Stage‐Specific Differences in Circulating Resistin From Premalignant to Advanced Colorectal Cancer: A Cross‐Sectional Cohort Study

ABSTRACT

Background

Resistin is a macrophage‐associated cytokine linked to obesity‐driven inflammation, a recognised risk factor for colorectal cancer. Prior studies of circulating resistin in colorectal cancer have reported inconsistent findings, largely because patients were analysed as a single group without accounting for tumour stage. Whether resistin follows a stage‐specific pattern across the adenoma–carcinoma sequence, independent of insulin resistance and conventional tumour markers, has not been systematically examined.

Methods

In this single‐centre study, serum resistin levels were measured in 179 participants, including 137 patients with colorectal cancer (CRC), 11 with colorectal adenomas and 31 healthy controls. Tumour markers and metabolic variables were analysed. Moreover, stage‐stratified comparisons were performed using non‐parametric testing and ordinal logistic regression.

Results

Circulating resistin concentrations varied significantly across diagnostic categories and tumour stage. Elevated levels were observed in adenomas and early‐stage CRC (stage I–II; median ~7600 pg/mL), whereas significantly lower levels were detected in advanced‐stage disease (stage III–IV; median ~2500 pg/mL; p < 0.001). In multivariable ordinal regression analysis, resistin remained independently associated with tumour stage after adjustment for BMI, fasting insulin, CEA and CA19‐9. Resistin levels were not associated with insulin resistance as measured by HOMA‐IR.

Conclusions

Circulating resistin shows a stage‐dependent association across colorectal tumorigenesis, with higher levels in premalignant and early‐stage disease and lower levels in advanced‐stage disease. These cross‐sectional findings are consistent with a model in which resistin reflects host inflammatory activation rather than tumour burden, although this interpretation remains hypothesis‐generating. This pattern may help explain previously inconsistent associations reported for resistin in colorectal cancer; however, further prospective and mechanistic studies are required to clarify its role in tumour progression and validate our findings.

1. Introduction

Colorectal cancer (CRC) stands as one of the most common and deadly cancers around the globe [1, 2]. While the classic adenoma–carcinoma sequence, a stepwise process driven by genetic changes, remains central to our understanding of CRC development [3, 4, 5], there is now growing recognition that inflammation and immune system disturbances within the colon play a pivotal role in this disease [6]. In other words, CRC is shaped not only by genetic instability but also by ongoing, complex interactions between abnormal epithelial cells and their inflammatory surroundings.

Adipokines, which are signalling proteins secreted by fat tissue and immune cells, act as messengers between metabolism and inflammation [7, 8]. Resistin, for example, was first identified in mice as a hormone linking obesity to insulin resistance [9]. In humans, though, resistin is mainly produced by monocytes and macrophages, acting primarily as a pro‐inflammatory molecule rather than as a regulator of metabolism [10, 11].

Resistin has been linked to greater secretion of tumour necrosis factor‐α (TNF‐α) and interleukin‐6 (IL‐6), which are both central players in cancer‐related inflammation [11, 12]. This connection has sparked interest in whether resistin might serve as a bridge between immune activation and the development of colorectal tumours.

Multiple studies have reported that people with colorectal cancer often have higher circulating resistin levels than healthy individuals. Meta‐analyses and cohort studies lend support to this association [13, 14, 15]. Still, the relationship between resistin and tumour stage or progression has been inconsistent in the literature. A key limitation is that many studies do not break down results by stage, nor do they track changes in resistin from early adenomas through to advanced cancers, making it difficult to draw strong conclusions about what circulating resistin levels truly mean [15].

In light of these gaps, our study set out to measure serum resistin across different stages of colorectal neoplasia: in healthy people, in patients with colorectal adenomas and in those with CRC. We also examined how resistin levels relate to tumour stage, standard tumour markers (CEA and CA19‐9), body mass index (BMI) and measures of insulin resistance. Our central hypothesis was that resistin follows a non‐linear, stage‐dependent pattern across the progression of colorectal cancer, reflecting early immune and inflammatory activity rather than just the presence of advanced tumours.

2. Materials and Methods

2.1. Study Design and Setting

This was a single‐centre observational study with cross‐sectional biomarker assessment conducted at Agios Savvas Anticancer Hospital in Athens, Greece, between 2019 and 2020. Circulating biomarkers and clinical variables were obtained at a single pre‐treatment timepoint and analysed across diagnostic categories and colorectal cancer stage.

2.2. Study Population

Adult participants (≥ 18 years) were recruited into three groups: patients with histologically confirmed colorectal adenocarcinoma (CRC), individuals with colorectal adenomas requiring surgical excision due to endoscopic non‐resectability without histological evidence of invasive malignancy, and healthy controls undergoing screening evaluation with no evidence of colorectal neoplasia on colonoscopic assessment. All CRC patients were included consecutively during the study period. Healthy controls were recruited during the same time frame from individuals undergoing routine screening evaluation at the same institution and were broadly matched to the CRC/adenoma groups by age and sex; comorbidities, regular medication use and smoking status were recorded but were not used as formal matching criteria, and residual differences in these factors between groups cannot be excluded.

Demographic and anthropometric characteristics, including age, sex and body mass index (BMI), were recorded for all participants. Baseline characteristics were compared across diagnostic categories to assess comparability between groups.

Exclusion criteria were applied to minimise systemic inflammatory confounding and included acute infection at the time of sampling, autoimmune disease, chronic inflammatory conditions, prior or concurrent non‐colorectal malignancy and incomplete clinical or laboratory data. Participants with clinically evident hereditary colorectal cancer syndromes were excluded; however, formal germline testing was not routinely available for all participants.

The final analytic cohort comprised 179 participants: 137 patients with CRC, 11 with colorectal adenomas and 31 healthy controls. The study was approved by the Institutional Ethics Committee of Agios Savvas Anticancer Hospital, and all participants provided written informed consent.

2.3. Blood Sample Collection and Timing

Fasting venous blood samples (15 mL total) were collected from all participants following an overnight fast. For CRC patients, blood sampling was performed before any oncologic intervention, including surgery, chemotherapy or radiotherapy. Of the total volume, 10 mL was collected in serum separator tubes containing a clot activator, allowed to clot at room temperature for 20–30 min, centrifuged at 3000 rpm for 15 min at 8°C, aliquoted, and stored at −80°C until batch analysis. An additional 5 mL was collected in EDTA tubes, centrifuged at 4400 rpm for 15 min at 8°C for plasma and buffy coat separation, and stored at −80°C.

2.4. Laboratory Measurements

Serum resistin concentrations were measured in duplicate using a commercially available enzyme‐linked immunosorbent assay (Human Resistin ELISA Kit, Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions, with absorbance measured using a Multiscan FC Microplate Photometer (Thermo Fisher Scientific, Waltham, MA, USA). Concentrations were calculated from standard curves and the mean of the two duplicate readings was used for statistical analysis. Intra‐assay coefficients of variation were not formally calculated across all plates; however, wells exhibiting greater than 15% deviation between duplicates were flagged for review. Raw duplicate measurement data are not included in this report but are available from the corresponding author upon reasonable request. Inter‐assay coefficients of variation were not formally assessed across plates, and no external validation cohort was available for this assay; this is acknowledged as a limitation below.

Carcinoembryonic antigen (CEA), carbohydrate antigen 19‐9 (CA19‐9) and fasting insulin were measured using standard clinical immunoassays in the hospital laboratory. Insulin resistance was estimated using the Homeostatic Model Assessment of Insulin Resistance (HOMA‐IR), calculated as follows:

$$\text{HOMA}-\mathrm{IR}=\left(\text{fasting insulin}\left[\mathrm{\mu U}/\mathrm{mL}\right]\times \text{fasting glucose}\left[\mathrm{mg}/\mathrm{dL}\right]\right)/405.$$

2.5. Statistical Analysis

Statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Normality of continuous variables was assessed using the Shapiro–Wilk test. As most variables were non‐normally distributed, results are presented as medians with interquartile ranges (IQR), and non‐parametric tests were applied.

Baseline demographic characteristics were compared across diagnostic groups to evaluate group comparability. Between‐group comparisons of continuous variables were conducted using the Kruskal–Wallis test, with Bonferroni‐adjusted post hoc analyses where appropriate. Associations between continuous variables were evaluated using Spearman's rank correlation coefficient. Categorical variables were compared using the χ ² test or Fisher's exact test, as appropriate.

To explore independent associations with tumour stage, ordinal logistic regression analysis was performed with tumour stage entered as an ordered dependent variable. Variables entered into the model included serum resistin, BMI, fasting insulin, CEA and CA19‐9. Results are reported as odds ratios (ORs) with 95% confidence intervals (CI). A two‐sided p‐value < 0.05 was considered statistically significant.

In addition to between‐group comparisons, an ordered non‐parametric trend test (Jonckheere–Terpstra) was performed to evaluate monotonic changes in serum resistin across tumour stages (I–IV). Effect sizes were reported alongside p‐values where appropriate. For ordinal regression models, the proportional odds assumption was assessed using a likelihood‐ratio comparison against a model permitting predictor effects to vary across stage cut‐points, performed independently as a supplementary check. Multicollinearity among model covariates was assessed using variance inflation factors. To facilitate clinical interpretation, resistin was modelled both per 1000 pg/mL increase and per interquartile range (IQR) increase. Analyses involving CEA and CA19‐9 were performed using complete‐case analysis due to partial availability of tumour marker data. As a sensitivity analysis to address potential confounding by nutritional status, the multivariable ordinal model was additionally repeated with serum albumin included as a covariate in the subset of patients with available albumin measurements.

3. Results

3.1. Patient Characteristics

The study included 179 participants, 137 patients with colorectal cancer (CRC), and 42 non‐cancer participants. The non‐cancer reference group comprised 31 healthy individuals and 11 patients with surgically managed colorectal polyps. The median age of the cohort was 70 years (interquartile range [IQR] 13), and 56.4% were male. Median body mass index (BMI) was 27.7 kg/m² (IQR 5.6) (Table 1).

TABLE 1.

Baseline demographic and clinical characteristics of the study population.

Characteristics	Value
Total participants	179
Colorectal cancer cases	137
Non cancer participants	42
Age, years	Median 70 (IQR 13)
Sex, n (%)
Male	101 (56.4%)
Female	78 (43.6%)
BMI, kg/m2	Median 27.7 (IQR 5.6)
Tumour location (cancer patients), n (%)
Rectum	63 (46.0%)
Left colon (incl. sigmoid)	41 (29.9%)
Right colon	33 (24.1%)
Tumour stage, n (%)
Stage I	33 (24.1%)
Stage II	33 (24.1%)
Stage III	33 (24.1%)
Stage IV	38 (27.7%)

Among patients with CRC, tumours were located in the rectum in 46.0%, the left colon in 29.9% and the right colon in 24.1%. Tumour stage was evenly distributed across stages I–III (each 24.1%), while 27.7% of patients presented with stage IV disease.

BMI differed significantly across tumour stages (p = 0.006), with lower values observed in stage IV disease. Age and sex distribution did not differ significantly between stages.

3.2. Serum Resistin and Tumour Stage

Serum resistin concentrations differed significantly across tumour stages (Kruskal–Wallis p < 0.001) (Figure 1). Median resistin levels were 7681 pg/mL (IQR 4246) in stage I and 7692 pg/mL (IQR 5353) in stage II, compared with 2556 pg/mL (IQR 1120) in stage III and 2426 pg/mL (IQR 2372) in stage IV. Post hoc analyses confirmed significantly lower resistin concentrations in stages III and IV compared with stages I and II (all p < 0.001), while no significant differences were observed within early‐stage (I vs II) or advanced‐stage (III vs IV) comparisons. A Jonckheere–Terpstra test demonstrated a significant decreasing trend in resistin levels with advancing stage (p < 0.001), supporting a stage‐dependent trajectory.

FIGURE 1.

Serum resistin concentrations across tumour stages (I–IV). Distribution of serum resistin levels (pg/mL) measured by ELISA in patients with colorectal cancer (stage I: n = 33, stage II: n = 33, stage III: n = 33, stage IV: n = 38). Open circles (○) denote mild outliers (values between 1.5 and 3 interquartile ranges [IQR] from the box edge); asterisks (*) denote extreme outliers (values > 3 IQR from the box edge).

3.3. Insulin and Metabolic Parameters

Fasting serum insulin levels did not differ significantly across tumour stages (p = 0.554), with substantial overlap between stages. Similarly, HOMA‐IR values displayed wide inter‐individual variability but did not differ significantly across diagnostic categories (CRC, polyps, healthy controls) or across tumour stages (all p > 0.05).

Correlation analyses revealed no meaningful association between serum resistin and HOMA‐IR, either in the overall cohort or within diagnostic subgroups (Spearman ρ ≈ 0), indicating that circulating resistin levels were independent of systemic insulin resistance (Figure 2).

FIGURE 2.

Scatter plot of serum resistin versus HOMA‐IR. Scatter plot illustrating the relationship between serum resistin and HOMA‐IR across all study participants. No meaningful correlation was observed overall or within diagnostic subgroups (Spearman ρ ≈ 0; n = 179).

3.4. Classical Tumour Markers

In contrast to resistin, classical tumour markers increased progressively with advancing disease. Serum carcinoembryonic antigen (CEA) levels rose significantly across tumour stages (p < 0.001), increasing from a median of 2.10 ng/mL in stage I to 6.40 ng/mL in stage IV. CA19‐9 demonstrated a similar stage‐related increase (p = 0.018), with the highest values observed in stage IV disease. CEA and CA19‐9 measurements were available for 97 of 137 CRC patients.

CEA and CA19‐9 were strongly correlated (Spearman ρ = 0.430, p < 0.001). Neither marker showed a significant association with BMI or serum resistin levels. Biomarker concentrations by tumour stage are summarised in Table 2.

TABLE 2.

CEA and CA19‐9 concentrations by tumour stage among CRC patients with available tumour marker measurements (n = 97).

		N	Median	IQR	Kruskal–Wallis	p
CEA	Stage I	24	2.10	1.08	24.499	< 0.001
	Stage II	21	3.00	1.90
	Stage III	20	5.45	8.72
	Stage IV	32	6.40	16.90
	Total	97
CA 19‐9	Stage I	24	11.15	10.20	10.024	0.018
	Stage II	21	16.70	9.70
	Stage III	20	18.70	21.77
	Stage IV	32	27.70	78.63
	Total	97

Note: Values are reported as median (IQR). p‐values are from Kruskal–Wallis tests. Bold values indicate statistically significant results (p < 0.05).

3.5. Serum Resistin Across Diagnostic Categories

Analysis by diagnostic category (rectal cancer, colon cancer, colonic polyps and healthy controls) revealed significant differences in serum resistin concentrations (Kruskal–Wallis p = 0.017) (Table 3, Figure 3). Median resistin levels were highest in individuals with colorectal polyps (6718 pg/mL, IQR 2093), followed by patients with rectal cancer (4637 pg/mL, IQR 4725) and colon cancer (3466 pg/mL, IQR 5874). The lowest concentrations were observed in healthy controls (2981 pg/mL, IQR 3326).

TABLE 3.

Serum resistin (pg/mL) concentrations across diagnostic categories (rectal cancer, colon cancer, colonic polyps, healthy controls).

		N	Median	IQR	Kruskal–Wallis	p
Resistin (pg/mL)	Rectum cancer	63	4637	4725	10.252	0.017
	Colon cancer	74	3466	5874
	Colonic polyps	11	6718	2093
	Healthy	31	2981	3326

Note: Kruskal–Wallis p‐value reflects an overall omnibus test across the four groups; pairwise group differences are reported in the Results text. Bold values indicate statistically significant results (p < 0.05).

FIGURE 3.

Serum resistin (pg/mL) across diagnostic categories. Resistin concentrations in rectal cancer (n = 63), colon cancer (n = 74), colonic polyps (n = 11) and healthy controls (n = 31). Kruskal–Wallis test, p = 0.017; Bonferroni‐adjusted post hoc comparison showed significantly higher levels in the polyp subgroup versus healthy controls (adjusted p = 0.017), while differences between polyps and colon/rectal cancer were not significant after correction.

Post hoc pairwise comparisons with Bonferroni adjustment confirmed that resistin levels were significantly higher in the polyp subgroup compared with healthy controls (adjusted p = 0.017). In contrast, no statistically significant differences were observed between the polyp subgroup and either colon or rectal cancer subgroups after correction for multiple testing.

These findings indicate that circulating resistin is elevated at the premalignant stage and does not increase further in established invasive disease, supporting a stage‐dependent inflammatory trajectory across colorectal neoplasia.

3.6. Serum Resistin Stratified by BMI Category

BMI‐stratified analyses revealed differential patterns between CRC patients and healthy controls (Table 4). Among individuals with BMI < 25 kg/m², serum resistin levels did not differ significantly between CRC patients and healthy controls. Median resistin concentrations were 4105 pg/mL in CRC patients (n = 41) and 3803 pg/mL in healthy individuals (n = 9).

TABLE 4.

Serum resistin concentrations stratified by BMI category (BMI < 25 kg/m² vs. BMI ≥ 25 kg/m²) in healthy individuals and patients with colorectal neoplasia.

	Median (pg/mL)	p
ΒΜΙ < 25 kg/m2
Patients (n = 41)	4105	ns
Healthy (n = 9)	3803
ΒΜΙ ≥ 25 kg/m2
Patients (n = 96)	4251	0.004
Healthy (n = 19)	2357

Note: Bold values indicate statistically significant results (p < 0.05).

In contrast, among individuals with BMI ≥ 25 kg/m², serum resistin levels were significantly higher in CRC patients compared with healthy controls. Median resistin concentrations were 4251 pg/mL in CRC patients (n = 96) versus 2357 pg/mL in healthy individuals (n = 19) (p = 0.004). These findings indicate that the association between higher circulating resistin and colorectal cancer is most evident in individuals with BMI ≥ 25 kg/m².

3.7. Multivariable Analysis

Ordinal logistic regression analysis, including the 93 CRC patients with complete data on resistin, BMI, fasting insulin, CEA and CA19‐9, revealed that serum resistin was independently and inversely associated with tumour stage after adjustment for the other four covariates (coefficient −0.181 per 1000 pg/mL; p = 0.001) (Table 5). The proportional odds assumption was assessed using a likelihood‐ratio comparison of the proportional‐odds model against a model permitting predictor effects to vary across stage cut‐points (analogous in principle to a parallel‐lines/Brant‐type test) and was not violated (χ ²(10) = 3.80, p = 0.956). The number of events per predictor was considered adequate for the five covariates included. Multicollinearity among resistin, BMI, fasting insulin, CEA and CA19‐9 was assessed using variance inflation factors, all of which were below 4.5 (range 1.07–4.34), indicating no problematic collinearity. Per 1000 pg/mL increase in resistin, the odds ratio for higher tumour stage was 0.83 (95% confidence interval 0.75–0.93; p = 0.001). When scaled to the interquartile range (5325 pg/mL), an interquartile increase in resistin corresponded to approximately 62% lower odds of advanced tumour stage (OR 0.38, 95% CI 0.22–0.66), indicating a clinically meaningful stage shift. BMI and CA19‐9 were also independently associated with tumour stage in this model (p = 0.016 and p = 0.037, respectively), whereas fasting insulin and CEA were not significant independent predictors once resistin, BMI and CA19‐9 were accounted for (p = 0.201 and p = 0.213, respectively). As a sensitivity analysis addressing potential confounding by nutritional status, serum albumin was added to the model in the subset of patients with available albumin measurements (n = 86). The association between resistin and tumour stage remained essentially unchanged in this albumin‐adjusted model (OR 0.83, 95% CI 0.75–0.92, p = 0.001), indicating that the resistin–stage association is not explained by nutritional or cachexia‐related confounding. Albumin itself was not independently associated with stage in this model (OR 0.52, 95% CI 0.24–1.16, p = 0.112).

TABLE 5.

Ordinal logistic regression analysis evaluating independent predictors of tumour stage.

Predictor	Odds ratio (OR)	95% CI	p
Serum resistin (per 1000 pg/mL increase)	0.83	0.75–0.93	0.001
Body mass index (per kg/m2)	0.90	0.82–0.98	0.016
Carcinoembryonic antigen (CEA, per ng/mL)	1.01	0.99–1.03	0.213
CA19‐9 (per U/mL)	1.02	1.00–1.04	0.037
Fasting insulin (per μU/mL)	1.01	1.00–1.02	0.201

Note: Bold values indicate statistically significant results (p < 0.05).

4. Discussion

4.1. A Stage‐Dependent Inflammatory Model of Resistin in Colorectal Tumorigenesis

This study demonstrates a non‐monotonic, stage‐related pattern of circulating resistin across the colorectal adenoma–carcinoma sequence, with higher cross‐sectional concentrations in premalignant lesions and early‐stage colorectal cancer and lower concentrations in advanced‐stage disease. Because the design is cross‐sectional, this pattern reflects differences between stage groups rather than a confirmed sequential or longitudinal trajectory within individual patients. This pattern contrasts with classical tumour burden markers such as carcinoembryonic antigen and carbohydrate antigen 19‐9, which increased progressively with advancing stage, consistent with their established roles as tumour burden indicators [16, 17]. Notably, in the multivariable ordinal regression model, CEA no longer reached statistical significance once resistin, BMI and CA19‐9 were accounted for, whereas CA19‐9 retained an independent association with stage; this is consistent with partial overlap between CEA and CA19‐9 as correlated tumour burden indicators (VIF 4.26 and 4.34, respectively), and suggests that CA19‐9 may carry more independent prognostic information than CEA in this cohort once resistin and other covariates are considered. These cross‐sectional findings are consistent with, but do not establish, a model in which circulating resistin reflects early host inflammatory activation during colorectal tumorigenesis rather than tumour mass.

Colorectal cancer develops through a multistep adenoma–carcinoma sequence involving cumulative genetic alterations and sustained interactions between transformed epithelial cells and the inflammatory microenvironment [3, 4, 5, 6]. Increasing evidence indicates that chronic inflammation and immune dysregulation play central roles in colorectal tumorigenesis [6]. Adipokines represent an important interface between metabolic status, immune regulation and systemic inflammation [7, 8]. Among these, resistin has been shown in humans to be predominantly expressed by monocytes and macrophages and to function as a pro‐inflammatory cytokine rather than a primary metabolic regulator [10, 11, 12].

Most previous clinical studies have evaluated resistin in a binary case–control framework and have frequently reported elevated circulating levels in colorectal cancer compared with healthy controls [13, 14, 15]. However, stage‐stratified analyses and inclusion of premalignant lesions have been limited. The inclusion of adenomas in the present study provides important biological context, suggesting that resistin elevation may occur early during epithelial transformation rather than increasing linearly with tumour burden. This stage‐dependent model may reconcile inconsistencies in prior reports and underscores the importance of disease timing when interpreting inflammatory biomarkers.

The higher circulating resistin observed in adenomas relative to healthy individuals and in stage I–II disease compared to more advanced stages may reflect heightened immune engagement within the colonic mucosa, although this study did not directly assess immune cell activity and this interpretation remains hypothetical. Macrophage recruitment and inflammatory cytokine signalling are recognised features of early tumour development [6, 18, 19, 20]. Resistin promotes inflammatory cascades and interacts with cytokines such as TNF‐α and IL‐6 [10, 11, 12], supporting a plausible role as a marker of early inflammatory activation. Conversely, advanced malignancy has been associated in other studies with immune remodelling, expansion of regulatory immune populations and altered macrophage polarisation [18, 19, 20]; we propose, as a hypothesis to be tested rather than a conclusion supported by our data, that the decline in circulating resistin in stage III–IV disease may reflect a transition from inflammatory activation to tumour‐associated immune modulation. No immune‐cell profiling, cytokine analysis or tissue expression studies were performed in this cohort, and this hypothesis requires direct mechanistic investigation.

4.2. Independence From Insulin Resistance and Metabolic Status

Although resistin was initially described in murine models as a mediator linking obesity to insulin resistance [9], human studies have demonstrated that resistin primarily functions as an inflammatory cytokine [10, 11, 12]. In this study, circulating resistin levels were not associated with fasting insulin or HOMA‐IR, and no significant correlation was observed between resistin and systemic insulin resistance indices. These findings indicate that resistin predominantly acts as an inflammatory mediator rather than a marker of metabolic dysregulation in colorectal cancer.

Importantly, the stage‐dependent decline in resistin persisted after adjustment for BMI in multivariable ordinal regression analysis. While BMI decreased in advanced‐stage disease, consistent with cancer‐associated weight loss, resistin retained an independent association with tumour stage. This suggests that the observed pattern cannot be explained solely by metabolic deterioration or cachexia, although residual confounding related to nutritional status cannot be fully excluded.

4.3. Implications

Collectively, the data are consistent with a model in which circulating resistin reflects early immune–inflammatory activation within the adenoma–carcinoma sequence rather than tumour burden, although this remains a hypothesis given the cross‐sectional, observational design. The stage‐related pattern observed offers one possible biological explanation for previously inconsistent associations reported in the literature [13, 14, 15]. These findings emphasise the importance of stage stratification and inclusion of premalignant lesions when evaluating circulating inflammatory mediators in oncology, and align with the growing recognition of dynamic, stage‐specific immune remodelling within the colorectal tumour microenvironment described in recent literature [21].

Further studies incorporating serial biomarker assessment and detailed immune phenotyping will be required to determine whether resistin plays a mechanistic role in tumour progression or primarily serves as a systemic marker of early inflammatory engagement.

5. Conclusion

Circulating resistin shows a stage‐related, non‐monotonic pattern across colorectal neoplasia in this cross‐sectional cohort, with higher levels in premalignant and early malignant stages and lower levels in advanced disease. This pattern is independent of insulin resistance and distinct from classical tumour burden markers. These findings are consistent with, though do not prove, a model in which resistin reflects early host inflammatory activation within the adenoma–carcinoma sequence rather than tumour mass or metastatic extent. Further mechanistic and longitudinal studies are necessary to elucidate the role of resistin in colorectal tumorigenesis.

6. Limitations

This study has several limitations. Its cross‐sectional design does not allow assessment of changes in circulating resistin levels over time and therefore precludes causal inference regarding disease progression; apparent differences across stages should not be interpreted as sequential or longitudinal change in individual patients. The single‐centre setting may limit the generalisability of our findings to other populations. The adenoma group was small (n = 11) relative to the CRC group (n = 137); a formal power calculation was not performed prospectively, and this disparity may introduce sampling bias and limit the robustness of conclusions drawn about the premalignant subgroup. We did not measure markers of systemic inflammation such as CRP, ESR, IL‐6, TNF‐α or leukocyte count; although participants with overt inflammatory or autoimmune disease were excluded, residual confounding from subclinical systemic inflammation cannot be excluded. We did not perform detailed immune phenotyping or cytokine profiling, which limits mechanistic insight into the biological processes underlying the observed stage‐dependent pattern; proposed mechanisms such as macrophage polarisation or immune remodelling in advanced disease are presented as hypotheses rather than established conclusions. Assay reproducibility was not formally assessed (e.g., inter‐assay coefficients of variation), and no external validation cohort was available, which limits assessment of measurement reliability. Although analyses were adjusted for major metabolic and tumour‐related variables, and a sensitivity analysis incorporating serum albumin supported the robustness of the resistin–stage association, residual confounding related to systemic inflammation more broadly cannot be entirely excluded, particularly given the lower BMI observed in stage IV disease. Finally, as only a single resistin measurement per patient was obtained, the term ‘biphasic’ should be interpreted as describing a cross‐sectional, stage‐related pattern rather than a confirmed temporal trajectory within individuals. Prospective studies incorporating longitudinal biomarker assessment, immune profiling and external assay validation are needed to clarify the role of resistin in colorectal tumorigenesis.

Author Contributions

Pantelis Vassiliu: conceptualization, writing – review and editing, supervision. Dimitris Tsakogiannis: conceptualization, methodology, data curation, investigation, writing – review and editing. Aliki Vaia Rompou: conceptualization, methodology, data curation, investigation, validation, formal analysis, writing – original draft, writing – review and editing. Stamatios Theocharis: conceptualization, methodology, supervision, writing – review and editing. Nikolaos Danias: conceptualization, investigation, data curation, supervision, writing – review and editing. Garyfalia Bletsa: conceptualization, methodology, writing – review and editing, data curation.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The datasets generated during the current study are available from the corresponding author upon reasonable request.