Cited by 64AI in cancer detectionClinical Laboratory Practices and Quality ControlArtificial Intelligence in Healthcare and Education

The performance of digital microscopy for primary diagnosis in human pathology: a systematic review

Araújo, Anna Luíza Damaceno, Arboleda, Lady Paola Aristizábal, Palmier, Natalia Rangel, Fonsêca, Jéssica Montenegro, de Pauli Paglioni, Mariana, Gomes-Silva, Wagner, Ribeiro, Ana Carolina Prado, Brandão, Thaís Bianca +8 more·2019·Virchows Archiv

Abstract

Validation studies of whole slide imaging (WSI) systems produce evidence regarding digital microscopy (DM). This systematic review aimed to provide information about the performance of WSI devices by evaluating intraobserver agreement reported in previously published studies as the best evidence to elucidate whether DM is reliable for primary diagnostic purposes. In addition, this review delineates the reasons for the occurrence of discordant diagnoses. Scopus, MEDLINE/PubMed, and Embase were searched electronically. A total of 13 articles were included. The total sample of 2145 had a majority of 695 (32.4%) cases from dermatopathology, followed by 200 (9.3%) cases from gastrointestinal pathology. Intraobserver agreements showed an excellent concordance, with values ranging from 87% to 98.3% (κ coefficient range 0.8-0.98). Ten studies (77%) reported a total of 128 disagreements. The remaining three studies (23%) did not report the exact number and nature of disagreements. Borderline/challenging cases were the most frequently reported reason for disagreements (53.8%). Six authors reported limitations of the equipment and/or limited image resolution as reasons for the discordant diagnoses. Within these articles, the reported pitfalls were as follows: difficulties in the identification of eosinophilic granular bodies in brain biopsies; eosinophils and nucleated red blood cells; and mitotic figures, nuclear details, and chromatin patterns in neuropathology specimens. The lack of image clarity was reported to be associated with difficulties in the identification of microorganisms (e.g., Candida albicans, Helicobacter pylori, and Giardia lamblia). However, authors stated that the intraobserver variances do not derive from technical limitations of WSI. A lack of clinical information was reported by four authors as a source for disagreements. Two studies (15.4%) reported poor quality of the biopsies, specifically small size of the biopsy material or inadequate routine laboratory processes as reasons for disagreements. One author (7.7%) indicated the lack of immunohistochemistry and special stains as a source for discordance. Furthermore, nine studies (69.2%) did not consider the performance of the digital method-limitations of the equipment, insufficient magnification/limited image resolution-as reasons for disagreements. To summarize the pitfalls of digital pathology practice and better address the root cause of the diagnostic discordance, we suggest a Categorization for Digital Pathology Discrepancies to be used in further validations studies. Among 99 discordances, only 37 (37.3%) had preferred diagnosis rendered by means of WSI. The risk of bias and applicability concerns were judged with the QUADAS-2. Two studies (15.4%) presented an unclear risk of bias in the sample selection domain and 2 (15.4%) presented a high risk of bias in the index test domain. Regarding applicability, all studies included were classified as a low concern in all domains. The included studies were optimally designed to validate WSI for general clinical use, providing evidence with confidence. In general, this systematic review showed a high concordance between diagnoses achieved by using WSI and conventional light microscope (CLM), summarizes difficulties related to specific findings of certain areas of pathology-including dermatopathology, pediatric pathology, neuropathology, and gastrointestinal pathology-and demonstrated that WSI can be used to render primary diagnoses in several subspecialties of human pathology.

Integrity signals

Registry

Crossref · 10.1007/s00428-018-02519-z

Cites

72 of 72 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Araújo, A. L. D., Arboleda, L. P. A., Palmier, N. R., Fonsêca, J. M., de Pauli Paglioni, M., Gomes-Silva, W., Ribeiro, A. C. P., Brandão, T. B., Simonato, L. E., Speight, P. M., Fonseca, F. P., Lopes, M. A., de Almeida, O. P., Vargas, P. A., Madrid Troconis, C. C., & Santos-Silva, A. R.
Affiliations
  • Universidade Estadual de Campinas (UNICAMP) (BR)
  • Universidade de São Paulo (BR)
  • Universidade Brasil (BR)
  • University of Sheffield (GB)
  • Universidade Federal de Minas Gerais (BR)
Venue
Virchows Archiv
Published
2019-01-26
Volume
474
Issue
3
Pages
269-287
ISSN
  • 0945-6317 (ISSN-L)
  • 1432-2307
OpenAlex
W2914038321
Cited by
64
References
72
Citations by year
2026: 3 · 2025: 3 · 2024: 8 · 2023: 6 · 2022: 11 · 2021: 18 · 2020: 11 · 2019: 4
FWCI
0.72
Type
journal article
Language
en
Open access
closed
Locations
  • Archiv für Pathologische Anatomie und Physiologie und für Klinische Medicin (publishedVersion)
  • PubMed (publishedVersion)
Funders
  • Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico (306699/2017-6)
  • Funda??o de Amparo ? Pesquisa do Estado de S?o Paulo (2009/53839-2)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
  • Health Sciences › Medicine › Physiology › Clinical Laboratory Practices and Quality Control
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
Keywords
whole slide imaging, digital pathology, intraobserver agreement, diagnostic discordance, primary diagnosis, dermatopathology, gastrointestinal pathology, neuropathology, image resolution, biopsy quality, diagnostic pitfalls, immunohistochemistry limitations, special stains, brain biopsy
MeSH terms
Biopsy, Humans, Image Interpretation, Computer-Assisted, Microscopy, Pathology, Predictive Value of Tests, Reproducibility of Results, Observer Variation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review