For decades, cancer research relied almost exclusively on two-dimensional (2D) monolayers of cells grown on flat plastic surfaces. While inexpensive and reproducible, these 2D models fail to capture the structural complexity, cell-cell interactions, and microenvironmental cues that define real tumors. The shift toward three-dimensional (3D) cell culture has been one of the most consequential advances in oncology, enabling scientists to observe how cancer cells behave in a spatially relevant context, how they respond to drugs, and how they evolve under therapeutic pressure. 3D cultures more faithfully recapitulate the architecture of solid tumors, including gradients of nutrients, oxygen, and waste products, as well as the physical constraints that influence cell signaling and migration. As a result, 3D models are now driving breakthroughs in personalized medicine, drug screening, and mechanistic studies of metastasis.

Foundations of 3D Cell Culture in Cancer Research

Traditional 2D cultures force cells to adhere to a flat substrate, altering their morphology, polarity, and gene expression. In contrast, 3D culture systems allow cells to grow in three dimensions, forming aggregates or organized structures that mirror in vivo tissues. The most common 3D approaches include scaffold-based systems, in which cells are seeded onto or into a porous matrix, and scaffold-free systems, in which cells self-assemble into spheroids or organoids. Each method has its own strengths and limitations, and researchers increasingly combine multiple techniques to achieve greater physiological relevance.

Key Techniques in 3D Cell Culture

Scaffold-Based Systems

Scaffold-based 3D culture uses natural or synthetic biomaterials to support cell attachment, proliferation, and differentiation. Natural scaffolds such as Matrigel, collagen, and alginate provide biochemical cues that mimic the extracellular matrix (ECM) of tumors. Synthetic scaffolds made from polymers like polyethylene glycol (PEG) or poly(lactic-co-glycolic acid) (PLGA) offer tunable mechanical properties and degradation rates. These scaffolds enable the formation of tumor-like structures with defined porosity and stiffness—parameters known to influence cancer cell invasion and drug resistance. Researchers have used scaffold-based models to study breast cancer, glioblastoma, and pancreatic ductal adenocarcinoma, finding that gene expression profiles more closely match patient tumors than 2D cultures do. A comprehensive review in Nature Reviews Chemistry highlights how synthetic polymers have evolved to allow precise control over the tumor microenvironment.

Scaffold-Free Systems: Spheroids and Organoids

Scaffold-free methods rely on cells' intrinsic ability to aggregate when cultured under non-adherent conditions. Tumor spheroids are compact, multicellular clusters that develop gradients of oxygen, nutrients, and waste products—similar to avascular tumor nodules. They are relatively simple to produce in high throughput and are widely used for drug screening and radiation studies. Organoids take this concept further by incorporating stem cell differentiation to create miniature, self-organizing tissues that retain key features of the original organ. Patient-derived organoids (PDOs) have become powerful tools for personalized oncology, as they can be generated from a patient's biopsy and used to test drug sensitivity ex vivo. In colorectal, pancreatic, and lung cancers, PDO responses have been shown to correlate with clinical outcomes. A landmark study in Cell demonstrated that organoids from metastatic colorectal cancer patients could predict responses to chemotherapy regimens with high accuracy.

Bioprinting

3D bioprinting builds on additive manufacturing principles to deposit living cells and biomaterials layer by layer, creating tissue constructs with defined geometry. This technique allows researchers to position multiple cell types, including cancer cells, stromal fibroblasts, and immune cells, in precise spatial arrangements. Bioprinted tumor models have been used to study invasion patterns, test anti-angiogenic drugs, and even replicate patient-specific tumor architectures from imaging data. Recent advances include the use of photo-crosslinkable hydrogels that can be rapidly polymerized under visible light, improving cell viability. Bioprinting also facilitates the creation of vascularized tumor models, which are critical for studying drug delivery and metastasis. As reviewed in a paper in Biomaterials, bioprinted breast cancer models have been used to screen combinatorial therapies and identify synergistic drug pairs.

The field is accelerating as researchers integrate 3D culture with complementary technologies. These emerging trends are not only improving the fidelity of tumor models but also enabling new types of experiments that were previously impossible in dish-based systems.

Integration with Microfluidics

Microfluidic devices, often called "tumor-on-a-chip" platforms, connect 3D culture chambers with precisely controlled fluid flow. This setup mimics the dynamic microenvironment of tumors, including interstitial fluid pressure, shear stress from blood flow, and gradients of oxygen and nutrients. Microfluidics allows real-time imaging of cell behavior and collection of secreted factors for analysis. Researchers have used these systems to study how fluid flow affects cancer cell migration, how drug gradients influence resistance, and how immune cells infiltrate tumors. For example, a microfluidic model of the blood-brain barrier combined with a glioblastoma spheroid revealed that drug penetration is altered by the presence of pericytes. A 2020 study in Lab on a Chip demonstrated a multi-compartment microfluidic device that recapitulated liver metastasis, allowing the authors to track circulating tumor cells as they extravasated into a 3D liver model.

Patient-Derived Models

Personalized oncology has driven the development of patient-derived models that retain the genetic and phenotypic diversity of individual tumors. Beyond organoids, patient-derived xenografts (PDXs) are created by implanting human tumor tissue into immunocompromised mice. PDXs maintain the stromal architecture and heterogeneity of the original tumor, but they are expensive, time-consuming, and low-throughput. 3D culture of patient-derived cells offers a faster, more scalable alternative. Biopsy cubes, tumor slices, and dissociated cells can be cultured in 3D matrices to screen drugs in a timeline that matches clinical decision-making. A growing number of clinical trials now incorporate 3D patient-derived models alongside genomic profiling to guide therapy selection. A trial reported in the Journal of Clinical Oncology used organoid cultures from ovarian cancer patients to predict platinum sensitivity with over 80% accuracy, suggesting a path toward routine use in the clinic.

Artificial Intelligence and Machine Learning

3D cell cultures produce vast amounts of high-dimensional data, including confocal microscopy images, time-lapse movies, and gene expression profiles. Machine learning algorithms have become essential for extracting meaningful patterns from these datasets. Deep neural networks can segment spheroids, measure invasion distance, classify organoid morphologies, and predict drug responses based on image features alone. AI also helps identify subtle changes in cellular organization that precede drug resistance. For instance, a convolutional neural network trained on brightfield images of pancreatic tumor organoids could predict which organoids would respond to a specific combination therapy. In addition, generative models are being used to simulate how tumors might evolve under different treatment regimens. A perspective in Nature Biomedical Engineering argues that integrating AI with organoid platforms could enable virtual clinical trials, drastically reducing the time and cost of drug development.

High-Throughput 3D Screening

Historically, 3D cultures were limited by low throughput compared to 2D monolayer assays. Advances in 384- and 1536-well plate formats, automated dispensing, and confocal imaging have now made truly high-throughput 3D screening possible. Acoustic droplet ejection and microfluidic mixing can create uniform spheroids in each well, and automated microscopes capture images at multiple z-planes for quantitative analysis. Pharmaceutical companies are adopting these platforms to screen compound libraries against tumor spheroids and organoids, identifying hits that are more likely to translate in vivo. The U.S. National Cancer Institute has supported the development of 3D screening pipelines that incorporate patient-derived models, and several biotech companies now offer commercial organoid screening services. This shift is crucial because drugs that are effective in 2D often fail in 3D due to differences in penetration, metabolism, and signaling.

Vascularization and Perfusion Systems

A major limitation of many 3D cultures is the lack of a functional vasculature. Tumors larger than a few hundred micrometers develop hypoxic cores and necrotic centers, which limit their growth and alter drug distribution. To overcome this, researchers have developed methods to co-culture endothelial cells within 3D scaffolds, encouraging the formation of primitive vascular networks. Some groups use bioprinting to print channels lined with endothelial cells, which can then be perfused with culture medium. More advanced systems incorporate microfluidic pumps to generate pulsatile flow, mimicking the hemodynamics of real blood vessels. These vascularized tumor models have been used to study how cancer cells intravasate into vessels, how immune cells extravasate into tumors, and how anti-angiogenic drugs affect tumor growth. A study in PNAS showed that a vascularized glioblastoma-on-a-chip model could predict patient responses to temozolomide more accurately than non-vascularized spheroids.

Co-Culture and Immune Microenvironment Modeling

Solid tumors are not just collections of cancer cells; they include stromal fibroblasts, immune cells, endothelial cells, and pericytes that all contribute to tumor progression and therapy response. 3D co-culture systems allow researchers to incorporate these components in a controlled manner. For example, tri-cultures of cancer cells, cancer-associated fibroblasts (CAFs), and macrophages have revealed that CAFs can promote invasion while macrophages facilitate immune suppression. Immune checkpoint inhibitors, such as anti-PD-1 and anti-CTLA-4 antibodies, can be tested in 3D co-culture models that include patient-derived tumor-infiltrating lymphocytes (TILs). These models are gaining traction as alternatives to animal models for immuno-oncology research. A particularly exciting development is the use of 3D "tumor microenvironment" chips that incorporate multiple cell types, extracellular matrix, and fluid flow, enabling the study of T cell trafficking, activation, and killing in a physiologically relevant context. A review in Trends in Cancer discusses how such systems are being used to optimize combination immunotherapies.

Challenges and Practical Considerations

Despite their promise, 3D cell culture methods face several hurdles. Reproducibility remains a concern, especially for scaffold-based systems where batch-to-batch variability of natural matrices can introduce experimental noise. Standardization of protocols, quality control metrics, and data analysis pipelines is still evolving. Many 3D cultures also require specialized equipment, such as bioprinters, microfluidic pumps, and confocal microscopes, which can be cost-prohibitive for smaller labs. In addition, imaging and analyzing 3D structures is more complex than 2D monolayers, often requiring advanced image processing and computational resources. Nevertheless, the field is making rapid progress through collaborative efforts like the Human Organoid Initiative and the development of open-source software for spheroid analysis.

Future Directions

Looking ahead, several promising avenues are likely to define the next decade of 3D cancer research. One is the integration of multi-omics profiling (transcriptomics, proteomics, metabolomics) with 3D culture readouts, enabling a systems-level understanding of tumor behavior. Another is the development of "4D" culture systems that incorporate time as a variable, such as through real-time monitoring of drug responses with biosensors. Organoids derived from liquid biopsies (circulating tumor cells) could allow non-invasive, repeated monitoring of tumor evolution during treatment. The convergence of 3D culture with CRISPR-based functional genomics will enable systematic mapping of gene dependencies in a 3D context, revealing new therapeutic targets. Finally, regulatory agencies are beginning to accept data from validated 3D models for drug approval decisions, which could reduce the reliance on animal testing and accelerate the drug development pipeline.

As these trends mature, 3D cell culture is poised to become a routine tool in both basic cancer biology and clinical decision-making. By bridging the gap between simple 2D assays and complex animal models, these systems offer a scalable, human-relevant platform for understanding cancer and improving patient outcomes. The continued collaboration between engineers, biologists, and clinicians will be essential to overcome current limitations and fully realize the potential of 3D cell culture in the fight against cancer.