Cataract surgery is one of the most frequently performed procedures worldwide, and its demand is rising as the population ages. The integration of AI and machine learning (ML) is reshaping ophthalmic microsurgery, particularly in relation to IOL selection, cataract grading, and patient counseling. These technologies leverage large datasets to predict outcomes, personalize treatment, and improve communication, with the goal of optimizing refractive results and patient satisfaction. In the preoperative phase, AI excels in IOL power calculation, a critical step for achieving the target refraction. Traditional formulas rely on optical models, whereas AI-driven approaches, such as neural networks, enhance accuracy by analyzing complex, patient-specific variables. Beyond power calculations, AI can transform patient counseling by simulating outcomes, explaining risks, and tailoring discussions to the individual patient.
Over the next 5 to 10 years, advances in deep learning (DL), big data integration, and multimodal AI could make these processes even more precise and accessible, but data limitations and ethical concerns remain. This article explores the current state, practical applications, opportunities, and limitations of AI in individualized IOL selection and patient counseling.
CURRENT APPLICATIONS
Personalized IOL Calculations
AI has revolutionized IOL power calculations by transitioning from rigid optical formulas to data-driven predictions. An example of a pure AI model is the Hill-RBF (Radial Base Function) formula,1 which utilizes neural networks to predict IOL power based on biometric data. Compared to classical formulas, the Hill-RBF has been shown to achieve high accuracy in diverse eyes, including those with high axial myopia. The PEARL-DGS formula, which I helped develop, employs a hybrid strategy that uses ML to estimate parameters such as posterior corneal curvature or effective lens position while retaining an optical backbone for reliability.2 Both tools draw on large databases of postoperative refractions, IOL types, and biometric measurements.
A notable advantage of AI is its capacity to analyze extensive clinical data and identify empiric patterns from real-world outcomes.
Objective Quantification of Lens Opacification
Accurate assessment of cataract severity is crucial for surgical planning and patient counseling. Traditionally, grading has relied on subjective clinical scales such as the Lens Opacities Classification System III, which are prone to interobserver variability.
Recent advances in AI enable objective quantification of lens opacification using slit-lamp video analysis,3 swept-source OCT imaging,4 and retroillumination photography. DL models trained on large datasets can classify cataract severity, detect subtle progression, and provide reproducible metrics for surgical decision-making. Such tools could standardize cataract grading worldwide, improve longitudinal follow-up, and facilitate earlier intervention when indicated.
Enhanced Patient Counseling
Counseling patients before cataract surgery involves educating them about IOL options, surgical risks, and expected visual outcomes. AI can augment this process by providing data-backed insights and interactive tools. For example, AI-powered chatbots or virtual assistants can simulate patients’ postoperative vision, helping them visualize differences between monofocal and premium IOLs. Systems such as Dora, an AI-driven voice assistant currently used to manage postoperative queries, could be adapted to guide preoperative discussions on recovery timelines and visual outcomes.5 ML algorithms can stratify patient-specific risk to predict complications such as posterior capsular opacification with up to 92.2% accuracy,6 thereby enabling surgeons to present personalized risk probabilities.
Today’s AI tools, such as predictive dashboards, display individualized outcome probabilities based on patient data to support informed consent. A study of TempSeq-Net, a convolutional neural network–based model, analyzed slit-lamp images to forecast complication progression, which could assist in planning postoperative follow-up.7
FUTURE FORECAST
Over the next decade, AI is expected to evolve toward hyperpersonalized IOL selection through advanced DL and integration with emerging technologies. By 2030 to 2035, widespread adoption of generative AI for real-time biometric adjustments is anticipated that incorporates genetic data and lifestyle factors. Enhanced hybrid formulas could leverage cloud-based, real-world data for continuous prediction updates. Multimodal AI combining OCT, fundus imaging, and electronic health records could predict effective lens position with greater precision, enabling spectacle-free outcomes in more than 90% of cases.8
In counseling, AI-driven virtual reality simulations might allow patients to preview their postoperative vision, while large language model–based chatbots could offer culturally sensitive, multilingual discussions and predict satisfaction based on psychometric data. Intraoperative AI feedback, such as DeepPhase9 for surgical phase recognition, could inform preoperative planning in real time. Federated learning—training AI models across institutions without sharing raw data—could address privacy concerns and accelerate innovation.10
OPPORTUNITIES
AI presents substantial opportunities to improve precision and equity in cataract care.
In IOL selection, ML could reduce postoperative refractive errors by 20% to 30% compared with traditional formulas.2 AI-based cataract grading3,4 could standardize surgical thresholds globally.
For patient counseling, AI can democratize information, support shared decision-making, and reduce anxiety through personalized visualizations.
On a global scale, mobile AI tools using smartphone imaging11,12 could extend access to remote populations.
Potential training benefits are also significant: PhacoTracking uses ML to analyze surgeons’ intraoperative movements to promote more effective IOL planning by novice surgeons.13
Cost-effectiveness is another advantage; automated platforms such as Dora could streamline postoperative follow-up, freeing up clinical resources for patient counseling.5
LIMITATIONS AND PRACTICAL REALITIES
Data scarcity limits AI generalizability. Many datasets are restricted to ultrasonic phaco cases, excluding femtosecond laser–assisted and manual techniques.2 Bias in training data can reduce accuracy for underrepresented populations.
Ethical issues arise if AI replaces rather than augments clinician judgment. Implementing AI requires robust infrastructure, high-quality biometry, and interoperable electronic health records systems, which are lacking in many settings.
Costs and the black box nature of AI predictions might limit clinician trust. Developing explainable AI and combining empirical models with optical theory could improve acceptance and clinical impact.
CONCLUSION
AI and ML are transforming cataract surgery planning and counseling, from precise IOL calculations to objective cataract grading. Integrating these tools into routine workflows could enhance outcomes, standardize care, and expand access. Achieving these benefits, however, will require rigorous validation, regulatory oversight, and equitable deployment.
1. Wan KH, Chan TCY, Law SM, et al. 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.
2. Debellemanière G, Dubois M, Gauvin M, et al. The PEARL-DGS formula: the development of an open-source machine learning-based thick IOL calculation formula. Am J Ophthalmol. 2021;232:58-69.
3. Shimizu E, Tanji M, Nakayama S, et al. AI-based diagnosis of nuclear cataract from slit-lamp videos. Sci Rep. 2023;13(1):22046.
4. Zéboulon P, Panthier C, Rouger H, et al. Development and validation of a pixel-wise deep learning model to detect cataract on swept-source OCT images. J Optom. 2022;15(suppl 1):S43-S49.
5. Hatamnejad A, Higham A, Somani S, et al. Feasibility of an artificial intelligence phone call for postoperative care following cataract surgery in a diverse population: two phase prospective study protocol. BMJ Open Ophthalmol. 2024;9(1):e001475.
6. Mohammadi SF, Abdulrahman ZM, Patel DV, et al. Using artificial intelligence to predict the risk for posterior capsule opacification after phacoemulsification. J Cataract Refract Surg. 2012;38(3):403-408.
7. Jiang J, Deng B, Liu L, et al. Predicting the progression of ophthalmic disease based on slit-lamp images using a deep temporal sequence network. PLoS One. 2018;13(7):e0201142.
8. Choi JY, Choi H, Cho Y, Yoo TK. Artificial intelligence and refractive surgeries including laser vision correction and phakic IOL implantation—a narrative review. Ann Eye Sci. 2025;10:7. doi:10.21037/aes-24-40
9. Zisimopoulos O, Flouty E, Luengo I, et al. DeepPhase: surgical phase recognition in CATARACTS videos. In: Frangi AF, Schnabel JA, Davatzikos C, Alberola-López C, Fichtinger G, eds. Medical Image Computing and Computer Assisted Intervention – MICCAI 2018: 21st International Conference, Proceedings, Part IV. Springer; 2018:265-272.
10. Tabuchi H, Shimizu K, Ito Y, et al. Real-time artificial intelligence evaluation of cataract surgery: a preliminary study. Taiwan J Ophthalmol. 2022;12(2):147-154.
11. Grammatikopoulou M, Flouty E, Kadkhodamohammadi A, et al. CaDIS: cataract dataset for surgical RGB-image segmentation. Med Image Anal. 2021;71:102053.
12. Lindegger DJ, Wawrzynski J, Saleh GM. Evolution and applications of artificial intelligence to cataract surgery. Ophthalmol Sci. 2022;2(3):100164.
13. Smith P, Tang L, Balntas V, et al. “PhacoTracking”: an evolving paradigm in ophthalmic surgical training. JAMA Ophthalmol. 2013;131(5):659-661.
