Prediction models for overall and cancer-specific survival in patients with metastatic early-onset colorectal cancer: a population-based study

Optimizing prediction of metastasis among colorectal cancer patients using machine learning technology

BACKGROUND AND AIM

Colorectal cancer is among the most prevalent and deadliest cancers. Early prediction of metastasis in patients with colorectal cancer is crucial in preventing it from the advanced stages and enhancing the prognosis among these patients. So far, previous studies have been conducted to predict metastasis in colorectal cancer patients using clinical data. The current research attempts to leverage a combination of demographic, lifestyle, nutritional, and clinical factors, such as diagnostic and therapeutical factors, to construct an ML model with more predictive insights and generalizability than previous ones.

MATERIALS AND METHODS

In this retrospective study, we used 1156 CRC patients referred to the Masoud internal clinic in Tehran City from January 2017 to December 2023. The chosen machine learning algorithms, including LightGBM, XG-Boost, random forest, artificial neural network, support vector machine, decision tree, K-Nearest Neighbor and logistic regression, were utilized to establish prediction models for predicting metastasis among colorectal cancer patients. We also assessed features based on the best-performing model to improve clinical usability. To show the generalizability of the established prediction model for predicting CRC metastasis, we leveraged the data of 115 CRC patients from Imam Khomeini Hospital in Sari City. We assessed the predictive ability of LightGBM as the best-performing model based on external data.

RESULTS

The LightGBM model with a PPV of 97.32%, NPV of 84.67%, sensitivity of 83.14%, specificity of 93.14%, accuracy of 88.14%, F1-score of 87.51%, and an AU-ROC of 0.9 [Formula: see text]0.01 obtained satisfactory performance for prediction purposes on this topic. Factors including the history of IBD, family history of CRC, number of lymph nodes involved, fruit intake, and tumor size were considered as more strengthful predictors for metastasis in colorectal cancer and clinical usability. The external validation cohort showed a PPV of 0.8, NPV of 0.85, sensitivity of 0.78, specificity of 0.86, accuracy of 0.834, F1-score of 0.795, and AU-ROC of 0.77[Formula: see text]0.03, demonstrating satisfactory generalizability when leveraging external data from other clinical settings.

CONCLUSION

The current empirical results indicated that LighGBM has predictive competency that can be leveraged by physicians in clinical environments for early prediction of metastasis and enhanced prognosis in patients with colorectal cancer.

CLINICAL TRIAL NUMBER

Not applicable.

Construction and validation of a nomogram for predicting overall survival of patients with stage III/IV early-onset colorectal cancer

Purpose

This study aimed to identify prognostic factors and develop a nomogram for predicting overall survival (OS) in stage III/IV early-onset colorectal cancer (EO-CRC).

Methods

Stage III/IV EO-CRC patients were identified from the Surveillance, Epidemiology, and End Results (SEER) database between 2010 and 2015. The datasets were randomly divided (2:1) into training and validation sets. A nomogram predicting OS was developed based on the prognostic factors identified by Cox regression analysis in the training cohort. Moreover, the predictive performance of the nomogram was assessed using the receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA). Subsequently, the internal validation was performed using the validation cohort. Finally, a risk stratification system was established based on the constructed nomogram.

Results

Of the 10,387 patients diagnosed with stage III/IV EO-CRC between 2010 and 2015 in the SEER database, 8,130 patients were included. In the training cohort (n=3,071), sex, marital status, race/ethnicity, primary site, histologic subtypes, grade, T stage, and N stage were identified as independent prognostic variables for OS. The 1-, 3-, and 5-year area under the curve (AUC) values of the nomogram were robust in both the training (0.751, 0.739, and 0.723) and validation cohorts (0.748, 0.733, and 0.720). ROC, calibration plots, and DCA indicated good predictive performance of the nomogram in both the training and validation sets. Furthermore, patients were categorized into low-, middle-, and high-risk groups based on the nomogram risk score. Kaplan-Meier curve showed significant survival differences between the three groups.

Conclusion

We developed a prognostic nomogram and risk stratification system for stage III/IV EO-CRC, which may facilitate clinical decision-making and individual prognosis prediction.

Post-operative C-reactive protein as a strong independent predictor of long-term colorectal cancer outcomes: consistent findings from two large patient cohorts

BACKGROUND

Post-surgery blood-based biomarkers may be useful for guiding treatment and surveillance decisions among colorectal cancer (CRC) patients. However, most candidate biomarkers provide little if any predictive value beyond stage at diagnosis. We aimed to investigate the independent prognostic value of post-operative serum C-reactive protein (CRP), a highly sensitive biomarker of inflammation, for long-term CRC outcomes in two large patient cohorts.

MATERIALS AND METHODS

CRP levels were measured from serum samples of CRC patients collected ≥1 month post-surgery in the German DACHS (n = 1416) and the UK Biobank (n = 1149) cohorts. Associations of post-operative CRP with overall survival (OS) and CRC-specific survival (CSS) were assessed using Cox regression and presented as hazard ratios (HRs) with 95% confidence intervals (CIs), adjusted for key sociodemographic and clinical covariates.

RESULTS

In both cohorts, consistent strong dose-response relationships between post-operative CRP and both OS and CSS were observed. Adjusted HRs (95% CI) for CRP >10 versus <3 mg/l were 1.93 (1.58-2.35) and 2.70 (2.03-3.59) in the DACHS cohort, and 2.70 (1.96-3.71) and 2.61 (1.83-3.72) in the UK Biobank cohort, respectively. Associations between post-operative CRP and OS were particularly strong among younger patients (<65 years at diagnosis; P value for interaction by age <0.01).

CONCLUSIONS

Serum CRP determined a month or more after surgery may be useful as a strong independent prognostic biomarker for guiding therapeutic decisions and for surveillance of the course of disease of CRC patients, particularly those <65 years of age at diagnosis.