Ensemble Deep-Learning-Based Prognostic and Prediction for Recurrence of Sporadic Odontogenic Keratocysts on Hematoxylin and Eosin Stained Pathological Images of Incisional Biopsies.
Source check: Source checked. Not independent scientific confirmation.
Finding: Full MDPI text reviewed in Chrome: an ensemble of DenseNet-121, Inception-V3 and Inception-ResNet-V3 classifies recurrence of sporadic odontogenic keratocysts from histology images. The 60-patient cohort supplied about 1,660 annotated images; the 70/15/15 split and held-out results report up to 97% accuracy (abstract also reports ensemble AUC 0.9872). Approve narrowly as a small-cohort image-classification result; possible image-level leakage, no external validation and no prospective utility remain limitations.
Technology: learned CNN classifiers (DenseNet-121, Inception-V3, Inception-ResNet-V3) operate on H&E incisional-biopsy images; ensemble operation links histology to recurrence classification.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- learned CNN classifiers (DenseNet-121, Inception-V3, Inception-ResNet-V3) operate on H&E incisional-biopsy images; ensemble operation links histology to recurrence classification.
- Test and result
- recurrent/non-recurrent histologic features were distinguished; ensemble reports 97% accuracy, with source abstract reporting AUC 0.9872. Study has only 60 patients and about 1,660 images.
- Limitation / still unknown
- Methods/Results: 70/15/15 train/validation/test split within 60 patients; no independent cohort. Image-level partitioning may overstate generalization if images from a patient cross splits; interpret as preliminary.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.mdpi.com/2075-4426/12/8/1220 · https://europepmc.org/article/MED/35893314 · https://pubmed.ncbi.nlm.nih.gov/35893314/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.