A LASSO-based survival prediction model for patients with synchronous colorectal carcinomas based on SEER.

BACKGROUND

The nomogram for postoperative prediction of overall survival (OS) in patients' synchronous colorectal carcinomas (SCC) was developed and validated by least absolute shrinkage and selection operator (LASSO)-based Cox regression.

METHODS

The data was obtained from the SEER database of patients diagnosed with colorectal cancer (CRC) more than one time between 2004 and 2013. Patients who had CRC more than 3 times or multiple metachronous primary carcinomas were excluded. The cut-off points for the continuous variable were identified by the K-adaptive partitioning algorithm and x-tile software. Using LASSO-based Cox regression, a model for predicting the OS of SCC was built, internally and externally validated, and measured through a calibration curve, C-index, Akaike information criterion (AIC), Bayesian information criterion (BIC), net reclassification improvement (NRI), integrated discrimination improvement (IDI), time-dependent receiver operating characteristic (timeROC), time-dependent area under curve (timeAUC), and decision curve analysis (DCA), and results compared to the model developed by the Cox regression.

RESULTS

Patients with SCC were found to be older, more often men, and likely to have a depth of invasion by T3. In addition, there were no significant differences between the model developed by LASSO-based Cox regression and the Cox regression in the C-index (0.712 and 0.710), AIC (33,420 and 33,431), BIC (4.49), IDI (0.002), NRI (-0.009), timeROC, and DCA. Besides, the model developed by LASSO-based Cox regression was found to perform better than the Cox regression in the timeAUC. Moreover, the model developed by LASSO-based Cox regression showed good C-index (0.712, 0.637, and 0.651), AIC (33,420, 34,043, and 33,994), BIC (1,178.76 and 1,098.57), IDI (-0.072 and -0.064), NRI (0.525 and 0.466), timeROC, timeAUC and had a larger net benefit compared to both the first time TNM staging and the combination of two times TNM staging.

CONCLUSIONS

This present study indicates that a close follow-up of older patients, male, and T3 should be made. Compared with the traditional Cox regression model, LASSO-based Cox regression decreases the variables of the model, avoids overfitting and collinearity and has clinical significance.