Christopher R. Shea
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
Christopher R. Shea has a total of 116 co-authors. Here are the top co-authors by number of shared publications:
- Aly Azeem Khan
University of Chicago
5 shared publications - Derek Reiman
Toyota Technological Institute at Chicago
5 shared publications - Habibul Ahsan
Institute for Population and Precision Health (IPPH), University of Chicago, Chicago, IL 60637, USA
5 shared publications - Maria Argos
Boston University
5 shared publications - Mohammad Hasan Shahriar
Institute for Population and Precision Health, University of Chicago, Chicago, Illinois
5 shared publications - Mohammed Kamal
Department of Pathology, Bangabandhu Sheikh Mujib Medical University, Dhaka, Bangladesh
5 shared publications
Latest Publications
- 7 February 2026
- Published by Elsevier in Journal of the American Academy of Dermatology
- Vol. 94 (6) , e485-e486
In memoriam: Eugene Joseph Van Scott, MD (1922-2025)
- 7 February 2026
- Published by Elsevier in Journal of the American Academy of Dermatology
- Vol. 94 (6) , e485-e486
In memoriam: Eugene Joseph Van Scott, MD (1922-2025)
- Open Access
- 23 December 2025
- Published by Wiley in Journal of Cutaneous Pathology
- Vol. 53 (3) , 293-301
The Impact of Next‐Generation Sequencing on Interobserver Agreement and Diagnostic Accuracy of Deep Penetrating Melanocytic Neoplasms
Background: Next‐generation sequencing (NGS) is becoming more commonly used for diagnosis in dermatopathology. It's critical to appraise its efficacy and limitations. Distinguishing benign deep penetrating nevi (DPN) from deep penetrating like‐melanoma (DPN‐M) is a challenging diagnostic scenario even for experienced dermatopathologists. Methods: We sent a two‐phase survey (pre‐and postgenomics) to 32 experienced dermatopathologists to evaluate 39 diagnostically challenging cases from the DPN/WNT‐activated family of melanocytic neoplasms. Results: With NGS data, interobserver agreement improved from 0.41 to 0.51 ( p < 0.0001) in distinguishing DPN‐M from nonmelanoma cases. Overall diagnostic accuracy improved, mostly driven by a 16% increase in accurate diagnosis of DPN‐M. However, in two cases, the inclusion of genomics shifted the majority vote from a correct to an incorrect diagnosis. A total of 218 diagnostic changes occurred between Survey 1 and 2. Among the changes, 132 votes moved toward the correct diagnosis while 86 moved toward an incorrect diagnosis. The shift in voting which resulted in improved diagnostic accuracy was statistically significant ( p = 0.0001). Conclusions: NGS has the potential to improve interobserver agreement and diagnostic accuracy. We provide guidance on the utilization of bioinformatic data to maximize its benefits and improve diagnostic accuracy and interobserver agreement.
- 23 December 2025
- Published by Wiley in Journal of Cutaneous Pathology
- Vol. 53 (3) , 293-301
Open AccessThe Impact of Next‐Generation Sequencing on Interobserver Agreement and Diagnostic Accuracy of Deep Penetrating Melanocytic Neoplasms
Background: Next‐generation sequencing (NGS) is becoming more commonly used for diagnosis in dermatopathology. It's critical to appraise its efficacy and limitations. Distinguishing benign deep penetrating nevi (DPN) from deep penetrating like‐melanoma (DPN‐M) is a challenging diagnostic scenario even for experienced dermatopathologists. Methods: We sent a two‐phase survey (pre‐and postgenomics) to 32 experienced dermatopathologists to evaluate 39 diagnostically challenging cases from the DPN/WNT‐activated family of melanocytic neoplasms. Results: With NGS data, interobserver agreement improved from 0.41 to 0.51 ( p < 0.0001) in distinguishing DPN‐M from nonmelanoma cases. Overall diagnostic accuracy improved, mostly driven by a 16% increase in accurate diagnosis of DPN‐M. However, in two cases, the inclusion of genomics shifted the majority vote from a correct to an incorrect diagnosis. A total of 218 diagnostic changes occurred between Survey 1 and 2. Among the changes, 132 votes moved toward the correct diagnosis while 86 moved toward an incorrect diagnosis. The shift in voting which resulted in improved diagnostic accuracy was statistically significant ( p = 0.0001). Conclusions: NGS has the potential to improve interobserver agreement and diagnostic accuracy. We provide guidance on the utilization of bioinformatic data to maximize its benefits and improve diagnostic accuracy and interobserver agreement.
- 22 August 2025
- Published by Wolters Kluwer Health in American Journal of Gastroenterology
- Vol. 121 (1) , 268-273
New Eczematous Eruption in Patients With Inflammatory Bowel Disease Who Stop Janus Kinase Inhibitors
INTRODUCTION: Janus kinase inhibitors (JAKi) are small-molecule therapies used in inflammatory bowel disease (IBD). We describe 8 patients with IBD who developed an eczematous eruption, presumed atopic dermatitis (AD), after stopping JAKi therapy. METHODS: This case series describes 8 patients with IBD who developed de novo AD after withdrawal of JAKi therapy. RESULTS: Median time to AD onset was 7 days (interquartile range 2–12). Two cases with biopsy-confirmed AD resolved after reinitiation of JAKi therapy and adjunct AD treatments. DISCUSSION: A subset of patients with IBD may have latent susceptibility to AD, unmasked by therapy cessation and potentially triggered by an immune rebound effect.
- 22 August 2025
- Published by Wolters Kluwer Health in American Journal of Gastroenterology
- Vol. 121 (1) , 268-273
New Eczematous Eruption in Patients With Inflammatory Bowel Disease Who Stop Janus Kinase Inhibitors
INTRODUCTION: Janus kinase inhibitors (JAKi) are small-molecule therapies used in inflammatory bowel disease (IBD). We describe 8 patients with IBD who developed an eczematous eruption, presumed atopic dermatitis (AD), after stopping JAKi therapy. METHODS: This case series describes 8 patients with IBD who developed de novo AD after withdrawal of JAKi therapy. RESULTS: Median time to AD onset was 7 days (interquartile range 2–12). Two cases with biopsy-confirmed AD resolved after reinitiation of JAKi therapy and adjunct AD treatments. DISCUSSION: A subset of patients with IBD may have latent susceptibility to AD, unmasked by therapy cessation and potentially triggered by an immune rebound effect.
- Open Access
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Table 1 from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Important concepts and terms used in this study.
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Open AccessTable 1 from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Important concepts and terms used in this study.
- Open Access
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Table 2 from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Dataset characteristics broken down by NMSC subtype at the individual patient level and the biopsy level.
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Open AccessTable 2 from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Dataset characteristics broken down by NMSC subtype at the individual patient level and the biopsy level.
- Open Access
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Data from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Background:Early and precise diagnosis is vital to improving patient outcomes and reducing morbidity. In resource-limited settings, cancer diagnosis is often challenging due to shortages of expert pathologists. We assess the effectiveness of general-purpose pathology foundation models (FM) for the diagnosis and annotation of nonmelanoma skin cancer (NMSC) in resource-limited settings.Methods:We evaluated three pathology FMs (UNI, PRISM, and Prov-GigaPath) using deidentified NMSC histology images from the Bangladesh Vitamin E and Selenium Trial to predict cancer subtype based on zero-shot whole-slide embeddings. In addition, we evaluated tile aggregation methods and machine learning models for prediction. Lastly, we employed few-shot learning of PRISM tile embeddings to perform whole-slide annotation.Results:We found that the best model used PRISM’s aggregated tile embeddings to train a multilayer perceptron model to predict NMSC subtype [mean area under the receiver operating characteristic curve (AUROC) = 0.925, P < 0.001]. Within the other FMs, we found that using attention-based multi-instance learning to aggregate tile embeddings to train a multilayer perceptron model was optimal (UNI: mean AUROC = 0.913, P < 0.001; Prov-GigaPath: mean AUROC = 0.908, P < 0.001). We finally exemplify the utility of few-shot annotation in computation- and expertise-limited settings.Conclusions:Our study highlights the important role FMs may play in confronting public health challenges and exhibits a real-world potential for machine learning–aided cancer diagnosis.Impact:Pathology FMs offer a promising pathway to improve early and precise NMSC diagnosis, especially in resource-limited environments. These tools could also facilitate patient stratification and recruitment for prospective clinical trials aimed at improving NMSC management.
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Open AccessData from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Background:Early and precise diagnosis is vital to improving patient outcomes and reducing morbidity. In resource-limited settings, cancer diagnosis is often challenging due to shortages of expert pathologists. We assess the effectiveness of general-purpose pathology foundation models (FM) for the diagnosis and annotation of nonmelanoma skin cancer (NMSC) in resource-limited settings.Methods:We evaluated three pathology FMs (UNI, PRISM, and Prov-GigaPath) using deidentified NMSC histology images from the Bangladesh Vitamin E and Selenium Trial to predict cancer subtype based on zero-shot whole-slide embeddings. In addition, we evaluated tile aggregation methods and machine learning models for prediction. Lastly, we employed few-shot learning of PRISM tile embeddings to perform whole-slide annotation.Results:We found that the best model used PRISM’s aggregated tile embeddings to train a multilayer perceptron model to predict NMSC subtype [mean area under the receiver operating characteristic curve (AUROC) = 0.925, P < 0.001]. Within the other FMs, we found that using attention-based multi-instance learning to aggregate tile embeddings to train a multilayer perceptron model was optimal (UNI: mean AUROC = 0.913, P < 0.001; Prov-GigaPath: mean AUROC = 0.908, P < 0.001). We finally exemplify the utility of few-shot annotation in computation- and expertise-limited settings.Conclusions:Our study highlights the important role FMs may play in confronting public health challenges and exhibits a real-world potential for machine learning–aided cancer diagnosis.Impact:Pathology FMs offer a promising pathway to improve early and precise NMSC diagnosis, especially in resource-limited environments. These tools could also facilitate patient stratification and recruitment for prospective clinical trials aimed at improving NMSC management.
- Open Access
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Figure 1 from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Overview of generating whole-slide embeddings using FMs. Each WSI is split into 256 × 256 px tiles, and tiles with visible tissue are selected for further processing. Tile features are extracted by a tile encoder, tile embeddings for a WSI are aggregated into a slide embedding, and NMSC is classified using the slide-level embedding. Tile aggregation methods include ABMIL, FM-specific aggregators, or GAP. Classification is performed using LR, XGBoost, or a shallow MLP. ViT, vision transformer.
- 1 July 2025
- Published by American Association for Cancer Research (AACR)
Open AccessFigure 1 from AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Overview of generating whole-slide embeddings using FMs. Each WSI is split into 256 × 256 px tiles, and tiles with visible tissue are selected for further processing. Tile features are extracted by a tile encoder, tile embeddings for a WSI are aggregated into a slide embedding, and NMSC is classified using the slide-level embedding. Tile aggregation methods include ABMIL, FM-specific aggregators, or GAP. Classification is performed using LR, XGBoost, or a shallow MLP. ViT, vision transformer.
- Open Access
- 25 April 2025
- Published by American Association for Cancer Research (AACR) in Cancer Epidemiology, Biomarkers & Prevention
- Vol. 34 (7) , 1080-1088
AI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Background: Early and precise diagnosis is vital to improving patient outcomes and reducing morbidity. In resource-limited settings, cancer diagnosis is often challenging due to shortages of expert pathologists. We assess the effectiveness of general-purpose pathology foundation models (FM) for the diagnosis and annotation of nonmelanoma skin cancer (NMSC) in resource-limited settings. Methods: We evaluated three pathology FMs (UNI, PRISM, and Prov-GigaPath) using deidentified NMSC histology images from the Bangladesh Vitamin E and Selenium Trial to predict cancer subtype based on zero-shot whole-slide embeddings. In addition, we evaluated tile aggregation methods and machine learning models for prediction. Lastly, we employed few-shot learning of PRISM tile embeddings to perform whole-slide annotation. Results: We found that the best model used PRISM’s aggregated tile embeddings to train a multilayer perceptron model to predict NMSC subtype [mean area under the receiver operating characteristic curve (AUROC) = 0.925, P < 0.001]. Within the other FMs, we found that using attention-based multi-instance learning to aggregate tile embeddings to train a multilayer perceptron model was optimal (UNI: mean AUROC = 0.913, P < 0.001; Prov-GigaPath: mean AUROC = 0.908, P < 0.001). We finally exemplify the utility of few-shot annotation in computation- and expertise-limited settings. Conclusions: Our study highlights the important role FMs may play in confronting public health challenges and exhibits a real-world potential for machine learning–aided cancer diagnosis. Impact: Pathology FMs offer a promising pathway to improve early and precise NMSC diagnosis, especially in resource-limited environments. These tools could also facilitate patient stratification and recruitment for prospective clinical trials aimed at improving NMSC management.
- 25 April 2025
- Published by American Association for Cancer Research (AACR) in Cancer Epidemiology, Biomarkers & Prevention
- Vol. 34 (7) , 1080-1088
Open AccessAI-assisted Diagnosis of Nonmelanoma Skin Cancer in Resource-Limited Settings
Background: Early and precise diagnosis is vital to improving patient outcomes and reducing morbidity. In resource-limited settings, cancer diagnosis is often challenging due to shortages of expert pathologists. We assess the effectiveness of general-purpose pathology foundation models (FM) for the diagnosis and annotation of nonmelanoma skin cancer (NMSC) in resource-limited settings. Methods: We evaluated three pathology FMs (UNI, PRISM, and Prov-GigaPath) using deidentified NMSC histology images from the Bangladesh Vitamin E and Selenium Trial to predict cancer subtype based on zero-shot whole-slide embeddings. In addition, we evaluated tile aggregation methods and machine learning models for prediction. Lastly, we employed few-shot learning of PRISM tile embeddings to perform whole-slide annotation. Results: We found that the best model used PRISM’s aggregated tile embeddings to train a multilayer perceptron model to predict NMSC subtype [mean area under the receiver operating characteristic curve (AUROC) = 0.925, P < 0.001]. Within the other FMs, we found that using attention-based multi-instance learning to aggregate tile embeddings to train a multilayer perceptron model was optimal (UNI: mean AUROC = 0.913, P < 0.001; Prov-GigaPath: mean AUROC = 0.908, P < 0.001). We finally exemplify the utility of few-shot annotation in computation- and expertise-limited settings. Conclusions: Our study highlights the important role FMs may play in confronting public health challenges and exhibits a real-world potential for machine learning–aided cancer diagnosis. Impact: Pathology FMs offer a promising pathway to improve early and precise NMSC diagnosis, especially in resource-limited environments. These tools could also facilitate patient stratification and recruitment for prospective clinical trials aimed at improving NMSC management.
- 10 April 2025
- Published by Elsevier in Laboratory Investigation
- Vol. 105 (7) , 104169
Multicenter Evaluation of Certainty and Reliability in Calciphylaxis Diagnosis Using a Digital Imaging Platform
- 10 April 2025
- Published by Elsevier in Laboratory Investigation
- Vol. 105 (7) , 104169
Multicenter Evaluation of Certainty and Reliability in Calciphylaxis Diagnosis Using a Digital Imaging Platform