Removing the ‘Noise’ of Cancer Signals in Blood, HYU Research Team Develops Breast Cancer Classification Technology
Filters background signals just like removing radio noise to clearly detect micro extracellular vesicles
Utilizes general fluorescence microscopes without expensive equipment, outstanding performance in distinguishing breast cancer patients from healthy individuals
A research team led by Professor Choi Sung-young of the Major in Biomedical Engineering Major, School of Electrical and Biomedical Engineering at Hanyang University's College of Engineering, in collaboration with Professor Park Kyung-hwa's research team at Korea University College of Medicine, has developed 'EV-FFT', a computational imaging technology that precisely detects the minute fluorescence signals of extracellular vesicles (EVs) in the blood using a standard fluorescence microscope. The research team also confirmed its performance in distinguishing breast cancer patients from healthy individuals by jointly analyzing the expression patterns of multiple surface proteins.
Extracellular vesicles are nanometer-sized particles secreted by cells to the outside, and containing the cell's proteins and genetic material, they are gaining attention as promising biomarkers for cancer diagnosis through liquid biopsy. However, because they are extremely small, the fluorescence signals generated from individual extracellular vesicles are weak and easily buried in background noise, making them difficult to detect with standard fluorescence microscopes. High resolution microscopes can be utilized, but they carry limitations such as requiring expensive special equipment like high-power lasers and high-sensitivity detectors, as well as having a restricted area that can be observed at once.
The research team solved this problem by computationally processing the fluorescence signals instead of upgrading the equipment's performance. EV-FFT utilizes 'photobleaching'—a phenomenon where fluorescent materials gradually darken the more they are exposed to light—as analytical information rather than noise. By converting the fluorescence signals obtained through continuous photographing of extracellular vesicles into the frequency domain via Fast Fourier Transform (FFT), the gradually decreasing signal of the extracellular vesicles can be separated from the fast and irregularly changing background noise. Afterward, only the signals related to the extracellular vesicles are extracted and reconstructed into high-contrast fluorescence images.
As a result of verification using fluorescent nanoparticles, 87.0% of the reference nanoparticles were detected in EV-FFT while only 7.5% were detected in a typical single fluorescence image. The signal-to-noise ratio also increased from 4.4 to 14.1, effectively restoring weak nanoparticle signals that were obscured by noise in existing images.
It also showed high detection sensitivity in the analysis of extracellular vesicles derived from actual cancer cells. The limit of detection for CD9-positive extracellular vesicles was 69.1 per microliter, allowing detection down to a concentration approximately 13 times lower than typical single fluorescence imaging and about 50 times lower than ELISA, an existing enzyme-linked immunosorbent assay. In addition, as a result of analyzing eight protein markers across three types of cancer cell lines, the EV-FFT measurement values showed a high correlation with the ELISA results.
To confirm its potential for clinical application, the research team analyzed the plasma samples of 8 healthy individuals and 26 breast cancer patients. The breast cancer patients included 13 individuals each from stages 2 and 4. When the research team integrated and analyzed the expression information of the 8 protein markers, the average Area Under the Curve (AUC) distinguishing the breast cancer patient group from the healthy individual group was found to be 0.992. The sensitivity was 97.2% and the specificity was 99.3%, demonstrating a higher classification performance than when utilizing existing single fluorescence images.
As the number of markers was reduced, the classification performance gradually decreased. Through this, it was also confirmed that analyzing the information of multiple extracellular vesicle populations together is more important in distinguishing cancer patient groups than analyzing just a single biomarker.
Professor Choi Sung-young explained, "This research is significant in that it has transformed photobleaching, which is generally recognized as a phenomenon that degrades the performance of fluorescence imaging, into a useful signal for detecting extracellular vesicles. We can sensitively detect nanometer-sized particles using only a standard fluorescence microscope and computational imaging technology without adding expensive super-resolution equipment."
He continued, "By jointly analyzing information obtained from multiple populations of extracellular vesicles, we were able to more effectively reflect the heterogeneity of cancer in individual patients. If verified on a larger clinical cohort and across various types of cancer in the future, it has the potential to be utilized as an analytical tool that complements non-invasive cancer classification and treatment monitoring."
This research was conducted with support from the Ministry of Science and ICT and the National Research Foundation of Korea's Basic Research Program and Pioneer Research Center Program. The research results were published online in the international academic journal ACS Sensors on July 29, 2026.
The title of the paper is 'Frequency-Domain Photobleaching Denoising for Sensitive Extracellular Vesicle Detection Using Fluorescence Microscopy'. Hanyang University researchers Kang Ji-soo, Yoon Hyo-geun, and Shin Soo-yeon participated as co-first authors, while Professor Choi Sung-young of Hanyang University and Professor Park Kyung-hwa of Korea University College of Medicine participated as co-corresponding authors.
Source : NewsH(https://www.newshyu.com/news/articleView.html?idxno=1025750)
Utilizes general fluorescence microscopes without expensive equipment, outstanding performance in distinguishing breast cancer patients from healthy individuals
A research team led by Professor Choi Sung-young of the Major in Biomedical Engineering Major, School of Electrical and Biomedical Engineering at Hanyang University's College of Engineering, in collaboration with Professor Park Kyung-hwa's research team at Korea University College of Medicine, has developed 'EV-FFT', a computational imaging technology that precisely detects the minute fluorescence signals of extracellular vesicles (EVs) in the blood using a standard fluorescence microscope. The research team also confirmed its performance in distinguishing breast cancer patients from healthy individuals by jointly analyzing the expression patterns of multiple surface proteins.
Extracellular vesicles are nanometer-sized particles secreted by cells to the outside, and containing the cell's proteins and genetic material, they are gaining attention as promising biomarkers for cancer diagnosis through liquid biopsy. However, because they are extremely small, the fluorescence signals generated from individual extracellular vesicles are weak and easily buried in background noise, making them difficult to detect with standard fluorescence microscopes. High resolution microscopes can be utilized, but they carry limitations such as requiring expensive special equipment like high-power lasers and high-sensitivity detectors, as well as having a restricted area that can be observed at once.
The research team solved this problem by computationally processing the fluorescence signals instead of upgrading the equipment's performance. EV-FFT utilizes 'photobleaching'—a phenomenon where fluorescent materials gradually darken the more they are exposed to light—as analytical information rather than noise. By converting the fluorescence signals obtained through continuous photographing of extracellular vesicles into the frequency domain via Fast Fourier Transform (FFT), the gradually decreasing signal of the extracellular vesicles can be separated from the fast and irregularly changing background noise. Afterward, only the signals related to the extracellular vesicles are extracted and reconstructed into high-contrast fluorescence images.
As a result of verification using fluorescent nanoparticles, 87.0% of the reference nanoparticles were detected in EV-FFT while only 7.5% were detected in a typical single fluorescence image. The signal-to-noise ratio also increased from 4.4 to 14.1, effectively restoring weak nanoparticle signals that were obscured by noise in existing images.
It also showed high detection sensitivity in the analysis of extracellular vesicles derived from actual cancer cells. The limit of detection for CD9-positive extracellular vesicles was 69.1 per microliter, allowing detection down to a concentration approximately 13 times lower than typical single fluorescence imaging and about 50 times lower than ELISA, an existing enzyme-linked immunosorbent assay. In addition, as a result of analyzing eight protein markers across three types of cancer cell lines, the EV-FFT measurement values showed a high correlation with the ELISA results.
To confirm its potential for clinical application, the research team analyzed the plasma samples of 8 healthy individuals and 26 breast cancer patients. The breast cancer patients included 13 individuals each from stages 2 and 4. When the research team integrated and analyzed the expression information of the 8 protein markers, the average Area Under the Curve (AUC) distinguishing the breast cancer patient group from the healthy individual group was found to be 0.992. The sensitivity was 97.2% and the specificity was 99.3%, demonstrating a higher classification performance than when utilizing existing single fluorescence images.
As the number of markers was reduced, the classification performance gradually decreased. Through this, it was also confirmed that analyzing the information of multiple extracellular vesicle populations together is more important in distinguishing cancer patient groups than analyzing just a single biomarker.
Professor Choi Sung-young explained, "This research is significant in that it has transformed photobleaching, which is generally recognized as a phenomenon that degrades the performance of fluorescence imaging, into a useful signal for detecting extracellular vesicles. We can sensitively detect nanometer-sized particles using only a standard fluorescence microscope and computational imaging technology without adding expensive super-resolution equipment."
He continued, "By jointly analyzing information obtained from multiple populations of extracellular vesicles, we were able to more effectively reflect the heterogeneity of cancer in individual patients. If verified on a larger clinical cohort and across various types of cancer in the future, it has the potential to be utilized as an analytical tool that complements non-invasive cancer classification and treatment monitoring."
This research was conducted with support from the Ministry of Science and ICT and the National Research Foundation of Korea's Basic Research Program and Pioneer Research Center Program. The research results were published online in the international academic journal ACS Sensors on July 29, 2026.
The title of the paper is 'Frequency-Domain Photobleaching Denoising for Sensitive Extracellular Vesicle Detection Using Fluorescence Microscopy'. Hanyang University researchers Kang Ji-soo, Yoon Hyo-geun, and Shin Soo-yeon participated as co-first authors, while Professor Choi Sung-young of Hanyang University and Professor Park Kyung-hwa of Korea University College of Medicine participated as co-corresponding authors.
Source : NewsH(https://www.newshyu.com/news/articleView.html?idxno=1025750)