This article summarizes key observations from the peer-reviewed study “Validation of a Portable Whole-Slide Imaging System for Frozen Section Diagnosis.”
Authors: Rajiv Kumar Kaushal, Sathyanarayanan Rajaganesan, Vidya Rao, Akash Sali, Balaji More, Sangeeta B. Desai
Published: Journal of Pathology Informatics, Volume 12, Issue 1, 2021
DOI: https://doi.org/10.4103/jpi.jpi_95_20
Digital pathology has advanced rapidly, but frozen section diagnosis remains one of the most demanding applications in surgical pathology. During surgery, pathologists must deliver rapid interpretations that directly influence clinical decisions within minutes. This research reflects the growing role of digital pathology technology in healthcare and surgical decision-making.
This study evaluates whether whole slide imaging (WSI) using a portable scanner can support frozen section diagnosis with accuracy comparable to optical microscopy (OM). Because frozen section diagnosis guides immediate decisions such as margin assessment and tumor identification, any digital solution must maintain both speed and diagnostic reliability. The findings contribute valuable data to ongoing discussions about how digital pathology systems can enable accurate remote consultation while maintaining the speed and precision required during surgery.
Frozen section analysis plays a central role in intraoperative care. Tissue samples are rapidly frozen, sectioned, stained, and examined to provide immediate diagnostic feedback while the patient remains under anesthesia.
This process supports key clinical decisions, including evaluating surgical margins, determining lymph node involvement, and identifying unexpected tumors. Traditionally, this workflow depends on direct evaluation of glass slides using optical microscopy within the pathology laboratory.
However, increasing demand for subspecialty expertise and continuous diagnostic availability has created interest in digital alternatives. While digital pathology enables remote viewing of digital slides, frozen section workflows introduce additional complexity due to time constraints, tissue artifacts, and the need for rapid interpretation.
Whole slide imaging (WSI) converts traditional glass slides into high-resolution digital images that can be reviewed through standard viewing software. In this study, researchers conducted a blinded prospective comparison between conventional optical microscopy and WSI to evaluate how well digital pathology performs in real frozen section workflows.
The analysis included 60 frozen section cases (120 slides) representing routine clinical scenarios. Each case contained both hematoxylin and eosin and toluidine blue-stained slides, and cases were distributed across common intraoperative use cases: lymph node assessment (45%), margin assessment (37%), and primary diagnosis (18%). The validation followed College of American Pathologists guidelines, requiring WSI to demonstrate non-inferiority compared to conventional microscopy.
The study evaluated a portable slide scanner (Grundium Ocus®) operating at 20× magnification (0.48 µm/pixel). The system demonstrated consistent technical performance in handling frozen section slides.
The first-time successful scanning rate was 89.1%, with remaining slides successfully rescanned on a second attempt. Mean scanning time per slide was approximately 1 minute 47 seconds for hematoxylin and eosin slides and 1 minute 46 seconds for toluidine blue slides.
Digital image size averaged 0.83 GB per hematoxylin and eosin slide and 0.71 GB per toluidine blue slide, resulting in a total data volume of 87.8 GB for all cases. These results indicate that scanning adds approximately 3–4 minutes per case, which remains within acceptable limits for intraoperative workflows.
A total of 480 diagnostic reads were analyzed across both modalities, demonstrating high concordance between whole slide imaging (WSI) and optical microscopy (OM). Overall diagnostic accuracy was 93.5% for optical microscopy and 92.5% for WSI when compared with the reference standard.
When considering only major discrepancies, diagnostic accuracy increased to 95.42% for optical microscopy and 95.83% for WSI. The difference in clinically significant discrepancies between the two methods was minimal (0.41%) and not statistically significant. In addition, both inter-observer and intra-observer agreement were near perfect, with kappa values greater than 0.8 across all participating pathologists. These results indicate that whole slide imaging is non-inferior to optical microscopy for frozen section diagnosis across multiple specimen types.
Turnaround time is a critical factor in frozen section diagnosis. In this study, the average workflow included approximately 10 minutes for slide preparation, an additional 3 minutes and 33 seconds for scanning, and about 1 minute and 12 seconds for image interpretation.
This resulted in a total average turnaround time of approximately 14 minutes and 45 seconds per case. This falls within the College of American Pathologists recommended limit of 20 minutes for frozen section reporting.
Although interpretation time using WSI was slightly longer than with optical microscopy, the overall workflow remained clinically acceptable. The additional time required for scanning did not significantly impact the ability to deliver timely intraoperative diagnoses.
Image quality was rated as average to high in most cases, with no significant difference between staining methods. The resolution achieved by the portable scanner was consistent with standard digital pathology systems. However, pathologists reported lower confidence when using whole slide imaging compared to optical microscopy. Confidence levels varied by specimen type, with higher confidence observed in lymph node assessment and lower confidence in primary diagnosis cases.
These findings suggest that experience and training play an important role in adapting to digital pathology workflows, particularly for time-sensitive applications such as frozen section diagnosis.
The portable slide imaging system demonstrated several practical advantages for clinical implementation. Its compact design makes it suitable for space-constrained environments such as frozen section rooms, and setup time is minimal, requiring less than five minutes to begin scanning.
Key practical benefits include:
Efficient handling of wet frozen section slides
Access to digital slides through standard web-based viewing software
Support for digital workflows and image review across networked environments
While the study was conducted within a local network, such systems may support remote consultation where appropriate infrastructure is available. In addition, potential limitations such as pathologist training, workflow integration, and data management should be considered when implementing digital pathology systems in routine clinical practice. Successful implementation depends on these factors as well as consistent system use.
This validation study demonstrates that whole slide imaging using a portable scanner is non-inferior to optical microscopy for frozen section diagnosis. Diagnostic accuracy, concordance, and turnaround time were all within clinically acceptable ranges.
These findings support the role of digital pathology in frozen section workflows, particularly where access to pathology expertise may be limited. However, effective adoption depends on appropriate training, workflow integration, and supporting technical infrastructure.
A curated collection of digital pathology studies and references is available on Grundium’s website.
