Researchers at Ludwig Harvard have developed the Orion platform, which integrates multiplex imaging with clinical pathology to enhance cancer diagnostics. Orion allows the visualization of multiple molecular markers in individual cells, providing valuable insights into cancer aggressiveness and therapy response. By combining human expertise and artificial intelligence, the platform accurately predicts progression-free survival in colon cancer patients. The innovative approach bridges the gap between academic research and clinical settings, promising to revolutionize cancer diagnosis and improve patient outcomes.
Scientists at the Ludwig Center at Harvard University have developed a groundbreaking imaging technology that combines the longstanding microscopic analysis techniques used in pathology labs with the emerging field of multiplex imaging, which allows the visualization of multiple molecular markers within individual cells. This innovative approach has the potential to revolutionize cancer diagnostics by uncovering molecular characteristics related to cancer aggressiveness and susceptibility to therapy. While multiplex imaging has mainly been limited to academic research settings due to its incompatibility with the standardized processes of clinical pathology, a team of researchers led by Sandro Santagata, Peter Sorger, Jia-Ren Lin, and Yu-An Chen from Ludwig Harvard has developed a platform named Orion that bridges the gap between these two realms of analysis.
In an article published in Nature Cancer, the researchers not only present the development of the Orion platform but also demonstrate its effectiveness in predicting progression-free survival in colon cancer patients. The platform was successfully used by both human experts and artificial intelligence algorithms to identify cellular and molecular features in tumors. Santagata emphasized that Orion was specifically designed to be a practical tool for use in real-life clinical settings, with the primary goal of enhancing the diagnosis of human diseases.
Clinical pathology traditionally relies on microscopic analysis of tissues that are embedded in wax, sliced into thin strips, and stained with hematoxylin and eosin (H&E) to reveal the cellular morphology. Nowadays, researchers are increasingly incorporating machine learning and artificial intelligence algorithms to analyze the pink-and-purple images produced by H&E staining. This approach, coupled with access to a vast archive of H&E images and samples, enables the identification of patterns that may not be easily discernible to the human eye, thereby enhancing cancer diagnosis and treatment management.
On the other hand, the use of antibodies to capture molecular characteristics of tumors has evolved significantly and is now an essential component of cancer diagnostics. Recent advancements involve simultaneously applying multiple antibodies to tissues, generating “immunofluorescence” images that provide insights into gene expression and cellular interactions within tumors. Santagata explained that by gathering molecular information, distinct cell types can be identified, their current state determined, and their interactions with neighboring cells assessed to better understand tumor growth.
However, the application of multiplex imaging in clinical practice has been limited by a couple of challenges. Firstly, the methods generally do not allow for the evaluation of whole slides, which is a requirement for clinical diagnostics. Additionally, these methods are not compatible with the established techniques of H&E microscopy, which serves as the foundation of clinical pathology.
To overcome these limitations, the Ludwig Harvard team collaborated with a Seattle-based startup called RareCyte and received support from the Ludwig Tumor Atlas project and a Small Business Innovation Research grant from the National Cancer Institute. Together, they developed the Orion platform, which integrates multiple antibody images with H&E staining on the same tissue section, enabling the comprehensive analysis of information across an entire slide.
To validate the platform, the researchers analyzed colorectal cancer specimens from 40 patients, identifying molecular features closely associated with negative outcomes. Through extensive analysis of approximately 15,000 combinations of biomarkers, they identified the most predictive indicators of patient prognosis. These biomarkers were then applied to samples from an additional 34 colorectal cancer patients, accurately predicting poor prognosis with a high level of accuracy (a 1 in 20 chance of being wrong).
The combined use of multiplex and H&E imaging yielded illuminating findings, revealing relationships between molecular markers, cell morphology, and tumor topography. For example, inflammation or immune activity at the tumor’s edge was found to have pathological significance. Additionally, the molecular basis of a specific tissue morphology associated with metastasis propensity was discovered.
Moving forward, the researchers plan to further establish the effectiveness of the Orion platform by exploring markers related to treatment response and resistance. Ultimately, the discoveries made using Orion will need to be validated through large-scale clinical trials.
Santagata expressed excitement about the prospect of incorporating advanced imaging tools into routine clinical analysis, as it has the potential to unlock a wealth of knowledge about cancer tissues.
