EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Classification of Oral Cancer Through Artificial Intelligence-Based Histopathological Image Analysis: A Comprehensive Review of Diagnostic Methods and Clinical Challenges

Dilovan Asaad Zebari, Nechirvan Asaad Zebari

Qubahan Journal of Medical Sciences · 08.11.2026 · 3 · 35-55

DOI: 10.48161/qjms.v3a90 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Qubahan Journal of Medical Sciences, online erschienen am 08.11.2026. Verfasst von Dilovan Asaad Zebari und 1 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Artificial intelligence (AI) is increasingly applied to oral squamous cell carcinoma (OSCC) detection and oral epithelial dysplasia (OED) grading in digital histopathology.

Objective

To characterize diagnostic tasks, model families, reported performance, and evidence relevant to clinical generalizability.

Methods

We developed a structured, task-stratified evidence map from 39 topic-specific bibliographic sources, supplemented by two methodological guidance references and three cross-modal methodological examples. Sources were classified by evidence role, and published performance estimates were interpreted according to endpoint and validation setting. Comprehensive multi-database retrieval records, independent eligibility screening, and completed study-level risk-of-bias assessments were not available; the synthesis is therefore provisional.

Results

The topical inventory comprised 27 candidate classification or grading studies, three adjacent original investigations, five secondary reviews, two dataset sources, and two clinical or pathology background publications. Several classifiers reported high internal accuracy, but differences in case mix, diagnostic labels, prediction units, and independent testing precluded a valid cross-study ranking. No meta-analysis was undertaken because comparable patient-level diagnostic contingency data were not consistently available.

Conclusion

The principal barrier to clinical translation is not model discrimination alone but reproducible patient-independent evaluation, dependable reference diagnoses, external validation, and demonstrated benefit in clinical workflows. The review defines evidence requirements for more reliable future investigations.

Zitieren

APA

Zebari, D. A., & Zebari, N. A. (2026). Classification of Oral Cancer Through Artificial Intelligence-Based Histopathological Image Analysis: A Comprehensive Review of Diagnostic Methods and Clinical Challenges. Qubahan Journal of Medical Sciences, 3, 35-55. https://doi.org/10.48161/qjms.v3a90

Vancouver

Zebari DA, Zebari NA. Classification of Oral Cancer Through Artificial Intelligence-Based Histopathological Image Analysis: A Comprehensive Review of Diagnostic Methods and Clinical Challenges. Qubahan Journal of Medical Sciences. 2026;3:35-55. doi:10.48161/qjms.v3a90

BibTeX
@article{Zebari2026,
  author = {Zebari, Dilovan Asaad and Zebari, Nechirvan Asaad},
  title = {Classification of Oral Cancer Through Artificial Intelligence-Based Histopathological Image Analysis: A Comprehensive Review of Diagnostic Methods and Clinical Challenges},
  journal = {Qubahan Journal of Medical Sciences},
  year = {2026},
  volume = {3},
  pages = {35--55},
  doi = {10.48161/qjms.v3a90}
}
Kostenlos starten →