Introduction

The integration of artificial intelligence (AI) into healthcare has led to a new era of insight, efficiency, and personalization. From radiology and dermatology to cardiology and ophthalmology, AI-driven tools are redefining how clinicians diagnose, treat, and monitor disease.1,2 In eye care, AI has already demonstrated value in detecting diabetic retinopathy, age-related macular degeneration, and glaucoma through automated image analysis.3-7 However, its application in contact lenses is only now beginning to unfold.

For example, it is feasible that in the future, a range of smart contact lenses with embedded biosensors will emerge. Such innovations are still in early-stage development; however, current iterations have already attracted interest for their potential to collect physiological data such as glucose concentration or intraocular pressure (IOP).8-11 Whilst these devices do not employ AI directly, artificial intelligence methods can be applied to analyze the continuous data streams they generate, helping clinicians interpret patterns, predict disease progression, or personalize treatment.12-14 More immediately, AI is being applied for contact lens fitting, particularly for complex lenses such as scleral or orthokeratology (ortho-k) designs. By analyzing large datasets of ocular topography and prior fitting outcomes, AI algorithms can assist practitioners in selecting initial lens parameters, reducing chair time, and improving first fit success rates.15-23

This editorial examines the ways in which AI is shaping the future of contact lenses, from advanced design and manufacturing processes to clinical use cases such as personalized fitting and real-time ocular health monitoring.

Understanding AI in Eye Care

AI broadly refers to computational systems that can perform tasks traditionally associated with human cognition, such as pattern recognition, decision making, and problem solving. Within healthcare, most AI applications rely on machine learning, a method where algorithms improve performance by learning from data rather than being explicitly programmed. A subset of this, known as deep learning, uses layered artificial neural networks to model complex, non-linear relationships, particularly useful for image, audio, and time series data.24 These algorithms learn patterns from data, allowing them to perform tasks such as classification, regression, and segmentation that are common in both clinical diagnostics and device development.25 For example, in contact lens fitting, models may be trained to classify corneal shapes, estimate lens parameters, or predict fitting success based on biometric inputs.15,16,18 In research in this field, AI techniques are increasingly explored for interpreting continuous sensor data, recognizing anomalies, or forecasting changes in physiological variables over time.12,14 These capabilities depend not only on data volume and quality, but also on choosing models appropriate to the task, such as convolutional neural networks (CNN) for image analysis or recurrent architectures for time-series signals.25

AI in Contact Lens Fitting and Clinical Decision Making

AI-assisted fitting systems have been explored most actively in the context of ortho-k,15,17–23,26 where lens selection is highly dependent on corneal geometry and individual ocular responses. Ortho-k fitting traditionally relies on empirical tables or trial lens sets guided by topography, yet significant variability between eyes and lens designs can make the process time consuming and clinician-dependent.

Fan et al. initiated their AI-focused work in ortho-k fitting by developing models to predict key lens design parameters, specifically the return zone depth (RZD) and landing zone angle. Using linear regression and support vector regression models trained on corneal topography and refractive data from 1037 Chinese eyes fitted with a corneal refractive therapy lens system, their models outperformed the sliding card method (a traditional rule-based approach using manufacturer-provided fitting charts to estimate lens parameters), particularly in predicting RZD. Their approach yielded improved fitting estimates in East Asian populations, where standard fitting guides may be less representative.26 In follow-up work during the COVID-19 pandemic, the group applied their machine learning approach to retrospective topography data from a vision shaping treatment design, enabling safe remote fitting with minimal contact while maintaining accuracy.23

Expanding the scope beyond lens parameter selection, Xu et al. developed machine learning models using data from over 2600 eyes to predict both alignment curve radius and one-year axial length change. Among the algorithms tested, random forest achieved the best performance for both tasks, outperforming traditional formula-based methods.22 In a related study, these authors combined deep learning and image processing to segment topographic features such as the pupil and treatment zone. The resulting quantitative metrics, including decentration and effective defocus range, were used to classify topography types with more than 98% of agreement compared to expert assessments, supporting potential clinical use in lens evaluation.21

Other research groups have explored complementary AI strategies to support lens fitting and improve clinical outcomes.17-20 Xiao et al. developed a predictive model for lens decentration after one month of wear, identifying age, 8 mm sagittal height difference, 5 mm Kx1 (flat curvature difference), and 7 mm Kx2 (oblique curvature difference) as key predictors.17 Yang et al. applied a deep neural network to predict curvature, power, and lens diameter using corneal and refractive data. Their model achieved high accuracy overall, though performance declined in more atypical corneal profiles, highlighting the need for broader training datasets.18 Koo et al. used supervised learning to predict a full set of lens parameters, including toricity and sagittal depth, in Korean adolescents, developing a web-based application that outperformed manufacturer calculators.20 Finally, Lan et al. trained an AI-assisted prescription system on over 11000 multibrand fitting records and demonstrated improved lens selection efficiency and reduced practitioner variability, especially among junior clinicians.19

Although most of the advances in AI in contact lens practice have focused on ortho-k for myopia management, similar approaches are now being applied to the more complex task of fitting rigid lenses in keratoconus.16,27,28 Given the irregular corneal shapes and high interpatient variability, selecting suitable lens parameters in keratoconus remains challenging and often requires multiple trial lenses.

Hashemi et al. developed a Multiview deep learning model that combined four topographic maps from the Pentacam to predict the best base curve for rigid corneal contact lenses. Their system used features from both custom-built and pretrained CNNs and showed strong agreement with parameters selected by experts.28 In a related study, Risser et al. trained a neural network on curvature maps obtained from Scheimpflug imaging to recommend Rose K2 base curves. Their model performed better than standard selection based on average corneal curvature and maintained accuracy even in more advanced cases, although performance declined slightly with decentered cones.27 Abadou et al. explored a broader set of AI methods using raw topographic maps from the MS39 device. Their work directly compared these models to conventional keratometry method in a real clinical setting using Rose K2 lenses to predict the posterior curvature of the most suitable lens. Their findings showed that models trained on full imaging data gave clearly better predictions of the best posterior lens curvature, supporting the use of detailed topography in guiding contact lens selection.16

Together, these studies highlight the expanding role of AI in specialty lens fitting, offering data driven support where conventional methods are least reliable.

AI Use in Sensor-Enabled Contact Lenses

For a time, the promise of truly intelligent contact lenses captured public imagination, driven in part by companies like Mojo Vision. Their concept lens integrated a micro display, sensors, and wireless communication into a scleral platform, aiming to provide real-time information overlays directly to the eye. Although the project attracted significant attention for blending advanced optics with digital functionality, development was paused in 2023 due to financial constraints and a shift in corporate focus, highlighting the commercial and technical challenges facing this emerging field.

Despite this setback, momentum around smart lenses has not faded. Deeptech startups such as XPANCEO, InWith Corporation, and Innovega continue to develop contact lenses with augmented reality and sensing functions, pushing the boundaries of miniaturized electronics, biocompatible materials, and embedded optics. As these platforms evolve, AI is expected to play a central role by interpreting complex bio signals, managing real-time data capture, and enabling adaptive visual interfaces. These advances point toward a future where contact lenses do not just ‘see’ but also understand and respond.

Beyond the conceptual realm of augmented reality contact lenses, some devices already function as truly smart lenses in a clinical context, even if they do not yet rely on AI directly. One example is the Triggerfish® lens from Sensimed, a soft contact lens embedded with strain gauges that measure subtle changes in ocular dimensions over 24 hours.29–31 While originally developed to monitor IOP indirectly, the lens produces complex time-series data that has traditionally lacked advanced tools for clinical interpretation.

In an early effort to address this gap, Martin et al. explored whether parameters derived from Triggerfish recordings could support glaucoma diagnosis.12 Their study applied machine learning to signals reflecting ocular volume dynamics, demonstrating that even without conventional IOP readings, certain lens-derived features could distinguish between healthy individuals and those with primary open-angle glaucoma. The results suggested that these continuous signals captured biomechanical fluctuations relevant to disease processes beyond pressure alone.

More recently, Świerczyński et al. built on this concept by designing a classification model based solely on Triggerfish output to aid in glaucoma diagnosis. Recognizing that most AI research in ophthalmology focuses on fundus images or structural scans,32 their work signaled a shift toward analyzing longitudinal bio signals captured from the ocular surface. Their findings confirmed that variability patterns and nocturnal signal behaviors contained clinically useful information, especially when interpreted with machine learning.13,14

Together, these studies highlight the potential of smart lenses not only as tools for collecting data but also as sources of clinically useful insight. As methods for time series analysis advance and more AI models are trained specifically on this type of data, contact lens-based monitoring may become an important part of glaucoma diagnostics and other areas of care.

Opportunities, Barriers, and Ethical Considerations

The integration of AI into contact lens technologies offers a compelling vision for the future of eye care. Beyond automating fitting procedures, AI systems can support consistent lens selection, reduce chair time, and assist clinicians in managing complex cases.15,16,20 Looking further ahead, lenses equipped with sensors may allow real-time monitoring of ocular or systemic changes, enabling earlier diagnosis and personalized interventions.12–14

However, the path to full adoption presents practical, scientific, and ethical challenges. These include the availability and quality of training data, transparency and interpretability of algorithms, clinical validation, data privacy, and the fair distribution of new technologies.33 Most AI models are trained on retrospective data from narrow populations, limiting generalizability.34, 35 Prospective validation across diverse settings is essential. Continuous data capture also introduces challenges around ownership, consent, and privacy, especially when biometric signals are processed externally.33,36 In parallel, the shift toward smart sensing platforms brings unique engineering constraints. Embedding electronics and sensors into soft lenses requires precise fabrication on the nanoscale, along with stable, biocompatible materials that maintain comfort and safety over extended wear.37–39 Equally important is the continued role of the clinician. AI should support, not replace, clinical judgment, and its adoption must be guided by clear protocols that define responsibility when automated suggestions inform decision-making.40–42

REFERENCES:

  1. Alowais SA, Alghamdi SS, Alsuhebany N, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC medical education. 2023:23(1):689.
  2. Bajwa J, Munir U, Nori A, Williams B. Artificial intelligence in healthcare: transforming the practice of medicine. Future Healthcare Journal. 2021:8(2):e188–e194.
  3. De Fauw J, Ledsam JR, Romera-Paredes B, et al. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nature Medicine. 2018:24 (9):1342–1350.
  4. Nielsen KB, Lautrup ML, Andersen JKH, et al. Deep learning–based algorithms in screening of diabetic retinopathy: a systematic review of diagnostic performance. Ophthalmology Retina. 2029:3(4):294–304.
  5. Pead E, Megaw R, Cameron J, et al. Automated detection of age-related macular degeneration in color fundus photography: a systematic review. Survey of Ophthalmology. 2019:64(4):498–511.
  6. Hwang DK, Hsu CC, Chang KJ, et al. Artificial intelligence-based decision-making for age-related macular degeneration. Theranostics. 2019:9(1): 232, 2019.
  7. Romond K, Alam M, Kravets S, et al. Imaging and artificial intelligence for progression of age-related macular degeneration. Experimental Biology and Medicine. 2021:246(20):2159–2169.
  8. Roostaei N, Hamidi SM. Plasmonic smart contact lens based on etalon nanostructure for tear glucose sensing. Scientific Reports. 2025:15(1):14948.
  9. Fardoost A, Karimi K, Singh J, Patel H, Javanmard M. Enhancing glaucoma care with smart contact lenses: An overview of recent developments. Biomedical Microdevices. 2025:27(2):18.
  10. Pan M, Zhang Z, Shang L. Smart contact lenses: Disease monitoring and treatment. Research, 2025;8:0611.
  11. Wu KY, Dave A, Carbonneau M, Tran SD. Smart contact lenses in ophthalmology: Innovations, applications, and future prospects. Micromachines,. 2024:15(7):856.
  12. Martin KR, Mansouri K, Weinreb RN, et al. Use of machine learning on contact lens sensor–derived parameters for the diagnosis of primary open-angle glaucoma. American Journal of Ophthalmology. 2018:194:46–53.
  13. Świerczyński H, Pukacki J, Szczęsny S, Mazurek C, Wasilewicz R. Sensor data analysis and development of machine learning models for detection of glaucoma. Biomedical Signal Processing and Control. 2023:86:105350.
  14. Świerczyński H, Pukacki J, Szczęsny S, Mazurek C, Wasilewicz R. Application of machine learning techniques in glaucoma system for glaucoma diagnosis and collaborative research support. Scientific Reports. 2025:15(1):7940.
  15. Zhou X, Zeng L, Shen Y, et al.. Artificial intelligence-assisted fitting method using corneal topography outcomes enhances success rate in orthokeratology lens fitting. Cont Lens Anterior Eye. 2025:48(3):102350.
  16. Abadou J, Dahan S, Knoeri J, et al. Artificial intelligence versus conventional methods for RGP lens fitting in keratoconus. Cont Lens and Anterior Eye. 2025:8(1):102321.
  17. Xiao K, Lu W, Zhang X, et al. An integrative predictive model for orthokeratology lens decentration based on diverse metrics. Frontiers in Medicine. 2024:11:1490525.
  18. Yang HW, Liang CL, Chou S, Wang H, Chiang HK. Development and evaluation of a deep neural network model for orthokeratology lens fitting. Ophthalmic and Physiological Optics. 2024:44(6):1224–1236.
  19. Lan W, Tang H, Wen L, et al. Artificial intelligence–assisted prescription determination for orthokeratology lens fitting: From algorithm to clinical practice. Eye & Contact Lens, pages 10–1097, 2022.
  20. Koo S, Kim WK, Park YK, et al. Development of a machine–learning–based tool for overnight orthokeratology lens fitting. Translational Vision Science & Technology. 2024:13(2):17–17.
  21. Xu S, Yang X, Zhang S, et al. Evaluation of the corneal topography based on deep learning. Frontiers in Medicine. 2024:10:1264659.
  22. Xu S, Yang X, Zhang S, et al. Machine learning models for orthokeratology lens fitting and axial length prediction. Ophthalmic and Physiological Optics. 2023:43(6):1462–1468.
  23. Fan Y, Yu Z, Tang T, et al. Machine learning algorithm improves accuracy of ortho-k lens fitting in vision shaping treatment. Cont Lens and Anterior Eye. 2022:45(3):101474.
  24. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015:521(7553):436–444.
  25. Esteva A, Chou K, Yeung S, et al. Deep learning-enabled medical computer vision. NPJ digital medicine. 2021:4(1):5.
  26. Fan Y, Yu Z, Peng Z, et al. Machine learning based strategy surpasses the traditional method for selecting the first trial lens parameters for corneal refractive therapy in Chinese adolescents with myopia. Cont Lens and Anterior Eye. 2021:44(3):101330.
  27. Risser G, Mechleb N, Muselier A, Gatinel D, Zéboulon P. Novel deep learning approach to estimate rigid gas permeable contact lens base curve for keratoconus fitting. Cont Lens and Anterior Eye. 2023:46(6):102063.
  28. Hashemi S, Veisi H, Jafarzadehpur E, Rahmani R, Heshmati Z. Multi-view deep learning for rigid gas permeable lens base curve fitting based on Pentacam images. Medical & Biological Engineering & Computing. 2020:58:1467–1482.
  29. Leonardi M, Leuenberger P, Bertrand D, Bertsch A, Renaud P. First steps toward noninvasive intraocular pressure monitoring with a sensing contact lens. Invest Ophthal Vis Sci. 2004:45(9):3113–3117.
  30. Leonardi M, Pitchon EM, Bertsch A, Renaud P, Mermoud A. Wireless contact lens sensor for intraocular pressure monitoring: assessment on enucleated pig eyes. Acta ophthalmol. 2009:87(4): 433–437.
  31. Mansouri K, Weinreb RN, Liu JH. Efficacy of a contact lens sensor for monitoring 24-h intraocular pressure-related patterns. PloS One. 2015:10(5):e0125530.
  32. Thompson AC, Jammal AA, Medeiros FA. A review of deep learning for screening, diagnosis, and detection of glaucoma progression. Transl Vis Sci Technol. 2020:9(2):42–42.
  33. Ting DS, Peng L, Varadarajan AV, et al. Deep learning in ophthalmology: the technical and clinical considerations. Prog Retin Eye Res. 2019:72:100759.
  34. Rashidisabet H, Sethi A, Jindarak P, et al. Validating the generalizability of ophthalmic artificial intelligence models on real-world clinical data. Transl Vis Sci Technol. 2023:12(11):8–8.
  35. Li Z, Wang L, Wu X, et al. Artificial intelligence in ophthalmology: The path to the real-world clinic. Cell Reports Medicine. 2023:4(7).
  36. Murdoch B. Privacy and artificial intelligence: challenges for protecting health information in a new era. BMC Medical Ethics. 2021:22:1 5.
  37. Kazanskiy NL, Khonina SN, Butt MA. Smart contact lenses—a step towards non-invasive continuous eye health monitoring. Biosensors. 2023:13(10):933.
  38. Abdulamier AA, Shaker LM, Al-Amiery AA, et al. Advancements and applications of smart contact lenses: a comprehensive review. Results in Engineering. 2024:103268.
  39. Psoma S, Ndege G. Emerging smart contact lens technology for wearable biosensors and drug delivery: Biomarkers in tears. EMJ Innovations. 2025:9(1):95–105.
  40. Vayena E, Blasimme A, Cohen IG. Machine learning in medicine: addressing ethical challenges. PLoS Medicine. 2018:15(11): e1002689.
  41. Hatherley J, Kinderlerer A, Bjerring JC, Munch LA, Threlfall L. The FHIJ debate: Will artificial intelligence replace clinical decision-making within our lifetimes? Future Healthc J. 2024:11(3):100178.
  42. Akingbola A, Adeleke O, Idris A, Adewole O, Adegbesan A. Artificial intelligence and the dehumanization of patient care. Journal of Medicine, Surgery, and Public Health. 2024:3:100138.


Get the latest Optical News delivered to you. Click here to subscribe to our FREE print magazine and weekly newsletter