Document Type


Publication Date



This is the final published article from the Journal of Medical Imaging and Case Reports, 2020, 4(2): 48-53.

The full text of the article can also be found at the journal's website:

Copyright. The Authors.

Publication made possible in part by support from the Jefferson Open Access Fund


Purpose of Study: To evaluate the left superior pulmonary venous ridge length (RL) and the left atrial appendage fractal dimension (LAA FD) as predictors of cardioembolic ischemic stroke (CVA) and transient ischemic attack (TIA) in patients with atrial fibrillation (AF).

Materials, Methods and Procedures: A multicenter, retrospective multicenter study was conducted on patients with AF who underwent cardiac CT prior to catheter ablation between 01/01/2010 and 12/31/2015. Patients were grouped by the presence or absence of prior CVA/TIA. Patients with mitral stenosis, prior mechanical aortic valve replacement, ascending aortic arch atheroma, highrisk patent foramen ovale, history of atrial septal repair or device, preexisting LAA thrombus or intra-cardiac tumor, or prior open-heart surgery were excluded. Further exclusion was performed for patients with sub-optimal or unavailable cardiac CT imaging by investigators blinded to clinical data. Blinded investigators obtained RL by measuring the distance between the left superior pulmonary venous ostium and the internal ostium of the left atrial appendage. LAA FD was calculated using semi-automated volume rendering and processing software (ImageJ, Bethesda, MD) [1] and LAA FD are presented as means (95% confidence interval) and were compared between groups using unpaired t-tests. Logistic regression analysis was used to construct receiver operating curves and to assess the abilities of RL and LAA FD to predict prior CVA/TIA.

Results: 225 patients met inclusion criteria. Mean RL were 4.32 (3.80-4.93) and 5.20 (4.93-5.48) for patients with (n=24; mean age: 59.4; 70.8% male) and without (n=165; mean age: 59.3; 75.4% male) prior CVA/TIA, respectively(p=0.033). Mean LAA FD were 2.29 (95% CI: 2.24-2.34) and 2.33 (2.32-2.34) for patients with (n=22; mean age: 60.3; 68.2% male) and without (n=171; mean age: 59.3; 74.9% male) prior CVA/TIA, respectively (p=0.052). In a regression model including LAA FD, RL, and established predictive markers, only RL predicted prior CVA/TIA (OR 0.73; 0.54 to 0.98; p=0.034).

Discussion: Lower RL values were associated with prior CVA/TIA, whereas LAA FD values were similar between patients with AF with and without prior CVA/TIA. RL is a novel marker that may refine clinical decision-making regarding anticoagulation goals and treatment decisions for patients with AF. Future studies with larger samples should investigate the clinical utility of RL to improve CVA/TIA risk stratification of patients with AF and prospectively reduce the incidence of CVA/TIA in this population.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.



Included in

Radiology Commons