The abundance of tumor-infiltrating lymphocytes has been associated with a favorable prognosis in estrogen receptor-negative breast cancer. However, a high degree of spatial heterogeneity in lymphocytic infiltration is often observed and its clinical implication remains unclear. Here we combine automated histological image processing with methods of spatial statistics used in ecological data analysis to quantify spatial heterogeneity in the distribution patterns of tumor-infiltrating lymphocytes. Hematoxylin and eosin-stained sections from two cohorts of estrogen receptor-negative breast cancer patients (discovery: n = 120; validation: n = 125) were processed with our automated cell classification algorithm to identify the location of lymphocytes and cancer cells. Subsequently, hotspot analysis (Getis-Ord Gi*) was applied to identify statistically significant hotspots of cancer and immune cells, defined as tumor regions with a significantly high number of cancer cells or immune cells, respectively. We found that the amount of co-localized cancer and immune hotspots weighted by tumor area, rather than number of cancer or immune hotspots, correlates with a better prognosis in estrogen receptornegative breast cancer in univariate and multivariate analysis. Moreover, co-localization of cancer and immune hotspots further stratified patients with immune cell-rich tumors. Our study demonstrates the importance of quantifying not only the abundance of lymphocytes but also their spatial variation in the tumor specimen for which methods from other disciplines such as spatial statistics can be successfully applied.

IntroductionAbundance of immune cells has been shown to have prognostic and predictive significance in many tumor types. Beyond abundance, the spatial organization of immune cells in relation to cancer cells may also have significant functional and clinical implications. However there is a lack of systematic methods to quantify spatial associations between immune and cancer cells.MethodsWe applied ecological measures of species interactions to digital pathology images for investigating the spatial associations of immune and cancer cells in breast cancer. We used the Morisita-Horn similarity index, an ecological measure of community structure and predator–prey interactions, to quantify the extent to which cancer cells and immune cells colocalize in whole-tumor histology sections. We related this index to disease-specific survival of 486 women with breast cancer and validated our findings in a set of 516 patients from different hospitals.ResultsColocalization of immune cells with cancer cells was significantly associated with a disease-specific survival benefit for all breast cancers combined. In HER2-positive subtypes, the prognostic value of immune-cancer cell colocalization was highly significant and exceeded those of known clinical variables. Furthermore, colocalization was a significant predictive factor for long-term outcome following chemotherapy and radiotherapy in HER2 and Luminal A subtypes, independent of and stronger than all known clinical variables.ConclusionsOur study demonstrates how ecological methods applied to the tumor microenvironment using routine histology can provide reproducible, quantitative biomarkers for identifying high-risk breast cancer patients. We found that the clinical value of immune-cancer interaction patterns is highly subtype-specific but substantial and independent to known clinicopathologic variables that mostly focused on cancer itself. Our approach can be developed into computer-assisted prediction based on histology samples that are already routinely collected.Electronic supplementary materialThe online version of this article (doi:10.1186/s13058-015-0638-4) contains supplementary material, which is available to authorized users.

SUMMARY:Immunohistochemical expression analysis of mismatch repair gene products has been suggested for the prediction of hereditary nonpolyposis colorectal cancer (HNPCC) carrier status in cancer families and the selection of microsatellite instability (MSI)-positive tumors in sporadic colorectal cancer. In this study, we aimed to evaluate hMSH2 and hMLH1 immunohistochemistry in familial and sporadic colorectal cancer. We found that immunohistochemistry allowed us to identify patients with germline mutations in hMSH2 and many cases with germline mutations in hMLH1. However, some missense and truncating mutations may be missed. In addition, hMLH1 promoter methylation, commonly occurring in familial and sporadic MSI-positive colorectal cancer, can complicate the interpretation of immunohistochemical expression analyses. Our results suggest that immunohistochemistry cannot replace testing for MSI to predict HNPCC carrier status or identify MSI-positive sporadic colorectal cancer. (Lab Invest 2001, 81:535-541).

We have identified a gene, ST18 (suppression of tumorigenicity 18, breast carcinoma, zinc-finger protein), within a frequent imbalanced region of chromosome 8q11 as a breast cancer tumor suppressor gene. The ST18 gene encodes a zinc-finger DNA-binding protein with six fingers of the C2HC type (configuration Cys-X 5 -Cys-X 12 -His-X 4 -Cys) and an SMC domain. ST18 has the potential to act as transcriptional regulator. ST18 is expressed in a number of normal tissues including mammary epithelial cells although the level of expression is quite low. In breast cancer cell lines and the majority of primary breast tumors, ST18 mRNA is significantly downregulated. A 160 bp region within the promoter of the ST18 gene is hypermethylated in about 80% of the breast cancer samples and in the majority of breast cancer cell lines. The strong correlation between ST18 promoter hypermethylation and loss of ST18 expression in tumor cells suggests that this epigenetic mechanism is responsible for tumor-specific downregulation. We further show that ectopic ST18 expression in MCF-7 breast cancer cells strongly inhibits colony formation in soft agar and the formation of tumors in a xenograft mouse model.