Digital Interventional Tools for Cataract Surgery
Cataract is the leading cause of curable blindness, affecting an estimated 52.6 million individuals globally with moderate or severe vision impairment (1). Several reasons, including early intervention, increased eye surgery frequency, and population aging, have recently contributed to a rise in the incidence of cataract procedures (2, 3) However, limited resources and lengthy wait times are key impediments to cataract therapy in countries that rely predominantly on public healthcare systems. As a result of growing aging populations, this load is likely to grow significantly (4). However, eyecare services have not expanded in tandem, resulting in a gap that is becoming increasingly difficult to solve. Moreover, compared to affluent nations, low to middle-income economies have a greater prevalence of cataract-related visual impairment (1). Indeed, cataract accounts for more than half of all incidences of blindness in middle- and low-income nations (2, 5). Cataract surgery is one of the most cost-effective healthcare therapies, resulting in physical and psychological benefits (6). It is generally established that early detection and treatment of this condition helps to avert complications and save healthcare expenditures (2).
In the recent years, several artificial intelligence (AI) technologies have been developed to help in various aspects of cataract care. Artificial intelligence (AI)-based solutions have shown considerable promise in a variety of patient-care settings (7). As a result, as emerging nations progress and the general population ages, AI systems may become a significant aspect of the screening, staging, and treatment of eye disorders. This, in turn, might lead to faster population coverage, more accurate diagnosis and treatment, and a reduction in the amount of time-consuming activities performed by specialists (2). In this case, AI might help to bridge these gaps, sparking interest in its potential in cataract care.
Digital Tools for Diagnosis
Cataracts are currently clinically diagnosed at the slit-lamp by ophthalmologists, necessitating a face-to-face consultation. As a result, undiagnosed cataracts continue to pose a significant concern for many developing nations and rural communities due to a lack of access. An AI-assisted telemedicine platform for the preliminary diagnosis of cataract would remove barriers to access and hence reduce healthcare costs. A prompt diagnosis is critical, especially for pediatric cataracts, which can cause irreparable amblyopia (4). Some authors proposed methods to overcome these problems. In 2015, Gao et al.(8) suggested a method to score cataracts using slit-lamp pictures and a mixture of convolutional neural networks (CNN), recurrent neural networks (RNN), and support vector regression (SVR). For detecting referable cataracts, the system attained a 70.7% agreement ratio. Li et al. (9) proposed using slit-lamp pictures to diagnose and annotate cataracts as well as other anterior segment disorders using a proprietary model (Visionome). Visionome's performance was equivalent to that of ophthalmologists while outperforming an ophthalmologist with one year of clinical experience (accuracy 79.47-99.22%). Wu et al. (10) developed a ResNet DL algorithm with a three-step procedure for diagnosing and referral cataracts. It then distinguished between a cataractous lens, an IOL, and a normal crystalline lens, with an area under receiver operating characteristic curve > 0.99. In another work, Xu et al. (11) used fundus pictures as input to create a CNN-based ensemble algorithm (AlexNet and VisualDN) that detected and graded cataracts with an accuracy of 86.2%.
Digital Tools for Intraocular lens power calculation
Cataract surgeons have long strived for precise postsurgical refractive results. Despite substantial advancements in IOL formulations, individuals with a history of refractive surgery, severe or unusual biometry continue to pose a difficult dilemma. Nowadays, AI is being used by IOL formulae to improve prediction outcomes. Sramka et al. (12) assessed the SVM regression model and multilayer neural network ensemble model (MLNN-EM) and found that both ML algorithms outperformed conventional clinical approaches in terms of prediction accuracy. Ladas et al. (13) investigated the use of AI to improve current IOL formulas (SRK, Holladay I, and Ladas Super formula), combining supervised learning algorithms (SVR, extreme gradient boost [XGB], and ANN) with the aforementioned existing formulae to refine the anticipated refractive result.
Furthermore, novel formulae that are either AI-based or use AI-incorporated elements have been devised. These AI formulas have a bright future because several have demonstrated great prediction accuracy when compared to existing formulae (4).
The Kane model is theoretically based and uses regression and AI components to enhance predictions. In comparative tests, it has emerged as one of the best-performing formulas, outperforming Barrett Universal II, Haigis, Olsen, and other third-generation formulae (14-17). Even among newer generation formulations, it has continuously been a top three performance, and these results were relevant to both extremities of axial lengths (14-16).
The Hill-radial basis function (RBF) is an ANN IOL calculator that use regression analysis to analyze a large data set of refractive outcomes. It predicts refractive outcomes using pattern recognition and data interpolation (18). The PEARL-DGS formula predicts effective lens position and adjusts for extreme biometric values using ML modeling and output linearization.
Karmona (19) is another novel data-driven IOL power calculation approach. It predicts IOL power using several ML models (e.g., K-Nearest Neighbor, ANN, SVM, and random forest) with particular preoperative parameters.
Digital Tools for Surgery
AI can also help with cataract surgery training, intraoperative decision-making, and postoperative analysis to improve surgical techniques.
In a recent study, an AI program was able to correctly distinguish the various phases of cataract surgery (20). Automatic surgical tool detection (21) developed into automated phase detection on cataract surgery movies (22). Yu et al. (23) discovered that modeling instrument labels (alone or with video pictures) rather than video images alone provided the greatest accuracy for automated phase detection in cataract surgery. The identification of distinct phases of cataract surgery has the potential to translate into phase-specific assessments of surgical technical abilities and enable procedure-tracking throughout surgery. This will provide real-time feedback and will improve intraoperative decision-making (24).
Quellec and colleagues have created an autonomous video analysis system based on a novel algorithm capable of real-time surgical task recognition (25).
Furthermore, Artificial intelligence may be used to anticipate the risk of intraoperative problems and enhance surgical procedures. Lanza et al. (26) trained an AI model to detect risk variables for intraoperative problems and predict total surgical length after studying 1229 operations with 73 errors.
Moreover, AI and virtual reality can work together to create intelligent teaching systems for cataract surgery training (4). Eyesi (Haag-Streit, Köniz, Switzerland) (27), a commercially accessible ophthalmological simulation-based training system, is na exemple of an approach to give a comprehensive training experience, allowing learners to achieve competency before real patient exposure.
Optimizing cataract patient treatment is more important than ever (2, 4). AI has the ability to change cataract management in terms of evaluation and monitoring, IOL computation, intraoperative feedback, and postoperative care (4). Successful clinical translation can provide long-term benefits, particularly for low-income populations, such as increased healthcare efficiency, accessibility, scalability, and cost savings (4). Nonetheless, several problems must be overcome in order to do this, including ethical data management, ensuring security and privacy, demonstrating clinically acceptable performance, enhancing generalizability across different populations, and boosting user acceptability. Furthermore, only a few clinical trials have assessed the efficacy of AI systems in real-world scenarios to date. In fact, only a few algorithms have been demonstrated to be reliable in a clinical environment. However, given the likelihood of biases, additional randomized controlled studies are required to test the usefulness of AI systems (2).
1. Flaxman SR, Bourne RRA, Resnikoff S, Ackland P, Braithwaite T, Cicinelli MV, et al. Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis. Lancet Glob Health. 2017;5(12):e1221-e34. 2. Tognetto D, Giglio R, Vinciguerra AL, Milan S, Rejdak R, Rejdak M, et al. Artificial intelligence applications and cataract management: A systematic review. Surv Ophthalmol. 2022;67(3):817-29. 3. Wan Zaki WMD, Abdul Mutalib H, Ramlan LA, Hussain A, Mustapha A. Towards a Connected Mobile Cataract Screening System: A Future Approach. J Imaging. 2022;8(2). 4. Gutierrez L, Lim JS, Foo LL, Ng WY, Yip M, Lim GYS, et al. Application of artificial intelligence in cataract management: current and future directions. Eye Vis (Lond). 2022;9(1):3. 5. Pratap T, Kokil P. Efficient network selection for computer-aided cataract diagnosis under noisy environment. Comput Methods Programs Biomed. 2021;200:105927. 6. Tognetto D, Brézin AP, Cummings AB, Malyugin BE, Evren Kemer O, Prieto I, et al. Rethinking Elective Cataract Surgery Diagnostics, Assessments, and Tools after the COVID-19 Pandemic Experience and Beyond: Insights from the EUROCOVCAT Group. Diagnostics (Basel). 2020;10(12). 7. Taylor M, Liu X, Denniston A, Esteva A, Ko J, Daneshjou R, et al. Raising the Bar for Randomized Trials Involving Artificial Intelligence: The SPIRIT-Artificial Intelligence and CONSORT-Artificial Intelligence Guidelines. J Invest Dermatol. 2021;141(9):2109-11. 8. Gao X, Lin S, Wong TY. Automatic Feature Learning to Grade Nuclear Cataracts Based on Deep Learning. IEEE Trans Biomed Eng. 2015;62(11):2693-701. 9. Li W, Yang Y, Zhang K, Long E, He L, Zhang L, et al. Dense anatomical annotation of slit-lamp images improves the performance of deep learning for the diagnosis of ophthalmic disorders. Nat Biomed Eng. 2020;4(8):767-77. 10. Wu X, Huang Y, Liu Z, Lai W, Long E, Zhang K, et al. Universal artificial intelligence platform for collaborative management of cataracts. The British journal of ophthalmology. 2019;103(11):1553-60. 11. Xu X, Zhang L, Li J, Guan Y, Zhang L. A Hybrid Global-Local Representation CNN Model for Automatic Cataract Grading. IEEE J Biomed Health Inform. 2020;24(2):556-67. 12. Sramka M, Slovak M, Tuckova J, Stodulka P. Improving clinical refractive results of cataract surgery by machine learning. PeerJ. 2019;7:e7202. 13. Ladas J, Ladas D, Lin SR, Devgan U, Siddiqui AA, Jun AS. Improvement of Multiple Generations of Intraocular Lens Calculation Formulae with a Novel Approach Using Artificial Intelligence. Transl Vis Sci Technol. 2021;10(3):7. 14. Kane JX, Melles RB. Intraocular lens formula comparison in axial hyperopia with a high-power intraocular lens of 30 or more diopters. J Cataract Refract Surg. 2020;46(9):1236-9. 15. Hipólito-Fernandes D, Elisa Luís M, Gil P, Maduro V, Feijão J, Yeo TK, et al. VRF-G, a New Intraocular Lens Power Calculation Formula: A 13-Formulas Comparison Study. Clin Ophthalmol. 2020;14:4395-402. 16. Melles RB, Kane JX, Olsen T, Chang WJ. Update on Intraocular Lens Calculation Formulas. Ophthalmology. 2019;126(9):1334-5. 17. Savini G, Hoffer KJ, Balducci N, Barboni P, Schiano-Lomoriello D. Comparison of formula accuracy for intraocular lens power calculation based on measurements by a swept-source optical coherence tomography optical biometer. J Cataract Refract Surg. 2020;46(1):27-33. 18. Wan KH, Lam TCH, Yu MCY, Chan TCY. Accuracy and Precision of Intraocular Lens Calculations Using the New Hill-RBF Version 2.0 in Eyes With High Axial Myopia. Am J Ophthalmol. 2019;205:66-73. 19. Carmona González D, Palomino Bautista C. Accuracy of a new intraocular lens power calculation method based on artificial intelligence. Eye (Lond). 2021;35(2):517-22. 20. Al Hajj H, Lamard M, Cochener B, Quellec G. Smart data augmentation for surgical tool detection on the surgical tray. Annu Int Conf IEEE Eng Med Biol Soc. 2017;2017:4407-10. 21. Al Hajj H, Lamard M, Conze PH, Cochener B, Quellec G. Monitoring tool usage in surgery videos using boosted convolutional and recurrent neural networks. Med Image Anal. 2018;47:203-18. 22. Lecuyer G, Ragot M, Martin N, Launay L, Jannin P. Assisted phase and step annotation for surgical videos. Int J Comput Assist Radiol Surg. 2020;15(4):673-80. 23. Yu F, Silva Croso G, Kim TS, Song Z, Parker F, Hager GD, et al. Assessment of Automated Identification of Phases in Videos of Cataract Surgery Using Machine Learning and Deep Learning Techniques. JAMA Netw Open. 2019;2(4):e191860. 24. Navarrete-Welton AJ, Hashimoto DA. Current applications of artificial intelligence for intraoperative decision support in surgery. Front Med. 2020;14(4):369-81. 25. Quellec G, Charrière K, Lamard M, Droueche Z, Roux C, Cochener B, et al. Real-time recognition of surgical tasks in eye surgery videos. Med Image Anal. 2014;18(3):579-90. 26. Lanza M, Koprowski R, Boccia R, Krysik K, Sbordone S, Tartaglione A, et al. Application of Artificial Intelligence in the Analysis of Features Affecting Cataract Surgery Complications in a Teaching Hospital. Front Med (Lausanne). 2020;7:607870. 27. Menozzi M, Ropelat S, Köfler J, Huang YY. Development of Ophthalmic Microsurgery Training in Augmented Reality. Klin Monbl Augenheilkd. 2020;237(4):388-91.