Ovarian Cancer Screening: Recommendations and Future Prospects

Abstract

Background

Ovarian cancer remains a significant cause of mortality among women, largely due to challenges in early detection. Current screening strategies, including transvaginal ultrasound and CA125 testing, have limited sensitivity and specificity, particularly in asymptomatic women or those with early-stage disease. The European Society of Gynaecological Oncology, the European Society for Medical Oncology, the European Society of Pathology, and other health organizations currently do not recommend routine population-based screening for ovarian cancer due to the high rates of false-positives and the absence of a reliable early detection method.

Method

This review examines existing ovarian cancer screening guidelines and explores recent advances in diagnostic technologies including radiomics, artificial intelligence, point-of-care testing, and novel detection methods.

Results and Conclusion

Emerging technologies show promise with respect to improving ovarian cancer detection by enhancing sensitivity and specificity compared to traditional methods. Artificial intelligence and radiomics have potential for revolutionizing ovarian cancer screening by identifying subtle diagnostic patterns, while liquid biopsy-based approaches and cell-free DNA profiling enable tumor-specific biomarker detection. Minimally invasive methods, such as intrauterine lavage and salivary diagnostics, provide avenues for population-wide applicability. However, large-scale validation is required to establish these techniques as effective and reliable screening options.

Key Points

• Current ovarian cancer screening methods lack sensitivity and specificity for early-stage detection.

• Emerging technologies like artificial intelligence, radiomics, and liquid biopsy offer improved diagnostic accuracy.

• Large-scale clinical validation is required, particularly for baseline-risk populations.

Citation Format

• Chiu S, Staley H, Jeevananthan P et al. Ovarian Cancer Screening: Recommendations and Future Prospects. Rofo 2025; DOI 10.1055/a-2589-5696

 

Scroll to top

Zusammenfassung

Hintergrund

Ovarialkarzinome sind nach wie vor eine der häufigsten Todesursachen bei Frauen, was vor allem auf die Schwierigkeiten bei der Früherkennung zurückzuführen ist. Aktuelle Screening-Strategien, darunter die transvaginale Sonografie und die CA125-Bestimmung, haben eine geringe Sensitivität und Spezifität insbesondere bei asymptomatischen Frauen oder bei einer im Frühstadium befindlichen Erkrankung. Die „European Society of Gynaecological Oncology“, die „European Society for Medical Oncology“, die „European Society of Pathology“ und andere medizinische Fachgesellschaften sprechen aktuell keine Empfehlung für ein Ovarialkarzinom-Routinescreening in der Bevölkerung aus, da die Rate falsch-positiver Ergebnisse hoch ist und eine zuverlässige Früherkennungsmethode fehlt.

Methoden

In dieser Übersicht werden die aktuellen Leitlinien für das Ovarialkarzinom-Screening vorgestellt und die jüngsten Fortschritte bei diagnostischen Technologien wie Radiomics, künstliche Intelligenz, Point-of-Care-Tests und neuartige Nachweismethoden analysiert.

Ergebnisse und Schlussfolgerung

Die neuen Technologien sind vielversprechend, da sie verglichen mit herkömmlichen Methoden höhere Sensitivität und Spezifität aufweisen und somit die Diagnose von Ovarialkarzinomen verbessern. Künstliche Intelligenz und Radiomics können die Früherkennung von Ovarialkarzinomen durch die Identifizierung subtiler diagnostischer Muster revolutionieren, während auf Liquid-Biopsy basierende Ansätze und Profiling zellfreier DNA den Nachweis tumorspezifischer Biomarker ermöglichen. Minimalinvasive Methoden wie intrauterine Lavage und die Speicheldiagnostik bieten die Möglichkeit für eine breite Anwendbarkeit. Es ist jedoch eine umfangreiche Validierung erforderlich, um diese Techniken als wirksame und zuverlässige Screening-Optionen zu etablieren.

Kernaussagen

• Die aktuellen Methoden zur Früherkennung von Ovarialkarzinomen sind nicht zuverlässig.

• Neue Technologien wie künstliche Intelligenz, Radiomics und Liquid-Biopsy bieten eine verbesserte diagnostische Genauigkeit.

• Eine groß angelegte klinische Validierung ist erforderlich, insbesondere für Populationen mit erhöhtem Risiko.

 

Scroll to top

Keywords

Ovarian cancer - Biomarkers - Early detection - Artificial intelligence - Radiomics - screening

Introduction

Ovarian cancer (OC) is the 8^th most prevalent female cancer affecting approximately 324,000 females, with 200,000 deaths annually worldwide [1]. Despite advancements in genomic profiling and targeted therapies, the overall 5-year survival rate remains below 50%. This is secondary to the asymptomatic nature of early-stage and low-volume disease in addition to its extensive tumoral heterogeneity and pathogenesis [2]. Early detection of OC is crucial, as survival rates improve dramatically when the disease is diagnosed at a low-volume with limited tumor dissemination patterns [3] [4]. However, effective screening methods for early detection have proven elusive, with current strategies showing limited success in reducing mortality as summarized in [Table 1].

Table 1 Summary of key studies on ovarian cancer screening. This table summarizes major studies evaluating the effectiveness of ovarian cancer screening methods. Key parameters such as population size, follow-up duration, primary outcomes, and limitations are highlighted. The findings demonstrate earlier-stage detection in some trials but consistently reveal challenges in achieving significant reductions in ovarian cancer mortality. ADNEX = Assessment of Different Neoplasias in the Adnexa, AS = annual screening, CA-125 = Cancer Antigen 125, CI = confidence intervals, FTC = fallopian tube cancer, HR = hazard ratio, IOTA =  International Ovarian Tumor Analysis, IQR = interquartile range, MMS = multimodal screening, NACT = neoadjuvant chemotherapy, NPV = negative predictive value, NS=no screening, OC = tubal or ovarian cancer, PDS = primary debulking surgery, PPV = positive predictive value, RMI = risk malignancy index, ROCA = risk of ovarian cancer algorithm, ROMA = risk of malignancy algorithm, RCT = randomized control trial, RRSO = risk-reducing salpingo-oophorectomy, TVUSS = transvaginal ultrasound, UC = usual care, USS = ultrasound scan.

Study and year	Study type and size	Screening method	Positive test result definition	Median follow-up period (years)	Key findings
Current recommendations
UKCTOCS (2021) [5]	RCT n=202,562 MMS=50,625 (OC=522) USS=50,623 (OC=517) NS=101,314 (OC=1,016)	MMS: Longitudinal serum CA125 triaged based on ROCA algorithm If high risk TVUSS is performed USS: primarily TVUSS, with additional follow-up ultrasound examinations conducted in cases of unsatisfactory or abnormal initial results. No screening	CA-125 (ROCA) Intermediate risk (≥1/1818) Elevated risk (≥1/500) USS: One or both ovaries with complex morphology, simple cysts greater 60 cm3, or ascites	16.3 (IQR 15·1–17.3)	Earlier diagnosis did not translate into improved mortality MMS vs. USS: Sensitivity 89.4% vs. 84.9% Specificity 99.8% vs. 98.2% PPV 43.3% vs. 5.3% HR (95% CI) 0.96 (0.83–1.10) vs. 0.94 (0.82–1.08)
UKCTOCS (2023) [6]	RCT HGSOC=779 MMS=259 NS=520			9.5 (IQR 6·04–13.00)	Fewer advanced-stage HGSOC were diagnosed in the MMS vs. the NS group (p=0·0003). Greater proportion of patients underwent: PDS (61% vs. 42%, p<0.0001) Achieved complete tumor clearance following cytoreductive surgery (46% vs. 30%, p<0.0001) Combination of surgery and chemotherapy (74% vs. 64%, p=0.0032) MMS group demonstrated a 6.9% improvement in survival at 18 years compared to the no screening group (21% vs. 14%, p=0.042)
PLCO (2011) [8]	RCT n=78,216 AS=39,105 (OC=212) UC=39,111 (OC=176)	AS: CA125 for 6 years and TVUSS for 4 years UC: Not offered annual screening with CA125 for 6 years or TVUSS but received routine medical care	CA125: ≥35 IU/mL USS: Ovarian volume >10 cm3 Cyst volume >10 cm3 Any solid area or papillary projection extending into the cavity of a cystic ovarian tumor of any size; or any mixed (solid and cystic) component within a cystic ovarian tumor	12.4 (range 10.9–13.0)	Combined screening with CA125 and TVUSS did not lead to a reduction in OC mortality compared to standard care. Mortality RR, 1.18 (95% CI, 0.82–1.71) OC-related deaths AS = 3.1 per 10,000 person-years US = 2.6 per 10,000 person-years False-positive screening occurred in 3,285 cases with: Surgical follow-up = 1,080 (32.9%) Serious complication in those who underwent surgical intervention = 163 (15%)
ROCkeTS (2024) [13]	Multicenter, prospective diagnostic accuracy study n=1,242 (OC=215)	RMI 1 (CA125, USS findings, menopausal status) IOTA ADNEX ROMA (CA125, HE4, menopausal status) IOTA SRRISK IOTA simple rules CA125	RMI1 at ≥ 200 ROMA at multiple thresholds 14.4 25.3 27.7 29.9 IOTA ADNEX at 3% and 10% IOTA SRRisk model at 3% and 10% IOTA Simple Rules (malignant vs. benign, or inconclusive) CA125 at ≥35 IU/mL	1	IOTA ADNEX at 3% and 10% and ROMA at 14.4 have the highest sensitivity of the assessed tests (>96%). IOTA ADNEX at 10% had the highest sensitivity (96.1% vs. 82.9%) with less specificity (58.5% vs. 87.4%) than RMI at 250 in postmenopausal patients. AUC RMI1 0·92 (0·89–0·94) ROMA 0·92 (0·90 to 0·95) IOTA ADNEX 0·93 (0·91 to 0·95) IOTA SRRisk 0·88 (0·85 to 0·91) IOTA Simple Rules 0·89 (0·85 to 0·92) CA125 0·91 (0·88 to 0·93)
High-risk population
UKFOCSS (2017) [17]	Prospective multicenter cohort Women with an estimated lifetime OC/FTC risk ≥ 10%=4,348	CA125 every 4 months triaged based on ROCA algorithm TVUSS occurred Yearly if ROCA results were normal Within 2 months of abnormal ROCA results with repeat CA125 after 2 months	ROCA value Normal = <85th percentile Intermediate = 85th percentile up to 1 in 5 ROCA Elevated = ROCA > 1 in 5	4.8 (range 0–8.7)	OC/FTC detection within 1 year: Sensitivity 94.7% (CI, 74.0% to 99.9%) PPV 10.8% (CI, 6.5% to 16.5%) NPV 100% (CI, 100% to 100%) ROCA-based screening is an option for high risk of OC/FTC women who opt to defer or decline RRSO, due to its high sensitivity and notable shift to earlier stages. The long-term survival benefits among high-risk women who undergo screening are still unclear.

 

Scroll to top

Current Recommendations

Several large-scale randomized controlled trials have investigated the potential of OC screening in the general population. The UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) enrolled 202,638 women to evaluate ultrasound alone and multimodal screening using serial cancer antigen 125 (CA125) measurements interpreted by the Risk of Ovarian Cancer Algorithm (ROCA), along with transvaginal ultrasound (TVUSS). Although screening resulted in earlier-stage detection, with a median follow-up of 16.3 years, it did not show a statistically significant reduction in OC mortality [5]. A recent sub-analysis of the UKCTOCS study focused on patients with high-grade serous ovarian cancer (HGSOC) sought to understand why this earlier-stage diagnosis observed in the multimodal screening group did not correlate with a decrease in mortality. The findings revealed a 6.9% improvement in survival at 18 years in the multimodal screening group compared to the no screening group (21% vs. 14%, p=0.042). This modest survival benefit illustrates the limited gains from early detection likely due to the inherent tumor characteristics and minimal advantage derived from OC treatments [6]. This underscores the difficulty of achieving meaningful survival benefits, even when cancers are detected earlier. Similarly, the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial enrolled 78,216 women to evaluate the impact of annual CA125 and TVUSS screening. This trial also failed to demonstrate a reduction in mortality, and it highlighted the high rates of false positives and unnecessary surgical interventions, leading to considerable morbidity, such as surgical complications and emotional distress [7] [8].

In light of these results, the European Society of Gynaecological Oncology, the European Society for Medical Oncology, the European Society of Pathology (ESGO-ESMO-ESP), the American College of Obstetricians and Gynecologists, and the US Preventive Services Task Force (USPSTF) do not recommend routine OC screening for asymptomatic women with a baseline risk. This stems from concerns regarding the potential consequences of current screening approaches, such as invasive procedures following false-positive results [9] [10] [11] [12].

Recently, the Refining Ovarian Cancer Test accuracy Scores (ROCkeTS) study assessed multiple risk prediction models for their effectiveness in distinguishing between low- and high-risk OC cases among postmenopausal patients. They compared six indices and identified that the International Ovarian Tumor Analysis – Assessment of Different Neoplasms (IOTA ADNEX) model offered higher sensitivity but lower specificity than the Risk of Malignancy Index (RMI) at 250. The IOTA ADNEX model is a risk model used to classify ovarian masses into benign and 4 grades of ovarian tumors. Whereas the RMI is a numerical calculation based on menopausal status, CA125 level, and presence of abnormal US characteristics. These results suggest that the IOTA ADNEX at the 10% threshold may now represent the preferred diagnostic approach for evaluating OC risk in postmenopausal women [13]. Although these stratification tools provide valuable guidance, their limitations necessitate combining them with other diagnostic modalities and clinical judgment to optimize patient management.

The low prevalence of OC in the general population and the absence of reliable early-stage biomarkers emphasize the limitations of current screening strategies. As such, no population-wide screening program has been implemented, highlighting the need for more effective targeted approaches.

 

Scroll to top

High-Risk Populations

Hereditary OC accounts for approximately 20% of all cases and is predominantly attributable to pathogenic germline mutations in high-penetrance genes such as BRCA1 and BRCA2. Carriers of BRCA1 mutations are estimated to have a lifetime OC risk of 39–58%, while BRCA2 mutation carriers face a risk of 13–29% [3] [14] [15] [16]. These individuals are a key target for tailored screening strategies, as their elevated risk makes them unsuitable for general population screening, which has proven ineffective.

In response, the UK Familial Ovarian Cancer Screening Study (UKFOCSS) explored the use of the ROCA, which interprets longitudinal CA125 measurements, as opposed to relying on a single-threshold CA125 value. This dynamic approach enables the detection of subtle yet clinically significant deviations in CA125 levels that might otherwise be overlooked in static models, particularly in high-risk populations. ROCA-based screening demonstrated high sensitivity of 94.7% and a negative predictive value of 100%. Additionally, early-stage detection was higher in occult cancers (83.3%) compared to screen-detected cancers (38.5%). This underlines the potential of dynamic algorithms for more personalized screening particularly in those that decline or defer risk-reducing surgery. However, its effect on survival remains uncertain [17].

While preliminary data suggests that this approach may facilitate earlier diagnosis and intervention, the impact on OC-specific mortality remains unclear. Further longitudinal studies with robust follow-up are essential to confirm whether these screening strategies improve long-term outcomes in high-risk populations.

 

Scroll to top

Cost-Effectiveness of Screening

Currently, population screening for OC is not recommended either for the general population or high-risk groups by USPSTF and ESGO-ESMO-ESP [9] [10] [11] [12]. The health-economic benefit of OC screening is a factor, particularly given the high rates of false positives and the associated burden of unnecessary procedures [5] [6] [7] [8]. Large trials highlight that screening in the general population often results in non-beneficial interventions, leading to patient distress, increased healthcare costs, and potential surgical complications. In contrast, targeted screening for high-risk groups, such as those with BRCA mutations, may offer better clinical and health-economic value, as these patients are more likely to benefit from early detection [17] [18]. Clinicians must consider these economic and clinical trade-offs when recommending screening, ensuring that strategies are personalized to individual risk profiles to balance early detection with cost-efficient patient care.

 

Scroll to top

Role of Artificial Intelligence

Artificial intelligence (AI) is increasingly being explored for its potential to enhance OC screening by improving diagnostic accuracy and risk stratification. One innovative application involves the analysis of online search activity and consumer behavior as non-traditional screening tools. Research has shown that people often search online for health-related information before seeking medical advice [19]. Barcroft et al. demonstrated that certain online search patterns, such as frequent queries about symptoms like abdominal bloating or pelvic pain, can be predictive of OC among individuals referred by GPs with suspected cases. Notably, differences in online search behavior between individuals with benign and malignant diagnoses were detectable up to 360 days before GP referral when search data was analyzed directly. However, when the data was categorized by health-related queries, these differences were apparent only 60 days prior [20]. AI models analyzing large datasets from search engines may facilitate earlier detection by identifying individuals at risk, thus prompting timely medical consultations before clinical symptoms manifest.

The CLOCS (Consumer Lifestyle and Online Behavior) study introduces a novel approach by utilizing supermarket loyalty card data to identify lifestyle patterns associated with OC risk. Through the analysis of purchasing behaviors and dietary patterns, AI systems can detect subtle changes in consumer choices that may correlate with increased cancer risk, such as shifts in diet or the purchase of symptom-related products. Notably, increased purchases of pain and indigestion medications were observed up to 8 months prior to diagnosis (OR 2.9), with initial symptoms reported around 4.6 months and first medical presentations occurring approximately 3.6 months before diagnosis. This noninvasive, everyday data could serve as an additional layer of screening, helping to identify individuals who may require further medical evaluation based on emerging risk patterns [21] [22].

Moreover, machine learning (ML) models applied to web-based surveys on shopping habits have also shown promise in predicting OC. Of the women surveyed, 58.4% had purchased over-the-counter medications (e.g. analgesics and abdominal-related products) over a year before diagnosis. Women who self-medicated based on their doctor’s advice were seven times more likely to have experienced symptoms more than a year before being diagnosed with OC, compared to those who self-medicated independently. Predictive modelling indicates that these women who self-medicated under medical advice, exhibit distinct shopping behaviors that could be identified through purchasing data analysis [23].

In addition to analyzing non-traditional data, AI has also been applied to medical data for OC prediction. One study assessed the utilization of blood test results and tumor markers to develop predictive models. These models, including random forest (RF), support vector machine, decision trees, and artificial neural networks, were evaluated through 10-fold cross-validation. The RF model demonstrated the highest accuracy (>86%). The key predictive factors that were identified included human epididymal secretory protein 4 (HE4), CA125, and neutrophils, underscoring the potential of AI tools to predict OC with high accuracy and sensitivity using clinical data [24].

 

Scroll to top

Radiomics

Radiomics is a noninvasive technique that examines imaging features imperceptible to the naked eye, offering promise for enhancing OC screening and diagnosis. By applying advanced mathematical algorithms to medical images, radiomics extracts complex features such as texture, shape, wavelet, and intensity patterns. This quantitative insight may help to distinguish between benign and malignant lesions, potentially leading to earlier detection and more accurate OC diagnosis [25]. Traditional ultrasound-based models for characterizing adnexal masses are often subjective and dependent on the operator’s expertise, as shown in [Fig. 1]. To address these challenges, a recent study developed an end-to-end radiomics model using convolutional neural networks and radiomics features, achieving diagnostic accuracy (area under the curve (AUC) 0.90) comparable to expert ultrasound assessments and existing risk models [26].

Fig. 1 Ultrasound images of ovarian masses. (A) Physiologically normal ovary: characterised by a simple, anechoic cyst with well-defined borders and no solid components. (B) Mature cystic teratoma: presents as a cystic solid mass, including echogenic foci (e.g. fat, teeth or other calcifications). (C) Borderline serous ovarian tumour: features a multilocular cystic solid mass with thin, smooth or irregular septations and possible papillary projections within the cyst. (D) High-grade serous ovarian carcinoma: appears as a large, predominately solid mass with irregular septations, solid nodules, and increased vascularity on Doppler imaging, often accompanied by ascites. (E) Ovarian clear cell carcinoma: displays a cystic mass with thickened walls, solid areas, and possible areas of haemorrhage or necrosis, with increased vascularity in the solid components. (F) Endometrioid ovarian carcinoma: typically shows a multilocular cystic mass with solid or papillary nodules and irregular septations, along with increased vascularity on Doppler imaging. These examples highlight the diverse sonographic appearances encountered in ovarian cancer screening, emphasising the importance of accurate interpretation for distinguishing benign from malignant lesions.

Similarly, a decision support system (DSS) combining radiomics and ML was used to predict the malignancy risk of ovarian masses based on TVUSS and CA125 levels. The DSS demonstrated high accuracy (91%) and sensitivity (100%) in predicting the risk of malignancy in patients with suspicious ovarian masses, further highlighting the utility of radiomics and ML models in clinical decision-making [27]. This emphasizes the growing potential of radiomic models for enhancing OC screening, though further prospective evaluation is required to validate their clinical performance.

 

Scroll to top

Point of Care Tests

Point-of-care (POC) tests are critical for improving OC screening by reducing the long turnaround times inherent in traditional laboratory-based methods. Developing a simple, cost-effective, and noninvasive method for early-stage OC detection could significantly expedite diagnosis and be especially valuable in resource-limited settings where there is often limited access to specialized care. Despite this potential, validated POC methods with reliable results are lacking.

Currently, CA125 is used as a POC test in primary care, but this marker alone is not diagnostic of OC. Raamanathan et al. developed a promising solution with their programmable bio-nano-chip (p-BNC) platform, designed for rapid and multiplexed quantification of CA125 in serum. This microfluidic-based system uses a fluorescence-based sandwich immunoassay, achieving high precision (1.9% intra-assay, 1.2% inter-assay) and a detection limit of 1.0 U/mL in under 45 minutes. Validation with OC patient serum (n=20) showed excellent correlation (R²=0.97) with the gold standard ELISA, highlighting the p-BNC’s potential to enhance CA125 testing and offer rapid, accurate measurements that may improve early detection of OC [28].

Advancements in label-free immunosensors further expand the potential for rapid detection of various biomarkers including CA125 and HE4. These electrochemical sensors detect antigens in blood serum with high sensitivity across multiple linear ranges (1–100 pg/mL to 50–500 ng/mL) and provide results within 20–30 seconds. With impressive stability (60 days of use, 16 weeks of storage) and reusability (up to 9 cycles), these sensors offer a practical, user-friendly solution for POC OC screening. Despite these promising capabilities, widespread use of these immunosensors has yet to be realized, likely due to regulatory or technical challenges [29].

Urinary biomarkers, specifically CA125 and HE4, have shown promise, with significantly higher levels observed in OC patients compared to controls (p<0.05). When combined, urinary CA125 and HE4 demonstrated a sensitivity of 82.4%, comparable to serum CA125 alone (88.2%). Although further validation is required, these preliminary results suggest urinary biomarkers may serve as valuable noninvasive tools for triaging symptomatic women for more comprehensive OC assessment [30].

Another innovative approach integrates mobile technology with diagnostic platforms. Wang et al. demonstrated the integration of a cell phone-based microchip ELISA to detect urine HE4 which is noninvasive and easily obtainable. This portable detection method, leveraging a mobile application and a charge-coupled device, identified significantly elevated HE4 levels in urine from OC patients compared to healthy controls (p<0.001). This system achieved a high sensitivity (89.5%) and specificity (90%). However, as the study focused on urine samples from late-stage OC patients, further research is needed to validate its applicability in early-stage OC detection [31].

Alongside POC testing, the PROTECT-C (Population-based Risk Evaluation and Testing for Early Detection of Cancer) study, seeks to broaden the scope of genetic testing for early cancer detection, including OC. This study examines the feasibility of offering genetic testing to individuals regardless of their family or personal cancer history. Currently, NHS genetic testing is restricted to those meeting specific criteria, which leaves 50–80% of individuals with genetic mutations undiagnosed. These mutations are associated with increased risks for OC, as well as breast, bowel, and endometrial cancers. Expanding access to genetic testing could improve early detection and offer personalized approaches to cancer prevention [32]. The results from this study are eagerly anticipated, as they may complement existing POC diagnostic methods by providing a comprehensive genetic risk-based strategy for the early detection and prevention of OC.

 

Scroll to top

Novel Detection Methods

In addition to traditional biomarker testing, novel detection technologies are being explored to enhance the sensitivity and accuracy of OC screening. Salivary biomarker analysis offers a convenient, cost-effective, and noninvasive assessment, with the added benefit of stable sample storage and transport. Scebba et al. demonstrated proof-of-concept of its utility, conducting a proteomic analysis of saliva samples using mass spectrometry from OC patients, breast cancer patients, and healthy controls, identifying differential expression of several ionic markers. Proteomic assessment demonstrated a high level of accuracy in differentiating between OC and healthy subjects (specificity 97.1%, sensitivity 74.3%), and AUC of 0.905. A validated predictive model similarly achieved a sensitivity of 60% and specificity of 100%, correctly identifying all healthy subjects as true negatives [33]. While promising, these findings require replication in larger cohorts to confirm diagnostic reliability.

Intrauterine lavage (IUL) represents a more targeted approach for biomarker detection, where CA125 and HE4 are measured in collected uterine cavity fluid. While IUL can capture these markers, current results indicate that distinguishing between benign and malignant cases based solely on these concentrations remains challenging. Further optimization of IUL for early-stage detection is needed before clinical integration [34].

Proteomic analysis of cervicovaginal fluid is another novel approach, using a five-protein panel (serine proteinase inhibitor A1, periplakin, profilin1, apolipoprotein A1, thymosin beta4-like protein) demonstrated strong potential for early OC detection (AUC 0.86). Additionally, home-based collection may be feasible due to a moderate agreement (kappa=0.6) between physician-collected and self-collected samples. However, larger cohorts are required for validation and evaluation of its use for screening [35].

Emerging evidence also highlights the potential of cell-free DNA (cfDNA) as a promising biomarker for early OC detection. A study performed low-pass whole-genome sequencing of cfDNA to analyze key genomic features (including copy number variation, nucleosome footprinting, and DNA fragmentation) in 59 OC patients and 100 healthy controls. This demonstrated excellent diagnostic accuracy (specificity 98.0%, sensitivity 94.7%, AUC 0.98). Notably, the high sensitivity (85.7%) for early-stage OC, highlighted its potential for early detection. These results suggest that cfDNA holds promise as a tool for early OC screening and warrants further clinical validation [36]. Another study developed a cfDNA methylation-based liquid biopsy for early-stage HGSOC. The OvaPrint classifier, built on differentially methylated regions, achieved high diagnostic accuracy (PPV 95%, NPV 88%) with respect to distinguishing benign masses from HGSOC. While less sensitive for non-HGSOC epithelial OC, OvaPrint may help detect low-grade and borderline tumors with higher malignant potential, thus offering a promising tool for early risk assessment in symptomatic women [37].

Extracellular vesicle-based liquid biopsy has also emerged as a promising approach for early OC detection, offering enhanced sensitivity and specificity through abundant vesicle release and the assessment of multiple cancer-related markers on the vesicle surface [38]. Manning et al. evaluated an extracellular-based OC test in a blinded case-control study within the UKCTOCS trial analyzing HGSOC, controls, and false-positive TVUSS results. The test showed high specificity (97.7%) and sensitivity (82%) for detecting HGSOC in blood samples up to 12 months before diagnosis, outperforming CA125 (95.5% specificity, 63% sensitivity). Additionally, the OC test demonstrated a lower positivity rate (2.3%) in false-positive TVUSS cases compared to CA125 (10.4%), suggesting superior performance and warranting further evaluation in OC screening [39].

 

Scroll to top

Challenges, Controversies, Future Directions

A central challenge in OC screening is the balance between early detection and clinical outcomes. Despite evidence of earlier-stage detection, large trials have not demonstrated a significant reduction in OC mortality, thus raising doubts about the impact of routine screening on survival. Additionally, TVUSS-dependent screening strategies are subject to inter-operator variability, which can lead to misdiagnosis or missed lesions, potentially limiting the effectiveness of screening. Standardizing user-dependent factors is crucial to ensure robust and reliable implementation of screening programs. Reporting schemes such as IOTA simple rules, Society of Radiologists in Ultrasound consensus, and O-RADS-US have been devised to support non-expert ultrasound readers to achieve the best diagnostic performance [40] [41] [42]. Further work, if required, will help achieve consistency in the US scanning community which includes sonographers, radiologists, and sonographic gynecologists. We believe that regardless of the person scanning they must have the competency to perform a specialized gynecological scan in order to provide a robust and reliable OC screening program.

The heterogeneity of OC, including its diverse subtypes and molecular profiles, further complicates the development of a universal screening strategy, as one approach may not be applicable to all cases. Furthermore, screening asymptomatic, baseline-risk women may lead to misdiagnosis and unnecessary interventions, thus highlighting the difficulty in identifying appropriate populations for screening. Performing O-RADS MRI in women with an indeterminate or suspicious adnexal mass on ultrasound has been shown to improve the specificity for the detection of OC and is likely to emerge as a tool to prevent overdiagnosis and over-treatment as shown in [Fig. 2] [43].

Fig. 2 Comparison of USS findings with multi-parametric MRI in an indeterminate or suspicious adnexal mass. The USS was reported as a complex ovarian cyst with features consistent with an immature teratoma (predominantly solid with cystic components and internal vascularity in the solid component). MRI confirmed an 8cm dermoid cyst containing an enhancing Rokitansky nodule (as depicted by the arrow), with an O-RADS MRI score of 2. (A) Pelvic USS, (B) T1-weighted MRI, (C) T1 fat-saturated MRI, (D) T2-weighted MRI, (E) T1 fat-saturated dynamic contrast-enhanced MRI.

Future possibilities in OC screening lie in advances in precision medicine. New biomarkers, particularly those identifying genetic alterations linked to malignancy, may enable earlier, noninvasive detection. Increasing the use of O-RADS-MRI in suspicious cases may help reduce misdiagnosis and subsequent unnecessary intervention. Additionally, the integration of AI and ML with imaging and molecular profiling could improve diagnostic accuracy and risk stratification, facilitating the identification of high-risk individuals for more targeted screening.

 

Scroll to top

Conclusion

The landscape of OC screening is rapidly evolving, with numerous emerging technologies offering significant potential for earlier, more accurate detection. From advancements in biomarkers and liquid biopsies to AI-enhanced imaging and personalized genomic approaches, these innovations could revolutionize OC detection. However, further research and clinical validation are required to assess their effectiveness in real-world settings, particularly for populations at baseline risk.

 

Scroll to top

 

Scroll to top

Contributorsʼ Statement

SC and HS were the main contributors to preparation, writing, revision of the manuscript, and the gathering of data. SM and PJ assisted in data interpretation and provided critical revisions. CF contributed to the revision for important intellectual content. AR was responsible for the original manuscript design and drafting. In addition, she is also the guarantor for this paper and accepts full responsibility for the work and/or the conduct of the study.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 WCRF. Ovarian cancer statistics. Accessed September 02, 2024 at: https://www.wcrf.org/cancer-trends/ovarian-cancer-statistics/
• 2 CRUK. Ovarian cancer survival. Accessed June 21, 2024 at: https://www.cancerresearchuk.org/about-cancer/ovarian-cancer/survival#:~:text=Survival%20for%20all%20stages%20of%20ovarian%20cancer&text=almost%2045%20out%20of%20100,for%2010%20years%20or%20more
• 3 Sundar S, Neal RD, Kehoe S. Diagnosis of ovarian cancer. Bmj 2015; 351: h4443
• 4 NICE. Ovarian cancer: recognition and initial management. Accessed September 03, 2024 at: https://www.nice.org.uk/guidance/cg122/
• 5 Menon U, Gentry-Maharaj A, Burnell M. et al. Ovarian cancer population screening and mortality after long-term follow-up in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a randomised controlled trial. The Lancet 2021; 397: 2182-2193
• 6 Menon U, Gentry-Maharaj A, Burnell M. et al. Tumour stage, treatment, and survival of women with high-grade serous tubo-ovarian cancer in UKCTOCS: an exploratory analysis of a randomised controlled trial. Lancet Oncol 2023; 24: 1018-1028
• 7 Gohagan JK, Prorok PC, Hayes RB. et al. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: history, organization, and status. Control Clin Trials 2000; 21: 251s-272s
• 8 Buys SS, Partridge E, Black A. et al. Effect of Screening on Ovarian Cancer Mortality: The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial. JAMA 2011; 305: 2295-2303
• 9 González-Martín A, Harter P, Leary A. et al. Newly diagnosed and relapsed epithelial ovarian cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Annals of Oncology 2023; 34: 833-848
• 10 Burke W, Barkley J, Barrows E. et al. Executive Summary of the Ovarian Cancer Evidence Review Conference. Obstet Gynecol 2023; 142: 179-195
• 11 Grossman DC, Curry SJ, Owens DK. et al. Screening for Ovarian Cancer: US Preventive Services Task Force Recommendation Statement. Jama 2018; 319: 588-594
• 12 Ledermann JA, Matias-Guiu X, Amant F. et al. ESGO–ESMO–ESP consensus conference recommendations on ovarian cancer: pathology and molecular biology and early, advanced and recurrent disease. Annals of Oncology 2024; 35: 248-266
• 13 Sundar S, Agarwal R, Davenport C. et al. Risk-prediction models in postmenopausal patients with symptoms of suspected ovarian cancer in the UK (ROCkeTS): a multicentre, prospective diagnostic accuracy study. The Lancet Oncology 2024; 25: 1371-1386
• 14 Imyanitov EN, Kuligina ES, Sokolenko AP. et al. Hereditary cancer syndromes. World J Clin Oncol 2023; 14: 40-68
• 15 Suszynska M, Ratajska M, Kozlowski P. BRIP1, RAD51C, and RAD51D mutations are associated with high susceptibility to ovarian cancer: mutation prevalence and precise risk estimates based on a pooled analysis of ~30,000 cases. Journal of Ovarian Research 2020; 13
• 16 NCCN. National Comprehensive Cancer Network: NCCN Clinical Practice Guidelines in Oncology: Genetic/Familial High-Risk Assessment: Breast, Ovarian, and Pancreatic. Online: last access date: 10th December 2024.
• 17 Rosenthal AN, Fraser LSM, Philpott S. et al. Evidence of Stage Shift in Women Diagnosed With Ovarian Cancer During Phase II of the United Kingdom Familial Ovarian Cancer Screening Study. J Clin Oncol 2017; 35: 1411-1420
• 18 Sroczynski G, Gogollari A, Kuehne F. et al. A Systematic Review on Cost-effectiveness Studies Evaluating Ovarian Cancer Early Detection and Prevention Strategies. Cancer Prev Res (Phila) 2020; 13: 429-442
• 19 Jia X, Pang Y, Liu LS. Online Health Information Seeking Behavior: A Systematic Review. Healthcare (Basel) 2021; 9
• 20 Barcroft JF, Yom-Tov E, Lampos V. et al. Using online search activity for earlier detection of gynaecological malignancy. BMC Public Health 2024; 24
• 21 Brewer HR, Hirst Y, Sundar S. et al. Cancer Loyalty Card Study (CLOCS): protocol for an observational case–control study focusing on the patient interval in ovarian cancer diagnosis. BMJ Open 2020; 10: e037459
• 22 Brewer HR, Hirst Y, Chadeau-Hyam M. et al. Association Between Purchase of Over-the-Counter Medications and Ovarian Cancer Diagnosis in the Cancer Loyalty Card Study (CLOCS): Observational Case-Control Study. JMIR Public Health Surveill 2023; 9: e41762
• 23 Dolan EH, Goulding J, Tata LJ. et al. Using Shopping Data to Improve the Diagnosis of Ovarian Cancer: Computational Analysis of a Web-Based Survey. JMIR Cancer 2023; 9: e37141
• 24 Ayyoubzadeh SM, Ahmadi M, Yazdipour AB. et al. Prediction of ovarian cancer using artificial intelligence tools. Health Sci Rep 2024; 7: e2203
• 25 van Timmeren JE, Cester D, Tanadini-Lang S. et al. Radiomics in medical imaging—“how-to” guide and critical reflection. Insights into Imaging 2020; 11: 91
• 26 Barcroft JF, Linton-Reid K, Landolfo C. et al. Machine learning and radiomics for segmentation and classification of adnexal masses on ultrasound. npj Precision Oncology 2024; 8: 41
• 27 Chiappa V, Interlenghi M, Bogani G. et al. A decision support system based on radiomics and machine learning to predict the risk of malignancy of ovarian masses from transvaginal ultrasonography and serum CA-125. European Radiology Experimental 2021; 5
• 28 Raamanathan A, Simmons GW, Christodoulides N. et al. Programmable Bio-Nano-Chip Systems for Serum CA125 Quantification: Toward Ovarian Cancer Diagnostics at the Point-of-Care. Cancer Prevention Research 2012; 5: 706-716
• 29 Bilgi Kamaç M, Altun M, Yılmaz M. et al. Point-of-care testing: a disposable label-free electrochemical CA125 and HE4 immunosensors for early detection of ovarian cancer. Biomed Microdevices 2023; 25: 18
• 30 Barr CE, Njoku K, Owens GL. et al. Urine CA125 and HE4 for the Detection of Ovarian Cancer in Symptomatic Women. Cancers (Basel) 2023; 15
• 31 Wang S, Zhao X, Khimji I. et al. Integration of cell phone imaging with microchip ELISA to detect ovarian cancer HE4 biomarker in urine at the point-of-care. Lab on a Chip 2011; 11: 3411
• 32 Research YC. PROTECT-C. Accessed November 11, 2024 at: https://www.yorkshirecancerresearch.org.uk/research-story/protect-c
• 33 Scebba F, Salvadori S, Cateni S. et al. Top-Down Proteomics of Human Saliva, Analyzed with Logistic Regression and Machine Learning Methods, Reveal Molecular Signatures of Ovarian Cancer. Int J Mol Sci 2023; 24
• 34 Sia TY, Yaari Z, Feiner R. et al. Uterine washings as a novel method for early detection of ovarian cancer: Trials and tribulations. Gynecologic Oncology Reports 2024; 51: 101330
• 35 Rocconi RP, Wilhite AM, Schambeau L. et al. A novel proteomic-based screening method for ovarian cancer using cervicovaginal fluids: A window into the abdomen. Gynecologic Oncology 2022; 164: 181-186
• 36 Zhou H, Zhang X, Liu Q. et al. Can circulating cell free DNA be a promising marker in ovarian cancer? – a genome-scale profiling study in a single institution. J Ovarian Res 2023; 16: 11
• 37 Buckley DN, Lewinger JP, Gooden G. et al. OvaPrint—A Cell-free DNA Methylation Liquid Biopsy for the Risk Assessment of High-grade Serous Ovarian Cancer. Clinical Cancer Research 2023; 29: 5196-5206
• 38 Winn-Deen ES, Bortolin LT, Gusenleitner D. et al. Improving Specificity for Ovarian Cancer Screening Using a Novel Extracellular Vesicle–Based Blood Test. The Journal of Molecular Diagnostics 2024;
• 39 Manning B, Banerjee S, Bortolin LT. et al. Evaluation of a novel extracellular vesicle (EV) based ovarian cancer (OC) screening test in asymptomatic postmenopausal women. Journal of Clinical Oncology 2024; 42: 5553-5553
• 40 Timmerman D, Ameye L, Fischerova D. et al. Simple ultrasound rules to distinguish between benign and malignant adnexal masses before surgery: prospective validation by IOTA group. Bmj 2010; 341: c6839
• 41 Levine D, Patel MD, Suh-Burgmann EJ. et al. Simple Adnexal Cysts: SRU Consensus Conference Update on Follow-up and Reporting. Radiology 2019; 293: 359-371
• 42 Strachowski LM, Jha P, Phillips CH. et al. O-RADS US v2022: An Update from the American College of Radiology’s Ovarian-Adnexal Reporting and Data System US Committee. Radiology 2023; 308: e230685
• 43 Thomassin-Naggara I, Poncelet E, Jalaguier-Coudray A. et al. Ovarian-Adnexal Reporting Data System Magnetic Resonance Imaging (O-RADS MRI) Score for Risk Stratification of Sonographically Indeterminate Adnexal Masses. JAMA Network Open 2020; 3: e1919896-e1919896

Publication History

Received: 02 January 2025

Accepted after revision: 07 April 2025

Article published online:
23 May 2025