Not yet recruitingNCT07883551OvaClassUpdated Oct 20, 2026

AI Analysis of CT Scans to Classify Ovarian Tumours as Benign, Borderline or Malignant Before Surgery

An observational study in Ovarian Cancer, Ovarian Tumor and Borderline Ovarian Tumor, sponsored by Catharina Ziekenhuis Eindhoven. Not yet recruiting at 7 sites in 2 countries. Open to female participants aged 18 Years and older. Per ClinicalTrials.gov, last updated 2026-10-20.

Sponsored by Catharina Ziekenhuis Eindhoven · Observational

Updated Oct 20, 2026Newly registeredGo to Updates ↓

Study type: Observational
Model: Cohort
Time perspective: Retrospective
Enrollment: 1,300
Ages: 18 Years and older
Sex: Female

Study summary

The goal of this observational study is to learn if a computer program can tell what type of ovarian tumour a person has from a CT scan taken before surgery. The program uses artificial intelligence (AI). It sorts tumours into three groups:

benign: not cancer
borderline: abnormal cells that usually behave less aggressively than cancer
malignant: cancer

Before surgery, doctors try to judge which group a tumour belongs to. This guides where and how the operation is done. This judgement is often hard, especially for borderline tumours. An AI program that does it well could help doctors plan surgery.

The main questions it aims to answer are:

How well does the AI program sort ovarian tumours into benign, borderline and malignant on CT scans?
Does the AI program work as well on scans from hospitals it did not learn from?

Researchers will compare the AI program's results with the final diagnosis from the tumour tissue removed during surgery. They will also compare the program with methods doctors use today, such as ultrasound-based risk scores and the radiologist's report. In a smaller group, researchers will test the program on MRI scans.

This study only uses information that already exists. It includes adults who had surgery for a mass in or near the ovary at hospitals in the Netherlands and Belgium. Researchers will use:

CT or MRI scans made before surgery as part of regular care
information from medical records
the tissue diagnosis after surgery

Participants will not be contacted. They will not have any extra scans, tests or visits. Each hospital codes the data before sharing them, so researchers cannot see participants' names. The study does not change anyone's care.

Read the detailed description

Background and rationale Preoperative distinction between benign, borderline and malignant ovarian tumours determines referral, surgical approach and extent of surgery. Ultrasound-based models such as IOTA-ADNEX and the Risk of Malignancy Index (RMI) are the current triage tools, but their performance varies between populations and users. CT is often performed in indeterminate cases or for staging. There are no validated diagnostic criteria for CT, and interpretation depends on the reader. Published CT-based computer-aided diagnosis (CADx) models are mostly single-centre, binary (benign versus malignant), dependent on manual tumour segmentation and rarely externally validated. Borderline tumours need different management but are usually excluded or merged with malignant tumours. Earlier work from the Catharina Cancer Institute and Eindhoven University of Technology (TU/e) showed considerable domain shift between hospitals in a multicentre CT dataset (Koch et al., 2026).
Index test The index test is a CADx pipeline built on SPECTRE, a CT foundation model developed at TU/e and pre-trained with self-supervised learning on more than 250,000 CT scans. SPECTRE is fine-tuned with parameter-efficient adaptation. The pre-trained encoder is kept frozen or partially unfrozen, and lightweight task heads are trained for tumour localisation and three-class classification. The pipeline works directly on the CT volume. Tumour localisation is learned within the model, so no tumour masks or bounding boxes are needed for training or evaluation. An uncertainty quantification module produces calibrated confidence estimates alongside each prediction. The overall architecture is fixed. Training details such as the fine-tuning schedule, loss weighting and calibration method may be adapted based on interim results. Radiomics-based machine learning (handcrafted features with classifiers such as logistic regression, support vector machines, random forests or gradient boosting) is used only as a benchmark comparator. The model is applied retrospectively and its output is never available to treating clinicians.
Reference standard Definitive postoperative histopathology, classified as benign, borderline or malignant according to the 2020 WHO Classification of Female Genital Tumours. Ambiguous cases are adjudicated by an experienced gynaecological pathologist to keep labelling consistent across centres. In bilateral tumours with discordant histology, a predefined rule is applied (for example, selection of the lesion with the highest malignant potential).
Data acquisition and preprocessing Participating centres identify eligible patients and pseudonymise and code imaging, clinical and pathology data locally before transfer under formal data-sharing agreements. The key linking codes to identities stays at the local centre. Images are exported in DICOM format and may be converted to volumetric formats (NIfTI or NRRD). CT preprocessing includes intensity windowing (for example -100 to +300 HU), normalisation and voxel resampling to a standard resolution. Scanner manufacturer and model, slice thickness, contrast phase and reconstruction kernel are recorded for analysis of domain shift. For the MRI pilot, T2-weighted, T1-weighted with and without fat suppression, diffusion-weighted with ADC maps and contrast-enhanced T1-weighted sequences are used when available.
Model development and validation Within the development cohort, stratified cross-validation (or an equivalent resampling method adapted to the final sample size) is used, stratified by outcome class. Hyperparameter tuning and model selection take place only within the training data of each resampling iteration. All scans and lesions from one patient are kept in the same split to prevent data leakage. External validation uses one or more centres that are excluded entirely from model development. The allocation of centres to development and external validation is documented before any performance analysis. Calibration methods such as Platt scaling or temperature scaling may be applied.
Statistical analysis Discrimination is assessed with one-vs-rest AUROC per class and macro- and micro-averaged multiclass AUROC. Classification metrics include sensitivity, specificity, positive and negative predictive values, F1-score, balanced accuracy and confusion matrices. 95% confidence intervals are estimated by bootstrapping where appropriate. Calibration is assessed with calibration curves and the Brier score. Net benefit across clinically relevant risk thresholds is assessed with decision curve analysis. Comparison with RMI, IOTA-ADNEX, radiologist reports and, in the MRI pilot, O-RADS MRI is restricted to patients in whom the comparator is available or can be reconstructed. Paired comparisons use the DeLong test for ROC curves, McNemar's test for paired proportions, kappa statistics for agreement and the Brier score for probabilistic predictions. Baseline characteristics are compared with chi-squared or Fisher's exact tests and t-tests or Mann-Whitney U tests as appropriate. Two-sided p-values are reported with a nominal significance level of 0.05, with emphasis on effect sizes and confidence intervals. Missing clinical data are described and not imputed. Model behaviour is explored with attention and saliency maps (for example Grad-CAM).
Exploratory analyses Where data can be linked in coded form and are covered by ethics approval and data-sharing agreements, exploratory analyses will assess whether adding blood-based biomarkers (CA-125, CA-125 glycovariants, CEA, ctDNA, proteomics) or radiologist report content to CT imaging changes classification performance. In a subset with digitised histopathology slides, AI-based analysis of whole slide images will be explored to find histological correlates of CT-based tumour characteristics. These analyses do not change the primary imaging-based objective.
Reporting Results will be reported following TRIPOD+AI, supplemented by CLAIM and STROBE where applicable.

Conditions studied

Ovarian Cancer
Ovarian Tumor
Borderline Ovarian Tumor

Browse trials for

Ovarian Neoplasms 2,757

Keywords

Ovarian neoplasms
adnexal mass
borderline ovarian tumour
ovarian cancer
artificial intelligence
deep learning
transfer learning
foundation model
computer-aided diagnosis
computed tomography
magnetic resonance imaging
preoperative diagnosis
diagnostic accuracy
multiclass classification
external validation
domain shift

In context

Ovarian Neoplasms

2,757 studies on the registry are indexed under Ovarian Neoplasms; 735 are open to participants now.

This study's planned enrollment of 1,300 is above the median of 200 across 537 observational studies indexed under Ovarian Neoplasms.

Browse Ovarian Neoplasms studies →

Lead sponsor

Catharina Ziekenhuis Eindhoven is the lead sponsor of 85 studies on the registry; 16 are open to participants now.

Counted across the registry records on this site, refreshed daily.

Who can participate

Ages eligible: 18 Years and older
Sexes eligible: Female
Accepts healthy volunteers: No
Sampling method: Non-probability sample

Study population

Adult patients evaluated for an ovarian or adnexal mass of unknown nature at academic and teaching hospitals in the Netherlands and Belgium, who had a preoperative CT and/or MRI scan as part of routine care and subsequently underwent surgical removal of the mass with definitive histopathology. Benign, borderline and malignant tumours are included, as is advanced-stage ovarian cancer when the primary adnexal mass can be identified on imaging. CT or MRI was performed in routine care, mostly after indeterminate or suspicious ultrasound findings or for staging. The case mix therefore reflects a second-line population referred for cross-sectional imaging rather than all patients with an adnexal mass. Data come from a previous multicentre CADx cohort, the OVIDETECT registry and additional participating centres.

Go through the criteria one by one — yes, no or not sure — and take the list to your doctor or the study team. Kept in this browser only.

Inclusion criteria

Age 18 years or older at the time of imaging
Ovarian or adnexal mass on clinical and/or radiological assessment that can be reliably identified as the index lesion on CT and/or MRI
Preoperative CT and/or MRI scan available, acquired as part of routine clinical care
Surgical removal of the ovarian or adnexal mass, with definitive postoperative histopathology available as reference standard
Bilateral ovarian tumours are eligible. If histopathology differs between sides, a predefined rule is applied (for example, selection of the lesion with the highest malignant potential)
Advanced ovarian malignancy, including ascites or peritoneal disease, is eligible if the primary adnexal mass can be clearly identified on imaging

Exclusion criteria

Exclusion Criteria:

No definitive postoperative histopathology available
Imaging of insufficient diagnostic quality (for example incomplete series, corrupted files or non-diagnostic sequences)
Extra-ovarian disease that prevents reliable identification or attribution of the adnexal lesion as the index tumour (for example diffuse peritoneal carcinomatosis or a confirmed non-ovarian primary malignancy)
Previous ovarian malignancy or prior oncological treatment affecting the adnexal region
Pathology report judged indefinite or inconclusive by the reviewing clinician

Study design

Observational model: Cohort
Time perspective: Retrospective
Enrollment: 1,300 participants (estimated)
Patient registry: No

Groups and cohorts

CT development cohort
* Description: Adult patients with an adnexal mass who had preoperative CT and surgery with definitive histopathology at the development centres (Catharina Hospital Eindhoven, Amphia Hospital, Netherlands Cancer Institute-Antoni van Leeuwenhoek, Leiden University Medical Centre and Amsterdam UMC). Includes patients from a previous multicentre CADx cohort and the OVIDETECT registry, approximately 930 scans, of which around 118 are borderline tumours. Used for model training, tuning and internal cross-validation. * Intervention: SPECTRE-based CADx model
CT external validation cohort
* Description: Adult patients meeting the same criteria at centres held out entirely from model development (anticipated: Maastricht UMC+ and/or Ghent University Hospital), approximately 400 scans. Used only to assess generalisability across institutions. Centre allocation is fixed and documented before analysis. * Intervention: SPECTRE-based CADx model
MRI pilot cohort (exploratory)
* Description: Adult patients from participating centres with a preoperative MRI of the adnexal mass and definitive histopathology, approximately 100 to 200 scans. Used to explore preliminary CADx feasibility on MRI and to compare with O-RADS MRI. Not part of the primary CT-based objective. * Intervention: SPECTRE-based CADx model

What researchers measure

Primary outcomes

Discrimination of the CADx model for three-class classification (benign, borderline, malignant): macro-averaged AUROC
Area under the receiver operating characteristic curve (AUROC) of the SPECTRE-based CADx model on preoperative CT. Calculated per class (one-vs-rest) and macro-averaged across the three classes, with histopathology according to the 2020 WHO classification as reference standard. Reported for internal cross-validation in the development cohort and for the centre-held-out external validation cohort, with 95% confidence intervals estimated by bootstrapping. Range 0 to 1, higher values indicate better discrimination.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery

Secondary outcomes

Sensitivity, specificity and predictive values per tumour class
Sensitivity, specificity, positive predictive value and negative predictive value of the CADx model for each class (one-vs-rest), against histopathology. Decision thresholds are set in the development cohort and fixed before external validation. Reported with 95% confidence intervals for internal and external validation.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
F1-score, balanced accuracy and confusion matrix for three-class classification
Per-class and macro-averaged F1-score, balanced accuracy and the full three-by-three confusion matrix of predicted versus histopathological class, for internal and external validation.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Calibration of predicted class probabilities
Agreement between predicted probabilities and observed histopathological outcomes, assessed with calibration curves and the Brier score, before and after any recalibration (for example Platt or temperature scaling). Reported for internal and external validation.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Net benefit of CADx-supported triage (decision curve analysis)
Net benefit of using the CADx model to identify borderline or malignant tumours across a range of clinically relevant risk thresholds, compared with treat-all and treat-none strategies and with IOTA-ADNEX where available.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Diagnostic performance of the CADx model compared with IOTA-ADNEX
Paired comparison in patients with an available or reconstructable ADNEX result. ADNEX risk categories are collapsed to benign, borderline and malignant. Compared by AUROC (DeLong test), sensitivity and specificity (McNemar's test), Brier score and agreement (kappa).
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Diagnostic performance of the CADx model compared with the Risk of Malignancy Index (RMI)
Paired comparison in patients with a calculable RMI. Because RMI is a binary score, comparison is made for benign versus borderline or malignant tumours at the threshold used in routine practice (RMI ≥200). Compared by AUROC (DeLong test) and sensitivity and specificity (McNemar's test).
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Diagnostic performance of the CADx model compared with the radiologist's CT report
Paired comparison with the classification derived from the original clinical CT report using a predefined mapping rule. Compared by sensitivity, specificity (McNemar's test) and agreement (kappa).
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Inter-centre generalisability of CADx performance
Macro-averaged AUROC, sensitivity and specificity per centre, and the difference between internal cross-validated and external validation performance, to quantify domain shift related to centre, scanner and acquisition protocol.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
CADx performance by histological subtype (exploratory)
Proportion of correctly classified tumours and class-specific sensitivity per histological subtype (2020 WHO classification), for example serous, mucinous, endometrioid, clear cell and non-epithelial tumours.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
CADx performance by menopausal status and tumour size
Macro-averaged AUROC, sensitivity and specificity in subgroups by menopausal status (pre- versus postmenopausal) and by maximum tumour diameter on imaging.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
CADx performance on MRI and comparison with O-RADS MRI (exploratory pilot)
Within the MRI pilot cohort, AUROC, sensitivity and specificity of the CADx model on MRI against histopathology, and paired comparison with the O-RADS MRI score where available.
Time frame: Preoperatively (MRI scan acquired before surgery), compared with histopathology of the tumour removed at surgery

Other outcomes

Added value of blood-based biomarkers to CT-based CADx (exploratory)
Change in macro-averaged AUROC when CA-125, CA-125 glycovariants, CEA, ctDNA or proteomic data are combined with CT imaging, in the subset with linkable biomarker data.
Time frame: Preoperatively (CT scan and blood-based biomarkers obtained before surgery), compared with histopathology of the tumour removed at surgery
Added value of radiologist report content as auxiliary input (exploratory)
Change in macro-averaged AUROC when structured or free-text radiologist report content is added to CT imaging as model input.
Time frame: Preoperatively (CT scan acquired before surgery), compared with histopathology of the tumour removed at surgery
Correspondence between whole slide image features and CT-based tumour characteristics (exploratory)
Descriptive and quantitative association between AI-derived histopathology features on whole slide images and CT-based model representations, in a subset of tumours with digitised slides.
Time frame: Preoperatively (CT scan) and at surgery (histopathology specimen of the removed tumour)

Study locations

7 sites

Ghent University Hospital
Ghent, 9000, Belgium
Netherlands Cancer Institute - Antoni van Leeuwenhoek
Amsterdam, 1066 CX, Netherlands
Amsterdam University Medical Centre
Amsterdam, 1081 HV, Netherlands
Amphia Ziekenhuis
Breda, 4818 CK, Netherlands
Catharina Ziekenhuis
Eindhoven, 5623 EJ, Netherlands
Leiden University Medical Centre
Leiden, 2333 ZA, Netherlands
Maastricht UMC+
Maastricht, 6229 HX, Netherlands

References and documents

Publications

Koch AH, Claessens CHB, Nies I, Schultz EWR, Hellstrom TAE, Boers T, Dekker EML, Muntinga CLP, Hoogstad-van Evert JS, Niers-Stobbe I, Bruining A, Lok CAR, Nederend J, van der Sommen F, Piek JMJ. Evaluating Performance of a Deep Learning-based Artificial Intelligence Model for Ovarian Tumor Classification Using a Multicenter CT Dataset. Radiol Imaging Cancer. 2026 Mar;8(2):e250027. doi: 10.1148/rycan.250027. PubMed 41686074 ↗
Koch AH, Jeelof LS, Muntinga CLP, Gootzen TA, van de Kruis NMA, Nederend J, Boers T, van der Sommen F, Piek JMJ. Analysis of computer-aided diagnostics in the preoperative diagnosis of ovarian cancer: a systematic review. Insights Imaging. 2023 Feb 15;14(1):34. doi: 10.1186/s13244-022-01345-x. PubMed 36790570 ↗

Individual participant data

Plan to share: No — Individual participant data will not be shared outside the study consortium. The dataset consists of retrospective CT and MRI scans with linked clinical and pathology data from several hospitals in the Netherlands and Belgium. Data are shared within the consortium under a research agreement with joint controllership and the ethics approvals of each centre. Neither these agreements nor the consent obtained cover sharing with other researchers. Volumetric CT data also carry a residual risk of re-identification after pseudonymisation. Any reuse beyond the protocol would need new ethics review and agreement from all participating centres. Where these agreements allow, the study team plans to share model code, trained model weights and derived, non-identifiable features through public or institutional repositories. Researchers interested in collaboration can contact the study team.

Updates

1 registry update since Sep 25, 2026

Registered

First appeared on the registry. No changes since

Oct 20, 2026

Show all 1 update

Oct 20, 2026
First appeared on the registry

From the registry record's own update history. This site started tracking changes on Sep 25, 2026; for anything earlier, see the record history on ClinicalTrials.gov ↗

Registry details

Key details

Study ID: NCT07883551
Lead sponsor: Catharina Ziekenhuis Eindhoven
Collaborators: Eindhoven University of Technology, Maastricht University, Maastricht University Medical Center, Amphia Hospital, Leiden University Medical Center, Amsterdam UMC, location VUmc, University Hospital, Ghent, Antoni van Leeuwenhoek Hospital
Responsible party: Basiel Weyers (Fellow gynaecological oncology, Catharina Ziekenhuis Eindhoven) — Principal investigator
First posted: Oct 20, 2026
Start date: Nov 2026 (estimated)
Primary completion: Oct 2031 (estimated)
Completion: Oct 2031 (estimated)
Last update: Oct 20, 2026

Study contacts

Basiel Weyers, Medical Doctor

Contact

basiel.weyers@catharinaziekenhuis.nl

+3140 239 93 00

Jurgen Piek, M.D., Ph.D.

study chair · Catharina Ziekenhuis Eindhoven

Oversight

Data monitoring committee: No
FDA-regulated drug: No
FDA-regulated device: No

View the source record on ClinicalTrials.gov ↗

Not currently enrolling

This study is not yet recruiting, as verified in Oct 2026. You cannot join it, but the record below documents what was studied.

Follow this study

Get an email when the registry record changes — status, dates, results — or when someone posts here.

Discussion

Questions and observations about this study, from anyone following it. Not medical advice, and not a channel to the study team — their contact details are on the registry record.

Sign in to join the discussion. Reading takes no account; posting does. You choose a display name, and a pseudonym is the default.

Nothing here yet. If you are running this trial, taking part in it, or weighing whether to, this is the place to say so.

Other studies in this condition

Ovarian Neoplasms →

Other studies from this sponsor

Catharina Ziekenhuis Eindhoven

Start the discussion