A Joint Multitask Learning Model for Cross-sectional and Longitudinal Predictions of Visual Field Using OCT

Ryo Asaoka; Linchuan Xu; Hiroshi Murata; Taichi Kiwaki; Masato Matsuura; Yuri Fujino; Masaki Tanito; Kazuhiko Mori; Yoko Ikeda; Takashi Kanamoto; Kenji Inoue; Jukichi Yamagami; Kenji Yamanishi

doi:10.1016/j.xops.2021.100055

A Joint Multitask Learning Model for Cross-sectional and Longitudinal Predictions of Visual Field Using OCT

Ophthalmol Sci. 2021 Sep 7;1(4):100055. doi: 10.1016/j.xops.2021.100055. eCollection 2021 Dec.

Authors

Ryo Asaoka^{1

2

3

4

5}, Linchuan Xu^{6

7}, Hiroshi Murata⁵, Taichi Kiwaki⁶, Masato Matsuura^{5

8}, Yuri Fujino^{5

8

9}, Masaki Tanito⁹, Kazuhiko Mori¹⁰, Yoko Ikeda^{10

11}, Takashi Kanamoto^{12

13}, Kenji Inoue¹⁴, Jukichi Yamagami¹⁵, Kenji Yamanishi⁶

Affiliations

¹ Department of Ophthalmology, Seirei Hamamatsu General Hospital, Shizuoka, Hamamatsu, Japan.
² Seirei Christopher University, Shizuoka, Hamamatsu, Japan.
³ Nanovision Research Division, Research Institute of Electronics, Shizuoka University, Shizuoka, Japan.
⁴ The Graduate School for the Creation of New Photonics Industries, Shizuoka, Japan.
⁵ Department of Ophthalmology, Graduate School of Medicine and Faculty of Medicine, The University of Tokyo, Tokyo, Japan.
⁶ Graduate School of Information Science and Technology, The University of Tokyo, Tokyo, Japan.
⁷ Department of Computing, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong.
⁸ Department of Ophthalmology, Graduate School of Medical Sciences, Kitasato University, Kanagawa, Japan.
⁹ Department of Ophthalmology, Shimane University Faculty of Medicine, Izumo, Japan.
¹⁰ Department of Ophthalmology, Kyoto Prefectural University of Medicine, Kyoto, Japan.
¹¹ Oike-Ikeda Eye Clinic, Kyoto, Japan.
¹² Department of Ophthalmology, Hiroshima Memorial Hospital, Hiroshima, Japan.
¹³ Department of Ophthalmology, Hiroshima Prefectural Hospital, Hiroshima, Japan.
¹⁴ Inouye Eye Hospital, Tokyo, Japan.
¹⁵ Department of Ophthalmology, JR General Hospital, Tokyo, Japan.

Abstract

Purpose: We constructed a multitask learning model (latent space linear regression and deep learning [LSLR-DL]) in which the 2 tasks of cross-sectional predictions (using OCT) of visual field (VF; central 10°) and longitudinal progression predictions of VF (30°) were performed jointly via sharing the deep learning (DL) component such that information from both tasks was used in an auxiliary manner (The Association for Computing Machinery's Special Interest Group on Knowledge Discovery and Data Mining [SIGKDD] 2021). The purpose of the current study was to investigate the prediction accuracy preparing an independent validation dataset.

Design: Cohort study.

Participants: Cross-sectional training and testing data sets included the VF (Humphrey Field Analyzer [HFA] 10-2 test) and an OCT measurement (obtained within 6 months) from 591 eyes of 351 healthy people or patients with open-angle glaucoma (OAG) and from 155 eyes of 131 patients with OAG, respectively. Longitudinal training and testing data sets included 7984 VF results (HFA 24-2 test) from 998 eyes of 592 patients with OAG and 1184 VF results (HFA 24-2 test) from 148 eyes of 84 patients with OAG, respectively. Each eye had 8 VF test results (HFA 24-2 test). The OCT sequences within the observation period were used.

Methods: Root mean square error (RMSE) was used to evaluate the accuracy of LSLR-DL for the cross-sectional prediction of VF (HFA 10-2 test). For the longitudinal prediction, the final (eighth) VF test (HFA 24-2 test) was predicted using a shorter VF series and relevant OCT images, and the RMSE was calculated. For comparison, RMSE values were calculated by applying the DL component (cross-sectional prediction) and the ordinary pointwise linear regression (longitudinal prediction).

Main outcome measures: Root mean square error in the cross-sectional and longitudinal predictions.

Results: Using LSLR-DL, the mean RMSE in the cross-sectional prediction was 6.4 dB and was between 4.4 dB (VF tests 1 and 2) and 3.7 dB (VF tests 1-7) in the longitudinal prediction, indicating that LSLR-DL significantly outperformed other methods.

Conclusions: The results of this study indicate that LSLR-DL is useful for both the cross-sectional prediction of VF (HFA 10-2 test) and the longitudinal progression prediction of VF (HFA 24-2 test).

Keywords: CNN, convolutional neural network; CNN-TR, convolutional neural network and tensor regression; DL, deep learning; DLLR, deeply regularized latent space linear regression; GCL, ganglion cell layer; Glaucoma; HFA, Humphrey Field Analyzer; IPL, inner plexiform layer; LSLR-DL, latent space linear regression and deep learning; MLR, multiple linear regression; OAG, open-angle glaucoma; OCT; OS, outer segment; PLR, pointwise linear regression; Progression; RMSE, root mean square error; RNFL, retinal nerve fiber layer; RPE, retinal pigment epithelium; SVR, support vector regression; VF, visual field; Visual field; mTD, mean total deviation.