Predicting Immunotherapy Outcomes in Glioblastoma Patients through Machine Learning
1Mount Sinai Hospital, New York, NY 10029, USA.
Cancers
|January 23, 2024
Summary
Predicting glioblastoma treatment response is crucial. This study identifies immune resistance markers and uses machine learning to forecast patient progression after immune checkpoint inhibitors (ICIs).
Area of Science:
- Oncology
- Immunology
- Computational Biology
Background:
- Glioblastoma is an aggressive brain cancer with poor prognosis.
- Current treatments including immune checkpoint inhibitors (ICIs) have limited efficacy.
- Understanding and predicting resistance to ICIs is critical for glioblastoma patients.
Purpose of the Study:
- To identify immune profiles associated with glioblastoma resistance to PD-1 blockade.
- To develop predictive models for glioblastoma patient response to ICIs.
- To enable personalized immunotherapy strategies for glioblastoma.
Main Methods:
- Analysis of two glioblastoma patient cohorts treated with PD-1 blockade.
- Characterization of immune profiles in patients with progressive disease post-ICIs.
- Application of machine learning algorithms to develop predictive models based on immune characteristics.
Main Results:
- 60% of glioblastoma patients showed disease progression despite ICIs.
- Resistance was linked to impaired macrophage, monocyte, T follicular helper, and antigen presentation responses.
- Aberrant regulatory T cell (Tregs) activity and elevated immunosuppressive molecules (TGFB, IL2RA, CD276) were observed.
- A machine learning model accurately predicted progression status in 82.82% of patients based on immune profiles.
Conclusions:
- Glioblastoma resistance to ICIs involves multifaceted immune defects.
- Machine learning can predict treatment response using patient-specific immune signatures.
- Personalized immunotherapy approaches hold promise for improving glioblastoma patient outcomes.
