Ultradense Electrochemical Chip and Machine Learning for High-Throughput, Accurate Anticancer Drug Screening.

Area of Science:

• Electrochemistry
• Biosensing
• Machine Learning in Drug Discovery

Background:

• Electrochemical sensors offer potential for drug susceptibility testing but face challenges in throughput and accuracy.
• A lack of suitable platforms hinders drug candidate analysis during preclinical trials, slowing medicine development.

Purpose of the Study:

• To develop a high-throughput, user-friendly, and accurate method for determining 2D tumor cell viability for drug susceptibility assays.
• To integrate ultradense electrochemical chips with machine learning (ML) to overcome current limitations in drug screening.

Main Methods:

• Utilized ultradense electrochemical chips and machine learning (ML) for drug susceptibility assays.
• Employed cell detachment electrochemical assays with Ru(NH3)63+ and square wave voltammetry (SWV) to monitor cell death.
• Assessed the effect of doxorubicin on breast and colorectal cancer cells using drop-casting and microfluidic formats.

Main Results:

• Achieved high-throughput analysis by combining rapid SWV measurements (9 s) with serial chip analysis.
• Demonstrated accurate cell viability (98-104%) and half-maximal lethal concentration determination using ML-based data fitting.
• Validated the approach in both drop-casting and microfluidic assay formats.

Conclusions:

• The combined electrochemical chip and ML approach enables accurate and high-throughput drug susceptibility testing.
• This method shows significant potential for advancing preclinical drug screening and development.
• The findings pave the way for practical applications of electrochemical sensors in drug discovery.