Article Text

Download PDFPDF

#864 A nomogram combing mri and serum inflammatory biomarkers predicts postoperative vaginal invasion in IB-IIA stage cervical cancer —— a single institutional retrospective study of 580 patients
  1. Ning Xie,
  2. Jie Lin,
  3. Sufang Deng,
  4. Linying Liu,
  5. Haijuan Yu and
  6. Yang Sun
  1. Department of Gynecology, Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital, Fuzhou, China


Introduction/Background In cervical cancer (CC), pelvic examination has long been considered the standard method for clinical stage classification. However, it may easily misjudge and bias, including the occult vaginal invasion (VI). Insufficient preoperative assessment of VI often leads to vaginal lesions residues or inferior tumor-free distance during the operation. Recently studies showed MRI has the potential to detect occult tumors. At the same time, serum inflammatory biomarkers have been demonstrated to correlate with the tumor migration in various tumors such as lung cancer, esophageal cancer, and gastric cancer. Combining MRI and inflammatory biomarkers is meaningful to predict occult VI in CC patients with surgical procedures.

Methodology Our study was designed one-center and retrospectively. 580 CC patients with FIGO2018 stages IB-IIA2 were enrolled between January 2013 and December 2021. All patients underwent preoperative MRI and radical hysterectomy. The demographic, bimanual examination, MRI, and laboratory data were analyzed based on logistic regression analysis. Then the nomogram was developed to predict the probability occurrence of postoperative VI.

Results All patients were randomly divided into training set (n = 290) and validating set (n = 290). Parameters including MRI-derived vaginal invasion (P < 0.018), clinical vaginal invasion (P < 0.038), systemic inflammatory response index (SIRI) (P < 0.001), and platelet/albumin ratio (PAR) (P < 0.013) were the independent diagnostic factor for postoperative VI by univariate and multivariate logistic regression showed. The nomogram predicts the occurrence of VI had a robust predict performance, which was demonstrated by the AUC (AUCtraining=0.775 and AUCvalidating=0.769) and the calibration curves and DCA.

Conclusion Our study is the first to develop a multifactorial diagnostic model integrated preoperative parameters of MRI, serum SIRI, and PAR level to predict VI in CC patients with early stage. Our nomogram may assist surgeons in determining the length of vaginal resection when performing a procedure.

Disclosures The authors have no financial conflicts of interest.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.