Medizinische Universität Graz - Research portal
Selected Publication:
SHR Neuro Cancer Cardio Lipid Metab Microb
Strauss, RW; Scholz, F; Ulbig, MW; Kampik, A; Neubauer, AS.
Artifacts in optical coherence tomography (OCT) imaging of the retina.
Klin Monbl Augenheilkd. 2007; 224(1):47-51
Doi: 10.1055/s-2006-927274
Web of Science
PubMed
FullText
FullText_MUG
- Leading authors Med Uni Graz
- Strauß Rupert
- Altmetrics:
- Dimensions Citations:
- Plum Analytics:
- Scite (citation analytics):
- Abstract:
- Purpose: The purpose of this study was to investigate artifacts of OCT scans and of software analysis for retinal cross-section scans in a specialised retina clinic setting. Methods: A total of 205 vertical cross-hair OCT scans of retinal Stratus OCTs were randomly chosen from the database. All scans had been performed by one experienced technician. There were 75 (37%) normal findings, the remaining scans showed various types of retinal pathology: All scanning artifacts were analysed. Retinal thickness of all scans was measured automatically at the centre of the macula using two different software algorithms: the instrument's built-in "Stratus OCT Viewer V 4.01" and the stand-alone application "Datamedical OCTview V 3.5" (Datamedical Consulting, Hamburg). Errors of the software to correctly identify the retinal surface and the outer highly reflective layer were assigned into three categories: none, minor error (no influence on measurements) and major error. Results: A total of 7.3% of all OCT scans showed scanning artifacts: 5 motion artifacts, 9 scans with low signal intensity and I decentred scan. Scanning artifacts significantly increased the frequency of software errors (p = 0,012). The presence of retinal pathology also increased the number of errors (p = 0,004). Software analysis yielded a total of 20 major and 2 minor errors for the Stratus OCT (overall 10.7%) and 32 major and 64 minor errors for the Datamedical Viewer (p < 0.001). Measurements by Datamedical OCT view were a mean of 57 micron higher due to a different definition of the outer retinal border. Retinal pathologies significantly increased the likelihood of software errors for both algorithms (both p < 0.01), most critical were macular holes and changes in age-related macular degeneration. Conclusion: Scanning artifacts were associated with a significantly higher frequency of software errors. As artifacts of scans and software occur frequently, the interpretation of OCT scans requires special attention to artifacts.
- Find related publications in this database (using NLM MeSH Indexing)
- Artifacts -
- Female -
- Humans -
- Image Enhancement - methods
- Image Interpretation, Computer-Assisted - methods
- Male -
- Middle Aged -
- Reproducibility of Results -
- Retina - cytology
- Retrospective Studies -
- Sensitivity and Specificity -
- Software -
- Software Validation -
- Tomography, Optical Coherence - methods
- Find related publications in this database (Keywords)
- optical coherence tomography (OCT)
- artifact
- error
- image processing