitable for critical and irritable patients. We suggest that the MRI protocol should be determined according to the patient's condition, and it is not necessary to perform the whole TRANCE-MRI protocol. Lastly, TRANCE-MRI is expensive and not widely used at our institution yet.
The major limitation of this investigation was that it was a nonrandomized study with few patients. This study also limited with lacking in comparison of inter-observer variability and adequate validation with other image studies. However, we attempted to identify the values and pitfalls of TRANCE-MRI in venous pathology. This was the rst prospective study to apply TRANCE-MRI for assessing venous pathology in the lower extremities. Further evaluation of pelvic/abdominal assessment and accuracy for TRANCE MRI is needed before versatile clinical applications. The TRANCE-MRI may provide more useful information regarding optimal therapeutic protocols in treating complicated vascular diseases.
Conclusions TRANCE-MRI provided not only images of the arteries and veins in the lower extremities but also information about the pelvis and abdomen. However, false positive results may occur in iliac vessels. TRANCE-MRI could outline venous thrombi and collateral veins from the abdomen to both calves, and thus, could be a powerful tool in the treatment of venous pathology in the lower extremities.

Consent for publication:
This original study was approved by the Chang Gung MedicalFoundation IRB. Written informed consent forms for publication from study participants were obtained for all patients.
Availability of data and materials: The dataset(s) supporting the conclusions of this article is(are) included within the article (and its additional le(s)).

Authors' contributions:
YKH, YHT, CHL and CWC designed the study. YKH, YHT, CHL, SCW, YCH and CWC are involved with methodology and data analysis. YKH, YHT, SCW, YCH and CWC are involved with writing the manuscript. YKH and CWC were responsible for the study conception, design, data analysis and drafting of the manuscript. All authors read and approved the nal manuscript. labeling inversion pulses: comparison of imaging with the short tau inversion recovery method and the chemical shift selective method. Magn Reson Imaging 2015, 33 (1) Figure 1 Summarized principle of TRANCE-MRI technique All images of the arterial systems were evaluated by 3D TSE sequences at systole and diastolic period. In systole, arterial blood is owing fast and leading to ow voids. Subtraction of the two phased scans will make up a 3D data set with only arteries. Another one images of the venous systems were evaluated by 3D TSE STIR at systole period. STIR provides extra background suppression because fat and bones are also suppressed.  Multiplanar MRI was helpful for comprehensive diagnosis. Our MRI protocol provide coronal and axial images, as well as 3D MRA and MRV images. (A) High signal intensity re ects slow venous (v) blood ow in TSE sequence. In contrast, the ow void effect re ects a very high velocity of systolic arterial (a) blood ow. (B) 3D TSE STIR sequence triggered in diastole shows both venous and arterial structures with background subtraction. A retroperitoneal tumor (t) was observed in the left iliopsoas region, causing venous compression but still maintaining arterial patency. Coronal (C) and axial (D) images were then be examined for comprehensive diagnosis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplement1.mp4