Spectral-domain versus time domain optical coherence tomography before and after ranibizumab for age-related macular degeneration

Artificial intelligence in retinal screening using OCT images: A review of the last decade (2013-2023)

BACKGROUND AND OBJECTIVES

Optical coherence tomography (OCT) has ushered in a transformative era in the domain of ophthalmology, offering non-invasive imaging with high resolution for ocular disease detection. OCT, which is frequently used in diagnosing fundamental ocular pathologies, such as glaucoma and age-related macular degeneration (AMD), plays an important role in the widespread adoption of this technology. Apart from glaucoma and AMD, we will also investigate pertinent pathologies, such as epiretinal membrane (ERM), macular hole (MH), macular dystrophy (MD), vitreomacular traction (VMT), diabetic maculopathy (DMP), cystoid macular edema (CME), central serous chorioretinopathy (CSC), diabetic macular edema (DME), diabetic retinopathy (DR), drusen, glaucomatous optic neuropathy (GON), neovascular AMD (nAMD), myopia macular degeneration (MMD) and choroidal neovascularization (CNV) diseases. This comprehensive review examines the role that OCT-derived images play in detecting, characterizing, and monitoring eye diseases.

METHOD

The 2020 PRISMA guideline was used to structure a systematic review of research on various eye conditions using machine learning (ML) or deep learning (DL) techniques. A thorough search across IEEE, PubMed, Web of Science, and Scopus databases yielded 1787 publications, of which 1136 remained after removing duplicates. Subsequent exclusion of conference papers, review papers, and non-open-access articles reduced the selection to 511 articles. Further scrutiny led to the exclusion of 435 more articles due to lower-quality indexing or irrelevance, resulting in 76 journal articles for the review.

RESULTS

During our investigation, we found that a major challenge for ML-based decision support is the abundance of features and the determination of their significance. In contrast, DL-based decision support is characterized by a plug-and-play nature rather than relying on a trial-and-error approach. Furthermore, we observed that pre-trained networks are practical and especially useful when working on complex images such as OCT. Consequently, pre-trained deep networks were frequently utilized for classification tasks. Currently, medical decision support aims to reduce the workload of ophthalmologists and retina specialists during routine tasks. In the future, it might be possible to create continuous learning systems that can predict ocular pathologies by identifying subtle changes in OCT images.

Retinal Nerve Fiber Layer Alterations After Photocoagulation: A Prospective Spectral-Domain OCT Study

PURPOSE

To evaluate the effect of panretinal photocoagulation (PRP) on retinal nerve fiber layer thickness (RNFLT) in treatment-naive patients with proliferative diabetic retinopathy (PDR).

METHODS

Fifty eight previously untreated eyes of 30 patients with PDR who underwent PRP treatment were enrolled prospectively. All patients had at least six months of follow-up. Detailed ophthalmologic examinations including macular thickness and RNFLT assessments with spectral-domain type optic coherence tomography were performed at baseline as well as the third and sixth posttreatment months. Initial RNFLT and macular thickness of laser administered patients were compared with two separate control groups that were consisted of either nondiabetic patients or diabetics without PDR.

RESULTS

The mean age of study patients was 52.4±7.1 years (Range, 32-66 years) and 16 of them (53.3%) were female. At the sixth post-PRP month, visual stabilization or improvement was achieved in 54 eyes (93.1%). No significant difference was demonstrated in initial RNFLT measurements between the study patients and two control groups (p=0.478). Mean RNFLT was measured as 108.5±17.5µm, 115.8±17.6µm, and 103.0±16.4µm at baseline, third and sixth months of the follow-up, respectively. Although RNFLT increase noted at the third post-laser month was statistically significant compared to its baseline values (p<0.001), there was a significant reduction in RNFLT at the sixth post-laser month compared to its baseline values (p<0.001).

CONCLUSION

RNFLT increase in the third month of follow-up may be related to ensuing axonal edema. Significant RNLFT decrease at the sixth month of follow-up may be attributed to axonal loss secondary to the laser treatment.

Guidelines for the management of neovascular age-related macular degeneration by the European Society of Retina Specialists (EURETINA)

Age-related macular degeneration (AMD) is still referred to as the leading cause of severe and irreversible visual loss world-wide. The disease has a profound effect on quality of life of affected individuals and represents a major socioeconomic challenge for societies due to the exponential increase in life expectancy and environmental risks. Advances in medical research have identified vascular endothelial growth factor (VEGF) as an important pathophysiological player in neovascular AMD and intraocular inhibition of VEGF as one of the most efficient therapies in medicine. The wide introduction of anti-VEGF therapy has led to an overwhelming improvement in the prognosis of patients affected by neovascular AMD, allowing recovery and maintenance of visual function in the vast majority of patients. However, the therapeutic benefit is accompanied by significant economic investments, unresolved medicolegal debates about the use of off-label substances and overwhelming problems in large population management. The burden of disease has turned into a burden of care with a dissociation of scientific advances and real-world clinical performance. Simultaneously, ground-breaking innovations in diagnostic technologies, such as optical coherence tomography, allows unprecedented high-resolution visualisation of disease morphology and provides a promising horizon for early disease detection and efficient therapeutic follow-up. However, definite conclusions from morphologic parameters are still lacking, and valid biomarkers have yet to be identified to provide a practical base for disease management. The European Society of Retina Specialists offers expert guidance for diagnostic and therapeutic management of neovascular AMD supporting healthcare givers and doctors in providing the best state-of-the-art care to their patients.

TRIAL REGISTRATION NUMBER

NCT01318941.

Influence of image compression on the interpretation of spectral-domain optical coherence tomography in exudative age-related macular degeneration

PURPOSE

To evaluate the effect of image compression of spectral-domain optical coherence tomography (OCT) images in the examination of eyes with exudative age-related macular degeneration (AMD).

METHODS

Thirty eyes from 30 patients who were diagnosed with exudative AMD were included in this retrospective observational case series. The horizontal OCT scans centered at the center of the fovea were conducted using spectral-domain OCT. The images were exported to Tag Image File Format (TIFF) and 100, 75, 50, 25 and 10% quality of Joint Photographic Experts Group (JPEG) format. OCT images were taken before and after intravitreal ranibizumab injections, and after relapse. The prevalence of subretinal and intraretinal fluids was determined. Differences in choroidal thickness between the TIFF and JPEG images were compared with the intra-observer variability.

RESULTS

The prevalence of subretinal and intraretinal fluids was comparable regardless of the degree of compression. However, the chorio-scleral interface was not clearly identified in many images with a high degree of compression. In images with 25 and 10% quality of JPEG, the difference in choroidal thickness between the TIFF images and the respective JPEG images was significantly greater than the intra-observer variability of the TIFF images (P=0.029 and P=0.024, respectively).

CONCLUSIONS

In OCT images of eyes with AMD, 50% of the quality of the JPEG format would be an optimal degree of compression for efficient data storage and transfer without sacrificing image quality.

Is spectral-domain optical coherence tomography essential for flexible treatment regimens with ranibizumab for neovascular age-related macular degeneration?

Purpose. To evaluate the ability of spectral-domain optical coherence tomography to detect subtle amounts of retinal fluid when the choroidal neovascularization is detected as inactive via time-domain optical coherence tomography and clinical examination in neovascular age-related macular degeneration (nAMD) patients. Methods. Forty-nine eyes of 49 patients with nAMD after ranibizumab treatment were included in this cross-sectional, prospective study. All patients were imaged with TD-OCT and SD-OCT at the same visit one month after a ranibizumab injection. The presence of subretinal, intraretinal, and subretinal pigment epithelium fluid (subRPE) in SD-OCT was evaluated; also mean central retinal thickness (CRT) and the rate of vitreoretinal surface disorders detected via the two devices were evaluated. Results. The mean CRT via TD-OCT and SD-OCT was 218.1 ± 51.3 and 325.7 ± 78.8 microns. Sixteen patients (32.6%) showed any kind of retinal fluid via SD-OCT. In detail, 8 patients (16.3%) showed subretinal fluid, 10 patients (20.4%) showed intraretinal fluid, and 3 patients (6.1%) showed SubRPE fluid. The ability of detecting vitreoretinal surface disorders was comparable between the two devices, except vitreomacular traction. Conclusion. SD-OCT is essential for the nAMD patients who are on an as-needed treatment regimen with ranibizumab. Only TD-OCT and clinical examination may cause insufficient treatment in this group of patients.

Comparison of central macular thickness measured by three OCT models and study of interoperator variability

Purpose. To compare central macular thickness (CMT) measurement on healthy patient using 3 different OCT devices by two operators. Methods. Prospective, monocentricstudy. Right eye's central macular thickness (CMT) of 30 healthy patients has been measured three times using a time-domain (TD) OCT (Stratus OCT, Carl Zeiss Meditec, Dublin, Ca) and two spectral domain (SD) OCTs (Cirrus HD-OCT, Carl ZeissMeditec, Dublin, Ca) and 3D-OCT 1000 (Topcon, Tokyo, Japan) by two operators. Six measurements were taken randomly for each patient the same day. Results. No significant difference between measurements obtained by the two operators has been observed, whatever the studied OCT. P value was 0.164, 0.193, and 0.147 for Stratus OCT, Cirrus HD-OCT and 3D-OCT, respectively. Mean CMT significantly differed from instrument to instrument (P < 0.001) and was, respectively, 197 μm, 254 μm, and 236 μm using Stratus OCT, Cirrus HD-OCT, and 3D-OCT 1000. Using Cirrus OCT and 3D-OCT 1000, CMT was, respectively, 57 μm and 39 μm thicker than using Stratus OCT (P < 0.05). Conclusions. Whatever the OCT device, on healthy patients CMT was not operator dependent. CMT measurements obtained by SD-OCTs are greater than those obtained by TD-OCT. These data imply that the different OCT devices cannot be used interchangeably in clinical monitoring.

Therapeutic efficacy of bevacizumab for age-related macular degeneration: what are the implications of CATT for routine management?

CATT (Comparison of Age-related macular degeneration [AMD] Treatment Trials) examined the efficacy of ranibizumab and bevacizumab for the treatment of neovascular AMD. This prospective, randomized, but unblinded trial revealed a significant improvement in vision with both treatments in terms of visual acuity; importantly, patients with juxtafoveal choroidal neovascularization (CNV) and retinal pigment epithelial detachments were not excluded from the study. Monthly treatment with the drugs resulted in similar increases in visual acuity, although angiograms indicated that ranibizumab was superior in terms of reducing retinal fluid and leakage. As the study also differentiated between a fixed regimen and an as-needed (pro re nata [PRN]) dosing regimen, a larger sample size and Bonferroni statistical correction were necessary. The equivalence of the PRN dosing of bevacizumab to the monthly treatment could not be confirmed. Almost all of the frequent deviations from the protocol (referring to retreatment criteria: 25.7-28.5%) resulted in under-treatment. Since this applied to both drugs equally, under-treatment alone could not explain the larger loss of visual acuity observed in the bevacizumab PRN arm. The PRN regimen was generally associated with a larger lesion size after 12 months compared with the fixed treatment regimens. The investigators accepted the drawbacks of an incomplete masking to allow co-payment by Medicare. As assessments of drug trials are often politically motivated, the higher demands of a non-inferiority trial compared with a superiority design must be emphasized. A comparison of the per-protocol and last-observation-carried-forward analysis has not yet been published; ongoing subgroup analysis might highlight the impact of different lesion characteristics. While CATT provided further evidence for the efficacy of bevacizumab treatment, differences in adverse events between the two treatments (e.g. a higher rate of serious adverse events with bevacizumab compared with ranibizumab) were reported; however, these still have to be analysed, with the larger sample sizes of previous ranibizumab studies needing to be taken into account. Preclinical studies imply some differences between the drugs in terms of their adverse event profiles. A possible increased risk of adverse events could not be ruled out by previous clinical case series and CATT because the sample sizes and the follow-up intervals were not adequate. The large discrepancy in the price of bevacizumab versus ranibizumab in the US means a cost-benefit analysis is warranted. A lack of quality-of-life data has prevented calculation of an appropriate bevacizumab price in the context of its performance in the ophthalmological setting. Thus, CATT suggests that a favourable visual acuity might be achieved by very frequent administration of bevacizumab in patients with neovascular AMD. Although there are certain safety caveats, increased focus on subgroup analyses and obtaining longer follow-up data are expected to yield additional information of clinical relevance.