The role of urine and serum biomarkers in the early detection of ovarian epithelial tumours

Abstract Ovarian cancer (OC) is one of the leading causes of gynaecological cancer mortality in women worldwide. If detected at an early stage (I, II), OC has a 90% 5-year survival rate; nevertheless, symptoms are often hidden, leading to late-stage (III, IV) diagnosis and a poor prognosis. The current diagnostic procedures, such as a pelvic exam, transvaginal ultrasound, CA-125 blood tests, serum HE4 tests and multivariate index assays (MIA), are insufficient. Sadly, surgery is frequently required to confirm a positive diagnosis. Therefore, there has been an increased interest in different biomarkers using a non-invasive test as a tool for the earlier diagnosis of OC to resolve the need for precise and non-invasive diagnostic methods. This review article aims to investigate how biomarkers influence early OC detection and to emphasise the role of using a combination of serum biomarkers panel rather than a single biomarker. In addition, this review provides insights into the current serum biomarkers, urine biomarkers and other emerging biomarkers in the early detection of OC for better specificity and sensitivity and to improve the overall survival (OS) rate.


Introduction
Ovarian cancer (OC) is a typical clinically malignant tumour in women, with a morbidity rate of 6.31/10 million cases and a mortality rate of 2.73/10 million cases (Siegel et al. 2018). Approximately 60-70% of OC patients have progressed to stages III-IV or have developed abdominal metastases because the patients' early symptoms are often insignificant (Reed et al. 2010). According to studies from 2013 to 2015, roughly 1.3% of women will be diagnosed with OC at some point. ( 'Seer, Surveillance, Epidemiology, And End Results Program--Seer, 2016). In addition, statistics from 2011 to 2015 show 11.6 per 100,000 annual cases of OC in women and 7.2 per 100,000 annual fatalities from OC (Seer, Surveillance, Epidemiology, And End Results Program-Seer, 2016). This cancer has a bad prognosis because of its advanced metastases at the time of diagnosis and the difficulties in detecting it early. More than 60% of cases are diagnosed after cancer has spread to other body organs. When the cancer is localised in the ovary at the time of diagnosis, the 5-year survival rate is 92%; however, when the cancer is diagnosed late, the 5-year survival rate drops to 29% (Siegel et al. 2018). Only 25% of ovarian tumours are identified at stage I . OC can be classified into genetic and nongenetic risk factors linked to a number of factors, including reproductive history, hormonal contraceptives use, family history, lifestyle and environment (Webb and Jordan 2017). The highest risk factors for OC are a family history of OC, especially if a relative has been diagnosed under 50 years old with germline BRCA1/BRCA2 mutations. The mutations of MSH2, MSH6, MLH1, PMS2, EPCAM, BRIP1, PALB2, RAD51C and RAD51D associated with Lynch syndrome may contribute to OC. It is estimated that about 18% of epithelial cancers, particularly high-grade serous carcinomas, are caused by inherited mutations (Weiderpass andTyczynski 2015, Lincoln et al. 2020). Despite advancements in surgical techniques, radiotherapy, and chemotherapy, the 5-year survival rate for OC is still just 30%, and the prognosis is poor (Ozols et al. 2004). OC cure rates have remained unchanged for the past 30 years. Patients with epithelial OC (EOC) undergo cytoreductive surgery (debulking) as well as platinum and paclitaxel chemotherapy in the initial stages of treatment. Recent developments in genomic analysis and cutting-edge molecular characterisation of tumour stroma, as well as the emergence of targeted therapies like bevacizumab and poly (ADPribose) polymerase (PARP) inhibitors, have made it possible to stratify OC patients to provide the best possible treatment. These breakthroughs will allow for the minimisation of side effects while increasing clinical efficacy. Furthermore, individuals in the early phases of the disease (FIGO I and II) had a better prognosis than those in the advanced stages (Narod 2016). OC progresses asymptomatically in its early stages due to its location within the pelvis (Norouzi-Borough et al. 2018a, 2018b. Due to late diagnosis and inadequate screening of OCs, they have a high mortality rate. As a result, OC is known as a 'silent killer' or 'whispering disease' (Speeckaert et al. 2013, Norouzi-Borough et al. 2018a, 2018b, Momenimovahed et al. 2019. Therefore, early detection of OC is crucial to patients' prognosis. Research has shown that several biomarkers could be used to detect OC at an early stage. Current tumour markers include the human epididymis protein 4 (HE4) (Speeckaert et al. 2013) and carbohydrate antigen-125 (CA-125) (Terlikowska et al. 2016). The risk of ovarian malignancy algorithm (ROMA) and risk malignancy index (RMI) (Wei et al. 2016, Xu et al. 2016, Lu 2018. These are essential tools for the differential diagnosis of patients with abdominopelvic masses.
This study overviews the present state of the science concerning OC biomarkers; HE4 and CA-125, how they can be used to manage OC patients clinically, and the direction of future research. In addition, this review explores how biomarkers have impacted OC detection, treatment and mortality and investigate the demand for novel biomarkers for early-stage I and II OC. Finally, this review also attempts to draw the importance of combining biomarker panels rather than a single biomarker approach in the early diagnosis of OC for higher specificity and sensitivity.

Search strategy and inclusion and exclusion criteria
According to Worldwide cancer data and World Cancer Research Fund International (wcrf.org), OC is one of the leading causes of a gynaecologically related death in women worldwide, and there is still no efficient screening method for detecting OC in its early stages, and the current screening procedures are insufficient. Hence the main objective of this literature review is to explore how biomarkers have impacted OC detection, treatment and mortality and to investigate the demand for novel biomarkers for early-stage I and II OC. The keywords and terms of the central concepts for this review included OC, ovarian signalling pathway, diagnosis and prognosis, and current and new OC biomarkers, which were developed and combined to form the search strategy.
A systematic search of Google Scholar, PubMed, Web of Science and Sci-Hub databases was combined for relevant publications, from the time of their inception to November 2021, using the key terms 'ovarian cancer screening', 'biomarkers for ovarian cancer', urine biomarkers, the role of HE4 in OC screening, and CA-125 for OC screening. 'Ovarian cancer early detection biomarkers in urine', 'ovarian cancer early detection biomarkers in serum' and 'new developing biomarkers for OC early detection' are among the topics covered. This study intends to provide insights to clinicians and researchers to develop non-invasive strategies for OC to aid in early detection and improve the overall survival (OS) rate. Additionally, it could benefit women and health care workers worldwide by improving the quality of their lives. The references provided in this review were organised using Endnote, a reference management software.

Current biomarkers for ovarian cancer
Diagnostic tools are being improved to detect OC at the earliest possible stage in its progression, allowing treatment to begin as soon as possible and enhancing the related cure rate (Lu 2018). High specificity is also crucial to reduce the necessity for surgical procedures, which can be costly and result in problems like postoperative infections (Raja et al. 2012). Moreover, a highly sensitive and specific screening procedure is associated with a high positive predictive value (PPV) and a high negative predictive value (NPV), thus minimising the occurrence of false-positive/negative readings which result in a false referral or failure of cancer detection, respectively (Simon 2015). In an attempt to test high-risk women for OC early detection, procedures like pelvic examination, transvaginal ultrasonography (TVUS) and laparoscopy have been developed (Midulla et al. 2012). Biomarkers are biological characteristics that can be objectively measured to indicate a healthy or pathological state, the stage of a disease and predict therapy response) (Cesano and Warren 2018). According to these definitions, biomarkers can be diagnostic, prognostic and predictive. A biomarker, however, now refers to any substance, structure, or process that can be measured and used to influence or predict disease occurrence or outcome (United Nations Environment Programme. International Labour Organization. World Health Organization. International Program on Chemical Safety 2001). The number of potential biomarkers has been growing exponentially, thanks to advances in proteomics and molecular biology, with blood plasma and serum (plasma excluding clotting factors) having long been the main focus for biomarker research. However, due to its lower complexity proteome and lack of homeostatic processes as compared to blood, urine, a readily accessible waste material, is receiving more interest (Jing and Gao 2018).

Commonly used serum and urine biomarkers for ovarian cancer detection
In 1976, the carcinoembryonic antigen (CEA) biomarker was utilised as the first OC biomarker. However, it was shortly substituted by CA125 due to CEA's low sensitivity and specificity for detecting OC in patient serum (Guo et al. 2017). CA125 seems to be the only serological biomarker now in routine use for the management of patients with OC (Herzog et al. 2011, Wang et al. 2011) and it has the most research behind it. CA125 levels are increased ( � 35units/mL) in 83% of OC patients (Coticchia et al. 2008). A higher level of CA125 serum is linked to a higher risk of death (CA125 serum levels can be used to assess malignant potential prior to surgery (Micha et al. 2009). Despite CA125 being one of the highest in terms of specificity and sensitivity among existing OC biomarkers, many weaknesses render CA125 ineffective during early-stage screening (Micha et al. 2009). It is a poor biomarker due to its sensitivity and specificity, particularly in early-stage OC development. HE4 is a proteinase inhibitor that was found to have increased expression in OC in 1999 and has been argued as a superior biomarker when compared with CA125 for several years (Schummer et al. 1999, Lowe et al. 2008, Jia et al. 2017a, 2017b. A systematic analysis was conducted to evaluate the diagnostic value for HE4 in detecting OC. The authors found urinary HE4 to be sensitive and highly specific for the diagnosis of OC. They suggested that diagnosis of OC using this non-invasive method could be highly efficient (Liao et al. 2015, Hellstrom et al. 2010). In addition, HE4 was detected in the urine of OC patients with a specificity of 94.4%, including those with stage I/II disease. These results suggest that assessing HE4 in the urine may be helpful for early-stage diagnosis and provide valuable information on treatment response, allowing the implementation of effective treatment plans for patients. However, fewer than 16% of the 413 OC patients had stage I disease in this study. To investigate the ability of HE4 to detect early-stage OC, it would be necessary to conduct further studies on a much larger sample size that would focus on early disease stages (Hellstrom et al. 2010, Wang et al. 2011). Blood has traditionally been the main subject for biomarker discovery due to the wealth of knowledge of the dynamic blood proteome, which is more complex than the urine proteome, relative uniformity of volume and concentration between patients, and a large number of blood samples are collected as part of routine tests (Jing and Gao 2018). Blood is often the first choice for the discovery of biomarkers, but it has several disadvantages; for example, homeostasis mechanisms are involved in regulating body temperature, pH, composition and many other factors (Chovatiya andMedzhitov 2014, Modell et al. 2015). It could be surmised that some biomarkers are removed from the blood as part of these natural processes. This is a particular issue for earlystage disease biomarkers as they may only be present and detectable in the blood for a limited amount of time, and most likely in low concentrations, at any time point in the disease progression (Jing and Gao 2018). Most blood biomarkers for cancer have been antibodies to cancer rather than cancer-specific proteins ( Table 1). As urine is the product of the homeostasis mechanism, it is more likely to have chemical changes from most body sites, making it a viable source of biomarkers. In addition, protein modification and degradation occur in the blood and during the collection procedure. However, degradation is not thought to occur during the urine-collecting process (Kolch et al. 2005, Decramer et al. 2008. Therefore, urine is the more favourable body fluid for biomarker detection of the ease of access, the volume available and a protein profile that is less complex than that of plasma or serum. In addition, the proteins or peptides excreted with urine tend to maintain more excellent stability (Ye et al. 2007, Wang et al. 2011. Therefore, urine biomarkers for OC, such as HE4 and Bcl-2 (B-cell lymphoma 2), may be able to differentiate between benign and malignant tumours (Li et al. 2009). In Table 2, we illustrate common urine biomarkers used for early OC detection (Table 2).
Despite the abundance of biomarkers and multivariate index assays (MIA) currently available for OC, most are thought to be ineffective in detecting early-stage disease. Therefore, a statement was made that an effective screening technique must have a specificity greater than 99.6%, a PPV greater than 10% and a better than 75% sensitivity to get an early-stage diagnosis ). In addition, the assay would be easily quantifiable, clinically validated and non-invasive, with many current biomarkers failing to meet these criteria (Blyuss et al. 2015) (Table 3).

Single biomarkers for the early detection of ovarian cancer
Over 70% of OCs are identified late (Holschneider and Berek 2000). The challenge is to discover the biomarkers for the early detection of OC that might benefit clinical output. Owing to the rare prevalence of OC, maximum specificity and sensitivity have to be the goal for screening for earlystage disease Jacobs 2001, Jacobs andMenon 2004,) that is, attaining 99.6% of specificity and >75% sensitivity to overcome insupportable false-positive results and thereby achieve a PPV of 10% (Yang et al. 2017). Despite CA125 being one of the highest in specificity and sensitivity amongst existing OC biomarkers, many weaknesses render CA125 ineffective during early-stage screening (Micha et al. 2009). CA125 is an imperfect biomarker, in terms of specificity and sensitivity, most significantly in early-stage OC development of CA125 (Nowak et al. 2015). In addition, the histology type of OC affects the concentration of CA125, with most EOC displaying the highest levels and mucinous tumours displaying low expression levels of CA125 (Duffy Sensitivity and specificity of commonly used biomarkers of ovarian cancer diagnosis.  . 2005). Therefore, using CA125 as the only biomarker for analysis will leave out those not expressing this antigen. Circulating concentrations of CA-125, HE4, prolactin, IL-2R, CA 15-3, CA 19-9, CA 72-4, MIF, Cyfra 21-1, TNFR1, TNFR2, IL-6, IL-7, IL-10, IGFBP1, TSH, TNF-a, GH, TIMP-1, ACTH and osteopontin had been stated to be significantly (p < 0.001) better in serum of patients with early-stage OC as compared to healthy women. While serum levels of HE4, IL-2R, prolactin, CA 15-3, CA 19-9, CA 72-4, Cyfra 21-1, TNFR1, TNFR2, IL-6, IL-7, IL-10, TNF-a, TSH, IGFBP1, MMP-7, VCAM-1, eotaxin-1, FSH, LH, ErbB2, ApoA1, TTR, adiponectin and CD40L differed significant (p < 0.01) among patients with early-stage (degrees I and II) and late-stage (degrees III and IV) OC (Yurkovetsky et al. 2010). However, serum biomarkers apart from CA125 are not presently used as a detection tool for early-stage disease because of their lower sensitivity or specificity (Baron et al. 2003, Perkins et al. 2003. Consequently, the quest is on early biomarkers with high specificity and sensitivity to continue to predict metastasis's prevalence before it manifests within the patients. Based on sensitivity and specificity, several potential single serum biomarkers for early analysis of OC are summarised and listed in (Table 4).

Panels of biomarkers for early-stage detection of ovarian cancer
Nevertheless, none of the single biomarkers for early-stage cancer have accomplished the required specificity and sensitivity. Despite the wealth of biomarkers and MIA already available for OC, most are considered inadequate in detecting early-stage disease. A proposal was made that for a practical screening approach to attain an early-level diagnosis, it must obtain a specificity of more than 99.6%, to achieve a PPV larger than 10% and sensitivity greater than 75% (Badgwell and Bast 2007). In addition, the assay would be effortlessly quantifiable, clinically validated and non-invasive, with many current biomarkers failing to meet these standards (Blyuss et al. 2015). To fill this gap, extraordinary approaches and sources of biomarkers must be investigated further to increase the possibility of patients' benefits and novel biomarker discovery. Combining biomarkers to be used in a panel has been proven to have better diagnostic sensitivity than single markers. Several studies based on multi-biomarker methods had been reported to enhance the sensitivity over the single biomarker at comparable specificity in diagnosing early-stage OC. A study of serum from patients with benign (n ¼ 262) or malignant (n ¼ 196) ovarian tumours was analysed alongside healthy donors (n ¼ 386) for CEA, CA125, CA19-9 and HE4 levels, alone and in combination with each other. This study cited the highest sensitivity achieved while CA125 and HE4 were used in combination (Chen et al. 2018). CA125, in combos with numerous other serum tumour biomarkers, was tested. For example, CA-125 and HE4 have been proven to be excellent amongst all biomarker combos in distinguishing benign cells from the early stage of OC at 74.2% sensitivity, and 85% specificity, whereas CA-125, HE4 and EGFR drastically distinguish the benign from malignancy at 75.9% sensitivity and 87.5% specificity (Nolen et al. 2010). For instance, a panel of six biomarkers inclusive of CA-125, osteopontin, leptin, prolactin, MIF and IGF-II improved the sensitivity to 95.3% and specificity to 99.4% for OC detection (Visintin et al. 2008). Studies that combine biomarkers in diagnostic screens now include the usage of MIA, consisting of OVA1, that detect biomarkers CA125, pre-albumin, apolipoprotein A-1, b2-microglobulin and transferrin and generate an index score, from 0 to 10, to determine whether there is a high or low probability of malignancy. The sensitivity of OVA1 has been proven to be between 92.2% and 94% in studies and 98.1% in combination with TVUS and pelvic examination, which is significantly better than the standard test for CA125 at 76% (Brodsky et al. 2017, Ueland 2017). However, although OVA1 confirmed a very high sensitivity rate, the specificity Sensitivity and specificity and positive predictive value (PPV) of common serum Biomarkers of ovarian cancer used for early diagnosis. for detecting OC was far lower than in comparison with testing for CA125 in patient serum at 35%. The second-generation MIA (MIA2G), named Overa, was released in 2016 and combines CA125, HE4, apolipoprotein A-1, follicle-stimulating hormone and transferrin to maintain the assay sensitivity and increase specificity for OC (Coleman et al. 2016). Recent research reported the best overall performance of CA-125 with CA 19-9, EGFR, G-CSF, Eotaxin, IL-2R, cVCAM and MIF that, improved the sensitivity by 98.2% and sensitivity of 98.7% in diagnosing the early-stage OC as highlighted in (Table 5). Here, we propose that a multi-serum biomarker panel is a better version for the early detection of OC (Table 5).

Discussion and future directions
The ability of current OC tests to provide an early-stage diagnosis is currently lacking. However, existing diagnostic approaches only identify 30-45% of women with early-stage OC. Most women continue to be diagnosed later (Anaya 2016) when there are observable signs and symptoms of OC. In recent years, several studies have discovered various serological biomarkers in various combinations. However, there is currently no reliable, validated biomarker(s) in terms of specificity, sensitivity, and stability. As a result, developing appropriate multiple biomarkers that promote early detection with high accuracy is crucial. Currently, in OC patients, there is no predictive method for the response to chemotherapy, and all patients are treated with standard chemotherapy without adverse effects. This represents an unreasonable approach; selecting patients based on prognostic data would be attractive to tailor treatment to each patient. Biomarkers such as HE4 seem able to do this, but further studies are needed to confirm this opportunity. In addition, these biomarkers seem to be able to successfully identify patients with poor survival and higher risk of death, also playing an essential role in disease monitoring. Furthermore, we know that prognosis is highly correlated with histology and BRCA mutations. This information may change the clinical and surgical management of each patient. Therefore, it may be important to investigate HE4 levels and their correlation with the above characteristics. Despite the high rate of OC recurrence, surveillance strategies have not been standardised. This has led to various follow-up strategies, all sharing CA-125 regularly assessed. In addition, the FDA recently approved HE4 together with CA-125 for EOC follow-up, even though there are few recent studies on its use in this setting. Only a few prospective controlled studies have been published, and all analysed studies had a small number of women. Relative to CA-125, HE4 was shown to be an earlier indicator of OC recurrence with a lead time of 5-8 months. In addition, the sensitivity and specificity of HE4 alone or in combination with other markers (CA-125, ca-72.4) seem to be higher in diagnosing OC recurrence relative to CA-125 alone (Plotti et al. 2012). New markers such as HE4 should be tested in the follow-up of patients with EOC to improve the surveillance program performance. The challenge is to anticipate the diagnosis of OC recurrence and translate the early diagnosis of relapse into a quality of life and survival Improvement (Rustin et al. 2010). To better assess biomarker performance, it is important to focus attention on menopausal status. Higher HE4 concentrations are detectable in postmenopausal women. Second, HE4 may differ in EOC histological subtypes and EOC FIGO stages. Third, these results had significant publication bias in HE4 studies, and the retrieved studies were very heterogeneous. Moreover, the prognostic role of HE4 appears to be the most promising and provides us with interesting insights. Even if there is no optimal treatment for recurrence, this feature should be investigated to better select patients into risk groups and tailor surgical and clinical treatment to their characteristics.
In the near future, innovative technologies based on very small samples are projected to alter medical practice radically. However, liquid biopsy tests that are now available are not yet ready for clinical usage. In addition, there is still much work to be done to develop effective assays for early OC diagnosis. The underlying genetic instability of OC is one of the problems for future management. Gene mutations that can be targeted are scarce in high-grade serous OC (HGSC) (Nguyen et al. 2013). Except for TP53 mutations and BRCA1/2 abnormalities, all additional genetic changes are found in less than 10% of HGSC. Instead, in HGSC, chromosome instability is typical: it causes widespread gains and losses in DNA-activating oncogenes and inactivates tumour suppressors, resulting in tumour growth and resistance to chemotherapy (Cancer Genome Atlas Research Network 2011, Birrer et al. 2015, Mirza et al. 2016. Characterising and validating potential druggable targets in big amplicons remains a significant problem in turning molecular discoveries from multiple comprehensive genomic research into clinically valuable treatments (Ciriello et al. 2013). Thorough biology and laboratory-based research will be required to discover the tumour's driving factors and how they evolve throughout the disease's natural history.
Several miRNAs, small non-coding RNAs that downregulate protein expression of target genes, has been detected in whole blood, plasma, serum and exosomes of OC patients. MiRNAs showed differential expression in different histological types of OC, and in a study by Prahm et al., it was shown that many different miRNAs acted as predictors of OS, time to progression (TTP), progression-free survival (PFS) and chemoresistance (Prahm et al. 2018). Recent studies have hypothesised a role for miRNAs as prognostic markers for recurrent EOC after chemotherapy. The expression of hsamir-1273g-3p was significantly downregulated in recurrent EOC compared with healthy controls (G€ unel et al. 2018). miR-200a, miR-200b and miR-429 have significant potential as diagnostic markers in relapse (Hu et al. 2009). Therefore, identifying a series of miRNAs as prognostic biomarkers may also lead to novel screening tools. However, although several reports have demonstrated the applicability of circulating miRNAs as cancer biomarkers, these molecules are still considered insufficient for clinical applications, mainly due to the lack of large-scale validation and inconsistency between detection devices (Yokoi et al. 2018). This strategy has already led to identifying important genes with therapeutic potential. Exosomes, which carry tumour-specific antigens and nucleic acids (especially microRNAs), can be extracted easily and utilised as non-invasive diagnostic and prognostic biomarkers. Furthermore, they can be utilised for prognosis and prediction of therapeutic efficacy as well as the development of metastatic disease based on their distinct molecular patterns between different stages of the disease and healthy controls, in addition to early detection (Bast et al. 2009, Jia et al. 2017a, 2017b. These subcellular nano-particles are detectable in practically all physiological fluids; nevertheless, selecting a suitable isolation technique based on downstream analysis, type and volume of the initial sample is crucial to achieving the best results considering the cancer type (Taylor and Gercel-Taylor 2008). Aptamers that target specific biomarkers found in OC could provide new hope for diagnosis and treatment. In addition, uterine lavage procedures are simple and safe, and this method appears to have much potential for use in clinical practice. Combining blood biomarkers with nucleic acids, such as free DNA, mRNA, microRNAs and circulating tumour DNA (ctDNA), is also growing rapidly as a malignancy diagnostic tool (Bettegowda et al. 2014, Halvaei et al. 2018, Schwarzenbach et al. 2011. A matched set of protein and nucleic acid biomarkers for non-invasive OC screening and diagnosis could be beneficial. Gene-wide transcriptome profiling may help researchers find new OC biomarkers by identifying abnormal genes. Multivariate studies, including tumour biomarkers, ultrasonography, and appropriate algorithms, would be effective. Since there are few circulating biomarkers in the early stages of the disease, as a point of care diagnostic tool, the development of a lowcost, reliable, robust, and rapid detection multiplexed biosensor system could aid in early diagnosis and timely management. Although there is still a long way to go to develop a reliable method with high specificity for early detection of these cancers, cancer prevention, and intervention methods will become more effective in the near future, especially with the development of new cancer-specific screening tools. Recent evidence in OC defines CSCs as a significant contributor to disease aggressiveness, drug resistance, and tumour recurrence. However, the presence of CSCs has been proposed long ago and evaluated in several studies (Lagan� a et al. 2015). This concept states that tumour growth is driven by a small population of stem cells that appear to be involved in all stages of tumorigenesis: from initial development to metastasis and, therefore, tumour recurrence. Regarding OC, ovarian CSCs have been isolated from OC cell lines, ascites, and primary and metastatic tumours (Trott 1994). Discussion of the reasons for CSCs concerning ovarian biomarkers is based on the fact that the antigens and molecular targets of CSCs can be used as OC markers. For example, adhesion molecule for epithelial cells (EpCAM) is a transmembrane protein essentially expressed in human adenocarcinomas (Munz et al. 2009). EpCAM was significantly expressed in EOC tissues compared with normal ovarian tissues. Expression along with CD44 was associated with EOC's stage, grade, and metastasis. Metastatic and recurrent tumours express higher levels of EpCAM than primary OC s (Bellone et al. 2009). ALDH enzymes belong to a family of enzymes involved in metabolic processes and are responsible for the oxidation of aldehydes in carboxylic acids (Marcato et al. 2011). CSCs with high ALDH activity were shown to correlate with advanced tumour stage, grade, and poor outcomes in OC patients (Sun et al. 2016). It has been demonstrated that antigens like CD44, EpCAM and ALDH1 can be used as OC biomarkers of recurrence (Muinao et al. 2018). However, the most significant limit as biomarkers is that none of these current CSC markers are expressed exclusively by OC tissues, thus the necessity to combine in different panel markers. Instead, using markers expressed by CSCs in OC could be essential to discover the mechanisms of chemoresistance, find therapeutic targets and develop new treatment modalities, especially in the case of metastasis or recurrence.

Recommendations and conclusions
OC is a high-mortality gynecological condition affecting women all over the world. Unfortunately, while substantial improvement has been made in OC patients' detection and overall 5-year survival rate, both remain very low. This is because successful early-stage diagnostic biomarkers and therapeutic goals remain inadequate. Therefore, it is also essential to identify novel early detection molecules or therapeutic targets, including pathways with a minimally invasive approach and high sensitivity and specificity, which can significantly increase OC patients' OS rate and quality of life.
In our opinion, there is enough data to develop clinical trials in which treatment decisions are based on the biomarkers in combination/alone to confirm their role in OC management. However, more extensive cooperation between institutions and countries is needed to perform prospective, controlled, and randomised studies to confirm all these preliminary data and the roles of these biomarkers in clinical practice. The upcoming goal will be to find the most potent combination of biomarkers for screening that can detect recurrence as early as possible with high sensitivity and specificity, considering that early treatment of disease recurrence seems to be the optimal way to improve survival. In the future, the definition of new biomarkers and the use of newgeneration sequencing technology for malignant OC could also help to identify specific markers for molecular targets therapy, as OC stem cell-associated biomarkers suggest.