Patient-derived tumor organoids: A preclinical platform for personalized cancer therapy

Highlights

• •PDTOs represent a significant advancement in cancer research and personalized medicine, retaining genetic and molecular characteristics of original tumors.
• •The success rates of establishing PDTOs vary widely, influenced by factors such as cancer type, tissue quality, and media composition.
• •Recent studies have supported the correlation between PDTOs and corresponding patient responses, promoting their integration in clinical trials.
• •PDTOs preserve oncogenic signatures of their cancer type, maintaining disrupted transcriptional pathways associated with specific cancers.
• •Genomic evolution can occur in PDTOs during extended culture periods, potentially causing drift from the original tumors.
• •PDTOs can be engineered using CRISPR technology to model cancer development and progression.
• •Challenges in PDTO research include standardization of protocols, addressing tumor heterogeneity, and improving culture conditions for certain cancer types.
• •The development of organoid biobanks has facilitated large-scale studies and drug screening efforts, contributing to advancements in cancer research and drug discovery.

Abstract

Patient-derived tumor organoids (PDTOs) represent a significant advancement in cancer research and personalized medicine. These organoids, derived from various cancer types, have shown the ability to retain the genetic and molecular characteristics of the original tumors, allowing for the detailed study of tumor biology and drug responses on an individual basis. The success rates of establishing PDTOs vary widely and are influenced by factors such as cancer type, tissue quality, and media composition. Furthermore, the dynamic nature of organoid cultures may also lead to unique molecular characteristics that deviate from the original tumors, affecting their interpretation in clinical settings without the implementation of rigorous validation and establishment of standardized protocols. Recent studies have supported the correlation between PDTOs and the corresponding patient response. Although these studies involved a small number of patients, they promoted the integration of PDTOs in observational and interventional clinical trials to advance translational cancer therapies.

Graphical abstract

Introduction

The development of neoplasia results from the constant and evolving interaction between the malignant cells and their environment. The intricacy of these interplays has been gradually revealed by the progress made in genomics and proteomics. Over the last two decades since the publication of the first draft of the Human Genome Project, subsequent projects, including the Human Cancer Genome Project, have significantly influenced cancer research and care implementation. As of July 2024, the Precision Oncology Knowledge Base developed by Memorial Sloan Kettering contains annotations for 7810 alterations in 875 cancer-associated genes across 139 cancer types [1]. Between June 1998 and January 2024, 217 novel oncology drugs received FDA approval. Notably, 83.4% of these drugs (181 out of 217) fall into the category of targeted therapies. More than half (51.9%) of targeted therapies are classified as precision oncology therapies. However, as precision medicine in oncology expands its horizons to encompass genomics, proteomics, transcriptomics, and molecular imaging, significant challenges remain. Patients harboring these characterized actionable targets show varying responses to personalized therapies [2,3]. These variabilities underscore the incomplete grasp of the various genetic backgrounds and the identification of the patients to whom the targeted therapy would be beneficial [3]. In addition, a large number of cancers fail to express clinically actionable oncogenic drivers, lack well-defined biomarkers, or vary widely in their molecular characteristics. Examples include pancreatic, ovarian, esophageal, gastric, and triple-negative breast cancers, as well as sarcomas [[4], [5], [6], [7], [8], [9], [10]]. The challenge lies in bridging the gap between a patient tumor molecular profile and an effective treatment. Recently, patient-derived tumor organoids have proven fidelity toward the original tumor tissues [11], and have been tested as a preclinical platform, showing scalability [12]. In this review, we provide the latest data regarding tumor organoids, including their cultures, representativity, and their implications and relevance in clinical trial testing.

Origin of organoids

Cell lines and animal models have been pivotal in biomedical research from the late twentieth century to the present. They have significantly advanced the understanding of cellular signaling pathways, aided in identifying potential drug targets, and validated candidate drugs for pathologies like cancer and infectious diseases. Historically, disease investigation in animal models has progressed from genetic screens in invertebrates, followed by mammalian models, and ultimately clinical translation to humans [13]. These molecular principles, conserved through evolution, have yielded detailed insights into many human diseases. However, the critical challenge lies in correlating the results obtained from animal disease models to humans. Recent studies have highlighted biological processes unique to the human body, which cannot be accurately modeled in other animals [14,15]. Examples include distinctive organ and/or tissue development, cellular metabolism [[16], [17], [18]], and cellular physiology, especially in the context of cancers [19,20]. The emergence of organoids as in vitro 3D cell culture systems based on stem cells offers solutions to overcome some limitations.

Organoids are self-organized miniature, simplified versions of organs or tissues developed in vitro. To some extent, the organoids mimic the structure, cellular composition, and function of the intended modeled organs or tissues [[21], [22], [23]]. All organoids arise from either pluripotent stem cells or adult stem cells [24]. The development of organoid culture owes much to the groundbreaking work of James G. Rheinwald and Howard Green in 1975. Their research demonstrated that a co-culture of primary human keratinocytes and mouse 3T3 fibroblasts could be self-organize in vitro, producing squamous epithelial colonies with a proliferative basal layer and keratinized cells in the upper layer that closely mimicked the human epidermis architecture [25]. Subsequent efforts focused on reproducing the environmental cues necessary for precise control over cell-cell and cell-matrix interactions. This research led to the creation of matrigel, consisting of a basement membrane extracellular matrix composed of components and growth factors, extracted from mouse sarcoma tumors, which was shown to support cell culture in vitro [26]. Matrigel supported the 3D development of murine breast epithelial cells, which form lumens with milk protein secretion [27]. Matrigel also supported the growth of normal and malignant human breast epithelial cells [28]. Despite these early breakthroughs, progress in organoid technology remained slow for nearly two decades due to challenges related to culture conditions, tissue-specific factors, and long-term viability. In 2009, Hans Clevers’ group achieved a significant breakthrough by successfully producing mouse intestine organoids [21]. These organoids, derived from the intestinal epithelium, contain all differentiated cell types and can be initiated from single Lgr5+ stem cells. Remarkably, these self-organizing structures can form in vitro even without a non-epithelial cellular niche [21]. Sato et al. identified the Wingless-related integration site (Wnt) signaling pathway as a pivotal requirement for stem cell proliferation, differentiation, and long-term maintenance of the organoids, which includes a Wnt agonist (R-spondin-1), Noggin to block BMP (bone morphogenic protein) activity, and epidermal growth factor (EGF) as a mitogen [21]. Later, the organoid media was revamped to accommodate human intestinal stem cell origin by adding growth factors such as gastrin and Wnt3A, small molecule precursors such as nicotinamide, and small molecule inhibitors like A83-01, specifically targeting the TGF-β (transforming growth factor-beta) signaling pathway, and SB202190 which selectively inhibits p38 MAPK (mitogen-activated protein kinase) signaling [29]. Following these landmark studies, multiple growth factor formulations were optimized to promote the development of organoids from various tissues and organs (Table 1). The organoids are generated from either adult stem cells (AdSCs), pluripotent stem cells (PSCs), including embryonic stem cells or their synthetic equivalent, induced pluripotent stem cells (iPSCs), or cancer stem cells (CSCs). Organoids were developed to model organ/tissue development, physiology, diseases, or cancers. They were also used to recapitulate the air-liquid interface in specific organs such as the intestine, stomach, pancreas, lung, and brain [[30], [31], [32], [33]]. They recapitulate the intra- and inter-heterogeneity of tissues and are valuable in mimicking cancer development in an approach tailored to each patient [[34], [35], [36], [37]]. The intra-heterogeneity in each patient-derived cancer organoid reflects the varying complexity of the tumor cell clonal population, ranging from stem cell-like properties to more differentiated states, and instills new knowledge about the interactions of these various clonal populations and their selection following treatment. Organoids further advanced our understanding of cancer progression and therapeutic responses, highlighting the importance of tailored organoid models in cancer research as highlighted by the work in esophageal cancer and pancreatic cancer [38,39].

Table 1.

Development of organoids to model organ and disease function using various stem cell sources.

Organ models	Cell Source	Tissue	Cancer	Reference
Adrenal glands	AdSC, iPSC	✓		[40,41]
Appendix	AdSC	✓		[42]
Bladder	AdSC, CSC	✓	✓	[34,[43], [44], [45], [46]]
Blood Vessels (Arteries, Veins, Capillaries)	iPSC	✓		[47]
Bone Marrow	iPSC	✓		[48]
Brain (Forebrain, Midbrain, Hindbrain, Brainstem, spinal cord)	iPSC	✓	✓	[[49], [50], [51], [52], [53]]
Breast Tissue	AdSC, CSC	✓	✓	[37,54,55]
Cartilage	iPSC	✓		[56]
Cervix	AdSC, CSC	✓	✓	[57,58]
Colon	AdSC, CSC, iPSC	✓	✓	[[59], [60], [61]]
Cornea	iPSC, AdSC	✓		[62,63]
Esophagus	AdSC, CSC	✓	✓	[[64], [65], [66]]
Eyes (Retina, Lens)	iPSC	✓		[67,68]
Fallopian tube	iPSC, AdSC, CSC	✓	✓	[69,70]
Gallbladder	AdSC, CSC	✓	✓	[71,72]
Heart	PSC	✓		[73]
Small intestine (duodenum, jejunum, ileum)	AdSC	✓		[74,75]
Kidneys	AdSC, PSC, CSC	✓	✓	[[76], [77], [78]]
Liver	AdSC, PSC, CSC	✓	✓	[[79], [80], [81]]
Lungs	AdSC, PSC, CSC	✓	✓	[[82], [83], [84]]
Lymphoid Tissues (lymph nodes, tonsil, spleen)	AdSC	✓		[85]
Muscle (skeletal)	iPSC	✓		[86]
Nasal epithelium	AdSC	✓		[87]
Ovaries	AdSC, iPSC, CSC	✓	✓	[[88], [89], [90]]
Pancreas	PSC, AdSC, CSC	✓	✓	[[91], [92], [93]]
Parathyroid glands	PSC, AdSC	✓		[94,95]
Placenta	AdSC	✓		[96]
Prostate	iPSC, AdSC, CSC	✓	✓	[97,98]
Rectum	AdSC, CSC	✓	✓	[99]
Salivary glands	iPSC, AdSC	✓		[100,101]
Skin	PSC, AdSC, CSC	✓	✓	[102,103]
Stomach	PSC, AdSC, CSC	✓	✓	[[104], [105], [106]]
Testes	PSC, AdSC, CSC	✓	✓	[107,108]
Thymus	PSC	✓		[109]
Thyroid gland	PSC, AdSC, CSC	✓	✓	[110,111]
Urethra	PSC, AdSC, CSC	✓	✓	[43,112]
Uterus	PSC, AdSC, CSC	✓	✓	[23,113]

Abbreviations: AdSC: adult stem cell; CSC: cancer stem cell; PSC: Pluripotent stem cell; iPSC: induced pluripotent stem cell

Patient-derived tumor organoids (PDTOs) for cancer research

The widespread adoption of organoids can be attributed to their exceptional establishment rates from primary cancer tissues. Furthermore, organoid models can be directly derived from affected patients without needing prior knowledge of the specific genes involved. This approach is particularly useful for multigenic disorders like cancers, assuming the pathology originates in the affected epithelium. This innovative technology facilitates the detailed characterization of cancer cells on a patient-specific basis. Additionally, organoids can be maintained in a state enriched with stem cells, owing to their intrinsic physiological properties, making them invaluable for studying cancer stem cells (CSCs).

The first patient-derived tumor organoids (PDTOs) were developed from primary colon adenocarcinoma [29], and this paved the way for the generation of organoids from a wide variety of tumors (Table 1). The success rate of establishing organoids from cancer tissues varies significantly depending on the type of cancer and the quality of the tumor sample, the method of tissue processing, and the specific culture conditions used. Generally, patient-derived tumor organoids (PDTOs) have shown an establishment success rate above 50% for most tumors (for review see [114,115]. Sachs et al. created a biobank of breast cancer organoids from 155 specimens, achieving a success rate of over 80% in generating viable organoid lines that reflected the heterogeneity of the original tumors [37]. However, organoids derived from highly metastatic prostate cancer samples have a low establishment success rate of about 16% [116].

These PDTOs closely represent the genetic and molecular characteristics of the original tumors. This fidelity allows for more accurate modeling of cancer behavior and treatment responses. PDTOs exhibit morphological characteristics that closely resemble the original patient tissues. For instance, colorectal cancer organoids maintain cystic or solid features similar to those observed in patient samples [117]. The histology of breast cancer patient-derived organoids (bcPDTOs) shows notable similarities to the breast cancer from which they were derived, matching cell dimensions, shapes, and intercellular organization, indicating a complex structure that is more representative of in vivo conditions than traditional 2D cultures [118]. The bcPDTOs and the tumor tissues lacked a basement membrane and myoepithelial layer, typically in healthy mammary tissue [118]. This absence is a common feature in tumor tissues, suggesting that the bcPDTOs accurately reflect the histological state of the cancerous tissues [118]. Histological similarities between the PDTOs and original cancer tissues were also observed for many cancers, including gynecologic [119], colorectal [120], pancreatic [121], lung [122], prostate [123]and gastric cancers [124]. Additionally, these PDTOs retained the expression patterns found in the original patient cancers, preserving the identity of the original tumors’ histological subtypes as evidenced by specific markers such as p53, Human Epidermal growth factor Receptor 2 (HER2), and the estrogen receptor (ER) (for review [125]). As an example, the status of ER, progesterone receptor (PR), and HER2, the commonly assessed markers in breast cancers, as well as Ki-67, were assessed in both organoids and matched tumors [126]. Immunohistochemical staining results demonstrated that the expression patterns of these breast cancer markers were well preserved in both drug-treated and treatment-naïve tumor-derived organoids [126]. The same study demonstrated the stability of these markers in some long-term organoid cultures, while a drift was noticed in others with cells changing from ER-negative to ER-positive [126]. These findings underscore the fidelity of PDTOs in maintaining the histological and molecular characteristics of the original tumors. Many studies found that newly established PDTOs preserved the oncogenic signatures of their cancer type, preserving the disrupted transcriptional pathways associated with specific cancers. This characteristic is crucial for studying the biology of different cancer types and their responses to therapies. PDTOs established from head and neck squamous cell carcinoma (HNSCC) retain their tumorigenic potential and recapitulate the epithelial tumors' histology, genetic, and molecular characteristics. The organoids maintain their identity and do not drift away from their original characteristics during culture and experimentation [127]. Larsen et al. created PDTOs from over 1,000 cancer patients, successfully recapitulating the genomic and transcriptomic profiles of the source tumors. On average, the somatic recapitulation rates for tumor variants (with a variant allele fraction [VAF] >10%) were greater than 76.9% and copy number variation (CNV) concordance was over 77.6% [128]. Another study demonstrated that organoids derived from pancreatic ductal adenocarcinoma conserved the mutations found in the primary tumors at rates ranging from 82.49 to 99.46% in the early passages [92]. However, PDTOs frequently experience genomic evolution during extended culture periods and drift from the original tumors. For instance, liver cancer organoids maintained approximately 92% of their original somatic mutations after two months of in vitro culture; this preservation rate dropped to 80% after four months [129]. The nature of the mutations also influences the stability of the cancer-derived organoids. For example, colon cancer organoids derived from tumors with mismatch-repair deficiency accumulated numerous new mutations after six months in culture, whereas their mismatch-repair proficient counterparts maintained a much more stable genome over the long term [130]. In essence, these patterns could be attributed to stochastic neutral drift, might result from selection pressures in culture, and/or could indicate the activity of one or more oncogenic drivers within tumor subclones that facilitate clonal competition [130]. Despite these challenges, the ability of PDTOs to preserve key oncogenic signatures makes them invaluable for personalized medicine and drug development. Their use in large-scale studies has demonstrated their potential to provide insights into tumor behavior and treatment efficacy, underscoring their importance in advancing cancer research.

CRISPR engineering in organoids for cancer modeling

The utilization of CRISPR/Cas9 technology has significantly advanced the field of organoid research, providing innovative methods for studying genes associated with the onset of cancer and resistance to treatments [131]. Table 2 provides a snapshot of the critical mutations commonly observed in different cancer types studied using PDTOs. CRISPR/Cas9 implementation has provided valuable insights into cancer initiation, progression, and treatment resistance by targeting essential genes associated with various cancer types. For instance, Matano et al. generated mutations using CRISPR-Cas9 genome-editing system in multiple tumor suppressor genes APC (Adenomatous Polyposis Coli), SMAD4 (SMAD Family Member 4) and TP53 (Tumor Protein p53), and in the oncogenes KRAS (Kirsten Rat Sarcoma Viral Oncogene Homolog) (G12V) and/or PIK3CA (Phosphoinositide-3-Kinase Catalytic Subunit Alpha) (E545K) to understand their implications in colorectal cancer development [132]. The study developed organoids expressing all five mutations that acquired a cancer behavior and grew independently of niche factors in vitro and formed tumors when implanted under the kidney subcapsule in mice. However, after being injected into the spleen of mice, they only produced micrometastases with dormant tumor-initiating cells and did not colonize the liver. In contrast, organoids derived from chromosome-instable human adenomas successfully formed macrometastatic colonies. The study identified that while ‘driver’ pathway mutations support stem cell maintenance in the challenging tumor microenvironment, additional molecular changes are necessary for invasive behavior [132]. This highlights the complexity of colorectal cancer progression and the need for multiple genetic alterations to achieve full malignancy. Several studies have subsequently explored the application of CRISPR technology for gene editing in PDTOs (Table 2). Two studies successfully generated isogenic models for PDAC using human adult stem cell (AdSC)-derived ductal pancreas organoids. They employed CRISPR-based multiplexed mutation to target genes such as TP53, CDKN2A, and SMAD4, combined with the oncogenic KRAS G12V mutation, resulting in organoids that exhibited PDAC phenotypes [133,134]. CRISPR-Cas9-mediated knockouts of tumor suppressor genes like TP53, SMAD4, PTEN, NF1, and BAP1 were created in cholangiocyte organoids [135]. CRISPR-Cas9 was used to knock out four breast cancer-associated tumor suppressor genes (TP53, PTEN, RB1, NF1) in mammary progenitor cells from six donors [136]. The resultant mutant organoids developed estrogen-receptor-positive luminal tumors when transplanted into mice, which occurred in one of six P53/PTEN/RB1–mutated lines and three of six P53/PTEN/RB1/NF1–mutated lines. These organoids also showed responses to endocrine therapy and chemotherapy, highlighting the model's potential for advancing our understanding of the molecular events that drive specific breast cancer subtypes [136]. Independent studies utilized CRISPR-Cas9 to create knockout models of DNA repair genes in human colonic organoids. For instance, loss-of-function mutations in MLH1 (MutL homolog 1) were analyzed to understand mutational signatures associated with DNA replication errors [137]. In human colon organoids, researchers modeled common fusion genes (e.g., DLG1-BRAF, PTPRK-RSPO3) using CRISPR. Co-transfection of two sgRNAs targeting both loci resulted in complex genomic rearrangements, particularly in organoids lacking TP53 expression [138]. CRISPR-Cas9 strategies were also applied to create single or double-mutant isogenic knockout models of TP53 in human hepatocyte organoids, ARID1A in gastric organoids, and RB1 in intestinal organoids to model various cancers [105,139,140]. More recently, the CRISPR-Cas9 strategy was also initiated to study the genes involved in the onset of lung cancer [141].

Table 2.

Cancer progression models using CRISPR-Cas9 in PDTO studies.

Cancer Type	Key Mutations	References
Colorectal Cancer	APC, SMAD4, TP53, KRAS, PIK3CA	[132]
	MHL1, NTHL1	[137]
	PTPRK-RSPO3 fusion, DLG1-BRAF fusion	[138]
	TP53, RB1	[140]
Liver cancer	TP53	[139]
Pancreatic Cancer	KRAS, TP53, CDKN2A, SMAD4	[133,134]
Gastric cancer	ARID1A	[105]
Cholangiocarcinoma	TP53, SMAD4, PTEN, NF1, BAP1	[135]
Breast Cancer	TP53, PTEN, RB1, NF1	[136]
Lung Cancer	TP53, RB1	[141]

Abbreviations: APC - Adenomatous Polyposis Coli; SMAD4 - SMAD Family Member 4; TP53 - Tumor Protein p53; KRAS - Kirsten Rat Sarcoma Viral Oncogene Homolog; PIK3CA - Phosphatidylinositol-4,5-Bisphosphate 3-Kinase Catalytic Subunit Alpha; MLH1 - MutL Homolog 1; NTHL1 - Nth Like DNA Glycosylase 1; PTPRK- Protein Tyrosine Phosphatase Receptor Type K; RSPO3 - R-spondin 3; DLG1- Discs Large MAGUK Scaffold Protein 1; BRAF - B-Raf Proto-Oncogene; RB1 - RB Transcriptional Corepressor 1; CDKN2A - Cyclin Dependent Kinase Inhibitor 2A; ARID1A - AT-Rich Interaction Domain 1A; PTEN - Phosphatase and Tensin Homolog; NF1 - Neurofibromin 1; BAP1 - BRCA1 Associated Protein 1.

A different approach utilized CRISPR-based editing to correct oncogenic mutations in PDTOs. Sayed et al. corrected oncogenic mutations in KRAS and TP53 genes [142]. The KRAS mutation G12D was edited significantly, with over 30% of the targeted adenine bases being correctly edited to guanine (A>G). The correction indicated an alteration of the growth of cancer organoids [142]. The mutation of TP53 (p.R175H, c.524G>A) in a colorectal carcinoma also indicated that the corrected cells were effectively depleted and outcompeted by non-infected cells, highlighting the functional impact of restoring the wild-type TP53 sequence. The recovery of even one wild-type allele of the hemizygous TP53 mutation was sufficient to reactivate TP53 function. This reactivation likely triggered a sharp depletion of the corrected cells through the upregulation of cell-cycle and apoptosis regulator genes, such as p21 and PUMA [142]. Overall, the study demonstrates that base editing can effectively correct mutations in PDTOs, providing insights into potential therapeutic strategies.

PDTOs and cancer stem cells (CSCs)

Two principal models have been proposed to explain the cellular origin of cancer. The stochastic model suggests that a series of mutational events in differentiated somatic cells gradually enables reprogramming and the acquisition of a malignant genotype. This process either confers a stem cell-like status to these cells or, in a few cancers such as pancreatic cancers, creates transiently active tumor-initiating cells [[143], [144], [145]]. In contrast, the second hypothesis suggests that mutations occur in stem cells or progenitors [146]. It is also plausible that both models could coexist. The longevity, plasticity, and quiescence of stem cells, and to some extent progenitors, are crucial for accumulating serial mutational events [147]. Mutations in progenitor cells may also lift their restricted potential and promote reprogramming and tumor initiation when provided with a stimulatory microenvironment. These opportunistic stem cell niches disrupt the traditional hierarchical organization of stem cells [148]. Most tumors harbor CSCs in dedicated niches, yet their identification and targeting have not been obvious [149]. Organoids provide the desired ex-vivo platform to study these rare cell populations. Cancer stem cells (CSCs) are critical in tumor recurrence and resistance to conventional therapies like chemotherapy and radiation, as they often remain quiescent and are not effectively targeted by these treatments [150]. Recent advancements in PDTOs have emerged as a promising tool for studying CSCs and their interactions with therapies. PDTOs closely mimic the original tumor's genetic and molecular characteristics, allowing for personalized drug testing and a better understanding of treatment responses [151]. They have shown potential in predicting patient-specific responses to chemotherapies, particularly in colorectal cancer, where they can model the tumor microenvironment and CSC-driven resistance mechanisms [151].

In colorectal PDTOs, CSCs were characterized by the expression of LGR5 (Leucine-rich repeat-containing G protein-coupled receptor 5). The study demonstrated the contribution of LGR5+ CSCs to tumor growth and maintenance but also highlighted the complex dynamics of tumor biology [152]. The study showed the capacity of this cell population to self-renew and differentiate, demonstrated the plasticity of the colorectal cancer cells in their ability to revert to a stem cell-like state upon the ablation of LGR5+ CSCs, and the formation of a new population of LGR5+ CSCs [152]. Overall, the study illustrated the challenge of targeting the CSC population and the plasticity of cancer cells. The presence of CSCs in PDTOs was demonstrated across various tumor types and identified by the expression of markers, including CD44, CD24, CD133, CD326, CD54, Aldehyde dehydrogenase 1 (ALDH1), CD166, CD49f, and CD9, expressed either alone or in combination [153,154]. These PDTOs reflect the heterogeneity of cancers and underscore the relevance of organoids in modeling CSC behavior and their response to therapies.

Different types of PDTO culture

PDTOs provide numerous benefits, such as high throughput, efficiency, and preservation of epithelial tumor heterogeneity. PDTO has traditionally utilized basement membrane extract (BME), such as laminin-rich Matrigel or BME-2, to embed cancer cells or tissues [29]. This material provides a supportive extracellular matrix that mimics the natural tissue environment, promoting the growth and differentiation of cells within a 3D culture system. The organoid cultures are maintained submerged in the tissue culture medium. This technique has been crucial for cancer biology studies, although it mainly focuses on epithelial cells and lacks stromal components [155]. PDTOs offer several advantages when compared to traditional preclinical models such as organotypic slice cultures and patient-derived xenograft (PDX) models. PDX model involves implanting patient tumor tissue into immunocompromised mice and are crucial for translational studies as they capture genetic diversity and patient tumor physiology. However, the high costs and extended time requirements over several months make serial PDX modeling through treatment and tumor evolution impractical [156]. Additionally, the murine stromal tumor microenvironment and the need for immune reconstitution in PDXs present challenges for immuno-oncology research, necessitating the development of other preclinical models. Organotypic slice cultures maintain the heterogeneity of the tumor and stroma, thereby accurately representing the tumor in its native environment, including epithelial/tumor cells, the entire extracellular matrix (ECM), and stromal and immune cells [157]. In this manner, all cells preserve their functions such as hormone secretion, vascular contractility, cytokine secretion, and proteome and [158]. When compared to PDTOs, organotypic slice culture faces several limitations, including limited growth potential, reproducibility, scalability, and tissue heterogeneity (for review, see [159])

The establishment of PDTOs faces challenges such as obtaining sufficient tumor tissue, variable success rates in establishing PDTOs, and the need for standardized protocols. Establishing PDTOs can take several weeks, causing potential delays in treatment initiation, and growth rates vary significantly between patients. Potential solutions to address these issues include developing standardized processes and integrating advanced imaging and modeling techniques to understand and predict PDTO growth dynamics better. Overcoming these bottlenecks is crucial for effectively incorporating PDTOs into clinical practice and enhancing personalized treatment strategies for cancer patients.

For achieving physiological relevance in organoid culture systems, air-liquid interface (ALI) systems are particularly effective for modeling tissues interacting with air and liquid, such as lung and gut tissues [155]. ALI systems have been successfully used with various PDTOs, including lung cancer [160], esophageal cancer [161], pancreatic cancer [162], colorectal cancer [162,163], and renal cancer [164]. These systems maintain the architecture of the tumor microenvironment (TME), encompassing both tumor parenchyma and stroma, along with functional, tumor-specific tumor-infiltrating lymphocytes (TILs) [162]. In an ALI system, tumor fragments are embedded in a collagen I matrix seeded on a transwell with their basal surface submerged in culture medium, while the apical surface remains exposed to air [165]. Unlike the tumor epithelium, which can be repeatedly passaged and cryopreserved, the immune component and other cellular components of ALI PDTOs diminish over time [162].

Microfluidic systems, or organ-on-a-chip technologies, offer high physiological relevance by precisely controlling the microenvironment and simulating organ physiology through fluid flow, mechanical forces, and nutrient gradients [166]. These systems use biocompatible materials to create scaffold-based cultures replicating the native tissue's mechanical and chemical properties, thus supporting cells' structured growth and differentiation (for review [167]). PDTOs have been integrated into various organ-on-a-chip systems; for example, glioblastoma PDTOs have been used to assess PD-1 immunotherapy efficacy [168], lung cancer PDTOs to test drug sensitivity [169], and HER2+ breast cancer PDTOs associated with cancer-associated fibroblasts, immune cells, and endothelial cells to test Trastuzumab response [170]. Current organ-on-a-chip systems face challenges such as a lack of standardization procedures and often have low throughput, making it difficult to conduct large-scale studies or high-throughput drug screening. Despite these limitations, organ-on-a-chip technology holds significant potential for advancing PDTO studies and enhancing our understanding of tumor biology and treatment responses.

Relevance of PDTOs in the clinic

In recent years, PDTOs have emerged as a promising tool for extensive drug screening aimed at personalizing treatment for various cancer types. A standardized, high-quality PDTO biobank is essential for preserving valuable genetic resources. It is pivotal in basic and clinical research on major diseases, the development of clinical diagnostic and treatment technologies, drug research and development, and the advancement of precision and translational medicine. These biobanks provide a robust platform for drug screening and testing, allowing the evaluation of the efficacy of established and new therapies in a personalized manner, as PDTOs often recapitulate the drug response observed in patients (Table 3). Additionally, PDTO biobanks facilitate the study of tumor heterogeneity and the genetic underpinnings of cancer, enabling the exploration of individual drug resistance mechanisms and the identification of valuable targets [92]. The first PDTO biobank was established for colorectal cancer in 2015 and was used for drug screening [117]. Biobanks were then derived from many types of cancer, including brain [171], breast [37], gastrointestinal [130,172,173], lungs [122,174], and others. There are now several prominent PDTO biobanks, both public and private; one of the largest public organoid biobanks was established by The Hubrecht Institute, the University Medical Center Utrecht, and the Royal Netherlands Academy of Arts and Sciences (KNAW) as part of Hubrecht Organoid Technology (HUB, huborganoids.nl). This collection includes over 1,000 organoid lines derived from various organs and diseases, such as breast, colon, head, neck, intestine, liver, lung, ovarian, and pancreatic tumors [175].

Table 3.

Correlation between PDTO treatment effect and patient clinical outcomes across various cancer types.

Cancer type	Classification	Sample size	Comments	Reference
Bladder	Non-muscle invasive Muscle invasive	2	PDTOs can predict drug responses and identify effective therapies for patients.	[177]
Brain tumor metastasis	Oligodendroglioma (OA), Grade III OA, Grade II. Colon cancer metastasis to the brain.	3	Treatments identified through PDTO testing can improve patient outcomes. For example, one patient treated based on PDTO results experienced complete tumor regression and remained free of recurrence for over 15 months.	[183]
Breast	Her2+ (2) TNBC (2) Luminal B (1)	5	In HER-2+ patients, the response of the tissue-derived organoids to the drugs was consistent with the patient's clinical response. The organoids derived from triple-negative and luminal B-types responded like parental tumors.	[178]
	Luminal B TNBC Her2+	6	PDTO pharmaco-phenotyping results guided treatments. The clinical outcomes were partial response, stable disease, or disease-free.	[126]
Colorectal	Metastatic	22	Patients with high levels of circHAS2 expression demonstrated a notable sensitivity to anlotinib treatment. In a phase 2 clinical trial, 14 patients expressing high levels of circHAS2 yielded meaningful clinical remission, with a DCR of 100.0% and an ORR of 85.7%	[181]
	Metastatic	9	Sensitivity was limited to 66.7%, with five false negative cases in total. The most significant potential for standard-of-care PDTO testing lies in selecting first- and second-line treatments, especially when multiple viable therapy options exist.	[185]
	Metastatic	34	At the two-month mark, 17 out of 34 patients achieved progression-free survival, surpassing the pre-defined minimal relevant difference compared to historical controls from randomized trials.	[184]
		11	The research reported an approximately 80% concordance rate between the responses of PDTOs and the clinical responses observed in donor patients.	[190]
	Metastatic	3	The study demonstrates a clear correlation between PDTO drug treatment responses and patients' original tumor responses, supporting using PDTOs as a preclinical platform for precision medicine in colorectal cancer.	[191]
	Metastatic	6	No lasting clinical responses from treatment decisions guided by organoids. No correlation between PDTOs and patient outcomes.	[188]
	Metastatic	71	The patient-derived tumor organoid model demonstrated a sensitivity of 63.33%, a specificity of 94.12%, and an accuracy of 79.69% in predicting responses to chemotherapy regimens.	[180]
	Metastatic	9	A correlation between patient outcomes and PDTO results was observed with Gemcitabine and Regorafenib. FOLFOX did not correlate strongly with patient outcomes based on PDTO results.	[187]
	Metastatic	10	PDTOs to predict the outcomes of irinotecan-based chemotherapy is clinically feasible.	[186]
	Metastatic	1	This study is among the first to prospectively investigate the use of PDTOs to predict treatment responses for individual cancer patients.	[192]
Esophagogastric adenocarcinoma		13	The study defined a PDTO-derived threshold value that accurately classified patients into pathological responders and non-responders with high sensitivity (90%), specificity (100%), and accuracy (92%) based on the FLOT response. The receiver operating characteristic (ROC) analysis for FLOT testing yielded an area under the curve (AUC) of 0.994.	[179]
Gastric		3	An example of direct correlation is when one patient who received a 5-FU-based drug post-operatively exhibited progressive disease, and the PDTO culture also did not respond to the same treatment. In another example, two patients. who were treated with cisplatin and 5-FU-based drugs after surgery showed substantial tumor shrinkage.	[193]
Gastrointestinal	Metastatic	14	The PDTOs showed 100% sensitivity, 93% specificity, 88% positive predictive value, and 100% negative predictive value in forecasting response to targeted agents or chemotherapy in patients.	[172]
Glioblastoma		7	Three out of seven patients exhibited a reduction in Ki67+ cells following treatment with temozolomide and radiation, indicating a positive response to the therapy. However, the study did not consistently correlate PDTO responses with MGMT methylation status, a recognized predictive marker for temozolomide response. This implies that, although some individual cases showed encouraging correlations, a broader, definitive connection between PDTO treatment responses and the original tumor characteristics was not established.	[171]
Head and neck		21	In the adjuvant treatment setting, PDTO viability at 2 Gy showed the strongest correlation with clinical response (n = 15), although this correlation was not statistically significant. For patients who received primary RT, no statistically significant correlation was observed between organoid response and clinical relapse (n = 6).	[194]
Lung	Adenocarcinoma Squamous cell adenocarcinoma Small cell lung cancer Other types	21	The study successfully correlated the drug responses of lung PDTOs derived from 21 patients with their respective clinical outcomes and genetic profiles, highlighting the potential of this approach in personalized medicine for lung cancer treatment.	[122]
	Adenocarcinoma Squamous cell adenocarcinoma	34	The PDTOs maintained the sensitivity of the corresponding parental tumors to targeted therapies, demonstrating their usefulness in validating or identifying biomarker-drug combinations.	[174]
Mesothelioma		2	Patient 1 showed a strong response to cisplatin/ pemetrexed, which aligned with the patient's clinical response to this treatment. Patient 2 did respond to carboplatin and pemetrexed; the response was not strong.	[195]
Ovary	High grade serous	1	Drug responses in matched organoid lines derived from a single patient's primary chemosensitive and recurrent chemoresistant ascites.	[88]
		5	Same PDTOs as [88]. PDTO drug response often reflects patients’ clinical response. However, PDTOs derived from different tumor locations within the same patient demonstrated differential drug responses. Additionally, organoids established from the same patient at different time points showed variability in their responses to drugs, indicating that tumor heterogeneity can significantly affect treatment outcomes.	[189]
Pancreas		9	Six patients with progression-free survival (PFS) longer than the median were treated with at least one drug that their PDTO was sensitive to. This resulted in a mean PFS of 332 days, significantly higher than the expected 180 days.	[92]
Prostate	Neuroendocrine (metastatic)	4	Two of the four PDTOs exhibited treatment responses consistent with the patient's clinical outcomes. One showed significant sensitivity to the aurora kinase inhibitor alisertib, which aligned with the patient's response in a Phase 2 trial of the same drug. The other one did not respond to alisertib, mirroring the lack of response observed in the patient during the clinical trial. However, this organoid did respond to the MEK inhibitor cobimetinib, indicating that while not all PDTOs matched perfectly, there were notable correlations in treatment responses.	[182]
Rectal		80	Irradiation: 16 patients with organoids sensitive to irradiation responded well to neoadjuvant chemoradiation, while 64 patients with resistant organoids had varied responses. 5-FU: Exposure to 10 mM 5-FU showed significant variability in PDTO responses. Among 53 patients with 5-FU-resistant PDTOs, 38 had poor responses, and 15 had good responses. Irinotecan: 32 PDTOs were sensitive, with most patients responding well, except for 7 cases. Overall, 34 patients had good clinical responses, and their PDTOs were sensitive to at least one treatment (irradiation, 5-FU, or irinotecan). Conversely, 34 patients responded poorly, with their organoids resistant to all treatments. The analysis showed a strong correlation (AUC of 0.882) between PDTO data and clinical outcomes.	[173]
		7	Seven had sufficient clinical follow-up to provide complete progression-free survival (PFS) data. PDTOs exhibited a range of radiation sensitivities, corresponding to the clinical responses observed in the patients.	[99]

Abbreviations: PDTO: patient-derived tumor organoid; OA: oligodendroglioma; Her2: human epidermal growth factor receptor 2; TNBC: triple-negative breast cancer; DCR: disease control rate; ORR: objective response rate; FOLFOX: folinic acid, fluorouracil (5-FU), oxaliplatin; ROC: receiver operating characteristic, AUC: area under the curve; FLOT: Fluorouracil, leucovorin, oxaliplatin, Taxotere; MGMT: O-6-methylguanine-DNA methyltransferase; RT: radiation therapy; MEK: mitogen-activated protein kinase kinase; PFS: patient-free survival.

Chemotherapy is currently used either as a standalone treatment or in combination with radiation or surgical resection. Although some patients initially respond well to chemotherapy, tumors often develop resistance during treatment. Consequently, research on personalized anticancer treatment has intensified across all cancer types [176]. PDTOs have been extensively used to assess drug sensitivity and resistance in various cancers, hoping to offer a more relevant and accurate testing platform. The correlation between PDTO responses and clinical outcomes has been demonstrated across various cancer types, exemplifying the capacity of PDTOs to inform patient treatments and predict outcomes (Table 3). For instance, PDTOs can identify effective therapies in bladder cancer, leading to better treatment outcomes for non-muscle-invasive and muscle-invasive cases [177]. Consistency between PDTOs and patient outcomes has also been demonstrated in breast cancers [126,178] (Table 3). A study on metastatic gastrointestinal cancers involving 14 patients showed that their respective PDTOs displayed 100% sensitivity, 93% specificity, 88% positive predictive value, and 100% negative predictive value in forecasting response to targeted agents or chemotherapy in patients [172]. A correlation above 60% in sensitivity and 90% in specificity was also reported for PDTOs derived from esophagogastric adenocarcinoma [179] and metastatic colorectal cancer [180,181] (Table 3). Correlation was also established for chemotherapeutic agents and radiation therapy (Table 3).

The potential of PDTOs to inform patient outcomes is promising. However, it is hindered by the limited amount of validated data linking the efficacy of PDTO treatments to corresponding patient outcomes (Table 3). In addition, the representativeness of the PDTO treatment to patient response is hampered because some cancers have low PTDO establishment efficiency from primary tissues, such as neuroendocrine metastatic prostate cancer with an efficiency of 16% [182], metastatic brain tumors with an establishment rate of only 50% [183], or colorectal cancer, where PDTO establishment efficiency varies from 49 to 81% depending on the study [180,[184], [185], [186], [187], [188]]. The ability to predict patient-specific treatment outcomes diminishes, as the limited number of organoids may not capture the full range of tumor heterogeneity. Ovarian PDTO heterogeneity in drug responses between patients and within the same tumor was demonstrated by de Witte et al. [189]. The PDTOs exhibited varying responses to different drugs, which were partially correlated with the mutational profiles of the patients. This suggests that genetic variations among patients can influence how their cancer responds to treatment [189]. Furthermore, PDTOs derived from different tumor locations within the same patient demonstrated differential drug responses. Additionally, PDTOs established from the same patient at different time points showed variability in their responses to drugs, indicating that tumor heterogeneity can significantly affect treatment outcomes and the interpretation of patient sensitivity [189].

Moreover, single-cell RNA sequencing has identified distinct expression patterns in older organoids, indicating that extended culture periods may result in the development of unique molecular characteristics not found in the original tumors. These findings underscore the dynamic nature of organoid cultures and their potential limitations in accurately portraying the original tumor in culture conditions lasting several months (for review, see [125]).

While PDTOs hold significant promise for personalizing cancer treatment and advancing our understanding of tumor biology, several challenges remain. The limited amount of validated data linking PDTO treatment efficacy to patient outcomes and the low establishment efficiency for certain cancer types hinder their full potential. Addressing these issues through improved biobanking practices and more extensive research will be crucial for maximizing the clinical relevance of PDTOs and enhancing their role in precision medicine. Large prospective studies are necessary to demonstrate the versatility of PDTOs for predicting patient response. Many clinical trials have been registered for various types of cancers in recent years.

Clinical trial landscape of PDTOs

As of September 2024, 153 clinical trials involving the application of PDTO were registered on clinical trial.gov (Fig. 1). The first clinical trial was an observational study registered in 2014 by the Mayo Clinic (NCT06217874), planning the recruitment of 5000 breast cancer patients and generating PDTOs from primary tumors and metastasis. The clinical trial is still open, and patients are still being recruited. Clinical trials involving PDTOs have increased since 2014 (Fig. 1A). Most of these studies are in the recruitment phase of patients, and only 5 clinical trials have been completed (Fig. 1B), including NCT04859166 titled “Prospective primary human lung cancer organoids to predict treatment response”. This study uses PDTOs from lung cancer patients to predict how these tumors will respond to various treatments. Another trial is NCT04371198 entitled “Patient-derived organoids for rectal cancer”. This study was conducted by Duke University and focused on establishing rectal cancer organoids from patient biopsies. The primary aim was to create a model that closely mimics the patient's tumor to study its characteristics and test potential treatments. Out of the 20 patients enrolled, 17 (85%) had patient-derived organoids created from their pre-treatment specimens. Of these, 15 (88%) were successfully established, meaning they could be passed at least twice. All established organoids underwent standard of care and high-throughput drug screens, revealing variations in drug sensitivities among the samples. Additionally, within two weeks of receiving the samples, four organoids were established quickly enough to complete drug screening with oxaliplatin, SN38, and 5-Fluorouracil [196]. Another study, NCT04342286, entitled “To establish a reproducible organoid culture model with human kidney cancer,” was conducted by the Chinese University of Hong Kong and aimed to develop a reliable organoid culture model using human kidney cancer tissues. Similarly, NCT03307538, titled “Stereotactic body radiation therapy for unresectable perihilar cholangiocarcinoma (STRONG),” focuses on evaluating the safety and efficacy of stereotactic body radiation therapy (SBRT) in combination with chemotherapy for treating patients with unresectable perihilar cholangiocarcinoma. PDTOs were used to better understand tumor biology and test the efficacy of various treatments, including the combination of SBRT and chemotherapy. Lastly, NCT04072445, titled “A study to evaluate trifluridine/tipiracil in combination with irinotecan to treat biliary tract cancers”, was conducted by the Mayo Clinic and aimed to assess the effectiveness of the combination therapy in treating advanced, refractory biliary tract cancers (BTCs). This trial involved using PDTOs to test drug responses, providing a personalized approach to treatment. No results have been published from these completed clinical trials.

Fig. 1.

Clinical Trial Landscape of Patient-Derived Tumor Organoids (PDTOs). (A) Annual registration of clinical trials specifically utilizing PDTOs from 2014 to 2024, showcasing a substantial rise in PDTO-related research activities. (B) Status distribution of these PDTO-related clinical trials, illustrating the number categorized as recruiting, active (not recruiting), completed, unknown, and terminated. Notably, most studies are in the recruitment phase, with only a few completed. (C) Pie chart demonstrating the distribution of cancer types from which PDTOs are derived, with percentages highlighting the diversity and prevalence of specific cancers in these trials. (D) Classification of PDTO-related study types into observational and interventional, with interventional studies further broken down into clinical trial phases (Phase 1, Phase 1/2, Phase 2, Phase 3, and not specified), reflecting the phases of therapeutic exploration. (E) Number of patients enrolled in each PDTO-related clinical trial, showing the size distribution of patient enrollment across various studies, indicating the scale and reach of these trials.

Overall, these clinical trials aim to test the feasibility of developing PDTOs from tumor tissues, biopsies, or metastasis, to assess their genetic and drug sensitivity relevance to the original tissue, and to test new treatments. Only 20 trials are focused on establishing biobanks of PDTOs derived from various cancer types. These trials include NCT05464082, NCT06315868, NCT05134779, NCT05404321, and NCT05317221 for breast cancer; NCT04859166 for lung cancer; NCT05384184 for colorectal cancer and hepatocellular carcinomas; NCT02436564 for liver, biliary, and pancreatic cancers; NCT04927611 for neuroendocrine endoplasm; NCT05727020 and NCT06246630 for pancreatic cancer; NCT06077591, NCT05734963, NCT03896958, and NCT06350539 for various cancers; NCT06155370 for gynecological cancers; NCT06229522 for ovarian cancer; NCT04865315 for glioma; NCT04342286 for kidney cancer; and NCT03890614 for myeloma.

Various cancers were proposed to be modeled in the clinical trials; the most common were colorectal, breast, pancreatic, and lung (Fig. 1C). Most of these clinical trials are observational, and the remaining are interventional (Fig. 1D). Interventional clinical trials focus on predicting therapeutic responses in patients. They were classified across all phases, with 5 in clinical trial phase 1, 2 in phase 1/2, 20 in phase 2, and 7 in phase 3, while 24 were unspecified (Fig. 1D). Most clinical trials enrolled between 21 and 50 patients (30%) (Fig. 1E). Several clinical trials were dedicated solely to research purposes. For instance, the NCT06355700 trial focuses on developing hepatocellular carcinoma organoids and examining their integration with the host gut microbiota and peripheral blood mononuclear cells. The NCT04868396 trial aims to create primary patient-derived organoid cultures from gliomas (GM) to explore the mechanisms contributing to aggressive tumor growth and treatment resistance in both primary and recurrent gliomas. The NCT05571956 trial is designed to simultaneously establish a coculture of pancreatic ductal adenocarcinoma organoids and cancer-associated fibroblasts (CAFs) to study the pancreatic cancer microenvironment. The NCT05577689 trial seeks to develop a translational research platform to identify novel drug targets for metastatic prostate cancer using PDTOs. Lastly, the NCT05038358 trial plans to investigate the role of the immunological microenvironment in chemoresistant colorectal cancer PDTOs.

The increasing number of clinical trials involving PDTOs underscores their growing importance in cancer research. There is a high unmet need for a faithful preclinical model to help predict the cancer patient response and emergence of treatment resistance.

Conclusion

The pressing need to improve cancer therapies has brought great attention to PDTOs for modeling and predicting patient responses. Their eventual integration into a standard of care offers Brobdingnagian potential for advancing basic and translational research. However, the process is fraught with challenges. Creating and maintaining these PDTOs and storing them is complex. Standardized protocols, rigorous validation, and quality control are essential to establish best practices. The need for thorough validation to ensure organoids retain disease-specific genetic characteristics and the effect of long-term in vitro culturing on molecular and cellular integrity is paramount. Despite these hurdles, the future of PDTOs in translational biomedical research is promising. The integration of tumor microenvironments and immune cell population will lead to the potential for groundbreaking discoveries. Achieving this will require the scientific community's coordinated and collaborative effort to drive timely advancements in this exciting field.

CRediT authorship contribution statement

Sebastien Taurin: Writing – review & editing, Writing – original draft, Supervision, Formal analysis, Conceptualization. Reem Alzahrani: Writing – original draft, Investigation. Sahar Aloraibi: Writing – original draft, Resources, Formal analysis. Layal Ashi: Writing – original draft, Data curation, Conceptualization. Rawan Alharmi: Writing – original draft, Investigation, Data curation. Noora Hassani: Writing – original draft.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.