CA-125 AUC as a predictor for epithelial ovarian cancer relapse

Abstract

Purpose:

The aim of the present work was to evaluate the usefulness of CA-125 normalized in time area under the curve (CA-125 AUC) to signalise epithelial ovarian cancer relapse.

Patients and Methods:

Data from a hundred and eleven patients were submitted to two different approaches based on CA-125 AUC increase values to predict patient relapse. In Criterion A total CA-125 AUC normalized in time value (AUC ${}_{\mathit{i}}$ ) was compared with the immediately previous one (AUC ${}_{\mathit{i}-1}$ ) using the formulae ${\mathrm{AUC}}_{\mathit{i}}\ge \mathit{F}\ast {\mathrm{AUC}}_{\mathit{i}-1}$ (several F values were tested) to find the appropriate close related increment associated to patient relapse. In Criterion B total CA-125 AUC normalised in time was calculated and several cut-off values were correlated with patient relapse prediction capacity.

Results:

In Criterion A the best accuracy was achieved with a factor (F) of 1.25 (increment of 25% from the previous status), while in Criterion B the best accuracies were achieved with cut-offs of 25, 50, 75 and 100 IU/mL. The mean lead time to relapse achieved with Criterion A was 181 days, while with Criterion B they were, respectively, 131, 111, 63 and 11 days.

Conclusion:

Based on our results we believe that conjugation and sequential application of both criteria in patient relapse detection should be highly advisable. CA-125 AUC rapid burst in asymptomatic patients should be firstly evaluated using Criterion A with a high accuracy (0.85) and with a substantial mean lead time to relapse (181 days). If a negative answer was obtained then Criterion B should performed to confirm the absence of relapse.