Eyelid Tumor Images for Telediagnosis Comparison, 2016-2017

From January 2016 to July 2017, 36 patients with eyelid lesions were photographed using a conventional digital camera and a smartphone camera. The dataset, associated with a research paper by Lísia Aoki, compares the diagnostic accuracy of these image types for telediagnosis against a histological gold standard. It includes diagnoses for lesions such as basal cell carcinoma, actinic keratosis, and nevus.

Use Cases

• Benchmarking smartphone vs. conventional camera image quality for medical diagnosis based on the described comparison study.
• Training or validating image classification models for eyelid tumor types like basal cell carcinoma, actinic keratosis, and nevus.
• Studying inter-rater agreement in telediagnosis based on the reported kappa coefficients for different image sources.

Strengths

• Includes a histological diagnosis gold standard for 36 lesions.
• Provides specific accuracy metrics for in-person (94.4%) and remote evaluations (80.6% to 86.1%).
• Reports kappa coefficients for inter-rater agreement on malignant lesion diagnosis (0.68 for conventional, 0.78 for smartphone images).

Limitations

• Row count and column-level documentation are unknown, which limits suitability assessment.
• Data scale is small with 36 patients, which may limit statistical power for model training.
• Last update date is unknown; freshness unverified.

Provenance

Source

paperswithcode

Collection Method

Clinical study involving patient examinations, imaging, and remote analysis by specialists.

Time Range

January 2016 to July 2017