Can the preoperative Ca-125 level predict optimal cytoreduction in patients with advanced ovarian carcinoma? A single institution cohort study

2009 
Abstract Background Preoperative Ca-125 level has been used as a predictor of optimal cytoreduction in advanced ovarian carcinoma. Yet, controversy exists regarding the ability of the tumor marker to predict optimal debulking and moreover of the proper cut-off limit to do so. Methods The preoperative Ca-125 levels of 426 patients with Stage III/IV ovarian carcinoma from a single institution were correlated with surgical outcome. Optimal was considered the cytoreduction if the largest residual tumor was ≤1 cm in diameter. Receiver operation characteristic (ROC) curve data were combined with interval likelihood ratios at various Ca-125 levels to determine the cut-off level with the maximum prognostic power. Sensitivity, specificity, positive and negative predictive values and accuracy were also calculated. Results Preoperative Ca-125 proved to be a reliable predictor for optimal cytoreduction. The area under curve of the ROC curve was 0.89, 95% C.I.=[0.828–0.952], indicating very good discriminating capability. The level of 500 IU/ml was found to have the most predictive power. The sensitivity of Ca-125 at that level was 78.5%, the specificity 89.6%, the positive predictive value 84.2%, the negative predictive value 85.4% and its accuracy 85%. Furthermore, the likelihood ratio for correct discrimination between optimal and sub-optimal cytoreduction, dropped sharply from 6.33, 95% C.I. [5.19–10.91] at the level of 500 IU/ml to 0.58, 95% C.I. [0.21–1.63] at the level of 600 IU/ml. Conclusions Our data indicate that preoperative Ca-125 is a good predictor for optimal cytoreduction. The best threshold for this prediction proved to be 500 IU/ml. These patients may be candidates for neo-adjuvant chemotherapy treatment. Nevertheless, all clinical and radiological findings must be co-evaluated.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    30
    References
    50
    Citations
    NaN
    KQI
    []