The presence of adherent perinephric fat (APF) is a significant challenge for surgeons during renal tumor resection. It can restrict renal mobility, hinder the separation of renal masses, and cause complications such as renal capsule injury, iatrogenic vascular injury, fat tissue hemorrhage, and vascular damage, thereby complicating intracavity urological surgical procedures. Additionally, APF may increase the risk of tumor invasion, extend surgery duration, and increase intraoperative blood loss.. However, there is a notable lack of noninvasive predictive models to objectively assess the presence of preoperative APF. Our study aimed to develop and validate a predictive model utilizing computed tomography (CT) radiomics combined with deep learning to accurately identify preoperative APF.
In this multicenter study, data were collected from 417 renal tumor patients at a tertiary hospital. The dataset was randomly divided into a cross-validation group (70%) and an in-house test set (30%). Additionally, prospective data from three other tertiary hospitals, comprising 43 patients, were used as an external test set. We employed a 3D-UNet deep learning model for kidney segmentation, evaluating intra-observer and inter-observer reproducibility using the Dice similarity coefficient. Logistic regression models were constructed to distinguish between APF and non-APF cases.
The study analyzed 460 patients with renal tumors (mean [SD] age, 60.1[8.7] years). The Dice similarity coefficients for both intra-observer and inter-observer assessments exceeded 0.80. Twenty-eight radiomics features and three clinical features were selected for modeling. The optimal model achieved an area under the curve (AUC) of 0.95(95% CI, 0.94-0.97) in the training set, 0.872 (95% CI,0.807-0.937) in the internal test set, and 0.805(95% CI,0.676-0.935) in the external test set.
CT radiomics combined with deep learning shows significant promise for predicting perirenal fat status in patients with renal tumors. Accurate and reliable preoperative prediction of APF enhances the predictability of surgical complexity in renal tumor cases, thereby improving patient selection and surgical outcomes.