EUVIMEDEUVIMED entdecken
Crossref

Archetypal analysis for visual field loss characterization of leber hereditary optic neuropathy

Catarina Coutinho, Ferdinando Zanchetta, Michele Carbonelli, Alice Galzignato, Marco Battista, Federico Fantaguzzi, Giulia Amore, Valerio Carelli, Luigi Brotto, Paolo Nucci, Lisa Checchin, Giacomo Savini und 5 weitere

Acta Ophthalmologica · 2025-01 · 103 (S284)

DOI: 10.1111/aos.17183 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Acta Ophthalmologica, online erschienen am 19.01.2025. Verfasst von Catarina Coutinho und 16 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Aims/Purpose: Characterization of Visual Field (VF) loss patterns of chronic stage Leber Hereditary Optic Neuropathy (LHON) patients by application of machine learning (ML).

Methods

For chronic stable LHON patient, at least 2 years after disease onset, 24‐2 and 30‐2 VFs tests performed by SITA standard Humphrey VF analyzer (Carl Zeiss Meditec) were collected. Two groups were considered: with visual acuity (VA) recovery, when a change of more than 0.2 decimal or from off‐chart to on‐chart vision after initial loss of vision; and without VA recovery. Resorting to Python, archetypal analysis (AA) was implemented to each group using the VFs total deviations (TDs). Spearman correlations were performed between each archetype (AT) relative weight and the VFs mean deviation (MD).

Results

From 117 patients (203 eyes), 104 VFs made up the Recovered Group, and 115 VFs the Not Recovered Group. Both AA models were made up of 7 ATs. For the Recovered Group, the normal VF AT had the highest weight (33.40%), followed by different severity central scotoma ATs (AT2, 16.32%; AT3, 11.82%; AT5, 10.37%). For the Non‐Recovered Group, first the widespread loss AT (47.87%), followed by enlarged central scotoma ATs (AT2, 11.13%; AT5, 8.97%). For the Recovered Group, AT3 (Cecocentral Scotoma), AT4 (Fenestration) and AT6 (Widespread loss) weights decreased with increasing MD; and AT1 (normal VF) and AT5 (Central Scotoma) increased with increasing MD. The AT1 (Widespread Loss) weight of the Non‐Recovered Group decreased with increasing MD, while AT5 (Enlarged Central Scotoma) weight increased with increasing MD.

Conclusions

Implementation of AA‐ML allowed the association of more severe and widespread visual loss patterns to LHON patients without visual recovery, whereas less severe and more centralized patterns to recovered patients, as expected. Therefore, the presented VF analysis might present as a standardized method for LHON visual loss characterization, providing further characterization of patients with and without vision recovery.

Zitieren

APA

Coutinho, C., Zanchetta, F., Carbonelli, M., Galzignato, A., Battista, M., Fantaguzzi, F., Amore, G., Carelli, V., Brotto, L., Nucci, P., Checchin, L., Savini, G., Bandello, F., Morgia, C. L., Cascavilla, M. L., Fioresi, R., & Barboni, P. (2025). Archetypal analysis for visual field loss characterization of leber hereditary optic neuropathy. Acta Ophthalmologica, 103(S284). https://doi.org/10.1111/aos.17183

Vancouver

Coutinho C, Zanchetta F, Carbonelli M, Galzignato A, Battista M, Fantaguzzi F, et al. Archetypal analysis for visual field loss characterization of leber hereditary optic neuropathy. Acta Ophthalmologica. 2025;103(S284). doi:10.1111/aos.17183

BibTeX
@article{Coutinho2025,
  author = {Coutinho, Catarina and Zanchetta, Ferdinando and Carbonelli, Michele and Galzignato, Alice and Battista, Marco and Fantaguzzi, Federico and Amore, Giulia and Carelli, Valerio and Brotto, Luigi and Nucci, Paolo and Checchin, Lisa and Savini, Giacomo and Bandello, Francesco and Morgia, Chiara La and Cascavilla, Maria Lucia and Fioresi, Rita and Barboni, Piero},
  title = {Archetypal analysis for visual field loss characterization of leber hereditary optic neuropathy},
  journal = {Acta Ophthalmologica},
  year = {2025},
  volume = {103},
  number = {S284},
  doi = {10.1111/aos.17183}
}
Kostenlos starten →