Pallavi Tiwari
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Pallavi Tiwari has a total of 252 co-authors. Here are the top co-authors by number of shared publications:
- Anant Madabhushi
Department of Biomedical Engineering, Emory University, Atlanta, GA
23 shared publications - Vivek Prabhakaran
Department of Radiology, School of Medicine and Public Health, University of Wisconsin-Madison
18 shared publications - Ankush Bhatia
Neurology, University of Wisconsin-Madison School of Medicine and Public Health, Madison, Wisconsin, USA
17 shared publications - Anna Skakodub
Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY
17 shared publications - Anne S. Reiner
Memorial Sloan Kettering Cancer Center, New York, New York, USA
17 shared publications - Henry Walch
Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY
17 shared publications
Latest Publications
- 2 February 2027
- Published by Springer Nature
- p. 462-471
Active Learning Strategy for Automatic Classification of H&E-Stained Histopathologic Brain Tumor Sub-regions (BraTS-Path Challenge 2024)
- 2 February 2027
- Published by Springer Nature
- p. 462-471
Active Learning Strategy for Automatic Classification of H&E-Stained Histopathologic Brain Tumor Sub-regions (BraTS-Path Challenge 2024)
- Open Access
- 11 September 2026
- Published by Springer Nature in npj Precision Oncology
Graph-radiomic learning (GrRAiL) descriptor to characterize imaging heterogeneity in confounding tumor pathologies
Distinguishing benign confounding pathologies from malignant neoplasms on routine imaging remains a significant challenge in solid tumors. Radiomics approaches attempt to capture lesion heterogeneity on clinical imaging (CT,MRI), but typically aggregate feature values across the region of interest (ROI), losing the complex spatial relationships among intensity compositions within the lesion. We present Graph-Radiomic Learning (GrRAiL), a new descriptor for characterizing intralesional heterogeneity (ILH) on clinical MRI. GrRAiL (1) identifies clusters of sub-regions from per-voxel radiomic measurements and (2) computes graph-theoretic measurements that quantify the spatial associations of these radiomic clusters. The resulting weighted graphs capture higher-order spatial relationships within the ROI to reliably distinguish confounding pathologies from malignancies. GrRAiL was evaluated on n = 947 subjects across three multi-institutional use-cases: distinguishing radiation effects from tumor recurrence in glioblastoma ( n = 106) and brain metastasis ( n = 233), and differentiating no+low- versus high-risk pancreatic intraductal papillary mucinous neoplasms (IPMNs) ( n = 608). GrRAiL consistently outperformed Graph Neural Networks, textural radiomics, and intensity graph analysis, achieving cross-validation and test accuracies of 89% and 78% (glioblastoma), 84% and 74% (metastasis), and 85% and 79% (IPMN), representing >10%, >13%, and >10% test-accuracy improvements over comparative approaches, respectively.
- 11 September 2026
- Published by Springer Nature in npj Precision Oncology
Open AccessGraph-radiomic learning (GrRAiL) descriptor to characterize imaging heterogeneity in confounding tumor pathologies
Distinguishing benign confounding pathologies from malignant neoplasms on routine imaging remains a significant challenge in solid tumors. Radiomics approaches attempt to capture lesion heterogeneity on clinical imaging (CT,MRI), but typically aggregate feature values across the region of interest (ROI), losing the complex spatial relationships among intensity compositions within the lesion. We present Graph-Radiomic Learning (GrRAiL), a new descriptor for characterizing intralesional heterogeneity (ILH) on clinical MRI. GrRAiL (1) identifies clusters of sub-regions from per-voxel radiomic measurements and (2) computes graph-theoretic measurements that quantify the spatial associations of these radiomic clusters. The resulting weighted graphs capture higher-order spatial relationships within the ROI to reliably distinguish confounding pathologies from malignancies. GrRAiL was evaluated on n = 947 subjects across three multi-institutional use-cases: distinguishing radiation effects from tumor recurrence in glioblastoma ( n = 106) and brain metastasis ( n = 233), and differentiating no+low- versus high-risk pancreatic intraductal papillary mucinous neoplasms (IPMNs) ( n = 608). GrRAiL consistently outperformed Graph Neural Networks, textural radiomics, and intensity graph analysis, achieving cross-validation and test accuracies of 89% and 78% (glioblastoma), 84% and 74% (metastasis), and 85% and 79% (IPMN), representing >10%, >13%, and >10% test-accuracy improvements over comparative approaches, respectively.
- 1 May 2026
- Published by Radiological Society of North America (RSNA) in Radiology: Imaging Cancer
- Vol. 8 (3) , e250435
Radiomics-based Differentiation of Recurrent Brain Metastases from Treatment Effects: A Multi-Institutional Comparative Study with Advanced Imaging
A radiomic classifier constructed from multi-institutional posttreatment MRI data reliably differentiated tumor recurrence from treatment effects in patients with brain metastases and accurately classified lesions that were indistinguishable at clinical imaging.
- 1 May 2026
- Published by Radiological Society of North America (RSNA) in Radiology: Imaging Cancer
- Vol. 8 (3) , e250435
Radiomics-based Differentiation of Recurrent Brain Metastases from Treatment Effects: A Multi-Institutional Comparative Study with Advanced Imaging
A radiomic classifier constructed from multi-institutional posttreatment MRI data reliably differentiated tumor recurrence from treatment effects in patients with brain metastases and accurately classified lesions that were indistinguishable at clinical imaging.
- Open Access
- 28 March 2026
- Published by Springer Nature in Abdominal Radiology
- Vol. 51 (10) , 1-2
Correction to: Multi-center evaluation of radiomics and deep learning to stratify malignancy risk of IPMNs
- 28 March 2026
- Published by Springer Nature in Abdominal Radiology
- Vol. 51 (10) , 1-2
Open AccessCorrection to: Multi-center evaluation of radiomics and deep learning to stratify malignancy risk of IPMNs
- Open Access
- 12 January 2026
- Published by Springer Nature in Abdominal Radiology
- Vol. 51 (8) , 1-16
Multi-center evaluation of radiomics and deep learning to stratify malignancy risk of IPMNs
Purpose: Distinguishing high-risk intraductal papillary mucinous neoplasms (IPMNs) from low-risk lesions remains a clinical challenge, often resulting in unnecessary procedures due to limited specificity of current methods. While radiomics and deep learning (DL) have been explored for pancreatic cancer, cyst-level malignancy risk stratification of IPMNs remains untapped. Methods: Our multi-institutional assessed the feasibility of AI for predicting IPMN dysplasia grade using cyst-level image features using 359 T2-weighted (T2W) MRI images from seven centers. We developed and compared 2D and 3D radiomics-only, DL-only, and radiomics-DL fusion models using expert radiologist scoring as a baseline reference. Model performance was evaluated using held-out test data. Results: The radiomics-DL fusion model showed the highest discriminatory ability on the test set AUC of 69.2%, outperforming the radiomics-only model, AUC of 66.5%. Expert accuracy varied widely from 37.4% to 66.7%, and the inter-rater agreement varied as well with weighted Cohen’s kappa coefficients of 0.33–0.67. Conclusion: The fusion model, which combines DL with radiomics features from routine T2W MRI, shows promise for objective, cyst-level risk stratification of IPMNs in a multi-center cohort, outperforming radiomics-only models and nearly matching expert radiologists using only T2W and T1-weighted (T1W) sequences. While performance requires improvement for standalone clinical use, this approach offers a scalable, non-invasive method to potentially improve diagnostic accuracy and reduce unnecessary surgical interventions.
- 12 January 2026
- Published by Springer Nature in Abdominal Radiology
- Vol. 51 (8) , 1-16
Open AccessMulti-center evaluation of radiomics and deep learning to stratify malignancy risk of IPMNs
Purpose: Distinguishing high-risk intraductal papillary mucinous neoplasms (IPMNs) from low-risk lesions remains a clinical challenge, often resulting in unnecessary procedures due to limited specificity of current methods. While radiomics and deep learning (DL) have been explored for pancreatic cancer, cyst-level malignancy risk stratification of IPMNs remains untapped. Methods: Our multi-institutional assessed the feasibility of AI for predicting IPMN dysplasia grade using cyst-level image features using 359 T2-weighted (T2W) MRI images from seven centers. We developed and compared 2D and 3D radiomics-only, DL-only, and radiomics-DL fusion models using expert radiologist scoring as a baseline reference. Model performance was evaluated using held-out test data. Results: The radiomics-DL fusion model showed the highest discriminatory ability on the test set AUC of 69.2%, outperforming the radiomics-only model, AUC of 66.5%. Expert accuracy varied widely from 37.4% to 66.7%, and the inter-rater agreement varied as well with weighted Cohen’s kappa coefficients of 0.33–0.67. Conclusion: The fusion model, which combines DL with radiomics features from routine T2W MRI, shows promise for objective, cyst-level risk stratification of IPMNs in a multi-center cohort, outperforming radiomics-only models and nearly matching expert radiologists using only T2W and T1-weighted (T1W) sequences. While performance requires improvement for standalone clinical use, this approach offers a scalable, non-invasive method to potentially improve diagnostic accuracy and reduce unnecessary surgical interventions.
- Open Access
- 13 November 2025
- Published by Springer Nature in Research Square
Graph-Radiomic Learning (GrRAiL) Descriptor to Characterize Imaging Heterogeneity in Confounding Tumor Pathologies
A significant challenge in solid tumors is the reliable distinction of confounding pathologies from malignant neoplasms on routine imaging scans. While radiomics approaches have attempted to capture surrogate markers of lesion heterogeneity on clinical imaging (CT, MRI), these approaches often aggregate feature values across the region of interest (ROI) and may not capture the complex spatial relationships across various intensity compositions in the ROI. We present a new Graph-Radiomic Learning (GrRAiL) descriptor for characterizing intralesional heterogeneity (ILH) on clinical MRI scans. GrRAiL involves (1) identifying clusters of sub-regions based on per-voxel radiomic measurements, followed by (2) computing graph-theoretic measurements that analyze the spatial associations of radiomic clusters. The weighted graphs constructed via the GrRAiL descriptor capture the complexity of the higher-order spatial relationships within the ROI, in an attempt to reliably characterize ILH and distinguish confounding pathologies from malignant neoplasms. To evaluate its efficacy and clinical feasibility, GrRAiL descriptor was evaluated on a total of n=947 subjects spanning 3 different clinical use-cases including distinguishing radiation-effects from tumor recurrence in glioblastoma (n = 106) and brain metastasis (n = 233), as well as distinguishing no+low- versus high-risk pancreatic intraductal papillary mucus neoplasms (IPMNs) (n = 608). In a multi-institutional setting, GrRAiL consistently outperformed state-of-the-art approaches (Graph Neural Networks (GNNs), textural radiomics, and intensity graph analysis). For the glioblastoma use-case, the cross validation (CV) and test accuracies obtained to classify tumor recurrence from pseudo-progression, were 89% and 78%, respectively, with >10% improvement in test accuracy over comparative approaches. In the brain metastasis cohort, GrRAiL achieved CV and test accuracies of 84% and 74% in classifying tumor recurrence from radiation necrosis (>13% improvement over other approaches). Finally, when GrRAiL was used to classify between no+low- and high-risk IPMNs, CV and test accuracies of 85% and 79% were obtained, demonstrating >10% improvement over comparative approaches.
- 13 November 2025
- Published by Springer Nature in Research Square
Open AccessGraph-Radiomic Learning (GrRAiL) Descriptor to Characterize Imaging Heterogeneity in Confounding Tumor Pathologies
A significant challenge in solid tumors is the reliable distinction of confounding pathologies from malignant neoplasms on routine imaging scans. While radiomics approaches have attempted to capture surrogate markers of lesion heterogeneity on clinical imaging (CT, MRI), these approaches often aggregate feature values across the region of interest (ROI) and may not capture the complex spatial relationships across various intensity compositions in the ROI. We present a new Graph-Radiomic Learning (GrRAiL) descriptor for characterizing intralesional heterogeneity (ILH) on clinical MRI scans. GrRAiL involves (1) identifying clusters of sub-regions based on per-voxel radiomic measurements, followed by (2) computing graph-theoretic measurements that analyze the spatial associations of radiomic clusters. The weighted graphs constructed via the GrRAiL descriptor capture the complexity of the higher-order spatial relationships within the ROI, in an attempt to reliably characterize ILH and distinguish confounding pathologies from malignant neoplasms. To evaluate its efficacy and clinical feasibility, GrRAiL descriptor was evaluated on a total of n=947 subjects spanning 3 different clinical use-cases including distinguishing radiation-effects from tumor recurrence in glioblastoma (n = 106) and brain metastasis (n = 233), as well as distinguishing no+low- versus high-risk pancreatic intraductal papillary mucus neoplasms (IPMNs) (n = 608). In a multi-institutional setting, GrRAiL consistently outperformed state-of-the-art approaches (Graph Neural Networks (GNNs), textural radiomics, and intensity graph analysis). For the glioblastoma use-case, the cross validation (CV) and test accuracies obtained to classify tumor recurrence from pseudo-progression, were 89% and 78%, respectively, with >10% improvement in test accuracy over comparative approaches. In the brain metastasis cohort, GrRAiL achieved CV and test accuracies of 84% and 74% in classifying tumor recurrence from radiation necrosis (>13% improvement over other approaches). Finally, when GrRAiL was used to classify between no+low- and high-risk IPMNs, CV and test accuracies of 85% and 79% were obtained, demonstrating >10% improvement over comparative approaches.
- Open Access
Graph-Radiomic Learning (GrRAiL) Descriptor to Characterize Imaging Heterogeneity in Confounding Tumor Pathologies
A significant challenge in solid tumors is reliably distinguishing confounding pathologies from malignant neoplasms on routine imaging. While radiomics methods seek surrogate markers of lesion heterogeneity on CT/MRI, many aggregate features across the region of interest (ROI) and miss complex spatial relationships among varying intensity compositions. We present a new Graph-Radiomic Learning (GrRAiL) descriptor for characterizing intralesional heterogeneity (ILH) on clinical MRI scans. GrRAiL (1) identifies clusters of sub-regions using per-voxel radiomic measurements, then (2) computes graph-theoretic metrics to quantify spatial associations among clusters. The resulting weighted graphs encode higher-order spatial relationships within the ROI, aiming to reliably capture ILH and disambiguate confounding pathologies from malignancy. To assess efficacy and clinical feasibility, GrRAiL was evaluated in n=947 subjects spanning three use cases: differentiating tumor recurrence from radiation effects in glioblastoma (GBM; n=106) and brain metastasis (n=233), and stratifying pancreatic intraductal papillary mucinous neoplasms (IPMNs) into no+low vs high risk (n=608). In a multi-institutional setting, GrRAiL consistently outperformed state-of-the-art baselines - Graph Neural Networks (GNNs), textural radiomics, and intensity-graph analysis. In GBM, cross-validation (CV) and test accuracies for recurrence vs pseudo-progression were 89% and 78% with >10% test-accuracy gains over comparators. In brain metastasis, CV and test accuracies for recurrence vs radiation necrosis were 84% and 74% (>13% improvement). For IPMN risk stratification, CV and test accuracies were 84% and 75%, showing >10% improvement.
Open AccessGraph-Radiomic Learning (GrRAiL) Descriptor to Characterize Imaging Heterogeneity in Confounding Tumor Pathologies
A significant challenge in solid tumors is reliably distinguishing confounding pathologies from malignant neoplasms on routine imaging. While radiomics methods seek surrogate markers of lesion heterogeneity on CT/MRI, many aggregate features across the region of interest (ROI) and miss complex spatial relationships among varying intensity compositions. We present a new Graph-Radiomic Learning (GrRAiL) descriptor for characterizing intralesional heterogeneity (ILH) on clinical MRI scans. GrRAiL (1) identifies clusters of sub-regions using per-voxel radiomic measurements, then (2) computes graph-theoretic metrics to quantify spatial associations among clusters. The resulting weighted graphs encode higher-order spatial relationships within the ROI, aiming to reliably capture ILH and disambiguate confounding pathologies from malignancy. To assess efficacy and clinical feasibility, GrRAiL was evaluated in n=947 subjects spanning three use cases: differentiating tumor recurrence from radiation effects in glioblastoma (GBM; n=106) and brain metastasis (n=233), and stratifying pancreatic intraductal papillary mucinous neoplasms (IPMNs) into no+low vs high risk (n=608). In a multi-institutional setting, GrRAiL consistently outperformed state-of-the-art baselines - Graph Neural Networks (GNNs), textural radiomics, and intensity-graph analysis. In GBM, cross-validation (CV) and test accuracies for recurrence vs pseudo-progression were 89% and 78% with >10% test-accuracy gains over comparators. In brain metastasis, CV and test accuracies for recurrence vs radiation necrosis were 84% and 74% (>13% improvement). For IPMN risk stratification, CV and test accuracies were 84% and 75%, showing >10% improvement.
- Open Access
- 20 March 2025
- Published by Wiley in Alzheimer's & Dementia
- Vol. 21 (3) , e70095
Association of neighborhood disadvantage with cognitive function and cortical disorganization in an unimpaired cohort: An exploratory study
INTRODUCTION: Neighborhood disadvantage has been shown to impact health and cognitive outcomes, while morphological similarity network (MSN) can elucidate structural morphological patterns underlying cognitive functions. We hypothesized MSNs could provide cortical patterns linked with neighborhood disadvantage and cognitive function, explaining the potential risk of cognitive impairment in disadvantaged neighborhoods.METHODS: For cognitively unimpaired participants from the Wisconsin Alzheimer's Disease Research Center or Wisconsin Registry for Alzheimer's Prevention (n = 524), and the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort (n = 100), neighborhood disadvantage was obtained using Area Deprivation Index (ADI) and its association with cognitive performance and MSN features was analyzed using linear regression and mediation analysis.RESULTS: Neighborhood disadvantage was associated with worse cognitive performance on memory, executive function, processing speed, and preclinical Alzheimer's tests on both datasets. Local morphological organization of predominantly the frontal and temporal regions showed association trends with ADI.DISCUSSION: Morphological patterns associated with ADI, in‐part, may explain the risk for poor cognitive functioning in a neighborhood disadvantaged population.Highlights: Social determinants of health such as neighborhood context can be studied using ADI. High neighborhood disadvantage was related to worse performance on category fluency, implicit learning speed, story recall memory and pre‐clinical Alzheimer's cognitive composite. In this exploratory study, using morphological brain networks that indicate similarity in distribution of cortical thickness between regions, we observed that centrality of predominantly frontal and temporal regions was marginally linked with neighborhood disadvantage status and also partially mediated its association with preclinical Alzheimer's composite test. There is a potential role for considering neighborhood status in early screening of cognitive impairment and dementia.
- 20 March 2025
- Published by Wiley in Alzheimer's & Dementia
- Vol. 21 (3) , e70095
Open AccessAssociation of neighborhood disadvantage with cognitive function and cortical disorganization in an unimpaired cohort: An exploratory study
INTRODUCTION: Neighborhood disadvantage has been shown to impact health and cognitive outcomes, while morphological similarity network (MSN) can elucidate structural morphological patterns underlying cognitive functions. We hypothesized MSNs could provide cortical patterns linked with neighborhood disadvantage and cognitive function, explaining the potential risk of cognitive impairment in disadvantaged neighborhoods.METHODS: For cognitively unimpaired participants from the Wisconsin Alzheimer's Disease Research Center or Wisconsin Registry for Alzheimer's Prevention (n = 524), and the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort (n = 100), neighborhood disadvantage was obtained using Area Deprivation Index (ADI) and its association with cognitive performance and MSN features was analyzed using linear regression and mediation analysis.RESULTS: Neighborhood disadvantage was associated with worse cognitive performance on memory, executive function, processing speed, and preclinical Alzheimer's tests on both datasets. Local morphological organization of predominantly the frontal and temporal regions showed association trends with ADI.DISCUSSION: Morphological patterns associated with ADI, in‐part, may explain the risk for poor cognitive functioning in a neighborhood disadvantaged population.Highlights: Social determinants of health such as neighborhood context can be studied using ADI. High neighborhood disadvantage was related to worse performance on category fluency, implicit learning speed, story recall memory and pre‐clinical Alzheimer's cognitive composite. In this exploratory study, using morphological brain networks that indicate similarity in distribution of cortical thickness between regions, we observed that centrality of predominantly frontal and temporal regions was marginally linked with neighborhood disadvantage status and also partially mediated its association with preclinical Alzheimer's composite test. There is a potential role for considering neighborhood status in early screening of cognitive impairment and dementia.
- Open Access
Advances for Managing Pancreatic Cystic Lesions: Integrating Imaging and AI Innovations
Pancreatic cystic lesions (PCLs) represent a spectrum of non-neoplasms and neoplasms with varying malignant potential, posing significant challenges in diagnosis and management. While some PCLs are precursors to pancreatic cancer, others remain benign, necessitating accurate differentiation for optimal patient care. Conventional approaches to PCL management rely heavily on radiographic imaging, and endoscopic ultrasound (EUS) guided fine-needle aspiration (FNA), coupled with clinical and biochemical data. However, the observer-dependent nature of image interpretation and the complex morphology of PCLs can lead to diagnostic uncertainty and variability in patient management strategies. This review critically evaluates current PCL diagnosis and surveillance practices, showing features of the different lesions and highlighting the potential limitations of conventional methods. We then explore the potential of artificial intelligence (AI) to transform PCL management. AI-driven strategies, including deep learning algorithms for automated pancreas and lesion segmentation, and radiomics for analyzing heterogeneity, can improve diagnostic accuracy and risk stratification. These advanced techniques can provide more objective and reproducible assessments, aiding clinicians in decision-making regarding follow-up intervals and surgical interventions. Early results suggest that AI-driven methods can significantly improve patient outcomes by enabling earlier detection of high-risk lesions and reducing unnecessary procedures for benign cysts. Finally, this review emphasizes that AI-driven approaches could potentially reshape the landscape of PCL management, ultimately leading to improved pancreatic cancer prevention.
Open AccessAdvances for Managing Pancreatic Cystic Lesions: Integrating Imaging and AI Innovations
Pancreatic cystic lesions (PCLs) represent a spectrum of non-neoplasms and neoplasms with varying malignant potential, posing significant challenges in diagnosis and management. While some PCLs are precursors to pancreatic cancer, others remain benign, necessitating accurate differentiation for optimal patient care. Conventional approaches to PCL management rely heavily on radiographic imaging, and endoscopic ultrasound (EUS) guided fine-needle aspiration (FNA), coupled with clinical and biochemical data. However, the observer-dependent nature of image interpretation and the complex morphology of PCLs can lead to diagnostic uncertainty and variability in patient management strategies. This review critically evaluates current PCL diagnosis and surveillance practices, showing features of the different lesions and highlighting the potential limitations of conventional methods. We then explore the potential of artificial intelligence (AI) to transform PCL management. AI-driven strategies, including deep learning algorithms for automated pancreas and lesion segmentation, and radiomics for analyzing heterogeneity, can improve diagnostic accuracy and risk stratification. These advanced techniques can provide more objective and reproducible assessments, aiding clinicians in decision-making regarding follow-up intervals and surgical interventions. Early results suggest that AI-driven methods can significantly improve patient outcomes by enabling earlier detection of high-risk lesions and reducing unnecessary procedures for benign cysts. Finally, this review emphasizes that AI-driven approaches could potentially reshape the landscape of PCL management, ultimately leading to improved pancreatic cancer prevention.