E Joyeux1, T Miras2, I Masquin3, P-E Duglet4, K Astruc5, S Douvier3. 1. Service de gynécologie obstétrique et médecine de la reproduction, CHU de Dijon, 14, rue Paul-Gaffarel, 21079 Dijon, France. Electronic address: joyeuxemi@gmail.com. 2. Service de gynécologie obstétrique, centre hospitalier de Chalon-sur-Saône, 71100 Chalon-sur-Saône, France. 3. Service de gynécologie obstétrique et médecine de la reproduction, CHU de Dijon, 14, rue Paul-Gaffarel, 21079 Dijon, France. 4. Département de médecine générale, faculté des sciences de santé de Dijon, 21000 Dijon, France. 5. Service d'épidémiologie et d'hygiène hospitalière, CHU de Dijon, 21000 Dijon, France.
Abstract
OBJECTIVE: The principal aim of this study was the predictability of malignant ovarian tumors and to determine a cut-off value for this score to indicate the risk of malignancy that would be easy to use in clinical practice. METHODS: We retrospectively calculated the ADNEX score for all patients who underwent surgery for ovarian tumours in two Burgundy hospitals (Dijon University Hospital and Chalon-sur-Saône Hospital). We used the nine criteria of the ADNEX model. The inclusion criteria were the presence of all of the ADNEX criteria and a histology result. We analysed the sensitivity, specificity, PPV and PNV of four cut-offs (3%, 5%, 10% and 15%) for the entire pool then by age groups; from 14 to 42 (group 1) and 43 and more (group 2) RESULTS: Two hundred and eighty-four patients managed for an ovarian tumour were included between the 1st January 2013 and the 31st December 2015. Our AUC was of 0.94 (95% CI [0.903-0.977]) for discrimination between benign and malignant ovarian tumors. For a cut-off of 10%, sensitivity was 90%, specificity was 81.1%, PPV was 34.6% and PNV 98.5%. Results were lower for young women than for the second group. For a cut-off of 10%, group 1 had a sensitivity of 77.7% and specificity of 89.6%, PPV of 46.6% and PNV 97.5%. For the group 2, sensitivity was 95.2%, specificity was 76.6%, PPV was 33.8% and PNV was 99.2%. The most reasonable cut-off for the whole pool was 10%. For group 1 a cut-off of 5% was retained due to the less satisfying detection of "borderline" tumours more frequent in younger patients. For group 2 the cut-off of 10% gave the best results. CONCLUSION: In our study, a lower cut-off for younger women seemed better suited to discriminate borderline tumours. In practice, the ADNEX score associated with the peroperative laparoscopic examination seems to be the best way to use the ADNEX model. Our study showed that the ADNEX model allows a good predictability of malignant ovarian tumours. The predictability becomes less satisfying for the youngest patients. A cut-off malignity value allowing surgical treatment of patients in a specialised facility was reached for two age groups: a cut-off of 5% for women under 42 years old and a cut-off of 10% for women over 43 years old.
OBJECTIVE: The principal aim of this study was the predictability of malignant ovarian tumors and to determine a cut-off value for this score to indicate the risk of malignancy that would be easy to use in clinical practice. METHODS: We retrospectively calculated the ADNEX score for all patients who underwent surgery for ovarian tumours in two Burgundy hospitals (Dijon University Hospital and Chalon-sur-Saône Hospital). We used the nine criteria of the ADNEX model. The inclusion criteria were the presence of all of the ADNEX criteria and a histology result. We analysed the sensitivity, specificity, PPV and PNV of four cut-offs (3%, 5%, 10% and 15%) for the entire pool then by age groups; from 14 to 42 (group 1) and 43 and more (group 2) RESULTS: Two hundred and eighty-four patients managed for an ovarian tumour were included between the 1st January 2013 and the 31st December 2015. Our AUC was of 0.94 (95% CI [0.903-0.977]) for discrimination between benign and malignant ovarian tumors. For a cut-off of 10%, sensitivity was 90%, specificity was 81.1%, PPV was 34.6% and PNV 98.5%. Results were lower for young women than for the second group. For a cut-off of 10%, group 1 had a sensitivity of 77.7% and specificity of 89.6%, PPV of 46.6% and PNV 97.5%. For the group 2, sensitivity was 95.2%, specificity was 76.6%, PPV was 33.8% and PNV was 99.2%. The most reasonable cut-off for the whole pool was 10%. For group 1 a cut-off of 5% was retained due to the less satisfying detection of "borderline" tumours more frequent in younger patients. For group 2 the cut-off of 10% gave the best results. CONCLUSION: In our study, a lower cut-off for younger women seemed better suited to discriminate borderline tumours. In practice, the ADNEX score associated with the peroperative laparoscopic examination seems to be the best way to use the ADNEX model. Our study showed that the ADNEX model allows a good predictability of malignant ovarian tumours. The predictability becomes less satisfying for the youngest patients. A cut-off malignity value allowing surgical treatment of patients in a specialised facility was reached for two age groups: a cut-off of 5% for women under 42 years old and a cut-off of 10% for women over 43 years old.
Authors: Ben Van Calster; Lil Valentin; Wouter Froyman; Chiara Landolfo; Jolien Ceusters; Antonia C Testa; Laure Wynants; Povilas Sladkevicius; Caroline Van Holsbeke; Ekaterini Domali; Robert Fruscio; Elisabeth Epstein; Dorella Franchi; Marek J Kudla; Valentina Chiappa; Juan L Alcazar; Francesco P G Leone; Francesca Buonomo; Maria Elisabetta Coccia; Stefano Guerriero; Nandita Deo; Ligita Jokubkiene; Luca Savelli; Daniela Fischerová; Artur Czekierdowski; Jeroen Kaijser; An Coosemans; Giovanni Scambia; Ignace Vergote; Tom Bourne; Dirk Timmerman Journal: BMJ Date: 2020-07-30