A Korean research team has developed a new imaging system that combines light and ultrasound to more accurately distinguish thyroid cancer without a biopsy. When a nodule is found in the thyroid, most patients have to undergo a needle biopsy, but that burden is expected to drop significantly.
Kim Cheolhong, professor in the Department of Electrical and Electronic Engineering, the Department of IT Convergence Engineering, the Department of Mechanical Engineering, and the Graduate School of Convergence at Pohang University of Science and Technology POSTECH, along with Professors Lim Dongjun and Lee Jaekyung at the Catholic University Seoul St. Mary's Hospital and Professor Park Byeolli at Sungkyunkwan University, published these findings in the international journal Science Advances on Aug. 27.
In general, thyroid cancer diagnosis proceeds by first performing an ultrasound exam and, if a suspicious malignant nodule is found, collecting tissue using a needle. However, ultrasound alone has low accuracy in distinguishing benign from malignant, so many nodules that are not actually cancer undergo unnecessary biopsy. As a result, patients bear physical and psychological burdens, and medical staff face concerns about diagnostic accuracy.
With support from the POSTECH-Catholic University Biomedical Engineering Institute (POGA Institute), the team has developed photoacoustic imaging (PAI) technology. Malignant nodules have active metabolism and low oxygen saturation, so the method measures blood oxygen saturation from the faint ultrasound signals emitted by red blood cells when exposed to a laser (light), then determines whether the nodule is benign or malignant. However, this approach alone had limitations in classifying the various types of thyroid cancer.
The team used data from a total of 106 patients, including 45 with papillary thyroid carcinoma, 32 with follicular tumors, and 29 with benign nodules. From their photoacoustic images, they extracted various parameters such as oxygen saturation, asymmetry of distribution, and spectral slope, and analyzed them with machine learning (AI) techniques to devise a new diagnostic metric, the ATA-Photoacoustic (ATAP) score.
The results showed a sensitivity of 97% for detecting malignant nodules, which remained very high. At the same time, the specificity for excluding benign nodules from unnecessary testing was 38%, more than double that of conventional ultrasound diagnosis (17%). This indicates the potential to reduce unnecessary tests, lessen patient burden, and even cut medical costs.
Kim Cheolhong said, "This study is significant in that it combined photoacoustic and ultrasound to distinguish malignancy, including follicular tumors that were previously difficult to diagnose." Park Byeolli noted, "Through follow-up research, we will improve the completeness of the technology and continue large-scale clinical validation to develop it into a medical device that can be used in real-world clinical settings."
References
Science Advances (2025), DOI: https://doi.org/10.1126/sciadv.ady6173