Suvai Gunasekaran

Suvai Gunasekaran

Auto-generated profile, based on publicly available information

Suvai Gunasekaran

Department of Radiology, Northwestern University, Chicago, IL, USA

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 March 2024 in Heliyon
Verified
Journal Article
Open access
HLhui lin
hui lin
SLSantiago López-Tapia
Santiago López-Tapia
FSFlorian Schiffers
Florian Schiffers
YWYunan Wu
Yunan Wu
SGSuvai Gunasekaran
Suvai Gunasekaran
AKAggelos K. Katsaggelos
Aggelos K. Katsaggelos
Left atrial (LA) fibrosis plays a vital role as a mediator in the progression of atrial fibrillation. 3D late gadolinium-enhancement (LGE) MRI has been proven effective in identifying LA fibrosis. Image analysis of 3D LA LGE involves manual segmentation of the LA wall, which is both lengthy and challenging. Automated segmentation poses challenges owing to the diverse intensities in data from various vendors, the limited contrast between LA and surrounding tissues, and the intricate anatomical structures of the LA. Current approaches relying on 3D networks are computationally intensive since 3D LGE MRIs and the networks are large. Regarding this issue, most researchers came up with two-stage methods: initially identifying the LA center using a scaled-down version of the MRIs and subsequently cropping the full-resolution MRIs around the LA center for final segmentation. We propose a lightweight transformer-based 3D architecture, Usformer, designed to precisely segment LA volume in a single stage, eliminating error propagation associated with suboptimal two-stage training. The transposed attention facilitates capturing the global context in large 3D volumes without significant computation requirements. Usformer outperforms the state-of-the-art supervised learning methods in terms of accuracy and speed. First, with the smallest Hausdorff Distance (HD) and Average Symmetric Surface Distance (ASSD), it achieved a dice score of 93.1% and 92.0% in the 2018 Atrial Segmentation Challenge and our local institutional dataset, respectively. Second, the number of parameters and computation complexity are largely reduced by 2.8x and 3.8x, respectively. Moreover, Usformer does not require a large dataset. When only 16 labeled MRI scans are used for training, Usformer achieves a 92.1% dice score in the challenge dataset. The proposed Usformer delineates the boundaries of the LA wall relatively accurately, which may assist in the clinical translation of LA LGE for planning catheter ablation of atrial fibrillation.

1 April 2023 in Radiology: Cardiothoracic Imaging
Verified
Journal Article
Closed access
AMAnthony Maroun
Anthony Maroun
JBJustin J. Baraboo
Justin J. Baraboo
SGSuvai Gunasekaran
Suvai Gunasekaran
JHJulia M. Hwang
Julia M. Hwang
SLSophia Z Liu
Sophia Z Liu
MPMaurice Pradella
Maurice Pradella
To compare maximum left atrial (LA) volume (LAV) from the routinely used biplane area-length (BAL) method with three-dimensional (3D)–based volumetry from late gadolinium-enhanced MRI (3D LGE MRI) and contrast-enhanced MR angiography (3D CE-MRA) in patients with atrial fibrillation (AF). Sixty-four patients with AF (mean age, 63 years ± 9 [SD]; 40 male patients) were retrospectively included from a prospective cohort acquired between October 2018 and February 2021. All patients underwent a research MRI examination that included standard two- and four-chamber cine acquisitions, 3D CE-MRA, and 3D LGE MRI performed prior to the atrial kick. Contour delineation on cine imaging and LA 3D segmentations were performed by a radiologist. Maximum LAV (BALmax) was extracted from the BAL volume-time curve and compared with LAV from 3D CE-MRA and 3D LGE MRI. The Kruskal-Wallis test was performed, followed by the Dunn post hoc test and Bland-Altman analyses. Interobserver variability was assessed in 10 patients. BALmax underestimated LAV compared with 3D CE-MRA (bias: -23.5 mL ± 46.2, P < .001) and 3D LGE MRI (bias: -31.3 mL ± 58.3, P < .001), whereas 3D LGE MRI volumes showed no evidence of a difference from 3D CE-MRA (bias: 7.8 mL ± 45.7, P = .38). Interobserver variability yielded excellent agreement for each method (intraclass correlation coefficient, 0.96–0.98). BALmax underestimated LAV in patients with AF compared with 3D LGE MRI and 3D CE-MRA, suggesting that the geometric assumption of an ellipsoidal LA shape in BAL does not reflect LA geometry in patients with AF. Keywords: Left Atrial Volume, Biplane Area-Length, Late Gadolinium-enhanced 3D MRI, Contrast-enhanced 3D MR Angiography, Atrial Fibrillation Supplemental material is available for this article. © RSNA, 2023

1 December 2023 in Radiology: Cardiothoracic Imaging
Verified
Journal Article
Closed access
DMDomenico Mastrodicasa
Domenico Mastrodicasa
SGSuvai Gunasekaran
Suvai Gunasekaran
SASamer Alabed
Samer Alabed
GGGaurav S Gulsin
Gaurav S Gulsin
KHKate Hanneman
Kate Hanneman

1 December 2024 in Radiology: Cardiothoracic Imaging
Verified
Journal Article
Closed access
DMDomenico Mastrodicasa
Domenico Mastrodicasa
SGSuvai Gunasekaran
Suvai Gunasekaran
SASamer Alabed
Samer Alabed
KHKate Hanneman
Kate Hanneman
GGGaurav S Gulsin
Gaurav S Gulsin

29 October 2026 in Journal of Cardiovascular Magnetic Resonance
Verified
Journal Article
Open access
SGSuvai Gunasekaran
Suvai Gunasekaran
GTGeorge D. Thornton
George D. Thornton
AKAlan C. Kwan
Alan C. Kwan
HLHaoran Li
Haoran Li
SRSébastien Roujol
Sébastien Roujol
DBDavid A. Bluemke
David A. Bluemke