Ovarian cancer, especially its type II according to the dualistic model proposed by Kurman and Shih (1), represented mostly by the high-grade serous ovarian cancer (HGSOC) is the most lethal tumor of the female genital tract. The cumulative 5-year survival in the population of patients with all clinical stages does not exceed 48% (2). Despite the fact that some cases of the HGSOC are primarily chemo-refractory, the most of the cancers belonging to this group are chemosensitive to first-line chemotherapy; however, they quickly acquire the secondary chemoresistance that constitutes the main problem in effective management. Moreover, the HGSOC possesses unique behavior that allows spreading of the tumor cells, mostly in the form of cellular spheroids, from the primary tumor into the distant localizations in the peritoneal cavity. Therefore, the HGSOC is a highly malignant, rapidly progressive tumor characterized by poor prognosis and mortality reaching 90% of all ovarian cancer cases (3).

One of the main problems in the treatment of HGSOC is the existence of ovarian cancer stem cells (OCSCs) that reside inside tumor niches and cooperate with surrounding cells that compose tumor microenvironment (TME). The character of this cooperation shapes tumor behavior and influences several processes including dormancy, proliferation, metastasis, and, most of all, chemoresistance (4). Cancer stem cells are a population of cells capable of self-renewal and reproduction of the original phenotype of the tumor and are enriched especially in the advanced, disseminated, and recurrent tumors (5). There are two functionally distinct populations of CSCs, proliferating and quiescent, which occupy different niches inside the tumor. The proliferative CSCs are chemoresistant but vulnerable to overdoses of the chemotherapeutics; however, quiescent CSCs are in the autophagic state and could survive even high doses of anti-cancer drugs, thus enabling tumor relapse (6). One of the key phenomena responsible for regulation of stemness is epithelial–mesenchymal transition (EMT) viewed as a continuum of phenotype cellular states from complete epithelial and proliferative state, through several intermediate hybrid states to complete mesenchymal and invasive phenotype. Cancer stem cells could represent any of these steps due to the outstanding plasticity (7). This plasticity of CSCs is highly dependent on the patient’s immunosurveillance as well as on epigenetic and environmental signals from the TME (6). The most recognized stressors that could influence both phenotype and function of CSCs are hypoxia, acidity, mechanical stress, immunological response, epigenetic changes like DNA methylation, histone and non-coding RNA modifications, and, finally, activation of stemness signaling pathways (8–11).

The problem of the stemness is directly connected to the cancer dormancy that is dependent on the presence of circulating tumor cells (CTCs) and disseminated tumor cells (DTCs) that have partially overlapping functions and are enriched by the population of quiescent CSCs (12). CTCs, DTCs, and CSCs are able to produce micrometastases that migrate and home inside the target organs in the pre-metastatic niches composed from tumor cells and recruited local stromal and immune cells from the environment. Quiescent CSCs and dormant DTCs inside pre-metastatic niche show overexpression of signaling pathways, enabling them to survive in stressful conditions, including chemotherapeutics (13–15).

Ovarian CSCs are characterized by cell surface CD44, CD117, CD133, CD24, MyD88, epithelial cell adhesion molecule (EpCAM), leucine-rich repeat containing G protein–coupled receptor-5 (LGR5), and LGR6] and intracellular [aldehyde dehydrogenase (ALDH), sex determining region Y-box 2 (SOX2), octamer-binding transcription factor-4 (OCT4), homeobox protein NANOG transcription factor (NANOG), and forkhead box protein M1 (FOXM1)] markers, as well as by their specific behavior (“side-population” cells). The markers for characterization of OCSCs, their function, and clinical significance are presented in Table 1 . The OCSC markers unfortunately are not cancer stem cell specific, as they are also present on normal stem cells. Another feature of OCSCs is activation of signaling pathways upregulating their stemness, cancer proliferative capability, and chemoresistance. The most important and studied pathways for preservation of OCSC function are Wnt/β-catenin, Hedgehog, Hippo/yes-associated protein (YAP), neurogenic locus notch homolog (NOTCH), nuclear factor-κ-light chain enhancer of activated B cells (NF-κB), hypoxia-induced factor-1α (HIF-1α), PI3K/protein kinase B (AKT), Janus kinase (JAK)/signal transducer and activator of transcription protein (STAT), transforming growth factor–β (TGF-β), and Rho/Rho-associated protein kinase (Rho/ROCK) pathways. The functional and clinical characterization of these pathways is included in Table 2 .

Ascites is a unique microenvironment for OCSCs and is responsible for exceptional biology of ovarian cancer, shaped by the transcoelomic spread of peritoneal implants. The EMT process enables the tumor cells from primary localization to seed in the form of multiform cellular conglomerates, mostly adopting the form of spheroids enriched in OCSCs. They are transported in fluid into distant places of peritoneal cavity, with the predisposition to home into the adipose tissue collections inside peritoneum, like “milky spots”, omental fat, mesentery, or bowel appendices (125). Sphere-forming cells express OCSC markers CD44v6, CD117, ALDH1, and NANOG and are resistant to anoikis despite lack of anchorage to the surface (16, 126). The presence of cytokines [interleukin-6 (IL-6), IL-8, IL-10, and vascular endothelial growth factor (VEGF)], osteoprotegerin, and exosomes containing micro RNAs (miRNAs), cytokines, and growth factors further enhances stemness in the spheroids (38, 68, 127–129). Spheroids adhere to and destroy the mesothelium, go through the mesenchymal/epithelial transition, and start to proliferate (130, 131). TGF-β, CD133, and CD44 from spheroids stimulate mesothelium to produce fibronectin for cancer cells adhesion, enhance attachment of floating cells to the epithelial surface, and stimulate secretion of metalloproteinase-9 (MMP-9) that supports mesothelial invasion (108, 132). The initial opinions on random transportation of cellular conglomerates have been replaced by the theory of collective invasion, according to that clusters of cancer cells migrate in a directed and coordinated way (133). Collective invasion is described by some characteristic features, mainly preservation of cell-cell junctions, interaction with surface cells and ECM on their way, cooperative cytoskeleton dynamics enabling migration of clusters as a single unit, and multicellular polarity (120, 133–135). Despite a collective behavior, not all cells in the cluster are invasion-competent, and the population of cells that rule invasion is called “leader cells”. These cells delineate the way, change cellular contractility, and grow invadopodia, as well as respond to environmental signals (120, 136–138). Their presence at the front of the cluster results in its polarization. The coordinated movement requires rearrangement of the cytoskeleton, actinomyosin contraction, and activation of PI3K and Rho/ROCK pathways (120, 135, 139). After adhesion to the mesothelial surface, “leader cells” express proteolytic enzymes and penetrate the basement membrane (120, 140). The phenotype of “leader cells” is characterized by the keratin-14 (KRT14) expression. Their functional phenotype resembles the OCSC phenotype but does not correlate to EMT. The KRT14+ cells are able to re-establish the epithelial cells, show clonogenicity, are abundant in metastases, are enriched in response to chemotherapy, and promote the chemoresistance (120, 140–143). Cancer-associated fibroblasts (CAFs) present in TME play important role in collective invasion by regulation of TME remodeling to “pave” the routs for migrating cell clusters (120, 144). After exposition to chemotherapy, the population of apoptosis-resistant “leader cells” increases and shows expression of ALDH1 and CD44v6 stemness markers together with chemoresistance. Functional impairment of the “leader cells” restores chemosensitivity in vitro (145). After homing into peritoneal environment, OCSCs reside inside the “metastatic niche” composed of several cell populations, ECM components, lipids, exosomes, regulatory RNAs, and hypoxia that are orchestrated to support the OCSCs. Table 3 presents the function and clinical significance of the main components of the TME inside the “metastatic niche”.

The treatment of ovarian cancer is based on debulking cytoreductive surgery, platinum-taxane–based first line chemotherapy, second-line chemotherapy, and targeted therapy approved by the FDA (Food and Drug Administration) and EMA (European Medicines Agency) using bevacizumab and poly-ADP-ribose polymerase (PARP) inhibitors. Several others drugs are being tested in clinical trials including programmed death-1 (PD-1)/ programmed death ligand-1 (PD-L1) inhibitors. HGSOC is initially a chemosensitive tumor, especially in the cases of positive BRCA germinal or somatic mutations. However, recurrent tumors are mostly chemoresistant due to activation of many mechanisms associated with the exceptional function and proliferative activity of OCSCs or reverse BRCA mutations occurring during the treatment. Moreover, the unique pattern of cancer spread inside peritoneal cavity that utilizes both collective invasion and sanguiferous route is relatively early phenomena in the course of the disease. The important obstacle in the effective treatment of HGSOC is also tumor heterogeneity comprehended as spatial heterogeneity in the different areas of the tumor, the inter-patient heterogeneity, and temporal heterogeneity between primary tumors, metastases, and recurrent disease. Even OCSCs themselves exhibit unexpected phenotypic plasticity and may differ in the same patient or among different patients depending on the cancer molecular type, advancement of the disease, patient health, and treatment scheme. The conclusion from those observations is that the use of the uniform treatment for all patients or for all temporal stages of the tumor is an oversimplification that results in observed unsatisfactory results in the context of both OS and PFS. The complexity of interaction between tumor cells, OCSCs, and TME in metastatic niche is another factor of great importance for supporting tumor growth, enhancing chemoresistance and the immune attack defiance. Therefore, tumor environment with all its components should also be treated as a target for anti-cancer therapy.

Taking the abovementioned reflections into consideration, the interesting targets for multidirectional treatment are OCSCs themselves and the components of OCSC microenvironment, particularly metastatic niche. One of the most explored areas of anti-OCSCs therapy is drugs directed against OCSC markers, signaling pathways, and epigenetic regulators. Targeting OCSC markers is important as chemotherapy, whereas decreasing tumor burden simultaneously increases the number of OCSCs. After exposition to chemotherapy, increased numbers of ascitic EpCAM+, CD44+, and OCT4+ cells were noted (248). Similarly, recurrent tumors contain more ALDHA1+, CD133+, and CD44+ OCSCs than primary tumors (49). These phenomena are observed not only in standard platinum/taxol-based chemotherapy but also in the tumors treated with PARP inhibitors (PARPis), where increased numbers of CD133+ and CD117+ OCSCs precede the acquired PARP resistance (249). However, targeting OCSC markers has to overcome two problems. The first one is that OCSC markers are not able to distinguish cancer stem cells exclusively, as about 75% of known cancer stem cell markers are also present on the surface of embryonic and adult stem cells (250). For instance, CD44 is present on hematopoietic cells, MSCs, and adipose-derived stem cells (251–253). CD117 is positive on 25% of embryonic stem cells (254), whereas CD166 is also found on epithelial cells, MSCs, and intestinal stem cells (255, 256). Intracellular cancer stem cell markers, like NANOG, OCT4, and SOX2, are also present in normal stem cells (257, 258). The second problem is associated with the fact that there is no universal cancer stem cell marker known. Tumor heterogeneity, differentiation status, and environment are reasons for OCSC different types. Therefore, the more effective strategy of elimination of OCSCs relies on targeting of at least two OCSC markers simultaneously. Targeting the signaling pathways used by OCSCs is also reasoned by the fact that many of them are likewise OCSC markers, activated after exposition to chemotherapy (259). Epigenetic regulation in ovarian cancer is associated with both hypermethylation and hypomethylation of DNA, as well as with histone methylation and acethylation. Hypermethylation of DNA contributes to formation of OCSCs (260). The CpG islands of many onco-suppressor genes were shown to be hypermethylated in ovarian cancer, leading to the loss of DNA-repair function and cell cycle control desynchronization (261). Upon chemotherapy, hypermethylation of genes responsible for cell resistance to apoptosis was detected (262). Gene hypomethylation is frequently observed in advanced HGSOC and correlates with worse survival (263). Histone methylation is engaged in upregulation of ATP-binding cassette drug membrane (ABC) transporters in chemoresistant OCSCs (264). Disturbed function of histone deacetylases promotes tumor progression (265). Table 4 contains data on both the experimental and clinical trials of targeting OCSCs.

One of the most important targets in TME is CAFs. However, the past experience with anti-CAFs therapy has indicated that the aim in this approach should be to revert CAFs functionally back to normal fibroblasts, rather than eradicating them completely from the TME. Eradication of CAFs has proved to change the tumor into more aggressive phenotype, instead of eliminating tumor cells (350). It is even more important taking into consideration that CAF populations of different tumor-promoting abilities and phenotype (CD49e+, fibroblast activation protein FAP-high or FAP-low) have been identified (351). The reprogramming of M2 tumor-associated macrophages (TAMs), another key population of tumor-supporting cells, into M1 phenotype could be similarly to CAFs, which is a better option than eliminating them completely (352). Another, recently identified population of cells in TME is cancer-associated mesothelial cells (CAMs) that originate from peritoneal normal mesothelial cells activated by cancer-derived promoting factors that induce mesothelial–mesenchymal transition and secretion of factors, enhancing peritoneal metastases and chemoresistance (353). Hepatocyte growth factor (HGF) released from ovarian cancer cells in hypoxic conditions induces the senescence of mesothelial cells and downregulates the expression of junctional proteins that results in disintegration of mesothelial integrity and enables cancer invasion through the mesothelial barrier (354, 355). Phenotypic changes of mesothelial cells to CAMs are mediated by TGF-β and CD44 and annexin A2 secreted inside exosomes from cancer cells (356–358). In response to those changes, CAMs secrete VEGF and upregulate fibronectin expression in ECM, thus promoting tumor vascularization and binding of tumor cells’ integrins to ECM to support metastases (108, 359). Moreover, CAMs increase secretion of IL-8 and CCL2 that stimulate pyruvate dehydrogenase kinase-1 in cancer cells followed by increased expression of integrins to enhance adhesion and migration (360, 361). Interaction between intelectin-1 on CAMs and lipoprotein receptor–related protein-1 on cancer cells also contributes to invasion by upregulation of MMP-1 (362). CAMs pre-stimulated by cancer cell–derived TGF-β secret osteopontin, which, in turn, activates CD44/PI3K/AKT pathway in OCSCs, leading to ABC transporters’ overexpression and chemoresistance (363). M2-shifted TAMs also support CAMs activity by macrophage inflammatory protein-1β that activates P-selectin secretion by CAMs, followed by stimulation of CD24 on the cancer cells’ surface and increased adhesion (364). CAMs are, in turn, able to polarize the TAM phenotype into M2 type (365). CAMs are also capable to regulate the expression of glucose transporter type 4, resulting in increased glucose intake by cancer cells and growth promotion (362). Because of all above functions, CAMs are an interesting target for anti-TME therapy in ovarian cancer.

The next promising target for the therapy is metabolism of cancer cells. Cancer cells use both aerobic glycolysis (the Warburg effect) and oxidative phosphorylation (OXPHOS). Aerobic glycolysis protects cells from oxydative stress and fuels proliferation. However, OXPHOS and resistance to glucose deprivation in tumor environment are a metabolic adaptation enabling chemoresistance. Both ways of glucose metabolism are therefore used by cancer cells, including OCSCs and are another sign of their plasticity (366–368). The metabolic interactions between omental adipocytes and OCSCs are another reason for cancer progression and chemoresistance. Fatty acids could be very efficient source of energy that fuels the spread and growth of peritoneal implants (369). Adipocytes are stimulated by cancer cells to release fatty acids into metastatic niche, and, in turn, adipocytes induce expression of fatty acid receptor CD36 on cancer cells, thus enhancing uptake of fatty acids by cancer (370). Colonization of omental tissue depends on expression of salt-inducible kinase 2 (SIK2) in cancer cells. SIK2 kinase stimulates cell proliferation in PI3K/AKT-mediated manner and enhances paclitaxel resistance in HGSOC cells (371). Moreover, fatty acid oxidase and fatty acid synthase (FASN) have been shown to sustain survival of cancer cells in TME and increase resistance to anoikis and chemotherapy and spheroid formation in HGSOC lines (347, 372). Ovarian cancer CSCs indicate increased concentration of unsaturated lipids and what enhances cell membrane fluidity and facilitates OCSC plasticity and self-renewal. Inhibition of desaturases inhibits spheroid formation and abrogates tumor growth and metastases (373).

Another potential target for anti-TME therapy in HGSOC is exosomes. The identification of their origin inside TME and the recognition of their cargo have the key role in exosome-directed therapy. Exosomes could be also used as potential vehicles for the transportation of drugs into the tumor. It was also found that exosomes secreted from untreated tumors have a significant influence on the expression of many genes involved in functional change of fibroblasts into CAFs and in stimulation of tumor metastases. Such ability was less evident in exosomes secreted by pre-treated tumors (374). The situation is, therefore, complicated, as it seems that exosomes differ depending not only on the type of secreting cell but also on its functional status and temporal changes during therapy. Exosomes are able to influence several mechanisms of tumor growth. Their cargo, including proteins, neoantigens, cytokines, growth factors, and miRNAs, is responsible for cancer progression, metastases, and chemoresistance. Exosomes contain also modulators of immune response capable of inhibition of macrophages: natural killer (NK) cells, dendritic cells (DCs), and B and T lymphocytes (375, 376). Exosomes negatively regulate immunosurveillance of the host against tumor, through inhibition of T lymphocytes, NK cells, DCs, and monocytes in tumor environment and ascites (227, 377–379). Exosomes stimulate tumor angiogenesis affecting the VEGF and HIF-1α expression and by activation of Wnt/β-catenin and NF-κB signaling pathways (380, 381). Exosomes influence also stroma remodeling by cooperation with CAFs and adipocyte-derived stem cells (165). Recently, tumor-derived exosomal miR-141 was identified as a regulator of stromal-tumor interactions and inducer of tumor-promoting stromal niche by activation of YAP/chemokine (C-X-C motif) ligand 1 (GROα)/CXCR signaling pathway (382). One of the most interesting vectors of information between cancer cells and TME is non-coding miRNAs and long non-coding RNAs (lncRNAs). They were found in the serum and ascites of patients with ovarian cancer (227, 228); however, their presence in tumor-derived exosomes ensures safe and undisturbed transportation to the target cells. Non-coding RNAs play extremely important functions. Exosomes loaded with miR-1246 are able to enhance pro-tumorigenic effects of M2-shifted TAMs and to facilitate paclitaxel resistance (383). Cancer cell–derived miR-21-3p, miR-222, miR-125b-5p, miR-181d-5p, and miR-940 target TAMs and polarize them into M2 phenotype (172, 384). miR-99a-5p affects human peritoneal mesothelial cells and enhances cancer cell invasion (385). The Let-7a and miR-200a regulate tumor invasiveness (386). Exosomes containing lncRNAs ENST00000444164 and ENST0000043768 are responsible for activation of NF-κB signaling in cancer cells (387). Table 5 presents data on targeting the components of TME and OCSC niche.

The urgent need for improvement of efficacy in the HGSOC treatment is obvious, and many researchers have called attention to the novel approaches in diagnosis, monitoring, and management of patients with ovarian cancer. We have learned from the experience from therapy of hematologic cancers and several solid tumors that the individual approach to the treatment based on genetic, molecular, or metabolic signatures of the patients and the cancer itself usually results in better treatment efficacy and improved outcome. However, such individualization of therapy is much more difficult to be used in solid tumors, compared to hematologic malignancies, and ovarian cancer due to its unique biology is even more demanding and challenging target.

In the recent article devoted to OCSCs and OCSC-targeted treatment (470), we proposed that the novel complex standard of ovarian cancer therapy called the “DEPHENCE” system (“Dynamic PHarmacologic survEillaNCE”) should be worked out. In our opinion, it ought to be based on the following rules:

We think that the necessary components incorporated into the DEPHENCE system should also be

The practical implementation of the “DEPHENCE” system in the diagnosis and therapy of ovarian cancer is still awaiting, although the first signs of its use can be seen in the attempts to classify the molecular signatures of the tumors and TME components (158, 458, 471–476), to personalize therapy according to the tumor origin, histology, and most of all to genomic and epigenomic disturbances. The first such studies grouped HGSOC tumors T into four subtypes: C1, high stromal response; C2, high immune signature; C4, low stromal response; and C5, mesenchymal, with low immune signature. These subtypes differed in the extent of immune infiltration, desmoplasia, and EMT predisposition, and what could suggest different approach to the treatment, including immunotherapy, and patients from the C1 and C5 subtypes showed poor survival compared with other subtypes (3). Another genomic classification was proposed by The Cancer Genome Atlas Research Network, which, based on the genomic pattern, divided the ovarian cancer into four subtypes: mesenchymal, immunoreactive, proliferative, and differentiated. Mesenchymal and proliferative subtypes showed profound desmoplasia and invasive gene expression pattern, with limited immune infiltration and activation of stemness markers. Both were characterized by unfavorable prognosis. Immunoreactive subtype showed extensive immune infiltration and, similar to differentiated more mature tumors, had better prognosis (477–479). The next analysis of tumor genome identified three novel ovarian cancer subtypes named tumor-enriched, immune-enriched, and mixed. The meaning of these subtypes for therapy implies that tumor-enriched tumors should be treated with tumor killing therapy, whereas immune-enriched tumors with immunotherapy or mixture of both approaches (480). Molecular characterization of platinum-refractory and platinum-resistant ovarian tumors identified three tumor clusters: cluster 1 with overrepresentation of growth factor signaling pathways, cluster 2 with pathways regulating cell survival in hypoxic conditions and senescence, and cluster 3 related to cellular senescence. A possible treatment of choice for cluster 1 could be tyrosine kinase or angiokinase inhibitors, cluster 2 could theoretically response to mTOR inhibitors, whereas cluster 3 could be treated with the deacetylase inhibitors (87, 481, 482). Another single-cell transcriptome study revealed the heterogeneity of HGSOC, which was found to be composed of several cell clusters. The first one called EC1 showed gene enrichment for glycolysis/gluconeogenesis and ECM-receptor interactions. The EC2 subtype expressed genes, suggesting their origin from tube epithelium. The EC3 subtype showed overexpression of genes associated with function of ABC transporters, suggesting a potential to be a drug-resistant subtype. EC4 subtype was characterized by the immune response-related pathways indicating the activity of EC4 cells in immune response. The chemoresistance responsible genes were strongly represented in EC5 cell population (483).

Epigenomic analysis of immune-related lncRNAs revealed RNAs having the potential to divide the population of patients with ovarian cancer into high-risk and low-risk groups characterized by a shorter or longer overall survival (OS), respectively. High-risk score tumors were positively correlated with abundant representation of checkpoint and immunosuppressive molecules, indicating the group of patients with compromised anti-tumor immune response (484). The DNA methylation signatures represent another epigenetic point of interest in ovarian tumors. The hypomethylated upregulated tumor necrosis factor (TNF), estrogen receptor 1 (ESR1), mucin 1 (MUC1) genes, and hypermethylated downregulated forkhead box O1 (FOXO1) gene could serve as targets for epigenetic therapy and were correlated with patients’ prognosis (485).

According to the TME components, the four different CAF subsets (S1 to S4) were identified in ovarian tumors. The HGSOC of mesenchymal subtype, defined by stromal gene signatures and poor survival, had high numbers of CAF-S1 cells, which attracted and sustained immunosuppressive infiltration of Treg CD25+FoxP3+ T lymphocytes (475). The study of immunological profile of HGSOC showed the presence of activated‐immune and CAF‐immune subtypes. Activated-immune subtype showed anti-tumor features exemplified by active immune response and better prognosis. The CAF‐immune subtype was characterized by tumor‐promoting signals like, activated stroma, M2 macrophages, and a poor prognosis. The activated‐immune subtype was more likely than the CAF‐immune subtype to respond to checkpoint blockade immunotherapy (486).

The most painful problem in ovarian cancer therapy is the acquired chemoresistance following the initial good response to the first-line chemotherapy. Therefore, identification of the biomarkers of chemoresistance is one of the most important activities in ovarian cancer surveillance. The classic biomarkers of platinum and PARP chemosensitivity are the germinal and somatic mutations of BRCA1/2 genes (487). However, the reversion mutations in BRCA genes and in other homologous recombination repair (HR) genes were found to be responsible for secondary resistance to platinum- and PARPi–based therapy (488, 489). On the basis of the homologous recombination deficiency, insertions and deletions, copy number changes, and mutational signatures, a combined predictor of platinum resistance, named DRDscore, was established, and, when validated in a cohort of patients with HGSOC, it reached sensitivity of 91% (490). Four miRNA biomarkers (miR-454-3p, miR-98-5p, miR-183-5p, and miR-22-3p) identified in ovarian cancer tissues were able to discriminate between platinum-sensitive and platinum-resistant patients with HGSOC (491). Treatment using PARPis results in acquired PARPi resistance. The reason for this is a promotion of STAT3 activity both in tumor cells and populations of immune and CAF cells, followed by creation of an immunosuppressive environment. Treatment of olaparib-resistant ovarian cancer cell line with napabucasin, the STAT3 inhibitor, improved PARPi sensitivity (492). Hypoxia and therapy-induced senescence are the key drivers of primary chemo-refractoriness and secondary chemoresistance of HGSOC (493). Hypoxic TME induces the M2-phenotype in TAMs, which, in turn, secrete exosomes containing miR-223 that, when transported into ovarian cancer cells, makes them chemoresistant (494). To overcome chemoresistance, there are plenty of different drug combinations tested in both experimental and clinical settings ( Tables 4 , 5 ). Simultaneously, identification of potentially resistant tumors is of the utmost importance for successful therapy. Identification of ovarian cancer cells with high-stress signature and disturbed drug responsiveness could optimize the subsequent therapy to attenuate their function or eliminate them from the tumor (493, 495, 496). Moreover, as HGSOC tumors are characterized by temporal heterogeneity, the repetitive circulating tumor DNA (ctDNA)/CTCs testing should be performed to have the most actual picture of the disease.

The exploration of the infection factors in the origin or predisposition to ovarian cancer is also being realized in the analysis of microbiome and viral infections (497–499). Another field of intensive investigation is searching for prognostic biomarkers (500–503). It is a lot of work to do to safely and effectively combine different drugs, but the practical use of the “DEPHENCE” system philosophy could, in our opinion, lead doctors and researchers in proper direction.

All authors contributed to the conception of the review. Data collection and analysis were performed by JW, MW, and EP. The first version of manuscript was written by JW and MW, and all authors commented on previous versions of the paper. EP and MW read and approved the final version of the manuscript, made linguistic corrections, and revised the text. All authors contributed to the article and approved the submitted version.