CORDIS - EU research results

Development of corneal biomechanical model. Dynamic topographical characterization based on 3D plenoptic imaging

Final Report Summary - POPCORN (Development of corneal biomechanical model. Dynamic topographical characterization based on 3D plenoptic imaging.)

Executive Summary:
The European project Popcorn has led to the development of a system allowing registering changes on the shape of the overall cornea during a short time (around 20 ms) while a controlled force with an air jet is applied. These changes are captured by a dynamical topographer based on plenoptic imaging. The plenoptic imaging system consists of an array of microlenses coupled with a video sensor designed for high speed video imaging. The micro-lens array makes a sampling of the entire ray map reflected from the cornea without needing time for scanning. Fast and robust algorithms for depth map estimation allow the reconstruction of the 3D corneal elevation and the dynamic changes during the application of the air pressure. This information is uploaded to a biomechanical numerical model that includes both the patient-specific corneal geometry and the patient-specific tissue properties. Furthermore, the model allows computing predictive models for different corneal surgical procedures, including laser in situ keratomileusis (LASIK) or small-incision lenticule extraction (SMILE).

Project Context and Objectives:
Successful laser-based refractive eye surgery provides efficacious and safe correction of refractive errors which are the second cause of avoidable blindness in the world. However, this type of procedures cannot be recommended to all patients and a careful preoperative examination is necessary, including a detailed analysis of corneal topography and thickness. In spite of this detailed examination and the high precision of laser systems, residual refractive errors still appears in some cases that are not easy to explain in all cases. On the other hand, as the laser refractive correction is not possible in cases of corneal ectasia, other corrective options such as the implantation of intracorneal ring segments have been developed. This option, although potentially useful, lacks of sufficient predictability. The main explanation for the presence of unexpected residual refractive errors or the limited predictability of corneal implants is the extremely limited knowledge of the corneal biomechanical response to these surgical treatments. To this date, only the Ocular Response Analyzer (ORA) from Reichert and the CorVis ST system from Oculus have been developed for the clinical evaluation of the corneal biomechanical properties which have been shown to be of limited validity and to provide metrics with no direct relation with the standard physical properties of the viscoelastic materials. The main barrier to in-vivo measurement of the mechanical properties of the cornea is the difficulty of registering dynamically the changes occurring in the topographic profile of the corneal surface in response to the application of a load. The introduction of a plenoptic system for
3D imaging of the cornea and the corresponding algorithms to analyze 3D changes in its shape may be the one potential solution to overcome this barrier.
The first objective of the current study is to develop a new non-invasive and accurate system based on plenoptic imaging combined with an advanced biomechanical model to characterize the biomechanical properties of the cornea in the clinical setting. The personalized and in-vivo
biomechanical modeling provided by this methodology will allow the clinician to predict the biomechanical behavior and status of the cornea after a surgical treatment and to avoid risks and bad effects that can cause the need of new treatments. This modeling will also allow surgeons to define and select more predictable implants to correct refractive errors and corneal aberrations in ectatic corneas. Furthermore, the standardization of the corneal parameters provided by this technology will convert it into a useful clinical tool for the diagnosis of a great variety of corneal disorders.
The second objective is to validate clinically the measurements provided by the device developed by evaluating the consistency of repeated corneal measurements (intraobserver repeatability) taken in different groups of eyes with specific characteristics, the distribution of the corneal biomechanical parameters obtained in the normal healthy population in order to characterize it, the ability of the device to detect significant differences in corneal biomechanics between normal healthy corneas and those with keratoconus (ectatic corneal disease) or previous excimer laser surgery for the correction of refractive errors (LASIK, laser assisted in situ keratomileusis), and the capability of the biomechanical model combined with the plenoptic system of predicting biomechanical alterations after corneal refractive surgery (LASIK).

Project Results:
The POPCORN project is a European project within the 7th Framework (Grant Agreement 606634) that has just finished with interesting findings on the clinical characterization of the corneal biomechanics ( The aim of the project has been the development of a device for characterizing the mechanical properties of the cornea based on the plenoptic imaging technique, three-dimensional (3D) mapping algorithms and an advanced biomechanical model that has been patented. The project was possible by a consortium including small and medium enterprises (SMEs) and technological institutes, which were coordinated by Alicante Oftalmologica OFTALMAR, an ophthalmological clinic within the Vithas Medimar International Hospital founded in 1996. The Popcorn consortium includes the SMEs Biotronics 3D from United Kingdom, Optoelectronica 2001 from Romania, and Costruzione Strumenti Oftalmici (CSO) from Italy, as well as the Applied Mechanics and Bioengineering Research Group from the University of Zaragoza (Spain), and the UK Intelligent Systems Research Institute (Figure 1).

The use of plenoptic for 3D mapping of the cornea
The transparent cornea is a living tissue that constitutes the most important component of the outer ocular layers. The peculiar layered structure of the stromal collagen lamellae provides a specific mechanical behaviour to the cornea, being very dependent on the direction of the collagen fibres. Before predicting its mechanical behaviour, it is essential to understand the consequences of modifying the overall corneal geometry when it is subjected to an intracorneal ring segments or by laser refractive surgery.1,2 The analysis of corneal mechanics (i.e. biomechanics) is also crucial to improve the diagnosis and management of ectatic corneal disorders, such as keratoconus.1,2 The Popcorn project has led to the development of a dynamical topographer suitable for 3D mapping the deformation of the corneal surface when subjected to a controlled force. The 3D imaging system is based on plenoptic imaging, with a corneal analysis not limited to a specific section of the cornea. This emerging technique can overcome the limitations of current topographers that are unable to dynamically measuring the deformation of the cornea as changes of the elevation maps. With plenoptic technique, the entire surface of cornea can be analyzed in real-time, avoiding the need of scanning the surface. Specifically, the developed system uses Focused Plenoptic Imaging, in which a pattern is reflected by the cornea and it is focused on an intermediate plane just before of a micro-lens array. 3D information of the cornea is coded in a plenoptic image an array of multiples perspectives of the pattern, each one recorded with the corresponding micro-lens (Figure 3). Eventually, the elevation of the cornea is reconstructed by triangulating equivalent points between different views (Figure 3).

Advanced Biomechanical Model
Data analysis and patient-specific corneal modeling have been carried out with advanced computational techniques (i.e. finite element, FE, method, artificial neural networks and mathematical optimization). The in-silico model takes into account the non-linear behavior of the material, being able to reproduce large deformations as well as the anisotropy of the material due to the preferential orientation of the collagen fibers. Our simulations outline that an interplay between the corneal response to an air jet, the geometry, the intraocular pressure (IOP) and the corneal tissue properties exists.3 A ‘soft’ cornea coupled with a high IOP may present the same corneal response that a ‘stiff’ cornea with a low IOP, when an air load is applied onto its surface (Figure 4).3 As the cornea is bending during the air-puff, the anterior surface works in compression whereas the posterior surface works in tension. Hence, collagen fibers located closer to the anterior surface do not contribute to load bearing. Moreover, the diurnal variations in IOP could have also an influence on this mechanical response.4 The maximum variability of the corneal displacement related to diurnal changes in pachymetry (i.e. Central Corneal Thickness, CCT) and IOP was found to reach 5%.4
Since when the cornea is measured by a topographer is yet pre-stressed (i.e. is naturally deformed), the impact of neglecting the physiological pre-stress due to the IOP has been also evaluated.5 An iterative free-stress algorithm including a fiber’s pull-back has been applied to incorporate the pre-stress field to the model. Finally, the simulation of a general non-commercial non-contact tonometry diagnostic test (air-puff) over a large set of 130 patients (53 healthy, 63 keratoconic and 14 post-LASIK surgery eyes) has been addressed.5 The results have confirmed the influence of the CCT, the IOP and the fibers stiffness on the numerical corneal displacement (i.e. they define a 87% of the numerical response using a Pareto analysis), being in good agreement with clinical results. Likewise, these simulations have also confirmed the importance of considering the corneal pre-stress in the FE analysis (Figure 5).5 Specifically, the inclusion of the free-stress configuration of the eyeball results on a stiffening of the mechanical corneal response (Figure 5.A) i.e. the obtained displacements are smaller than those obtained with a numerical model not accounting for the pre-stress.
This advanced biomechanical model is used for the analysis of the corneal deformation captured and analyzed by the plenoptic imaging system. To obtain a real-time clinical application, machine learning algorithms have been used for both the patient classification and the material parameter estimation.

Data storing
A protocol for 3Dnet remote processing has been also developed and planned, allowing for a more efficient use of the data provided by the Popcorn system. Additionally, it skips the necessity of a powerful computer in clinic that would significantly increase the manufacturing cost of the system. All the information is stored, analyzed and processed in the cloud. As a final step, the clinic information will train the machine-learning tool, refining its decisions concerning the classification and the type of biomechanical behavior present in each cornea.

Potential Impact:
The expected final result is the commercialization of the device designed for evaluating the corneal biomechanical properties (material behavior of corneal tissue) in clinical practice. As commented, it is based on the plenoptic imaging system and the finite element patient specific eyeball model developed in the current project. It should be considered that a better understanding of corneal mechanic behavior allows the surgeon to improve the predictability of laser refractive surgery outcomes and the efficacy in the identification of eyes at risk of developing ectasia after refractive surgery. Likewise, obtaining clinical information on corneal biomechanics optimize clinical decisions since the clinician is provided with more information about the mechanic stability of the corneal tissue, which is essential to obtain a better understanding about the etiology of corneal ectasia. In addition, more help for a better diagnosis of some corneal diseases is also obtained, the detection of corneal diseases at incipient stages to avoid and prevent from their evolution and worsening and, finally, providing more clinically useful information in glaucoma, such as consistent compensation factors for IOP measurements. Nowadays, there is no method or system determining the biomechanical behavior of the cornea that would allow the clinician to better estimate the corneal response to a laser-based refractive surgical procedure, avoiding those with potential bad long-term effects or reorienting some cases to other treatment options. Therefore, the corneal examination with the POPCORN system can become a widely used device in clinical practice, with many users worlwide. The POPCORN system will allow careful evaluation of a patient before a laser based surgical procedure, avoiding future poor long-term effects and allowing to reduce by 90% the percentage of retreatments after refractive surgery. The biomechanical response of the cornea to laser treatments will be predicted in advance, allowing a better surgery planning. It must be considered that the global market for refractive surgery devices was valued at $666.5m in 2010, and it is forecast to grow at a Compound Annual Growth Rate (CAGR) of 3% to reach $805.2m in 2017. The previous evaluation of the biomechanical properties of the cornea before refractive surgery will also allow to detect corneas with apparently normal topography patterns but a significant mechanical weakening compatible with the presence of subclinical ectasia. In such cases, the refractive surgery must be avoided as it will accelerate the progress of the disease. Therefore, the POPCORN device is also a crucial tool for prevention and early diagnosis of corneal ectasia. Finally, the implantation of intracorneal ring segments in keratoconus or other corneal ectasias to reduce the level of corneal irregularity could be improved with a system providing information about the mechanical properties of the corneal structure. Currently, these implants are selected according to empiric nomograms without a consistent scientific basis. The POPCORN system may be also be useful for the implementation of this type of implants.
Concerning dissemination activities, many of them have been performed during the project:
-Press releases in local newspapers of the countries of each participant of the project (Spain, Italy, Romania, and United Kingdom). Some examples: Información, Heraldo de Aragón, Europa Sur, El Mundo, El Periódico de Aragón...
-Web page: This web page has been continuously updated with news and the inclusion of innovative content
-Congresses: participation i several national and international congresses
*XXXIV Congress of the ESCRS. 2016. Copenhaguen (Denmark): "New methodology for characterization of corneal mechanical properties: the POPCORN European project." (Oral presentation)
* 31 Congreso de la Sociedad Española de Cirugía Ocular Implanto Refractiva (SECOIR). 2016. Murcia (Spain). "Influencia de la presión intraocular, espesor y curvatura corneal en la medida clínica con dispositivos de pulso de aire de las propiedades biomecánicas de la córnea." (Oral presentation)
* 91 Congreso de la Sociedad Española de Oftalmología (SEO). 2015. Seville (Spain). "Tecnologías actuales para evaluar la biomecánica corneal." (Oral presentation)
*3rd Congress of the European Academy of Orthokeratology (eurOK). 2015. Budapest (Hungary). "Implications of corneal biomechanics in orthokeratology." (Oral presentation)
*21th Congress of the European Society of Biomechanics. 2015. Prague (Czech Republic). "Automatized patient-specific corneal modeling. Application to simulation of a general non-contact tonometry test." (Oral presentation)
*XXXIII Congress of the ESCRS. 2015. Barcelona (Spain) "Could corneal biomechanics diurnal variations be affecting corneal response to an air puff in healthy eyes: preliminary evidence." (Poster)
*2º Congreso Internacional Online de Jóvenes Optometristas SIYO. 2014. Valencia (Spain). "Avances en la caracterización clínica de las propiedades biomecánicas de la córnea." (Oral presentation)
*Vision Sciences and Eye Research Meeting (VISER). 2014. Santiago de Compostela (Spain). "Avances en la modelización y caracterización de la biomecánica de la córnea en el ámbito clínico." (Oral presentation)
*I Jornadas Clínicas de Ortoqueratología Nocturna. 2014. Madrid (Spain). "Biomecánica corneal." (Oral presentation)
*III Jornadas Optoinnova. 2014. Alicante (Spain). "Avances en la caracterización clínica de la biomecánica corneal." (Oral presentation)
*XXIII Congreso Internacional de Optometría, Contactología y Óptica Oftálmica (Optom 2014). 2014. Madrid (Spain). "Evaluación de la biomecánica corneal: avances recientes en técnicas de medida." (Oral presentation)
*XXXII Congress of the ESCRS. 2014. London (United Kingdom)."Numerical approach for interpreting non-contact tonometer results in terms of corneal biomechanical behaviour." (Poster)
*29 Congreso SECOIR. 2014. Alicante (Spain). "Interpretación de la biomecánica corneal a partir de los resultados proporcionados por el sistema Corvis ST." (Poster)
-Scientific publications in peer-reviewed journals:
*Ariza-Gracia MA, Zurita J, Piñero DP, Calvo B, Rodríguez-Matas JF. Automatized patient-specific methodology for numerical determination of biomechanical corneal response. Ann Biomed Eng 2016; 44: 1753-72.
*Ariza-Gracia MÁ, Zurita JF, Piñero DP, Rodriguez-Matas JF, Calvo B. Coupled biomechanical response of the cornea assessed by non-contact tonometry. A simulation study. PLoS One 2015; 10(3): e0121486.
*Ariza-Gracia MA, Piñero DP, Rodríguez JF, Pérez-Cambrodí RJ, Calvo B. Interaction between diurnal variations of intraocular pressure, pachymetry, and corneal response to an air puff: preliminary evidence. JCRS Online Case Reports 2015; 3: 12-5.
*Piñero DP, Alcon N. Corneal biomechanics: a review. Clin Exp Optom 2015; 98: 107-16.
*Piñero DP, Alcon N. In vivo characterization of corneal biomechanics. J Cataract Refract Surg 2014; 40: 870-87.

Regarding the exploitation of the results, this is the final scheme stablished:
• Oftalmar will retain the ownership of the overall assembled product through a patent that will transfer to the consortium for exploitation following a royalty scheme. It will be the first ophthalmic clinic in having the new system which will provide it with a competitive advantage in terms of costs saving, service and prestige. Oftalmar will receive the 1% of the total sales in a royalty concept from CSO.
• CSO will retain the distribution rights of the equipment in Europe, USA, China and India which will maintain its leadership as a very advanced provider winning prestige in front of the large companies that operate in the ophthalmological field. In addition CSO will share with B3D the ownership of the plenoptic sensor which can apply to other products in a different market field.
• Optoelectronica 2001 will retain the rights for manufacturing the new system increasing its business and specialised knowledge by means of assembling very complex optoelectronic and mechanical components.
• Biotronics 3D will retain the ownership of the copyright for the images software processing that could license through its netCloud service, specialising its knowledge in a very high-tech application which will increase its business and market position. In addition B3D will share with CSO the ownership of the plenoptic sensor which can apply to other products in a different market field.

List of Websites:

Relevant contacts: Riccardo Nicoletti, David P Piñero, Begoña Calvo