Dormant estrogen receptor positive (ER+) breast cancer micrometastases in the bone marrow survive adjuvant chemotherapy and recur stochastically for more than 20 years. We hypothesized that inflammatory cytokines produced by stromal injury can re-awaken dormant breast cancer cells. We used an established in vitro dormancy model of Michigan Cancer Foundation-7 (MCF-7) breast cancer cells incubated at clonogenic density on fibronectin-coated plates to determine the effects of inflammatory cytokines on reactivation of dormant ER+ breast cancer cells. We measured induction of a mesenchymal phenotype, motility and the capacity to re-enter dormancy. We induced secretory senescence in murine stromal monolayers by oxidation, hypoxia and estrogen deprivation with hydrogen peroxide (H2O2), carbonyl-cyanide m-chlorophenylhydrazzone (CCCP) and Fulvestrant (ICI 182780), respectively, and determined the effects on growth of co-cultivated breast cancer cells. Exogenous recombinant human (rh) interleukin (IL)-6, IL-8 or transforming growth factor β1 (TGFβ1) induced regrowth of dormant MCF-7 cells on fibronectin-coated plates. Dormant cells had decreased expression of E-cadherin and estrogen receptor α (ERα) and increased expression of N-cadherin and SNAI2 (SLUG). Cytokine or TGFβ1 treatment of dormant clones induced formation of growing clones, a mesenchymal appearance, increased motility and an impaired capacity to re-enter dormancy. Stromal injury induced secretion of IL-6, IL-8, upregulated tumor necrosis factor alpha (TNFα), activated TGFβ and stimulated the growth of co-cultivated MCF-7 cells. MCF-7 cells induced secretion of IL-6 and IL-8 by stroma in co-culture. Dormant ER+ breast cancer cells have activated epithelial mesenchymal transition (EMT) gene expression programs and downregulated ERα but maintain a dormant epithelial phenotype. Stromal inflammation reactivates these cells, induces growth and a mesenchymal phenotype. Reactivated, growing cells have an impaired ability to re-enter dormancy. In turn, breast cancer cells co-cultured with stroma induce secretion of IL-6 and IL-8 by the stroma, creating a positive feedback loop.

Differentiated epithelial cells reside in the homeostatic microenvironment of the native organ stroma. The stroma supports their normal function, their G₀ differentiated state, and their expansion/contraction through the various stages of the life cycle and physiologic functions of the host. When malignant transformation begins, the microenvironment tries to suppress and eliminate the transformed cells, while cancer cells, in turn, try to resist these suppressive efforts. The tumor microenvironment encompasses a large variety of cell types recruited by the tumor to perform different functions, among which fibroblasts are the most abundant. The dynamics of the mutual relationship change as the sides undertake an epic battle for control of the other. In the process, the cancer “wounds” the microenvironment through a variety of mechanisms and attracts distant mesenchymal stem cells to change their function from one attempting to suppress the cancer, to one that supports its growth, survival, and metastasis. Analogous reciprocal interactions occur as well between disseminated cancer cells and the metastatic microenvironment, where the microenvironment attempts to eliminate cancer cells or suppress their proliferation. However, the altered microenvironmental cells acquire novel characteristics that support malignant progression. Investigations have attempted to use these traits as targets of novel therapeutic approaches.

Up to 40% of patients with breast cancer (BC) have metastatic cells in the bone marrow (BM) at the initial diagnosis of localized disease. Despite definitive systemic adjuvant therapy, these cells survive in the BM microenvironment, enter a dormant state and recur stochastically for more than 20 years. Once they begin to proliferate, recurrent macrometastases are not curable, and patients generally succumb to their disease. Many potential mechanisms for initiating recurrence have been proposed, but no definitive predictive data have been generated. This manuscript reviews the proposed mechanisms that maintain BC cell dormancy in the BM microenvironment and discusses the data supporting specific mechanisms for recurrence. It addresses the well-described mechanisms of secretory senescence, inflammation, aging, adipogenic BM conversion, autophagy, systemic effects of trauma and surgery, sympathetic signaling, transient angiogenic bursts, hypercoagulable states, osteoclast activation, and epigenetic modifications of dormant cells. This review addresses proposed approaches for either eliminating micrometastases or maintaining a dormant state.

BACKGROUND: African American race negatively impacts survival from localized breast cancer but co-variable factors confound the impact.

Background and objectives: Deep learning (DL)-based models for predicting the survival of patients with local stages of breast cancer only use time-fixed covariates, i.e., patient and cancer data at the time of diagnosis. These predictions are inherently error-prone because they do not consider time-varying events that occur after initial diagnosis. Our objective is to improve the predictive modeling of survival of patients with localized breast cancer to consider both time-fixed and time-varying events; thus, we take into account the progression of a patient’s health status over time. Methods: We extended four DL-based predictive survival models (DeepSurv, DeepHit, Nnet-survival, and Cox-Time) that deal with right-censored time-to-event data to consider not only a patient’s time-fixed covariates (patient and cancer data at diagnosis) but also a patient’s time-varying covariates (e.g., treatments, comorbidities, progressive age, frailty index, adverse events from treatment). We utilized, as our study data, the SEER-Medicare linked dataset from 1991 to 2016 to study a population of women diagnosed with stage I–III breast cancer (BC) enrolled in Medicare at 65 years or older as qualified by age. We delineated time-fixed variables recorded at the time of diagnosis, including age, race, marital status, breast cancer stage, tumor grade, laterality, estrogen receptor (ER), progesterone receptor (PR), and human epidermal receptor 2 (HER2) status, and comorbidity index. We analyzed six distinct prognostic categories, cancer stages I–III BC, and each stage’s ER/PR+ or ER/PR− status. At each visit, we delineated the time-varying covariates of administered treatments, induced adverse events, comorbidity index, and age. We predicted the survival of three hypothetical patients to demonstrate the model’s utility. Main Outcomes and Measures: The primary outcomes of the modeling were the measures of the model’s prediction error, as measured by the concordance index, the most commonly applied evaluation metric in survival analysis, and the integrated Brier score, a metric of the model’s discrimination and calibration. Results: The proposed extended patients’ covariates that include both time-fixed and time-varying covariates significantly improved the deep learning models’ prediction error and the discrimination and calibration of a model’s estimates. The prediction of the four DL models using time-fixed covariates in six different prognostic categories all resulted in approximately a 30% error in all six categories. When applying the proposed extension to include time-varying covariates, the accuracy of all four predictive models improved significantly, with the error decreasing to approximately 10%. The models’ predictive accuracy was independent of the differing published survival predictions from time-fixed covariates in the six prognostic categories. We demonstrate the utility of the model in three hypothetical patients with unique patient, cancer, and treatment variables. The model predicted survival based on the patient’s individual time-fixed and time-varying features, which varied considerably from Social Security age-based, and stage and race-based breast cancer survival predictions. Conclusions: The predictive modeling of the survival of patients with early-stage breast cancer using DL models has a prediction error of around 30% when considering only time-fixed covariates at the time of diagnosis and decreases to values under 10% when time-varying covariates are added as input to the models, regardless of the prognostic category of the patient groups. These models can be used to predict individual patients’ survival probabilities based on their unique repertoire of time-fixed and time-varying features. They will provide guidance for patients and their caregivers to assist in decision making.