Predictive models for chronic kidney disease after radical or partial nephrectomy in renal cell cancer using early postoperative serum creatinine levels

Abstract

Background: Several predictive factors for chronic kidney disease (CKD) following radical nephrectomy (RN) or partial nephrectomy (PN) have been identified. However, early postoperative laboratory values were infrequently considered as potential predictors. Therefore, this study aimed to develop predictive models for CKD 1 year after RN or PN using early postoperative laboratory values, including serum creatinine (SCr) levels, in addition to preoperative and intraoperative factors. Moreover, the optimal SCr sampling time point for the best prediction of CKD was determined.

Methods: Data were retrospectively collected from patients with renal cell cancer who underwent laparoscopic or robotic RN (n = 557) or PN (n = 999). Preoperative, intraoperative, and postoperative factors, including laboratory values, were incorporated during model development. We developed 8 final models using information collected at different time points (preoperative, postoperative day [POD] 0 to 5, and postoperative 1 month). Lastly, we combined all possible subsets of the developed models to generate 120 meta-models. Furthermore, we built a web application to facilitate the implementation of the model.

Results: The magnitude of postoperative elevation of SCr and history of CKD were the most important predictors for CKD at 1 year, followed by RN (compared to PN) and older age. Among the final models, the model using features of POD 4 showed the best performance for correctly predicting the stages of CKD at 1 year compared to other models (accuracy: 79% of POD 4 model versus 75% of POD 0 model, 76% of POD 1 model, 77% of POD 2 model, 78% of POD 3 model, 76% of POD 5 model, and 73% in postoperative 1 month model). Therefore, POD 4 may be the optimal sampling time point for postoperative SCr. A web application is hosted at https://dongy.shinyapps.io/aki_ckd .

Conclusions: Our predictive model, which incorporated postoperative laboratory values, especially SCr levels, in addition to preoperative and intraoperative factors, effectively predicted the occurrence of CKD 1 year after RN or PN and may be helpful for comprehensive management planning.

Keywords: Chronic kidney disease; Creatinine; Nephrectomy; Predictive factors; Renal cell cancer.

Fig. 1

Schematic diagram of the overall analysis workflow. Candidate features were first grouped into preoperative, intraoperative, and postoperative feature sets. Preoperative and intraoperative features were merged into a set of $F_{\mathit{pre}}$, which were used in all the tested models. Postoperative features were categorized based on the sampling time points (postoperative day [POD] 0, 1, 2, 3, 4, 5, and postoperative 1 month). Each of these feature sets was combined with $F_{\mathit{pre}}$ to yield 7 different feature sets ($F_{0d}$, …, $F_{5d}$, and $F_{1m}$). The 8 feature sets were used to fit 8 different Lasso regression models. Features of each set with non-zero coefficients, $F_{\mathit{Lasso}}$, were then passed onto a partial correlation filter that evaluated the correlation of each of the features with the target variable $\mathrm{\Delta }{\mathit{SCr}}_{1y}$. The final features, $F_{\mathit{final}}$, were then used to train the final multivariate linear regression models. The final step used all possible combinations of the predictions generated by the 7 final models, ${\mathit{Model}}_{0d-1m}$, to yield 120 meta-models

Fig. 2

Goodness-of-fit plots of the 8 final models. The ordinate and abscissa represent the observed and predicted $\mathrm{\Delta }{\mathit{SCr}}_{1y}$, respectively. The red lines indicate the lines of unity

Fig. 3

Screenshot of the developed web application (https://dongy.shinyapps.io/aki_ckd). The minimum required information were patient characteristics and preoperative and intraoperative factors. As postoperative measurements of SCr (and BUN if measured on postoperative days 4 and 5) become available, they can be entered into the newly appearing widget after clicking on the checkboxes under the sampling time point(s) heading. The predictions are updated in real-time with the incremental addition of newly acquired information. In the graph of SCr versus time, the different percentiles of longitudinal trajectories of SCr in our patients are shown as different shades of yellow-colored bands, median value as a red line, and SCr measurements as blue open circles