CHAPTER 2 30 METHODS Patients The samples of patients were retrospectively collected within the European Network for Individualized Treatment of Endometrial Cancer (ENITEC) from a previous study including 1199 EC patients.15 Patients were only included when they were diagnosed by an expert gynecological pathologist of the participating hospitals, with complete data on treatment and histopathology. Clinical and pathological data were recorded from the patient files into a database; including patient age, date of diagnosis, preoperative sampling method, surgical treatment, original pre- and postoperative tumor grade and histological subtype, myometrial invasion (MI), cervical invasion (CI), lymphovascular space invasion (LVSI), FIGO (International Federation of Gynecology and Obstetrics) stage, adjuvant treatment, recurrent disease and death.15 The sole additional inclusion criterion used for this study was the availability of preoperative EC tissue samples, resulting in 598 patients. Tumor classification In addition to the FIGO three-tiered tumor grade, EC tissue samples were classified into low- and high-grade EC as recommended by the recent ESGO-ESTRO-ESP guideline and the World health organization (WHO) classification of tumors.9, 18 Low-grade EC was defined as grade 1 and 2 EEC, and included samples with mucinous histology as well, since prognosis and molecular characterization are similar to low-grade EECs.15 High-grade EC included grade 3 EEC and NEEC, i.e. serous, clear cell carcinoma, carcinosarcoma and mixed carcinomas.9, 18 Endometrial tissue samples were defined as upgraded if the preoperative sample was lowgrade and postoperative high-grade EC. Downgraded was defined as preoperative high-grade and postoperative low-grade EC. Biopsies initially diagnosed as premalignant, but EC on final hysterectomy specimen were included in this study. Scoring All the preoperative endometrial sampling slides were digitalized using Pannoramic Scanner 250 Flash III (3DHISTECH, Budapest, Hungary). As described previously by Reijnen et al., images were saved as a JPEG-compressed file and the area of endometrial tissue was digitally calculated using ImageJ software, selecting only benign, premalignant and malignant endometrial epithelium (Supplementary Figure S1).14 Thresholds 24-bit RGB images based on Hue Saturation and Brightness (HSB) were used to select the endometrial tissue surface, by adjusting the different threshold values to segment the image into the area of interest and the background. The Pannoramic Viewer software was used to examine the original-size digital slide in order to ensure ImageJ correctly selected the proper tissue. Subsequently, analysis was performed on the area selection to count and measure pixels in the threshold images and calculate the total area of endometrial tissue. A set of 50 slides were scored
RkJQdWJsaXNoZXIy MTk4NDMw