Premium
EP19.03: Interobserver agreement between expert and trainee sonologist in differentiating adnexal mass using ultrasound pattern recognition
Author(s) -
Gepaya M.R.
Publication year - 2018
Publication title -
ultrasound in obstetrics and gynecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.202
H-Index - 141
eISSN - 1469-0705
pISSN - 0960-7692
DOI - 10.1002/uog.20061
Subject(s) - medicine , adnexal mass , kappa , radiology , ultrasound , cohen's kappa , adnexal diseases , laparoscopy , philosophy , linguistics , machine learning , computer science
Objectives: To quantify the interobserver agreement between an expert and trainee in examining adnexal masses on real-time ultrasound using pattern recognition in our institution. Methods: This study utilised cross sectional research design. Forty-six women diagnosed as having ovarian mass were evaluated and examined by transvaginal/ transabdominal ultrasound. Examination was performed first by the trainee and afterwards by an expert sonologist. Both used ultrasound pattern recognition to described the mass as benign and malignant and to identify the specific benign adnexal lesions and verified it with the histopath result. Agreement was assessed by calculating weighted and standard kappa index values with 95% CI and the percentage of agreement between the observers. Results: A total of 46 subjects were included in the study. All underwent surgery with histopathologic diagnosis. 87% (40/6) of the ovarian mass were confirmed to be benign and 13% (6/46) malignant by histopathologic examination. The diagnostic performance between expert and trainee in differentiating adnexal mass as benign or malignant using pattern recognition showed no significant difference with a p value of 1.00. The inter observer agreement between the trainee and expert was very good as shown by the Kappa value of 0.91. The accuracy of diagnostic performance between expert and trainee in differentiating adnexal mass as benign or malignant showed 100% sensitivity for both while specificity was 100% by the expert and 97.5% by the trainee. Conclusions: The results showed that there was no significant difference in the diagnosis of both the trainee and expert compared to histopath in differentiating adnexal mass using pattern recognition as benign and malignant. The interobserver agreement ranged from very good to excellent agreement with histopath results. Confidence in making a specific histological diagnosis using pattern recognition can be done in dermoid cyst, endometrial cyst, mucinous cystadenoma, and serous cystadenoma.