Objective To analyze the early prediction value of serum miR-155 and miR-24 expression levels for cer-vical cancer residual cancer after concurrent chemoradiotherapy(CCRT).Methods A total of 300 cervical cancer pa-tients treated with CCRT were retrospectively selected as the study objects and divided into the modeling group(210 cas-es)and the validation group(90 cases)according to the 7∶3 allocation ratio.Patients in the modeling group were divid-ed into residual group(81 cases)and non-residual group(129 cases)according to postoperative magnetic resonance im-aging(MRI)results.Clinical baseline data and laboratory test indicators before CCRT of patients in the modeling group were collected,independent influencing factors were screened to construct a prediction model of early tumor residual no-mogram after CCRT of cervical cancer,and the validation and value analysis of the prediction model were completed through data collection in the verification group.Results The serum miR-155,parametrial infiltration,and FIGO stag-ing are independent risk factors affecting early tumor residue after CCRT,while serum miR-24 is an independent protec-tive factor affecting early tumor residue after CCRT(OR=1.77,3.22,8.55,0.18,P＜0.05).The early tumor residue predic-tion model for cervical cancer after CCRT based on these independent influencing factors was constructed.The AUC of the ROC curve for the modeling group is 0.84(95%CI 0.79-0.89),with good discriminability.Through external validation of the validation group data,the ROC curve AUC of the validation group was 0.90(95%CI 0.81-0.98).The internal and external calibration curves have a good fit with the standard curve.The internal and exter-nal decision curves indicate that the model can pro-vide significant clinical net benefits.Conclusion Se-rum miR-155 and miR-24 expression levels have good predictive value for early tumor residue after CCRT of cervical cancer.The construction of a nomogram prediction model combined with clinical characteristics such as parametrial infiltration and FIGO staging can provide data support for effective screening of high-risk patients with early tumor residue.
维普
万方数据