Polarization of tumor-infiltrating macrophages predicts complete response to neoadjuvant treatment in patients with rectal cancer

Characterizing the composition of immune cells in the tumor microenvironment (TME) has shown promise in understanding variable responses to neoadjuvant treatment in rectal cancer. Despite this, no method has been established that can confidently predict the response to pre-operative chemoradiotherapy. Tumor-associated macrophages (TAMs) in the TME can be polarized by the biochemical milieu to adopt pro-inflammatory (M1-like) and anti-inflammatory (M2-like) phenotypes. Utilizing the spectrum of TAM polarization as a biomarker to predict patient response to neoadjuvant therapy in rectal cancer patients has not been fully explored. We address this using 7-plex immunofluorescence staining to quantify 16 subsets of TAMs in rectal adenocarcinoma pretreatment biopsies. Pretreatment biopsies were obtained from 128 patients with known tumor regression grades (TRG). Through quantifying the prevalence of TAM subsets, we identified a higher density of the completely M1-polarized subset and a lower density of the completely M2-polarized subset in patients with a complete pathological response, the absence of detectable tumor cells following treatment. Predictive modeling using TAM subsets showed that only the densities of two completely polarized subsets were predictors of therapeutic response in patients with rectal cancer. Significantly, we demonstrate that the full panel of markers we employed were required to define predictive populations, as total TAMs or TAM subsets defined by fewer markers were incapable of predicting treatment response. Taken together, we validate the utility of a multiplex approach to define the spectrum of TAM polarization, which can be leveraged to identify patients that will have a complete pathological response to neoadjuvant treatment.