Objective:

To establish a radiomics model for predicting the lymph-vascular space invasion (LVSI) status of cervical cancer patients using multi-sequence magnetic resonance imaging (MRI).

Methods:

Double sequence MRI data of 151 patients obtained from Liaoning Provincial Cancer Hospital from November 2015 to August 2018 were selected. Features from the two MRI sequences were extracted and selected. A nomogram model for predicting the LVSI status were constructed and evaluated using receiver

operating characteristic (ROC) curve

area under curve (AUC)

sensitivity

specificity and calibration curve. The clinical value of nomogram model was evaluated by a decision curve analysis model.

Results:

The AUC values of established nomogram model were 0.812 (specificity=0.705

sensitivity=0.818) and 0.775 (specificity=0.759

sensitivity=0.765) in the train and test cohorts

respectively. The calibration curve shows that the nomogram has a good prediction ability. The decision curve analysis suggested that model has good clinical value.

Conclusion:

Radiomics model based on double sequence MRI features has good potential inprediction of LVSI in cervical patients.