June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
Analysis of numerical feature extraction from automated geographic atrophy segmentation
Author Affiliations & Notes
  • Theodore Spaide
    Genentech, Inc., South San Francisco, California, United States
  • Jasmine Patil
    Genentech, Inc., South San Francisco, California, United States
  • Jiaxiang Jiang
    Genentech, Inc., South San Francisco, California, United States
    University of California, Santa Barbara, Santa Barbara, California, United States
  • Neha Anegondi
    Genentech, Inc., South San Francisco, California, United States
  • Michael Kawczynski
    Genentech, Inc., South San Francisco, California, United States
  • Verena Steffen
    Genentech, Inc., South San Francisco, California, United States
  • Simon S. Gao
    Genentech, Inc., South San Francisco, California, United States
  • Footnotes
    Commercial Relationships   Theodore Spaide, Genentech, Inc. (E); Jasmine Patil, Genentech, Inc. (E); Jiaxiang Jiang, Genentech, Inc. (F); Neha Anegondi, Genentech, Inc. (E); Michael Kawczynski, Genentech, Inc. (E); Verena Steffen, Genentech, Inc. (E); Simon Gao, Genentech, Inc. (E)
  • Footnotes
    Support  Genentech, Inc., South San Francisco, CA, provided support for the study and participated in the study design; conducting the study; and data collection, management, and interpretation
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 2124. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Theodore Spaide, Jasmine Patil, Jiaxiang Jiang, Neha Anegondi, Michael Kawczynski, Verena Steffen, Simon S. Gao; Analysis of numerical feature extraction from automated geographic atrophy segmentation. Invest. Ophthalmol. Vis. Sci. 2021;62(8):2124.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : To examine deep learning-based segmentation of geographic atrophy (GA) lesions and the reliability of derived features.

Methods : Nine hundred and forty pairs of images were taken from 194 patients, with each pair comprising a fundus autofluorescence (FAF) and near-infrared (NIR) image from 1 eye (Proxima B, NCT02399072). Lesions were annotated on the FAF by a grader, and the data were split at the patient level into training (n=155) and validation (n=39) sets. A test set comprising 90 FAF-NIR pairs from 90 patients (Proxima A, NCT02479386) was annotated by 2 graders (G1 and G2). Two multimodal deep learning networks (UNet and YNet) were trained on the training set and tuned on the validation set. The final network was applied to the test set. For each segmentation mask, the lesion area, perimeter, circularity, Feret diameters, and number of lesions were extracted. As a numerical proxy for the FAF pattern, the excess rim intensity (ERI), equal to the mean FAF intensity in a 0.5-mm rim around the lesion minus the mean FAF intensity in a 0.5- to 1-mm rim around the lesion, was also extracted. For all measures except for number of lesions, the relevant metric was computed for the whole segmented area without separating it into different components.

Results : The average Dice score between the network and G1 on the test set was 0.92. The Pearson correlation (r) of area, perimeter, circularity, major Feret diameters, minor Feret diameters, ERI, and number of lesions between the YNet and G1 was 0.98, 0.93, 0.86, 0.87, 0.93, 1.00, and 0.46, respectively. Analogous statistics for the network and G2 and for G1 and G2 are given in Table 1.

Conclusions : Networks trained to segment GA lesions could produce accurate segmentations. When the segmentations were used to obtain the values of area and ERI, the agreement between the networks and human graders was similar to the agreement between two graders. Inferred values of perimeter, circularity, and Feret diameters were less similar, and often varied between models despite similar Dice scores. The inferred number of lesions matched human grading poorly. The variable accuracy of the examined features could be an important factor for their use in predictive models of GA growth.

This is a 2021 ARVO Annual Meeting abstract.

 

Table 1. Comparison of different networks and graders on the test set

Table 1. Comparison of different networks and graders on the test set

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×