A research team from The Hong Kong University of Science and Technology (HKUST), together with an HKUST-incubated medtech startup, PhoMedics Limited, has developed Glanzir®, the world's first artificial intelligence (AI)-enabled, slide-free pathology imaging system. The system enables direct imaging of fresh tissue without the need for conventional procedures such as freezing, sectioning, and staining, producing histological images in approximately three minutes in an operating room setting for intraoperative assessment.
Compared with using formalin-fixed and paraffin-embedded (FFPE) tissue, which is widely regarded as the gold standard in pathology, Glanzir® has demonstrated approximately 85% or higher diagnostic concordance. Upon completion of large-scale clinical trials, the team expects this concordance to improve further to around 95%. The technology has the potential to shorten intraoperative diagnostic time while preserving intact tissue for downstream analyses. The team plans to collaborate with public and private hospitals to advance clinical adoption, with the goal of improving overall diagnostic workflow efficiency.
With an aging population and rising cancer rates, demand for pathology services continues to grow. Histopathology remains fundamental to cancer diagnosis. In current clinical practice, two primary techniques are used for tissue assessment: FFPE and frozen section analysis. While FFPE is considered the gold standard due to its high accuracy, it typically requires several days to a week to complete. Frozen section analysis, by contrast, can provide preliminary results within 30 to 45 minutes during surgery, but its diagnostic accuracy is suboptimal compared to FFPE, occasionally necessitating repeat surgery for patients.
To address these constraints, a research team led by Prof. Terence WONG Tsz-Wai, Associate Head and Associate Professor in the Department of Chemical and Biological Engineering at HKUST, together with his medtech startup PhoMedics Limited, has developed Glanzir®. The system operates based on autofluorescence imaging (AFI) using specific wavelengths of light to excite naturally occurring fluorophores within the tissue. These signals are then processed through a deep learning-based virtual staining algorithm to generate images comparable to those produced by conventional hematoxylin and eosin (H&E) staining.
In practice, healthcare professionals only need to place fresh tissue samples into a dedicated holder and insert them into the system to initiate scanning. The device automatically performs region of interest (ROI) selection, point-by-point focusing and scanning, followed by real-time image processing and virtual staining. A built-in progress display allows clinicians to monitor scanning status and align the process with surgical workflows.
Key Advantages of Glanzir®
Compact for operating room use: Measuring only 56 cm (L) × 56 cm (W) × 30 cm (H), with a total weight of 30 kg.
Rapid, consistent imaging: Completes the imaging process in approximately three minutes, which is around ten times faster than frozen section, while achieving diagnostic concordance comparable to FFPE (~85%) which is expected to be improved to 95% after the large-scale clinical trials.
Streamlined operation: One-click scanning with automated image acquisition and processing, helping to shorten intraoperative turnaround time.
Non-destructive analysis with downstream compatibility: Eliminates the need for sectioning while preserving intact tissue structure, allowing samples to remain available for immunohistochemistry (IHC), fluorescence in situ hybridization (FISH), and other molecular analyses.
The technology was developed by Prof. Wong, an expert in histological imaging. His earlier work on the Computational High-throughput Autofluorescence Microscopy by Pattern Illumination (CHAMP) secured funding under the first batch of the Innovation and Technology Commission's Research, Academic and Industry Sectors One-plus (RAISe+) Scheme, laying the foundation for subsequent development.
Prof. Wong said, "In cancer surgery, incomplete tumour resection may lead to the need for repeat operations. Clinical data suggest that around 10‒20% of cases require a second operation. Our imaging technology is designed to provide timely reference during surgery. By generating histological images within about three minutes, it helps clinicians assess whether residual malignant cells remain. This may reduce the risk of repeat surgery, while also improving surgical workflow and supporting more effective use of healthcare resources."
The team has been collaborating with major public and private hospitals and healthcare institutions in Hong Kong and the Chinese Mainland, including Queen Mary Hospital, Prince of Wales Hospital, and Hong Kong Sanatorium & Hospital, as well as Tongji Hospital at Tongji Medical College of Huazhong University of Science and Technology and Sun Yat-sen Memorial Hospital of Sun Yat-sen University. To date, more than 2,000 lung and breast cancer patient samples have been collected for system training and validation. In the coming three months, the team will conduct large-scale clinical trials to further advance the system towards clinical deployment.
