Diagnosing Ovarian Cysts - the DOC Study

Background:

Ovarian cysts are commonly found in women with an estimated prevalence up to 35 % in premenopausal women and 17 % in postmenopausal women. The majority are of benign character and sometimes self-limiting; however, malignancy occurs as ovarian cancer (OC). More than 500 women are diagnosed with OC every year in Denmark, and Denmark has the highest OC incidence rate in the Nordic countries. Today, more than 70% of women with OC are diagnosed with advanced stage disease, where the five-year overall survival rate is 16-29% compared to 70-88% in early stages. Symptoms are often present months prior to diagnosis, however, they are vague and un-specific of character leading to delay of diagnosis. Attempts to improve patient awareness and correct symptom interpretation are encouraged. OC rates the fifth most deadly cancer among women and causes more deaths than any other cancer of the female reproductive system in more developed countries. Centralized treatment of OC9, OC guidelines10, fast-track OC programs, and a gynecologic cancer database (Danish Gynecological Cancer Database (DCGD)) are all national approaches that has contributed to a decreased mortality rate in Denmark during the past decades.

Yet further approaches to improve OC survival are highly needed. Given that OC survival is considerably better in early stage disease, methods to correctly differentiate benign ovarian tumors from OC at an early stage is a high research priority. In Denmark, gynecologists currently use the Risk of Malignancy Index (RMI) to detect women at high risk of OC. RMI calculation is based on transvaginal ultrasound findings, serum Cancer Antigen 125 (CA125) level, and menopausal status. Women in high risk of OC (RMI ≥ 200) are referred in the national fast-track cancer program for further diagnostics at a tertiary center specialized in gynecologic malignancies, whereas women in low risk of OC (RMI < 200) can be observed or treated at a local department of gynecology or in a gynecology practice. However, RMI is not optimal to distinguish between benign and malignant tumors with a sensitivity of 92 %, specificity of 82 %, negative predictive value of 97 %, and a positive predictive value of 62 % in a Danish population.

Considering the high prevalence of benign ovarian cysts in comparison to OC, it is of utmost importance, that efforts to detect OC at an early stage is not accompanied by an unintended increase of false positive tests, resulting in overdiagnosis and overtreatment. New promising studies suggest an improvement in the diagnostics and risk assessment models by adding the serum biomarker Human Epididymis protein 4 (HE4) and systemized ultrasound imaging in accordance to the International Ovarian Tumor Analysis (IOTA).

The biomarker HE4 is a secretory gluco-protein that is highly elevated especially in serum from women with OC. The diagnostic performance of HE4 is equal to CA125 and improves further when HE4 and CA125 is combined In the newly developed diagnostic Copenhagen Index (CPH-I), HE4 is combined with age and CA125. On tertiary center data, CPH-I has shown to distinguish benign gynecologic disease from OC significantly better than RMI in the international, multicenter, validating cohort demonstrating an area-under-curve (AUC) at 0.951 and 0.935, respectively.

IOTA is an internationally accepted classification system for describing a pelvic mass; and was developed by the IOTA group and published by Timmerman et al. in 2001. The classification system is based on standardized definitions and terms regarding ultrasound morphological features, doppler analysis, and presence/absence of ascites. Several promising models for risk analysis of a pelvic mass (e.g. Logistic Regression 2 model, Simple Rules, ADNEX model) have been developed based on the IOTA classification system. An increasing number of gynecologists in Denmark are IOTA certified; however, IOTA has never been validated in a Danish cohort and is not yet implemented . HE4, CPH-I, and the IOTA classification appear to be more accurate in differentiating ovarian cysts. However, these methods all need further prospective evaluations in the primary- and secondary sector before clinical implementation in Denmark.

Purpose:

The purpose is to evaluate symptoms and the diagnostic performance of HE4, CPH-I and the IOTA classification in Danish women diagnosed with an ovarian cyst. Furthermore, the aim is to develop and validate an optimized diagnostic algorithm to differentiate benign ovarian cysts from OC, based on symptoms, biomarkers, and IOTA.

Methods:

Study design and population: The study is a prospective, observational study. After oral and written project information and consent is given, women are included by gynecologist from a gynecology practice or a department of gynecology in the Capital Region of Denmark. Women with ovarian cysts are eligible for study inclusion. Exclusion criteria are age < 18 years and unable to provide consent. We expect to include around 2.500 women in a period from ultimo 2019 to May 2022.

Power calculation: A preliminary power calculation was performed based on specificities of RMI and CPH at 0.78 and 0.81, respectively. Sample size was estimated to 665 cases using a 2 sample non-inferiority or superiority calculation with type I error set at 1 %, power at 0.80, a non-inferiority or superiority margin of -0.1, and a sampling ratio of 1.

Baseline data: Included women will be asked to answer a questionnaire with detailed information on health data and symptoms, that previously have been shown to be correlated to a pelvic mass.

Ultrasound: The ultrasound examination will be performed as part of the routine examination. It is performed in lithotomy position, with an empty bladder, using a transvaginal probe and high-end equipment. Ultrasound beam frequencies are set between 5.0 and 9.0 MHz, and the internal female reproductive organs are systematically examined with transvaginal gray scale. Doppler settings regarding pulse repetition frequency are fixed at 0.3 to 0.9 MHz, and color gain is gradually increased until appearance of flash artifacts; followed by a frequency decrease to a level, where slow flow in blood vessels is still visible when examining for tumor vascularity. A transabdominal probe is also used when necessary. The ovarian cyst is described by the gynecologist in accordance to IOTA terminology in the patient's medical record. The definition of a lesion and how to classify, distinguish, and measure the morphological features, including septae, solid tissue, papillary projections, regularity of internal walls, cystic content, acoustic shadows, and ascites, are specifically defined by the IOTA group.25 Furthermore, five specific ultrasound images per patient will be stored for later validation.

Blood sampling, biomarkers, and analyses: Blood sampling for HE4 analysis will be performed as part of the routine examination in women tested for CA125. Blood sampling takes place at the gynecology practice or at in- or out-hospital laboratories in the Capital Region. Blood samples will be collected in 4-milliliter blood collection tubes. After sampling, the blood is centrifuged and fractionated; and serum is extracted before analysis. After routine CA125 examination is made, serum will be safely stored in the Danish CancerBiobank (DCB), a part of the Bio- and GenomeBank, Denmark (RBGB).26 Samples will be stored at -80 degrees Celsius until batch analysis. Levels of HE4 will be quantitatively measured using Elecsys HE4 fully automated chemiluminescent immuno-assays.

Data collection and storage: Information on symptoms and health care information will be obtained from questionnaires answered by the patient. Information about routine diagnostics and follow-up are is retrieved from the electronical patient files used in the gynecology practice and departments of gynecology and Labka. Information on diagnosis is retrieved from Patobank and Danish Gynecological Cancer Database (DGCD). Together with ultrasound images and results from HE4 analyses, all data will be continuously transferred to an electronic database as prescribed by the Capitol Region in Denmark.

Statistical considerations:

Descriptive statistics of the study will be listed and compared between groups using the Chi square test or the Fisher's exact test for categorical variables and the Mann-Whitney test for continuous variables. Categorical variables will be listed as numbers (n) and percentages (%), whereas continuous variables will be listed in medians and ranges. P-values for the differences will be calculated, and a p-value ≤ 0.05 will be considered statistically significant.

CPH-I is calculated using the formula defined by Karlsen et al.:

CPH-I = - 14.0647 + 1.0649 x log 2 (HE4) + 0.6050 x log 2 (CA125) + 0.2672 x age / 10.

RMI is calculated by the formula by Jacobs et al.:

RMI = Menopause (yes=3 points; no=1 points) x ultrasonic findings (unilocular=0; bi- or multilocular=1, solid areas=1, excrescences=1, bilateral=1, ascites=1, other ovarian disease=1 --> 0-1=1 points and 2-6=3 points) x CA125 (U/mL).

The ability to differentiate between healthy women, women with non-gynecologic disease, benign gynecologic disease, and OC will be examined using univariate and multivariate logistic regression analyses.

Dissemination of study results:

All projects with positive, inconclusive, as well as negative results will be published in international peer-reviewed scientific journals. The study group and other relevant collaborators will be co-authors, if they meet the Vancouver criteria. Intellectual property of the study belongs to the study group and can be shared after consent.

Significance of the project:

Most ovarian cysts are benign and can be followed or surgically treated at the local gynecology practice or department of gynecology. Sometimes, however, it can be challenging to differentiate from OC with the diagnostic tools available. We expect to develop an optimized diagnostic tool used to detect women in low risk of OC from women in high risk of OC with a higher precision than existing risk models. Furthermore, the implementation of the DOC study will contribute to a quality improvement in the daily clinical practice applying consistent use of systematic ultrasound and standardized terminology in patient file templates.