Significance  Gastrointestinal cancer represents a major global health challenge and is among the leading causes of death worldwide. By the time symptoms manifest, the disease is often at an advanced stage, making early detection and treatment critical for improving patient survival rates. Endoscopic surveillance is vital in identifying early lesions; However, traditional white-light endoscopy has limitations in accurately detecting early-stage lesions. This is primarily due to its reliance on replicating human vision, which can make it difficult for clinicians to precisely identify target tissue areas requiring resection. Although advanced endoscopic technologies, such as magnifying endoscopy combined with narrow-band imaging (ME-NBI), have improved the visualization of early gastrointestinal cancers, these techniques are still complex, time-consuming, and heavily dependent on the expertise of endoscopists and pathologists. In this context, Light Scattering Spectroscopy (LSS) emerges as a non-invasive and highly sensitive optical technique that has been widely applied for early cancer detection. LSS is particularly effective in detecting subcellular changes in epithelial tissues, which are often the first indicators of malignancy. By capturing single-scattered light and analyzing it through Mie scattering theory, LSS can infer crucial parameters such as cell nucleus size, shape, and refractive index—key factors in accurate and early disease diagnosis. The significance of LSS lies in its ability to provide real-time, non-destructive diagnostic information, thereby guiding clinical decisions and improving patient outcomes. However, the clinical application of LSS is currently constrained by interference caused by diffuse scattering from underlying tissues. Overcoming these challenges is essential to unlock the full potential of LSS in cancer diagnostics.

Progress  Over the past two decades, substantial advancements in LSS have focused on improving the accuracy and reliability of this technology. Four experimental techniques have been introduced for extracting single-scattered light, each contributing to the enhancement of LSS in various applications (Fig.4). The first significant development is the coherent interference technique, which combines the depth-resolving capabilities of Optical Coherence Tomography (OCT) with low-coherence interferometry. It excels in obtaining single-scattered signals from specific tissue depths. Despite the system's complexity and the high demand for data processing, this method offers a significant advantage in signal-to-noise ratio, making it particularly suitable for high-resolution imaging scenarios where precise depth information is crucial. Another important technique is the azimuthal technique, which measures the differences in reflected signals at 0° and 90° azimuthal angles. This approach takes advantage of the differences in single-scattering signals from large particles, such as cell nuclei, when viewed from different angles. However, the high demands on optical path design and detector precision present certain challenges in practical applications. The third technique, spatial gating, involves adjusting the source-detector distance to differentiate between shallow and deep tissue scattering signals. Although the design of spatial gating probes is complex, the small size of the probe makes it ideal for applications in confined spaces, such as in the detection of pancreatic cancer. The fourth technique is polarization gating, which leverages the polarization state preservation of single-scattered light to distinguish it from multiple-scattered signals. Among these, the Polarized Light Scattering Spectroscopy (PLSS) systems (Fig.21, Fig.23), based on polarization gating, have been the most extensively researched and have shown high early cancer detection rates in clinical settings (Fig.24), demonstrating immense potential for clinical application. However, traditional PLSS techniques rely on the rotation of polarizers or the orthogonal arrangement of two polarizers to capture single-scattered light, which reduces measurement efficiency and increases system complexity. To overcome these challenges, the snapshot PLSS technique was developed by combining polarization spectrum modulation technology with PLSS (Fig.25). This technique uses a spectral modulation polarizer to directly record the single-scattered light from the sample, allowing the system to operate without the need for rotating polarizers. Building on this, the dual-optical-path snapshot PLSS endoscopic system was developed (Fig.26). This system uses a single retarder to modulate the intensity of polarized scattered light, which is then transmitted in intensity form through a multimode optical fiber to the other end of the endoscope. Compared to traditional three-optical-path PLSS endoscopic systems, this system expands the effective sampling area of PLSS and eliminates the problem of signal aliasing from non-overlapping regions (Fig.26(a)-(b)). Integrating this innovative system with advanced computational algorithms has significantly improved the analysis and interpretation of scattering data, enabling more precise identification of early-stage malignancies. Clinical studies using these enhanced systems have shown high sensitivity and specificity, validating the practicality of PLSS in early cancer diagnostics (Fig.27) and providing new tools and methods for early cancer detection.

Conclusions and Prospects  Recent advancements in LSS underscore its potential in non-invasive cancer diagnostics. The introduction of the snapshot PLSS technique has improved the efficiency and accuracy of single-scattered light measurements while simplifying traditional PLSS systems, offering promising solutions for biomedical imaging. However, both traditional PLSS endoscopic systems and the dual-optical-path snapshot PLSS endoscopic system lack image resolution, which can lead to missed diagnoses when cancerous cell signals are masked by healthy cells in the same area. Thus, developing pixel-level PLSS imaging is a crucial research direction. While LSS, particularly when combined with advanced computational methods, shows significant promise, it currently relies on complex mathematical models that may not fully account for tissue morphology and optical properties. Future research should focus on refining these models by incorporating quantitative measurements and more accurate tissue characterization, improving their clinical applicability. Continued efforts are needed to enhance light scattering measurement accuracy and develop robust data processing algorithms. Additionally, integrating LSS with other imaging modalities and creating more user-friendly, portable systems will be vital for widespread clinical adoption.