Identification of Cancer cell types by Electrical Impedance Spectroscopy Based on Principal Component Analysis Integrated with Equivalent Circuit Model (ECM-PCA).

Objective: This study aims to enhance the identification of cancer cell types using electrical impedance spectroscopy (EIS) by introducing a novel analysis method, ECM-PCA, which integrates an equivalent circuit model with principal component analysis.

Methods: The ECM-PCA method addresses the limitations of conventional PCA and kernel PCA (kPCA) in handling non-linear and frequency-dependent data. Impedance data of four cancer cell types (DLD-1, T.Tn, U138, and U87) were acquired across a frequency range of 0.1 MHz to 300 MHz. The ECM-PCA method was applied to analyze the frequency-dependent impedance behaviour and compare its clustering performance with PCA and kPCA.

Results: ECM-PCA demonstrated clustering performance comparable to kPCA while capturing the frequency-dependent features of impedance spectra, which kPCA lacks. The phase angle component as the ECM-PCA input achieved the highest Calinski-Harabasz (CH) score of 935, and the method achieved an identification accuracy of 93.6% in the PC1 and PC2 plane.

Conclusion: ECM-PCA improves the accuracy and interpretability of cancer cell type identification based on electrical impedance data.

Significance: This study highlights the potential of ECM-PCA in advancing cancer diagnostics through enhanced analysis of impedance spectra.