Svjetlost tests the reliability of advanced vision diagnostics across more than 30,000 scans

A new paper by our team in the international journal Translational Vision Science & Technology.

Copied

A scientific paper by Ivan Gabrić, Ivan Samuel Modrić and Filip Grulović, based on a registry of 2,533 eyes, shows how important standardised image acquisition and quality control are for a reliable assessment of the eye’s optics.

At Svjetlost Eye Clinic, the large number of examinations and the systematic follow-up of patients also form the basis for scientific research. A new paper by our team, published in October 2026 in the international journal Translational Vision Science & Technology, presents a comprehensive analysis of the repeatability of aberrometry measurements – advanced diagnostics used to assess the optical irregularities of the eye. The authors of the study are Ivan Gabrić, Ivan Samuel Modrić and Filip Grulović.

The paper is based on a clinical registry comprising 2,533 eyes, 30,915 raw acquisitions and 10,416 measurement sessions. The data were collected over almost four years, from April 2022 to March 2026, during the pre-operative assessment and follow-up examinations of myopic patients before and after laser vision correction by lenticule extraction.

Quality of vision is not just a matter of prescription

The refractive prescription is important, but it is not the only parameter that describes the optics of the eye. Aberrometry makes it possible to measure finer optical irregularities, including so-called higher-order aberrations. Their analysis plays an important role in planning refractive procedures and in understanding visual disturbances after surgery.

For such diagnostics to be useful, it is necessary to know how repeatable their results are. If two measurements of the same eye show a difference, it is important to understand whether it represents a real change or the expected variability of the measurement process itself. This was one of the central questions of the Svjetlost study.

A large clinical registry, a detailed analysis

What sets this paper apart is its analysis of data from everyday clinical practice. Rather than relying only on a small, specially recruited group of participants, the researchers examined thousands of measurement sessions, including pre-operative measurements and those taken during post-operative follow-up.

This approach also captures the real-world challenges of diagnostics: varying pupil size, the quality of individual acquisitions and the completeness of the data exported by the device. Measurements were performed according to a standardised protocol, in a darkened room and without pharmacological pupil dilation. Four specially trained optometrists took part in data collection.

The size of the registry reflects the large number of diagnostic measurements and the systematic patient follow-up at Svjetlost. Its scientific value, however, lies not only in the amount of data, but in the detailed verification of its quality and reliability.

Precision is not just a property of the device

The study showed excellent repeatability of aberrometry measurements with the OSIRIS device when the described acquisition protocol and quality control were applied. For total higher-order aberrations, the coefficient of repeatability was 0.033 micrometres for analysis over a 4 mm diameter and 0.050 micrometres for analysis over a 6 mm diameter, when pupil size allowed such a measurement. These values define the expected range of differences between two measurements of the same eye within the same session.

A particularly important finding concerns how the data are handled. Without capture selection and the exclusion of invalid zero-value records, the coefficient of repeatability for total higher-order aberrations at 6 mm would increase approximately fivefold – meaning considerably greater variability in the results.

The message is clear: advanced technology delivers the best results when it is supported by a consistent acquisition procedure, a trained team and validation of every result.

Why it matters to recognise invalid data

The Svjetlost team also examined in detail how pupil size affects the availability of results. When the available measurement diameter did not reach 6 mm, the device exported only zero values for the 6 mm analysis in 96.8% of such sessions. Such a record does not mean that the eye has no optical irregularities, but that the corresponding data are not available.

The paper therefore defines rules for identifying such records and excluding them from analysis. This makes a practical contribution to the processing of aberrometry data: it prevents a missing result from being misinterpreted as a normal one.

Different measurements provide different information

The study also compared whole-eye aberrations measured with OSIRIS with total corneal aberrations calculated with the MS-39 device. The results were correlated but not interchangeable. This reflects the fact that different parts of the eye’s optical system are being analysed, not that one of the devices is necessarily inaccurate.

For clinical practice, this means that each result should be interpreted according to what it measures and the decision it will inform. Detailed diagnostics is not just about collecting results, but about understanding their meaning, relationships and limitations.

Knowledge shared with the international scientific community

Alongside the publication, the authors have made the anonymised analysis data, results tables and code publicly available. This allows other researchers to verify and reproduce the analyses and to make further scientific use of the data. The device manufacturer had no role in the study design, data analysis or the decision to publish.

For Svjetlost, this paper brings together everyday clinical experience and the scientific evaluation of its own procedures. Its central message is precision: it is not enough simply to measure – one must also understand how far the results can be trusted.