Driving innovation for rare skin cancers: utilizing common tumours and machine learning to predict immune checkpoint...

Area of Science:

• Oncology
• Immunotherapy
• Biomarker Discovery

Background:

• Metastatic Merkel cell carcinoma (MCC) and cutaneous squamous cell carcinoma (cSCC) respond well to immune checkpoint inhibitors (ICIs).
• However, 40% of patients do not respond to ICI therapy, necessitating predictive biomarkers.
• Developing biomarkers and combination treatments for these rare cancers is challenging.

Purpose of the Study:

• To review the pathophysiology and treatment of MCC and cSCC.
• To focus on potential biomarkers for immunotherapy response.
• To explore the use of transfer learning and deep phenotyping for biomarker development and novel treatment targets.

Main Methods:

• Literature review of MCC and cSCC pathophysiology and treatment.
• Discussion of potential predictive biomarkers for immunotherapy response.
• Exploration of transfer learning with big data from common tumors and deep phenotyping of rare tumors.

Main Results:

• Identified challenges in predicting ICI response in rare cancers.
• Highlighted the need for novel biomarkers and combination therapies.
• Proposed transfer learning and deep phenotyping as methods for biomarker discovery.

Conclusions:

• Effective prediction of ICI response in MCC and cSCC requires advanced strategies.
• Transfer learning and deep phenotyping offer promising avenues for developing predictive biomarkers.
• These approaches can elucidate novel therapeutic targets for rare cancers.