Jessica Loo1, Traci E Clemons2, Emily Y Chew3, Martin Friedlander4, Glenn J Jaffe5, Sina Farsiu6. 1. Department of Biomedical Engineering, Duke University, Durham, North Carolina. 2. Ophthalmology, Emmes, Rockville, Maryland. 3. Division of Epidemiology and Clinical Applications, National Eye Institute, Bethesda, Maryland. 4. Department of Molecular Medicine, The Scripps Research Institute, La Jolla, California; The Lowy Medical Research Institute, La Jolla, California. 5. Department of Ophthalmology, Duke University Medical Center, Durham, North Carolina. 6. Department of Biomedical Engineering, Duke University, Durham, North Carolina; Department of Ophthalmology, Duke University Medical Center, Durham, North Carolina. Electronic address: sina.farsiu@duke.edu.
Abstract
PURPOSE: To validate the efficacy of a fully automatic, deep learning-based segmentation algorithm beyond conventional performance metrics by measuring the primary outcome of a clinical trial for macular telangiectasia type 2 (MacTel2). DESIGN: Evaluation of diagnostic test or technology. PARTICIPANTS: A total of 92 eyes from 62 participants with MacTel2 from a phase 2 clinical trial (NCT01949324) randomized to 1 of 2 treatment groups METHODS: The ellipsoid zone (EZ) defect areas were measured on spectral domain OCT images of each eye at 2 time points (baseline and month 24) by a fully automatic, deep learning-based segmentation algorithm. The change in EZ defect area from baseline to month 24 was calculated and analyzed according to the clinical trial protocol. MAIN OUTCOME MEASURE: Difference in the change in EZ defect area from baseline to month 24 between the 2 treatment groups. RESULTS: The difference in the change in EZ defect area from baseline to month 24 between the 2 treatment groups measured by the fully automatic segmentation algorithm was 0.072±0.035 mm2 (P = 0.021). This was comparable to the outcome of the clinical trial using semiautomatic measurements by expert readers, 0.065±0.033 mm2 (P = 0.025). CONCLUSIONS: The fully automatic segmentation algorithm was as accurate as semiautomatic expert segmentation to assess EZ defect areas and was able to reliably reproduce the statistically significant primary outcome measure of the clinical trial. This approach, to validate the performance of an automatic segmentation algorithm on the primary clinical trial end point, provides a robust gauge of its clinical applicability.
RCT Entities:
PURPOSE: To validate the efficacy of a fully automatic, deep learning-based segmentation algorithm beyond conventional performance metrics by measuring the primary outcome of a clinical trial for macular telangiectasia type 2 (MacTel2). DESIGN: Evaluation of diagnostic test or technology. PARTICIPANTS: A total of 92 eyes from 62 participants with MacTel2 from a phase 2 clinical trial (NCT01949324) randomized to 1 of 2 treatment groups METHODS: The ellipsoid zone (EZ) defect areas were measured on spectral domain OCT images of each eye at 2 time points (baseline and month 24) by a fully automatic, deep learning-based segmentation algorithm. The change in EZ defect area from baseline to month 24 was calculated and analyzed according to the clinical trial protocol. MAIN OUTCOME MEASURE: Difference in the change in EZ defect area from baseline to month 24 between the 2 treatment groups. RESULTS: The difference in the change in EZ defect area from baseline to month 24 between the 2 treatment groups measured by the fully automatic segmentation algorithm was 0.072±0.035 mm2 (P = 0.021). This was comparable to the outcome of the clinical trial using semiautomatic measurements by expert readers, 0.065±0.033 mm2 (P = 0.025). CONCLUSIONS: The fully automatic segmentation algorithm was as accurate as semiautomatic expert segmentation to assess EZ defect areas and was able to reliably reproduce the statistically significant primary outcome measure of the clinical trial. This approach, to validate the performance of an automatic segmentation algorithm on the primary clinical trial end point, provides a robust gauge of its clinical applicability.
Authors: Stephanie J Chiu; Joseph A Izatt; Rachelle V O'Connell; Katrina P Winter; Cynthia A Toth; Sina Farsiu Journal: Invest Ophthalmol Vis Sci Date: 2012-01-05 Impact factor: 4.799
Authors: Sina Farsiu; Stephanie J Chiu; Rachelle V O'Connell; Francisco A Folgar; Eric Yuan; Joseph A Izatt; Cynthia A Toth Journal: Ophthalmology Date: 2013-08-29 Impact factor: 12.079
Authors: Zhichao Wu; David Cunefare; Elizabeth Chiu; Chi D Luu; Lauren N Ayton; Cynthia A Toth; Sina Farsiu; Robyn H Guymer Journal: Invest Ophthalmol Vis Sci Date: 2016-07-01 Impact factor: 4.799
Authors: Thomas Schlegl; Sebastian M Waldstein; Hrvoje Bogunovic; Franz Endstraßer; Amir Sadeghipour; Ana-Maria Philip; Dominika Podkowinski; Bianca S Gerendas; Georg Langs; Ursula Schmidt-Erfurth Journal: Ophthalmology Date: 2017-12-08 Impact factor: 12.079
Authors: Emily Y Chew; Traci E Clemons; Tunde Peto; Ferenc B Sallo; Avner Ingerman; Weng Tao; Lawrence Singerman; Steven D Schwartz; Neal S Peachey; Alan C Bird Journal: Am J Ophthalmol Date: 2014-12-19 Impact factor: 5.258
Authors: Peter Charbel Issa; Mark C Gillies; Emily Y Chew; Alan C Bird; Tjebo F C Heeren; Tunde Peto; Frank G Holz; Hendrik P N Scholl Journal: Prog Retin Eye Res Date: 2012-12-03 Impact factor: 21.198
Authors: Jessica Loo; Matthias F Kriegel; Megan M Tuohy; Kyeong Hwan Kim; Venkatesh Prajna; Maria A Woodward; Sina Farsiu Journal: IEEE J Biomed Health Inform Date: 2021-01-05 Impact factor: 5.772
Authors: Tisileli S Tuifua; Joseph R Abraham; Sunil K Srivastava; Peter K Kaiser; Jamie Reese; Justis P Ehlers Journal: Retina Date: 2022-02-01 Impact factor: 4.256