Original contributionEvaluation of 2 whole-slide imaging applications in dermatopathology
Introduction
In anatomic pathology, whole-slide imaging (WSI; the process of digitally capturing an entire glass slide at high magnification) holds tremendous promise for transforming current practice. The integration of WSI may encourage more efficient workflow, image storage, collaboration, interactive teaching tools, and the possibility of computer-assisted diagnostic devices similar to those available in radiology. Already, studies illustrate the benefits of digital slides in facilitating novel pathology education programs [1], [2]. In subspecialty fields such as dermatopathology where access to an expert is limited, studies have documented the success of both static (store and forward) [3], [4], [5] and dynamic telepathology [6].
Whereas previous validation studies comparing digital slides to glass slides did suggest that pathologists were capable of achieving reliable diagnoses with digital slides [7], [8], [9], [10], [11], current applications have not gained widespread acceptance. Infrastructure is costly and creation of the digital slides is time consuming. Despite advances in scanning systems, perfect focus is not consistently achieved across the entire slide. A major challenge is that viewing a digital slide is inherently different from evaluating a glass slide with a microscope. As a result, current viewing applications are often difficult to use and inefficient. Krupinski et al [12] tracked eye movements of pathologists using telepathology slides. They found differences in the visual behaviors of viewers based on their experience and suggested their results may help guide more intuitive designs of the workstations [12]. Schrader et al [13] developed a tool to document the “observation path” or image sequence viewed by the pathologist and correlated these findings with the pathology oral dictation report. Such an instrument might help us better understand how pathologists view digital slides and assist in quality control [13]. More such studies that document workflow with digital slides are warranted, with rather expanded study focus.
To better understand the current limitations of digital slide systems, and guide the future development of more usable systems, we were interested in describing and comparing baseline pathologist workflow behavior with both traditional glass microscopy and digital applications. We were particularly interested in the dermatopathology specialty as the unique nature of this field might greatly benefit from digital viewing systems. The traditional process of transferring a sample from the dermatologist to the histology laboratory and then to the pathologist is tedious. An intuitive digital slide acquisition system might facilitate processing of the biopsy. When comparing the traditional microscope to digital applications, we specifically expected to find differences in (1) time spent per slide and (2) time spent at high-power magnification. Time spent per slide is thought to reflect both the level of case complexity as well as the efficiency of the viewing application in facilitating diagnosis. Time spent at high magnification is thought to provide insight into the quality of the image produced by the viewing application as well as the pathologists' viewing preferences with that modality. Finally, we were interested in testing whether a digital application that had been designed in-house with the goal of better approximating dermatopathologist behavior was in fact superior to commercial applications.
Section snippets
Subjects
Three board-certified pathologists with dermatopathology training participated in the study: 2 attending physicians and one fellow. One subject had extensive experience using digital slides, one had limited experience, and one had no prior experience. We chose to restrict our subjects to pathologists trained specifically in dermatopathology to decrease the variability in physician expertise and skill. These design decisions limited the study size.
Case selection
The process of case selection was designed to
Digital slide acquisition
A total of 81 digital slides were captured using the system's default settings. Of these, 7 (8.6%) were identified to have large out-of-focus areas visible at low magnification in the subsampled thumbnail image. The slides were rescanned with manually placed focus points. Of this subset of glass slides, none had to be rescanned a second time. Image imperfections only visible at greater than ∼2 × magnification did not result in a reacquisition of the digital slide.
Time per slide
Overall, between the 3
Digital slide acquisition
A total of 8.6% (7/81) of glass slides scanned by the system needed to be rescanned because of major failures of the focusing algorithms to acquire an adequate digital slide. Several glass slides contained imperfections, including tissue folds, small air bubbles, and dust particles that most likely interfered with the system's algorithm. However, several of the glass slides that produced poor quality scans did not seem to have any imperfections and would have passed standard quality guidelines.
Conclusion
Some routine tasks in dermatopathology practice, such as identifying eosinophils, neutrophils, apoptosis, and grading dysplastic nevi, are not easily accomplished with these hardware and software configurations. This may result in diagnostic discrepancies and ultimately affect patient outcomes. A study performed with digital slide acquisitions at 40× magnification might be helpful in further examining these issues. Experimenting with other image quality compression settings may also be of
References (13)
- et al.
An array microscope for ultrarapid virtual slide processing and telepathology. Design, fabrication, and validation study
Hum Pathol
(2004) - et al.
Use of whole slide imaging in surgical pathology quality assurance: design and pilot validation studies
Hum Pathol
(2006) - et al.
Eye-movement study and human performance using telepathology virtual slides. Implications for medical education and differences with experience
Hum Pathol
(2006) - et al.
Teaching medical histology at the University of South Carolina School of Medicine: transition to virtual slides and virtual microscopes
Anat Rec
(2003) - et al.
Virtual microscopy for learning and assessment in pathology
J Pathol
(2004) - et al.
Teledermatopathology: a controlled study about diagnostic validity and technical requirements for digital transmission
Am J Dermatopathol
(2006)
Cited by (62)
Digitization of Pathology Labs: A Review of Lessons Learned
2023, Laboratory InvestigationComparative Assessment of Digital Pathology Systems for Primary Diagnosis
2021, Journal of Pathology InformaticsDigital pathology as a platform for primary diagnosis and augmentation via deep learning
2020, Artificial Intelligence and Deep Learning in PathologyConcordance, intra- and inter-observer agreements between light microscopy and whole slide imaging for samples acquired by EUS in pancreatic solid lesions
2019, Digestive and Liver DiseaseCitation Excerpt :The median time needed to perform WSI evaluation was significantly longer than the one occurred to examine LM slides (p < 0.0001). This result is in accordance with three previous studies all showing an advantage of LM regarding the time spent to review slides [42–44]. An increased time to diagnose may be of particular concern for financially pressed health care systems.
Validation of whole-slide digitally imaged melanocytic lesions: Does z-stack scanning improve diagnostic accuracy?
2019, Journal of Pathology InformaticsCitation Excerpt :For this purpose, the college of American pathologists (CAP) provided a guideline for validation of WSI.[5] A number of dermatopathology WSI validation studies have been published.[6–10] These studies show considerable variation in sample sizes, case selection, and types of lesions studied.
Digital pathology: semper ad meliora
2019, Pathology