AI Enhances Lung Cancer Biopsy Accuracy

In July 2026, an article published by CanHealth discussed the critical nature of timely lung cancer staging results for patients, highlighting that delays can significantly impact treatment decisions, which may include surgery, chemotherapy, or palliative care. The standard procedure for staging lung cancer is the endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), a minimally invasive method used to sample lymph nodes in the chest. While EBUS-TBNA is well-established and effective, variability in outcomes can occur depending on clinician experience and procedural volume. Inconclusive results often necessitate repeat procedures, prolonging patient uncertainty.

To address the variability in EBUS-TBNA diagnostic performance, a new clinical trial led by McMaster University is underway, with support from the University Health Network’s Toronto General Hospital. The trial aims to evaluate an AI-assisted intervention developed by Node AI, a Hamilton-based medical AI company co-founded by Dr. Wael Hanna, Dr. Anthony Gatti, and CEO Mackensey Bacon. Their platform is designed to integrate seamlessly into existing EBUS workflows without requiring additional hardware, aiming to support all major bronchoscope manufacturers.

During EBUS-TBNA procedures, the AI platform analyzes ultrasound video in real-time to assist clinicians in identifying lymph node anatomy and making targeting decisions. This technology is based on robust clinical research data and an extensive EBUS video dataset compiled over seven years. Dr. Hanna emphasized the need to investigate whether AI can help enhance consistency in EBUS-TBNA performance and lead to improved patient outcomes.

The clinical trial, which will enroll 100 participants over three months, has received Research Ethics Board approval and aims to assess the AI platform's capability to process EBUS imaging and return real-time predictions for more than 90% of captured images. Meeting this threshold would demonstrate the technical feasibility of real-time AI integration in the procedure. Additionally, secondary endpoints will evaluate improvements in diagnostic yield and reductions in variability among different operators.

The selection of UHN’s Toronto General Hospital as a study site stems from its extensive expertise in EBUS-TBNA and its history of advancing the procedure through research and training. With a focus on integrating AI assistance into clinical workflows, the hospital provides a fitting environment to rigorously evaluate the technology's impact.

Accurate and timely lung cancer staging is crucial for effective treatment, and research has indicated that consistent, high-quality results in EBUS-TBNA may not be universally accessible across all healthcare settings. The trial by Node AI represents an early exploration of AI’s potential to address this disparity.

If the trial demonstrates that the AI performs as intended in a controlled clinical setting, further studies could explore its validation across diverse institutional backgrounds. The implications of ensuring equitable access to consistent staging outcomes could be significant for lung cancer patients, and the trial's results will guide future directions in this area of research. Dr. Kazuhiro Yasufuku, who has extensively contributed to EBUS-TBNA development, now leads the study at Toronto General Hospital.

In conclusion, the ongoing trial signifies a crucial step in understanding the role of AI in enhancing lung cancer staging procedures and potentially improving patient outcomes in a critical area of healthcare.