Risk stratification in colorectal cancer: A retrospective cohort study of nomograms based on inflammatory and immunohematological biomarkers Authors: Jiadai Tang, Mengying Xiang, Guangrui Xiong, Liuhan Chen, Yedan Liao, Xin Shen, Rong Li, Yanping Lin, Ke Zhang, Lin Xie Published: 2026 ABSTRACT ObjectiveInflammation and immunohematological biomarkers are recognized as contributing factors in the progression of solid tumors. This study aimed to develop nomograms incorporating inflammatory and immunohematological biomarkers to predict the prognosis of colorectal cancer.MethodsWe retrospectively enrolled 432 patients with colorectal cancer at Yunnan Cancer Hospital from January 2019 to December 2022. Clinicopathologic data and blood samples were obtained at initial diagnosis before any treatment, and therapy was initiated within 3 days. Patients with prior therapy, secondary cancers, or systemic conditions affecting hematology were excluded. Metastases were confirmed by pathology or imaging, and overall survival and progression-free survival were recorded through follow-up. Multivariate logistic and Cox regression analyses, nomogram construction, and internal validation with 1000-bootstrap resampling were performed using R, with risk groups defined by receiver operating characteristic curve–derived cutoffs.ResultsIn the training cohort, the model accuracy for predicting distant metastasis was 66.25%. The concordance index values for the three nomograms were 0.791 (95% confidence interval, 0.745–0.838), 0.752 (95% confidence interval, 0.699–0.806), and 0.687 (95% confidence interval, 0.647–0.726), respectively. Calibration plots showed strong agreement between predicted and observed outcomes, while decision curve analysis indicated potential clinical utility. Subgroup analyses confirmed that the nomograms were effective prognostic indicators across stage I–IV colorectal cancer patients (P < 0.05).ConclusionSingle or combined hematological indicators are readily obtainable and feasible and possess high prognostic value. Nomograms based on cancer-associated inflammatory factors and immunohematological biomarkers demonstrated robust accuracy in predicting distant metastasis, overall survival, and progression-free survival in patients with colorectal cancer. These tools may assist clinicians in risk stratification and treatment decision-making, ultimately supporting improved oncological outcomes. Introduction Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer-associated deaths globally, and its epidemiology varies markedly across regions, age groups, sexes, and races. 1 Although some countries in Europe and the United States have seen a decrease in CRC incidence, China has witnessed a substantial rise in both incidence and mortality. 2 Xi et al. 3 predicted that the number of new CRC cases worldwide will reach 3.2 million by 2040, driven by population aging and growth. The tumor–node–metastasis (TNM) staging system, which considers tumor volume, depth of infiltration, and regional lymph node metastasis, is commonly used to predict prognosis. However, due to individual patient disparities and tumor heterogeneity, patients with the same TNM stage may have different survival outcomes. Therefore, identifying simple, rapid, and cost-effective prognostic factors for risk stratification is necessary for predicting the biological behavior of CRC and for early determination of prognosis. With advances in the understanding of tumor microenvironment mechanisms, the systemic inflammatory response has been recognized as an important immunological mechanism involved in cancer cell proliferation, migration, metastasis, and chemotherapy resistance. 4 Several studies have shown that chronic inflammation could inhibit tumor cell apoptosis, increase tumor angiogenesis, and enhance the likelihood of tumor metastasis. Immunoinflammatory cells, which play crucial roles in the tumor microenvironment, are also important in tumor growth, progression, invasion, and metastasis. Peripheral blood parameters that can be routinely measured, such as lymphocyte (LYM), platelets (PLT), systemic immune-inflammation index (neutrophil*platelet-to-lymphocyte ratio, SII), derived neutrophil-to-lymphocyte ratio (dNLR), CD4+ T-helper cells, CD8+ T cells, and other immune-inflammation indexes, have been shown to be reliable predictors and surrogates of cancer-associated immunological inflammation and have progressively been regarded as important prognostic predictors of certain solid tumors.5–9 In CRC, host inflammation is associated with tumor characteristics and disease progression. Persistently progressing cancer can induce tissue inflammation, creating a microenvironment favorable for tumor survival. Elevated systemic inflammatory responses have been linked to poor outcomes independent of tumor stage. 10 The recently emerged concept of cancer immunoediting also suggests that the immune system has a dual role in cancer development. Studies have demonstrated that either CD4+ T-cell or CD8+ T-cell infiltrates in breast cancer tumors participate in the immune response but mediate different clinical roles. Patients with high levels of CD4+ T-cell infiltration have an increased cumulative risk of mortality and recurrence over time, whereas high levels of tumor-infiltrating CD8+ T cells are prognostically beneficial. The ratio of the two, known as the CD4+ T-helper cell-to-CD8+ T-cell ratio (CD4+Th/CD8+T), is a valuable clinical indicator for evaluating the risk of late recurrence and survival in breast cancer. 7 Meanwhile, natural killer (NK) cells can attack tumor cells, and patients with CRC and gastric cancer may benefit from high NK cell levels in controlling tumor progression.11–12 However, some studies suggest that NK cell activity is influenced by tumor burden, with a heavier tumor burden leading to lower NK cell activity and a subsequent rebound of NK cells to destroy tumor cells. 13 This study aimed to develop nomograms incorporating inflammatory and immunohematological biomarkers to predict the prognosis of CRC. Methods Sample collection and laboratory analysis We retrospectively enrolled 432 patients with CRC from January 2019 to December 2022 at Yunnan Cancer Hospital. Inclusion criteria were as follows: (a) CRC confirmed by pathology, imaging, and clinical evaluation, with data reviewed by at least two independent observers; (b) no prior treatment and complete clinical and imaging records; and (c) absence of secondary cancers. Exclusion criteria included the following: (a) systemic diseases affecting hematological parameters; (b) inability or refusal to cooperate with clinical data collection or follow-up; and (c) shortened survival or death due to factors unrelated to the primary tumor. All laboratory samples were collected at diagnosis, before any therapeutic intervention. Treatment was initiated within 3 days of diagnosis and sample collection, resulting in a short time to treatment initiation (TTI) that is unlikely to influence baseline laboratory measurements or their predictive validity. Metastases were confirmed by pathology or imaging, and overall survival (OS) and progression-free survival (PFS) were documented during follow-up. Survival time was defined as the period from diagnosis to either death or December 2023. Collected clinicopathologic indicators included sex, age, Karnofsky Performance Status (KPS), body mass index (BMI), smoking history, drinking history, family history, primary tumor site, and pathologic TNM stage, determined according to the American Joint Committee on Cancer (AJCC) 8th edition (2017). The study was conducted and reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines. 14 The Ethics Committee of the Third Affiliated Hospital of Kunming Medical University, Yunnan Cancer Hospital, approved the study on 9 December 2022. The review number was KY LX 2022210. Statistical analysis All statistical analyses, including development and validation of nomograms, were performed using R software version 4.3.1 (https://www.r-project.org/). Predictive models were internally validated using the bootstrap method with 1000 resamples, and the concordance index (C-index) was calculated. Calibration curves were generated based on the Hosmer–Lemeshow test. For baseline variables with missing values, multiple imputation was performed using the fully conditional specification discriminant function for categorical variables and fully conditional specification regression for continuous variables. The sample size of 432 patients was deemed adequate for developing and internally validating the 3 prognostic models based on established methodological guidelines. Specifically, the number of outcome events for distant metastasis, OS, and PFS satisfied the recommended events-per-variable ratio, ensuring sufficient statistical power and stability of model estimates. The number of candidate predictors relative to the number of events supports the reliability of the multivariate analyses, and internal validation using 1000-bootstrap resampling further mitigated potential overfitting. Survival outcomes were estimated using the Kaplan–Meier method and compared using the log-rank test. Variables associated with distant metastasis, OS, and PFS in univariate analyses were subsequently entered into multivariate models. Distant metastasis was defined as tumor spread beyond the primary site, confirmed by pathological examination or imaging modalities, including ultrasound, computed tomography, or magnetic resonance imaging, and included both synchronous and metachronous events. OS was defined as the interval from diagnosis to death from any cause or last follow-up, while PFS was defined as the time from diagnosis to disease progression, recurrence, metastasis, or death. Patients were monitored through scheduled clinical visits or telephone follow-up. Multivariable logistic regression and Cox regression analyses were used to identify independent prognostic factors. Predictors with P < 0.05 in univariate analyses were included in the multivariate regression models. Due to the retrospective design and heterogeneity of clinical management across disease stages, detailed treatment-related variables (e.g. surgical margin status, adjuvant chemotherapy, targeted therapy, and immunotherapy) were not consistently available and were therefore not incorporated into the multivariable analyses. Optimal cutoff values for risk score–based grouping were determined using receiver operating characteristic (ROC) curves, with thresholds selected according to the Youden index and derived exclusively from the training set. Missing data were handled using multiple imputation, with all candidate variables included, and the proportion of missingness was reported. Model stability was ensured by generating multiple imputations and evaluating convergence and plausibility of the imputed datasets. A P-value <0.05 was considered statistically significant. All patient details were de-identified to ensure anonymity. Results Clinicopathological characteristics A total of 432 patients with CRC who met all inclusion criteria were enrolled in the study. Their clinical characteristics are summarized in Table 1. The cohort comprised 252 males (58.33%) and 180 females (41.67%), with a median age of 61.25 (range: 23–86) years. More than half of the patients presented with higher KPS (>90; n = 236, 54.63%), a BMI (>18.33; n = 403, 93.29%), and a history of smoking (n = 235, 54.40%) and alcohol consumption (n = 255, 59.03%) at initial diagnosis. Regarding primary tumor location, 106 tumors (24.54%) were detected in the right colon, 93 (21.53%) in the left colon, and 226 (52.31%) in the rectum, while the tumor site remained undetermined in 7 patients (1.62%). The distribution of disease stages I, II, III, and IV was 51 (11.81%), 151 (34.95%), 127 (29.40%), and 103 (23.84%), respectively. Regional lymph node metastasis was identified in 254 patients (58.80%). Among the 432 patients, 102 (23.61%) had synchronous metastases at initial diagnosis. During follow-up, 182 patients (42.13%) who were metastasis-free at baseline developed new metastatic lesions, whereas 148 patients (34.26%) remained free of metastasis throughout the study period (Table 2, Figure 1). Table 1.Demographic and clinicopathological characteristics of the study population.NumberFrequency (%)Sex Male25258.33 Female18041.67Age (years) <53.512428.70 ≥53.530871.30KPS ≤9019645.37 >9023654.63BMI ≤18.33296.71 >18.3340393.29Smoking history Yes23554.40 No19745.60Drinking history Yes25559.03 No17740.97Family history Yes31071.76 No12228.24Primary tumor site Right-side colon10624.54 Left-side colon9321.53 Rectum22652.31 Indistinguishable71.62Clinical stage I5111.81 II15134.95 III12729.40 IV10323.84T stage T181.85 T24911.34 T313531.25 T424055.56N stage N017841.20 N112328.48 N213030.09 N310.23KPS: Karnofsky Performance Status; BMI: body mass index. Demographic and clinicopathological characteristics of the study population. KPS: Karnofsky Performance Status; BMI: body mass index. Table 2.Metastatic condition.Metastatic timeNumberFrequency (%)Metastatic group28465.74Initial diagnosis10235.92Follow-up18264.08Non-metastatic group14834.26Total432100 Metastatic condition. Group composition. Independent prognostic factors for distant metastasis, OS, and PFS Univariate and multivariate logistic regression analyses demonstrated that the presence of regional lymph node metastasis at initial diagnosis (odds ratio (OR): 9.104, 95% confidence interval (CI): 5.567–14.890), LYM (OR: 1.973, 95% CI: 1.139–3.416), the CD4+/CD8+ T-cell ratio (OR: 1.722, 95% CI: 1.045–2.837), and the proportion of NK cells (NK%) (OR: 2.094, 95% CI: 1.228–3.572) were independent predictors of distant metastasis (all P < 0.05) (Table 3, Figure 2(a)). Interestingly, higher peripheral lymphocyte counts and NK cell percentages were associated with an increased risk of distant metastasis in our cohort. Although these findings appear counterintuitive given the known antitumor roles of these cells, sensitivity analyses confirmed the robustness of the associations. We emphasize that these results should be interpreted cautiously and require validation in independent cohorts and mechanistic studies. Tables 4 and 5 present the univariate and multivariate Cox regression analyses, in which all variables were dichotomized according to predefined cutoff values. Patient characteristics included age (<53.5 vs. ≥53.5 years), KPS (≤90 vs. >90), and BMI (≤18.33 vs. >18.33 kg/m2), together with tumor-related factors such as infiltration depth (below the mucous layer vs. above the mucous layer) and initial regional lymph node metastasis (yes vs. no). Hematological parameters comprised white blood cell count (WBC; <4.75 vs. ≥4.75 × 109/L), LYM (<1.42 vs. ≥1.42 × 109/L), monocytes (MNO; <0.25 vs. ≥0.25 × 109/L), neutrophils (NEUT; <3.06 vs. ≥3.06 × 109/L), and PLT (<224.5 vs. ≥224.5 × 109/L). Inflammatory ratios and indices included the neutrophil-to-WBC ratio (NWR; <3.06 vs. ≥3.06), neutrophil-to-lymphocyte ratio (NLR; <1.08 vs. ≥1.08), lymphocyte-to-monocyte ratio (LMR; ≤3.26 vs. >3.26), platelet-to-lymphocyte ratio (PLR; <92.74 vs. ≥92.74), SII (<435.93 vs. ≥435.93), aggregate index of systemic inflammation (AISI; <474.39 vs. ≥474.39), dNLR (<0.62 vs. ≥0.62), monocyte-to-lymphocyte ratio (MLR; ≤0.46 vs. >0.46), neutrophil-to-(lymphocyte × platelet) ratio (NLPR; <110.67 vs. ≥110.67), and systemic inflammation response index (SIRI; <1.95 vs. ≥1.95). Nutritional and composite indices included albumin (ALB; ≤41.5 vs. >41.5 g/L), prognostic nutritional index (PNI; ≤51.53 vs. >51.53), and advanced lung cancer inflammation index (ALI; ≤171.38 vs. >171.38). Immune cell subsets were defined as CD45+T/CD3+T-cell ratio (≤87.55% vs. >87.55%), CD3+T/CD4+ T-helper (Th) cell ratio (≤26.95% vs. >26.95%), CD3+T/CD8+ T suppressor (Ts) cell ratio (≤8.0% vs. >8.0%), CD4+Th/CD8+T ratio (<1.36 vs. ≥1.36), NK cells (<13.55% vs. ≥13.55%), and cytokine-induced killer (CIK) cells (<7.1% vs. ≥7.1%). Cytokine levels included interleukin (IL)-10 (<12.15 vs. ≥12.15 pg/mL), IL-6 (<13.35 vs. ≥13.35 pg/mL), IL-4 (<30.15 vs. ≥30.15 pg/mL), and IL-2 (<22.25 vs. ≥22.25 pg/mL). Univariate and multivariate Cox regression analyses identified three and four independent prognostic factors for OS and PFS, respectively. TNM stage (hazard ratio (HR): 6.905, 95% CI: 4.554–10.472), PLT (HR: 2.616, 95% CI: 1.498–4.569), and AISI (HR: 0.486, 95% CI: 0.280–0.844) were independently associated with OS. For PFS, TNM stage (HR: 3.949, 95% CI: 2.875–5.424), PLT (HR: 1.528, 95% CI: 1.087–2.147), SII (HR: 1.515, 95% CI: 1.087–2.114), and dNLR (HR: 1.453, 95% CI: 1.082–1.951) were identified as independent risk factors (all P < 0.05) (Tables 4 and 5, Figure 2(b) and 2(c)). Table 3.Univariate and multivariate logistic regression analysis of prognostic factors for predicting distant metastasis.VariablesUnivariate analysisMultivariate analysisOR (95% CI)POR (95% CI)PSex0.895 (0.599–1.340)0.591Age (years)1.520 (0.988–2.339)0.057KPS1.080 (0.725–1.609)0.706BMI1.386 (0.643–2.985)0.405Infiltration depth1.139 (0.665–1.952)0.635Initial regional LN metastasis8.709 (5.524–13.730)<0.0019.104 (5.567–14.890)<0.001WBC (×109 L)1.775 (1.025–3.072)0.0400.999 (0.475–2.103)0.998LYM (×109 L)2.015 (1.285–3.161)0.0021.973 (1.139–3.416)0.015MNO (×109 L)1.804 (1.023–3.184)0.0421.289 (0.637–2.610)0.480NEUT (×109 L)1.688 (1.095–2.602)0.0181.593 (0.934–2.717)0.087PLT (×109 L)1.270 (0.839–1.923)0.258NWR1.357 (0.823–2.237)0.232NLR1.500 (0.786–2.863)0.219LMR1.150 (0.734–1.801)0.543PLR1.804 (1.023–3.184)0.0421.966 (0.912–4.242)0.085SII1.335 (0.886–2.012)0.167AISI1.004 (0.669–1.509)0.983dNLR1.348 (0.901–2.015)0.146MLR1.940 (0.932–4.037)0.076NLPR1.508 (0.946–2.403)0.084SIRI1.727 (0.959–3.110)0.069ALB1.327 (0.828–2.128)0.239PNI1.275 (0.840–1.934)0.253ALI1.207 (0.785–1.855)0.391CD45+T/CD3+T (%)3.443 (0.001–1.998)0.980CD3+T/CD4+Th (%)1.445 (0.842–2.477)0.181CD3+T/CD8+Ts (%)2.917 (0.482–17.655)0.244CD4+Th/CD8+T (%)1.504 (1.002–2.257)0.0491.722 (1.045–2.837)0.033NK cell (%)1.825 (1.185–2.810)0.0062.094 (1.228–3.572)0.007CIK cell (%)1.049 (0.589–1.870)0.871IL10 (pg/mL)1.518 (1.018–2.263)0.0411.216 (0.723–2.045)0.460IL6 (pg/mL)1.523 (1.021–2.271)0.0391.382 (0.828–2.305)0.216IL4 (pg/mL)1.519 (0.979–2.357)0.062IL2 (pg/mL)1.647 (1.060–2.560)0.0271.372 (0.772–2.436)0.281OR: odds ratio; CI: confidence interval; KPS: Karnofsky score; BMI: body mass index; LNs: lymph nodes; WBC: white blood cells; LYM: lymphocytes; MNO: monocytes; NEUT: neutrophils; PLT: platelets; NWR: neutrophil-to-white blood cell ratio; NLR: neutrophil-to-lymphocyte ratio; LMR: lymphocyte-to-monocyte ratio; PLR: platelet-to-lymphocyte ratio; SII: systemic immune-inflammation index; AISI: aggregate systemic inflammation index; dNLR: derived neutrophil-to-lymphocyte ratio; MLR: monocyte-to-lymphocyte ratio; NLPR: neutrophil-to-(lymphocyte × platelet) ratio; SIRI: systemic inflammation response index; ALB: albumin; PNI: prognostic nutritional index; ALI: advanced lung cancer inflammation index; CD45+T/CD3+T: CD45+T/CD3+T ratio; CD3+T/CD4+Th: CD3+T/CD4+Th ratio; CD3+T/CD8+Ts: CD3+T/CD8+Ts ratio; CD4+Th/CD8+T: CD4+Th/CD8+T ratio; NK: natural killer cells; CIK: cytokine-induced killer cells; IL-10: interleukin-10; IL-6: interleukin-6; IL-4: interleukin-4; IL-2: interleukin-2. Univariate and multivariate logistic regression analysis of prognostic factors for predicting distant metastasis. OR: odds ratio; CI: confidence interval; KPS: Karnofsky score; BMI: body mass index; LNs: lymph nodes; WBC: white blood cells; LYM: lymphocytes; MNO: monocytes; NEUT: neutrophils; PLT: platelets; NWR: neutrophil-to-white blood cell ratio; NLR: neutrophil-to-lymphocyte ratio; LMR: lymphocyte-to-monocyte ratio; PLR: platelet-to-lymphocyte ratio; SII: systemic immune-inflammation index; AISI: aggregate systemic inflammation index; dNLR: derived neutrophil-to-lymphocyte ratio; MLR: monocyte-to-lymphocyte ratio; NLPR: neutrophil-to-(lymphocyte × platelet) ratio; SIRI: systemic inflammation response index; ALB: albumin; PNI: prognostic nutritional index; ALI: advanced lung cancer inflammation index; CD45+T/CD3+T: CD45+T/CD3+T ratio; CD3+T/CD4+Th: CD3+T/CD4+Th ratio; CD3+T/CD8+Ts: CD3+T/CD8+Ts ratio; CD4+Th/CD8+T: CD4+Th/CD8+T ratio; NK: natural killer cells; CIK: cytokine-induced killer cells; IL-10: interleukin-10; IL-6: interleukin-6; IL-4: interleukin-4; IL-2: interleukin-2. Multivariate forest map (a) for distant metastasis, (b) OS, and (c) PFS. AISI acts as a protective factor for OS, consistent with the regression analysis. AISI: aggregate index of systemic inflammation; OS: overall survival; PFS: progression-free survival. Table 4.Univariate and multivariate Cox regression analysis of prognostic factors for OS.VariablesUnivariate analysisMultivariate analysisHR (95% CI)PHR (95% CI)PSex0.910 (0.606–1.366)0.649Age (years)1.151 (0.730–1.816)0.544KPS0.967 (0.649–1.442)0.871BMI0.712 (0.345–1.470)0.359TNM stage6.887 (4.561–10.401)<0.0016.905 (4.554–10.472)<0.001WBC (×109 L)0.761 (0.451–1.284)0.306LYM (×109 L)1.064 (0.670–1.688)0.793MNO (×109 L)0.968 (0.539–1.737)0.913NEUT (×109 L)1.110 (0.700–1.761)0.656PLT (×109 L)1.790 (1.121–2.859)0.0152.616 (1.498–4.569)0.001NWR0.673 (0.393–1.151)0.148NLR0.876 (0.467–1.642)0.679LMR1.411 (0.854–2.330)0.178PLR1.014 (0.565–1.818)0.964SII0.889 (0.590–1.340)0.575AISI0.654 (0.439–0.975)0.0370.486 (0.280–0.844)0.010dNLR1.008 (0.676–1.503)0.968MLR0.924 (0.465–1.835)0.821NLPR0.985 (0.607–1.599)0.953SIRI0.866 (0.483–1.554)0.629ALB1.366 (0.809–2.306)0.243PNI1.725 (1.073–2.776)0.0251.306 (0.773–2.207)0.318ALI1.531 (1.019–2.300)0.040.781 (0.447–1.366)0.387CD45+T/CD3+T (%)0.817 (0.114–5.864)0.840CD3+T/CD4+Th (%)1.235 (0.674–2.261)0.494CD3+T/CD8+Ts (%)0.716 (0.100–5.148)0.740CD4+Th/CD8+T (%)1.202 (0.784–1.842)0.399NK cell (%)1.179 (0.748–1.859)0.478CIK cell (%)1.520 (0.900–2.567)0.117IL-10 (pg/mL)0.714 (0.478–1.065)0.874IL-6 (pg/mL)1.420 (0.940–2.144)0.166IL-4 (pg/mL)1.014 (0.664–1.549)0.099IL-2 (pg/mL)1.178 (0.777–1.786)0.096HR: hazard ratio; CI: confidence interval; KPS: Karnofsky Performance Status; BMI: body mass index; TNM: tumor–node–metastasis stage; WBC: white blood cells; LYM: lymphocytes; MNO: monocytes; NEUT: neutrophils; PLT: platelets; NWR: neutrophil-to-white blood cell ratio; NLR: neutrophil-to-lymphocyte ratio; LMR: lymphocyte-to-monocyte ratio; PLR: platelet-to-lymphocyte ratio; SII: systemic immune-inflammation index; AISI: aggregate index of systemic inflammation; dNLR: derived neutrophil-to-lymphocyte ratio; MLR: monocyte-to-lymphocyte ratio; NLPR: neutrophil-to-(lymphocyte × platelet) ratio; SIRI: systemic inflammation response index; ALB: albumin; PNI: prognostic nutritional index; ALI: advanced lung cancer inflammation index; CD45+T/CD3+T: CD45+ to CD3+ T-cell ratio; CD3+T/CD4+Th: CD3+ to CD4+ T-helper cell ratio; CD3+T/CD8+Ts: CD3+ to CD8+ T suppressor cell ratio; CD4+Th/CD8+T: CD4+ T-helper to CD8+ T-cell ratio; NK: natural killer cells; CIK: cytokine-induced killer cells; IL-10: interleukin-10; IL-6: interleukin-6; IL-4: interleukin-4; IL-2: interleukin-2. Univariate and multivariate Cox regression analysis of prognostic factors for OS. HR: hazard ratio; CI: confidence interval; KPS: Karnofsky Performance Status; BMI: body mass index; TNM: tumor–node–metastasis stage; WBC: white blood cells; LYM: lymphocytes; MNO: monocytes; NEUT: neutrophils; PLT: platelets; NWR: neutrophil-to-white blood cell ratio; NLR: neutrophil-to-lymphocyte ratio; LMR: lymphocyte-to-monocyte ratio; PLR: platelet-to-lymphocyte ratio; SII: systemic immune-inflammation index; AISI: aggregate index of systemic inflammation; dNLR: derived neutrophil-to-lymphocyte ratio; MLR: monocyte-to-lymphocyte ratio; NLPR: neutrophil-to-(lymphocyte × platelet) ratio; SIRI: systemic inflammation response index; ALB: albumin; PNI: prognostic nutritional index; ALI: advanced lung cancer inflammation index; CD45+T/CD3+T: CD45+ to CD3+ T-cell ratio; CD3+T/CD4+Th: CD3+ to CD4+ T-helper cell ratio; CD3+T/CD8+Ts: CD3+ to CD8+ T suppressor cell ratio; CD4+Th/CD8+T: CD4+ T-helper to CD8+ T-cell ratio; NK: natural killer cells; CIK: cytokine-induced killer cells; IL-10: interleukin-10; IL-6: interleukin-6; IL-4: interleukin-4; IL-2: interleukin-2. Table 5.Univariate and multivariate Cox regression analysis of prognostic factors for PFS.VariablesUnivariate analysisMultivariate analysisHR (95% CI)PHR (95% CI)PSex0.772 (0.569–1.048)0.097Age (years)1.225 (0.874–1.719)0.239KPS1.079 (0.803–1.451)0.613BMI1.574 (0.774–3.199)0.210TNM stage3.773 (2.757–5.163)<0.0013.949 (2.875–5.424)<0.001WBC (×109 L)1.153 (0.744–1.787)0.524LYM (×109 L)1.482 (1.028–2.138)0.0351.429 (0.990–2.062)0.057MNO (×109 L)1.358 (0.844–2.186)0.207NEUT (×109 L)1.386 (0.975–1.971)0.069PLT (×109 L)1.639 (1.174–2.287)0.0041.528 (1.087–2.147)0.015NWR0.852 (0.591–1.229)0.392NLR0.968 (0.594–1.576)0.895LMR0.908 (0.652–1.264)0.568PLR1.265 (0.786–2.036)0.333SII1.402 (1.015–1.937)0.0401.515 (1.087–2.114)0.014AISI1.126 (0.830–1.528)0.446dNLR1.403 (1.045–1.883)0.0241.453 (1.082–1.951)0.013MLR1.207 (0.765–1.903)0.419NLPR1.207 (0.830–1.754)0.324SIRI1.336 (0.919–1.941)0.129ALB1.082 (0.753–1.554)0.671PNI1.158 (0.841–1.594)0.368ALI1.070 (0.782–1.464)0.673CD45+T/CD3+T (%)1.301 (0.416–4.076)0.651CD3+T/CD4+Th (%)1.115 (0.731–1.703)0.613CD3+T/CD8+Ts (%)1.705 (0.239–12.178)0.595CD4+Th/CD8+T (%)1.066 (0.784–1.449)0.684NK cell (%)1.185 (0.847–1.658)0.322CIK cell (%)1.374 (0.923–2.047)0.612IL-10 (pg/mL)1.079 (0.803–1.450)0.271IL-6 (pg/mL)1.182 (0.878–1.591)0.318IL-4 (pg/mL)1.169 (0.860–1.590)0.176IL-2 (pg/mL)1.234 (0.910–1.674)0.612HR: hazard ratio; CI: confidence interval; KPS: Karnofsky Performance Status; BMI: body mass index; TNM: tumor–node–metastasis; WBC: white blood cells; LYM: lymphocytes; MNO: monocytes; NEUT: neutrophils; PLT: platelets; NWR: neutrophil/WBC ratio; NLR: neutrophil/lymphocyte ratio; LMR: lymphocyte/monocyte ratio; PLR: platelet/lymphocyte ratio; SII: systemic immune-inflammation index; AISI: aggregate systemic inflammation index; dNLR: derived neutrophil/lymphocyte ratio; MLR: monocyte/lymphocyte ratio; NLPR: neutrophil/(lymphocyte × platelet) ratio; SIRI: systemic inflammation response index; ALB: albumin; PNI: prognostic nutritional index; ALI: advanced lung cancer inflammation index; CD45+T/CD3+T: CD45+/CD3+ T-cell ratio; CD3+T/CD4+Th: CD3+/CD4+ T-helper ratio; CD3+T/CD8+Ts: CD3+/CD8+ T suppressor ratio; CD4+Th/CD8+T: CD4+/CD8+ T-cell ratio; NK: natural killer cells; CIK: cytokine-induced killer cells; IL-10: interleukin-10; IL-6: interleukin-6; IL-4: interleukin-4; IL-2: interleukin-2. Univariate and multivariate Cox regression analysis of prognostic factors for PFS. HR: hazard ratio; CI: confidence interval; KPS: Karnofsky Performance Status; BMI: body mass index; TNM: tumor–node–metastasis; WBC: white blood cells; LYM: lymphocytes; MNO: monocytes; NEUT: neutrophils; PLT: platelets; NWR: neutrophil/WBC ratio; NLR: neutrophil/lymphocyte ratio; LMR: lymphocyte/monocyte ratio; PLR: platelet/lymphocyte ratio; SII: systemic immune-inflammation index; AISI: aggregate systemic inflammation index; dNLR: derived neutrophil/lymphocyte ratio; MLR: monocyte/lymphocyte ratio; NLPR: neutrophil/(lymphocyte × platelet) ratio; SIRI: systemic inflammation response index; ALB: albumin; PNI: prognostic nutritional index; ALI: advanced lung cancer inflammation index; CD45+T/CD3+T: CD45+/CD3+ T-cell ratio; CD3+T/CD4+Th: CD3+/CD4+ T-helper ratio; CD3+T/CD8+Ts: CD3+/CD8+ T suppressor ratio; CD4+Th/CD8+T: CD4+/CD8+ T-cell ratio; NK: natural killer cells; CIK: cytokine-induced killer cells; IL-10: interleukin-10; IL-6: interleukin-6; IL-4: interleukin-4; IL-2: interleukin-2. Prognostic nomograms for distant metastasis, OS, and PFS Nomograms for predicting the probability of distant metastasis, OS, and PFS in CRC patients are presented in Figures 3(a), 4(a), and 5(a). These models were constructed based on the independent risk factors identified in the multivariate analyses. Among the variables included, the presence of regional lymph node metastasis at initial diagnosis, LYM ≥1.42 (×109/L), CD4+/CD8+ T-cell ratio ≥1.36, and NK cell percentage ≥13.55% were assigned higher scores for predicting distant metastasis. Furthermore, TNM stage ≥ IV, PLT ≥224.5, AISI <474.39, SII ≥435.93, and dNLR ≥0.62 were associated with poorer OS and PFS. The point allocation for each independent risk factor is detailed in Tables 6, 7, and 8. For example, in the metastasis nomogram, regional lymph node metastasis at diagnosis was assigned 100 points, while LYM, CD4+/CD8+ T-cell ratio, and NK cell percentage were assigned 29, 26, and 33 points, respectively, yielding a total possible score ranging from 0 to 188. Similarly, in the OS and PFS nomograms, TNM stage contributed 100 points in both models. PLT accounted for 47 points (OS) and 30 points (PFS), while AISI, SII, and dNLR contributed 33, 30, and 26 points, respectively. The total scores ranged from 0 to 180 for OS and 0 to 186 for PFS. Based on optimal cutoff values, patients were subsequently stratified into low- and high-risk groups. Development and performance of the nomogram for predicting distant metastasis. (a) Nomogram for predicting the probability. (b) ROC curves of the nomogram. (c) Calibration plots of the nomogram. (d) DCA of the nomogram. (e) NIR of the nomogram. ROC: receiver operating characteristic; DCA: decision curve analysis; NIR: net improvement rate. Development and performance of the nomogram for predicting OS. (a) Nomogram for predicting the probability. (b) ROC curves of the nomogram. (c) Time-AUC and C-index of the nomogram. (d) Calibration plots of the nomogram. (e) DCA of the nomogram. AISI acts as a protective factor for OS, consistent with the regression analysis. OS: overall survival; ROC: receiver operating characteristic; AUC: area under the curve; C-index: concordance index; DCA: decision curve analysis; AISI: aggregate index of systemic inflammation. Development and performance of the nomogram for predicting PFS. (a) Nomogram for predicting the probability. (b) ROC curves of the nomogram. (c) Time-AUC and C-index of the nomogram. (d) Calibration plots of the nomogram. (e) DCA of the nomogram. AISI acts as a protective factor for OS, consistent with the regression analysis. PFS: progression-free survival; ROC: receiver operating characteristic; AUC: area under the curve; C-index: concordance index; DCA: decision curve analysis; AISI: aggregate index of systemic inflammation. Table 6.Point assignment and prognostic score of the nomogram for distant metastasis.VariablesPrognostic score (points)NumberInitial lymph node metastasis Yes100254 No0178LYM (×109 L) <1.420105 ≥1.4229327CD4+Th/CD8+T (%) <1.360165 ≥1.3626267NK cell (%) <13.550122 ≥13.5533310Group Low (<113.17 points)198 High (>113.17 points)0–188234LYM: lymphocyte; CD: cluster of differentiation; CD4+Th/CD8+T: CD4+ T-helper cell-to-CD8+ T-cell ratio; NK: natural killer. Point assignment and prognostic score of the nomogram for distant metastasis. LYM: lymphocyte; CD: cluster of differentiation; CD4+Th/CD8+T: CD4+ T-helper cell-to-CD8+ T-cell ratio; NK: natural killer. Table 7.Point assignment and prognostic score of the nomogram for OS.VariablesPrognostic score (points)NumberTNM stage