This article is mentioned in:

Introduction

Ovarian cancer (OC) is a malignant tumor that poses a significant threat to women's health. Globally, OC is the third most prevalent gynecological malignancy, yet it is associated with the highest mortality rate among these types of cancer (1). In 2022, it was estimated that 324,398 new cases of OC would be diagnosed and 206,839 related deaths would occur worldwide (2). OC is often diagnosed in its advanced stages, with current screening techniques unable to facilitate an earlier diagnosis (3); thus, the overall 5-year relative survival rate for OC is 50% (4). Delayed clinical presentation and a late diagnosis also contribute to the high mortality rate associated with this disease, mainly due to a significant lack of sensitive and specific biomarkers for early-stage disease (5).

Currently, the standard chemotherapeutic regimen for OC commonly employs a combination of taxane-based (such as paclitaxel) and platinum-based drugs (6). Cisplatin was the first platinum compound used in the primary treatment of OC; however, it is associated with dose-limiting toxicity, including minor symptoms such as nausea, severe kidney impairment and peripheral neuropathy. To overcome the toxicity associated with cisplatin use, organic platinum analogs of cisplatin, such as carboplatin, have been developed and used instead of cisplatin in initial OC chemotherapy (7). In patients with advanced disease, carboplatin plus paclitaxel chemotherapy is considered an optimal choice (8). Over time, chemotherapy for OC has resulted in progressive enhancements in survival, and the use of platinum/paclitaxel combinations, particularly in combination with intraperitoneal delivery, has the potential to double or increase survival time (9). In addition, heat-intraperitoneal perfusion chemotherapy may improve the disease-free survival of patients with OC, especially when the residual tumor is ≤1 cm or invisible, but it may only improve overall survival (OS) in patients with recurrent disease with a residual tumor of ≤1 cm (10).

Several challenges remain in the treatment of OC, despite advances in chemotherapy. For example, all forms of maintenance therapy have thus far failed to prolong survival. Intraperitoneal therapy has been shown to improve survival in small-volume stage III disease but is associated with significant toxicity (11). Furthermore, a substantial proportion of patients diagnosed with advanced OC are at an increased risk of recurrence (12) and drug resistance (13), which presents a significant challenge for clinicians in terms of developing effective treatment strategies. It is estimated that 60–80% of patients with epithelial ovarian carcinoma (EOC) will achieve complete remission following a combination of surgical and postoperative chemotherapy treatments; however, ~80% of these patients face the risk of chemotherapeutic resistance and potential recurrence (14).

With advancements in technology and research, organoid models have been shown to demonstrate notable improvements over traditional models in multiple aspects, primarily benefiting from their ability to more accurately simulate the complexity and dynamics of biological systems. Traditional two-dimensional (2D) culture models often overlook the three-dimensional (3D) structure of cells and the interactions between cells (15), and xenograft models have drawbacks such as immune rejection and species differences (16). Although patient-derived xenograft (PDX) models retain the tumor characteristics of patients (17,18), they are costly and technically complex to establish and maintain (19). By contrast, organoids, which are 3D dynamic tumor models that can simulate tumor environments in vitro, can mimic the disease development process (20). Organoids can overcome the limitations of traditional models, and thus provide new perspectives for OC research and treatment. Thus, there is an urgent need to develop new in vitro cultivation techniques that can mimic the development of the disease and produce organoids, and can be successfully grown from ovarian tumors, ascites or pleural fluid (21). As a novel approach to cell culture, organoid technology offers significant potential for developing individualized treatments and screening anticancer drugs. This is due to the high degree of tumor replication, the preservation of tumor heterogeneity, the short growth cycle and the stable transmission observed in organoids, which collectively address the limitations of existing preclinical models, including cell lines and animal models (22). The efficacy and resistance of chemotherapeutic drugs can be more accurately predicted through organoid culture, providing a basis for individualized treatment; therefore, organoids may be considered a convenient and efficient platform for studying drug resistance mechanisms (23).

The objective of the present review was to present a thorough overview of the definition, formation and construction of OC organoids. Additionally, the current review considers the advancement of OC organoids in chemotherapy, and their applications, deficiencies and prospects.

Overview of OC organoids

Organoids represent a 3D in vitro cell culture system (23), which are derived from cellular or organ progenitors and, through in vitro differentiation, develop functional cell clusters that display the essential properties of their organ of origin (24). Furthermore, they can preserve the heterogeneity of the original tumor and mimic the tumor growth microenvironment (23); they are thus considered a powerful model for simulating human disease.

In a 2D cell model, cells are grown as a single cell layer in a petri dish. Notably, 2D cell culture models cannot accurately represent the heterogeneity of tumors since they lack the inherent complexity and the TME of the original tumor (25). In addition, researchers have transplanted patient-derived subcutaneous or in situ tumor cells into immunodeficient mice to simulate the heterogeneity of the original tumors; however, this PDX model is expensive, time-consuming, and has low transplantation rates (26,27) and ethical drawbacks (28), Patient-derived organoids (PDOs) have emerged as a pivotal preclinical model system in the field of cancer research and clinical studies (29,30). Organoids retain a high degree of genomic similarity to their corresponding tumors; moreover, they can exhibit the inheritance of certain diseases and the ability to epigenetically alter their phenotype, which may facilitate the development of specific interventions to promote personalized medicine (31,32). The primary advantages of in vitro organoids over other in vitro tumor models are: i) They retain a high degree of heterogeneity among tumor cells, ii) they can be cultivated in small samples, and iii) they maintain the ability for long-term expansion, cryopreservation and genetic modification (24). The advantages and disadvantages of in vitro models are shown in Table I.

Table I. Advantages and disadvantages of in vitro models.

Table I.

Advantages and disadvantages of in vitro models.

Model	Advantages	Disadvantages	(Refs.)
2D cell model	Straightforward and easy to use; low cost	Lack of tumor complexity; poor expression of tumor heterogeneity; lack of immune cells and tumor microenvironment	(19)
PDX model	Preservation of tumor heterogeneity; mimics original tumor characteristics	Time-consuming; low transplantation rate; ethical issues; cost issues	(20–22)
Organoid	Maintenance of high tumor heterogeneity; high genomic homogeneity; possible to culture small samples; retention of long-term amplification, cryopreservation and genetic modification	Technical limitations; individual differences; standardization issues; cost issues	(23–25,27,28,88–97)

[i] PDX, patient-derived xenograft; 2D, two-dimensional.

In order to generate organoids, tumor cells isolated from primary tumor material, such as surgically excised tissues, biopsies, ascites, pleural fluid or circulating tumor cells, need to be capsulated in a 3D matrix and cultured in medium containing variable growth factors and hormones (33) (Fig. 1). The composition and concentration of certain substances in the medium vary depending on the specific experiment or study; however, advanced Dulbecco's Modified Eagle Medium (DMEM)/Nutrient Mixture F-12 (F-12) is widely used (34,35). DMEM and F-12 are often used in combination to obtain higher concentrations of components of DMEM and a wider variety of components in the F-12 nutrient mix. DMEM/F-12, however, is free of growth factors, hormones and lipids; therefore, to achieve optimal cell proliferation when using DMEM/F-12, it is necessary to add an appropriate combination of growth factors, hormones, proteins or peptides, depending on the specific research needs and cell type. The mixture of growth factors utilized for OC organoid propagation is not standardized and may vary between studies. For OC organoids, it has been suggested that R-spondin 1 (an activator of the Wnt pathway), Noggin (an inhibitor of BMP-dependent differentiation) and 17β-estradiol are necessary for growth, while neuromodulin-1 (NRG1) may be needed for organoid expansion (36).

Figure 1. Formation of organoids from OC. Tumor cells from ovarian samples were mixed with ECM, aggregated into spheroids and placed in a medium with R-Spondin 1, Noggin, 17-β-estradiol, NRG1 and advanced DMEM/F-12, ultimately forming a fully developed OC organoid. ECM, extracellular matrix; NRG1, neuregulin 1; OC, ovarian cancer.

The development of OC organoids from multiple stages and subtypes (56 organoid lineages) has previously been successfully achieved (37). Organoid cultivation techniques can be applied to a wide range of patient samples, including tissues, biopsies, ascites and pleural effusions (38–40). Most OC organoids generated thus far have been high-grade serous ovarian carcinoma (HGSOC), but there are also other types of cancer, such as low-grade serous OC, mucinous OC, ovarian clear cell carcinoma (OCCC) and ovarian endometrioid carcinoma (ENOC) (41). Furthermore, the proliferation of primary ovarian organoids from different histological subtypes, including HGSOC, OCCC and ENOC, has been reported to have a success rate of 80% and to retain the original mutation pattern (42). Moreover, OC organoids are suitable for constructing biobanks to further improve organoid construction efficiency. For example, a study by Hoffmann et al (43) successfully generated 15 organoids (with an efficiency of ~30%) from 13–45 patients under conditions of low Wnt expression for long-term passaging culture. The OC organoids were cryopreserved for >5 months and passaged every 10–20 days. This previous study showed that there were no significant changes in the morphology or proliferation capacity of the organoids even after repeated thawing and freezing cycles.

OC organoid model applications

Application of OC organoid models for chemotherapeutic drug screening

Drug screening, which refers to testing the effects of a variety of drugs on several cell lines, is used to identify and develop medicines (44). In cancer therapy, drug screening aims to screen numerous compounds or new compounds for specific physiological activity against tumor cells. The present review aimed to summarize the differences in drug screening methods and their application in tumor therapy. For this purpose, models such as 2D cell cultures, multicellular tumor spheroids (MTSs) and organ tissues were assessed (Table II).

Table II. Multi-model perspective on ovarian cancer drug screening and sensitivity applications.

Table II.

Multi-model perspective on ovarian cancer drug screening and sensitivity applications.

A, Drug screening
Model	Key findings	(Refs.)
2D cell cultures	Prediction of drug activity, metabolism and toxicity; only 10% of drugs tested in vitro are clinically effective and only 5% of anticancer drugs; only 12 of 900 drugs identified were approved by the Food and Drug Administration in 2011	(45,50,52)
MTSs	Potential of screening antitumor drugs; MCTSs demonstrate increased resistance to chemotherapeutic agents and the ability to screen for active drug candidates	(61)
Organoid	Improvement of antitumor activity effect; the potential for higher growth rates, faster drug response and better drug penetration; efficient screening in a passive microfluidic platform	(62)
B, Carboplatin
Model	Key findings	(Refs.)
2D monolayer, 3D spheres, 3D tumor ex vivo, mouse xenografts	3D sphere and mouse xenograft models best correlate with in vivo response; Variability in therapeutic response across models	(40)
2D monolayer, 3D spheres, mouse xenografts	3D sphere systems are in best agreement with in vivo models; 2D super-resistant cell lines are more sensitive in 3D and mouse models vice versa	(65)
Four cell lines, two PDOs, one PDX	PDX and PDO models reproduce patient tumor heterogeneity; Resistance to standard therapeutic drugs; Consistent with the clinic, PDO provides support for therapeutic decisions	(66)
C, Paclitaxel
Model	Key findings	(Refs.)
2D monolayer versus 3D hydrogel	3D sphere survival (40–60%) is better than 2D single layer survival (20%)	(67)
3D spheres (methylcellulose)	Spheres maintain cell density, and exhibit higher apoptotic behavior and higher paclitaxel resistance	(68)
PDO	PDO cell viability: 48.7%	(69)
D, Bevacizumab (Avastin®)
Model	Key findings	(Refs.)
Mouse xenograft model versus PDO	Mouse xenograft model may prolong survival and reduce ascites formation; models in vivo better reflect actual treatment effects, and PDO may be used for initial drug sensitivity screening; low effect on cell viability when used as a single agent; possibility of improvement of therapeutic effect when used in combination with chemotherapy	(69,73)
E, Olaparib
Model	Key findings	(Refs.)
Monolayer AsPCs versus 3D AsPCs	Monolayer AsPCs responded significantly more than 3D AsPCs to PARPis treatment	(41)
PDO	Only 2 of 33 organoid cultures tested (6%) were sensitive to olaparib	(41)

[i] 2D, two-dimensional; PDX, patient-derived xenograft; MTSs, multicellular tumor spheroids; MCTSs, multicellular tumor spheres; PDO, patient-derived organoid; AsPCs, ascites-derived primary cell cultures; PARPis, poly(ADP-ribose) polymerase inhibitors.

Notably, 2D cell cultures are the mainstay of fundamental cellular research, and are used to predict drug activity, metabolism and toxicity in vitro (45). In cancer screening, cytotoxicity testing is often performed on cultured tumor cell lines in monolayer culture due to the rapid and uncontrolled growth of these cell lines (46). By contrast, the main disadvantage of 2D systems is their inability to reproduce complex 3D structures or to communicate with the TME (47,48). Thus, 2D culture models may differ greatly from growing tumors with regard to cell morphology, proliferation, and gene and protein expression (49). Therefore, only 10% of the drugs that pass in vitro testing positively impact the clinic or lead to approval. The proportion of anticancer drugs that demonstrate clinical efficacy is even lower, at ~5% (50). Furthermore, a number of drugs reported to have potent anticancer effects in 2D cell culture models have failed clinical trials (51). In 2011, ~900 anticancer drugs passed cell.based tests, but only 12 were approved by the Food and Drug Administration after clinical testing (52).

The cells in MTS models more closely mimic the cell shape and environment in vivo than monolayer culture models (53,54); however, the size of spheroids has a substantial influence on their survival and reaction to drugs (55). Spheroids <100 µm do not reveal the complexity of the tumor or tissue (56). On the other hand, spheroids with diameters in the range of 400–500 µm can reflect the behavior of tumors in vivo due to the presence of different cellular layers: A necrotic core surrounded by a region of quiescent cells and an outer region composed of proliferating cells. The peripheral cells may reflect the in vivo tumor situation next to the capillaries, while the distant internal cells remain quiescent or die due to apoptosis or necrosis (57). This may also be why a number of reports (58–60) have shown that antitumor drugs are less effective against 3D cultured cancer cells than against 2D cultures.

It has also been shown that the MTS model has the potential to screen for antitumor drugs. Hirst et al (61) previously grew a series of EOC cell lines in 3D cell cultures to form multicellular tumor spheres (MCTSs). These MCTSs were characterized on the basis of the molecular and cellular properties of EOC, and were screened against cells grown in 2D cell culture to identify previously underappreciated anticancer agents. The MCTSs demonstrated enhanced resistance to chemotherapy, showed symptoms of senescence and hypoxia, and expressed numerous stem-cell-related transcripts, including ALDH1A and CD133 (PROM1). This previous study identified licofidone as a candidate with more activity in the MCTSs than in the 2D-cultured cells using a clinically repurposed drug library. Licofidone has been shown to be synergized with paclitaxel in an ovarian MCTS model, as well as in a patient-derived tumor xenograft model. The combination of licofidone and paclitaxel significantly increased the median survival time (>141 days) in mice compared with paclitaxel (115 days), licofidone (37 days) or vector (30 days). In addition, the Mantel-Haenszel hazard ratio (HR) was confirmed to be superior to the vector (HR=0.037) and paclitaxel (HR=0.017). The results of this previous study identified the potential use of an underappreciated anti-inflammatory agent in OC treatment.

Although the MTS model has demonstrated its unique advantages in antitumor drug screening, most identified drugs have failed in clinical applications due to the differences between MTSs and the real in vivo environment. Therefore, researchers have begun to explore OC organoid models. Cavarzerani et al (62) derived organoids from the ascites or tissues of patients with OC. In this previous study, the development of an organoid culture model with a higher growth rate, a faster drug response and better drug penetration into the extracellular matrix (ECM) was revealed to be possible in a passive microfluidic platform that maintained the vital signs of the sample and collected data for up to 16 medicines on a single plate. Achieving successful cultivation of organoids originating from patients with HGSOC was considered feasible in Mimetas Dual Lane OrganoPlates®. Thus, OC organoids may potentially be used to assess the antitumor activity of drugs.

OC organoids in chemosensitivity testing

Organoids can also be applied to chemosensitivity testing. In vitro susceptibility and resistance tests are used to determine whether a sample of a patient's tumor tissue exhibits a response (i.e., a reduction in tumor survival) when exposed to a selected chemotherapeutic agent under laboratory conditions (63). Drug response measurements aim to assess the therapeutic effectiveness of a drug across a specified concentration range. Wherever possible, multi-drug response metrics should be used to account for possible experimental variations during the measurement, initial population and cell division count, including: Emax (maximal effect of the drug), EC50 (drug concentration that reaches Emax), IC50 (median inhibitory concentration), GI50 (concentration that decreases the overall cell growth by 50%), GR50 (50% inhibition of cell growth rate) and AUC (area under the dose-response curve, which represents the cumulative effect of the medicinal product) (64). Therefore, a multi-model perspective offers a comparative analysis of sensitivity testing in OC (Table II).

Carboplatin

Carboplatin is the most common first-line chemotherapy agent in the treatment of gynecological malignancies. The sensitivity of ovarian epithelial adenocarcinoma cell lines to carboplatin chemotherapy in vitro varies in 2D or 3D models (65). Compared with other ex vivo and in vivo models, the response of the 3D system to carboplatin has been revealed to be in best agreement with the in vivo model. To assess the current differences in treatment response in in vitro and in vivo models, and to deal with this problem, a comparative study was conducted by Maru et al (40) to distinguish the response of four different models to carboplatin chemotherapy, including 2D single layer, 3D sphere, 3D tumor ex vivo and mouse xenograft models. Six previously characterized EOC cell lines with different chemosensitivities were used and viability tests were carried out on each model. The in vivo results in the mouse model correlated with the 2D response in 3/6 cell lines, while they correlated with the 3D sphere response in 4/6 cell lines and the 3D ex vivo model in 5/5 cell lines. These results highlight the variability of treatment responses across models, and indicated that the response of carboplatin was best correlated with the response of EOC cell lines grown in a 3D ex vivo model.

Similarly, Brodeur et al (65) investigated the response of single-layer and spherical (3D) EOC cells to carboplatin, and then compared it with the in vivo response (xenografts). Mice were injected with six ovarian epithelial adenocarcinoma cell lines that received three different concentrations of carboplatin. Their responses were evaluated and classified based on tumor volume and immunofluorescence measurements. The same ovarian epithelial adenocarcinoma cell lines were seeded onto shallow adhesion plates to form spheroids and were processed. Flow cytometric analysis was performed to classify ovarian epithelial adenocarcinoma cell lines based on IC50; this was compared to previously published 2D IC50 results. The results showed that the 3D system was in the best agreement with the in vivo model; in particular, the 2D super-resistant ovarian epithelial adenocarcinoma cell lines became more sensitive in either the mouse model or the 3D system. By contrast, the ultrasensitive ovarian epithelial adenocarcinoma cell line in the 2D system was more resistant in xenografts and spherical tissues.

Additionally, Thorel et al (66) established seven models (four cell lines, two PDOs and one PDX), all derived from the same OCCC. In order to establish their relevance, a comprehensive characterization was conducted using morphological, histological and transcriptomic analyses, as well as an evaluation of the patient's response to treatment, which was compared with the patient's chemotherapeutic response and showed resistance to drugs, carboplatin, gemcitabine and doxorubicin. Only PDX and PDO models derived from tumors could replicate the heterogeneity of patients. Patients were refractory to carboplatin, doxorubicin and gemcitabine, whereas cancer cells were susceptible to these agents. The PDX and PDO models, on the other hand, were resistant to these three drugs. Transcriptome analysis was consistent with these results, as models that faithfully reproduced the clinical response differed from the classical 2D cell culture models. Subsequently, the researchers examined the potential of previously unused drugs and determined that the histone deacetylase inhibitor belinostat was a potential therapeutic agent based on the response of the PDO. These findings indicated that PDX and PDO models may be the most relevant; however, only the PDO model offered all of the necessary preconditions for prediction and could be used in conventionally refractory types of cancer, especially aggressive ones.

Paclitaxel

Paclitaxel is a commonly used medicine for the treatment of OC. In a previous study, 3D patient-derived tumor spheroid structures were tested for paclitaxel susceptibility in comparison to 2D monolayer cultures. Loessner et al (67) performed an extensive study of EOC using hydrogel as a 3D model. The two cell lines (OV-MZ-6 and SKOV-3) were revealed to exhibit similar proliferation in 2D cultures, but different in 3D cultures. Spherical cell formation could only be observed in 3D cultures when the cells were embedded in hydrogels. The proliferation of OV-MZ-6 and SK-OV-3 cells in polyethylene glycol hydrogel possessing arginyl-glycyl-aspartic acid functionalized sites with glutamine, a general nutritional addition to the cell culture process, and matrix metalloproteinase sensitive sites for 14 days, followed by paclitaxel treatment, demonstrated a higher survival rate (40–60%) 7 days post-treatment compared with the single cells (20%) in 2D culture. Thus, 2D chemosensitivity evaluation does not necessarily reflect patient pathophysiology.

In order to determine the optimal conditions for analyzing the drug response of SKOV-3 and OVCAR-3 spheroids with the expected properties (roundness approaching 1.0; diameter in the range of 200–500 µm) to paclitaxel, two methods were used to evaluate the spheroids in two 3D cultures, after the addition of methylcellulose (0.25 and 0.5%, w/v) to the culture medium. Compared with 2D culture, SKOV-3 and OVCAR-3 spheroids retained cell density, increased apoptosis and increased resistance to paclitaxel treatment. The results indicated that 3D cultures may provide more reliable drug response results than 2D monolayer cultures (68). Meanwhile, Bi et al (69) assessed paired tumors and adjacent normal tissues from the same patient. At tumor-killing doses, they reported minimal cell destruction with variable chemotherapeutic agents. The results indicated that, in comparison with other drug treatments, the OC organoids of the patient were more sensitive to paclitaxel, and the survival rate was 48.7% after exposure to this single agent. The patient OC organoids were also more sensitive to gemcitabine (65% decrease in viability) and topotecan (56% decrease in viability) than the previously received chemotherapy regimen carboplatin + paclitaxel. Thus, organoids were considered more sensitive to paclitaxel compared to 2D monolayers.

Other drugs

Bevacizumab (Avastin®) is a recombinant humanized monoclonal antibody against vascular endothelial growth factor (VEGF) (70), which binds to and neutralizes the biological properties of human VEGF by blocking VEGF interaction with its receptor (71). Bevacizumab is frequently used as an adjuvant treatment in gynecological tumors (72). In mice bearing HOC22 ×enografts, OS and complete remission were markedly prolonged when bevacizumab was used in combination with chemotherapy and continued at the end of chemotherapy. Bevacizumab, on its own, inhibited the formation of ascites and had a limited impact on tumor load (73). Bi et al (69) also tested the sensitivity of PDOs to either bevacizumab alone or in combination with first-line standard chemotherapy. Overall, bevacizumab as a monotherapy had a minimal effect on cell viability and, when combined with chemotherapy, had little additive effect on cell killing

Olaparib was the first approved poly (ADP-ribose) polymerase inhibitor (PARPi) used to treat patients with advanced BRCA-mutated OC (70,72,74). PARPis inhibit PARP in cells with BRCA-mutated tumors; they can cause ‘synthetic lethality’ by targeting two DNA repair pathways at the same time and do not affect normal cells (73), Hill et al (41) formulated a 3D spherical functional evaluation methodology aimed at quantifying the responsiveness of two PARPi agents, specifically niraparib and olaparib, within ascites-derived primary cell cultures (AsPCs) derived from patients diagnosed with HGSOC. An agarose-based protocol for AsPC preparation enabled efficient isolation and propagation of monolayer and 3D AsPCs. The results showed a distinct sensitivity difference to PARPis, with monolayer AsPCs exhibiting 88 and 52% sensitivity to niraparib and olaparib, respectively, compared with 66 and 38% sensitivity in 3D AsPCs. The latter aligns with prior homologous recombination deficiency (HRD) estimates in EOC (40–60%), emphasizing the relevance of 3D models.

Application of OC organoid models in individualized therapy

The application of organoids in personalized medicine has also begun to be implemented (74,75). The development of personalized medicine is about preventing, diagnosing and treating pathological conditions based on individual characteristics (38). The characteristics of the individual include the entire body, tissues and cells, including genetic, epigenetic, transcriptomic, proteomic and metabolomic markers (76). In addition, gene expression profiles can be studied and changes associated with tumors can be identified (77). Personalized medicine obtains patient-specific disease information to select drugs and the correct dosage, and to optimize therapeutic procedures (78). Personalized medicine also influences preclinical studies of drugs, not just regarding efficacy and safety, but also to reduce drug development costs (38). OC encompasses a diverse array of histotypes, each characterized by distinct origins, underlying genetic and epigenetic alterations, and varying survival outcomes. Moreover, the differing prevalence of these histotypes across various ethnicities adds another layer of intricacy (79). In light of these multifaceted variables, the use of traditional 2D cell lines, albeit immortalized and homogeneous, as in vitro models of OC, are inherently limited in their capacity to capture the complexity of the disease at any single time point or in a microenvironment that accurately mirrors the in vivo scenario. Consequently, organoid technology has emerged as a promising avenue for advancing cancer research and facilitating the development of tailored therapeutic strategies (80). In clinical practice, OC organoid models have demonstrated their unique value in personalized therapy. For example, Lui et al (81) provided prospective and retrospective evidence from ≥18 cases, which indicated that organoid drug screening can accurately predict clinical response to chemotherapy and targeted therapy. This previous study also reported on a patient with platinum-resistant plasma OC who responded to ibrutinib therapy after screening, with this drug identified as having an excellent response in their organoid. After 3 months of treatment, the CA125 levels of the patient had decreased from 250 to 125 U/ml.

Preclinical studies

It is becoming increasingly apparent that standard treatment does not apply to every patient and that it would be helpful to use pretreatments to provide a more individualized regimen. Drug testing of PDOs can be executed in a timely manner; in addition, PDOs have the potential to identify the most efficacious drug in each case, and to offer a promising preclinical platform for tailoring personalized cancer treatment strategies in patients with gynecological malignancies (69).

For example, in 2019, Phan et al (82) pioneered the use of PDOs from HGSOC and ovarian sarcoma, using an automated screening system to assess the unique responses of these organoids to a panel of 240 kinase inhibitors. The high-throughput methodology not only identified sensitive signaling pathways but also facilitated the selection of the most efficacious drugs within specific molecular classes. This previous study presented a scalable and functional precision medicine platform titled Drug Efficacy Testing in 3D Culture, which can quantify the responsiveness of patient-derived cells to various drugs and drug combinations through live-cell imaging. Employing this platform, 27 tailored drug combinations were evaluated, achieving results within 10 days post-surgery. Notably, the findings revealed that the combination of carboplatin and a Bcl-xL inhibitor exhibited synergistic effects in 4 out of 8 OC samples tested, whereas afatinib and the Bcl-xL inhibitor A-1331852 demonstrated synergism in 5 of 7 OC models. These results suggested that combining the Bcl-xL inhibitor A-1331852 with either afatinib or carboplatin may hold promise as a potential future therapeutic strategy for patients with OC (83).

Biomarkers

Predictive biomarkers of therapeutic response are vital for assessing clinical benefit in patients with OC, yet they remain an unmet need (84). Notably, the potential of RAD51 foci as predictive biomarkers in HGSOC warrants validation in clinical trials (85). RAD51 has emerged as a prime candidate biomarker for such assessments due to several reasons. Firstly, upon the occurrence of DNA double-strand breaks, the kinase ATM swiftly phosphorylates histone H2AX (ᵧH2AX), initiating a cascade that leads to the formation of single-stranded DNA with 3′ projections. These, in turn, engage replication protein A, which RAD51 subsequently displaces via BRCA1 and BRCA2, facilitating homologous recombination (HRec). Given the multitude of upstream events required for RAD51 binding and its central role in HRec, RAD51 provides a holistic readout of the intricate steps of HRec. Secondly, RAD51 binding results in the formation of nucleoprotein filaments that can be visualized as foci under a microscope, with RAD51 focus deficiency serving as a functional marker for HRD (86–88). Thirdly, RAD51 foci detection has been shown to possess predictive value for platinum chemotherapy and PARPi responses in OC xenografts (89–91), correlating with in vitro platinum responses in established and primary OC cell lines (Pearson r=0.96; P=0.01). Tumor tissue cells that did not respond to platinum have been reported to have significantly higher RAD51 scores than those that responded to platinum (P<0.001). Notably, tumors with low RAD51 scores exhibited heightened platinum sensitivity, with a significantly higher likelihood of achieving pathologic complete response [relative risk (RR) 5.28; P<0.001] and platinum sensitivity (RR, ∞; P=0.05). Furthermore, the RAD51 score could accurately predict chemotherapy response (AUC, 0.90; 95% CI, 0.78.1.0; P<0.001), with an automated quantification system mirroring manual test results at 92% accuracy. In the validation cohort, RAD51-low tumors displayed enhanced platinum sensitivity (RR, ∞; P<0.001), a 100% positive predictive value for platinum sensitivity, and improved progression-free survival (HR, 0.53; 95% CI, 0.33.0.85; P<0.001) and OS (HR, 0.43; 95% CI, 0.25.0.75; P=0.003) compared with RAD51-high tumors (85). Advancements have focused on assessing HRec proficiency through RAD51 foci formation at DNA damage sites, exemplified by the repair capability (RECAP) test (92). Initially devised for breast cancer HRD assessment, this test has been validated using tumor sections and has demonstrated applicability in OC (93), underlining the versatility of tumor sections and diagnostic/therapeutic biomarker potential. Additionally, functional HRec status in PDOs has been evaluated using the RECAP test, correlating with PDO drug sensitivity (including platinum chemotherapy and PARPis) (37,42) and clinical response (41). However, as evidenced by in vivo studies, HRD alone does not guarantee a response to platinum or PARPis (41), pointing to a more intricate landscape requiring further exploration.

Genomics

Precision healthcare represents a strategy that tailors therapeutic interventions by comprehensively assessing the interplay of an individual's genetic makeup, environmental factors and lifestyle patterns. Since genomics alone is insufficient to determine treatment options for most patients with advanced diseases, the creation of live biological specimens of tumor-like organs can facilitate the integration of genomic profiles and drug screening within an iterative framework, enabling the identification of personalized therapeutic strategies tailored to each patient (38). In comparison to traditional 2D cultures, 3D organoid cultures exhibit a broader spectrum of drug responses that more accurately mirror genomic variations (39), rendering them an ideal platform for drug sensitivity assessments in translational medicine and precision healthcare endeavors. PDX models of OC also reflect the heterogeneity of tumor genomic profiles and serve as a valuable tool for drug evaluation (94,95). Nevertheless, organoids emerge as a superior 3D culture system, surpassing PDX in terms of i) reduced tumor sample requirements, ii) expedited engraftment timelines and iii) enhanced engraftment success rates (96). Notably, de Witte et al (97) conducted in vitro drug screening utilizing 36 genome-wide profiled PDOs from 23 patients with OC, demonstrating the ability of OC PDOs to faithfully recapitulate individual patient responses to first-line chemotherapeutics. Notably, variations in sensitivity were observed, with low responsiveness to carboplatin/paclitaxel, PARPis and specific tyrosine kinase inhibitors (afatinib, adavotinib), contrasted by high responsiveness to gemcitabine, flavopiridol (a cell cycle-dependent kinase inhibitor) and vimofenib (a BRAF V600E kinase inhibitor). Therefore, OC organoid models may be mainly applied for drug screening, sensitivity and individualized therapy (Fig. 2).

Figure 2. Application of organoids for chemotherapy of OC. OC organoid models may be mainly applied for drug screening, sensitivity and individualized therapy. OC, ovarian cancer.

Deficiencies and prospects of OC organoid models in chemotherapy

Deficiencies

PDO cultures may not be representative of tumor settings because of the absence of stroma, vasculature and immune cells (35); therefore, the organoid model cannot mimic the response to immunotherapy and antiangiogenic therapy (98). Heterogeneity within PDO culture tumors is also incomplete (97). The ability of organoids to maintain biological properties and functions similar to those of the primary tumors in an in vitro environment holds paramount significance for developing clinical applications.

In vivo, tissue development is constrained by external stimulation delivered in an exact space and time sequence. Usually, this does not occur in conventional 3D culture systems, where cells are embedded in an isotropic matrix and uniformly flooded with biochemical and ecological niche signals. This limitation can be overcome by a 3D culture matrix that can release or present biomolecules under spatiotemporal control (99,100). Achieving accurate control of the growth and differentiation of organoids is a critical task in organic research, which carries profound significance across diverse areas, such as disease modeling, pharmaceutical screening and regenerative medical research.

Individual differences

Because of the significant individual differences in patients with OC and the wide variety of tumor mutant types (e.g. Trp53, Brca1, KRAS, BRAF, PTEN, PIK3CA), it is not possible to fully simulate each patient's individual conditions. This may result in some bias in the screening and evaluation of chemotherapeutic drugs, and also increases the difficulty and cost of organoid culture.

Tumor characteristics and physiological barriers, such as the blood-brain barrier (BBB), vary from patient to patient, affecting drug penetration and metabolism in cancer tissues and thus altering therapeutic efficacy. Therefore, the use of human equivalent dose (the in vivo concentration of a drug in tumor tissue) is recommended in cancer organ tissue therapeutic trials to more accurately assess drug efficacy. The development of microfluidics and an artificial BBB based on microchip organ technology may advance this field (101,102).

Cost

Compared with monolayer cell cultures, organoid cultures are more costly because of the need for specific media, including ECM (e.g., Matrigel) and cofactors (e.g., growth factors and hormones) (103,104). In addition, the existence of individual differences and the difficulty of standardization further increases the cost of experiments, as different laboratories and research groups may need to use different methods and materials to conduct experiments, leading to higher costs of duplicating experiments and validation.

The key to improving the efficacy and comparability of organoid techniques is standardization; however, there are currently notable differences in organoid culture protocols from different investigators (34,103), which may lead to inconsistent results. To achieve standardization, key factors such as tissue manipulation procedures, culture media and growth factors need to be harmonized (103). However, standardization is challenged by individual differences, unknown composition of animal-derived scaffolds and poor tunability (105). In addition, the animal-derived Engelbreth-Holm-Swarm substrates used in organoid cultures vary widely between batches and contain unknown contaminants (106), which also limits standardization. Next-generation organoid models, such as co-culture models that incorporate different cells, may partially address these issues but have variable success rates and have also not been standardized (107).

Prospects

3D bioprinting is a promising method for the construction of heterogeneous and reproducible cancer models with controlled shape and structure. The 3D-bioprinted structure can be used to simulate the TME. Different cell layers can be produced using bioprinting, including normal tissue-specific cells, connective tissue and cancer cells (108). This technology can accurately replicate the complex structure and microenvironment of cancer cells, creating organoid models that are more similar to the real-world situation of the tumor (109). Baka et al (110) described a 3D-bioprinted ovarian carcinoma model that combined cancer cells (SKOV-3) and cancer-associated fibroblasts (CAFs). The resulting tumor models demonstrated their ability to maintain cell viability and proliferation, and cells were observed to self-assemble in heterogeneous aggregates. Furthermore, it was observed that CAFs were recruited and multiplied in the surrounding tumor cells in an in vivo process occurring in the TME. Notably, this approach has also demonstrated an ability to produce a large number of reproducible tumor models, which can undergo conventional experimental methods (cell viability and metabolism assays, histology and immunology, and real-time imaging). 3D bioprinting of tumor constructs as personalized in vitro models may be a useful tool for screening for anticancer drugs and establishing accurate treatment regimens. As an example, MRC-5 fibroblasts and human OC cells (OVCAR-5) were previously printed on Matrigel by means of droplet-based bioprinting to form high-throughput, reproducible multi-cellular structures in a controlled space. This approach was shown to provide a platform for minimizing the size of 3D culture models at the macro scale; this could be considered an innovative platform for high-throughput and robust personalized drug screening, including tumor and stromal cell microenvironments (111).

Gene editing technologies

Gene editing techniques, such as CRISPR-Cas9, have the potential to introduce or modify mutations that mimic tumor-forming processes, and potentially identify and modify driver genes in cancer organoids (112). Through CRISPR-Cas9, organoids can be artificially altered by mutation of function or loss of function to induce malignancy (113). Notably, these gene editing techniques may be used to modify specific genes in organoid models to investigate the role of these genes in the occurrence, progression and treatment of OC. This could help to improve the understanding of the mechanisms of OC and to develop new therapeutic agents. Lõhmussaar et al (114) previously introduced mutations into common HGSOC genes, including Trp53, Brca1, Nf1 and Pten, using the genome editing method of CRISPR-Cas9. Organoids have also been artificially modified by gain-of-function or loss-of-function mutations to induce malignancy via CRISPR-Cas9 (115). Furthermore, Kopper et al (37) demonstrated the feasibility of this approach in OC using normal Fallopian tube organoids from high-risk OC donors that could be efficiently genome-edited via CRISPR-Cas9 and clonally amplified. In addition, in this previous study, CRISPR-Cas9 genome editing was used to introduce common HGSOC gene defects into mouse oviducts and the ovarian surface epithelium cell-like organs, showing that both organ types are potentially carcinogenic.

Genomics

Single-cell sequencing has emerged as an advanced biotechnological technique capable of decoding the OC landscape at single-cell resolution. It works at the gene, transcriptome, protein, epigenome and metabolism levels, and provides detailed information that differs from bulk sequencing methods, which only provide average data, representing the overall or average characteristics of the cell population in the entire sample or lesion area on specific lesions. Single-cell sequencing techniques provide detailed insights into the immune and molecular mechanisms that underlie tumorigenesis, progression, resistance and immune escape. These insights can be used to develop innovative diagnostic markers, treatment strategies and prognostic indicators (116). Wan et al (117) reported that HGSOC-like organoid co-cultures were treated with bispecific anti-PD-1 or PD-L1 antibodies, and then underwent single-cell RNA-sequencing. The results showed that this treatment reprogrammed inert natural killer cells and T cells into highly active cytotoxic states, thus indicating the potential benefit of bispecific antibody treatment in HGSOC. Gonzalez et al (118) performed an in-depth single-cell phenotypic characterization of HGSOC by multiparametric mass cytometry. The features of HGSOC biology were examined using a cytometry antibody panel and rare tumor subtypes of HGSOC were identified. It was shown that one of these rare subtypes was enriched in epithelial-mesenchymal transition signaling and was associated with increased tumor metastasis.

Conclusion

In summary, OC organoids maintain high tumor cell heterogeneity and are cultured in small samples, which can be maintained for long-term expansion and can be cryopreserved. OC organoids have been successfully developed from multiple stages and subtypes (56 organoid lines) (37). Thus far, OC organoid models have been widely used in drug screening, biomarker identification and personalized therapy; however, there are challenges to the use of these organoid models, including technical limitations, individual differences, cost issues and issues of standardization. To overcome these shortcomings, a further exploration into the application of technologies, such as 3D bioprinting, gene editing and genomics, is recommended in the optimization of organoid models to improve the future stability and reproducibility of these models. The development of these technologies is expected to make organoid models more stable and reproducible, which will better simulate the in vivo environment, and improve the accuracy and efficiency of drug screening. Therefore, the OC organoid model has a broad potential for application in chemotherapy. Along with the development of the technique, it is hoped that this model will be a more precise and efficient method to predict the prognosis of OC, and to improve therapeutic effects and the quality of life in patients with OC. In addition, research in this area may provide valuable information for other cancer treatments.

Acknowledgements

Not applicable.

Funding

This work was supported by the Natural Science Foundation General Project of Hubei Provincial (grant no. 2024AFB1044), the Development Fund Project of Yangtze University (grant no. WJ2019-22) and the Start Fund of First Affiliated Hospital of Yangtze University (grant no. 2022DIF01).

Availability of data and materials

Not applicable.

Authors' contributions

WZ designed and conceived the study, wrote the manuscript and acquired funding. YD designed and conceived the study and wrote the manuscript. HH and KC conducted the literature analysis. QZ and XC were involved in visualization. HZ and YX revised the manuscript. Data authentication is not applicable. All authors read and approved the final version of the manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References