Personalized Cancer Treatment With Organoid Models – A Highlighted Study at ESMO 2024

The Era of Precision Oncology

Precision oncology is an approach that embraces individualized cancer treatment. It aims to identify the most appropriate treatment based on each patient's genetic, environmental and lifestyle characteristics. In modern medicine, treatment often follows standard protocols; precision oncology instead focuses on tailoring care to the patient's unique circumstances. This can both improve treatment effectiveness and minimize side effects.

Current Precision Oncology Practices

Current practices in precision oncology include genetic sequencing, biomarker testing, targeted therapies and minimal residual disease testing. These methods identify specific genetic mutations and other biological markers in patients' tumors, allowing this information to guide treatment selection. For example, a tumor carrying a particular mutation may respond more effectively to a specific medicine.

What Is a Patient-Derived Organoid Model?

Patient-derived organoids (PDOs) are three-dimensional cell cultures grown in the laboratory from a patient's own tumor tissue. By retaining the tumor's genetic and phenotypic characteristics, they provide a model resembling the original tumor tissue. These models allow the effects of medicines on tumors to be observed directly, giving scientists and clinicians the opportunity to test personalized treatment options.

The Contribution of Organoid Models

Organoid models make an important contribution to precision oncology by allowing tumor-specific drug responses to be tested in the laboratory. Potential treatments can be assessed in advance for each patient, allowing the most effective option with the fewest side effects to be identified before clinical use. Organoid testing also plays an important role in developing new medicines and contributes to the discovery of targeted therapies. This integrated approach to precision oncology plays a critical role in shaping the future of cancer treatment.

Details of a New Study From France

In this study presented at the 2024 ESMO Congress, researchers predicted clinical responses using patient-derived organoids  obtained from patients with pancreatic and colorectal cancer. The research included 155 organoid models: 70 pancreatic cancer and 85 colorectal cancer samples. Each organoid was created in the laboratory using cells from the patient's tumor, then tested with the medicines the patient received clinically. These tests aimed to simulate individual responses and improve predictions of future treatment responses.

The research combined biochemical tests on organoid models with multivariable analyses to develop prospective predictions of how patients might respond to treatment. The model served as a tool for predicting which treatments would produce the best outcomes in clinical practice.

Research Results

The study showed that organoid models could accurately predict patients' treatment responses and the time before cancer progression:

Prediction of Overall Treatment Response:

• Sensitivity: 83%
• Specificity: 88%

Prediction of Progression-Free Survival:

The study used progression-free survival (PFS) predictions based on a Cox regression model and hazard ratios (HRs). 

• Cox model C-score: 0.59
  • The C-score is a statistic measuring a model's predictive accuracy. A value of 0.5 represents random prediction, while 1.0 represents perfect prediction. In this study, 0.59 indicates that the model's ability to determine treatment response was only slightly better, rather than perfect.
• Hazard ratio: 2.82 (95% confidence interval: 1.25–6.33)
  • The hazard ratio shows how much the risk of an event, here cancer progression, increases within a group. HR = 2.82 means that the event (cancer progression) is 2.82 times more likely in the presence of a particular factor, such as treatment. This ratio shows how much the treatment affects cancer progression.

Treatments producing good outcomes were evaluated using the growth modulation index (GMI):

GMI for treatments producing good outcomes: 1.81

GMI for treatments not producing good outcomes: 0.78

The Impact of Clinical Data

Adding clinical and molecular data improved the predictive accuracy of the organoid models:

• Cox model C-score: 0.69

The number of medicines received before treatment and the disease type affected responses:

• HR for the number of medicines: 1.37
• HR for disease type: 1.89

General health status (ECOG performance score) did not significantly affect treatment response.

• After controlling for clinical and molecular variables, predictive power increased markedly (HR: 4.36, p = 0.0027).
  • This indicates a 4.36-fold increase in the risk of an event, such as cancer progression or worsening disease. It means that this risk increases substantially when clinical and molecular variables are controlled for. In other words, once these factors are taken into account, the event is much more likely to occur.

Conclusion

This study shows that adding clinical and molecular variables to functional screening of organoid models substantially improves their predictive ability. Accurately predicting patient responses is a critical step toward more successful personalized cancer treatment. These findings indicate that organoid models should be used more effectively in precision oncology and could open new possibilities in future cancer treatment.