Electrical impedance-based tissue classification for bladder tumor differentiation.

Area of Science:

• Biomedical Engineering
• Surgical Technology
• Oncology

Background:

• Tumor recurrence in bladder cancer is a significant issue, often due to incomplete tumor removal.
• Intraoperative tissue characterization is crucial for surgeons to ensure complete tumor resection.
• Electrical impedance spectroscopy (EIS) shows promise for tissue differentiation but faces challenges in real-world surgical settings.

Purpose of the Study:

• To develop a robust, patient-based classification method for intraoperative bladder tissue characterization using EIS.
• To improve the accuracy of distinguishing cancerous tissue from healthy tissue during surgery.
• To mitigate challenges posed by inter-individual variations, radiotherapy, and mechanical disturbances.

Main Methods:

• A patient-based classification approach was developed, evaluating impedance measurements relative to a healthy reference from the same patient.
• Two feature extraction methods were explored: direct measurement points and model-based parameters.
• A Gaussian process classifier was trained based on the distance of feature vectors to the healthy reference.

Main Results:

• The proposed method achieved high classification accuracy, up to 100% on noise-free data under controlled conditions.
• Even with external disturbances, the approach maintained a classification accuracy of 80%.
• The patient-based relative classification demonstrated robustness against inter-individual differences and systematic errors.

Conclusions:

• The developed EIS-based approach offers a promising solution for intraoperative bladder tissue characterization.
• This technique has the potential to significantly reduce bladder cancer recurrence rates by aiding complete tumor removal.
• Further research and validation are warranted to implement this method in clinical practice.