A Multi-Parameter Corneal Tomography Dataset for Machine Learning–Driven IOL Selection and Corneal Subtype Classification

This dataset contains Pentacam-derived corneal tomographic and biometric parameters collected from 61 eyes of patients undergoing preoperative assessment for cataract surgery with intraocular lens (IOL) implantation. The data were acquired using a standardized Scheimpflug tomography protocol with the Pentacam HR system.

Each record represents a single eye and includes 35 variables encompassing anterior and posterior corneal curvature (K1, K2, KMAX), astigmatism magnitude and axis, corneal asphericity (Q-value), Total Corneal Refractive Power (TCRP), higher-order aberrations (HOA), spherical aberration (Z4,0), the Belin/Ambrósio Deviation index (BAD-D), wavefront root-mean-square (RMS) metrics, central corneal thickness (CCT), anterior chamber parameters, white-to-white distance (HWTW), corneal volume, axial length, refractive error classification, pupil diameter, and selected IOL type.

Eyes are clinically classified into six corneal subtypes:

Normal
High spherical aberration
Fuchs endothelial corneal dystrophy
High higher-order aberrations
Combined high HOA and high SA
High HOA with corneal ectasia

The dataset is provided as a structured Excel file (.xlsx) with one row per eye and one column per parameter. A companion parameter glossary document (.docx) defines all variables, units, and normative reference ranges.