Side-View Photoacoustic Endoscope for Early Cancer Staging

Abstract

Accurate staging of early gastrointestinal (GI) cancers is crucial for determining the appropriate treatment strategy, as the depth of tumor invasion into the mucosal wall dictates whether endoscopic resection or surgery is required. However, conventional imaging methods, such as endoscopic ultrasound (EUS), computed tomography (CT), and magnetic resonance imaging (MRI), encounter challenges in precisely assessing the invasion depth of early-stage GI cancers, particularly in differentiating between the mucosal and submucosal layers. This limitation underscores the need for an imaging modality that provides high spatial resolution and sufficient tissue penetration to visualize the thickness of the mucosa and assess the extent of tumor invasion into the submucosa, enabling accurate tumor staging and guiding treatment decisions.

Photoacoustic endoscopy (PAE) offers a promising minimally invasive approach to visualize early tumor invasion. This dissertation presents the design, development, and validation of a novel PAE system specifically tailored for early-stage colorectal cancer staging. The system incorporates a prism-based optical design, a piezoelectric (PZT) bender for lateral beam scanning, a step motor for circumferential imaging, and a custom ultrasound transducer for signal detection. This innovative design allows for cross-sectional imaging of the colon wall with a wide field of view and adequate imaging depth.

The use of targeted contrast agents, specifically near-infrared (NIR) fluorescent-labeled peptides, significantly enhanced the visualization of adenomas in both ex vivo and in vivo imaging studies. Pharmacokinetic studies were conducted to determine the optimal timing for imaging based on the peak accumulation of these agents within the target tissue. The side-view PAE system was successfully validated in a CPC;Apc mouse model of colorectal cancer. In vivo imaging demonstrated the ability of the system to detect and delineate adenomas with high contrast and accuracy, providing valuable information for early cancer staging.

PAE images were collected in vivo and ex vivo from n = 7 mice, including n = 14 adenomas with widths ranging from 0.746 – 6.793 mm and depths ranging from 0.114 – 1.62 mm. Analysis of these images demonstrated a strong correlation between PAE- measurements of adenoma dimensions and those obtained from histology (H&E), with Pearson correlation coefficients (ρ) ranging from 0.90 to 0.97. Furthermore, the relative root mean squared error (RRMSE) between the two methods ranged from 10.61% to 17.65%, indicating good agreement and low deviation. The results of this dissertation suggest that the side-view PAE system, with its compact design and targeted contrast enhancement capabilities, holds significant promise for clinical translation and has the potential to improve the accuracy of early cancer staging.