Morphology-Predicted Large-Scale Transition Number in Circulating Tumor Cells Identifies a Chromosomal Instability...

Area of Science:

• Oncology
• Genomics
• Computational Biology

Background:

• Chromosomal instability (CIN) drives tumor evolution and therapeutic resistance.
• Circulating tumor cells (CTCs) offer a dynamic window into tumor heterogeneity.
• Large-scale transitions (LSTs) are key genomic alterations indicative of CIN.

Purpose of the Study:

• To develop a rapid, image-based biomarker for CIN in CTCs.
• To correlate this biomarker with treatment outcomes in metastatic castration-resistant prostate cancer (mCRPC).

Main Methods:

• Direct sequencing of CTCs to quantify LSTs.
• Development of a computer vision algorithm to predict LST number from CTC morphology.
• Prospective analysis of 10,240 CTCs from 294 mCRPC patients.

Main Results:

• The computer vision algorithm accurately predicted high LST numbers.
• A high LST number, identified by the biomarker in pretreatment CTCs, strongly associated with poor overall survival.
• This association was observed in patients treated with standard therapies like androgen receptor signaling inhibitors and taxanes.

Conclusions:

• A computer vision-based biomarker of CIN in CTCs is a significant predictor of poor outcomes in mCRPC.
• This tool offers a rapid, non-invasive method to assess prognostic risk.
• The findings support the use of CTC-based CIN assessment for real-time treatment management.