PhD Thesis Proposal: Yao Chen

Optional ZOOM LINK
MeetingID: 980 9104 0548
Passcode: 818945

"Advancing Computational Analysis of Fluorescence Imaging for Head and Neck Cancer Surgical Guidance"

Abstract

The primary objective of surgery for head and neck squamous cell carcinoma (HNSCC) is to achieve complete tumor resection while preserving much normal tissue to increase patient survival and decrease morbidity. Unfortunately, it is challenging to achieve clear margins due to the inherent complexity of physiological structures and tumor infiltration, especially when relying solely on surgeons’ visual and palpation inspection. In recent years, NIR fluorescence molecular imaging has shown unique potential for tumor identification through specific binding to targets (e.g., EGFR in HNSCC) and a less-attenuating optical spectrum window, leading to the development of fluorescence-guided surgery (FGS).

Although promising, underperforming tumor identification remains a challenge with fluorescence imaging, limiting the widespread adoption of FGS. Modest tumor-to-normal imaging contrast was observed in multiple FGS clinical trials. Confounding effects are both physiological (i.e., imaging agents’ non-specific uptake and incomplete tumor penetration) and computational (i.e., existing contrast analysis based on fluorescence intensity). This unmet need for precise tumor identification calls for comprehensive work in advancing computational analysis of fluorescence images, including investigating the best-performing combination of imaging parameters to optimize contrast and developing or standardizing innovative fluorescence imagery analysis methods to improve the accuracy of tumor-to-normal differentiation.

Here, we present the development of computational models (based on pre-clinical mouse models of HNSCC) to investigate/predict the microscale tissue penetration and imaging agents binding in different physiological environments and on the resulting macroscale optical imaging contrast. This work aims to provide guidance on design of optimal imaging agents, dosing strategy, and administration-to-imaging time to achieve optimal imaging contrast for surgeries. Furthermore, clinical HNSCC fluorescence images from the ABY-029 Phase-0 clinical trials are used for the development of ML computer-assisted algorithms to improve the accuracy of HNSCC detection and margin assessment by providing the interpretation of fluorescence signal. Finally, clinical images obtained from a multi-institution collaboration that pools data from several FGS clinical trials using similar molecular imaging agents targeted to EGFR in HNSCC are used to perform imaging contrast analysis for standardization and automation of tumor detection. These advancement in computational analysis of fluorescence imaging would provide a wealth for translating FGS in clinics.

Thesis Committee

• Kimberley S. Samkoe (Chair)
• Scott. C. Davis
• Geoffrey P. Luke
• Greg M. Thurber (U Michigan, Ann Arbor)

Contact

For more information, contact Thayer Registrar at thayer.registrar@dartmouth.edu.