To investigate how optical coherence tomography (OCT) modifies the preset scan parameters to correct the errors resulting from ocular magnification, the influence of examiner's final correction of those already modified parameters on retinal nerve fiber layer (RNFL) thickness measurements, the induced change on RNFL thickness measurements and RNFL estimated integrals (RNFL(estimated integrals)) by adjusting the actual scan radius during RNFL examinations performed by OCT. Thirty-five healthy patients underwent an RNFL examination by OCT four times using different scan radii. The first scan was performed with the preset circular scan diameter of 3.46 mm; the actual scan diameter was different, however, because it was modified by the OCT instrument. The second, third, and fourth scans were generated after readjusting the already modified scan diameter by the examiner to 3.46, 3.20, and 3.60 mm. The relationship of axial length and refractive error with the actual scan radius (with ocular magnification calculated by OCT), with the influence of the examiner's final correction on RNFL thickness measurements, with the relationship between scan radius with RNFL thickness measurements, and with RNFL(estimated integrals) were investigated. The actual scan diameter was found to be primarily determined by axial length (R = 0.97, P < 0.0001), but the influence of refractive error was small (R = -0.26, P = 0.067). Final correction of the actual scan radius by the examiner had a significant influence on RNFL thickness measurements (P = 0.025). RNFL thickness measurements obtained without correction of the actual scan radius for magnification were found to be inversely correlated with axial length (R = -0.54, P = 0.001), whereas no similar relationship was found when RNFL thickness measurements were obtained with correction (R = 0.21, P = 0.11). A reciprocal relationship between 1/scan radius with RNFL thickness measurements (they tended to be thinner as scan radii were increased) was found (R = 0.41, P = 0.169), but RNFL(estimated integrals) areas were found to be independent of the scan radius (P = 0.521). To increase the accuracy of RNFL thickness measurements, it will be appropriate for the examiner to manually correct the actual scan parameters to the desired or preset ones after their automatic modification performed by the OCT instrument. Keeping the actual scan radius constant for repeated exams is also recommended because RNFL thickness measurements were found to depend on scan size. Alternatively, RNFL(estimated integrals) could be used because they were found to be independent of the scan size.
Read full abstract