Tumor organoids can retain some of the histological features, driver alterations, and clonal differences of a patient’s tumor, providing expandable human-derived models for studying tumor evolution, therapeutic sensitivity, and mechanisms of drug resistance.

Keywords: tumor organoid; patient-derived organoid; tumor heterogeneity; three-dimensional culture; malignant identity characterization; drug screening; resistance mechanism; precision therapy; tumor microenvironment

1 Biological Basis and Technical Characteristics of Tumor Organoid Models

1.1 Malignant Clones and Tumor Organoid Formation

Tumor organoids are generally established from malignant cells with sustained proliferative capacity in patient tumor tissues. Tumor cells differ in their tolerance to anoikis, oxidative stress, and exogenous growth signals, and only some cells can survive in a three-dimensional environment and form glandular lumens, cysts, solid aggregates, or branched structures. Organoid establishment therefore represents a process of clonal selection, and the resulting culture may not fully contain all cellular subpopulations present in the original tumor.

1.2 Retention and Loss of Heterogeneity From the Original Tumor

Patient-derived tumor organoids can retain some histological features, driver mutations, copy-number abnormalities, and differences in drug response. However, the degree to which individual subclones are preserved is influenced by the sampling location, initial cellular proportions, and culture conditions. Low-abundance clones may be lost during early establishment, whereas rapidly growing subclones may gradually dominate during serial passaging. Therefore, the presence of only a few shared mutations between the original tumor and organoid does not demonstrate complete preservation of heterogeneity.

1.3 Culture Selection Pressure and Subclonal Remodeling

Wnt, EGF, FGF, TGF-β pathway inhibitors, and hormonal components in the culture medium impose selective pressure on tumor cells. Clones dependent on particular growth signals may disappear when the corresponding factors are absent, whereas clones that have acquired pathway autonomy may become enriched. The composition of long-term organoid cultures reflects both the characteristics of the original tumor and adaptive evolution under in vitro conditions.

1.4 Competitive Expansion of Normal Epithelial Cells

Tumor tissues commonly contain normal glands, ducts, or basal cells, and some normal stem cells have a greater organoid-forming capacity than malignant cells. Normal cells may form regular, thin-walled, highly polarized cystic structures and expand rapidly in conventional expansion media. Distinguishing normal from tumor organoids requires integration of original tissue pathology, driver alterations, copy-number abnormalities, and growth-factor dependence and cannot rely solely on three-dimensional morphology.

1.5 Limits of Tumor Microenvironment Preservation

Conventional tumor organoid cultures are dominated by malignant epithelial cells, whereas immune cells, fibroblasts, vascular cells, and the original extracellular matrix gradually decrease. These models are therefore more suitable for studying tumor cell-intrinsic growth and drug responses than for directly representing the complete tumor microenvironment. Studies of immune escape, stromal barriers, or angiogenesis require coculture, air-liquid interface, microfluidic, or tissue-fragment culture systems.

2 Tumor Tissue Sampling, Dissociation, and Three-Dimensional Culture Workflow

2.1 Sample Types and Their Value for Model Establishment

Tumor organoids can be established from surgically resected tissue, endoscopic biopsy, core needle biopsy, pleural or peritoneal effusions, exfoliated urinary cells, or metastatic lesions. Large surgical specimens allow multiregional sampling and histological comparison, small biopsies better match the clinical decision-making window, and liquid samples facilitate access to tumor cells from patients with advanced disease or tumors that cannot be surgically sampled again. Because these sample types differ in cell number, necrotic content, and normal-cell contamination, identical processing intensity should not be applied to all samples.

2.2 Differences Between Pre-Treatment and Post-Treatment Samples

Untreated primary tumors generally contain more treatment-sensitive clones, whereas samples collected after chemotherapy, targeted therapy, or radiotherapy may be enriched in resistant subpopulations and may contain more fibrosis, necrosis, and cells with reduced viability. When establishing longitudinal organoid models, the sampling time, treatment regimen, treatment course, and clinical response should be recorded separately to avoid misinterpreting treatment-induced changes as intrinsic tumor characteristics.

2.3 Tumor Content and Sampling Regions

Before organoid establishment, the tumor-cell proportion should be determined through pathological evaluation, and regions containing extensive necrosis, hemorrhage, or normal tissue should be avoided whenever possible. For tumors with marked spatial heterogeneity, organoids can be established separately from the central region, invasive front, primary lesion, and metastatic lesion. Directly mixing multiple regions may increase the overall establishment rate but can obscure region-specific clonal composition and drug responses.

2.4 Mechanical Processing and Enzymatic Digestion Intensity

Tumor tissue can first be cut into small fragments, followed by digestion with collagenase or a mixed-enzyme system selected according to collagen content and tissue density. Highly fibrotic pancreatic cancers and some breast cancers generally require more extensive matrix digestion, whereas fragile lung cancers, ovarian cancers, or small biopsy samples require shorter processing times. Excessive dissociation into single cells increases anoikis and disrupts intercellular dependency signals, whereas retaining small tumor-cell clusters is often more favorable for primary organoid formation.

2.5 Control of Normal and Non-Target Cells

Red blood cells, adipose tissue, necrotic debris, and excessive fibroblasts can interfere with matrix embedding and imaging. However, excessive epithelial-cell enrichment may also eliminate low-abundance malignant subpopulations. Depending on the cancer type, differential sedimentation, mesh-based size fractionation, short-term differential adhesion, or culture-factor selection may be used. Nevertheless, the tumor identity of the culture must ultimately be confirmed by pathological and molecular analyses.

2.6 Primary Seeding and Criteria for Successful Establishment

Cell clusters are generally embedded in a basement membrane matrix or collagen-based system and cultured in a medium matched to the tissue of origin and molecular subtype. During primary culture, the time of organoid appearance, morphological type, number of structures formed, and proportion of normal structures should be recorded continuously. Formation of a three-dimensional structure does not by itself indicate successful establishment. Only cultures that maintain tumor characteristics after serial passaging and cryopreservation-recovery are suitable for use as stable models.

Figure 1 Sample Processing, 3D Establishment, and Research Applications of Patient-Derived Tumor Organoids

Table 1 Establishment Challenges and Processing Priorities for Different Tumor Samples

3 Regulation of Culture by Extracellular Matrix, Growth Factors, and Small-Molecule Compounds

3.1 Culture Medium as a Selection Environment for Tumor Clones

Tumor organoid culture medium does more than provide nutrients; it determines which clones can continue to expand. Cells strongly dependent on exogenous Wnt, EGF, or FGF have a growth advantage in media enriched with these factors, whereas malignant clones with strong signaling autonomy can continue expanding after these factors are reduced. Medium formulations should balance increased establishment efficiency against reduced clonal selection according to the research objective.

3.2 Growth-Factor Withdrawal and Pathway Autonomy

Some tumors reduce their dependence on exogenous signals because of abnormalities in APC, CTNNB1, KRAS, EGFR, or other pathways. Gradual withdrawal of Wnt, R-Spondin, EGF, or other growth factors can assist in identifying clones with pathway autonomy and suppress the growth of some normal cells. However, factor independence is not the exclusive consequence of a particular genetic alteration and must be interpreted together with genomic and pharmacological evidence.

3.3 Dual Effects of TGF-β and BMP Signaling

Inhibition of TGF-β- or BMP-associated signaling can improve the expansion efficiency of some epithelial tumor organoids but may also alter epithelial-mesenchymal states, stemness, and drug responses. In tumors with abnormalities in SMAD4, TGFBR2, or other pathway components, the effect of exogenous inhibitors may be substantially reduced. When studying pathway-directed drugs, prolonged inclusion of inhibitors acting in the same direction as the test drug should be avoided.

3.4 Matrix Composition and Invasive Phenotype

Basement membrane matrices favor epithelial structure formation, whereas type I collagen and stiffness-adjustable hydrogels are more suitable for observing migration, branching, and invasion. Increased matrix density also alters drug diffusion and cellular mechanosignaling, causing the same organoid to show different drug sensitivity in different matrices. Drug-screening experiments must therefore standardize matrix batch, concentration, and embedding volume.

3.5 Interference From Hormonal and Metabolic Components

Insulin, IGF, estrogen, androgen, glucocorticoids, and nicotinamide can directly affect the growth and treatment response of breast cancer, prostate cancer, liver cancer, and endocrine-related tumors. When evaluating endocrine therapies or metabolic inhibitors, the hormonal and nutritional background of the culture medium should be clearly defined, and withdrawal or graded supplementation strategies should be used to distinguish intrinsic tumor dependence from culture-condition effects.

3.6 Limits of ROCK Inhibitor Use

Y-27632 reduces cell death after tissue dissociation and recovery and is suitable for short-term use during primary seeding or after single-cell passaging. Continuous exposure may affect the cytoskeleton, adhesion, and invasive behavior and interfere with migration, mechanical, and drug-sensitivity assays. For studies involving these endpoints, the ROCK inhibitor should be removed after organoid recovery.

4 Culture Characteristics of Digestive and Genitourinary Tumor Organoids

4.1 Colorectal Cancer Organoids

Colorectal cancer organoids have a relatively high establishment rate, but normal intestinal epithelium also has strong expansion capacity. During culture, normal cystic structures must be distinguished from malignant glands, and abnormalities in APC, KRAS, TP53, SMAD4, or other genes shared with the original tumor should be evaluated. This model is particularly suitable for comparing responses between primary and liver metastatic lesions, analyzing sensitivity to anti-EGFR therapy, and studying resistant clones after chemotherapy.

4.2 Gastric and Esophageal Cancer Organoids

Gastric cancer and esophageal adenocarcinoma exhibit substantial histological and molecular subtype diversity, and a single culture formulation may preferentially support particular glandular clones. Model confirmation should integrate morphology of the original tissue, ERBB2 amplification, CDH1 abnormalities, and other tumor-specific alterations. Esophageal squamous cell carcinoma organoids require greater attention to expansion of basal-like cells, keratinization status, and responses to chemoradiotherapy and should not be cultured using adenocarcinoma formulations without modification.

4.3 Pancreatic Cancer Organoids

Pancreatic ductal adenocarcinoma samples commonly have a low tumor-cell proportion and extensive fibrosis, making tissue dissociation and competition from normal pancreatic ductal cells major establishment challenges. Organoid identity should be confirmed by comparing abnormalities in KRAS, TP53, SMAD4, CDKN2A, and other genes with the original tissue. Their principal translational applications include comparison of chemotherapy combinations, analysis of KRAS pathway dependence, and coculture with cancer-associated fibroblasts.

4.4 Liver Cancer and Cholangiocarcinoma Organoids

Hepatocellular carcinoma, intrahepatic cholangiocarcinoma, and combined hepatocellular-cholangiocarcinoma differ markedly in cellular origin and signaling dependence. Cultures should be evaluated using lineage information such as HNF4A, GPC3, and KRT19, and tumor identity should be confirmed using alterations in the TERT promoter, CTNNB1, TP53, IDH1/2, FGFR2, or related genes. These models are suitable for studies of molecular subtypes, metabolic dependence, and responses to targeted therapy.

4.5 Prostate Cancer Organoids

Primary prostate cancer cells have limited in vitro expansion capacity, and normal basal cells often become dominant during culture. Model confirmation should simultaneously evaluate androgen receptor signaling, tumor-specific genetic abnormalities, and luminal epithelial characteristics. For castration-resistant and neuroendocrine prostate cancers, absence of AR or PSA should not be interpreted as failed establishment.

4.6 Bladder and Kidney Cancer Organoids

Bladder cancer organoids can be established from resection specimens, biopsies, or urine and are suitable for comparing molecular subtypes and responses to FGFR pathway-directed drugs. Clear cell renal cell carcinoma cultures should be confirmed using PAX8, CA9, and VHL pathway abnormalities. Both model types require prevention of competitive expansion by normal urothelial or renal tubular epithelial cells.

4.7 Ovarian and Endometrial Cancer Organoids

High-grade serous ovarian cancer organoids can be used to evaluate platinum agents, PARP inhibitors, and homologous recombination repair status, but distinct tumor subtypes cannot be classified solely by PAX8 or WT1. Endometrial cancer organoids should be stratified according to mismatch repair status, POLE, TP53, and hormone receptor status before in vitro drug responses are associated with clinical molecular subtypes.

5 Culture Characteristics of Respiratory, Breast, and Head and Neck Tumor Organoids

5.1 Lung Cancer Organoids

Normal airway epithelial cells in lung cancer samples often have strong organoid-forming capacity and are a major cause of false establishment. Lung adenocarcinoma and lung squamous cell carcinoma models should be confirmed using the corresponding lineage markers together with EGFR, KRAS, ALK, ROS1, MET, or other driver alterations. These models are mainly used to study targeted therapy sensitivity, acquired resistance, and differences between primary lesions and pleural effusion samples.

5.2 Breast Cancer Organoids

Breast cancer organoid culture is readily influenced by estrogen, progesterone, insulin, and IGF signaling. Estrogen receptor-positive, HER2-positive, and triple-negative breast cancers should retain their corresponding receptor status and proliferation characteristics and should be compared with receptor testing from the original tissue. Studies of endocrine resistance should use a clearly defined hormonal background rather than culture systems containing complex and undefined hormonal components.

5.3 Head and Neck Squamous Cell Carcinoma Organoids

Head and neck squamous cell carcinoma organoids can form solid, stratified, or keratinized structures and are suitable for studies of radiotherapy, EGFR inhibition, and combination treatment. Oropharyngeal cancer samples should also be stratified according to HPV nucleic acid and p16 status. Because the original tumors often contain hypoxic regions and immune infiltration, epithelial-only organoids have limited capacity to model radiotherapy and immunotherapy responses.

5.4 Nasopharyngeal Carcinoma Organoids

Nasopharyngeal carcinoma is closely associated with Epstein-Barr virus infection, but the viral genome and latent expression state may change during culture. Viral indicators and driver characteristics of the original tumor should be monitored continuously after model establishment. The main value of this model lies in studying virus-associated tumor phenotypes, radiosensitivity, and immune recognition rather than merely evaluating organoid proliferation rate.

5.5 Salivary Gland Tumor Organoids

Salivary gland tumors comprise diverse histological types, and some subtypes are driven by specific gene fusions. Organoid characterization should prioritize confirmation of tumor-specific fusions and luminal or myoepithelial differentiation rather than relying on general epithelial markers. Because some subtypes grow slowly, subclonal selection and contamination by normal glandular cells require particular attention during long-term culture.

6 Culture Characteristics of Nervous System, Skin, and Other Tumor Organoids

6.1 Glioblastoma Organoids

Glioblastoma can be cultured by reaggregation of dissociated cells or by retaining tissue fragments. Tissue-fragment approaches better preserve local hypoxia, necrosis-like regions, and cellular-state differences but provide lower culture uniformity. Model confirmation should incorporate EGFR amplification, PTEN abnormalities, TERT promoter status, and other molecular features of the original tumor rather than relying solely on SOX2 or GFAP expression.

6.2 Brain Metastasis Organoids

Brain metastasis organoids should first be confirmed to share lineage identity with the primary tumor, after which differences in drug sensitivity and signaling dependence between primary and metastatic lesions can be compared. The absence of brain endothelial cells, astrocytes, and microglia during culture may weaken resistance phenotypes associated with the brain microenvironment. These models are therefore suitable for combination with brain-derived stromal cells or microfluidic systems.

6.3 Neuroendocrine Tumor Organoids

Neuroendocrine tumors differ markedly in differentiation, proliferation rate, and hormone secretion. Organoids should be confirmed using CHGA, SYP, INSM1, and tumor-specific molecular abnormalities, and function should be evaluated through hormone or neuropeptide secretion. Highly differentiated and slowly proliferating tumors may be difficult to expand over long periods, and establishment failure does not indicate that the original sample lacked tumor cells.

6.4 Melanoma Organoids

Melanoma is not a typical epithelial tumor, and its three-dimensional culture depends more strongly on cell aggregation, collagen environments, and stromal-cell support. SOX10, MITF, MLANA, and BRAF, NRAS, or NF1 status can be used to confirm model identity. These models are suitable for studies of MAPK-targeted therapy resistance and immune-cell cytotoxicity, but pure tumor-cell cultures cannot fully reproduce immunotherapy responses.

6.5 Sarcoma and Mesenchymal Tumor Models

Sarcomas are generally cultured as three-dimensional spheroids, tissue fragments, or in stiffness-adjustable hydrogels rather than in standard epithelial organoid systems. Individual subtypes should be characterized primarily using fusion genes, tissue of origin, and mesenchymal differentiation features. Research commonly focuses on matrix mechanics, cell migration, drug penetration, and local recurrence rather than gland formation or epithelial polarity.

6.6 Pediatric and Rare Tumor Organoids

Samples from pediatric and rare tumors are limited and frequently retain embryonic or developmental characteristics. Model establishment should prioritize preservation of driver fusions, amplifications, or specific differentiation programs while minimizing culture selection that favors normal developmental progenitors. These models are particularly valuable for supplementing the limited availability of conventional cell lines and validating drug responses associated with rare alterations.

7 Characterization of Tissue Origin, Lineage, and Functional Markers in Tumor Organoids

7.1 Histological Concordance Between the Original Tumor and Organoid

Organoids should be compared with the original tumor for glandular architecture, nuclear atypia, mucus secretion, keratinization, cellular arrangement, and necrotic features. Histological concordance supports model representativeness but cannot independently prove malignant origin because normal epithelial cells can also form polarized three-dimensional structures containing lumens.

7.2 Confirmation of Tissue Lineage

Tissue-specific markers can be used to determine organoid origin. For example, CDX2 and SATB2 support colorectal lineage, NKX2.1 and Napsin A support lung adenocarcinoma lineage, AR and NKX3.1 support prostate lineage, and PAX8 is expressed in renal and some female reproductive system tumors. Lineage markers are generally not malignancy-specific and should be interpreted together with pathological and genetic evidence.

7.3 Driver Alterations and Copy-Number Abnormalities

Driver mutations, gene fusions, amplifications, or large deletions shared between the original tumor and organoid provide important evidence of malignant identity. Variant allele frequency and subclonal proportion should be evaluated rather than only determining whether a mutation is present or absent. Normal-cell contamination, changes in tumor purity, and clonal selection can all alter the results.

7.4 Distinguishing Normal and Tumor Organoids

Normal epithelial organoids generally retain more complete tissue-differentiation programs and growth-factor dependence, but some tumor organoids can also display regular cystic structures. Reliable distinction requires integrated evaluation of histology, tumor-specific alterations, copy-number abnormalities, aneuploidy, factor independence, and drug responses and cannot be based on a single morphological or proliferation indicator.

7.5 Evaluation of Subclonal Composition and Heterogeneity

Multiregional sampling, single-organoid sequencing, single-cell analysis, or parallel establishment under different culture conditions can be used to assess subclone retention and selection. Pooled analysis of all organoids is useful for evaluating overall characteristics but may obscure low-frequency clones. Monoclonal expansion facilitates mechanistic studies but cannot represent the complete heterogeneity of the original tumor.

7.6 Functional Phenotypic Characterization

Functional evaluation of tumor organoids may include growth-factor independence, invasion, drug response, DNA damage, apoptosis, metabolism, and clonogenic regeneration. Functional readouts should directly correspond to the research objective and should be normalized to organoid number, area, cell number, or total protein. A single decrease in ATP-based viability is insufficient to distinguish cell death, growth arrest, and metabolic inhibition.

Table 2 Evidence Levels for Confirming the Malignant Identity of Tumor Organoids

8 Applications of Tumor Organoids in Disease Modeling, Drug Screening, and Personalized Therapy

8.1 Modeling Driver Genes and Tumor Evolution

Driver alterations can be introduced sequentially into normal tissue organoids, or specific abnormalities can be corrected in tumor organoids, to analyze the roles of individual mutations in tumor initiation and evolution. Isogenic models help distinguish the effects of a single alteration but still require exclusion of differences caused by clonal selection, gene-editing off-target effects, and culture adaptation.

8.2 Functional Drug Sensitivity Testing Within Clinical Time Windows

Biopsy-derived organoids may be used to compare candidate treatments while a patient is awaiting therapy, but the time required for establishment, expansion, and drug screening must fit the actual clinical decision window. Rapid drug testing should prioritize a limited number of clinically accessible drug combinations associated with the patient’s molecular subtype rather than attempting to test large numbers of compounds without clinical relevance.

8.3 Research on Resistance to Targeted Therapy and Chemotherapy

Pre- and post-treatment organoids can be used to compare molecular differences between sensitive and resistant clones. Resistance may arise from target mutations, activation of bypass pathways, drug efflux, enhanced DNA repair, or phenotypic transition. The generation of a resistant culture after prolonged drug exposure does not demonstrate that the same mechanism was already present in the patient’s original tumor.

8.4 Combination Therapy and Identification of Bypass Pathways

Tumor organoids are suitable for dose-matrix evaluation of drug combinations and can be integrated with phosphorylation signaling, cell-cycle, and apoptosis analyses to identify bypass pathways. A combination producing greater inhibition than either monotherapy does not necessarily indicate synergy. Drug interactions should be calculated across a complete dose range using an appropriate analytical model.

8.5 Differences Between Primary and Metastatic Lesions

Primary and metastatic lesions from the same patient may differ in subclonal composition, organ adaptation, and therapeutic sensitivity. Establishing organoids separately and testing them in parallel can identify resistance features found only in the metastatic lesion, but differences in sampling time and prior treatment must be included in interpretation.

8.6 Immune Cell Coculture

Tumor organoids can be cocultured with tumor-infiltrating lymphocytes, peripheral blood lymphocytes, macrophages, or engineered immune cells to evaluate immune recognition and cytotoxicity. Experiments must control HLA compatibility, effector-to-target ratio, antigen expression, and matrix barriers and should distinguish tumor-specific killing from nonspecific cellular injury.

8.7 Reconstruction of Tumor Stroma and Vascular Environments

Cancer-associated fibroblasts, endothelial cells, or defined matrices can be introduced to study paracrine signaling, drug penetration, and tumor invasion. Coculture outcomes depend strongly on the proportions of different cells and medium compatibility. Pure tumor organoid controls should therefore be retained to identify which cellular component is responsible for each observed change.

8.8 Limits of Correspondence Between Organoid Drug Sensitivity and Patient Outcomes

Organoids reflect the direct response of tumor cells to drugs but do not fully reproduce drug absorption, distribution, metabolism, immune effects, or dose-limiting toxicity. Personalized treatment decisions should integrate organoid drug responses with pathology, genomics, previous treatment, and clinically achievable drug exposure. In vitro sensitivity should not be interpreted as a guarantee of patient benefit.

9 Products Related to Tumor Organoid Establishment, Pathway Intervention, Drug Sensitivity, and Microenvironment Evaluation

9.1 Products for Tumor Tissue Dissociation and Three-Dimensional Model Establishment

9.2 Products for Tumor Clone Selection and Pathway Intervention

9.3 Products for Tumor Organoid Drug Sensitivity and Functional Evaluation

9.4 Products for Evaluating Tumor Matrix and Paracrine Signaling

The value of tumor organoids does not lie in simply reproducing the culture procedures used for normal organ organoids, but in preserving and identifying malignant clones, pathway dependencies, and therapeutic responses from patient tumors. By controlling culture selection pressure, confirming tumor identity, and comparing models established before and after treatment or from different lesions, the reliability of tumor organoids in resistance-mechanism studies and personalized drug evaluation can be improved.

For more related articles, please see below:

[1] Overview of Organs-on-Chips

[2] New Uses for Organoids