Colorectal cancer (CRC) represents a pressing global health burden. Contemporary CRC management adopts a multimodal approach integrating surgical resection, chemotherapy, radiotherapy, immunotherapy, and complementary modalities such as traditional Chinese medicine (TCM) (1). Surgical intervention remains the cornerstone for localized disease, while systemic therapy is predominantly based on chemotherapy regimens incorporating fluorouracil, oxaliplatin and irinotecan. Radiotherapy plays a crucial role in the management of rectal cancer, particularly in the neoadjuvant setting where it reduces local recurrence (3). Recent advances in immunotherapy, especially for mismatch repair-deficient tumors, have shown encouraging clinical outcomes (4). Additionally, TCM and novel nanomedicine approaches are being increasingly explored as complementary approaches to enhance treatment efficacy and reduce side effects (5,6). Despite these diverse treatment options, CRC remains one of the most common malignancies globally and ranks as the third leading cause of cancer-related mortality (7). In the United States alone, ~153,020 new CRC cases were projected in 2023, with 52,550 deaths, including 19,550 cases and 3,750 deaths among individuals under 50 years of age (8). Although regular screening, surveillance and high-quality treatment can prevent a substantial proportion of CRC morbidity and mortality (9), recurrence and metastasis remain frequent. Among patients with metastatic CRC, ~70-75% survive beyond 1 year, 30-35% remain alive at 3 years, while the 5-year survival rate falls below 20% (10). Therefore, there is an urgent need to identify reliable biomarkers that can enable early and accurate prognostic assessment.

In a variety of solid tumors, both the density and polarization status of TAMs are strongly associated with patient prognostic outcomes. In CRC, high infiltration of M1-polarized TAMs is associated with a favorable prognosis across various disease stages (11), whereas elevated levels of M2 TAMs often portend a poorer prognosis, a phenomenon observed in multiple cancer types, including thyroid, lung, stomach and breast cancers (12). While macrophage enrichment is usually linked to adverse tumor outcomes, this correlation appears to be reversed in CRC (13). Nevertheless, the role of TAMs within the CRC tumor microenvironment (TME) remains complex, and some studies have shown that TAMs are associated with poor prognosis in patients with CRC (14). Several investigations have found that CD68⁺ TAMs are predominantly distributed in the CRC tumor stroma, especially at the invasive front, and that CD68⁺ TAMs infiltration at these sites correlates with improved prognosis in patients with CRC (15). However, TAMs of distinct subtypes and spatial distributions exert different prognostic significance in CRC: For instance, infiltration of CD68⁺ TAMs and M2-type TAMs is linked to unfavorable outcomes (16). Collectively, TAMs play diverse roles in the prognosis of CRC, and their number, subtype, and spatial distribution within the TME, as well as their impact on patient outcomes, are complex. These findings emphasize the importance of in-depth studies of TAM properties and the search for specific TAM markers to inform therapeutic strategies and improve prognostic assessment.

In recent years, cancer stem cells (CSCs) have been extensively studied as key drivers underlying core hallmarks of tumor progression, including distant metastasis, recurrence and drug resistance (17). CSCs exhibit long-term self-renewal capacity and metastatic potential across a variety of malignancies, including CRC (18). In CRC, CD133⁺/CD44⁺ CSCs populations have been shown to correlate negatively with both disease-free survival (DFS) and overall survival (OS) (19). Combined detection of CD133/CD44 expression enhances the identification of colorectal CSCs (20). The TME engages in complex crosstalk with CSCs, which are often characterized by immune suppression and low immunogenicity in CRCs (21). Furthermore, Luo et al (22) reported that CSCs can recruit TAMs into the TME and accelerate their polarization into tumor-promoting phenotypes, whereas TAMs in turn maintain CSC stemness and construct niches unfavorable to the survival of patients with CSC. Although the prognostic values of TAMs or CSC markers have been investigated individually, studies analyzing their combined effects and synergistic interactions on patient prognosis are scarce. Most previous studies have examined TAMs or CSC markers in isolation, failing to capture the prognostic power embedded within their interplay. Therefore, it was hypothesized that a combined biomarker panel reflecting this TAM-CSC synergy may provide superior prognostic stratification compared with individual markers alone.

In the present study, the expression of key TAM markers (CD86 for M1-like and CD163 for M2-like phenotypes) and CSC markers (CD44 and CD133) was simultaneously investigated in a cohort of patients with CRC. The primary objective of the present study was to evaluate the prognostic impact of their combined expression and to determine whether this synergistic biomarker panel could serve as an independent predictor of survival, potentially offering a more refined tool for risk assessment in CRC.

CRC ranks as the third most commonly diagnosed gastrointestinal malignancy worldwide (25). Current treatments for CRC include surgical resection, chemotherapy, radiotherapy and immunomodulatory approaches (26). Despite these interventions, nearly 40% of patients develop disease recurrence or distant metastasis, solidifying CRC s status as the third leading cause of cancer-related mortality (27). TNM staging, while a major prognostic factor, has limitations in accurately predicting CRC patient prognosis (26). Therefore, it was aimed to identify biomarkers with a stronger prognostic correlation with CRC. Given that the combined detection of multiple markers can improve the specificity of CSC and TAM identification, the joint expression of CSC and TAM biomarkers was evaluated to determine a panel with improved prognostic correlation, thereby providing a potential basis for more precise clinical diagnosis and therapeutic decision-making.

CSCs regulate the TME by recruiting immune cells via paracrine signaling. For instance, glioma stem cells promote tumor progression by secreting periostin, recruiting TAMs, and constructing the TME (28). In addition, CSCs secrete cytokines, such as TGF-β, IL-10, IL-4 and IL-13 into the TME, which exert an inhibitory effect on multiple immune cells (29) and thereby promote tumor progression. Reciprocally, TAMs regulate CSC stemness and maintain self-renewal capacity through secretion of cytokines, chemokines, and participation in complex signaling networks. Key pathways mediating this bidirectional crosstalk include the IL-6/STAT3 axis, wherein TAM-derived IL-6 activates STAT3 signaling in CSCs to promote self-renewal and stemness maintenance (30). Additionally, TGF-β signaling plays a pivotal role in mediating the bidirectional communication between TAMs and CSCs. TAM-secreted TGF-β not only enhances CSC stemness but also induces epithelial-mesenchymal transition, further contributing to tumor progression and metastasis (31). Other important pathways, such as NF-κB, can be activated by TAM-derived IL-1 and TNF-α, creating an inflammatory niche that further supports CSC persistence (32).

The biological significance of these cytokine-mediated interactions lies in their ability to establish a sustainable niche that maintains CSC populations, drives tumor heterogeneity, and confers therapeutic resistance. For instance, in hepatocellular carcinoma, TAMs can induce STAT3 activation via IL-6, stimulating further cytokine release and forming a positive feedback loop that amplifies CSC self-renewal (33). Similarly, in breast cancer, TAMs have been reported to enhance the stemness characteristics of CSCs through Ephrin-EphA4 interactions, which in turn stimulate CSCs to produce inflammatory cytokines, including IL-1, IL-6 and IL-8(34).

In summary, a bidirectional communication exists between CSCs and TAMs. To investigate this relationship, the protein expression of TAM and CSC biomarkers in CRC tumor tissues was detected by IHC. It was found that high expression of CD163, CD44 and CD133 was associated with poor prognosis, while high CD86 expression was correlated with favorable prognosis. Indeed, the prognostic role of CD86⁺ TAMs in CRC is complex and context-dependent. Certain studies suggested that CD86 is associated with improved survival, and patients with high CD86 gene expression have higher OS rates than those with low CD86 expression. This indicates that low expression of this marker is linked to tumor progression and aggressiveness (35), which is in line with the present findings. While certain studies report CD86 as a general M1 marker associated with antitumor effects, the present data suggest its prognostic significance may be context-dependent (36,37). Discrepancies with previous studies could stem from heterogeneity in TMEs, differences in antibody clones, or variations in cohort demographics. By contrast, CD163⁺ TAMs are known to drive tumor progression by suppressing antitumor immunity, enhancing angiogenesis, and promoting metastasis. The association between CD163⁺ M2-like TAMs and poor prognosis has been more consistently reported across various cancer types, including CRC (38,39). Therefore, Certain studies suggested that combined analysis of CD86⁺ TAMs and CD163⁺ TAMs appears to be more suitable for determining relapse and mortality rates (40,41). In addition, the correlations between TAM and CSC markers, clinicopathologic parameters and patient prognosis were investigated. CD86 expression was markedly correlated with tumor size and distant metastasis, whereas CD163 expression was significantly correlated with tumor size and T-stage. Additionally, CD44 and CD133 expression were strongly linked to TNM staging and lymph node metastasis, respectively. Moreover, both CD163 and CD133 were identified as independent prognostic factors in CRC. The present analysis revealed significant prognostic value for both CD86 (M1-like) and CD133, as well as a negative correlation between them. While the combination of CD86 and CD133 is of interest, the present study was strategically designed to investigate the synergistic pro-tumorigenic interaction between TAMs and CSCs. In this context, the concurrent high expression of CD163 (M2-like TAMs) and CD133 represents a functionally coherent and biologically synergistic unit that collectively fosters an immunosuppressive and pro-stemness TME, leading to the most aggressive disease phenotype. This is substantiated by the present multivariate Cox regression analysis, which identified both CD163 and CD133 as the most robust and independent prognostic factors (P<0.001 for both DFS and OS). Therefore, prioritizing the CD163⁺/CD133⁺ profile was a deliberate strategy to most directly test our core hypothesis and to establish a clinically actionable biomarker for identifying the highest-risk patient subgroup.

The present analysis revealed significant prognostic value for both CD86 and CD133, as well as a negative correlation between them. While the combination of CD86 and CD133 is of interest, the present study was strategically designed to investigate the synergistic pro-tumorigenic interaction between TAMs and CSCs. In this context, the concurrent high expression of CD163 and CD133 represents a functionally coherent and biologically synergistic unit that collectively fosters an immunosuppressive and pro-stemness TME, leading to the most aggressive disease phenotype. This is substantiated by our multivariate Cox regression analysis, which identified both CD163 and CD133 as the most robust and independent prognostic factors (P<0.001 for both DFS and OS). Therefore, prioritizing the CD163⁺/CD133⁺ profile was a deliberate strategy to most directly test our core hypothesis and to establish a clinically actionable biomarker for identifying the highest-risk patient subgroup.

Current studies of TAMs and CSCs in CRC remain relatively limited, typically focusing either on the prognostic significance of different TAM subtypes or on the relationship between CSC expression and CRC occurrence and progression (42,43). By contrast, the present study combines the assessment of TAM and CSC biomarkers, specifically CD163 and CD133, providing a more robust prognostic tool than evaluating either marker alone in CRC. This integrative approach distinguishes the current study from prior works that primarily focused on single-marker systems. By assessing TAMs in conjunction with CSCs, the roles of different subtypes of TAMs and CSCs in CRC were more comprehensively and accurately explored. Both univariate and multivariate analyses identified CD163⁺ TAMs and CD133⁺ CSCs as independent prognostic factors for DFS and OS. The prognostic significance of their combined expression was therefore further examined, finding that concurrent high levels of CD163 and CD133 were strongly associated with shorter DFS and OS, indicating poor prognosis. These results suggested that the TAM-CSC functional unit, rather than either component in isolation, represents a key determinant of tumor aggressiveness in CRC.

In conclusion, the novel contribution of our research lies in establishing and validating a combined CD163/CD133 biomarker profile that captures the synergistic interaction between pro-tumorigenic TAMs and stem-like cancer cells. While TNM staging provides an anatomical framework, the CD163⁺/CD133⁺ profile directly reflects the pro-tumorigenic and treatment-resistant potential of the TME. This dual-marker panel identifies a high-risk patient subgroup with significantly shorter DFS and OS, thereby significantly improving the accuracy of prognostic predictions in patients with CRC. These findings carry important clinical implications for developing more effective prognostic assessment tools and personalized treatment strategies. Future research should focus on validating these biomarkers in larger, multicenter cohorts and exploring targeted therapies that disrupt the TAM-CSC interaction axis to improve patient outcomes. Furthermore, this approach holds promise for predicting responses to both conventional chemotherapy and emerging immunotherapies, potentially guiding more effective, personalized treatment strategies.