Abstract:

Objective To construct and verify a forecast nomogram for 1- and 2-year survival rate of patients with bone metastasis from rectal cancer. Methods Through the SEER*Stat software, the clinical data and follow-up data of 671 patients who had diagnosis with bone metastasis in SEER database between 1973 and 2017 were included in the modeling group, and the clinical data of 108 patients with rectal cancer bone metastasis in our hospital were included as the verification group. Independent forecast factors affecting the prognostic patients with bone metastasis from rectal cancer were obtained through Multivariate Lasso regression analysis, and constructed to predict the 1- and 2-year nomogram of survival rate of patients with bone metastasis from rectal cancer. The forecast model was evaluated through C-index, clinical decision curve (DCA) and the calibration curve of nomogram prognostic model was applied to evaluate the consistency, and decision tree was drawn to help clinical diagnosis and treatment by the data of the verification group. Results Seven factors including the age, histological grade, M staging, surgery, lymph node metastasis, radiotherapy and chemotherapy were independent forecast factors for prognostic patients with bone metastasis from rectal cancer by univariate and multivariate analysis and these factors were successfully included construct the nomogram. The C-index value of the nomogram for internal validation in the modeling group was 0.75, and the C index obtained by the verification group was 0.74. Conclusion The forecast nomogram for the survival rate of prognostic patients with bone metastasis from rectal cancer constructed in this study has forecast accuracy and strong clinical applicability.

Key words: rectal tumor, neoplasm metastasis, prognosis