Digital image analysis and machine learning-assisted prediction of neoadjuvant chemotherapy response in triple-negative breast cancer.

BACKGROUND: Pathological complete response (pCR) is associated with favorable prognosis in patients with triple-negative breast cancer (TNBC). However, only 30-40% of TNBC patients treated with neoadjuvant chemotherapy (NAC) show pCR, while the remaining 60-70% show residual disease (RD). The role of the tumor microenvironment in NAC response in patients with TNBC remains unclear. In this study, we developed a machine learning-based two-step pipeline to distinguish between various histological components in hematoxylin and eosin (H&E)-stained whole slide images (WSIs) of TNBC tissue biopsies and to identify histological features that can predict NAC response. METHODS: H&E-stained WSIs of treatment-naïve biopsies from 85 patients (51 with pCR and 34 with RD) of the model development cohort and 79 patients (41 with pCR and 38 with RD) of the validation cohort were separated through a stratified eightfold cross-validation strategy for the first step and leave-one-out cross-validation strategy for the second step. A tile-level histology label prediction pipeline and four machine-learning classifiers were used to analyze 468,043 tiles of WSIs. The best-trained classifier used 55 texture features from each tile to produce a probability profile during testing. The predicted histology classes were used to generate a histology classification map of the spatial distributions of different tissue regions. A patient-level NAC response prediction pipeline was trained with features derived from paired histology classification maps. The top graph-based features capturing the relevant spatial information across the different histological classes were provided to the radial basis function kernel support vector machine (rbfSVM) classifier for NAC treatment response prediction. RESULTS: The tile-level prediction pipeline achieved 86.72% accuracy for histology class classification, while the patient-level pipeline achieved 83.53% NAC response (pCR vs. RD) prediction accuracy of the model development cohort. The model was validated with an independent cohort with tile histology validation accuracy of 83.59% and NAC prediction accuracy of 81.01%. The histological class pairs with the strongest NAC response predictive ability were tumor and tumor tumor-infiltrating lymphocytes for pCR and microvessel density and polyploid giant cancer cells for RD. CONCLUSION: Our machine learning pipeline can robustly identify clinically relevant histological classes that predict NAC response in TNBC patients and may help guide patient selection for NAC treatment.

Quantitative proteomics reveals serum proteome alterations during metastatic disease progression in breast cancer patients.

BACKGROUND: Tumor recurrence and metastatic progression remains the leading cause for breast cancer related mortalities. However, the proteomes of patient- matched primary breast cancer (BC) and metastatic lesions have not yet been identified, due to the lack of clinically annotated longitudinal samples. In this study, we evaluated the global-proteomic landscape of BC patients with and without distant metastasis as well as compared the proteome of distant metastatic disease with its corresponding primary BC, within the same patient. METHODS: We performed mass spectrometry-based proteome profiling of 73 serum samples from 51 BC patients. Among the 51 patients with BC, 29 remained metastasis-free (henceforth called non-progressors), and 22 developed metastases (henceforth called progressors). For the 22 progressors, we obtained two samples: one collected within a year of diagnosis, and the other collected within a year before the diagnosis of metastatic disease. MS data were analyzed using intensity-based absolute quantification and normalized before differential expression analysis. Significantly differentially expressed proteins (DEPs; absolute fold-change ≥ 1.5, P-value < 0.05 and 30% abundance per clinical group) were subjected to pathway analyses. RESULTS: We identified 967 proteins among 73 serum samples from patients with BC. Among these, 39 proteins were altered in serum samples at diagnosis, between progressors and non-progressors. Among these, 4 proteins were further altered when the progressors developed distant metastasis. In addition, within progressors, 20 proteins were altered in serum collected at diagnosis versus at the onset of metastasis. Pathway analysis showed that these proteins encoded pathways that describe metastasis, including epithelial-mesenchymal transition and focal adhesion that are hallmarks of metastatic cascade. CONCLUSIONS: Our results highlight the importance of examining matched samples from distant metastasis with primary BC samples collected at diagnosis to unravel subset of proteins that could be involved in BC progression in serum. This study sets the foundation for additional future investigations that could position these proteins as non-invasive markers for clinically monitoring breast cancer progression in patients.

A novel role for KIFC1-MYH9 interaction in triple-negative breast cancer aggressiveness and racial disparity.

African American (AA) women are twice as likely to develop triple-negative breast cancer (TNBC) as women of European descent. Additionally, AA women with TNBC present a much more aggressive disease course than their European American (EA) counterparts. Thus, there is an unmet clinical need to identify race-specific biomarkers and improve survival outcomes in AA patients with TNBC. The minus-end directed microtubule motor protein kinesin family member C1 (KIFC1) promotes centrosome clustering and chromosomal instability and is often overexpressed in TNBC. Previous findings suggest that KIFC1 plays a role in cell proliferation and migration in TNBC cells from AAs and that the levels of nuclear KIFC1 (nKIFC1) are particularly high in AA patients with TNBC. The nuclear localization of KIFC1 in interphase may underlie its previously unrecognized race-specific association. In this study, we found that in TNBC cells derived from AAs, nKIFC1 interacted with the tumor suppressor myosin heavy chain 9 (MYH9) over EA cells. Treatment of AA TNBC cells with commercial inhibitors of KIFC1 and MYH9 disrupted the interaction between KIFC1 and MYH9. To characterize the racial differences in the KIFC1-MYH9-MYC axis in TNBC, we established homozygous KIFC1 knockout (KO) TNBC cell lines. KIFC1 KO significantly inhibited proliferation, migration, and invasion in AA TNBC cells but not in EA TNBC cells. RNA sequencing analysis showed significant downregulation of genes involved in cell migration, invasion, and metastasis upon KIFC1 KO in TNBC cell lines from AAs compared to those from EAs. These data indicate that mechanistically, the role of nKIFC1 in driving TNBC progression and metastasis is stronger in AA patients than in EA patients, and that KIFC1 may be a critical therapeutic target for AA patients with TNBC.

Demographic disparities in the limited awareness of alcohol use as a breast cancer risk factor: empirical findings from a cross-sectional study of U.S. women.

BACKGROUND: Alcohol use is an established yet modifiable risk factor for breast cancer. However, recent research indicates that the vast majority of U.S. women are unaware that alcohol use is a risk factor for breast cancer. There is limited information about the sociodemographic characteristics and alcohol use correlates of awareness of the alcohol use and breast cancer link, and this is critically important for health promotion and intervention efforts. In this study, we assessed prevalence of the awareness of alcohol use as a risk factor for breast cancer among U.S. women and examined sociodemographic and alcohol use correlates of awareness of this link. METHODS: We conducted a 20-minute online cross-sectional survey, called the ABLE (Alcohol and Breast Cancer Link Awareness) survey, among U.S. women aged 18 years and older (N = 5,027) in the fall of 2021. Survey questions assessed awareness that alcohol use increases breast cancer risk (yes, no, don't know/unsure); past-year alcohol use and harmful drinking via the Alcohol Use Disorders Identification Test (AUDIT); and family, health, and sociodemographic characteristics. We conducted multivariate multinomial regression analysis to identify correlates of awareness that alcohol use increases breast cancer risk. RESULTS: Overall, 24.4% reported that alcohol use increased breast cancer risk, 40.2% reported they were unsure, and 35.4% reported that there was no link between alcohol use and breast cancer. In adjusted analysis, awareness of alcohol use as a breast cancer risk factor, compared to not being aware or unsure, was associated with being younger (18-25 years old), having a college degree, and having alcohol use disorder symptoms. Black women were less likely than white women to report awareness of the alcohol use and breast cancer link. CONCLUSIONS: Overall, only a quarter of U.S. women were aware that alcohol use increases breast cancer risk, although 40% expressed uncertainty. Differences in awareness by age, level of education, race and ethnicity and level of alcohol use offer opportunities for tailored prevention interventions, while the overall low level of awareness calls for widespread efforts to increase awareness of the breast cancer risk from alcohol use among U.S. women.

Integrative spatial omics reveals distinct tumor-promoting multicellular niches and immunosuppressive mechanisms in African American and European American patients with TNBC.

Racial disparities in triple-negative breast cancer (TNBC) outcomes have been reported. However, the biological mechanisms underlying these disparities remain unclear. We integrated imaging mass cytometry and spatial transcriptomics, to characterize the tumor microenvironment (TME) of African American (AA) and European American (EA) patients with TNBC. The TME in AA patients was characterized by interactions between endothelial cells, macrophages, and mesenchymal-like cells, which were associated with poor patient survival. In contrast, the EA TNBC-associated niche is enriched in T-cells and neutrophils suggestive of an exhaustion and suppression of otherwise active T cell responses. Ligand-receptor and pathway analyses of race-associated niches found AA TNBC to be immune cold and hence immunotherapy resistant tumors, and EA TNBC as inflamed tumors that evolved a distinctive immunosuppressive mechanism. Our study revealed the presence of racially distinct tumor-promoting and immunosuppressive microenvironments in AA and EA patients with TNBC, which may explain the poor clinical outcomes.

Predicting Neoadjuvant Treatment Response in Triple-Negative Breast Cancer Using Machine Learning.

BACKGROUND: Neoadjuvant chemotherapy (NAC) is the standard treatment for early-stage triple negative breast cancer (TNBC). The primary endpoint of NAC is a pathological complete response (pCR). NAC results in pCR in only 30-40% of TNBC patients. Tumor-infiltrating lymphocytes (TILs), Ki67 and phosphohistone H3 (pH3) are a few known biomarkers to predict NAC response. Currently, systematic evaluation of the combined value of these biomarkers in predicting NAC response is lacking. In this study, the predictive value of markers derived from H&E and IHC stained biopsy tissue was comprehensively evaluated using a supervised machine learning (ML)-based approach. Identifying predictive biomarkers could help guide therapeutic decisions by enabling precise stratification of TNBC patients into responders and partial or non-responders. METHODS: Serial sections from core needle biopsies (n = 76) were stained with H&E and immunohistochemically for the Ki67 and pH3 markers, followed by whole-slide image (WSI) generation. The serial section stains in H&E stain, Ki67 and pH3 markers formed WSI triplets for each patient. The resulting WSI triplets were co-registered with H&E WSIs serving as the reference. Separate mask region-based CNN (MRCNN) models were trained with annotated H&E, Ki67 and pH3 images for detecting tumor cells, stromal and intratumoral TILs (sTILs and tTILs), Ki67+, and pH3+ cells. Top image patches with a high density of cells of interest were identified as hotspots. Best classifiers for NAC response prediction were identified by training multiple ML models and evaluating their performance by accuracy, area under curve, and confusion matrix analyses. RESULTS: Highest prediction accuracy was achieved when hotspot regions were identified by tTIL counts and each hotspot was represented by measures of tTILs, sTILs, tumor cells, Ki67+, and pH3+ features. Regardless of the hotspot selection metric, a complementary use of multiple histological features (tTILs, sTILs) and molecular biomarkers (Ki67 and pH3) resulted in top ranked performance at the patient level. CONCLUSIONS: Overall, our results emphasize that prediction models for NAC response should be based on biomarkers in combination rather than in isolation. Our study provides compelling evidence to support the use of ML-based models to predict NAC response in patients with TNBC.