Researchers have developed a groundbreaking 3D-printed artificial cancer tissue that can cultivate cancer cells from actual patients in an environment closely mimicking in vivo conditions.
On Tuesday, the Ulsan National Institute of Science and Technology (UNIST) announced that a research team led by biomedical engineering professors, Tae Eun Park and Hyun Wook Kang, in collaboration with a research team led by professor Seung Jae Myung from Seoul Asan Medical Center, has created an artificial cancer tissue called “Eba-PDO.” This innovative tissue replicates real cancer tissues’ high stiffness and low oxygen conditions.
By employing AI to analyze the shape of this artificial cancer tissue, researchers can predict the expression of key prognostic marker genes for colorectal cancer with an impressive 99% accuracy.
Due to their rapid proliferation, cancer cells form denser, firmer structures than normal tissue and thrive in oxygen-deprived environments.
Despite being created from patient-derived cells, previous artificial cancer tissues showed distorted growth patterns and drug responses because they failed to replicate these conditions accurately.
The research team developed this novel artificial cancer tissue by culturing patient-derived cancer organoids in three dimensions, combining them with a specially designed bio-ink, and printing them in a bead-like formation.
The bio-ink, a mixture of gelatin and extracellular matrix components, was engineered to recreate the hard, oxygen-deprived environment in which cancer typically flourishes.
Artificial cancer tissues grown using this method maintained consistent shapes for individual patients, while exhibiting variations in size and form among different patients.
Capitalizing on these characteristics, the research team developed an AI system capable of predicting CEACAM5 gene expression using only microscopic images. CEACAM5, a protein commonly found in solid tumors, including colorectal cancer, is associated with increased metastasis risk and chemotherapy resistance.
When overexpressed in artificial cancer tissues, this protein weakens intercellular bonds, resulting in less dense and structurally imbalanced tissues. The AI was trained to recognize these morphological changes to predict gene expression levels.
Moreover, the gene expression similarity between this artificial cancer tissue and actual patient-derived cancer tissues improved from 70% to 90%, a significant 20% increase. The model also accurately reproduced differences in patients’ responses to the chemotherapy drug 5-fluorouracil (5-FU).
UNIST researchers Hye Jin Jeong and Jong Hyeuk Han served as co-lead authors of the study.
The research team expressed optimism that this method, replicating cancer cell growth ex vivo for analysis, will pave the way for more precise personalized cancer treatments. They plan to refine their artificial cancer models further by incorporating immune cells and vascular structures in future iterations.
This research was supported by the Korean ARPA-H project from the South Korean Ministry of Health and Welfare, the Industrial Technology Alchemist project from the South Korean Ministry of Trade, Industry and Energy, the Bio-Medical Technology Development Program, the Advanced Bio-Technology and Human Resource Exchange project, and the South Korean Ministry of Education’s Glocal University Project (University of Ulsan) COMPaaS joint research.
The research findings were published online on March 28 in the prestigious international journal Advanced Science.