To comprehensively review advances in this field, a systematic literature search of the PubMed database (https://pubmed.ncbi.nlm.nih.gov/) was conducted for publications from 2009 to 2025, prioritizing studies from the past 5 years, using the terms ‘lung cancer organoids’, ‘metastasis model’, ‘patient-derived organoids’, ‘tumor microenvironment’ and ‘drug screening’, alone and in combination, to identify original research articles, high-quality reviews and key methodological papers. Emerging lung cancer organoid-based metastasis models recapitulate critical steps of the metastatic cascade, yielding mechanistic insights and informing therapeutic discovery.

Since the pioneering development of intestinal organoids by Sato et al (34) in 2009, stem cell-derived organoid technology has rapidly expanded across multiple tissues, including the brain, liver and diverse tumor types, offering new opportunities for drug screening and disease modeling. Brain organoids serve as a prominent example, as they are constructed through the systematic differentiation of neural stem cells and precise cellular interactions, ultimately forming highly complex neural networks. These networks not only exhibit functions similar to those of the natural brain but also provide essential biological markers for investigating brain development and neurodegenerative diseases (35). Leveraging this advantage, researchers have systematically studied the pathogenesis, pathological evolution and potential therapeutic targets of neurodegenerative diseases such as Alzheimer's disease and Parkinson's disease using in vitro models that closely mimic in vivo physiological conditions (36,37).

In oncology research, tumor organoids derived from patient samples often retain the phenotypic characteristics, genomic heterogeneity and mutational landscape of the original tumor. Tumor organoids derived from patient samples preserve the architectural, molecular and genetic features of the primary tumor, including structure, function, mutation spectrum and gene expression profile. These organoids are effectively used for disease characterization, mechanism exploration, high-throughput drug screening and evaluation, discovery of innovative therapeutic targets and potential compounds and the development of personalized treatment strategies based on individual tumor characteristics (38).

The construction of lung cancer organoid metastasis models begins with the acquisition of suitable tissue samples. Primary lung cancer tissues, obtained from surgical resections or biopsies are commonly used (39,40). To investigate metastatic mechanisms, tissues from metastatic lesions are particularly valuable, including samples from brain, bone or liver metastases, which directly reflect the characteristics of tumor cells within site-specific metastatic microenvironments. For example, a previous study successfully established organoid models from surgical specimens of brain metastases in patients with lung cancer, providing essential tools for exploring the mechanisms of brain metastasis (41).

Malignant pleural effusions (MPEs), including pleural effusion and ascitic fluids, represent an important source for organoid construction. MPEs are rich in tumor cells, and their collection is relatively noninvasive (42). Mazzocchi et al (43) successfully developed 3D lung cancer organoids from pleural effusion. These organoids, composed of tumor and stromal cells embedded in a hydrogel that mimics the ECM, exhibited alveolar-like structures and in vivo-like drug responses. The study confirmed that pleural effusion-derived organoids are suitable for investigating lung adenocarcinoma (LUAD) and that 3D cultures outperform two-dimensional (2D) cultures in disease modeling and drug screening. In another study, Wang et al (44) established lung cancer organoids from both tumor tissue and MPE. It was found that MPE significantly enhanced organoid formation, reduced culture time by >50% and altered drug sensitivity profiles, resulting in increased resistance to certain drugs. Mechanistically, although MPE did not alter the genomic composition of the organoids, it influenced stem cell distribution and cell architecture, including expansion of the extracellular space. Gene expression analyses revealed significant enrichment of pathways related to the ECM and mitochondrial function.

The development of lung cancer organoids presents unique challenges that are not typically encountered in other organoid systems. A primary obstacle is the need for multiple, subtype-specific culture protocols to account for the substantial histological and genetic diversity between NSCLC and small cell lung cancer (SCLC). Furthermore, the high clinical mortality associated with metastasis necessitates modeling organ-specific metastatic niches, such as those in the brain and bone. In addition, given the focus on guiding precision therapies, particularly immunotherapy, there is a pressing need to develop sophisticated, immune-competent co-culture models. This combination of factors makes lung cancer organoids a highly specialized and complex platform.

Organoids provide physiologically relevant platforms for drug screening and sensitivity testing by closely replicating the in vivo architecture and functional characteristics of human tissues. Furthermore, organoids retain tissue-specific features and preserve critical cellular interactions, and capture tumor heterogeneity more accurately. This makes them especially valuable for personalized treatment strategies (45). Drug responses can be evaluated by treating patient-derived tumor organoids (PDTOs) with a range of therapeutic agents and measuring efficacy through multiple analytical readouts. In oncology, organoid-based models are widely used to assess cancer drug sensitivity and to predict therapeutic outcomes, covering a variety of agents, including targeted therapies, immune monotherapies, combination immunotherapies and anti-angiogenic drugs (46,47). For example, clinical studies have employed organoid drug sensitivity assays to anticipate chemotherapy efficacy (48,49).

To mimic the in vivo tumor microenvironment, lung cancer organoids are cultured in 3D matrices such as Matrigel, which provide both structural support and biochemical cues (50). Cultures are maintained in a base medium (such as advanced DMEM/F12) supplemented with growth factors and nutrients essential for organoid survival and expansion (51,52). Furthermore, media are tailored to histological subtypes. NSCLC organoids typically require the addition of EGF, FGF, R-spondin1 and Noggin, whereas SCLC organoids often need additional factors to maintain their neuroendocrine characteristics (53) Further optimization is necessary to model organ-specific metastasis more accurately. For brain metastasis models, neural-supportive medium components such as Neurobasal^™ (Gibco; Thermo Fisher Scientific, Inc.) are incorporated into the medium (54). Culture conditions, including oxygen tension, temperature and humidity, are tightly controlled. Hypoxia, in particular, can preserve cancer stem-like phenotypes and upregulate metastasis-associated programs, providing a more physiologically relevant context for studying tumor invasion, spread and therapy response (55,56).

The 3D scaffold not only supports organoid growth but also serves as a modifiable ECM surrogate. Tumor organoids dynamically remodel their surrounding matrix to facilitate invasion, which can be captured by live imaging using fluorescent ECM components (57,58). Parallel biochemical analyses of conditioned media reveals elevated levels of matrix metalloproteinases and other ECM-modifying enzymes, providing quantitative readouts of invasive potential (59).

EMT is a key transcriptional reprogramming event that confers migratory and invasive capacities. In organoid systems, EMT can be reproducibly elicited by microenvironmental stimuli. Exposure to hypoxia or exogenous transforming growth factor-β (TGF-β) drives molecular and morphological shifts, including downregulation of epithelial markers (such as E-cadherin), upregulation of mesenchymal markers (such as vimentin) and morphological transitions from smooth, spherical structures to irregular, invasive protrusions. This phenotypic change, often mediated by effectors, directly enhances tumor cell migration and invasion, thereby modeling early metastatic steps (60).

Immune-organoid co-culture systems are crucial for elucidating the dynamic interactions between tumor cells and the immune microenvironment, particularly the mechanisms by which cancer cells evade immune surveillance during metastasis. Lung cancer organoids can be directly co-cultured with autologous immune components, such as TILs or peripheral blood mononuclear cells (Fig. 1) (61), to recapitulate patient-specific immune responses. This setup enables real-time assessment of immune cell cytotoxicity and tumor cell lysis, and it captures adaptive immune evasion mechanisms, including the upregulation of checkpoint molecules such as PD-L1 on organoid cells. These co-culture models also provide a physiologically relevant platform for evaluating immunotherapies, including immune checkpoint inhibitors such as anti-PD-1/PD-L1 antibodies, by quantifying treatment-induced organoid death (62).

The successful establishment of lung cancer organoids critically depends on tailoring culture conditions to each subtype, reflecting their distinct cellular origins and oncogenic drivers.

LUAD often originates from alveolar type II (AT2) cells and thus requires culture conditions that sustain AT2 stemness and proliferative capacity. The base medium typically includes essential mediators such as EGF, FGF-7/10, Noggin and R-spondin 1 to promote proliferation and maintain an AT2-like state. This protocol was further optimized by inhibiting the BMP signaling pathway and supplementing with ligands such as heregulin-β1 to better mimic the alveolar niche, as demonstrated in feeder-free systems that enable long-term expansion and modeling of LUAD (63,64).

LUSC, which arises from bronchial basal cells, often relies more on ECM-derived physical and biochemical cues than exogenous growth factors. A robust LUSC model was established using a base medium without the addition of growth factors such as R-spondin, EGF or BMP-4. The success of this system hinged on a tailored ECM composed of Matrigel and collagen I, as well as the strategic initiation of cultures from pre-formed cell spheroids. This approach reproducibly generated organoids that self-organize into a characteristic hierarchical structure with a p63-positive basal cell layer and inward differentiation, providing a distinctive model for studying LUSC biology (65).

The high-grade neuroendocrine phenotype of SCLC necessitates specialized culture conditions. A base media typically includes B27 and N2 supplements, key mitogens (bFGF, EGF and insulin), hydrocortisone and ROCK inhibitors to prevent anoikis and support suspension growth. To model metastasis, particularly brain metastasis, advanced platforms co-culture SCLC cells with human cerebral organoids in a Matrigel-based 3D environment. When combined with modulation of key pathways such as Notch and longitudinal monitoring using live-cell imaging, this platform enables real-time assessment of invasion, colonization and subtype plasticity within the brain microenvironment (66,67).

Patient-derived lung cancer organoids faithfully replicate the pathological and genomic characteristics of their source tumors, preserving key driver alterations, while maintaining malignant cell phenotypes (68). Using gene editing technologies such as the CRISPR/Cas9 system (69), lung cancer organoids can be genetically modified to create metastasis models driven by specific gene mutations. By knocking out or inserting key genes associated with lung cancer metastasis, researchers can study gene function and its impact on the metastatic process.

For example, the CRISPR/Cas9 system has been used to knock out the SOX9 gene in human embryonic stem cells prior to directed differentiation into lung organoids. While SOX9 deletion did not affect epithelial cell differentiation, it impaired proliferation and increased apoptosis. Upon transplantation under the renal capsule of immunodeficient mice or orthotopically into bleomycin-injured mouse lungs, SOX9 inactivation was found to affect the expression of specific cell markers and cell maturation. Techniques such as immunofluorescence and histological analysis revealed these changes, which provided direct evidence for the role of SOX9 in human lung epithelial development (70).

Gmeiner et al (71) (2020) successfully established small cell lung cancer PDX-derived organoid models using HyStem-HP hydrogel scaffolds and demonstrated that the novel drug CF10 could overcome tumor chemoresistance. Functionalized hydrogels are defined scaffolds with well-defined compositions and precisely tunable properties. As a result, they are increasingly supplanting animal-derived matrices such as Matrigel, whose undefined composition and batch variability limit reproducibility and mechanistic control. Current advancements are primarily focused on two main directions. The first involves synthetic hydrogels, such as PEG and polyisocyanopeptides, which provide a highly controllable microenvironment for organoids through the grafting of cell-adhesive peptides (such as RGD and GFOGER), and through fine-tuning of mechanical properties (such as stiffness and viscoelasticity). This approach enables precise investigation of the role of mechanical signals in organoid growth, EMT and invasion (72). The other direction focuses on engineered natural polymers, such as organ-specific dECM, peptide-functionalized nanocellulose and alginate (73). These materials retain tissue-specific biological signals and enhance functionality through chemical modifications, effectively promoting structural maturity and functional complexity of organoids (such as vascularization). The establishment of these functionalized hydrogel scaffold systems has significantly improved the reliability and utility of organoids in standardized modeling, disease mechanism research, drug screening and clinical translation (74–76).

The metastatic cascade unfolds within a complex physiological milieu that is difficult to fully recapitulate in vitro. Consequently, the in vivo transplantation of organoids serves as a critical intermediary, validating their tumorigenic and metastatic capacities in a more physiologically relevant framework for investigation. For example, orthotopic transplantation of lung cancer organoids into the lungs of immunodeficient mice can reproduce the native tumor microenvironment, enabling comprehensive observation of the metastatic continuum from local invasion to distant colonization (77).

The utility of organoid transplantation has been extensively demonstrated across multiple cancer types. In colorectal cancer, organoid libraries retain stable pathological and genetic characteristics. GFP-labeled organoids derived from patient-derived xenografts (PDXs), when transplanted into immunodeficient mice, facilitate visualization and quantification of micrometastatic lesions (such as hepatic metastases following splenic injection) (78). Comparable strategies have been employed in models of lung, breast (79), pancreatic (80) and brain (81) cancer, effectively recapitulating tumor dynamics and supporting preclinical evaluation of therapeutic efficacy.

By constructing humanized lung cancer PDX models through engraftment of human hematopoietic stem cells into immunodeficient mice to reconstitute a human immune system, researchers can effectively evaluate the efficacy of PD-1/PD-L1 inhibitors such as pembrolizumab and nivolumab (82).

While organoid transplantation is a versatile tool in metastasis research, it is one of several available preclinical models. A comparative evaluation of these mainstream platforms underscores their distinct advantages and optimal applications, with PDOs occupying a unique and increasingly important niche in translational studies.

Comparative studies have highlighted the distinct advantages of PDOs over PDXs in several key aspects. PDOs can often be established within weeks, which is significantly faster than the months typically required for PDXs. PDOs are more cost-effective and offer greater scalability, making them suitable for high-throughput in vitro studies (83). A critical limitation of conventional PDX models, which are grown in immunodeficient mice, is the absence of a functional human immune system, restricting their application in immunotherapy research. Humanized PDX models address this issue to some extent, but still face challenges such as incomplete immune reconstitution. Additionally, murine stromal cells gradually replace the human tumor microenvironment in PDXs, potentially altering key aspects of tumor biology. Although PDOs also face challenges in fully replicating the tumor microenvironment, they offer greater flexibility for incorporating human immune components through co-culture systems (84).

Spheroids are simple 3D cell aggregates that are valuable for high-throughput drug screening and for studying basic nutrient gradients. However, they generally lack the complex and architecturally accurate tissue morphology, cellular diversity (including stromal components) and long-term functional differentiation that characterize PDOs derived from stem or progenitor cells. These shortcomings limit their ability to mimic the complexity of native tissues (85,86).

Organ-on-a-chip platforms are highly effective at engineering precise physiological microenvironments, incorporating fluid flow, mechanical forces and multi-cellular interactions to model dynamic processes such as vascular perfusion and immune cell trafficking. Their limitation often stems from the use of established cell lines, which may not fully represent the patient-specific genomic heterogeneity. PDOs complement these systems by offering a genetically accurate, patient-derived tissue source, which can be integrated into chip-based devices to create more personalized and representative models (87).

GEMMs are an optimal model for studying tumorigenesis and metastasis within an intact, immunocompetent organism and in the context of developing tissues. However, their primary limitations include long timelines, high costs and inherent species differences, which can hinder the direct translation of findings to human patients. PDOs offer a more rapid, human-based platform for personalized screening and genetic manipulation, providing a valuable complement to GEMMs (88).

These models are not mutually exclusive, and they are highly complementary. Insights gained from high-throughput drug screening in PDOs can be validated within the more physiologically intact contexts of PDXs or GEMMs. Conversely, PDOs can be integrated into organ-on-a-chip devices to create more complex, human-relevant organoid-on-a-chip systems. The choice of model depends on the specific research question, with PDOs offering a balanced platform that preserves human tumor heterogeneity while enabling scalable in vitro experimentation.

PDOs and patient-derived organotypic tissue cultures (PD-OTCs) have recently demonstrated notable value in recapitulating the lung cancer tumor microenvironment and elucidating its functional dynamics, particularly the interactions between metabolism and immunity. Using stable isotope-resolved metabolomics, Fan et al (89) found that PD-OTCs more accurately reproduce complex metabolic reprogramming features observed in patients, including enhanced glycolysis, nucleotide synthesis and anaplerotic reactions, compared with PDX models of NSCLC. Building on this platform, the study directly uncovered key regulatory mechanisms along the metabolism-immunity axis. It demonstrated that immunomodulators (such as β-glucan) and checkpoint inhibitors can reprogram the metabolic state of immune cells, effectively reversing immunosuppression by shifting TAMs from the M2 anti-inflammatory phenotype to the M1 pro-inflammatory phenotype. These experimental findings provide empirical support for the theoretical framework proposed in the study by Huang et al (90), which suggested that metabolic networks within the tumor microenvironment coordinate immune responses and drive immune evasion. In summary, PDOs and PD-OTCs have emerged as key bridges connecting the theory of metabolic reprogramming with the practice of immunotherapy, establishing a methodological foundation for developing individualized combination strategies that target the metabolism-immunity axis.

By contrast, CTC-derived organoids are established from tumor cells shed into the bloodstream from primary or metastatic sites, thereby capturing dynamic and comprehensive tumor-related information. Culturing organoids from sources such as intestinal tissues is relatively straightforward, as these models can be stably cultured and passaged in defined media and matrix gels using well-established protocols. However, culturing CTC organoids poses greater technical challenges. CTCs are extremely rare in peripheral blood, typically on the order of 1 to 10 cells/ml, making their isolation and enrichment particularly demanding. CTCs derived from different tumor types require tumor-specific culture conditions, further contributing to low success rates (91,92). In prostate cancer, for example, the reported success rate of culturing CTCs is ~5.8% (93).

Regarding practical applications, general organoids are widely used to study organ development, disease modeling and drug testing. CTC organoids are particularly valuable for investigating tumor metastasis, personalized cancer therapy and drug resistance mechanisms. In terms of cellular characteristics, organoids derived from normal tissue stem cells typically exhibit stable morphology and function, while those derived tumor tissues retain some malignant features of the original tumor cells. CTC organoids display a high degree of heterogeneity that reflects the diversity of tumor cells during the metastatic process, and they may exhibit stronger metastatic potential (94).

CTC organoids derived from blood samples of patients with SCLC can simulate the tumor microenvironment and support drug response testing. By preserving tumor cell heterogeneity, CTC organoids can improve prediction of drug efficacy relative to traditional preclinical models. This enables clinicians to identify effective therapeutic agents tailored to individual patients, prioritize drugs with higher sensitivity, avoid those with potential resistance and reduce adverse effects associated with ineffective treatments. Ultimately, this approach can improve clinical outcomes and enhance patient life quality. Beyond their utility in guiding personalized therapy, CTC organoids serve as a powerful tool for investigating the biological features of SCLC, including the mechanisms of tumor initiation, progression and metastasis (94,95). They also offer insights into the development of drug resistance and inform strategies to overcome it. From a translational perspective, molecular analysis of CTC organoids can facilitate the discovery of novel therapeutic targets, assist in evaluating the efficacy and safety of emerging drug candidates and support early-stage clinical trials due to their operational simplicity, cost-effectiveness and relatively short culture time (96).

Lung cancer organoid-based metastasis models provide an effective platform for in-depth exploration of metastatic mechanisms. Research on metastasis-related genes and signaling pathways has revealed that numerous genes serve critical roles in lung cancer metastasis. For example, deletion of the histone methyltransferase gene KMT2C promotes distant metastasis in SCLC through DNMT3A-mediated epigenetic reprogramming. In lung cancer organoid models, KMT2C knockout markedly increases organoid invasiveness and metastatic potential. Further analyses have revealed associated alterations in histone modification and DNA methylation patterns, which regulate downstream genes involved in metastatic progression (97).

These models are also instrumental for studying the impact of the tumor microenvironment on metastasis. Cellular components such as tumor-associated macrophages (TAMs) and fibroblasts interact with tumor cells and modulate their metastatic behavior. Co-culture experiments involving TAMs and lung cancer organoids have demonstrated that macrophages secrete cytokines that enhance organoid invasion and migration. This pro-metastatic effect has been further validated in vivo, where the presence of TAMs significantly promoted metastatic spread following transplantation (98).

Lung cancer organoid metastasis models are valuable for drug screening and efficacy evaluation. They support high-throughput screening of anti-metastatic drugs by evaluating treatment-induced changes in cancer cell migration, invasion and expression of metastasis-associated markers. Lung cancer organoid metastasis models are also well suited for assessing the synergistic effects of combination therapies, thereby informing optimization of clinical treatment strategies. In addition, correlating genetic characteristics and molecular marker expression in patient-derived lung cancer organoids with clinical metastasis data may facilitate the development of predictive models for metastatic risk. For example, elevated expression of specific genes or proteins could indicate increased metastatic risk and poor prognosis, aiding early intervention and improved outcomes (99).

Metastasis is a major determinant of prognosis in lung cancer. A previous study established lung cancer organoids from malignant serous effusions that retained key tumor characteristics and showed high genomic concordance with the original tumors. The Lung Cancer Organoids-Drug Sensitivity Test (LCO-DST) demonstrated strong predictive performance, with a sensitivity of 84.0%, specificity of 82.8% and accuracy of 83.3% for distinguishing drug-sensitive from drug-resistant patients. This test effectively predicted responses to targeted therapies and chemotherapy. Notably, lung cancer organoids exhibit both stability and inter-patient heterogeneity. Organoids derived at different time points from the same patient maintained consistent genomic features, while organoids from different patients showed distinct morphologies and drug sensitivities. These findings highlight the potential of LCO-DST for predicting the effectiveness of combination therapies in advanced lung cancer (100).

Using lung cancer brain metastasis organoid models as an example, researchers have evaluated the impact of various therapeutic agents on the growth and invasion of brain metastases. Treating lung cancer brain metastases (LCBM) remains a significant clinical challenge, underscoring the importance of drug screening. Traditional cancer cell lines and PDX models have notable limitations; cell lines often lack in vivo tumor heterogeneity and PDX models face interspecies incompatibilities and low establishment efficiency (101). By contrast, LCBM organoids recapitulate the molecular and phenotypic features of metastatic tumors. They can be used to screen targeted therapies based on specific genetic alterations such as EGFR or ALK mutations, to identify potential radiosensitizers and evaluate the combined radio-chemotherapy efficacy. With access to large-scale drug libraries, organoid models enable high-throughput drug screening, providing strong support for LCBM drug development and clinical treatment selection (101).

Tumor heterogeneity is a critical factor in the development, therapeutic resistance and metastasis of lung cancer. The dynamic evolution of tumor heterogeneity in space (between primary and metastatic sites and within different tumor regions) and in time (before and after treatment or at different disease stages), poses significant challenges for constructing effective metastasis models (102–104). Compared with traditional cell lines and GEMMs, lung cancer organoids more accurately preserve genomic diversity, phenotypic variations and key components of the tumor microenvironment, including co-culture systems that incorporate stromal and immune cells from the original tumors (105).

Single-cell sequencing studies (106–109) have demonstrated that lung cancer organoids retain subclonal structures and transcriptomic heterogeneity comparable to those of the primary tumors. These organoids are also enriched for invasive front cell populations, particularly cells undergoing EMT, which serve crucial roles in metastasis (110). However, replicating the dynamic evolution of heterogeneity within organoid-based metastasis models remains technically challenging. As the metastatic cascade involves selection of heterogeneous CTCs, current organoid systems are limited in their ability to mimic circulatory dynamics, including fluid shear stress and trans-endothelial migration. To address these limitations, integration with organ-on-a-chip technologies will be necessary to achieve more physiologically relevant spatiotemporal reconstruction (111–113).

LCBMs are a common and severe complication in advanced diseases. Brain metastases exhibit higher median absolute deviation and higher mutation allele tumor heterogeneity scores than primary tumors, indicating greater genomic diversity (114). In addition, single-cell RNA sequencing has revealed marked differences in gene expression between tumor cells from brain metastases and those from primary sites, highlighting cell-level heterogeneity that may influence disease progression and therapeutic response (115). In the future, the integration of spatial transcriptomics, live-cell imaging and AI-driven algorithms for quantifying heterogeneity will enable lung cancer organoid models to capture dynamic tumor evolution at single-cell resolution. These advances will support early prediction and precise intervention for metastatic risk.

The success rate of organoid culture remains suboptimal for certain lung cancer subtypes, such SCLC and sarcomatoid carcinoma, due to their limited stemness and strong dependence on the ECM (116). A major bottleneck is the widespread use of ill-defined matrices such as Matrigel, which are affected by batch-to-batch variability, undefined composition and non-physiological mechanical properties. Although dECM provides a more native biochemical environment, donor-dependent variability complicates standardization (117–119).

Reconstructing the tumor microenvironment also presents significant challenges. Co-culture systems that integrate immune cells are hindered by poor immune cell infiltration into dense hydrogel domes and by the difficulty of sourcing authentic TILs. These limitations impede accurate modeling of complex tumor microenvironment interactions that are crucial for studying tumor-immune dynamics and metastasis (120). In addition, current models are limited in ability to replicate full cellular diversity and complex stromal interactions, including functional vasculature, which limits the ability to model tumor growth, metastatic dissemination and dynamic interactions between tumor cells and their microenvironment.

From a clinical perspective, the absence of standardized protocols for model construction, culture and assessment severely limits reproducibility and hinders large-scale application. High costs associated with tissue acquisition, organoid culture and molecular analyses present significant economic barriers. Additionally, there is a lack of large-scale, multi-center clinical validation to confirm the predictive accuracy of these models for drug response and personalized therapy, which impedes the broader clinical adoption (121).

Ethically, the collection and use of patient-derived tissues requires rigorous informed consent processes. The application of advanced techniques such as gene editing introduces additional ethical complexities, requiring clear regulatory frameworks for safety assessment and intellectual property management.

To address these challenges, the field is moving towards a multi-faceted strategy that leverages cutting-edge technologies and promotes standardization.

Integrating multi-omics technologies, including genomics, transcriptomics, proteomics and metabolomics, with organoid models will provide a comprehensive molecular view of metastasis. This approach may reveal novel drivers, elucidate dysregulated metabolic pathways and uncover complex molecular interactions within the tumor microenvironment. Combining these high-throughput techniques with organoid models will deepen mechanistic insights and accelerate the development of more targeted and effective therapies (122).

A key goal in cancer research is the development of sophisticated immune-organoid models. By co-culturing lung cancer organoids with autologous immune cells, including T cells and natural killer cells, or by utilizing humanized mouse models, researchers can better replicate crucial tumor-immune interactions. These models provide a powerful platform for evaluating immunotherapy efficacy and for studying mechanisms of immune evasion. Understanding how tumors interact with immune cells in a more physiologically relevant context can accelerate the discovery of novel therapeutic strategies and improve the precision of cancer immunotherapy (123–128).

The synergy between organoid models and liquid biopsy is also highly promising. Analysis of circulating tumor DNA (ctDNA) from patient plasma or from organoid culture supernatants, provides a non-invasive method to track tumor evolution in real-time. This approach can guide drug selection in organoid models, enabling personalized and adaptive treatment strategies. By incorporating ctDNA analysis, researchers can continuously monitor treatment response and identify emerging resistance mechanisms, creating a closed-loop feedback system that enhances the precision and flexibility of oncology therapies. This combination has the potential to transform personalized cancer care by aligning treatments with the evolving molecular landscape of each tumor (91,129–131). Single-cell RNA sequencing is another pivotal tool; it can resolve cellular heterogeneity within organoids and their tumor microenvironment, validate model fidelity and identify rare cell populations critical to metastasis research (106).

The application of AI and machine learning is poised to transform the field. AI algorithms can analyze high-content imaging data from organoids to automate quantification of metastatic phenotypes, including invasion and migration, which streamlines the screening process (132–134). Furthermore, AI can integrate multi-omics and drug response data to build predictive models for metastatic risk and therapy optimization, thereby guiding personalized treatment decisions. These advances must be supported by rigorous standardization. Harmonized protocols for tissue processing, culture conditions and quality control are essential to ensure reproducibility, enable multi-center collaborations and facilitate clinical adoption. In parallel, optimizing workflows and reducing costs are important for improving accessibility.