Shohei Morikawa

Shohei Morikawa

Auto-generated profile, based on publicly available information

Shohei Morikawa

Department of Ophthalmology, University of Tsukuba, Tsukuba 305-8575, Japan

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications

3 March 2022 in BMJ Open Ophthalmology
Verified
Journal Article
Open access
PDF file
SMShohei Morikawa
Shohei Morikawa
Fumiki OkamotoFumiki Okamoto
Fumiki Okamoto
TMTomoya Murakami
Tomoya Murakami
YSYoshimi Sugiura
Yoshimi Sugiura
THTakahiro Hiraoka
Takahiro Hiraoka
TOTetsuro Oshika
Tetsuro Oshika
Background To investigate the relationship between stereopsis and vision-related quality of life (VR-QOL) in patients with branch retinal vein occlusion (BRVO) before and after treatment with intravitreal ranibizumab (IVR). Methods This prospective multicentred observational study included 37 patients undergoing IVR treatment for unilateral BRVO and 24 age-matched healthy controls. Stereopsis was evaluated using the TNO stereo test (TNO) and Titmus stereo test (TST) every month, and the 25-item National Eye Institute Visual Function Questionnaire (VFQ-25) was administered at baseline, then at 3, 6 and 12 months after treatment. Results Time course of the changes in stereopsis and VR-QOL. IVR treatment significantly reduced central fovea thickness and improved both the TNO and the TST from 2 to 12 months (both p<0.05). Stereopsis before and after IVR injection in the eyes with BRVO were significantly worse than those in control subjects (TNO, p<0.001; TST, p<0.001). The VFQ-25 composite score significantly improved from 3 to 12 months after IVR treatment (p<0.05). Univariate analysis showed that the TNO score at baseline was significantly correlated with the VFQ-25 composite score at baseline and after treatment (p<0.05, p<0.05, respectively). TST score was not associated with the VFQ-25 composite score at baseline or after treatment. Conclusions Treatment with IVR for BRVO improved cystoid macular oedema, which was correlated to improved stereopsis, although not to the control level. The TNO score at baseline was associated with VR-QOL in patients with BRVO.

9 February 2024 in PLOS ONE
Verified
Journal Article
Open access
PDF file
SMShohei Morikawa
Shohei Morikawa
FOFumiki Okamoto
Fumiki Okamoto
TMTomoya Murakami
Tomoya Murakami
YSYoshimi Sugiura
Yoshimi Sugiura
TOTetsuro Oshika
Tetsuro Oshika
Purpose: To compare stereopsis and foveal microstructure after internal limiting membrane peeling and inverted internal limiting membrane flap technique in patients with macular hole. Design: Retrospective observational study. Methods: Sixty-six patients with macular hole were included, of whom 41 underwent 25-gauge pars-plana vitrectomy with complete internal limiting membrane peeling (Peeling group) and 25 with the inverted flap technique (Inverted group). We evaluated stereopsis using the Titmus Stereo Test and the TNO stereo test, best-corrected visual acuity, macular hole closure rate, and foveal microstructure with optical coherence tomography before and at 3, 6, and 12 months after surgery. Main outcome measures: Stereopsis and foveal microstructure. Results: Preoperatively, no difference was observed in the base and minimum diameters of macular hole, Titmus Stereo Test score, TNO stereo test score, and best-corrected visual acuity between the Peeling and Inverted groups. The macular hole closure rate in the Peeling and Inverted groups were 97.6% and 100%, respectively, with no significant difference between groups. At 12 months postoperatively, Titmus Stereo Test score (2.1 ± 0.4 in the peeling and 2.2 ± 0.4 in the inverted groups), TNO stereo test score (2.3 ± 0.4 and 2.2± 0.5), and best-corrected visual acuity (0.20 ± 0.18 and 0.24 ± 0.25) were not significantly different between groups (p = 0.596, 0.332, respectively). The defect of the external limiting membrane was more common in the Inverted group than in the Peeling group at 6 months after surgery (5.4 vs. 28.0%; p < 0.05). No statistically significant inter-group differences were noted in the ellipsoid zone defect ratio throughout the follow-up period. Conclusions: There was no difference in postoperative stereopsis nor foveal microstructure between the internal limiting membrane peeling group and the inverted group in patients with macular hole.

24 February 2025 in Journal of Clinical Medicine
Verified
Journal Article
Open access
PDF file
TMTomoya Murakami
Tomoya Murakami
Fumiki OkamotoFumiki Okamoto
Fumiki Okamoto
TMTakeshi Matsueda
Takeshi Matsueda
YSYoshimi Sugiura
Yoshimi Sugiura
SMShohei Morikawa
Shohei Morikawa
TOTetsuro Oshika
Tetsuro Oshika
Background/Objectives: We sought to assess the relationship between contrast sensitivity (CS) and optical coherence tomography (OCT) findings, including ganglion cell–inner plexiform layer (GCIPL) thickness, in eyes with cystoid macular edema, secondary to branch retinal vein occlusion (BRVO-CME), treated with intravitreal ranibizumab (IVR). Methods: This prospective study included 44 patients with BRVO-CME who underwent treatment with IVR (three monthly injections and pro re nata) and were followed up for 12 months. We collected data on CS, best-corrected visual acuity (BCVA), and OCT findings (ellipsoid zone [EZ] and external limiting membrane status [ELM], central foveal thickness [CFT], and average GCIPL thickness) at the time of the final visit when macular edema was resolved. Multiple regression analysis was used to evaluate the relationship between visual functions and OCT findings, age, and lens status. Results: Multiple regression analysis revealed that lower GCIPL thickness was significantly associated with worse CS (β = 0.008; 95% CI, 0.002–0.014; p = 0.011), whereas this was not the case with BCVA. Lower CFT and mild cataracts were also associated with worse CS (CFT: β = 0.003; 95% CI, 0.001–0.004; p = 0.001; mild cataract: β = −0.182; 95% CI, −0.286–−0.078; p = 0.001) and worse BCVA (CFT: β = −0.002; 95% CI, −0.003–−0.001; p < 0.001; mild cataract: β = 0.079; 95% CI, 0.008–0.150; p = 0.029). Conclusions: GCIPL thickness may serve as a valuable biomarker for CS in eyes with BRVO-CME following IVR treatment.

7 October 2025 in medRxiv
Verified
Preprint
Open access
PDF file
STShumpei Takinami
Shumpei Takinami
SMShohei Morikawa
Shohei Morikawa
TOTetsuro Oshika
Tetsuro Oshika
Purpose To develop and evaluate a novel self-supervised learning approach using Masked Autoencoder (MAE) pre-trained Vision Transformer (ViT) for automated detection of diabetic macular edema (DME) from optical coherence tomography (OCT) images, addressing the critical need for scalable screening solutions in diabetic eye care. Study Design Artificial intelligence model training. Methods We utilized the publicly available Kermany dataset containing 109,312 OCT images, defining DME detection as a binary classification task (11,559 DME vs. 97,753 non-DME images). Five deep learning architectures were compared: MAE-pretrained ViT (MAE_ViT), standard ViT, ResNet18, VGG19_bn, and EfficientNetV2. MAE_ViT underwent two-stage training: (1) self-supervised pre-training with 75% patch masking for 1,000 epochs to learn robust visual representations, and (2) supervised fine-tuning for DME classification. Model performance was evaluated using accuracy, sensitivity, specificity, F1 score, and area under the receiver operating characteristic curve (AU-ROC) with 95% confidence intervals calculated via bootstrap resampling. Results MAE_ViT achieved superior performance with AU-ROC 0.999 (95% CI: 0.999-1.000), accuracy 98.5% (95% CI: 97.7-99.2%), sensitivity 99.6% (95% CI: 98.7-100%), and specificity 98.1% (95% CI: 97.2-99.1%). VGG19_bn showed the second-best performance (AU-ROC 0.997), while ResNet18 demonstrated poor specificity (28.3%) despite perfect sensitivity. The self-supervised approach of MAE_ViT outperformed standard supervised ViT (AU-ROC 0.995), demonstrating the effectiveness of learning from unlabeled data. Conclusion MAE pre-trained Vision Transformer establishes a new benchmark for automated DME detection, offering exceptional diagnostic accuracy and potential for deployment in resource-constrained settings through reduced annotation requirements.

4 December 2025 in Sensors
Verified
Journal Article
Open access
PDF file
HTHidenori Takahashi
Hidenori Takahashi
TTTaiki Tsuge
Taiki Tsuge
YKYusuke Kondo
Yusuke Kondo
YYYasuo Yanagi
Yasuo Yanagi
SISatoru Inoda
Satoru Inoda
TYToshihiko Yamasaki
Toshihiko Yamasaki
The relationship between retinal images and intraocular cytokine profiles remains largely unexplored, and no prior work has systematically compared fundus- and OCT-based deep learning models for cytokine prediction. We aimed to predict intraocular cytokine concentrations using color fundus photographs (CFP) and retinal optical coherence tomography (OCT) with deep learning. Our pipeline consisted of image preprocessing, convolutional neural network–based feature extraction, and regression modeling for each cytokine. Deep learning was implemented using AutoGluon, which automatically explored multiple architectures and converged on ResNet18, reflecting the small dataset size. Four approaches were tested: (1) CFP alone, (2) CFP plus demographic/clinical features, (3) OCT alone, and (4) OCT plus these features. Prediction performance was defined as the mean coefficient of determination (R2) across 34 cytokines, and differences were evaluated using paired two-tailed t-tests. We used data from 139 patients (152 eyes) and 176 aqueous humor samples. The cohort consisted of 85 males (61%) with a mean age of 73 (SD 9.8). Diseases included 64 exudative age-related macular degeneration, 29 brolucizumab-associated endophthalmitis, 19 cataract surgeries, 15 retinal vein occlusion, and 8 diabetic macular edema. Prediction performance was generally poor, with mean R2 values below zero across all approaches. The CFP-only model (–0.19) outperformed CFP plus demographics (–24.1; p = 0.0373), and the OCT-only model (–0.18) outperformed OCT plus demographics (–14.7; p = 0.0080). No significant difference was observed between CFP and OCT (p = 0.9281). Notably, VEGF showed low predictability (31st with CFP, 12th with OCT).