Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Prediction of Cellular Malignancy Using Electrical Impedance Signatures and Supervised Machine Learning

Shadeeb Hossain

TL;DR

The paper tackles non-invasive cancer classification from cellular bioelectrical signatures by aggregating dielectric properties from 33 studies and evaluating three supervised classifiers. They implement and tune Random Forest, SVM, and KNN on standardized datasets spanning 150 kHz to 20 GHz, using accuracy and F1 as metrics. Random Forest achieved the highest performance, about 90% accuracy with a maximum depth of 4 and 100 trees, while KNN and SVM reached F1 scores near 78% and 76.5%, respectively. The work demonstrates the viability of combining bioelectrical impedance analysis with supervised learning for in vitro and potential in-situ diagnostics, pointing to future hardware-enabled, real-time cell classification.

Abstract

Bioelectrical properties of cells such as relative permittivity, conductivity, and characteristic time constants vary significantly between healthy and malignant cells across different frequencies. These distinctions provide a promising foundation for diagnostic and classification applications. This study systematically reviewed 33 scholarly articles to compile datasets of quantitative bioelectric parameters and evaluated their utility in predictive modeling. Three supervised machine learning algorithms- Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN) were implemented and tuned using key hyperparameters to assess classification performance. Model effectiveness was evaluated using accuracy and F1 score as performance metrics. Results demonstrate that Random Forest achieved the highest predictive accuracy of ~ 90% when configured with a maximum depth of 4 and 100 estimators. These findings highlight the potential of integrating bioelectrical property analysis with machine learning for improved diagnostic decision-making. Similarly, for KNN and SVM, the F1 score peaked at approximately 78% and 76.5%, respectively. Future work will explore incorporating additional discriminative features, leveraging stimulated datasets, and optimizing hyperparameter through advanced search strategies. Ultimately, hardware prototype with embedded micro-electrodes and real-time control systems could pave the path for practical diagnostic tools capable of in-situ cell classification.

Prediction of Cellular Malignancy Using Electrical Impedance Signatures and Supervised Machine Learning

TL;DR

The paper tackles non-invasive cancer classification from cellular bioelectrical signatures by aggregating dielectric properties from 33 studies and evaluating three supervised classifiers. They implement and tune Random Forest, SVM, and KNN on standardized datasets spanning 150 kHz to 20 GHz, using accuracy and F1 as metrics. Random Forest achieved the highest performance, about 90% accuracy with a maximum depth of 4 and 100 trees, while KNN and SVM reached F1 scores near 78% and 76.5%, respectively. The work demonstrates the viability of combining bioelectrical impedance analysis with supervised learning for in vitro and potential in-situ diagnostics, pointing to future hardware-enabled, real-time cell classification.

Abstract

Bioelectrical properties of cells such as relative permittivity, conductivity, and characteristic time constants vary significantly between healthy and malignant cells across different frequencies. These distinctions provide a promising foundation for diagnostic and classification applications. This study systematically reviewed 33 scholarly articles to compile datasets of quantitative bioelectric parameters and evaluated their utility in predictive modeling. Three supervised machine learning algorithms- Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN) were implemented and tuned using key hyperparameters to assess classification performance. Model effectiveness was evaluated using accuracy and F1 score as performance metrics. Results demonstrate that Random Forest achieved the highest predictive accuracy of ~ 90% when configured with a maximum depth of 4 and 100 estimators. These findings highlight the potential of integrating bioelectrical property analysis with machine learning for improved diagnostic decision-making. Similarly, for KNN and SVM, the F1 score peaked at approximately 78% and 76.5%, respectively. Future work will explore incorporating additional discriminative features, leveraging stimulated datasets, and optimizing hyperparameter through advanced search strategies. Ultimately, hardware prototype with embedded micro-electrodes and real-time control systems could pave the path for practical diagnostic tools capable of in-situ cell classification.
Paper Structure (10 sections, 4 figures, 1 table)

This paper contains 10 sections, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background Information
  3. Bioelectrical Properties and Cole-Cole plot
  4. Machine Learning Models in Bioelectrical Cancer Classification
  5. Understanding Random Forest Algorithm
  6. Understanding Support Vector Machine (SVM)
  7. Understanding K-Nearest Neighbor
  8. Data Collection and Methodology
  9. Results and Discussion
  10. Conclusion and Future Work

Figures (4)

  • Figure 1: Schematic representation of the proposed models, in which bioelectrical measurements are obtained from test samples and then processed using pre-trained machine learning (ML) models.
  • Figure 2: (a) The Frickle-Morse equivalent circuit model used for impedance extraction. (b-c) Comparison of electric field lines between healthy and malignant cell morphologies.
  • Figure 3: Cole-Cole plot illustrating the separation in dielectric properties between cancerous and non-tumorous breast cell lines.
  • Figure 4: K-Nearest Neighbor (KNN) train vs. test accuracy model graph for 100 iterations showing the variation in evaluation metrics with number of k varied between 1 and 20.