Evaluating the Potential of Optical Coherence Elastography for Early Gastric Cancer Detection

Abstract

Gastric cancer remains a major clinical challenge, being the third most common cause of cancer death worldwide with an overall five-year survival rate of 35.7%. However, early detection substantially increases the survival rate to over 95%. Endoscopic screening or surveillance can be appropriate and cost effective options to save lives, but only if an effective protocol can be developed. Unfortunately, white-light endoscopy, which is the current method for gastric cancer detection, can have as low as 42% sensitivity to detecting cancerous tissues. Consequently, our team has been evaluating the promise of optical coherence elastography (OCE) for early gastric cancer detection. OCE is based on the optical coherence tomography (OCT), which enables visualization of the mechanical and microstructural properties of tissues. OCT is based on low-coherence interferometry to produce high-resolution sample structure images as a function of depth. In our work, we first conducted a patient study using ex vivo patient specimens and an “optical palpation” approach with tissue phantoms to estimate tissue stiffness. Our results suggested that tissue stiffness was positively correlated with disease. However, we observed significant issues with noise and outliers due to the phantom stiffness being too high. In addition, there were issues with the OCT image artifacts, preventing the image analysis algorithm from accurately detecting the phantom stiffness. To tackle these problems, we conducted phantom characterization to determine that using freshly made 38 kPa phantoms increase OCE accuracy and reliability. When tested these to compare normal tissues and gastric cancer ex vivo using a mouse model. The freshest 38 kPa phantoms enabled more distinct and consistent measurement of tissue stiffness and the results are consistent with the patient study. Ultimately, we show the potential of OCE for early gastric cancer detection through this project.