Introduction/Background Understanding biological characteristics of endometrial cancer (EC) has opened possibilities of treatment individualisation. Enabling non-invasive methods of evaluation in patients with EC can therefore aid decision making in the office setting. Herein, we present the feasibility study evaluating endometrial cytological sampling and mutational analysis of catenin beta-1 (CTNNB1) gene to aid integrated molecular classification of tumours prior to treatment.

Methodology Women were recruited at the University Medical Centre Maribor between November 2020 to May 2022. Prior to surgical treatment for benign disease or EC, endometrial cytological sample was obtained using Endobrush (Lab CCD, Paris, France) and stored in DNA/RNA Shield^TM. Tumour biopsies were stored following routine pathologic examination. DNA was extracted from tumours and cytological samples using QIAamp DNA Mini Kit and Qiuck-DNA/RNA MinPrep Plus Kit, respectively. Sanger sequencing was used to detect mutations in the exon 3 of CTNNB1. Cytological samples were compared to tumour tissue. Continuous variables were expressed as median, and proportions indicated as percentages. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for CTNNB1 mutational status determination.

Results Patient characteristics are presented in table 1. Among 24 women included in the study, 2 patients (8%) were identified having CTNNB1-mutated tumours. CTNNB1 mutational status was not confirmed in cytological samples. The current approach to tissue sampling resulted in 50% sensitivity and 100% sampling specificity. The positive predictive value was 100% and the negative predictive value 94.7%. The test diagnostic accuracy is currently 92.3%. Cytology DNA isolation failure was present in one women with FIGO IA disease and in a control sample.

Conclusion DNA isolation from endometrial cytology samples was successful in 91% of samples and isolation of CTNNB1 mutations showed an appropriate level of specificity, but optimisation of sensitivity is needed for clinical use implementation.