Preoperative prediction of p16 expression in cervical cancer by amide proton transfer imaging combined with diffusion kurtosis imaging: a retrospective study

Abstract

Background Immunohistochemical p16 expression is important for the assessment of risk factors and prognosis of cervical cancer. This study evaluated the correlation between preoperative amide proton transfer imaging (APT), diffusion kurtosis imaging (DKI), and p16 expression in cervical cancer. Methods Fifty-five records of patients with cervical cancer with preoperative magnetic resonance imaging (MRI) and complete postoperative pathology were retrospectively analyzed. Uterine MRI scans were examined, and the corresponding APT values, mean kurtosis (MK), and mean diffusivity (MD) were obtained. Pathology was used to determine p16 positivity. The predictive effects of the APT- and DKI-derived parameters applied alone or in combination with p16 expression were compared. Results Among the 55 cases of cervical cancer, 35 were p16 positive and 20 were p16 negative. The receiver operating characteristic curve area (AUC) of p16 was 0.809, 0.801, 0.790, 0.845, 0.866, and 0.871 when APT, MK, and MD were used individually and then jointly. The effect of APT combined with the DKI derivative parameters was more pronounced than when they were used separately. Conclusion Preoperative APT and DKI imaging are valuable in predicting the expression of p16 in cervical cancer, and the combination of the two can improve the prediction efficacy. These findings could help to guide treatment decisions and determine prognosis in cervical cancer.

Keywords

• Amide proton transfer imaging
• Diffusion kurtosis imaging
• Cervical cancer
• P16