Advanced lung cancer inflammation index as a new predictor for colon cancer in elderly patients: an NHANES-based study

Abstract

Background: Nutritional and inflammatory status have both been implicated in colon cancer risk. The advanced lung cancer inflammation index (ALI) is a composite prognostic index that incorporates body mass index (BMI), serum albumin, and neutrophil-to-lymphocyte ratio (NLR), reflecting both nutritional and systemic inflammatory states. However, its role in predicting colon cancer prevalence in elderly individuals remains unclear.

Methods: We used the ALI as a composite marker reflecting both inflammation status and nutritional health. The ALI is calculated as BMI × serum albumin/NLR, where higher values indicate better nutritional status and lower systemic inflammation. To evaluate the association between ALI and colon cancer prevalence, we conducted multivariate logistic regression, applied an Extreme Gradient Boosting (XGBoost) machine learning model, and performed subgroup analyses. Additionally, a smoothed two-piece logistic regression model was used to identify the ALI threshold predictive of colon cancer.

Results: The study included 10,137 elderly participants, with a colon cancer prevalence of 2.45%. The ALI was significantly lower in the colon cancer group compared to those without (p < 0.001). Multivariable logistic regression revealed a significant inverse association between ALI and colon cancer (p < 0.05), with individuals in the highest ALI tertile experiencing a 67% lower risk compared to those in the lowest tertile (p for trend = 0.008). Generalized additive models showed a linear relationship, identifying an inflection point at 4.73 and a predictive threshold of 113.3. Sensitivity analyses confirmed the robustness of these findings, particularly among individuals aged over 70 years, females, unmarried individuals, alcohol consumers, and those with a BMI below 30. In the XGBoost model, ALI demonstrated the highest predictive value for colon cancer (AUC = 0.717), outperforming traditional demographic and clinical variables such as age and BMI. Furthermore, ALI showed a positive association with dietary health status (p < 0.05) but was not significantly related to bowel habits.

Conclusion: This study demonstrated that ALI, a nutritional-inflammation prognostic index, is significantly and inversely associated with the risk of colon cancer in older adults. Thus, ALI may serve as a promising, non-invasive biomarker for risk stratification, particularly among high-risk subgroups such as unmarried females, alcohol consumers, and those with lower BMI. Its strong predictive value, confirmed by machine learning, supports its potential role in personalized prevention. Further studies are required to explore underlying mechanisms, including dietary and microbial factors.

Keywords: NHANES; colon cancer; inflammation; inflammation index of advanced cancer; nutrition.

Figure 1

Stratified association between log₁₀-transformed ALI and colon cancer across subgroups. Multivariable logistic regression analyses were conducted in various subgroups to assess the association between ALI and colon cancer. Estimates were adjusted for all covariates listed in the Methods section unless otherwise specified. Interaction p-values are provided for subgroup comparisons.

Figure 2

Top predictors of colon cancer identified by the XGBoost model, ranked by AUC. To compare the predictive ability of ALI and other variables, we used the XGBoost machine learning model with 10-fold cross-validation. The AUC values of the top variables are displayed. ALI ranked as the strongest predictor, followed by age and BMI.

Figure 3

SHAP summary plot demonstrating the contribution of each variable to the XGBoost model. SHAP values were used to interpret the contribution of each feature in the model. The plot illustrates both the magnitude and direction (positive or negative) of each variable’s impact on the predicted colon cancer risk.