By Dorys Lopez

After two consecutive years of virtually holding the annual meeting of the American Association for Cancer Research (AACR), the conference returned to an in-person format, while retaining a hybrid option for those who prefer to join virtually. There were over 19,000 registrants, with approximately 78% attending in-person, according to Margaret Foti, chief executive officer of the AACR.

One of our team members is a seasoned AACR attendee, with this being her 24th AACR annual meeting. For the rest of us newbies, we could not have imagined the extent of the conference: from the breath of topics covered by countless posters, talks and booths, to the size of the convention center. There was so much to take in with a three-day period (without counting the educational program). It was encouraging to see all the exciting research emerging in the cancer field and even more exciting to be part of it. Our team did a great job presenting nine posters, all highly attended, addressing high impact questions in cancer care across a spectrum of topics.

It would be impossible to cover the entire variety of topics presented at the conference, so we selected a couple of them to highlight:

Cancer disparities. This topic was discussed throughout many sessions. During a plenary session, Mariana Stern, Director for the Molecular Epidemiology MS/PhD Program at USC, emphasized the importance of progress in cancer research reaching all populations and individuals. Cancer research has mainly relied on samples that come from Caucasian (white) patients. Efforts are now in place to increase diversity, to understand how race and ethnicity influence patient care and outcomes. Melissa Davis, Scientific Director of the International Center Study of Breast Cancer Subtypes, talked about the impact of genetic ancestry and social determinants on disparities, focusing primarily on African ancestry and how it relates or differs from observations in African American cohorts. A full session was dedicated to breast cancer genomics in Latin America, where Laura Fejerman, Associated Professor in the Department of Public Health Sciences School of Medicine at the University of California Davis, emphasized the lack of diversity in the breast cancer genome-wide association studies (GWAS). She now leads the Latin America Genetics and Genomics of Breast Cancer Consortium (LAGENO-BC), which intends to expand the study of the genetics of breast cancer in Latina women.
Our team presented our latest research on racial can abstract and poster titled “Integrating imaging and multi-omics data to elucidate racial differences in breast tumor biology to optimize precision oncology approaches and patient outcome” where African American tumors in the analyzed cohort grew faster, had more adipose tissue around the tumor, were stiffer and more heterogenous in drug distribution with lower spatial blood flow distribution than tumors from Caucasian patients. We agree that the characterization of racial-specific biological factors is crucial as it may lead to new strategies for personalized treatment.

Triple Negative Breast Cancer (TNBC). A full session was dedicated to this breast cancer subtype. Edison Liu, professor and President Emeritus of The Jackson Laboratory (JAX), identified differences in TNBC response to platinum therapy based on BRCA1 promoter methylation or BRCA1/2 mutations. He found TNBC with BRCA1 promoter methylation to be associated to resistance to platinum therapy, while that is not the case for TNBC with BRCA1 mutations. In line with therapy resistance, Lisa Carey, L. Richardson and Marilyn Jacobs Preyer Distinguished Professor in Breast Cancer Research in the Department of Medicine at the University of North Carolina (UNC), talked about the challenges of TNBC therapy. She walked the audience through the therapies that showed preclinical and early clinical promise but failed to reach late clinical phase success, such as therapies targeting EGFR, the PI3K/AKT pathway and the Ras/Raf/MEK pathway. She mentioned the TNBC luminal-androgen receptor (LAR) subtype, and the potential to target this with endocrine therapy, although, until now the clinical rate benefit has been modest (19-20%). A new wave of antibody drug conjugates (ADCs) to treat TNBC is developing. These antibody-drug conjugates target Trop2, a protein expressed on the cell surface of about 80% of TNBCs, although it is also expressed in many healthy cell types. Scituzumab govitecan-hzyi (AB SN38) and Dato-DXd, are two of the ADCs that are being evaluated. Finally, she highlighted the need of better biomarkers for immunotherapy, where PD-L1 works as a biomarker in metastatic TNBC but not in non-metastatic TNBC.
This breast cancer subtype is considered the most aggressive type. Tumor aggressiveness plays an important role in escalation of care for TNBC patients. Our team presented an abstract and poster on “Accurate prediction of tumor growth and doubling times for triple negative breast cancer (TNBC) allows for patient-specific assessment of tumor aggressiveness”. Using TumorScope’s biophysical simulation platform, our team accurately calculated tumor growth (specific growth rate) using RNA-seq data, and tumor volume doubling time (TVDT) using metabolic models. This model can be used to assess the aggressiveness of the tumor in TNBC patients, providing valuable information that could potentially support treatment selection.

Single cell, spatial transcriptomics and spatial analysis of tumors. In the last couple of years, the spatial analysis of tumors has rapidly grown and become a main topic of interest. Research has shown that is not only important to know what type of cells are present in the tumor, it is equally important to understand how they are distributed across the tumor, as well as how they interact with each other and with the microenvironment.
A full minisymposium session and poster section were dedicated to single cell and spatial transcriptomics. Linghua Wang, tenure-track Assistant Professor at Department of Genomic Medicine, explained the many applications of single-cell RNA sequencing approaches during the introduction of the session. With single-cell RNA sequencing it is possible to assess the tumor cell heterogeneity by profiling diverse subpopulation of tumor cell types, cell states, cell composition and tissue distribution. Additionally, it can be used for correlative analysis of biomarkers. With spatial transcriptomics it is possible to do spatial clustering to study cell patterns and cell type decomposition. High-quality data can be extracted from FFPE tissue and integrated with histology to do spatial analysis of these samples. In the Deep Learning in imaging session, Jeremy Goecks, section Head for Cancer Data Science and an Associate Professor of Biomedical Engineering at Oregon Health & Science University (OHSU), talked precisely about deep learning in the context of multiplex tissue imaging as a next generation IHC. He highlighted the use of spatial proteomics assay in FFPE slides, which yield high-dimensional single-cell proteomics data, and spatial position of cells. The applications of this technique for precision oncology include understanding how these features/phenotypes correlate with therapy response, resistance and recurrence. In the same session, Joel Saltz, Cherith Professor and Founding Chair, Department of Biomedical Informatics at Stony Brook University, talked about the use of whole slide imaging to understand tumor lymphocyte infiltration (TILs). He highlighted the use of ‘pathomics’ tissue analytics for precision medicine, specifically to understand the spatial distribution of TILs and how spatial TIL patterns can be used to predict treatment response and outcome, and to stratify patients. Maryam Pourmaleki, a PhD candidate at the Tri-Institutional PhD Program in Computational Biology and Medicine at Weill Cornell Medical College, also talked about the importance of the spatial view of immune cell function in cancer to identify biomarkers of immunotherapy response. In her talk she described different spatial features, such as the presence of MHCI and a certain type of CD8 T cells (PD1+LAG3+TIM3+) in the tumor region of responders to intralesional IL-2 treatment of melanoma. She underscored the spatial topology of the tumor microenvironment as a key to understanding the interplay of the tumor and immune cells.
Our TumorScope technology is able to do spatio-temporal modeling of the tumor microenvironment, whereby modelling nutrient transport and drug behavior it can predict how different regions of the tumor will respond to therapy and how the tumor changes in size, shape and morphology over time. TumorScope’s model was presented in the abstract and poster “Spatio-temporal modeling of the tumor microenvironment for prediction of patient-specific response to chemotherapy”. We are also able to integrate RNA-seq data in a systems biology approach that allows for the identification of different metabolic subtypes that could have therapeutic implications. Our team presented three posters on this topic. An explanation to our systems-based approach can be found in the abstract and poster “Using systems medicine for comprehensive metabolic profiling of tumors: how tumor metabolism shapes prognosis and response to chemotherapy.”

Computational Oncology and artificial intelligence. Computational oncology has been widely used to develop predictive models for drug response. Erica Silva, an MD/PhD candidate at the University of California, talked about Nested Systems in Tumors (NeST) in combination with visible neural networks (VNN), which allow for an understanding of how and why models make their predictions on patients and clinical samples. She built a model for Palbociblib, a CDK4/6 inhibitor used to treat advanced breast cancer. In her study using the project GENIE cohort she showed that her model was specific for Palbociclib response rather than just prognostic of overall survival. Our team also presented data on our predictive model for drug response that is integrated in our TumorScope platform. This data was presented in collaboration with the University of Cincinnati. In the study, TumorScope predicted the response to neoadjuvant therapy (pathological complete response and residual disease) across all breast cancer subtypes with high accuracy (91.2%) in 81 patients. This model accounts for the different types of standard-of-care regimens that the patients received as neoadjuvant therapy, so the predictions are made on a patient-specific manner based on individual patient characteristics from data collected at diagnosis. This data was presented in the abstract and poster “SimBioSys® TumorScopeTM accurately models and predicts response to neoadjuvant therapy in Breast Cancer – Validation study.”

To read the full SimBioSys’ abstracts that were presented at the AACR22 you can click here.