University of Illinois develops compact camera to spot cancer-linked lymph nodes

A camera platform developed at the University of Illinois Urbana-Champaign (UIUC) could help surgeons identify lymph nodes connected to a cancer tumor and assess whether the cancer has spread to them.

Published in Optica, the research into a bioinspired imager suitable for clinical use has led to a compact camera that captures ultraviolet, near-infrared and visible images using a single chip. 

The researchers were inspired by the multiwavelength vision capability of the mantis shrimp and its ability to detect both linearly and circularly polarized light. UIUC previously drew inspiration from the animal when developing a bioinspired endoscope system capable of capturing color images and resolving two near-IR probes simultaneously. 

Assessing lymph nodes is a key consideration during breast cancer surgery, when surgeons must decide which lymph nodes to biopsy, remove or preserve. Removing the right ones helps make sure all the cancer is eliminated, while preserving unaffected nodes can reduce complications like lymphedema.

Near-IR imaging can find the lymph nodes, but does not necessarily reveal whether the nodes are metastatic. And while some label-free optical methods can provide the biochemical contrast needed to detect cancer, they are typically not integrated with other imaging modalities. To provide both the location and likely cancer status of a lymph node requires images using different wavelengths of light, while keeping those images carefully aligned.

"Existing tools can show where lymph is draining but can't reliably indicate whether a particular lymph node is involved with cancer while the operation is underway," said Viktor Gruev from UIUC's Biosensors Lab. "This can lead to over-treatment, under-treatment or the need for a second procedure later."

Helping patients receive more precise surgery

As described in the project's paper, the UIUC sensor architecture employs pixel-level multiband spectral filters combined
with vertically stacked photodiodes, to sample a deep UV emission band together with red, green, and blue (RGB) and dual near-IR sub-bands.

"Interleaving complementary pixel classes yields perfectly coregistered UV, color, and near-IR images without beam splitting or multiple cameras," noted the project in its paper. The researchers used a mirror-based lens to keep the wide range of wavelengths in focus together and developed image-reconstruction software to turn the different signals into clear, well-aligned images.

This design parallels that of the mantis shrimp eyes, which contain stacked rows of photoreceptors each tuned to different parts of the spectrum.

"We borrowed that idea to allow our camera to collect several kinds of optical information from exactly the same place at the same time," said Viktor Gruev. "This is very helpful when you want images to line up perfectly during surgery."

In a clinical workflow the infrared wavelengths could detect indocyanine green (ICG), a standard dye used to find lymph nodes, while a brief UV-light measurement detects the tissue's inherent fluorescence, to determine whether the lymph node is cancerous without requiring a cancer-targeted label. The visible light capability provides a standard camera view giving surgeons the anatomical context needed to understand what they are viewing.

In trials on cell samples and biopsies covering 94 lymph nodes from 33 patients, camera images were compared with standard pathology findings from pathologists who did not see the ultraviolet imaging results. The ultraviolet readout achieved 97 percent sensitivity and 89 percent specificity, while the near-infrared signal reliably located the lymph nodes.

"If future testing is successful, this system could help patients receive more precise surgery," commented Gruev. "Beyond breast cancer, the approach could be useful in other cancers where lymph node status matters or anywhere that rapid, label-free tissue assessment could help during surgery or pathology assessment."