University of Twente Student Theses

Login

Clinicopathologic factors predictive of platinum resistance in advanced stage epithelial ovarian cancer : development and validation of a prediction model using a nationwide

Said, S.A. (2019) Clinicopathologic factors predictive of platinum resistance in advanced stage epithelial ovarian cancer : development and validation of a prediction model using a nationwide.

[img] PDF
414kB
Abstract:Objective: To identify clinicopathologic factors predictive of platinum resistance in advanced stage epithelial ovarian cancer (EOC) and to develop and internally validate a risk prediction model for platinum resistance after first-line of treatment in advanced stage EOC. Methods: In this retrospective population-based study, all consecutive patients diagnosed with advanced stage EOC between 1 st January 2008 and 31st December 2015 were identified from the Netherlands Cancer Registry. Patients who underwent cytoreductive surgery combined with platinum-based chemotherapy as initial EOC treatment were selected. Two risk prediction models, a pretreatment and a postoperative model, were developed and validated. Candidate predictors of platinum resistance were based on expert opinion and previous studies. These candidate predictors were fitted into multivariable logistic regression models. Models’ selection was performed using a backward selection procedure based on the likelihood ratio test. Models’ discrimination was estimated with the area under the receiver operating characteristic curve (AUC) and models’ calibration was assessed using calibration plots and the Hosmer-Lemeshow goodness of fit test. Internal validation of the final models was performed using a bootstrap resampling method, which provided an estimate of model optimism and a shrinkage factor for each model. Results: A total of 4,557 advanced stage EOC patients were identified, including 3,196 platinum sensitive patients and 1,361 platinum resistant patients. Platinum resistant patients were more likely to have FIGO stage IV, mucinous or clear cell type of ovarian cancer, presence of ascites, suboptimal (> 1 cm of residual tumor) residual disease, and more likely to have undergone interval cytoreductive surgery. The final pretreatment prediction model included age at diagnosis, FIGO stage, histologic subtype, presence of ascites, and pretreatment serum levels of CA-125. The AUC of the final pretreatment model was 0.65 [95% CI 0.64 – 0.67]. Calibration plots and Hosmer-Lemeshow test might suggest the pretreatment model was not perfectly calibrated. Bootstrap validation revealed an estimate of 0.003 of optimism in the pretreatment model’s performance and a shrinkage factor of 0.94. The final postoperative prediction model included FIGO stage, histologic subtype, presence of ascites, type of surgery performed, and residual disease after surgical treatment. The AUC of the postoperative model was 0.72 [95% confidence interval 0.70 – 0.73]. Calibration plots and the Hosmer-Lemeshow test did not show evidence for miscalibration of the final postoperative model. Bootstrap validation revealed an estimate of 0.001 of optimism in the postoperative model’s performance and a shrinkage factor of 0.96. Conclusion: A good discriminative clinical model has been developed that predicts the risk of platinum resistance following first-line of treatment in advanced EOC based on FIGO stage, histologic subtype, presence of ascites, type of surgery performed, and residual disease after surgical treatment. Even though external validation is still required, this prediction model can support treatment decision making in daily clinical practice.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Subject:44 medicine, 50 technical science in general
Programme:Health Sciences MSc (66851)
Link to this item:https://purl.utwente.nl/essays/79472
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page