Causative Classification System for Ischemic Stroke (CCS)

Introduction

Causative Classification of Ischemic Stroke, CCS, is an automated version of the SSS-TOAST system, which is an evidence-based system that harmonizes multiple aspects of the diagnostic stroke evaluation in a regularized manner to devise the most likely mechanism in the presence of multiple potential causes. It has been developed to maximize inter-examiner reliability in stroke classification while utilizing the full diagnostic data set from a typical “stroke work-up”. The automated CCS system limits inter-examiner variability in interpretation of stroke-related characteristics, ensures consistency in data entry, and thereby, further adds in SSS-TOAST's reliability in etiologic stroke classification.

The CCS system uses standard computer languages used for content distribution and user interaction through the Internet: HyperText Markup Language (HTML), Cascading Style Sheets (CSS) and JavaScript. It uses JavaScript to handle the logic-check part of the application (input error checking, automatic disabling and enabling of dependent elements, automatic checking and un-checking of dependent elements, and the calculation of the resulting classification including a description of the decision for a particular subtype). Pop-up windows are used to guide users by providing additional information about the content. The application could be run as client-side application only or as client-server application, which allows integration into a database to keep track of multiple entries and to perform statistical analyses.

The CCS offers subtype information in two different formats: the causative subtype and phenotypic subtype. Identification of the causative subtype requires integration of multiple aspects of ischemic stroke evaluation including symptom characteristics, vascular risk factors, diagnostic test results, response to treatment, and prognosis. In other words, designation of the causative subtype requires a decision-making process. For instance, the diagnosis of small vessel occlusion requires not only the presence of a lacunar infarct in a clinically relevant location but also exclusion of conditions such as dissection, atherosclerosis, vasculitis, or vasospasm of the parent artery at the origin of the penetrating artery. Phenotypic subtypes, on the other hand, are a documentation of abnormal test findings. The process of phenotypic subtyping does not require any judgment on the part of the clinician-investigator. For instance, in a patient with a lacunar infarction in the pons, multiple stereotypic lacunar TIAs during the preceding days, patent foramen ovale, and moderate to severe stenosis in the origin of one of the vertebral arteries, causative CCS subtype is “probable small vessel occlusion”, whereas phenotypic subtype is “large artery atherosclerosis + lacunar infarction + cardiac embolism”. The CCS contains four phenotypic subtypes (large artery atherosclerosis, cardiac embolism, lacunar infarction, other uncommon causes). There are four possible states for large artery atherosclerosis and cardioembolism (major, minor, absent, incomplete evaluation), three for lacunar infarction (major, absent, incomplete evaluation), and two for “other uncommon causes” group (major, absent). The last category (other uncommon causes) does not include incomplete evaluation because there is no standard minimum diagnostic evaluation for diverse array of stroke etiologies included in this category. Overall, the CCS system offers 96 phenotypic combinations.

The CCS system classifies stroke etiology in 5 major causative categories: Supra-aortic large artery atherosclerosis, cardio-aortic embolism, small artery occlusion, other uncommon causes, and undetermined causes. The undetermined group is further divided into 4 subcategories as cryptogenic-embolism, unknown, incomplete evaluation, and unclassified groups. Some of the important criteria used in causative subtype assignment in the CCS system are listed below:

1. Each etiologic category in the CCS system is subdivided as “evident”, “probable”, or “possible” based on the weight of available evidence.
2. A mechanism is deemed to be “evident” only if it is the sole potential mechanism that carries an established risk for stroke.
3. A mechanism is classified as “possible” if it is associated with either low or uncertain risk for stroke and there is no “evident” cause of stroke.
4. The distinction between “evident” and “possible” mechanisms is made by using a 2% annual or one-time primary stroke risk threshold. The 2% primary risk criterion applies to all subtypes for which there is primary risk data. Additional criteria used to define evident and possible mechanisms are summarized in the Table.
5. When there is more than one “evident” stroke mechanism, the CCS system regularizes assignment to a “probable” stroke mechanism based upon the presence of specific characteristics of the stroke that make one mechanism more probable than the others. Probable features are defined using the following criteria:
   1. The positive likelihood ratio (defined as the probability that a person with a given stroke subtype will have a particular clinical or imaging feature divided by the probability that a person with no such mechanism will have the same clinical or imaging features) is used to describe the strength of associations among clinical and imaging features and particular stroke mechanisms. Features with a positive likelihood ratio of greater than a cut-off of 2 qualify a stroke mechanism as probable. These features are: 1) For supra-aortic large artery atherosclerosis: a) prior history of one or more transient monocular blindness, TIA, or stroke in the territory of index atherosclerotic artery within the last month, b) internal watershed infarction, c) multiple, temporally separate infarcts exclusively within the territory of the stenotic artery. 2) For cardio-aortic embolism: a) history of systemic embolism, b) multiple acute infarctions in both anterior circulations or in both anterior and posterior circulation. 3) For small artery disease: a) stereotypic lacunar TIAs within the last week, b) presentation with a lacunar syndrome, c) typical lacunar infarct in a relevant brain location.
   2. The presence of a temporal relationship between an etiology and the onset of stroke qualifies the etiology as the probable mechanism. Examples include stroke following cardiac or vascular surgery, acute myocardial infarction, arterial dissection, and the use of certain drugs.
   3. The presence of a spatial relationship between brain infarct and its vascular or cardiac cause qualifies the etiology as probable. Examples include multiple infarcts in both hemispheres at the setting of proximal cardiac or aortic sources, and a focal pathology in the parent artery at the origin of the penetrating artery supplying the territory of a lacunar infarct.
6. The category of “other uncommon causes” includes patients with a diverse array of stroke mechanisms. The CCS system regularizes other causes into two groups with respect to their temporal and spatial relationship with brain infarct.
   1. “Other uncommon causes” that bear clear and close temporal or spatial relationship include: acute arterial dissection, cerebral vasculitis, cerebral venous thrombosis, acute disseminated intravascular coagulation, heparin-induced thrombocytopenia type II, hypoperfusion syndromes, iatrogenic causes, thrombosed intracranial aneurysm, meningitis, primary infection of the arterial wall, thrombotic thrombocytopenic purpura - hemolytic uremic syndrome, Moyamoya disease, and segmental vasoconstriction or vasospasm. When disorders in this category coexist with another evident cause (for which there is no probable criterion), the CCS assigns the subtype as “probable other”. For example, the CCS subtype will be “probable other” in a patient with atrial fibrillation and cerebral vasculitis.
   2. “Other uncommon causes” that do not bear clear and close temporal or spatial relationship are chronic arterial dissection, fibromuscular dysplasia, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, hyperviscosity syndromes, MELAS, primary antiphospholipid antibody syndrome, sickle cell disease, Sneddon`s syndrome, abnormalities of thrombosis and hemostasis, and other causes. When disorders in this category coexist with another evident cause (for which there is no probable criterion), the CCS assigns the subtype as “undetermined unclassified.” For example, in a patient with carotid territory stroke, Sneddon`s syndrome, and ipsilateral carotid stenosis >50%, the CCS subtype will be “undetermined unclassified”.
7. There are 2 disorders in the category of “other uncommon causes” that are considered as diagnosis of exclusion. These are “drug-induced” and “migraine-related” stroke. They do not influence subtype assignments in the presence of another attributable cause. For example, in a patient with history of cocaine use and left atrial myxoma, the CCS subtype assignment will be evident cardio-aortic embolism.
8. The term “multiple acute ischemic lesions” in data entry field-2d includes any combination of acute infarct, perfusion defect, and non-chronic clinically-relevant abrupt arterial occlusion. For example, a patient with an acute infarct in the right MCA territory and a region of abnormal perfusion in the left MCA territory is considered to have multiple acute ischemic lesions. Other examples include abrupt occlusion (cut-off) in the left PCA and acute infarct in the right MCA territory or abrupt occlusion in the left MCA and hypoperfusion in the right superior cerebellar artery territory. Multiple chronic infarcts do not qualify for data entry field 2d.
9. The CCS requires identification of “clinically relevant artery” in several data entry fields. Clinically relevant artery is defined as the artery that supplies the territory where the infarct or ischemia has occurred. For example, occlusion or severe stenosis of right internal carotid artery in a patient with an absent left A1 segment of the anterior cerebral artery can cause multiple infarcts in both right and left hemispheres. In this circumstance, the right internal carotid artery is the clinically-relevant artery for all of the infarcts.
10. The CCS system is constructed on a basic level of etiologic investigations. These include imaging of the brain, imaging of the vessels, cardiac evaluation, and specific blood and CSF tests based on the level of suspect from a particular cause. Each investigation is specific for one particular subtype: “Brain imaging” is for evident small artery occlusion, “vascular imaging” for evident large artery atherosclerosis, and “cardiac evaluation” for evident cardio-aortic embolism. Failure to perform etiologic investigations in the absence of positive test results for any particular etiology is considered “incomplete evaluation”. Failure to investigate for an etiology in the presence of its probable criteria is also considered to be “incomplete evaluation”. For instance, in a patient with atrial fibrillation and unilateral internal watershed infarcts (probable criterion for large artery atherosclerosis) but no angiographic evaluation, the CCS classification will be “incomplete evaluation”.
11. The CCS system tones down the level of confidence from “evident” to “probable” or from “evident” to “possible” if etiologic investigations are stopped (or not done) when a positive test result is obtained based on potential importance of the missing test result on subtype assignment. Some of the formulations used for this in the CCS include:
    1. Input: X evident (>50% stenosis in the ipsilateral internal carotid artery) + Y incomplete evaluation (no cardiac evaluation)
       Output = X possible (“possible” supra-aortic large artery atherosclerosis)
    2. Input: X evident (>50% stenosis in the ipsilateral internal carotid artery) + X probable (ipsilateral internal watershed infarcts) + Y incomplete evaluation (no cardiac evaluation)
       Output: X probable (“probable” supra-aortic large artery atherosclerosis)
12. The CCS software allows the entry of high- or low-risk cardiac sources based on clinical assessment, EKG, or cardiac examination even when it is the clinical judgment that cardiac evaluation is incomplete because of the absence of echocardiography. In other words, the system requires echocardiography (or other more advanced cardiac investigations) only when there is clinical suspicion of cardiac embolism and if clinical history, cardiac examination, and EKG do not reveal a source. For instance, echocardiography is not a prerequisite for stroke subtyping in a patient with atrial fibrillation diagnosed by EKG or a patient with prior history of rheumatic mitral valve stenosis. The CCS allows classification of such patients with known sources of cardiac embolism into relevant subtypes regardless of the expert opinion for the need for echocardiography.
13. Complex aortic atheroma (CAA) is grouped under “large artery atherosclerosis” in the phenotypic classification, and “cardio-aortic embolism” in the causative classification. Inclusion of CAA in the cardiac category in the causative classification provides the system the ability to avoid considering large artery atherosclerosis as the evident mechanism when there are simultaneous acute infarcts in multiple circulations or concurrent systemic embolism (as aorta could be the source) as well as to prevent considering all proximal sources of multiple emboli as cardiac when indeed the aorta cannot be ruled out as the source. The CCS provides the flexibility to classify complex aortic arch atheroma as an independent subset in the database function of the system so that it can also be categorized with large artery atherosclerosis when needed.