The presence of tumor associated macrophages in tumor stroma as a prognostic marker for breast cancer patients

A macrophage is one of the immune system's primary weapons—a cell whose job is to recognize threats and destroy them. Here is a disquieting fact: inside a breast tumor, macrophages are often abundant, and they help the cancer grow. Not because they have been corrupted in some dramatic way, but because tumors are extraordinarily good at hijacking the normal biology of inflammation and repair. Here is the detail that makes this clinically actionable: it is not just that these cells are present. It is where they sit inside the tumor, within a few millimetres of tissue, that tells you whether the patient will survive. That spatial insight is what a study by Medrek and colleagues set out to establish, using one hundred forty-four invasive breast cancer cases and a careful comparison of two immune markers. The findings reframe how we should think about immune infiltration in solid tumors. To understand what Medrek and colleagues found, you need a quick map of macrophage biology. Macrophages can be polarized into two broad functional states. M1 macrophages are the fighters—pro-inflammatory, capable of killing pathogens, and tumoricidal, meaning they can kill cancer cells. M2 macrophages are the healers—they downregulate inflammation, promote the growth of new blood vessels, recruit fibroblasts, and remodel connective tissue. Tumor-associated macrophages, or TAMs, share most of their characteristics with M2 macrophages. Instead of attacking the tumor, they are effectively doing its renovation work—building the blood supply, loosening the extracellular matrix to make room for invasion, and suppressing the immune responses that might otherwise control the cancer. Bingle and colleagues, reviewing the literature back in two thousand two, found that the majority of publications on TAMs in cancer, including breast cancer, linked high TAM infiltration to poor patient outcomes. The association was consistent, but the measurement was not. Most studies in that earlier literature used a marker called CD68, which recognizes macrophages broadly—both M1 and M2 populations together. The problem with a pan-macrophage marker is that it lumps together cells with opposite functions. Medrek and colleagues proposed that a more specific marker might sharpen the signal. CD163 is a scavenger receptor upregulated specifically in anti-inflammatory environments, and it is regarded as a highly specific marker for M2 macrophages. If TAMs are functionally M2-like, CD163 should be a cleaner probe. Before this study, CD163 had not been evaluated as a TAM marker in primary breast cancer. The research team built tissue microarrays—thin slices of archived tumor samples from one hundred forty-four patients—and stained them by immunohistochemistry for both CD163 and CD68. Critically, they scored each marker separately in two compartments: the tumor stroma, which is the connective tissue surrounding the cancer cells, and the tumor nest, which is the cancer cells themselves and the tissue immediately among them. This distinction turned out to be everything. The cohort was predominantly luminal A subtype—seventy-nine percent of cases—with eleven percent triple-negative or basal-like. The median follow-up was six point five five years. During that time, forty-one patients died and twenty-nine had disease recurrence. Dense CD163 infiltration was present in seventeen percent of tumors in the stroma and only nine percent in the nest. Dense CD68 infiltration was even rarer—nine percent in stroma and six percent in nest. What those numbers revealed, once the team ran their correlations, was a stark spatial divide. CD163 in the tumor stroma correlated with virtually every marker of aggressive disease. Higher histological grade, larger tumor size, elevated Ki67—Ki67 being a protein that marks actively dividing cells, a proxy for how fast the tumor is growing—estrogen receptor negativity, progesterone receptor negativity, and a strong association with triple-negative and basal-like subtypes. The Spearman correlations for those subtype associations were above zero point six, which is high for biological data. The inverse was equally clear: CD163 in stroma was negatively associated with the luminal A subtype, the most indolent and hormone-responsive form of breast cancer. Meanwhile, CD163 in the tumor nest showed none of this—zero significant associations with any clinicopathological feature. The same location-dependent pattern held for CD68 in stroma, which correlated with tumor size and grade and inversely with luminal A status—though with somewhat weaker correlations than CD163. The team also cross-checked their immunohistochemistry findings against a public gene expression dataset, GEO dataset GDS806. Breast cancer patients who received endocrine therapy and subsequently had recurrence had significantly higher CD68 gene expression than those who did not recur. Basal-like tumors showed significantly higher CD163 gene expression than luminal tumors. The molecular data lined up with the protein-level staining. That kind of convergence across two independent measurement platforms is meaningful. Now for the survival data, which is where the clinical weight of this study sits. Medrek and colleagues used Kaplan-Meier analysis and Cox proportional hazards modeling—Cox modeling being the standard statistical method for figuring out which variables independently predict survival after accounting for everything else. In univariable analysis, dense CD68 infiltration in tumor stroma carried a hazard ratio of fifteen point two for breast cancer-specific survival and nine point two four for recurrence-free survival. To put that in plain terms: patients with high stromal CD68 were roughly fifteen times more likely to die of breast cancer and nine times more likely to have their disease recur compared to patients with low stromal CD68. Those are not subtle effects. Dense CD163 infiltration in stroma also predicted worse breast cancer-specific survival in univariable analysis, with a hazard ratio of three point two. The critical test, though, is the multivariate model—because you need to know whether stromal macrophages are predicting survival on their own, or whether they are just a proxy for other known risk factors like lymph node involvement, tumor size, grade, or hormone receptor status. When Medrek and colleagues adjusted for age, lymph node status, tumor size, Ki67, grade, HER status, and estrogen receptor status, stromal CD68 still independently predicted breast cancer-specific survival—with a p-value of zero point zero two. CD163 in stroma did not retain independent significance in multivariate analysis, which suggests it is correlated with those other clinical variables rather than adding information beyond them. So CD68 in the stroma is the independent prognostic marker. CD163 in the stroma is the richer biological signal—more tightly linked to the aggressive subtypes—but CD68 is what holds up when you account for the rest of the clinical picture. There is also a recruitment story beginning to emerge from this data. Elkabets and colleagues had previously reported that tumors attract granulin-expressing hematopoietic cells—granulin being a protein encoded by the GRN gene—that localize near stromal fibroblasts and activate the stroma. In the breast cancer cohort, stromal CD163 macrophage density correlated with GRN expression, with a p-value of zero point zero one. That correlation suggests GRN-associated recruitment may be one mechanism drawing the prognostically dangerous macrophage population into the stroma. Medrek and colleagues also note they cannot rule out that some CD163-positive stromal cells are myeloid-derived suppressor cells, or MDSCs—an immature myeloid population known for potent immunosuppression that can further shield tumors from immune clearance. One additional observation is worth holding onto. Some CD163-positive areas in the stroma lacked CD68 expression entirely. That finding suggests CD163 is not simply a subset of what CD68 captures — it may mark a distinct myeloid population that CD68 misses. This could explain why the two markers tell slightly different prognostic stories, and it opens the question of whether future studies should use both in combination. The takeaway from Medrek and colleagues is precise: it is not the presence of immune cells inside a breast tumor that matters for prognosis—it is their location and their identity. Macrophages in the tumor stroma, particularly those marked by CD163 and CD68, are the dangerous population. Macrophages inside the tumor nest, by contrast, are clinically inert by these measures. If you want to understand the immune microenvironment of a breast tumor, you have to look at where the cells are sitting. If stromal macrophage recruitment is driven in part by signals like granulin, then those recruitment pathways become potential targets—not just for understanding prognosis, but for disrupting the process that turns an immune defender into a tumor's unwitting accomplice. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.