Development and Validation of Nomograms to Predict Survival in Patients Undergoing Complete Cytoreduction and Hyperthermic Intraperitoneal Chemotherapy for Pseudomyxoma Peritonei of Appendiceal Origin.

IMPORTANCE

Pseudomyxoma peritoni, a rare condition characterized by mucinous ascites and peritoneal deposits, mainly originates from a ruptured mucinous appendix tumor and is considered an indolent disease but can progress and become fatal. Optimal treatment to improve cure and survival rates involves complete cytoreductive surgery (CCRS) and hyperthermic intraperitoneal chemotherapy (HIPEC). Accurate predictive models are useful in supporting and informing treatment strategies and stratifying patient follow-up.

OBJECTIVE

To evaluate the prognostic significance of clinically important variables and generate validated nomograms to predict overall (OS) and disease-free survival (DFS) following CCRS and hyperthermic intraperitoneal HIPEC for pseudomyxoma peritonei (PMP) of appendiceal origin.

DESIGN, SETTING, AND PARTICIPANTS: This retrospective study used prospectively collected data on patients who had cytoreductive surgery (CRS) and HIPEC in a single institution between 1994 and 2018. The cohort was randomly allocated into development (70%) and validation (30%) sets. Univariate and multivariate analyses were performed with Cox proportional hazards regression.

MAIN OUTCOMES AND MEASURES

A prediction model was developed with significant prognostic factors identified by multivariate analysis. The model's prognostic performance was evaluated with the concordance index (C index). The nomogram was calibrated by comparing the predicted and observed probabilities.

RESULTS

Of 2637 CRS and HIPEC operations, 1102 patients (female, 64.4%; median age [IQR], 57.0 [48.0-66.0] years) (41.8%) had CCRS for PMP of appendiceal origin. Elevated tumor markers, peritoneal carcinomatosis index, gastrectomy, and tumor grade were independent predictive factors for DFS. Gender, age, elevated tumor makers, peritoneal carcinomatosis index, and tumor grade influenced OS. The nomograms were generated with respective prognostic factors. The nomograms showed good performance in predicting survival. Median OS of the cohort was 16.5 years (95% CI, 13.7-19.2) with a 5-year probability of survival of 80.2%. The median DFS was 10.3 years (95% CI, 7.2- 13.3) and the 5-year probability of recurrence-free survival was 60.5%.

CONCLUSIONS AND RELEVANCE

Clinically important independent predictors for survival and recurrence were selected to develop the nomograms for OS and DFS. These 2 nomograms are user friendly and useful tools for patient management with clinical trial design applications.