Propensity Score-Matched Analysis of Lesion Patterns in Stroke Patients With Patent Foramen Ovale and Patients With Spontaneous Intracranial Artery Dissection

Yangyang Huang, Yifan Cheng, Bei Shao, Xuan-You Zhou, Huazheng Liang, Jianhua Zhuang, Yong BiFrontiers in Neurology 2019 · DOI

Introduction

In young and middle-aged adults, ischemic strokes often present as "cryptogenic," where the underlying cause is not immediately apparent. Two frequent contributors to these events are Patent Foramen Ovale (PFO)—a persistent opening between the heart's atria—and Spontaneous Intracranial Artery Dissection (SIAD)—a tear in the arterial wall. Distinguishing between these etiologies is critical, as they necessitate vastly different management strategies, ranging from percutaneous closure of the PFO to anticoagulation or endovascular intervention for SIAD.

Previous comparative studies have often been confounded by selection bias, as patient demographics and risk factors typically differ between these two groups. This study utilizes propensity score matching (PSM) to provide a more rigorous comparison of brain lesion patterns identified via Diffusion-Weighted Imaging (DWI).

Methodology

The researchers conducted a retrospective analysis of 69 patients (40 with PFO and 29 with SIAD) treated at the First Affiliated Hospital of Wenzhou Medical University between 2010 and 2018. To mitigate baseline differences, a propensity score-matching analysis was performed. Patients were matched based on eight key variables: age, sex, hypertension, diabetes mellitus, hypercholesterolemia, smoking status, prior stroke history, and NIHSS scores.

Following the matching process, 21 pairs of patients were analyzed. The primary focus was the distribution of acute ischemic lesions on DWI, categorized into specific patterns: single lesions (cortical, subcortical, or corticosubcortical) and multiple lesions (within one or multiple vascular territories).

Figure 2: Examples of single lesion patterns (Pattern 1 and 2) typically observed in PFO-related strokes.

Key Findings

The study revealed significant differences in the spatial distribution of infarcts between the two groups:

  • Single Lesions (Cortical/Subcortical): Isolated lesions were significantly more prevalent in the PFO group (57%) compared to the SIAD group (19%, $P = 0.026$).
  • Multiple Lesions in One Territory: This pattern was a hallmark of SIAD, occurring in 43% of cases compared to only 10% in the PFO group ($P = 0.035$).
  • Vascular Territory: SIAD was more frequently associated with posterior circulation involvement.
Figure 3: Multiple lesions clustered within a single vascular territory, a pattern more characteristic of SIAD.

Discussion: "Small Plug" vs. "Embolic Shower"

The observed patterns align with the hypothesized pathophysiological mechanisms. PFO-related strokes are thought to result from paradoxical embolism, where small venous thrombi bypass the pulmonary filter and act as isolated "plugs" in the cerebral vasculature. In contrast, SIAD often involves a local site of arterial injury that can lead to repeated thrombus formation or hemodynamic insufficiency, resulting in an "embolic shower" within a specific downstream territory.

The diagnostic utility of these patterns was quantified using Area Under the Curve (AUC) analysis. Single lesions yielded an AUC of 0.647 for predicting PFO, while multiple lesions in a single territory yielded an AUC of 0.632 for SIAD. While these imaging "signatures" are not definitive in isolation, they provide valuable clinical clues to guide diagnostic prioritization.

Conclusion and Limitations

This study suggests that DWI lesion patterns can serve as important indicators for the underlying cause of cryptogenic stroke in younger patients. However, the authors note several limitations, including the small sample size of the matched cohort (n=42) and the single-center retrospective design. Additionally, the reliance on 24-hour ECG monitoring may have missed paroxysmal atrial fibrillation in some patients.

Future large-scale, multi-center prospective studies are needed to validate these findings and further refine the imaging-based differentiation of PFO and SIAD.

Summary updated

Abstract

<b>Aims:</b> To investigate the relationship between clinical and imaging features of stroke patients with patent foramen ovale (PFO) and those with spontaneous intracranial artery dissection (SIAD). <b>Materials and methods:</b> We retrospectively examined both clinical and imaging results of 40 stroke patients with PFO and 29 with SIAD. To reduce selection bias, we conducted a propensity score-matching analysis. The patients' propensity scores were estimated using a logistic regression model based on the following variables: age, sex, hypertension, diabetes mellitus, hypercholesterolemia, cigarette smoking, stroke histories, and their NIHSS scores. We compared the pattern of cerebral DWI lesions between patients with PFO and those with SIAD. <b>Results:</b> After propensity score matching, 21 pairs of patients were selected. Clinical characteristics of the 2 groups were well matched. The distribution of DWI lesion patterns differed between the 2 groups. Single lesions (cortical or subcortical) were more frequently observed in the PFO group than in the SIAD group (<i>P</i> = 0.026). Multiple lesions in one vascular territory occurred more frequently in the SIAD group than in the PFO group (<i>P</i> = 0.035). <b>Conclusion:</b> The present study suggests that lesion patterns observed from DWI of patients with PFO and SIAD might provide clues to the etiology of infarcts. Single lesions (cortical or subcortical) might be a typical feature of PFO associated strokes, while multiple lesions in one vascular territory might be a specific feature of SIAD associated strokes.
BibTeX
@inproceedings{huang2019propensity,
  author = {Yangyang Huang and Yifan Cheng and Bei Shao and Xuan-You Zhou and Huazheng Liang and Jianhua Zhuang and Yong Bi},
  title = {{Propensity Score-Matched Analysis of Lesion Patterns in Stroke Patients With Patent Foramen Ovale and Patients With Spontaneous Intracranial Artery Dissection}},
  booktitle = {Frontiers in Neurology},
  year = {2019},
  doi = {10.3389/fneur.2019.00418},
}
Figure