A Ludwig Cancer Research study offers pathologists a quick, cheap and reliable tool to diagnose diseases such as early breast cancer with a conventional optical microscope. In a paper published in the current issue of the journal Nature Biotechnology, an interdisciplinary research team led jointly by investigators at the Ludwig Center at Harvard and their colleagues at the Massachusetts Institute of Technology (MIT) demonstrates the utility of the new technique for microscopic diagnosis.
Cellular features used to diagnose certain diseases can be a little too small to be studied reliably. Rather than focusing on increasing the power of microscopes themselves to improve observation of such structures, the researchers--led by Ludwig Harvard's Octavian Bucur and Andrew Beck, who recently left Ludwig Harvard to join PathAI as CEO, and Edward Boyden and Yongxin Zhao of MIT--tweaked the samples themselves. They clinically optimized a technique called expansion microscopy, developed by Boyden and his team at MIT, to expand the physical size of their specimens. This meant that the biopsies could be reliably studied with the kind of microscopes that are found in every pathology laboratory.
"The most exciting thing is that we can use physical tissue expansion to push conventional optical microscopes beyond their limits, with important applications in diagnostic pathology and research," says Bucur, an investigator at the Ludwig Center at Harvard and the Department of Pathology and Cancer Research Institute at the Beth Israel Deaconess Medical Center, who is one of the two lead authors of the paper.
"We can apply this method to any type of clinical sample and all types of human tissues, including normal and cancerous tissues," says Zhao, the other lead author of the paper.
The optical microscope is an essential tool of the diagnostic pathology laboratory. But certain disease structures, including features of the filtration systems of the kidneys and early cancerous lesions of the breast are too small to be accurately studied.
There are solutions to this problem, including electron microscopy. But such instruments are expensive and must be operated by specialized personnel. Boyden, an MIT professor of biological engineering and the co-director of the MIT Center for Neurobiological Engineering, pursued a different solution. His team figured out how to infuse biological specimens with swellable polymers--similar to the chemicals in baby diapers--in an even fashion, so that when water was added to the specimens, the cells or tissues would expand a hundredfold in volume.
Boyden and colleagues demonstrated the uniform expansion of cells and mouse brain tissues in a 2015 Science paper. The method exploits a polymer network that swells uniformly within a tissue sample. After enzymatically cleaving the proteins in the tissue to prevent cracking, water is added to the sample, which has been treated with the polymer, to physically enlarge its finest structures. The technique, which the researchers call expansion microscopy, significantly improves the resolution of conventional microscopes. "In this way, we can image large-scale biological structures--like cancers, or brain circuits--with nanoscale precision, on ordinary microscopes," says Boyden. "My hope is that with expansion microscopy, we can begin to map the building blocks of life systematically, in health and disease states."
The group at MIT teamed with Beck, Bucur and colleagues to optimize the method for diagnostic pathology and research. The team developed a pathology-optimized expansion microscopy to significantly improve the accuracy of computational discrimination between early pre-cancerous lesions with a high or a low risk for cancer transformation, testing their method on breast lesions of this type.
This is notoriously difficult for pathologists to do. "Recent studies show that pathologists differ significantly in their diagnosis of early proliferative lesions," says Humayun Irshad, a postdoctoral fellow at the Ludwig Center at Harvard who developed the computational pipeline for this study. This has a significant impact on the treatment choice, potentially leading to overtreatment, such as unnecessary surgeries, or the neglect of cancers that require early intervention.
"We think that an improved system for differentiating early lesions will potentially prevent hundreds of thousands of misdiagnoses every year in the US," says Beck.
The team also showed that diagnosing certain kidney diseases, which currently requires electron microscopy, can now be performed with over 90% accuracy using expanded clinical samples and a conventional optical microscope. "Being able to eliminate the need for an electron microscope in diagnosing certain diseases will save a lot of money and enable a faster and easier diagnosis for those particular diseases," says Bucur.
The team is currently working to make pathologists aware of the new technique and testing other applications for the method, including studying drug resistance in breast cancer.