Matthew D. Lee

Matthew D. Lee

Auto-generated profile, based on publicly available information

Matthew D. Lee

Department of Radiology, NYU Grossman School of Medicine, New York, United States

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
5 December 2022 in Nature Communications
Verified
Journal Article
Open access
PDF file
SPSarthak Pati
Sarthak Pati
Ujjwal BaidUjjwal Baid
Ujjwal Baid
BEBrandon Edwards
Brandon Edwards
MSMicah Sheller
Micah Sheller
SWShih-Han Wang
Shih-Han Wang
Spyridon BakasSpyridon Bakas
Spyridon Bakas
Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

8 July 2025 in Nature Communications
Verified
Journal Article
Open access
PDF file
MZMaximilian Zenk
Maximilian Zenk
Ujjwal BaidUjjwal Baid
Ujjwal Baid
SPSarthak Pati
Sarthak Pati
ALAkis Linardos
Akis Linardos
BEBrandon Edwards
Brandon Edwards
Spyridon BakasSpyridon Bakas
Spyridon Bakas
Computational competitions are the standard for benchmarking medical image analysis algorithms, but they typically use small curated test datasets acquired at a few centers, leaving a gap to the reality of diverse multicentric patient data. To this end, the Federated Tumor Segmentation (FeTS) Challenge represents the paradigm for real-world algorithmic performance evaluation. The FeTS challenge is a competition to benchmark (i) federated learning aggregation algorithms and (ii) state-of-the-art segmentation algorithms, across multiple international sites. Weight aggregation and client selection techniques were compared using a multicentric brain tumor dataset in realistic federated learning simulations, yielding benefits for adaptive weight aggregation, and efficiency gains through client sampling. Quantitative performance evaluation of state-of-the-art segmentation algorithms on data distributed internationally across 32 institutions yielded good generalization on average, albeit the worst-case performance revealed data-specific modes of failure. Similar multi-site setups can help validate the real-world utility of healthcare AI algorithms in the future.

1 August 2019 in American Journal of Neuroradiology
Verified
Journal Article
Open access
PDF file
MLM D Lee
M D Lee
GBG.L. Baird
G.L. Baird
LBL.C. Bell
L.C. Bell
Chad QuarlesChad Quarles
Chad Quarles
JBJerrold Boxerman
Jerrold Boxerman
BACKGROUND AND PURPOSE: The percentage signal recovery in non-leakage-corrected (no preload, high flip angle, intermediate TE) DSC-MR imaging is known to differ significantly for glioblastoma, metastasis, and primary CNS lymphoma. Because the percentage signal recovery is influenced by preload and pulse sequence parameters, we investigated whether the percentage signal recovery can still differentiate these common contrast-enhancing neoplasms using a DSC-MR imaging protocol designed for relative CBV accuracy (preload, intermediate flip angle, low TE). MATERIALS AND METHODS: We retrospectively analyzed DSC-MR imaging of treatment-naïve, pathology-proved glioblastomas (n = 14), primary central nervous system lymphomas (n = 7), metastases (n = 20), and meningiomas (n = 13) using a protocol designed for relative CBV accuracy (a one-quarter-dose preload and single-dose bolus of gadobutrol, TR/TE = 1290/40 ms, flip angle = 60° at 1.5T). Mean percentage signal recovery, relative CBV, and normalized baseline signal intensity were compared within contrast-enhancing lesion volumes. Classification accuracy was determined by receiver operating characteristic analysis. RESULTS: Relative CBV best differentiated meningioma from glioblastoma and from metastasis with areas under the curve of 0.84 and 0.82, respectively. The percentage signal recovery best differentiated primary central nervous system lymphoma from metastasis with an area under the curve of 0.81. Relative CBV and percentage signal recovery were similar in differentiating primary central nervous system lymphoma from glioblastoma and from meningioma. Although neither relative CBV nor percentage signal recovery differentiated glioblastoma from metastasis, mean normalized baseline signal intensity achieved 86% sensitivity and 50% specificity. CONCLUSIONS: Similar to results for non-preload-based DSC-MR imaging, percentage signal recovery for one-quarter-dose preload-based, intermediate flip angle DSC-MR imaging differentiates most pair-wise comparisons of glioblastoma, metastasis, primary central nervous system lymphoma, and meningioma, except for glioblastoma versus metastasis. Differences in normalized post-preload baseline signal for glioblastoma and metastasis, reflecting a snapshot of dynamic contrast enhancement, may motivate the use of single-dose multiecho protocols permitting simultaneous quantification of DSC-MR imaging and dynamic contrast-enhanced MR imaging parameters.

1 June 2024 in Nature Biomedical Engineering
Verified
Journal Article
Closed access
PDF file
MNMustafa Nasir-Moin
Mustafa Nasir-Moin
LWLisa Irina Wadiura
Lisa Irina Wadiura
VSVlad Sacalean
Vlad Sacalean
DJDevin Juros
Devin Juros
MMMisha Movahed-Ezazi
Misha Movahed-Ezazi
DODaniel A. Orringer
Daniel A. Orringer

1 July 2024 in Radiology: Artificial Intelligence
Verified
Journal Article
Open access
NTNghi C. D. Truong
Nghi C. D. Truong
CYChandan Ganesh Bangalore Yogananda
Chandan Ganesh Bangalore Yogananda
BWBenjamin C. Wagner
Benjamin C. Wagner
JHJames M. Holcomb
James M. Holcomb
DRDivya Reddy
Divya Reddy
JMJoseph A. Maldjian
Joseph A. Maldjian
"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To develop a radiomics framework for preoperative MRI-based prediction of IDH mutation status, a crucial glioma prognostic indicator. Materials and Methods Radiomics features (shape, first-order statistics, and texture) were extracted from the whole tumor or the combination of nonenhancing, necrosis, and edema regions. Segmentation masks were obtained via the federated tumor segmentation tool or the original data source. Boruta, a wrapper-based feature selection algorithm, identified relevant features. Addressing the imbalance between mutated and wild-type cases, multiple prediction models were trained on balanced data subsets using Random Forest or XGBoost and assembled to build the final classifier. The framework was evaluated using retrospective MRI scans from three public datasets (The Cancer Imaging Archive (TCIA, 227 patients), the University of California San Francisco Preoperative Diffuse Glioma MRI dataset (UCSF, 495 patients), and the Erasmus Glioma Database (EGD, 456 patients)) and internal datasets collected from UT Southwestern Medical Center (UTSW, 356 patients), New York University (NYU, 136 patients), and University of Wisconsin-Madison (UWM, 174 patients). TCIA and UTSW served as separate training sets, while the remaining data constituted the test set (1617 or 1488 testing cases, respectively). Results The best-performing models trained on the TCIA dataset achieved area under the receiver operating characteristic curve (AUC) values of 0.89 for UTSW, 0.86 for NYU, 0.93 for UWM, 0.94 for UCSF, and 0.88 for EGD test sets. The best-performing models trained on the UTSW dataset achieved slightly higher AUCs: 0.92 for TCIA, 0.88 for NYU, 0.96 for UWM, 0.93 for UCSF, and 0.90 for EGD. Conclusion This MRI radiomics-based framework shows promise for accurate preoperative prediction of IDH mutation status in patients with glioma. Published under a CC BY 4.0 license.

5 September 2023 in Bioengineering
Verified
Journal Article
Open access
PDF file
CYChandan Ganesh Bangalore Yogananda
Chandan Ganesh Bangalore Yogananda
BWBenjamin C. Wagner
Benjamin C. Wagner
NTNghi C. D. Truong
Nghi C. D. Truong
JHJames M. Holcomb
James M. Holcomb
DRDivya D. Reddy
Divya D. Reddy
JMJoseph A. Maldjian
Joseph A. Maldjian
Isocitrate dehydrogenase (IDH) mutation status has emerged as an important prognostic marker in gliomas. This study sought to develop deep learning networks for non-invasive IDH classification using T2w MR images while comparing their performance to a multi-contrast network. Methods: Multi-contrast brain tumor MRI and genomic data were obtained from The Cancer Imaging Archive (TCIA) and The Erasmus Glioma Database (EGD). Two separate 2D networks were developed using nnU-Net, a T2w-image-only network (T2-net) and a multi-contrast network (MC-net). Each network was separately trained using TCIA (227 subjects) or TCIA + EGD data (683 subjects combined). The networks were trained to classify IDH mutation status and implement single-label tumor segmentation simultaneously. The trained networks were tested on over 1100 held-out datasets including 360 cases from UT Southwestern Medical Center, 136 cases from New York University, 175 cases from the University of Wisconsin–Madison, 456 cases from EGD (for the TCIA-trained network), and 495 cases from the University of California, San Francisco public database. A receiver operating characteristic curve (ROC) was drawn to calculate the AUC value to determine classifier performance. Results: T2-net trained on TCIA and TCIA + EGD datasets achieved an overall accuracy of 85.4% and 87.6% with AUCs of 0.86 and 0.89, respectively. MC-net trained on TCIA and TCIA + EGD datasets achieved an overall accuracy of 91.0% and 92.8% with AUCs of 0.94 and 0.96, respectively. We developed reliable, high-performing deep learning algorithms for IDH classification using both a T2-image-only and a multi-contrast approach. The networks were tested on more than 1100 subjects from diverse databases, making this the largest study on image-based IDH classification to date.