Artur Czekierdowski1, Norbert Stachowicz2, Agata Smoleń2, Tomasz Kluz3, Tomasz Łoziński4, Andrzej Miturski5, Janusz Kraczkowski5. 1. Department of Gynecological Oncology and Gynecology, Medical University of Lublin, 20-081 Lublin, Poland. 2. Chair and Department of Epidemiology and Clinical Research Methodology, Medical University of Lublin, 20-080 Lublin, Poland. 3. Department of Obstetrics and Gynecology, Fryderyk Chopin University Hospital No 1, Faculty of Medicine, Rzeszow University, 35-060 Rzeszow, Poland. 4. Department of Obstetrics and Gynecology, Pro-Familia Hospital, 35-302 Rzeszow, Poland. 5. Department of Obstetrics and Pathology of Pregnancy, Medical University of Lublin, 20-081 Lublin, Poland.
Abstract
BACKGROUND: To evaluate the accuracy of subjective assessment (SA), the International Ovarian Tumor Analysis (IOTA) group Simple Rules Risk (SRR) and the Assessment of Different NEoplasias in the adneXa (ADNEX) model for the preoperative differentiation of adnexal masses in pregnant women. METHODS: The study population comprised 36 pregnant women (median age: 28.5 years old, range: 20-42 years old) with a mean gestation age of 13.5 (range: 8-31) weeks at diagnosis. Tumors were prospectively classified by local sonographers as probably benign or probably malignant using SA. Final tumor histological diagnosis was used as the reference standard in all cases. Logistic regression SRR and ADNEX models were used to obtain a risk score for every case. Serum CA125 and human epidydimis protein 4 (HE4) concentrations were also retrieved and the Risk of Ovarian Malignancy Algorithm (ROMA) value was calculated. The calculated predictive values included positive and negative likelihood ratios of ultrasound and biochemical tests. RESULTS: Final histology confirmed 27 benign and 9 malignant (including 2 borderline) masses. The highest sensitivity (89%) and specificity (70%) were found for the subjective tumor assessment. Although no malignancy was classified as benign using the SRR criteria (sensitivity = 100%), the specificity of this scoring system was only 37%. At the cut-off risk level of >20%, the ADNEX model had a sensitivity of 78% and a specificity of 70%. Serum levels of CA125, HE4 and the ROMA risk model correctly identified adnexal malignant tumors with a sensitivity of 67%, 25% and 25%, respectively. Corresponding specificities were 72%, 100% and 100%, respectively. The highest positive and negative likelihood ratios were found for SA (LR+ = 3.0 and LR- = 0.16, respectively). Overall diagnostic accuracy of all predictive methods used in this study were similar (range: 70-75%) except for SRR (53%). CONCLUSION: Subjective assessment remains the best predictive method in complex adnexal masses found at prenatal ultrasound in pregnant women. For less experienced sonographers, both the SRR and ADNEX scoring systems may be also used for the characterization of such tumors, while serum tumor markers CA125 and HE4, along with the ROMA algorithm appear to be less accurate.
BACKGROUND: To evaluate the accuracy of subjective assessment (SA), the International Ovarian Tumor Analysis (IOTA) group Simple Rules Risk (SRR) and the Assessment of Different NEoplasias in the adneXa (ADNEX) model for the preoperative differentiation of adnexal masses in pregnant women. METHODS: The study population comprised 36 pregnant women (median age: 28.5 years old, range: 20-42 years old) with a mean gestation age of 13.5 (range: 8-31) weeks at diagnosis. Tumors were prospectively classified by local sonographers as probably benign or probably malignant using SA. Final tumor histological diagnosis was used as the reference standard in all cases. Logistic regression SRR and ADNEX models were used to obtain a risk score for every case. Serum CA125 and human epidydimis protein 4 (HE4) concentrations were also retrieved and the Risk of Ovarian Malignancy Algorithm (ROMA) value was calculated. The calculated predictive values included positive and negative likelihood ratios of ultrasound and biochemical tests. RESULTS: Final histology confirmed 27 benign and 9 malignant (including 2 borderline) masses. The highest sensitivity (89%) and specificity (70%) were found for the subjective tumor assessment. Although no malignancy was classified as benign using the SRR criteria (sensitivity = 100%), the specificity of this scoring system was only 37%. At the cut-off risk level of >20%, the ADNEX model had a sensitivity of 78% and a specificity of 70%. Serum levels of CA125, HE4 and the ROMA risk model correctly identified adnexal malignant tumors with a sensitivity of 67%, 25% and 25%, respectively. Corresponding specificities were 72%, 100% and 100%, respectively. The highest positive and negative likelihood ratios were found for SA (LR+ = 3.0 and LR- = 0.16, respectively). Overall diagnostic accuracy of all predictive methods used in this study were similar (range: 70-75%) except for SRR (53%). CONCLUSION: Subjective assessment remains the best predictive method in complex adnexal masses found at prenatal ultrasound in pregnant women. For less experienced sonographers, both the SRR and ADNEX scoring systems may be also used for the characterization of such tumors, while serum tumor markers CA125 and HE4, along with the ROMA algorithm appear to be less accurate.
Authors: Ben Van Calster; Kirsten Van Hoorde; Lil Valentin; Antonia C Testa; Daniela Fischerova; Caroline Van Holsbeke; Luca Savelli; Dorella Franchi; Elisabeth Epstein; Jeroen Kaijser; Vanya Van Belle; Artur Czekierdowski; Stefano Guerriero; Robert Fruscio; Chiara Lanzani; Felice Scala; Tom Bourne; Dirk Timmerman Journal: BMJ Date: 2014-10-15