Active Learning to Characterize the Full Contrast Sensitivity Function in Cataracts

To characterize contrast sensitivity function (CSF) in cataractous and pseudophakic eyes compared to healthy control eyes using a novel quantitative CSF test with active learning algorithms.

This is a prospective observational study at an academic medical center. CSF was measured in eyes with visually significant cataract, at least 2+ nuclear sclerosis (NS) and visual acuity (VA) ≥ 20/50, in pseudophakic eyes and in healthy controls with no more than 1+ NS and no visual complaints, using the Manifold Contrast Vision Meter. Outcomes included Area under the Log CSF (AULCSF) and CS thresholds at 1, 1.5, 3, 6, 12, and 18 cycles per degree (cpd). A subgroup analysis as performed on cataract eyes with VA ≥ 20/25.

A total of 167 eyes were included, 58 eyes in the cataract group, 77 controls, and 32 pseudophakic eyes with respective median AULCSF of 1.053 (0.352) vs 1.228 (0.318) vs 1.256 (0.360). In our multivariate regression model, cataract was associated with significantly reduced AULCSF (P= 0.04, β= -0.11) and contrast threshold at 6 cpd (P= 0.01, β= -0.16) compared to controls. Contrast threshold at 6 cpd was significantly reduced even in the subgroup of cataractous eyes with VA ≥ 20/25 (P=0.02, β=-0.16).

The novel qCSF test detected disproportionate significant contrast deficits at 6 cpd in cataract eyes; this remained significant even in the cataractous eyes with VA ≥ 20/25. CSF testing may enhance cataract evaluation and surgical decision-making, particularly in patients with subjective visual complaints despite good VA.