Rectal cancer treatment relies heavily on surgery, but neoadjuvant therapy improves outcomes by reducing local recurrence and enabling radical resection. A significant portion of patients (up to 20%) achieve a clinical complete response (cCR) following neoadjuvant treatment, defined as the absence of clinically detectable tumors. However, neoadjuvant therapy carries risks of postoperative complications and long-term side effects, making the risk-benefit assessment crucial for individual patients. Currently, few clinically available predictive markers exist to identify patients who will respond well to neoadjuvant treatment. The density of tumor-infiltrating lymphocytes (TILs) in colorectal cancer (CRC) has shown promise as a prognostic indicator for survival, potentially surpassing the TNM classification. Immunoscore (IS), based on TILs, is a promising predictor of tumor regression and prognosis. While some studies suggest pre-therapeutic TIL density predicts response to neoadjuvant treatment, particularly CD8+ TILs, conflicting results exist. Granzyme B (GrzB) from cytotoxic T cells (CTLs) induces apoptosis in tumor cells, and increased intraepithelial GrzB+ CTLs have been linked to higher tumor regression. Type I Interferon (IFN) enhances CTL priming by promoting cross-presentation of antigens; radiotherapy boosts this process via the cGAS-STING pathway. Type I IFN induces IFN-stimulated genes (ISGs), with MxA serving as a reliable surrogate marker for type I IFN activity, independent of other ISG inducers. This study investigates whether the density and distribution of T cell subsets, including γδT cells, and MxA expression as a type I IFN footprint, in pre-operative rectal cancer biopsies can predict response to neoadjuvant treatment.

Existing literature showcases the prognostic value of Immunoscore (IS) in colorectal cancer, potentially exceeding the predictive power of the TNM staging system. A biopsy-adapted version, IS B, has demonstrated a correlation between TIL density and neoadjuvant treatment response. However, the field lacks consensus regarding the predictive value of TILs, with studies reporting both positive and negative correlations between pre-treatment TIL density and tumor regression. Discrepancies may stem from variations in TRG evaluation, cutoff values for TIL density, cohort sizes, and intra-cohort dependencies in quantification models. While some studies show a strong correlation between T cell infiltration and response to neoadjuvant treatment, others do not. This highlights the need for a more refined approach to predict treatment response using a combination of biomarkers and advanced imaging techniques.

This study analyzed 130 patients treated for rectal cancer at Sahlgrenska University Hospital between 2007 and 2019 who received neoadjuvant therapy with a surgical resection interval exceeding 14 days. Pre-operative biopsies were subjected to multiplex immunofluorescence (mIF) staining to quantify various immune cell populations, including CD3+ T cells, CD8+ T cells, CD8+ GrzB+ cytotoxic T cells, and γδTCR+ cells. MxA expression, indicating type I IFN response, was also assessed in the tumor stroma. Tumor regression grade (TRG) from surgical specimens, ranging from 0 (pathological complete response, pCR) to 3 (no response), determined treatment response. Image analysis using TissueFAXS and StrataQuest software quantified cell densities. Statistical analysis included Wilcoxon's rank-sum test, Pearson correlation coefficient analyses, and logistic regression. The study compared the predictive accuracy of the biopsy-adapted Immunoscore (IS B), based on CD3+ and CD8+ cell densities, with a novel stratification approach combining CD8+ T cell density in the tumor tissue and MxA+ cell density in the tumor stroma. A heatmap visualization method was employed to display patients based on the ranked percentiles of CD8+ and MxA+ cells, creating six categories reflecting combinations of high, intermediate, and low levels of both markers.

No significant differences were observed in the numbers of conventional T cells, γδT cells, or their subsets between complete responders (TRG 0) and non-responders (TRG 3). Including GrzB expression slightly improved the trend towards increased TILs in pCR patients. However, quantification of MxA+ cells in the tumor stroma revealed a significantly higher density in pCR patients compared to non-responders. The biopsy-adapted Immunoscore (IS B) showed a similar pattern as MxA density. A heatmap approach, stratifying patients based on CD8+ and MxA+ cell densities, significantly improved the prediction of tumor regression. This heatmap approach showed a significantly higher predictive ability compared to using the IS B score or CD8+ or MxA+ cells alone. Importantly, almost all patients (92.3%) achieving pCR were in the high quartile of either CD8 or MxA, and none were low in both. In contrast, only 2.3% of non-responders were high in both parameters. The heat map stratification, combining CD8+ and MxA+, achieved the highest diagnostic odds ratio, indicating a superior ability to classify patient response. The inclusion of other clinical parameters did not diminish the importance of the developed method.

This study confirms the limitations of relying solely on conventional TILs, especially CD8+ T cells, to predict response to neoadjuvant treatment in rectal cancer. The finding of a strong association between a type I IFN footprint (MxA expression) and pCR suggests that the presence of an active type I IFN response contributes to the effectiveness of neoadjuvant therapy. This observation is consistent with the known role of type I IFN in enhancing CTL function and fostering an anti-tumor immune response. The heatmap approach, integrating both CD8+ T cell density and MxA expression, offers a significant improvement in predictive accuracy over existing methods like IS B. This suggests that a combination of sufficient pre-existing CTL infiltration and an active type I IFN response are crucial for achieving pCR following neoadjuvant treatment, possibly indicating a more permissive tumor microenvironment for effective immune cell-mediated tumor eradication.

This study demonstrates the independent prognostic value of a type I IFN footprint (MxA expression) in predicting response to neoadjuvant treatment in rectal cancer. The combination of CD8+ T cell density and MxA+ cell density, using the heatmap-based approach, provides a more accurate predictive model than existing methods. This novel approach may facilitate personalized treatment strategies, potentially enabling identification of patients who are most likely to benefit from neoadjuvant treatment, as well as patients who would be candidates for immune-reinvigoration therapies.

The main limitation is the lack of validation in an independent cohort. The study's findings require external validation to confirm their generalizability. Additionally, the mIF approach may introduce slightly increased variability compared to colorimetric methods. The study did not explore the specific subtypes of type I interferons or the cellular sources of type I interferon within the tumor microenvironment. Further research to investigate these factors will enhance the understanding of type I interferon's role in the observed correlation.