As a microfabricated device, the organ-on-chip is designated to integrate the culture of extracellular matrix (ECM), living cells and microstructures imitating organs or tissue (43,44). Integrating living human cells into a synthetically produced microenvironment models physiological homeostasis and the process of complex diseases (Fig. 3A) (43). Demers et al (45) developed a versatile microfluidic platform that mimics in vivo spatial and temporal chemical environments during neural tube development. Using similar techniques, Wang et al (46) developed a brain organoid-on-a-chip system from human iPSCs, promoting 3D culture, in situ neural differentiation and self-organization of brain organoids under continuous perfusion of neural differentiation medium in a controlled manner.

The organ-on-a-chip has the advantages of specific stroma, requiring a small amount of tissue for analysis, high-resolution optical measurement, real-time tracking of organoid morphogenesis and inexpensive manufacture. Although tumor microsystems are used to explore the cancer-specific hallmarks, the TME complexity cannot be recapitulated due to simultaneous or successive occurrence of cancer-specific hallmarks. Shirure et al (47) designed a tumor-on-a-chip microfluidic platform and that this could simultaneously model the hallmark characteristics of tumor progression in both cell lines and patient-derived organoids (PDOs), such as proliferation, migration, angiogenesis and intravasation.

At present, the culture environment of organoids lacks vascularization, leading to decreased organoid lifespan and changeability in tissue-specific functionality and architecture (48). In a previous study, 3D vascularized liver organoids comprising induced hepatic cells and decellularized liver ECM were developed based on the microfluidic system, exhibiting improved liver functionality, biosynthetic and metabolic activity, as well as drug response; this study also confirmed the feasibility of vascularized liver organ-on-a-chip systems as a high-throughput drug screening platform (49). To study tumor angiogenesis, the human primary clear cell renal cell carcinoma (ccRCC) cells are combined with endothelial cells in a vascularized, flow-directed, 3D culture system. Under continuous flow, cRCC clusters preserve the key angiogenic signaling axis between ccRCC and endothelial cells by promoting endothelial cell sprouting. This system signifies a vascularized tumor model with adjustable perfusate, input cells and matrices (50). The PDOs established in a multicellular microfluidic chip may prolong cellular function and longevity and construct an intricate organotypic TME (Fig. 3B). Targeting stroma in a tumor-chip model notably increases response to chemotherapy in cancer cells, further verifying the application of the tumor-chip device in drug testing (51). Additionally, multiorgan models of coculture with SC-derived stomach, intestinal and liver organoids have also been established, promoting the discovery of interorgan crosstalk characteristics (52). The aforementioned findings indicate how the organoids combined with organ-on-a-chip technique replicate cell maturity.

3D bio-printing accurately controls the spatial arrangement of cells, biomaterials and soluble factors, forming intricate multicellular structures (53,54). By offering tumor-specific ECM, accurate geometric architecture and biophysical properties, bio-printing can replicate the TME, thus promoting the establishment of complicated and controllable 3D tissue models. In most studies, however, monodispersed tumor cells used as bio-printing building blocks do not effectively replicate the tumor progression due to the rare presence of volumetric tumor cells in isolation (55,56). The combination of 3D bio-printing and tumor organoids allows for the introduction of miniaturized tumor aggregations into a heterogeneous 3D niche containing stromal cells and hydrogels, which are more cell-specific for simulation of TME features and high-throughput drug screening (Fig. 4A) (57,58).

Mollica et al (59) revealed that 3D bio-printed organoids and tumoroid formation are preserved effectively using novel 3D culture substrates. Reid et al (60) applied bio-printing to analyze tumorigenesis and microenvironmental redirection in breast cancer cells and demonstrated that bio-printing could promote the formation of tumor organoids in 3D collagen gels and tumor organoid arrays. Moreover, in vivo findings are accurately simulated through bio-printed organoids, highlighting the feasibility of the 3D bio-printing technique in understanding tumorigenesis and TME control. To increase the throughput of 3D drug screening, an immersion bio-printing technique has been developed to bio-print tumor organoids in multi-well plates (Fig. 4B) (61). Additionally, a 3D-bioprinted construct is used to test the resistance of anti-cancer drugs (sorafenib, cisplatin and 5-fluorouracil) in patient-derived cholangiocarcinoma cells. Compared with 2D cultures, bio-printed cholangiocarcinoma cells exhibit stem-like properties and high resistance to the aforementioned anti-cancer drugs, indicating the potential of 3D bio-printed tumor models in the discovery of targeted drugs (62). Additionally, bio-printed organoids undergo hepatocytic differentiation, including liver-specific enzyme activity and albumin synthesis (63), creating novel possibilities for regenerative medicine and individualized drug screening.

Matrigel is key for the culture of most organoids, but it may elevate the risk of animal-derived microbial infection and batch-to-batch variability in organoids, leading to unreproducible experimental results (64,65). Tissue-engineered cell scaffolds support cell proliferation and attachment and simulate ECM function in vivo (66). For cell- and tissue-derived matrices, synthetic scaffolds, such as polyethylene glycol (PEG)-based hydrogel scaffolds, allow control over the growth conditions.

For conventional organoid cultures, generic matrices are usually applied, but they are difficult to adjust to replicate the unique TME. Ng et al (67) used gelatin-based hydrogels to demonstrate CRC organoid sensitivity to multiple drugs in vivo, and found that these hydrogels may be a promising platform for biochemically and mechanically defined matrices used in multiple types of tumor organoid. In a previous study, a fully synthetic hydrogel scaffold was constructed based on the 8-arm PEG and pancreatic cancer organoids were generated successfully (68). Through regulation of hydrogel properties, the proliferation of pancreatic cancer organoids is controlled, and the phenotypic traits of the TME in vivo are effectively replicated when stromal cells are incorporated into the hydrogels (68). These findings suggest that synthetic scaffolds replicate a pathologically remodeled TME for studying normal and pancreatic cancer cells in vitro. Another study showed that ECM hydrogels generated organoids appropriate for gastrointestinal disease modeling, tissue regeneration and drug development (Fig. 5A), which may serve as effective alternatives to Matrigel (69). Accordingly, tissue-engineered cell scaffolds are promising next-generation materials for organoid technology to understand organ-based developmental biology and predict drug response in tumor organoids (67–69).

CRISPR-Cas9 enables more efficient gene knockout and knock-in than other types of genome editor through introduction of DNA double-strand breaks at specific genomic loci (70). In addition to identification of novel targets for cancer therapy, CRISPR-Cas9 is also used to produce genetically inhibited animal models for drug development (71,72). CRISPR-Cas9 combined with 3D organoid systems facilitates development of precise cancer models for studying diverse mechanisms of tumor progression, metastasis, interactions and drug resistance. Organoid systems not only mimic the human disease and tailor therapeutic strategies, but also serve as an experimental platform for mechanistically studying the gene function in humans (73).

Matano et al (74) introduced mutations into organoids from normal human intestinal epithelium using CRISPR-Cas9 and demonstrated that the isogenic organoids with mutations show tumorigenicity in mice. Murine gallbladder organoids with KRAS mutations or overexpression of Erb-B2 receptor tyrosine kinase 2 cause gallbladder cancer in transplanted immunocompromised mice when CRISPR-Cas9 is used for p53 or PTEN deletion (75). In another study, specific subtypes of breast cancer organoids were established following targeted knockdown of four breast cancer-associated suppressor genes in mammary progenitor cells through CRISPR-Cas9; these organoids respond to endocrine therapy or chemotherapy (Fig. 5B and C) (76,77). Vaishnavi et al (78) used CRISPR/Cas9 technology to silence RNA binding motif, single stranded interacting protein 3 (RBMS3); this facilitated the growth of BRAFV600E lung organoids and malignant progression of lung cancer. Additionally, CRISPR-Cas9 was also used to validate CRC driver genes in mouse intestinal tumor organoids and human CRC-derived organoids (79). Notably, CRISPR-Cas9-mediated homology-independent organoid transgenesis, a genetic tool for labeling specific genes in human organoids, effectively generates genetically engineered human liver duct and fetal hepatocyte organoids within 2–3 months (80,81). This genetic tool can be used to study cell fates and differentiation and identify targets for drug development, showing promise for cancer research.

Tumorigenesis and progression primarily depend on the accumulation of genetic alterations. Understanding the mutational process is key to analyzing the mechanism of tumorigenesis. Multiple studies have demonstrated the feasibility of introducing pathological mutations into normal organoids using genetic modification to simulate tumorigenesis (74,76,82,83) (Fig. 6). As reported by Matano et al (74), isogenic organoids with mutations show tumorigenicity in mice when gene mutations from driver pathways are introduced into organoids derived from normal human intestinal epithelium (84). By establishing AT-rich binding domain protein 1A-deficient human gastric cancer organoids using CRISPR/Cas9 technology, modes of oncogenic transformation are revealed, including essential transcriptional forkhead box protein M1/baculoviral IAP repeat-containing 5-stimulated proliferation and non-essential Wnt-inhibited mucinous differentiation (82). As a key cause of mortality in patients with cancer, the migration and invasion of tumor cells also serve important roles in tumor progression. Through coculture with mammary tumor organoids, a tissue-engineered model with physiologically realistic microvessels was created, which allows quantitative and real-time evaluation of tumor-vessel interactions under the conditions that retain various in vivo characteristics to identify targetable mechanisms of vascular recruitment and intravasation (84). In the patient-derived breast cancer organoids, CD homophilic interactions and subsequent CD44-p21-activated kinase 2 interactions mediate tumor cluster migration and aggregation (85). Moreover, invasion could also be triggered in breast cancer subtypes by basal epithelial gene expression (86).

To investigate the role of tumor organoids in treatment resistance, Boos et al (87) modeled the acquisition of chemotherapy tolerance in metastatic CRC organoids during first-line combined chemotherapy, highlighting the potential of CRC organoids in modeling chemotherapy tolerance in vivo. Similarly, in intestinal cancer organoids, atypical expression of cyclin P facilitates stemness-like phenotypes (88), which usually results in tumor metastasis, relapse and treatment resistance. Through establishment of oxaliplatin-resistant organoids and assessment of their gene expression, myoferlin was shown to be involved in oxaliplatin resistance and tumor progression in gastric cancer (89). As commonly used anti-tumor drugs in prostate cancer, resistance to androgen receptor pathway inhibitors may develop due to epigenetic reprogramming turning castration-resistant adenocarcinoma to neuroendocrine prostate cancer. Bioengineered ECM regulates the response of prostate cancer organoids to small-molecule inhibitors of epigenetic targets and dopamine receptor D2, suggesting that synthetic organoids exert a regulatory effect on ECM in response to targeted therapy in prostate cancer and serve as a novel treatment strategy to overcome resistance (90). Overall, tumor organoids serve as a powerful platform for studying drug resistance mechanisms.

Tumor organoids not only contribute to understanding the mechanism of tumorigenesis, but also predict the response to therapies, including chemotherapy, radiotherapy and targeted therapy (91–93). Vlachogiannis et al (92) established organoids from metastatic gastrointestinal cancer to predict the response to targeted drugs or chemotherapy, showing sensitivity of 100 and specificity of 93%. Gastric cancer organoids faithfully reflect responses to commonly used chemotherapy drugs, such as irinotecan, oxaliplatin, docetaxel, epirubicin and 5-fluerouracil (91). Tiriac et al (94) generated a pancreatic cancer organoid library and found that PDO profiling based on the next-generation sequencing of DNA and RNA and pharmacotyping may predict responses to chemotherapy in pancreatic cancer. Ji et al (95) identified potential drug combination therapies based on pharmaco-proteogenomic profiling of liver cancer organoids, offering guidance for clinical patient selection and drug combination therapies. Ganesh et al (93) revealed the heterogeneity of rectal cancer organoids in chemoradiation and ex vivo responses to clinically relevant chemoradiation associated with clinical responses. Moreover, KRAS-wild-type rectal cancer organoids are sensitive to cetuximab, while KRAS-mutant organoids are resistant, which is consistent with the results of a clinical trial that KRAS mutations are associated with resistance to EGFR-targeted therapy (93). Importantly, in a prospective, interventional clinical trial of the last-line systemic therapy based on PDOs, improved clinical outcomes were observed in patients with CRC compared with those receiving the best supportive care alone (96). In addition, the association between tumor organoids and clinical response has been identified in other types of cancer. By establishing PDOs from different stages of bladder cancer, Minoli et al (97) demonstrated that the PDOs exhibit heterogenous drug responses to standard-of-care treatment and drug screening showed sensitivity to targeted therapy. In a real-world study, lung cancer organoids were used to validate the response to osimertinib, chemotherapy and dual-targeted therapy and a high concordance was identified between lung cancer organoids and clinical response (12). PDO pharmaco-phenotyping not only reflects previous treatment responses of patients with advanced breast cancer but also serves as a potential platform to guide personalized treatment (9). Breast cancer organoids may also be used to predict patient-specific response to drug treatment (98).

Coculture of tumor organoids with immune components may generate tumor-reactive T cells, which may promote the prediction and evaluation of tumor responses at an individual level by blocking PD-1/PD-L1 (98,99). Cattaneo et al (100) described the generation and function of tumor-reactive T cells based on the coculture of tumor organoids with peripheral blood mononuclear cells (PBMCs), demonstrating the feasibility of establishing ex vivo models of T cell immunotherapy at an individual level. Meng et al (101) developed a platform for the expansion of tumor-targeted T cells from peripheral blood and revealed that the coculture of tumor organoids with PBMCs generates tumor-reactive T cells, thereby promoting personalized immunotherapy. Moreover, tumor organoids cocultured with PBMCs are also used to enrich tumor-reactive T cells from peripheral blood of patients with mismatch repair-deficient CRC and non-small-cell lung cancer (SCLC); these T cells are useful for assessing the killing efficiency of matched tumor organoids (102). In certain organoids established from immunotherapy-responsive tumors, activation of T cells and tumor killing activity have been identified using PD-1/PD-L1 blockade (98). Meanwhile, Votanopoulos et al (103) developed an immune-enhanced tumor organoid model with a high clinical association between these organoids and response to checkpoint inhibitors. Collectively, tumor organoids can predict the response to immunotherapy in cancer.

In a longitudinal, observational co-clinical study, second mitochondria-derived activator of caspase mimetic LCL161 was shown to serve as a treatment target in the organoids derived from recurrent, KRAS-mutated liver metastases from rectal cancer (104). Compared with THZ1, YPN-005, a potent inhibitor of CDK7, shows potent antitumor effects in SCLC organoids, suggesting its treatment value in SCLC (105). Based on the interaction of breast cancer organoids and tumor-specific cytotoxic T cells, epigenetic inhibitors GSK-LSD1, CUDC-101 and BML-210 identified via high-throughput screening show antitumor activities (106). Furthermore, BML-210 promotes the efficacy of PD-1-based immune checkpoint blockage.

The tumor-suppressing function and efficient delivery of drugs have key roles in cancer treatment. In multicellular hepatocellular carcinoma organoids (MCHOs) with activated Yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding motif (TAZ) signaling, there is stromal activation and damaged penetration of verteporfin; inhibiting YAP/TAZ transcriptional activity in hepatocellular carcinoma (HCC) cells may facilitate the penetration of drugs into MCHOs. These findings suggest that the treatment targeting activated tumor stroma may promote drug delivery into HCC cells with increased YAP/TAZ activity (107).

To improve understanding of angiogenic signaling pathways and investigate effective treatment strategies, tumor vasculature must be preserved in organoid cultures. For organoid vascularization, implantation of organoids into highly vascularized tissue is a frequently used method (108,109). Another method is coculture with cells including vascular smooth muscle and epithelial cells (ECs) based on gene editing or microfluidic platforms (110). By integrating mesodermal progenitor cells into organoids, Wörsdörfer et al (111) found that the vascularized organoids formed following coculture with tumor cells or neural spheroids and the vessels in tumor organoids are associated with the host vessels after transplantation. Breast cancer organoids with ECs and immune cells show an obvious angiogenic response when cultured with the vascular network (112). Similarly, in the collagen- and hyaluronic acid-enriched ECM with human fibroblasts and MCF-7 cells, vascularized breast cancer organoids have been established successfully (113). Additionally, in the coculture system of organoids and ECs, vascularization is triggered by vascular endothelial growth factors and hypoxia gradients based on compartmentalized microfluidic chips (109,114). The aforementioned findings underscore the importance of coculture models in organoid vascularization.

Coculture of tumor organoids with immune components, such as fibroblasts, stroma, ECs and immune cells, models the tumor-immune interactions, which provides insights into cancer immunotherapy (115). Using the tumor organoid culture for expansion and characterization of tumor-reactive T cells, Dijkstra et al (102) developed a multifunctional platform to study tumor-immune interactions and concluded that CD8⁺ T cells in PBMCs of the same patient are activated in half of the CRC organoids, with similar results in non-SCLC organoids. Meanwhile, a platform for expanding tumor-targeted T cells has been reported in patients with pancreatic cancer (116). By coculturing PBMCs and autologous tumor organoids, this platform enables recognition and expansion of tumor-targeted cytotoxic T cells (99). Based on the coculture of tumor organoids with PBMCs, the establishment and functional assessment of tumor-reactive T cells has also been described (98).

By generating organoids from surgically resected types of cancer based on the air-liquid interface, Neal et al (98) demonstrated that these organoid cultures retain various endogenous immune cell types and non-immune matrix components; immune checkpoint blockade with anti-PD-1 and/or anti-PD-L1 kills the tumor cells through induction of the expansion and activation of tumor antigen-specific T cells in organoid cultures. Moreover, a previous study suggested the potential of the organoid culture system in predicting adoptive immunotherapy responses following incorporation of patient-specific mature lymph node antigen-presenting cells into organoids (103).