Joseph R. Peterson
Research Subjects
Joseph R. Peterson has a total of 74 co-authors. Here are the top co-authors by number of shared publications:
- Zaida Luthey-Schulten
University of Illinois at Urbana Champaign
5 shared publications - John Ryan Pfeiffer
SimBioSys, Inc, Chicago, IL
4 shared publications - Michael J. Hallock
SimBioSys, Inc, Chicago, IL
4 shared publications Patricia Carrigan
SimBioSys, Inc, Chicago, IL
4 shared publicationsThe SimBioSys Team
4 shared publications- Tushar Pandey
Department of Pathology, Post Graduate Institute of Medical Education and Research, Chandigarh, India
4 shared publications
Latest Publications
- 1 January 2025
- Published by American Society of Clinical Oncology (ASCO) in JCO Clinical Cancer Informatics
- Vol. 9 (9) , e2300248
Validation of Clinical Dynamic Contrast-Enhanced Magnetic Resonance Imaging Perfusion Modeling and Neoadjuvant Chemotherapy Response Prediction in Breast Cancer Using 18 FDG and 64 Cu-DOTA-Trastuzumab Positron Emission Tomography Studies
PURPOSE: Perfusion modeling presents significant opportunities for imaging biomarker development in breast cancer but has historically been held back by the need for data beyond the clinical standard of care (SoC) and uncertainty in the interpretability of results. We aimed to design a perfusion model applicable to breast cancer SoC dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) series with results stable to low temporal resolution imaging, comparable with published results using full-resolution DCE-MRI, and correlative with orthogonal imaging modalities indicative of biophysical markers. METHODS: Subsampled high-temporal-resolution DCE-MRI series were run through our perfusion model and resulting fits were compared for consistency. The fits were also compared against previously published results from institutions using the full resolution series. The model was then evaluated on a separate cohort for validity of biomarker indications. Finally, the model was used as a fundamental part of predicting response to neoadjuvant chemotherapy (NACT). RESULTS: Temporally subsampled DCE-MRI series yield perfusion fit variations on the scale of 1% of the tumor median value when input frames are varied. Fits generated from pseudoclinical series are within the variation range seen between imaging sites (ρ = 0.55), voxel-wise. The model also demonstrates significant correlations with orthogonal positron emission tomography imaging, indicating potential for use as a biomarker proxy. Specifically, using the perfusion fits as the grounding for a biophysical simulation of response, we correctly predict the pathologic complete response status after NACT in 15 of 18 patients, for an accuracy of 0.83, with a specificity and sensitivity of 0.83 as well. CONCLUSION: Clinical DCE-MRI data may be leveraged to provide stable perfusion fit results and indirectly interrogate the tumor microenvironment. These fits can then be used downstream for prediction of response to NACT with high accuracy.
- 1 January 2025
- Published by American Society of Clinical Oncology (ASCO) in JCO Clinical Cancer Informatics
- Vol. 9 (9) , e2300248
Validation of Clinical Dynamic Contrast-Enhanced Magnetic Resonance Imaging Perfusion Modeling and Neoadjuvant Chemotherapy Response Prediction in Breast Cancer Using 18 FDG and 64 Cu-DOTA-Trastuzumab Positron Emission Tomography Studies
PURPOSE: Perfusion modeling presents significant opportunities for imaging biomarker development in breast cancer but has historically been held back by the need for data beyond the clinical standard of care (SoC) and uncertainty in the interpretability of results. We aimed to design a perfusion model applicable to breast cancer SoC dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) series with results stable to low temporal resolution imaging, comparable with published results using full-resolution DCE-MRI, and correlative with orthogonal imaging modalities indicative of biophysical markers. METHODS: Subsampled high-temporal-resolution DCE-MRI series were run through our perfusion model and resulting fits were compared for consistency. The fits were also compared against previously published results from institutions using the full resolution series. The model was then evaluated on a separate cohort for validity of biomarker indications. Finally, the model was used as a fundamental part of predicting response to neoadjuvant chemotherapy (NACT). RESULTS: Temporally subsampled DCE-MRI series yield perfusion fit variations on the scale of 1% of the tumor median value when input frames are varied. Fits generated from pseudoclinical series are within the variation range seen between imaging sites (ρ = 0.55), voxel-wise. The model also demonstrates significant correlations with orthogonal positron emission tomography imaging, indicating potential for use as a biomarker proxy. Specifically, using the perfusion fits as the grounding for a biophysical simulation of response, we correctly predict the pathologic complete response status after NACT in 15 of 18 patients, for an accuracy of 0.83, with a specificity and sensitivity of 0.83 as well. CONCLUSION: Clinical DCE-MRI data may be leveraged to provide stable perfusion fit results and indirectly interrogate the tumor microenvironment. These fits can then be used downstream for prediction of response to NACT with high accuracy.
- Open Access
- 15 November 2024
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 10 (46) , eadq0856
Generative adversarial networks accurately reconstruct pan-cancer histology from pathologic, genomic, and radiographic latent features
Artificial intelligence models have been increasingly used in the analysis of tumor histology to perform tasks ranging from routine classification to identification of molecular features. These approaches distill cancer histologic images into high-level features, which are used in predictions, but understanding the biologic meaning of such features remains challenging. We present and validate a custom generative adversarial network—HistoXGAN—capable of reconstructing representative histology using feature vectors produced by common feature extractors. We evaluate HistoXGAN across 29 cancer subtypes and demonstrate that reconstructed images retain information regarding tumor grade, histologic subtype, and gene expression patterns. We leverage HistoXGAN to illustrate the underlying histologic features for deep learning models for actionable mutations, identify model reliance on histologic batch effect in predictions, and demonstrate accurate reconstruction of tumor histology from radiographic imaging for a “virtual biopsy.”
- 15 November 2024
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 10 (46) , eadq0856
Open AccessGenerative adversarial networks accurately reconstruct pan-cancer histology from pathologic, genomic, and radiographic latent features
Artificial intelligence models have been increasingly used in the analysis of tumor histology to perform tasks ranging from routine classification to identification of molecular features. These approaches distill cancer histologic images into high-level features, which are used in predictions, but understanding the biologic meaning of such features remains challenging. We present and validate a custom generative adversarial network—HistoXGAN—capable of reconstructing representative histology using feature vectors produced by common feature extractors. We evaluate HistoXGAN across 29 cancer subtypes and demonstrate that reconstructed images retain information regarding tumor grade, histologic subtype, and gene expression patterns. We leverage HistoXGAN to illustrate the underlying histologic features for deep learning models for actionable mutations, identify model reliance on histologic batch effect in predictions, and demonstrate accurate reconstruction of tumor histology from radiographic imaging for a “virtual biopsy.”
- Open Access
- 14 November 2024
- Published by Springer Nature in npj Breast Cancer
- Vol. 10 (1) , 1-13
Performance of an AI-powered visualization software platform for precision surgery in breast cancer patients
Surgery remains the primary treatment modality in the management of early-stage invasive breast cancer. Artificial intelligence (AI)-powered visualization platforms offer the compelling potential to aid surgeons in evaluating the tumor’s location and morphology within the breast and accordingly optimize their surgical approach. We sought to validate an AI platform that employs dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to render three-dimensional (3D) representations of the tumor and 5 additional chest tissues, offering clear visualizations as well as functionalities for quantifying tumor morphology, tumor-to-landmark structure distances, excision volumes, and approximate surgical margins. This retrospective study assessed the visualization platform’s performance on 100 cases with ground-truth labels vetted by 2 breast-specialized radiologists. We assessed features including automatic AI-generated clinical metrics (e.g., tumor dimensions) as well as visualization tools including convex hulls at desired margins around the tumor to help visualize lumpectomy volume. The statistical performance of the platform’s automated features was robust and within the range of inter-radiologist variability. These detailed 3D tumor and surrounding multi-tissue depictions offer both qualitative and quantitative comprehension of cancer topology and may aid in formulating an optimal surgical approach for breast cancer treatment. We further establish the framework for broader data integration into the platform to enhance precision cancer care.
- 14 November 2024
- Published by Springer Nature in npj Breast Cancer
- Vol. 10 (1) , 1-13
Open AccessPerformance of an AI-powered visualization software platform for precision surgery in breast cancer patients
Surgery remains the primary treatment modality in the management of early-stage invasive breast cancer. Artificial intelligence (AI)-powered visualization platforms offer the compelling potential to aid surgeons in evaluating the tumor’s location and morphology within the breast and accordingly optimize their surgical approach. We sought to validate an AI platform that employs dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to render three-dimensional (3D) representations of the tumor and 5 additional chest tissues, offering clear visualizations as well as functionalities for quantifying tumor morphology, tumor-to-landmark structure distances, excision volumes, and approximate surgical margins. This retrospective study assessed the visualization platform’s performance on 100 cases with ground-truth labels vetted by 2 breast-specialized radiologists. We assessed features including automatic AI-generated clinical metrics (e.g., tumor dimensions) as well as visualization tools including convex hulls at desired margins around the tumor to help visualize lumpectomy volume. The statistical performance of the platform’s automated features was robust and within the range of inter-radiologist variability. These detailed 3D tumor and surrounding multi-tissue depictions offer both qualitative and quantitative comprehension of cancer topology and may aid in formulating an optimal surgical approach for breast cancer treatment. We further establish the framework for broader data integration into the platform to enhance precision cancer care.
- Open Access
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 1217
Abstract 1217: Spatio-temporal modeling of the tumor microenvironment for prediction of patient-specific response to chemotherapy
Tumor heterogeneity, understood as the intrinsic (e.g., mutations, dysregulation, metabolic reprogramming) and extrinsic (e.g., nutrient/drug perfusion, interactions with surrounding tissues and the tumor microenvironment (TME)) to the cells that constitute the tumor, drives the variability of response to neoadjuvant chemotherapy. SimBioSys has developed TumorScope, a platform capable of capturing the chemical and biophysical interactions within the tumor and with the surrounding tissue by integrating these factors into a virtual tumor model. TumorScope makes individualized predictions of the response of each patient’s tumor to NACT. It utilizes three-dimensional biophysical simulations that model the dynamic tumor microenvironment (TME). The platform allows modelling of nutrient transport and drug behavior through perfusion and predicts when and where different regions of the tumor grow and die, and how that translates to overall response to treatment. TumorScope also incorporates multi-omics data, imaging data, pathology, and patient characteristics to simulate tumor biology of individual patient tumors. We have validated our approach in breast cancer by comparing TumorScope simulations against actual tumor response to neoadjuvant chemotherapy in over 400 patients. Simulations were initialized using each patient’s pretreatment MRI and pathology data and run from the beginning of therapy to the specified surgical date. The simulated tumor volumetric percent response (calculated as the ratio of change in tumor volume to initial volume) at the time of surgery was then compared with actual tumor volumetric percent response extracted from presurgical MRIs. Moreover, using TumorScope as a decision support system, we predicted patient response to multiple chemotherapies. For patients with event free survival data available we performed a Cox proportional hazard analysis. Using SimBioSys TumorScope we predicted a pre-surgical tumor volumetric response with a median error of 0.03% and median absolute deviation of 8.2%. For patients with available EFS data, we found a significant difference associated with having a final simulated volume greater than 0.01 cc. The SimBioSys TumorScope accurately predicts patient specific responses to NACT using only standard-of-care pre-treatment diagnostic data. These predictions can support decision making and enable the selection of less-toxic regimens for patients in which a robust response is predicted, and more aggressive treatments or alternative treatment strategies when response is likely to be poor. Citation Format: John A. Cole, Joseph R. Peterson, Tyler M. Earnest, Michael J. Hallock, Daniel Cook, Eduardo Braun, Anu Antony, Lance Munn, John Pfeiffer, Tushar Pandey. Spatio-temporal modeling of the tumor microenvironment for prediction of patient-specific response to chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1217.
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 1217
Open AccessAbstract 1217: Spatio-temporal modeling of the tumor microenvironment for prediction of patient-specific response to chemotherapy
Tumor heterogeneity, understood as the intrinsic (e.g., mutations, dysregulation, metabolic reprogramming) and extrinsic (e.g., nutrient/drug perfusion, interactions with surrounding tissues and the tumor microenvironment (TME)) to the cells that constitute the tumor, drives the variability of response to neoadjuvant chemotherapy. SimBioSys has developed TumorScope, a platform capable of capturing the chemical and biophysical interactions within the tumor and with the surrounding tissue by integrating these factors into a virtual tumor model. TumorScope makes individualized predictions of the response of each patient’s tumor to NACT. It utilizes three-dimensional biophysical simulations that model the dynamic tumor microenvironment (TME). The platform allows modelling of nutrient transport and drug behavior through perfusion and predicts when and where different regions of the tumor grow and die, and how that translates to overall response to treatment. TumorScope also incorporates multi-omics data, imaging data, pathology, and patient characteristics to simulate tumor biology of individual patient tumors. We have validated our approach in breast cancer by comparing TumorScope simulations against actual tumor response to neoadjuvant chemotherapy in over 400 patients. Simulations were initialized using each patient’s pretreatment MRI and pathology data and run from the beginning of therapy to the specified surgical date. The simulated tumor volumetric percent response (calculated as the ratio of change in tumor volume to initial volume) at the time of surgery was then compared with actual tumor volumetric percent response extracted from presurgical MRIs. Moreover, using TumorScope as a decision support system, we predicted patient response to multiple chemotherapies. For patients with event free survival data available we performed a Cox proportional hazard analysis. Using SimBioSys TumorScope we predicted a pre-surgical tumor volumetric response with a median error of 0.03% and median absolute deviation of 8.2%. For patients with available EFS data, we found a significant difference associated with having a final simulated volume greater than 0.01 cc. The SimBioSys TumorScope accurately predicts patient specific responses to NACT using only standard-of-care pre-treatment diagnostic data. These predictions can support decision making and enable the selection of less-toxic regimens for patients in which a robust response is predicted, and more aggressive treatments or alternative treatment strategies when response is likely to be poor. Citation Format: John A. Cole, Joseph R. Peterson, Tyler M. Earnest, Michael J. Hallock, Daniel Cook, Eduardo Braun, Anu Antony, Lance Munn, John Pfeiffer, Tushar Pandey. Spatio-temporal modeling of the tumor microenvironment for prediction of patient-specific response to chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1217.
- Open Access
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 1917
Abstract 1917: Accurate modeling of HER2 positive breast cancer disease progression with a biophysical modeling software
Breast cancer (BC) progression during NAT is associated with development of distant metastases, positive LN status, and decreased OS/RFS. These can occur in the context of clinical trials and therapy de-escalation, where the focus is on delivering effective NAT to patients while reducing drug toxicity. The risks added by disease progression underscore the need for early identification of NAT progressors. To this end, we replicated the NeoSphere study in silico using TumorScope (TS), a biophysical modeling software, focusing on predicting disease progression during NAT. The NeoSphere trial studied the efficacy of docetaxel (T), pertuzumab (P), and trastuzumab (H) in combination with one another over 254 operable BC patients distributed across four study arms. We selected past BC patients with accompanying standard of care clinical data that matched NeoSphere sample composition based on patient and tumor characteristics. A total, 144 patients were included across four study arms (TH, THP, HP, and TP). Parameters from the NeoSphere study were mirrored where possible. Simulation generated volume trajectories of individual tumor’s response to therapy. Disease progressors were identified based on tumor volume at the final simulation timepoint compared to the first simulation timepoint. We then compared group means and proportions between progressors and responders using Welch’s two-sample t-test, and Fisher’s exact test, respectively. We replicated the NeoSphere trial using TS. pCR rates across study arms closely mirrored those of the actual trial. In the HP arm of our trial, we identified 12 (12/144) progressors. No difference was found when comparing it to that observed in the NeoSphere trial (p=1.00, OR=1.12). As expected, percent change in tumor volume from initial to final timepoints for the progressor group was significantly higher than the responder group (n=121, t=19.2, p=1.5x10-10, mean progressor=38.7, mean responder=-75.9). The progressor group was enriched with higher grade tumors (t=2.85, p=0.01), as well as HR-negative tumors (p=0.002, OR=7.54) compared to the responder group, and had lower HER2 receptor FISH ratios (t=-3.4, p=0.002). There were no differences observed between groups age, cancer subtype, or AJCC tumor stage (p>0.05). After trial replication, we identified clinical features that separated progressors from responders, which are being assessed for development of individualized predictive biomarkers of disease progression. While work is ongoing in the field to identify biomarkers of BC progression, it is evident that single markers are not sufficient. Comprehensive, multi-modal biomarkers of disease progression must be developed and applied to patient sub-populations to garner effective predictions. Using biophysical simulations, we are able to investigate the impact of drug delivery/sensitivity, metabolism, and spatial heterogeneity on BC progression. Citation Format: John Pfeiffer, Tim Foley, Eduardo Braun, Anu Antony, Lance Munn, Joseph R. Peterson, John A. Cole, The SimBioSys Team. Accurate modeling of HER2 positive breast cancer disease progression with a biophysical modeling software [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1917.
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 1917
Open AccessAbstract 1917: Accurate modeling of HER2 positive breast cancer disease progression with a biophysical modeling software
Breast cancer (BC) progression during NAT is associated with development of distant metastases, positive LN status, and decreased OS/RFS. These can occur in the context of clinical trials and therapy de-escalation, where the focus is on delivering effective NAT to patients while reducing drug toxicity. The risks added by disease progression underscore the need for early identification of NAT progressors. To this end, we replicated the NeoSphere study in silico using TumorScope (TS), a biophysical modeling software, focusing on predicting disease progression during NAT. The NeoSphere trial studied the efficacy of docetaxel (T), pertuzumab (P), and trastuzumab (H) in combination with one another over 254 operable BC patients distributed across four study arms. We selected past BC patients with accompanying standard of care clinical data that matched NeoSphere sample composition based on patient and tumor characteristics. A total, 144 patients were included across four study arms (TH, THP, HP, and TP). Parameters from the NeoSphere study were mirrored where possible. Simulation generated volume trajectories of individual tumor’s response to therapy. Disease progressors were identified based on tumor volume at the final simulation timepoint compared to the first simulation timepoint. We then compared group means and proportions between progressors and responders using Welch’s two-sample t-test, and Fisher’s exact test, respectively. We replicated the NeoSphere trial using TS. pCR rates across study arms closely mirrored those of the actual trial. In the HP arm of our trial, we identified 12 (12/144) progressors. No difference was found when comparing it to that observed in the NeoSphere trial (p=1.00, OR=1.12). As expected, percent change in tumor volume from initial to final timepoints for the progressor group was significantly higher than the responder group (n=121, t=19.2, p=1.5x10-10, mean progressor=38.7, mean responder=-75.9). The progressor group was enriched with higher grade tumors (t=2.85, p=0.01), as well as HR-negative tumors (p=0.002, OR=7.54) compared to the responder group, and had lower HER2 receptor FISH ratios (t=-3.4, p=0.002). There were no differences observed between groups age, cancer subtype, or AJCC tumor stage (p>0.05). After trial replication, we identified clinical features that separated progressors from responders, which are being assessed for development of individualized predictive biomarkers of disease progression. While work is ongoing in the field to identify biomarkers of BC progression, it is evident that single markers are not sufficient. Comprehensive, multi-modal biomarkers of disease progression must be developed and applied to patient sub-populations to garner effective predictions. Using biophysical simulations, we are able to investigate the impact of drug delivery/sensitivity, metabolism, and spatial heterogeneity on BC progression. Citation Format: John Pfeiffer, Tim Foley, Eduardo Braun, Anu Antony, Lance Munn, Joseph R. Peterson, John A. Cole, The SimBioSys Team. Accurate modeling of HER2 positive breast cancer disease progression with a biophysical modeling software [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1917.
- Open Access
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 1916
Abstract 1916: Using systems medicine for comprehensive metabolic profiling of tumors: how tumor metabolism shapes prognosis and response to chemotherapy
Background: Metabolic disfunction is one of the hallmarks of cancer, and metabolism-based therapies are a key class of chemotherapeutics. Despite this, there has previously been no systematic way to identify metabolism-based targets for therapeutic intervention. This is especially concerning when considering that targeting tumor metabolism has been shown to increase the efficacy of standard chemotherapy in pre-clinical studies. We therefore developed a systems medicine approach to interrogate tumor metabolism systematically, identify the effects of metabolic heterogeneity on overall survival prognosis, and gain insights into new metabolism-associated targets or therapies to complement standard of care for cancer patients. Method: Our systems medicine framework statistically integrates metabolic network modeling with RNA-sequencing data (RNA-seq). This allows us to integrate molecular data about individual tumors (from RNA-seq) with a curated knowledge base of how these molecules interact within a patient’s tumor (using metabolic network models). This results in a mathematical description, or model, of a specific tumor’s metabolism that is able to be interrogated (i.e., it is high-dimensional like RNA-seq data) and able to be simulated (i.e., we can “drug” the model and investigate downstream effects). Results: We applied this systems medicine approach across a range of tumor types, including lung adenocarcinoma, pancreatic adenocarcinoma, metastatic melanoma, clear cell renal cell carcinoma, and salivary cystic adenoid carcinoma. Within each cancer type, we found significant patient-to-patient metabolic heterogeneity. Across cancer types, a consistent pattern that emerged was that the metabolic systems with the most patient-to-patient heterogeneity were also the systems most associated with patient overall survival. This result suggests that metabolic variability is a key driver of tumor response to therapy and overall patient survival. Furthermore, targeting metabolic pathways in a patient-specific manner could enhance chemotherapy in tumors with metabolic profiles associated with poor prognosis. Conclusion: This study shows that the specific metabolic profiles of tumors are key drivers of overall survival. We therefore believe that metabolic tumor characterization and intervention could be a useful strategy to enhance chemotherapy efficacy and overall survival in patients across tumor types. Citation Format: Daniel Cook, John A. Cole, Patricia Carrigan, Joseph R. Peterson, The SimBiosys Team. Using systems medicine for comprehensive metabolic profiling of tumors: how tumor metabolism shapes prognosis and response to chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1916.
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 1916
Open AccessAbstract 1916: Using systems medicine for comprehensive metabolic profiling of tumors: how tumor metabolism shapes prognosis and response to chemotherapy
Background: Metabolic disfunction is one of the hallmarks of cancer, and metabolism-based therapies are a key class of chemotherapeutics. Despite this, there has previously been no systematic way to identify metabolism-based targets for therapeutic intervention. This is especially concerning when considering that targeting tumor metabolism has been shown to increase the efficacy of standard chemotherapy in pre-clinical studies. We therefore developed a systems medicine approach to interrogate tumor metabolism systematically, identify the effects of metabolic heterogeneity on overall survival prognosis, and gain insights into new metabolism-associated targets or therapies to complement standard of care for cancer patients. Method: Our systems medicine framework statistically integrates metabolic network modeling with RNA-sequencing data (RNA-seq). This allows us to integrate molecular data about individual tumors (from RNA-seq) with a curated knowledge base of how these molecules interact within a patient’s tumor (using metabolic network models). This results in a mathematical description, or model, of a specific tumor’s metabolism that is able to be interrogated (i.e., it is high-dimensional like RNA-seq data) and able to be simulated (i.e., we can “drug” the model and investigate downstream effects). Results: We applied this systems medicine approach across a range of tumor types, including lung adenocarcinoma, pancreatic adenocarcinoma, metastatic melanoma, clear cell renal cell carcinoma, and salivary cystic adenoid carcinoma. Within each cancer type, we found significant patient-to-patient metabolic heterogeneity. Across cancer types, a consistent pattern that emerged was that the metabolic systems with the most patient-to-patient heterogeneity were also the systems most associated with patient overall survival. This result suggests that metabolic variability is a key driver of tumor response to therapy and overall patient survival. Furthermore, targeting metabolic pathways in a patient-specific manner could enhance chemotherapy in tumors with metabolic profiles associated with poor prognosis. Conclusion: This study shows that the specific metabolic profiles of tumors are key drivers of overall survival. We therefore believe that metabolic tumor characterization and intervention could be a useful strategy to enhance chemotherapy efficacy and overall survival in patients across tumor types. Citation Format: Daniel Cook, John A. Cole, Patricia Carrigan, Joseph R. Peterson, The SimBiosys Team. Using systems medicine for comprehensive metabolic profiling of tumors: how tumor metabolism shapes prognosis and response to chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1916.
- Open Access
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 2714
Abstract 2714: Systems medicine based metabolic profiling of metastatic melanoma reveals insights into tumor biology, prognosis, and new therapeutic strategies
Background: Tumor cells are able to reprogram their metabolism in order to sustain continuous growth and proliferation. Although, this has led to the development of metabolism-based therapies, until now there are no systematic ways to identify key metabolic targets for therapy. This area of research deserves more attention considering that targeting tumor metabolism has been shown to increase the efficacy of standard chemotherapy in pre-clinical studies. For tumors with poor prognosis, like late-stage metastatic melanoma, this untapped area for therapeutic discovery could lead to the development of novel metabolism-based therapies. Method: To address this challenge we designed a systems medicine-based technique to understand tumor metabolism in individual patients at a level of detail not previously achievable. In concrete, we statistically integrated metabolic network modeling with RNA-sequencing data (RNA-seq). This allows us to integrate molecular data about individual tumors (from RNA-seq) with a curated knowledge base of how these molecules interact within a patient’s tumor (using metabolic network models). This results in a mathematical description, or model, of a specific tumor’s metabolism that is able to be interrogated (i.e., it is high-dimensional like RNA-seq data) and able to be simulated (i.e., we can “drug” the model and investigate downstream effects). We applied this technique to 325 metastatic melanoma tumor RNA-seq profiles downloaded from The Cancer Genome Atlas to characterize metabolic differences across patients. Results: We found considerable differences in metabolic function across metastatic melanoma patient tumors. One of the main drivers of patient-to-patient metabolic variability appeared to be BRAF mutations, with BRAF mutated tumors showing higher levels of glutathione metabolism, lower levels of glutamate metabolism, and lower levels of oxidative phosphorylation. In addition to these differences, we found that several metabolic pathways were associated with poor overall survival prognosis regardless of BRAF mutations status including high specific growth rate, high cholesterol metabolism, and high serotonin and melatonin biosynthesis. Furthermore, simulating knockouts of enzymes involved in these pathways resulted in a set of actionable targets that could enhance chemotherapy for metastatic melanoma patients. Conclusion: Using systems medicine, we have characterized metabolism in metastatic melanoma at high resolution in 325 patient tumors. This characterization revealed the variability in metastatic melanoma across patients and how this variability is associated with clinical outcome. Furthermore, our approach suggests additional therapeutic targets to enhance patient care. Citation Format: Daniel Cook, Dorys A. Lopez-Ramos, Patricia Carrigan, Joseph R. Peterson, The SimBioSys Team. Systems medicine based metabolic profiling of metastatic melanoma reveals insights into tumor biology, prognosis, and new therapeutic strategies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2714.
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 2714
Open AccessAbstract 2714: Systems medicine based metabolic profiling of metastatic melanoma reveals insights into tumor biology, prognosis, and new therapeutic strategies
Background: Tumor cells are able to reprogram their metabolism in order to sustain continuous growth and proliferation. Although, this has led to the development of metabolism-based therapies, until now there are no systematic ways to identify key metabolic targets for therapy. This area of research deserves more attention considering that targeting tumor metabolism has been shown to increase the efficacy of standard chemotherapy in pre-clinical studies. For tumors with poor prognosis, like late-stage metastatic melanoma, this untapped area for therapeutic discovery could lead to the development of novel metabolism-based therapies. Method: To address this challenge we designed a systems medicine-based technique to understand tumor metabolism in individual patients at a level of detail not previously achievable. In concrete, we statistically integrated metabolic network modeling with RNA-sequencing data (RNA-seq). This allows us to integrate molecular data about individual tumors (from RNA-seq) with a curated knowledge base of how these molecules interact within a patient’s tumor (using metabolic network models). This results in a mathematical description, or model, of a specific tumor’s metabolism that is able to be interrogated (i.e., it is high-dimensional like RNA-seq data) and able to be simulated (i.e., we can “drug” the model and investigate downstream effects). We applied this technique to 325 metastatic melanoma tumor RNA-seq profiles downloaded from The Cancer Genome Atlas to characterize metabolic differences across patients. Results: We found considerable differences in metabolic function across metastatic melanoma patient tumors. One of the main drivers of patient-to-patient metabolic variability appeared to be BRAF mutations, with BRAF mutated tumors showing higher levels of glutathione metabolism, lower levels of glutamate metabolism, and lower levels of oxidative phosphorylation. In addition to these differences, we found that several metabolic pathways were associated with poor overall survival prognosis regardless of BRAF mutations status including high specific growth rate, high cholesterol metabolism, and high serotonin and melatonin biosynthesis. Furthermore, simulating knockouts of enzymes involved in these pathways resulted in a set of actionable targets that could enhance chemotherapy for metastatic melanoma patients. Conclusion: Using systems medicine, we have characterized metabolism in metastatic melanoma at high resolution in 325 patient tumors. This characterization revealed the variability in metastatic melanoma across patients and how this variability is associated with clinical outcome. Furthermore, our approach suggests additional therapeutic targets to enhance patient care. Citation Format: Daniel Cook, Dorys A. Lopez-Ramos, Patricia Carrigan, Joseph R. Peterson, The SimBioSys Team. Systems medicine based metabolic profiling of metastatic melanoma reveals insights into tumor biology, prognosis, and new therapeutic strategies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2714.
- Open Access
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 2715
Abstract 2715: Profiling the metabolism of pancreatic adenocarcinoma using a systems medicine-based approach reveals insights into tumor biology, prognosis, and new therapeutic strategies
Background: One of the hallmarks of cancer is metabolic dysregulation that has been therapeutically targeted with metabolism-based therapies, an important class of chemotherapeutics. However, there has previously been no systematic way to identify key targets for metabolism-based therapeutic interventions. Targeting tumor metabolism has been shown to increase the efficacy of standard chemotherapy in pre-clinical studies. Thus, more efforts need to be put in the discovery of novel targets especially for tumors with poor prognosis and response to chemotherapy, like pancreatic adenocarcinoma. Method: We designed a systems medicine-based technique to understand tumor metabolism in individual patients at a level of detail not previously achievable. We statistically integrated metabolic network modeling with RNA-sequencing data (RNA-seq). This allows us to integrate molecular data about individual tumors (from RNA-seq) with a curated knowledge base of how these molecules interact within a patient’s tumor (using metabolic network models). The result is a mathematical description, or model, of a patient-specific tumor’s metabolism that is able to be interrogated (i.e., it is high-dimensional like RNA-seq data) and able to be simulated (i.e., we can “drug” the model and investigate downstream effects). We applied this technique to 177 pancreatic adenocarcinoma tumor RNA-seq profiles downloaded from The Cancer Genome Atlas to characterize metabolic differences across patients. Results: We found that pancreatic adenocarcinoma tumors tended to be slow growing (median specific growth rate = 0.57 %/day, range = 0-1.4 %/day). This low growth rate helps explain why pancreatic adenocarcinoma tumors are resistant to chemotherapy - there is simply not enough cellular growth for chemo to be effective. However, this result suggests that a small subset of tumors with high growth rate will respond to chemo. In addition, we found that pancreatic adenocarcinoma tumors were separated into three metabolic subtypes: classical, glutamate-addicted, and oxidative. The classical metabolic phenotype tended to have better overall survival than glutamate-addicted or oxidative. Whereas high levels of glutamate metabolism (characterizing the glutamate-addicted tumors) or high levels of oxidative phosphorylation (characterizing the oxidative tumors) were associated with poor overall survival prognosis. Conclusion: Taken together, our systems medicine approach to characterizing patient specific tumor metabolism indicates that a one size fits all approach to chemotherapy in pancreatic adenocarcinoma is likely to fail; and different subsets of patients will likely respond to differently targeted therapies. Citation Format: Daniel Cook, Joseph R. Peterson, Patricia Carrigan, Tushar Pandey, John Cole, The SimBioSys Team. Profiling the metabolism of pancreatic adenocarcinoma using a systems medicine-based approach reveals insights into tumor biology, prognosis, and new therapeutic strategies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2715.
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 2715
Open AccessAbstract 2715: Profiling the metabolism of pancreatic adenocarcinoma using a systems medicine-based approach reveals insights into tumor biology, prognosis, and new therapeutic strategies
Background: One of the hallmarks of cancer is metabolic dysregulation that has been therapeutically targeted with metabolism-based therapies, an important class of chemotherapeutics. However, there has previously been no systematic way to identify key targets for metabolism-based therapeutic interventions. Targeting tumor metabolism has been shown to increase the efficacy of standard chemotherapy in pre-clinical studies. Thus, more efforts need to be put in the discovery of novel targets especially for tumors with poor prognosis and response to chemotherapy, like pancreatic adenocarcinoma. Method: We designed a systems medicine-based technique to understand tumor metabolism in individual patients at a level of detail not previously achievable. We statistically integrated metabolic network modeling with RNA-sequencing data (RNA-seq). This allows us to integrate molecular data about individual tumors (from RNA-seq) with a curated knowledge base of how these molecules interact within a patient’s tumor (using metabolic network models). The result is a mathematical description, or model, of a patient-specific tumor’s metabolism that is able to be interrogated (i.e., it is high-dimensional like RNA-seq data) and able to be simulated (i.e., we can “drug” the model and investigate downstream effects). We applied this technique to 177 pancreatic adenocarcinoma tumor RNA-seq profiles downloaded from The Cancer Genome Atlas to characterize metabolic differences across patients. Results: We found that pancreatic adenocarcinoma tumors tended to be slow growing (median specific growth rate = 0.57 %/day, range = 0-1.4 %/day). This low growth rate helps explain why pancreatic adenocarcinoma tumors are resistant to chemotherapy - there is simply not enough cellular growth for chemo to be effective. However, this result suggests that a small subset of tumors with high growth rate will respond to chemo. In addition, we found that pancreatic adenocarcinoma tumors were separated into three metabolic subtypes: classical, glutamate-addicted, and oxidative. The classical metabolic phenotype tended to have better overall survival than glutamate-addicted or oxidative. Whereas high levels of glutamate metabolism (characterizing the glutamate-addicted tumors) or high levels of oxidative phosphorylation (characterizing the oxidative tumors) were associated with poor overall survival prognosis. Conclusion: Taken together, our systems medicine approach to characterizing patient specific tumor metabolism indicates that a one size fits all approach to chemotherapy in pancreatic adenocarcinoma is likely to fail; and different subsets of patients will likely respond to differently targeted therapies. Citation Format: Daniel Cook, Joseph R. Peterson, Patricia Carrigan, Tushar Pandey, John Cole, The SimBioSys Team. Profiling the metabolism of pancreatic adenocarcinoma using a systems medicine-based approach reveals insights into tumor biology, prognosis, and new therapeutic strategies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2715.
- Open Access
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 469
Abstract 469: A multi-scale analysis and visualization platform for cancer data - deriving tumor microenvironment behavior from pathology and transcriptomics
Heterogeneity is a hallmark of cancer and perhaps one of the most important features associated with resistance to therapies and likelihood of recurrence and/or metastasis. Genomic instability contributes to genetic diversity, which leads to high levels of intratumoural heterogeneity. While genetic diversity is one contributor to phenotypic heterogeneity in tumors, the spatial variation in the tumor microenvironment (TME) also drives emergent phenotypic heterogeneity and intrinsic variation in resistance to drugs and the immune system. Ultimately, heterogeneity leads to higher likelihood of metastasis. Traditionally, pathology allowed for the assessment of heterogeneity. However, this was limited to gross assessment of the TME, through measurement of mitotic levels and severity of cancer invasion. We present SimBioSys PhenoScope, a multi-scale analysis and visualization platform that integrates cancer data across scales to extract cross modality trends that drive cancer invasion. As a demonstrated use case for the platform, we present a vignette of using the platform to analyze pathology slides at three scales. Three convolutional neural networks (CNN) are developed and validated. The outputs of these networks were combined with 2D simulations of the metabolic behavior and growth of cells within the TME. Two CNNs were developed and one implemented: one that identifies cells undergoing mitosis, one that segments individual cells and classifies their type, and one that segments five tissues from pathology slides. Additionally, transcriptional data was used to generate patient specific metabolic models. Each CNN was developed using training images and validated on the test images. The mitosis detection CNN was found to have an accuracy of 76.2% (precision=83%, recall=76%) in the test set. The classification CNN was found to have a Dice Similarity Coefficient (DSC) of 0.821 for segmenting cells and an F1 score for classifying cells ranging from 0.559 (F1i) to 0.756 (F1d). The segmentation CNN was found to have an accuracy of 78.1% with DSC ranging from 0.66 and 0.86 depending on tissue. The segmentations were input into a proprietary simulation framework along with patient-specific metabolic models to predict the spatial gradients of nutrients, and spatial organization of growth and metabolism. Simulations show multiple behaviors such as regions of high lactate production or consumption by cancer, and regions that differ by lactate production and alanine uptake in cancers. These behaviors were correlated with local cell mitoses and invasion of Tumor Infiltrating Lymphocytes. Tools to examine cancers across scales and within the TME are currently lacking. We demonstrate a proof-of-principle approach of combining data across scales in a fashion that allows for novel predictions of TME behavior. Citation Format: Joseph R. Peterson, Matthew Shin, Patricia Carrigan, Michael J. Hallock, Snehal Patel, The SimBioSys Team. A multi-scale analysis and visualization platform for cancer data - deriving tumor microenvironment behavior from pathology and transcriptomics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 469.
- 15 June 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (12_Supplem) , 469
Open AccessAbstract 469: A multi-scale analysis and visualization platform for cancer data - deriving tumor microenvironment behavior from pathology and transcriptomics
Heterogeneity is a hallmark of cancer and perhaps one of the most important features associated with resistance to therapies and likelihood of recurrence and/or metastasis. Genomic instability contributes to genetic diversity, which leads to high levels of intratumoural heterogeneity. While genetic diversity is one contributor to phenotypic heterogeneity in tumors, the spatial variation in the tumor microenvironment (TME) also drives emergent phenotypic heterogeneity and intrinsic variation in resistance to drugs and the immune system. Ultimately, heterogeneity leads to higher likelihood of metastasis. Traditionally, pathology allowed for the assessment of heterogeneity. However, this was limited to gross assessment of the TME, through measurement of mitotic levels and severity of cancer invasion. We present SimBioSys PhenoScope, a multi-scale analysis and visualization platform that integrates cancer data across scales to extract cross modality trends that drive cancer invasion. As a demonstrated use case for the platform, we present a vignette of using the platform to analyze pathology slides at three scales. Three convolutional neural networks (CNN) are developed and validated. The outputs of these networks were combined with 2D simulations of the metabolic behavior and growth of cells within the TME. Two CNNs were developed and one implemented: one that identifies cells undergoing mitosis, one that segments individual cells and classifies their type, and one that segments five tissues from pathology slides. Additionally, transcriptional data was used to generate patient specific metabolic models. Each CNN was developed using training images and validated on the test images. The mitosis detection CNN was found to have an accuracy of 76.2% (precision=83%, recall=76%) in the test set. The classification CNN was found to have a Dice Similarity Coefficient (DSC) of 0.821 for segmenting cells and an F1 score for classifying cells ranging from 0.559 (F1i) to 0.756 (F1d). The segmentation CNN was found to have an accuracy of 78.1% with DSC ranging from 0.66 and 0.86 depending on tissue. The segmentations were input into a proprietary simulation framework along with patient-specific metabolic models to predict the spatial gradients of nutrients, and spatial organization of growth and metabolism. Simulations show multiple behaviors such as regions of high lactate production or consumption by cancer, and regions that differ by lactate production and alanine uptake in cancers. These behaviors were correlated with local cell mitoses and invasion of Tumor Infiltrating Lymphocytes. Tools to examine cancers across scales and within the TME are currently lacking. We demonstrate a proof-of-principle approach of combining data across scales in a fashion that allows for novel predictions of TME behavior. Citation Format: Joseph R. Peterson, Matthew Shin, Patricia Carrigan, Michael J. Hallock, Snehal Patel, The SimBioSys Team. A multi-scale analysis and visualization platform for cancer data - deriving tumor microenvironment behavior from pathology and transcriptomics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 469.
- Open Access
- 15 February 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (4_Suppleme) , P5-13
Abstract P5-13-34: A multi-modal biomarker of immunotherapy response
Background Immunotherapy (IO) is expected to become a part of standard neoadjuvant therapy for early-stage, triple-negative breast cancer. However, only a minority of patients benefit from the addition of IO to standard neoadjuvant chemotherapy (NACT). Biomarkers of IO response in the metastatic setting, such as PD-L1 expression, fail to predict benefit from IO in the early-stage setting. Given the financial cost of IO and the potential for irreversible, immune-related toxicities, predictive biomarkers are desperately needed. SimBioSys TumorScope (TS), a commercially available biophysical model, has been previously validated as a highly accurate predictor of response to standard NAT regimens. TS has the potential to identify patients treated with IO who would have responded well to standard NACT, and those who truly benefited from IO, enabling biomarker discovery. Methods We identified 17 pts who received pembrolizumab in combination with standard NACT. From these cases, dynamic contrast-enhanced MRIs were assessed from multiple timepoints throughout treatment. Using a proprietary convolutional neural net, the tumor region from each MRI was segmented to assess volume over time. We compared actual tumor volumes to predicted response to standard NACT (without pembrolizumab) generated with the TS platform. By contrasting the tumor volume across the simulation and the longitudinal MRI timepoints, a metric of response directly attributable to pembrolizumab (“IO benefit”) was generated. Clinical, radiological, metabolic, tumor morphology, and microvasculature features were extracted from TS simulations, and after quality control and removal of redundant metrics, used in linear models as predictors of IO benefit. Results Nine clinical, seven metabolic, 16 microvasculature, ten tumor morphology, and 77 radiological features were used as individual predictors of IO benefit in linear models. Eight out of 17 cases had positive IO benefit values. In all, 12 out of the 119 metrics of interest nominally significant predictors of IO benefit (pacross simulation timepoints. Significant relationships were salient in hypoxic fraction measures taken at early simulation timepoints, including weeks three (p=0.017), two (p=0.024), and five (p=0.038). Importantly, between IO benefit/hypoxic fraction relationships, we observed concordant positive directions of effect (i.e., higher hypoxic cancer fraction, higher IO benefit). Conclusion IO is likely to soon gain regulatory approval for early-stage TNBC, but biomarkers to predict benefit are lacking. PD-L1 expression has previously been correlated with hypoxic tumor environments and high density of tumor infiltrating lymphocytes. Even within a limited dataset, TS provided a biologically explainable marker to identify patients who may benefit from IO. Work is underway to further validate these findings in a larger dataset, and to validate hypoxic tumor fraction as a predictor of IO response. Citation Format: JR Pfeiffer, Joseph R. Peterson, Frederick Howard, Tyler M. Earnest, Tushar Pandey, Gong He, Rita Nanda. A multi-modal biomarker of immunotherapy response [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P5-13-34.
- 15 February 2022
- Published by American Association for Cancer Research (AACR) in Cancer Research
- Vol. 82 (4_Suppleme) , P5-13
Open AccessAbstract P5-13-34: A multi-modal biomarker of immunotherapy response
Background Immunotherapy (IO) is expected to become a part of standard neoadjuvant therapy for early-stage, triple-negative breast cancer. However, only a minority of patients benefit from the addition of IO to standard neoadjuvant chemotherapy (NACT). Biomarkers of IO response in the metastatic setting, such as PD-L1 expression, fail to predict benefit from IO in the early-stage setting. Given the financial cost of IO and the potential for irreversible, immune-related toxicities, predictive biomarkers are desperately needed. SimBioSys TumorScope (TS), a commercially available biophysical model, has been previously validated as a highly accurate predictor of response to standard NAT regimens. TS has the potential to identify patients treated with IO who would have responded well to standard NACT, and those who truly benefited from IO, enabling biomarker discovery. Methods We identified 17 pts who received pembrolizumab in combination with standard NACT. From these cases, dynamic contrast-enhanced MRIs were assessed from multiple timepoints throughout treatment. Using a proprietary convolutional neural net, the tumor region from each MRI was segmented to assess volume over time. We compared actual tumor volumes to predicted response to standard NACT (without pembrolizumab) generated with the TS platform. By contrasting the tumor volume across the simulation and the longitudinal MRI timepoints, a metric of response directly attributable to pembrolizumab (“IO benefit”) was generated. Clinical, radiological, metabolic, tumor morphology, and microvasculature features were extracted from TS simulations, and after quality control and removal of redundant metrics, used in linear models as predictors of IO benefit. Results Nine clinical, seven metabolic, 16 microvasculature, ten tumor morphology, and 77 radiological features were used as individual predictors of IO benefit in linear models. Eight out of 17 cases had positive IO benefit values. In all, 12 out of the 119 metrics of interest nominally significant predictors of IO benefit (pacross simulation timepoints. Significant relationships were salient in hypoxic fraction measures taken at early simulation timepoints, including weeks three (p=0.017), two (p=0.024), and five (p=0.038). Importantly, between IO benefit/hypoxic fraction relationships, we observed concordant positive directions of effect (i.e., higher hypoxic cancer fraction, higher IO benefit). Conclusion IO is likely to soon gain regulatory approval for early-stage TNBC, but biomarkers to predict benefit are lacking. PD-L1 expression has previously been correlated with hypoxic tumor environments and high density of tumor infiltrating lymphocytes. Even within a limited dataset, TS provided a biologically explainable marker to identify patients who may benefit from IO. Work is underway to further validate these findings in a larger dataset, and to validate hypoxic tumor fraction as a predictor of IO response. Citation Format: JR Pfeiffer, Joseph R. Peterson, Frederick Howard, Tyler M. Earnest, Tushar Pandey, Gong He, Rita Nanda. A multi-modal biomarker of immunotherapy response [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P5-13-34.