American Heart Association

diagnosis and imaging

Cortical Venous Filling in Dynamic CTA Predicts Clinical Outcome in Proximal MCA Occlusion

Danny R. Rose, Jr., MD

As the scope and availability of endovascular treatment for large vessel occlusions in patients with ischemic stroke continues to increase, it is important for providers to have the ability to screen and predict which patients will benefit from these therapies. The most effective approach has been a topic of much debate and research, including utilizing ancillary testing to account for individual variations in collateral circulation and other factors that could potentially extend treatment indications and predict outcomes. Nearly all endovascular treatment trials use some combination of noncontrast head CT (NCCT), CT angiography (CTA) and/or CT perfusion (CTP) imaging to assist in clinical decision making due to the speed and availability of such scans. However, recent studies found that the currently utilized modalities may be insufficient to predict treatment effect or outcome. Wijngaard et al. sought to use cortical venous filling (CVF), a potential marker for collateral extent and perfusion, to assess whether the  extent or velocity of CVF as obtained through dynamic CTA predicts clinical outcome at 3 months in ischemic stroke patients with proximal middle cerebral artery occlusion.

All patients received a NCCT evaluated for early ischemic changes (ASPECTS); CTP/CTA was used for clot burden score, collateral status, cerebral-blood-flow (CBF), cerebral-blood-volume (CBV), mean-transit-time (MTT) and time-to-peak (TTP). Using the dynamic CTAHo, CVF was assessed visually—cortical venous contrast opacification and the number of cortical and anastomotic veins were evaluated in comparison with the contralateral hemisphere. The number of seconds to reach first CVF (appearance of any cortical vein draining into the superior sagittal sinus), optimal CVF (maximum contrast opacification of all cortical veins), and end of venous filling (complete absence of contrast) was measured. The velocity of CVF was calculated as the median differences between venous filling of the affected versus non-affected hemisphere. In addition, the extent of CVF was defined as either “good” or “poor,” based on comparing the extent of cortical vein filling at optimal CVF with the unaffected hemisphere using 50% as a cutoff.

A total of 88 patients were selected from two Dutch medical centers from previously published studies (DUST and MR-CLEAN). The mean age of participants was 67 years; median NIHSS was 15 and 50% of the participants were women. The median time from onset to imaging was 77 minutes. Most patients (59%) were treated with IV thrombolysis alone, with 25% of patients receiving intra-arterial therapy. At 3 months, 54 patients (61%) had a poor clinical outcome (mRS 3-6), and 22 of 88 patients (25%) had died. Venous predictors of poor outcome at 3 months were: poor extent of optimal CVF (RR 1.8; 95% CI 1.4-2.4) and slow velocity of optimal CVF (RR 1.6; 95% CI 1.1-2.2). Logistic regression models were utilized, using a layered approach in which successive models included more variables. Model 1 included age, NIHSS, treatment and ASPECT score, model 2 included CBF, model 3 included collateral status, and model 4 included extent and velocity of CVF. The addition of venous filling parameters resulted in a significant (p<0.02) improvement in prediction of clinical outcome. Although only 22 patients underwent mechanical thrombectomy, good CVF parameters (good extent and fast CVF) predicted good clinical outcomes in patients with mTICI 2b and 3 reperfusion (RR 3.3; 95% CI 1.1-10.6, p<0.05), compared with patient groups with poor CVF and/or poor perfusion status combined.

This study represents a novel use of venous filling parameters in predicting clinical outcome after ischemic stroke, resulting in a more accurate prediction when used in combination with previously studied measures. Based on these preliminary findings, one of the more promising applications may be its utility in predicting treatment effect of mechanical thrombectomy, but the small number of patients examined requires evaluation in a larger, prospective trial. If supported in additional research, CVF has the potential to become an important part in future trial design with respect to selecting patients for endovascular intervention. The authors’ choice to utilize dynamic CTA, however, limits its widespread application, as the more ubiquitous single-phase CTA cannot obtain time-resolved measurements. Parameters designed for use in single-phase CTA could be evaluated in a future study for its potential utility in facilities without dynamic CTA.  

Cerebral Microbleeds and Intracerebral Hemorrhage

Russell Mitesh Cerejo, MD

Pasquini M, Benedictus MR, Boulouis G, Rossi C, Dequatre-Ponchelle N, Cordonnier C. Incident Cerebral Microbleeds in a Cohort of Intracerebral Hemorrhage. Stroke. 2016

The authors studied prognostic factors of incident cerebral microbleeds (CMBs) in the PITCH study primary intracerebral hemorrhage (ICH) cohort with at least 2 MRIs and survival for 6 months post event, stratifying the findings according to the index ICH location. 

Amongst 168 included patients (median age 64 years) with ICH, 53% had at least 1 CMB. 

Anatomical distribution and number of CMBs were similar in patients with lobar and non-lobar ICH. Incident CMBs appeared in 80 patients (48%) in a median delay of 3.4 years (including 29 patients without CMBs at ICH onset), with a total number of 285 new CMBs, leading to an incidence rate of 14.2 per 100 person-years. Patients with CMBs at ICH onset had an almost 2.5 fold increased risk of developing incident CMBs during the follow-up period compared with patients without CMBs at ICH onset (aOR 2.27; 95% CI 1.18-4.35). In non-lobar ICH, incident CMBs were associated with incident lacunes (aOR 2.86; 95%CI 1.04-7.85) and with the use of antiplatelet agents during follow-up (aOR 2.89; 95%CI 1.14-7.32), while in lobar ICH incident CMBs were associated with incident radiological macro-hemorrhages (aOR 9.76; 95%CI 1.07-88.77). However, the proportion of antiplatelet agents use was similar in lobar and non-lobar ICH (31% and 32% respectively; p=0.89). 

This study provides long term MRI data and a better understanding into the possible etiologies of primary ICH as related to CMBs, which may be interesting surrogate markers.

Is a CT Within 6 Hours of Headache Onset Enough to Rule Out SAH?

Alexander E. Merkler, MD

Dubosh NM, Bellolio MF, Rabinstein AA, Edlow, JA. Sensitivity of Early Brain Computed Tomography to Exclude Aneurysmal Subarachnoid Hemorrhage: A Systematic Review and Meta-Analysis. Stroke. 2016

Subarachnoid hemorrhage (SAH) is the most devastating type of stroke – 50% of survivors are dead within six months and among those patients who survive, only 50% return to their previous level of functioning. For decades, classic neurology dogma has stated that in order to rule out a SAH, any patient who presents with a thunderclap headache (HA) must receive a lumbar puncture (LP) if the head CT is negative. However, recent data suggests that in neurologically intact patients, a CT is 100% sensitive to rule out SAH when performed within six hours using a modern generation CT scanner (16-slice or greater). Hence, is there is no longer a need to perform an LP after a negative head CT that is performed within six hours of HA onset?

In this manuscript, Dr. Dubosh et al perform a meta-analysis to determine the sensitivity of modern generation CT scanners to rule out SAH in patients presenting to an emergency department within six hours of thunderclap HA. The authors identified five articles that met their inclusion criteria; four were retrospective and one was prospective. In total, 8,907 patients with thunderclap HA underwent a CT within six hours.

Overall, thirteen out of the 8,907 patients had a missed SAH. The overall sensitivity of CT was 0.987 (95% CI 0.971-0.994), specificity was 0.999 (95% CI 0.993-1.0) and the likelihood ratio of a negative CT was 0.010 (95% CI 0.003-0.034). This equated to a miss rate of 1.5 per 1000 patients who present with thunderclap HA and receive a modern CT scan within six hours.

It is important to note that each of the five studies had certain limitations. For example, perimesencephalic hemorrhage1 and SAH caused by a cervical arteriovenous malformation2 were considered missed causes of SAH. In addition, in the one prospective study by Perry et al3 (in which there were no documented missed cases of SAH), an LP was not performed in every patient who presented with thunderclap HA and had a negative CT. Although there was close follow-up using telephone interviews and monitoring coroner’s records, there may have been missed cases of SAH.

Modern CT performed within 6 hours of patients presenting with thunderclap HA is an extremely sensitive tool to rule-out SAH. As with most tests, it is impossible to say that it is 100% sensitive, but it certainly approaches it. Although perhaps very few cases of SAH may be missed, clinicians must weigh this against the potential consequences of performing an LP including time, anxiety, post-LP complications, unnecessary vascular imaging (CTA, MRA, angiography) and probably most importantly subsequent ramifications such as inappropriate procedures for incidentally found vascular lesions. Of course, missing a SAH may be life threatening and can lead to significant consequences including death.


1. Blok KM, Rinkel GJ, Majoie CB, Hendrikse J, Braaksma M, Tijssen CC et al. CT within 6 hours of headache onset to rule out subarachnoid hemorrhage in nonacademic hospitals. Neurology. 2015;12:1927-193.

2. Backes D, Rinkel GJ, Kemperman H, Linn FH, Vergouwen MD. Time-dependent test characteristics of head computed tomography in patients suspected of nontraumatic subarachnoid hemorrhage. Stroke. 2012;43:2115-2119.

3. Perry JJ, Stiell IG, Sivilotti ML, Bullard MJ, Emond M, Symington C et al. Sensitivity of computed tomography performed within six hours of onset of headache for diagnosis of subarachnoid haemorrhage: prospective cohort study. BMJ. 2011;343:d4277.

Thrombus Length Estimation on Delayed Gadolinium-Enhanced T1

Russell Mitesh Cerejo, MD

Yan S, Chen Q, Xu M, Sun J, Liebeskind DS, Lou M. Thrombus Length Estimation on Delayed Gadolinium–Enhanced T1. Stroke. 2016

The authors, in keeping with their previous studies on thrombus morphology on MRI, have found a new method to assess for thrombus length. They further correlated this thrombus length with recanalization and outcomes with IV thrombolysis. They evaluated patients over ~5 years with acute stroke due to M1 occlusion, received IV thrombolysis and had MRI pre and post thrombolysis. 

They obtained delayed gadolinium-enhanced T1 (dGE-T1) after perfusion imaging and co-registered the imaging with the time of flight MRA. Thrombus length on dGE-T1 was 8.18 ± 4.56 mm (range 1.63-30.26 mm). Recanalization occurred in 38 (51.4%) patients and no recanalization in 36 (48.6%). Thrombus length on dGE-T1 was significantly longer in patients without recanalization in univariate analysis. Thrombus length and M1 occlusion were independent predictors for no recanalization of MCA. Thrombus length of 6.77 mm, yielded a sensitivity of 77.8% and a specificity of 57.9%, and odds ratio 4.81 (95% CI: 1.742 to 13.292; p=0.002). No one achieved recanalization after IVT when length of thrombus exceeded 14 mm on dGE-T1 (see figure).

This study highlights the utility of MRI in assessing thrombus length without increasing the scan time or the contrast load. This may also be useful in acute stroke therapy to determine thrombus length and success of mechanical thrombectomy. 

Arterial Spin Labeling MRI Estimation of Antegrade and Collateral Flow in Unilateral Middle Cerebral Artery Stenosis

Russell Mitesh Cerejo, MD

Lyu J, Ma N, Liebeskind DS, Wang DJJ, Ma L, Xu Y, et al. Arterial Spin Labeling Magnetic Resonance Imaging Estimation of Antegrade and Collateral Flow in Unilateral Middle Cerebral Artery Stenosis. Stroke. 2016

The authors in their above titled paper describe a novel way to non-invasively assess for

antegrade and collateral blood flow in intracranial stenosis patients. They evaluated 41 consecutive patients with symptomatic intracranial stenosis of the M1 segment of the middle cerebral artery due to atherosclerosis. These patients underwent three-dimensional pseudo-continuous arterial spin labeling (3D pCASL) with 3.0T MRI. The calculated perfusion on the CBF map of post labeling delay (PLD) 1.5s as early-arriving flow, and perfusion on the CBF map of PLD 2.5s as combination of early-arriving flow, late-arriving antegrade flow and late-arriving retrograde flow. The mean early arriving flow proportion was 78.3%±14.9%. The mean late-arriving retrograde flow proportion was 16.1%±10.2%. Half patients underwent cerebral angiography with calculation of Modified TICI scale and ASITN/SIR collateral grade.

The authors found significant correlations between early-arriving flow and late-arriving flow on two-PLD pCASL with conventional angiographic antegrade and collateral scales, suggesting that the early-arriving flow and late-arriving retrograde flow to the territory supplied by the stenotic MCA may primarily represent antegrade and collateral flow, respectively.

This is an interesting study, which has tried to use a novel non-invasive technique to assess for collaterals that are important not only in chronic stenoses but also acute occlusions.

Intracranial Arterial Remodeling and the Risk of Stroke

Alexander E. Merkler, MD

Qiao Y, Anwar Z, Intrapiromkul J, Liu L, Zeiler SR, Leigh R, et al. Patterns and Implications of Intracranial Arterial Remodeling in Stroke Patients. Stroke. 2016

Intracranial atherosclerotic disease (ICAD) is a common cause of stroke; patients with ICAD face high rates of recurrent stroke despite aggressive medical and lifestyle modification. Currently, the diagnosis of ICAD is based on the degree of vessel stenosis, but perhaps, as Drs. Qiao et al discuss, this is not the best measure of either plaque burden, or more importantly, stroke risk.

Based on coronary plaque research, Drs. Qiao et al studied the impact of vessel remodeling in ICAD. Although plaque may lead to hemodynamic stenosis, remodeling of the vessel may either preserve (positive remodeling) or further impair (negative remodeling) the degree of stenosis. In addition, although positive remodeling may preserve the vessel lumen, it may make plaque more vulnerable to rupture or lead to clinical symptoms.

In this study, the authors used High-resolution black blood MRI (BBMRI) to assess vessel remodeling in ICAD and its association with ischemic events. Forty-five patients with ICAD with >50% stenosis in a large intracranial artery who also had a stroke or TIA in the distribution referable to that stenosis were included. The authors identified 137 plaques, of which 56 exhibited positive remodeling, 53 negative remodeling, and 28 intermediate remodeling. There was higher burden of plaque within the posterior circulation as compared to the anterior circulation and positive remodeling was more frequent in the posterior circulation (58% vs 31%). Furthermore, positive remodeling was associated with a trend towards culprit plaque classification. Finally, the authors found that the lumen begins to narrow (ie when remodeling fails to preserve the lumen) when plaque burden reached 55.3%.

This study challenges the current theory that ICAD is purely based on degree of vessel stenosis. As in the coronary vessels, remodeling occurs when plaque is present and may either augment or decrease the vessel lumen size. Furthermore, intracranial remodeling may lead to an underestimation of plaque burden (based on the current definition of vessel stenosis >50%), and therefore an underdiagnosis of ICAD, particularly in the posterior circulation. Further research to assess remodeling and its impact on stroke risk is warranted.

Assessing Mismatch Using DWI and FLAIR Predicts Favorable Outcome Following Recanalization With IV-tPA

Jay Shah, MD

Legrand L, Tisserand M, Turc G, Edjlali M, Calvet D, Trystram D, et al. Fluid-Attenuated Inversion Recovery Vascular Hyperintensities–Diffusion-Weighted Imaging Mismatch Identifies Acute Stroke Patients Most Likely to Benefit From Recanalization. Stroke. 2016

Perfusion-Diffusion mismatch on MRI has been proposed to select ischemic stroke patients for revascularization therapy. However, this strategy is time consuming and requires gadolinium. The authors previously have reported using mismatch between Fluid-attenuated inversion recovery (FLAIR) vascular hyperintensities (FVH) and diffusion for penumbral evaluation. FVH represent slow retrograde flow in leptomeningeal collaterals and are thought to represent impaired but viable tissue. The authors hypothesize that recanalization after IV-tPA would have better outcomes within FVH-DWI mismatch patients than non-mismatch patients. 

This study was a retrospective analysis of a prospective registry of patients treated exclusively with standard IV-tPA dosing for acute stroke between 2004-14. Other inclusion criteria included proximal M1 occlusion, pre-treatment and 24-hour follow-up MRI, and 3 month modified rankin scale (mRS) score. FVH-DWI mismatch was considered present when FVH extended beyond boundaries of the cortical DWI lesion. In total, 164 patients were included in the analysis. 121 patients had FVH-DWI mismatch. Complete recanalization occurred in 50 patients. Association between recanalization and favorable outcome was significant in patients with FVH-DWI mismatch (OR= 16.2). 

This study shows that DWI and FLAIR images can identify patients who are more likely to benefit from recanalization and the authors propose that this modality can be used as a surrogate to perfusion imaging. However, it is clearly understood that clinical outcomes in acute ischemic stroke is strongly associated with recanalization of the occluded artery. Non-mismatch patients also demonstrated benefit, albeit to a lesser degree and therefore revascularization should not be withheld for a perceived lack of benefit. Furthermore, all patients within this study had M1 occlusion but did not undergo endovascular intervention which is now established as the standard of care. For such patients, if recanalization can be achieved according to guideline recommendations, there should not be a need for further penumbral evaluation. However, in patients with a prolonged presentation or an unknown time of onset, assessment of mismatch could provide utility in selecting appropriate patients. Further studies should focus on this patient population.

The MRI Spot Sign: A Marker of Clinical outcomes in Patients with Primary ICH

Peggy Nguyen, MD

Schindlbeck KA, Santaella A, Galinovic I, Krause T, Rocco A, Nolte CH, et al. Spot Sign in Acute Intracerebral Hemorrhage in Dynamic T1-Weighted Magnetic Resonance Imaging. Stroke. 2016

The spot sign on CTA has been previously correlated with hematoma expansion, mortality, and poor clinical outcomes in patients with primary ICH. In practice, the CTA spot sign is likely to be of greatest relevance, given hyperacute stroke imaging is still largely predicated on CT imaging. However, MR imaging has increasingly been used early in the course of acute stroke imaging, but there is no equivalent sign that has been validated. Here, the authors report on an equivalent MR spot sign on contrast enhance T1 weighted imaging in 50 consecutive primary ICH patients presenting within 24 hours of an acute stroke syndrome. 

Contrast enhancement within the hematoma on MR (spot sign) was demonstrated in 23 of 50 patients (46%) with primary ICH. Larger spot signs were seen with larger hematomas and correlated with the outcome based on mRS; specifically when spot signs were dichotomized as large (> 1 mL) vs small (< 1mL), large spot signs were characterized by larger hematoma volumes (36 mL vs 5 mL) and worse outcomes (median mRS 5). When patients were dichotomized according to presence or absence of a spot sign, patients with the spot sign had worse outcomes (median mRS 4) despite similar NIHSS on admission. On follow up imaging, however, no significant difference was seen in regards to hematoma expansion between those who had a spot sign and those who did not.

Although the MR spot sign was not demonstrated to be predictive of hematoma expansion, it was predictive of clinical outcome. Given that MR allows evaluation of hematoma age and often provides additional information on possible etiology of hemorrhage, the validation of an equivalent spot sign on MR adds another tool to the arsenal of MR interpretation for ICH.

Prediction models for functional outcome in acute ischemic stroke patients. Closer to the truth?

Luciana Catanese, MD

Leiva-Salinas C, Patrie JT, Xin W, Michel P, Jovin T, and Wintermark M. Prediction of Early Arterial Recanalization and Tissue Fate in the Selection of Patients With the Greatest Potential to Benefit From Intravenous Tissue-Type Plasminogen Activator. Stroke. 2016

To date, there are no available prediction models to guide physicians in the accurate selection of ischemic stroke patients that will benefit from acute recanalization therapies. CT perfusion (CTP) parameters such as ischemic core volume, penumbra and mismatch profile have shown promising results but not independent from recanalization rates. Therefore, the authors aimed to determine whether the prediction of functional outcomes in acute ischemic stroke (AIS) patients was more accurate when using prediction models including CTP and recanalization parameters versus those without.

Five different hypothetical prognosis algorithms were designed for this purpose, one based on whether or not patients received tPA, a second one based on ASPECTS score of >=7 versus <7, a third one based on the site of occlusion, a fourth one based on volume of ischemic core and penumbra and a fifth one based on a matrix of predicted recanalization (using the ‘iSTROKEMD’ application) and volume of ischemic core and penumbra at baseline. Comparison of sensitivity, specificity, positive and negative predictive value as well as accuracy was made between such algorithms to predict good clinical outcomes, defined as mRS of 0-2 at 90 days.

Retrospective data from 173 patients taken from a repository published previously who were >=18 years, presented within 4.5 hours from symptom onset, had available CT, CTA and CTP on admission as well as CTA or MRA between 1 and 48 hours and who were considered for tPA but did not undergo endovascular therapy were analyzed. Overall, about half of patients were male (median age ~70), had a median ASPECT score of 7, ~50% M1 occlusions and 71.6 % received tPA (54% recanalization versus 34.6% without). Good outcome inversely correlated with admission NIHSS and was overrepresented in those treated with tPA when adjusting for age and baseline NIHSS, as seen in prior studies. Overall, the algorithm that combined the a-priori calculation of the prediction of recanalization with the volume of baseline PCT ischemic core and penumbra was the most accurate in predicting good outcomes with an accuracy of 77.7%. The following table displays the statistical measures of the performance of the different approaches in predicting functional outcomes.

The combination of predicted recanalization and perfusion parameters may be superior in prognosticating good clinical outcomes when compared to other predictors in isolation. However, considering the evolving nature of both CTP thresholds and softwares in accurately and universally measuring ischemic core and penumbra volumes, the small and non-randomized sample, restrospective analysis and lack of inclusion of patients with and without endovascular recanalization, among others, the interpretation of these results is limited. Successful recanalization and baseline tissue fate seem to influence outcomes in AIS patients that undergo recanalization. Closer to the truth but not quite there yet.

Magnetic resonance “black blood” thrombus imaging found to have high sensitivity and specificity for cerebral venous sinus thrombosis in small prospective study

Danny R. Rose, Jr., MD

Yang Q, Duan J, Fan Z, Qu X, Xie Y, Nguyen C, et al. Early Detection and Quantification of Cerebral Venous Thrombosis by Magnetic Resonance Black-Blood Thrombus Imaging. Stroke. 2016

Cerebral venous sinus thrombosis (CVST) is a relatively rare form of stroke that typically affects young individuals. Improved methods of diagnosis and treatment have significantly improved outcomes of CVST, but current indirect noninvasive imaging that relies on visualizing venous flow perturbation can lead to delays that may result in increased morbidity and mortality. Magnetic resonance direct thrombus imaging (MRDTI) may improve upon this through its ability to visualize thrombus directly by exploiting the short T1 relaxation time of methemoglobin within the thrombus. To date there has been little study using this imaging modality in the cerebral venous system. Yang et al. sought to investigate the ability of MRDTI (using a so-called “black-blood MR technique” or “MRBTI”) to detect CVST in a prospective study comparing this technique to traditional diagnostic imaging.

Forty-seven patients with signs and symptoms suspicious for CVST were recruited for the study. All patients received CT, MR and Time of Flight MR venography, with five patients also receiving a contrast enhanced CT venogram as well. Two readers performed a consensus reading of all conventional imaging studies with full clinical and outcome information, with a total of 14 venous segments being assessed for thrombi based on the presence or absence of intraluminal filling defects. MRBTI was performed using 3D variable-flip-angle turbo spin echo on a 3.0T scanner. MRBTI images were randomized and presented to two independent readers who were blinded to the diagnostic/therapeutic management of the patients, clinical information and conventional imaging data. Image quality was rated on a 4 point scale with 4 being excellent with no relevant artifacts. The presence or absence of thrombus was recorded for each of the 14 segments assessed via conventional imaging. Patients with CVST detected by MRBTI were divided into two groups based on duration of clinical onset: ≤ 7 days and between 7-30 days.

With conventional imaging, a total of 116 thrombosed venous segments were identified among 23 patients. MRBTI correctly identified 113 out of 116 segments, resulting in a sensitivity of 97.4%. In 527 out of 531 segments, CVST was ruled out correctly with a specificity of 99.3%. Using patient based parameters MRBTI correctly identified all 23 patients with CVST. One false-positive and zero false-negatives were identified, resulting in a sensitivity of 100% and specificity of 95.8%. There were no issues with image quality—647 of 658 segments (98%) were diagnostic according to the 4 point rating scale, and the overall image quality score was 3.5±0.6. Signal-to-noise and contrast-to-noise ratio, another metric of image quality comparing the signal intensity of the thrombus to that of adjacent tissue, was found to have significant differences in all signal measurements taken.

Utilizing MRBTI in CVST was found to approach the high diagnostic accuracy of conventional imaging in this small prospective study. Patients who stand to benefit most include those with renal impairment, which may serve as a contraindication to contrasted imaging, as well as pregnant patients and others where radiation exposure should be avoided. The utility of MRBTI may also extend to monitoring for thrombus progression/resolution, as this imaging modality is able to calculate thrombus volume as well. One significant limitation of the study that limits its widespread use is the use of a 3.0T MRI, which is seldom found outside of academic institutions and tertiary care centers. Other limitations to this study stem primarily stem from the use of a single center and small sample size, thus validation of this study in a multicenter prospective study with a larger enrollment is crucial.