A team of Beckman researchers led by Kianoush Falahkheirkhah, a graduate student in chemical and biomolecular engineering, and Rohit Bhargava, director of the Cancer Center at Illinois, developed new software to boost infrared imaging-based cancer diagnosis. In addition to computationally enhancing image resolution, this software will integrate data analysis and accelerate recording speeds of pathological imagery.
Their research is published in Chemometrics and Intelligent Laboratory Systems.
“Most of the strategies for cancer diagnosis today are based on assessing the morphologic detail of the sample, which involves a pathologist analyzing the sample to ascertain the type and severity of the cancer. With IR imaging, we get the chemical composition of a sample, providing us with the volume of information necessary for automated tools to analyze tumors and their surrounding microenvironment,” Falahkheirkhah said.
There are limitations associated with IR imaging, which is much slower than morphological analysis. Added to this is IR imaging's diffraction limit, which results in a lower spatial resolution. These limitations motivated Falahkheirkhah and Bhargava to develop software to make IR imaging 20 times faster with only a 5% drop in accuracy.
Society is currently in the golden age of software development, where developing new software is much faster than developing new hardware. Because of this, scientists can develop software that extracts more out of today’s machinery while waiting on the machines of tomorrow.
“Even though developing new software and computational tools is important, the development of true next-gen machines will require both software and hardware to go hand in hand,” Falahkheirkhah said.
Having been in development for almost 20 years, IR imaging is approaching deployment as a staple technology for cancer diagnosis. With many of its pitfalls addressed, the main hurdles that lie in the way of its mainstream use include publicity and training.
As with every new technology, users must be trained in the software’s use. According to Falahkheirkhah, this software can be modified to a user’s preferences and ease of access in as short a time as one week.
This project exemplifies how forging ahead and breaking barriers in science requires the interdisciplinary efforts of medical professionals, engineers, and biologists alike. The Beckman Institute is at the heart of such collaborations.
Editor's note: the paper associated with this work can be found at: https://www.sciencedirect.com/science/article/abs/pii/S0169743921001581?via%3Dihub
Chemometrics and Intelligent Laboratory Systems
Method of Research
Subject of Research
Deep learning-based protocols to enhance infrared imaging systems
Article Publication Date
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.