Most cited article - PubMed ID 34112736
ESGO/ISUOG/IOTA/ESGE Consensus Statement on pre-operative diagnosis of ovarian tumors
Ovarian lesions are common and often incidentally detected. A critical shortage of expert ultrasound examiners has raised concerns of unnecessary interventions and delayed cancer diagnoses. Deep learning has shown promising results in the detection of ovarian cancer in ultrasound images; however, external validation is lacking. In this international multicenter retrospective study, we developed and validated transformer-based neural network models using a comprehensive dataset of 17,119 ultrasound images from 3,652 patients across 20 centers in eight countries. Using a leave-one-center-out cross-validation scheme, for each center in turn, we trained a model using data from the remaining centers. The models demonstrated robust performance across centers, ultrasound systems, histological diagnoses and patient age groups, significantly outperforming both expert and non-expert examiners on all evaluated metrics, namely F1 score, sensitivity, specificity, accuracy, Cohen's kappa, Matthew's correlation coefficient, diagnostic odds ratio and Youden's J statistic. Furthermore, in a retrospective triage simulation, artificial intelligence (AI)-driven diagnostic support reduced referrals to experts by 63% while significantly surpassing the diagnostic performance of the current practice. These results show that transformer-based models exhibit strong generalization and above human expert-level diagnostic accuracy, with the potential to alleviate the shortage of expert ultrasound examiners and improve patient outcomes.
- MeSH
- Deep Learning MeSH
- Adult MeSH
- Middle Aged MeSH
- Humans MeSH
- Ovarian Neoplasms * diagnostic imaging diagnosis pathology MeSH
- Neural Networks, Computer MeSH
- Retrospective Studies MeSH
- Aged MeSH
- Sensitivity and Specificity MeSH
- Ultrasonography methods MeSH
- Artificial Intelligence * MeSH
- Check Tag
- Adult MeSH
- Middle Aged MeSH
- Humans MeSH
- Aged MeSH
- Female MeSH
- Publication type
- Journal Article MeSH
- Multicenter Study MeSH
- Validation Study MeSH
BACKGROUND: Several diagnostic prediction models to help clinicians discriminate between benign and malignant adnexal masses are available. This study is a head-to-head comparison of the performance of the Assessment of Different NEoplasias in the adneXa (ADNEX) model with that of the Risk of Ovarian Malignancy Algorithm (ROMA). METHODS: This is a retrospective study based on prospectively included consecutive women with an adnexal tumour scheduled for surgery at five oncology centres and one non-oncology centre in four countries between 2015 and 2019. The reference standard was histology. Model performance for ADNEX and ROMA was evaluated regarding discrimination, calibration, and clinical utility. RESULTS: The primary analysis included 894 patients, of whom 434 (49%) had a malignant tumour. The area under the receiver operating characteristic curve (AUC) was 0.92 (95% CI 0.88-0.95) for ADNEX with CA125, 0.90 (0.84-0.94) for ADNEX without CA125, and 0.85 (0.80-0.89) for ROMA. ROMA, and to a lesser extent ADNEX, underestimated the risk of malignancy. Clinical utility was highest for ADNEX. ROMA had no clinical utility at decision thresholds <27%. CONCLUSIONS: ADNEX had better ability to discriminate between benign and malignant adnexal tumours and higher clinical utility than ROMA. CLINICAL TRIAL REGISTRATION: clinicaltrials.gov NCT01698632 and NCT02847832.
- MeSH
- Algorithms MeSH
- CA-125 Antigen MeSH
- Humans MeSH
- Ovarian Neoplasms * diagnosis surgery pathology MeSH
- Adnexal Diseases * diagnosis surgery pathology MeSH
- Retrospective Studies MeSH
- Sensitivity and Specificity MeSH
- Ultrasonography MeSH
- Check Tag
- Humans MeSH
- Female MeSH
- Publication type
- Journal Article MeSH
- Multicenter Study MeSH
- Names of Substances
- CA-125 Antigen MeSH
In recent years the role of diagnostic imaging by pelvic ultrasound in the diagnosis and staging of gynecological cancers has been growing exponentially. Evidence from recent prospective multicenter studies has demonstrated high accuracy for pre-operative locoregional ultrasound staging in gynecological cancers. Therefore, in many leading gynecologic oncology units, ultrasound is implemented next to pelvic MRI as the first-line imaging modality for gynecological cancer. The work herein is a consensus statement on the role of pre-operative imaging by ultrasound and other imaging modalities in gynecological cancer, following European Society guidelines.
- Keywords
- cervical cancer, cross-sectional studies, ovarian cancer, uterine cancer, vulvar and vaginal cancer,
- MeSH
- Gynecology * MeSH
- Consensus MeSH
- Humans MeSH
- Genital Neoplasms, Female * diagnostic imaging MeSH
- Pelvis MeSH
- Ultrasonography MeSH
- Check Tag
- Humans MeSH
- Female MeSH
- Publication type
- Journal Article MeSH
- Review MeSH
OBJECTIVE: The aim of this study was to analyze the clinical and reproductive outcomes of patients treated with myomectomy who were histologically diagnosed with uterine smooth muscle tumor of uncertain malignant potential (STUMP). METHODS: Patients who were diagnosed with STUMP and underwent a myomectomy at our institution between October 2003 and October 2019 were identified. Variables of interest obtained from the institution's database included patient age, relevant medical history, pre-operative appearance of the tumor on ultrasound, parameters of the surgical procedure, histopathological analysis of the tumor, post-operative clinical course, and course of follow-up, including reinterventions and fertility outcomes. RESULTS: There were a total of 46 patients that fulfilled the criteria of STUMP. The median patient age was 36 years (range, 18-48 years) and the mean follow-up was 47.6 months (range, 7-149 months). Thirty-four patients underwent primary laparoscopic procedures. Power morcellation was used for specimen extraction in 19 cases (55.9% of laparoscopic procedures). Endobag retrieval was used in nine patients and six procedures were converted to an open approach due to the suspicious peri-operative appearance of the tumor. Five patients underwent elective laparotomy due to the size and/or number of tumors; three patients had vaginal myomectomy; two patients had the tumor removed during planned cesarean section; and two underwent hysteroscopic resection.There were 13 reinterventions (five myomectomies and eight hysterectomies) with benign histology in 11 cases and STUMP histology in two cases (4.3% of all patients). We did not observe any recurrence as leiomyosarcoma or other uterine malignancy. We did not observe any deaths related to the diagnosis. Twenty-two pregnancies were recorded among 17 women, which resulted in 18 uncomplicated deliveries (17 by cesarean section and one vaginal), two missed abortions, and two pregnancy terminations. CONCLUSIONS: Our study found that uterus-saving procedures and fertility-preservation strategies in women with STUMP are feasible, safe, and seem to be associated with a low risk of malignant recurrence, even while maintaining the mini-invasive laparoscopic approach.
- Keywords
- Gynecologic Surgical Procedures, Pathology, Uterine Neoplasms,
- MeSH
- Cesarean Section MeSH
- Fertility MeSH
- Infant MeSH
- Laparoscopy * methods MeSH
- Humans MeSH
- Uterine Myomectomy * adverse effects MeSH
- Smooth Muscle Tumor * pathology MeSH
- Uterine Neoplasms * pathology MeSH
- Child, Preschool MeSH
- Retrospective Studies MeSH
- Pregnancy MeSH
- Uterus pathology MeSH
- Check Tag
- Infant MeSH
- Humans MeSH
- Child, Preschool MeSH
- Pregnancy MeSH
- Female MeSH
- Publication type
- Journal Article MeSH
- Research Support, Non-U.S. Gov't MeSH
Maximal-effort upfront or interval debulking surgery is the recommended approach for advanced-stage ovarian cancer. The role of diagnostic imaging is to provide a systematic and structured report on tumour dissemination with emphasis on key sites for resectability. Imaging methods, such as pelvic and abdominal ultrasound, contrast-enhanced computed tomography, whole-body diffusion-weighted magnetic resonance imaging and positron emission tomography, yield high diagnostic performance for diagnosing bulky disease, but they are less accurate for depicting small-volume carcinomatosis, which may lead to unnecessary explorative laparotomies. Diagnostic laparoscopy, on the other hand, may directly visualize intraperitoneal involvement but has limitations in detecting tumours beyond the gastrosplenic ligament, in the lesser sac, mesenteric root or in the retroperitoneum. Laparoscopy has its place in combination with imaging in cases where ima-ging results regarding resectability are unclear. Different imaging models predicting tumour resectability have been developed as an adjunctional objective tool. Incorporating results from tumour quantitative analyses (e.g., radiomics), preoperative biopsies and biomarkers into predictive models may allow for more precise selection of patients eligible for extensive surgery. This review will discuss the ability of imaging and laparoscopy to predict non-resectable disease in patients with advanced ovarian cancer.
