In our study, we found high grade, ipsilateral arterial pathologies in around 23% of all patients with TIA diagnosed by acute CTA and a need for invasive treatment in around 7% of cases. These numbers indicate that CTA in the acute setting of TIA is justified in selected patient subgroups. As predictors, we identified cardiovascular disease and a short period of symptoms which might be useful for patient selection. In contrast, established TIA-scores turned out to be not predictive for high risk vascular pathologies.
TIA and ischemic stroke, both indistinguishable in the acute setting in most cases, require vascular imaging. Besides CTA, there are other imaging methods to be considered for neurovascular status evaluation like magnetic resonance imaging and neurovascular ultrasound. Neurovascular ultrasound of the brain supplying arteries has a high spatial resolution and is able to detect atherosclerotic changes, as well as thrombi in the proximal carotid artery very well. However, it is limited by investigator dependency and availability. Neurovascular Ultrasound in the acute setting can also result in treatment delays due to longer investigation times and patient factors like incompliance, anatomical problems like insufficient transtemporal doppler window and the fact that distal intracranial arteries as well as distal parts of the internal carotid arteries cannot be visualized.  Therefore, neurovascular ultrasound is not performed on a usual basis in the acute phase of ischemic stroke if fast and targeted therapy like IVT and EVT is available. Magnetic resonance imaging (MRI) is one of the most important methods to assess the brain tissue. However, it has been shown that CT combined with CTA provides a comparable quality to diffusion-weighted MRI in stroke patients. Considering the fact that MRI resources are more limited compared to CT, which is available 24/7 in all stroke centres, also is significantly faster and has a high sensitivity for the diagnosis of vessel occlusions and haemorrhages, CT imaging is the technology of choice in the acute setting of stroke.
The evaluated ABCD2-score, ABCD3-score and SPI-II, that stratify the risk for stroke after TIA, were not predictive for ipsilateral vascular pathologies or the need for invasive treatment in our study. This observation corresponds to findings from Lou et al. concerning the ABCD2-Score as a possible predictor for intervention. In their study, including 121 patients with TIA, they found that the ABCD2-Score was equally distributed in patients no matter if they received an intervention or not. Reasons for the missing predictive value of these scores might be, that TIA scores have been developed to predict recurrence rates of TIA or stroke in those patients. Many factors incorporated in these scores (admission blood pressure, age, diabetes mellitus and heart failure) were equally distributed in our cohort, raising the suspicion, that these factors are likely to be usable to quantify instability (risk of recurrence) of the underlying pathology, but are not predictive for the underlying pathology itself. Also, most items are related to the neurologic symptoms presented at admission. As the most important predictors for vascular disease like chronic arterial hypertension, high heart rate, smoking habit and other vascular diseases are not considered in these scores, this might explain the lack of prediction. [22, 23]
Our results advise to consider a certain risk profile if to decide to perform a CTA or not. Except the fact that patients with acute stroke symptoms should undergo immediate CTA. Patients with coronary heart disease or obesity and therefore a high risk for macroangiopathic vessel disease and TIA should receive CTA to diagnose high risk vascular pathologies and to initiate targeted therapy. CTA in this case could contribute to a time gain to diagnosis and could contribute to reduce the risk for TIA recurrence or a manifest ischemic stroke. For example, patients with symptomatic high-grade stenosis and poor vascular flow and collateralisation status are at high risk to suffer from early disabling stroke.  A CTA could reveal the stenosis and vascular flow. Therefore, it possibly affects treatment decisions. Another study found de novo dual antiplatelet therapy (OR 3.24) as a predictor for spontaneous recanalization in patients with symptomatic acute extracranial carotid occlusion.  As well dual antiplatelet therapy reduces the risk of recurrent stroke after TIA or minor stroke. [26–28] Information like these could be gained by CTA and cause a vital impact on treatment. Nevertheless, currently there is no known study to the authors which displays a direct link between early CTA and a benefit from early or careful treatment. It can only be assumed and needs further research.
A short duration of Symptoms in TIA patients is likely to be associated with macroangiopathic arterial disease of brain supplying arteries and can be found in most patients with carotid artery stenosis. A prototype of a short lasting neurologic deficit associated with carotid artery stenosis is the Amaurosis fugax, a special form of TIA.[29, 30] Already in the seventies Pessin et al. and Harrison et al. found that TIA-patients with a symptom duration under one hour were more likely to have a carotid artery stenosis and should get an angiography.[31, 32] These findings go in line with our finding, that a short duration of symptoms increases the risk for a high grade, ipsilateral vascular pathology and the need for an invasive treatment in patients with TIA. In contrast, cardiac emboli are more likely to cause permanent and more severe focal neurological deficits.[33–35] The difference in duration of symptoms caused by arterial or cardiac emboli could be explained by a difference in spontaneous revascularization due to a variation of constitution and size of the thrombi.[32, 36, 37] Another point to consider is the collateralisation status in patients with arterial emboli caused by vessel stenosis compared to cardiac embolization with no associated vessel stenosis. Arterial stenosis formation take month to years inducing intracranial collateralization, which is not present in cardiac embolization. This fact could explain short symptom duration in patients with arterial stenosis, where hemodynamics can adapt in a few seconds. Therefore, short duration of symptom should not be discarded as trivial by the clinician, but raise the suspicion of macroangiopathic disease with associated findings in CTA. TIA in crescendo was previously shown to significantly correlate with large artery atherosclerosis as the origin of TIA.  This correlation couldn’t be confirmed by our observations, e. g. table 1.
A contralateral motor weakness is more often seen in patients needing an intervention than in the other subgroups,and therefore it is predicting this circumstance. This symptom is highly associated with the anterior circulation, which is affected in 97.5% of all cases in the subgroup that needed an intervention.
Strength of our study is the high number of patients included in our study cohort as well as the use of “real word” data from a large volume, tertiary stroke centre. Limitations include the monocentric, retrospective design of the study. Moreover, a selection bias must be assumed, as only 66% of patients with TIA received a CTA in the emergency department. The reason for the decision to perform or not to perform a CTA in the acute setting could be influenced by multiple factors like preference of the consultant neurologist in charge, patient factors like kidney disease and medical history of the patient. Another weakness is the lack of MRI data, therefore a mixture of TIA and minor-stroke patients is to assume for our population. If more MRI data would have been available, the results of our study would possibly change. But this displays everyday praxis and the current lack of MRI imaging even in large stroke centers. 
Beside the shown results our study indicates further research on important information’s like the follow up on patients and how CTA influences their therapy and outcome compared to other imaging methods like neurovascular ultrasound.