Related Experiment Video
Updated: Sep 14, 2025
Ovarian Cancer Patient-Derived Organoid Models for Pre-Clinical Drug Testing
Published on: September 15, 2023
Development and validation of deep learning for predicting the growth of ovarian cancer organoids
Hongji Wu1, Lifang Ma2,3,4, Ling Wang2,3,4
1Bioengineering College, Chongqing University, Chongqing 400044, China.
Background:
Organoids have attracted enormous interest in disease modeling, drug screening, and precision medicine. However, developing robust patient-derived organoids (PDOs) was time-consuming, costly, and had low success rates for certain cancer types, which limited their clinical utility. This study aimed to develop an interpretable deep learning-based model to predict the cultivation outcome of ovarian cancer organoids in advance.
Methods:
Longitudinal microscopy images of 517 ovarian cancer organoid droplets were divided into training ( n = 325), validation ( n = 88), and test ( n = 104) sets. Subsequently, growth prediction models were developed based on four neural network backbones (ResNet18, VGG11, ConvNeXt v2, and Swin Transformer v2), and specific optimization methods were designed for better prediction. Finally, 179 samples from multiple centers were collected for prospective validation, and the gradient-weighted class activation mapping (Grad-CAM) method was used for interpretability analysis of the deep model to reveal the basis of the model's decisions.
Results:
The test set showed that the deep learning models could achieve high-performance prediction at the third stage with area under the curve (AUC) values greater than 0.8 for all four models. The homogeneous transfer learning optimization method improved the AUC from 0.833 to 0.884 ( P = 0.0039). In prospective validation, the optimized model achieved an AUC of 0.832, a Brier score of 0.1919 in the calibration curve, and a greater net benefit in the decision curve. Interpretability analysis revealed that the area where organoids are being formed and have already formed is important for prediction.
Conclusions:
Our developed models achieved satisfactory results in predicting the growth of ovarian cancer organoids. There is potential for further development of the model toward process automation.
Insights
This study developed an interpretable deep learning model to accurately predict ovarian cancer organoid growth, improving clinical utility for disease modeling and drug screening.
Area of Science:
- Biomedical Engineering
- Computational Biology
- Oncology
Background:
- Organoids are crucial for disease modeling and drug screening but developing patient-derived organoids (PDOs) is challenging.
- Low success rates and high costs limit the clinical utility of PDOs, especially for ovarian cancer.
- Predicting organoid cultivation outcomes is essential for improving efficiency and success rates.
Purpose of the Study:
- To develop an interpretable deep learning model for predicting ovarian cancer organoid cultivation outcomes.
- To enhance the accuracy and reliability of organoid growth prediction.
- To address the limitations of current PDO development methods.
Main Methods:
- Trained deep learning models (ResNet18, VGG11, ConvNeXt v2, Swin Transformer v2) on longitudinal microscopy images of 517 ovarian cancer organoid droplets.
- Optimized models using homogeneous transfer learning and validated prospectively on 179 multi-center samples.
- Employed Grad-CAM for model interpretability to identify key predictive features.
Main Results:
- Deep learning models achieved high prediction performance (AUC > 0.8) on the test set.
- Optimization improved AUC from 0.833 to 0.884 (P = 0.0039).
- Prospective validation showed an AUC of 0.832, with good calibration and decision curve analysis, highlighting the importance of organoid formation areas.
Conclusions:
- The developed deep learning models effectively predict ovarian cancer organoid growth.
- Interpretability analysis provides insights into model decision-making.
- The models show potential for process automation in organoid development.
