Algorithm helps ophthalmologists decide best treatment option for patients based on data
Algorithm-based software can help ophthalmologists choose the best keratoconus treatment option for their patients based on patients’ unique needs.
By Mazen M. Sinjab, MD, PhD, Special to Ophthalmology Times
Damascus, Syria-When treating ectatic corneal disorders, such as keratoconus, ophthalmologists have a plethora of management options from which to choose. However, choosing among these treatment modalities is not easy, because different patient parameters (i.e., age, gender, corneal astigmatism, and visual acuity) are often indicative of different treatment options.
To address this, I developed an algorithm-based software (Sinjab Data-Based Algorithm for Managing Keratoconus, or DamaSKo-Gram [the software name reflects the city where it was developed, Damascus, Syria]) that can help ophthalmologists decide the best management option for their patients based on patient data. In this article, two case studies demonstrate how this software functions and how it can help improve keratoconus management.
A 42-year-old male patient was referred for keratoconus in his right eye. Upon presentation, he had a clear corneal surface and stable disease. Uncorrected distance visual acuity (UDVA) was counting fingers and best-corrected visual acuity (BCVA) was 0.7 D. He had a manifest refractive error of –3 D Sph/–4 D Cyl ´ 35 axis and a maximum K reading (Kmax) of 58 D. The thinnest location of his cornea was 459 mm and he had topographic astigmatism of 4 D, consistent with his refractive cylinder. To decide the best treatment option for this patient, I entered this information into the patient information window of the software.
With a simple click on the “management” button, the software analyzed the patient’s data and suggested four treatment options customized for this patient, listed in the order of most favorable to least favorable:
To help decide among these options, the software also provides a “discussion” that highlights the conditions under which one treatment option would be better than the other. For instance, in this case, the discussion suggested using contact lenses as the first option based on the patient’s tolerance due to the high refractive error and K readings. It also highlighted the fact that the patient’s parameters made him suitable for ICR implantation, CXL, and TG-PRK, with TG-PRK being preferable to CXL due to high astigmatism.
The discussion also warned that the patient’s high Kmax value increased the possibility of CXL failure. The last treatment option of using spectacles was suggested because of the good CDVA and CDVA-UDVA difference; however, it was discouraged due to high refractive error.
In this case, since the patient was intolerant to contact lenses, I chose the second recommendation and implanted two ICR segments (Keraring SI5, Mediphacos). Three months postoperatively, the patient’s UDVA was 0.7 D and CDVA was 1 D with –0.75 D Sph\–0.5 D Cyl ´ 43 axis. The patient was very satisfied with the results, even without glasses.
A 25-year-old female presented with a diagnosis of keratoconus in her left eye. She had previously had penetrating keratoplasty (PKP) for the same condition, but in her right eye. When she presented, she had progressive disease in her left eye and had a clear corneal surface, with UDVA of counting finger and a CDVA of 0.4 D. She had a high manifest refraction of –4 D Sph/–7 D Cyl ´ 160 axis, Kmax of 61 D, and topographical astigmatism of 8 D. The area of least corneal thickness in this patient was 420 mm. Based on this information, which I entered in the patient information window, the software recommended the following options starting with the most preferred option:
These recommendations were accompanied by a discussion as in the previous case to explain the choice of treatment options proposed and how they are suited to the patient’s parameters. In this case, I proceeded with option 2. Although the patient was tolerant to contact lenses, she had been using them for several years already and was seeking a more radical solution, hence excluding option 1.
PKP was also not a viable option in this case because of the patient’s experience with the graft in her right eye. Although she had good visual acuity in this eye, the graft embarrassed her because it was visible to close friends. In choosing between ICR and Myoring implantation, I prefer the Keraring because it gives more predictable results, especially with low sphere. Furthermore, the software predicted good prognosis with Keraring based on the patient’s topography pattern. Therefore, the patient and I chose to proceed with Keraring implantation.
An additional advantage of the software is that it provides guidelines for the different treatment options suggested based on the patient’s parameters. Therefore, I consulted the software to determine the best size of the Keraring for this patient. Due to the high topographic astigmatism in this case, the software recommended using a 6-mm Keraring instead of a 5-mm Keraring. Three months postoperatively, the patient’s UDVA improved to 0.2++, CDVA improved to 0.9 D, and refraction was –2 D Sph/–1 D Cyl ´ 150.
With such a significant improvement, we were able to avoid PIOL implantation and proceed directly to CXL. Three months post-CXL, we noticed an additional improvement in vision of UDVA 0.4 D, CDVA 1 D, refraction –1 D Sph/–0.5 D Cyl. The patient and I were both very satisfied with the results.
Therefore, the software is able to adapt to different cases and suggest treatment options based on the severity of the patient’s parameters.
Contact lenses and spectacles have been the mainstay of keratoconus treatment for most patients. However, there have been several recent developments in other surgical and non-surgical treatment options for keratoconus treatment, including CXL, PRK, ICRs, the Myoring, PIOLs, and deep anterior lamellar keratoplasty. No doubt, future innovations will only add to this list.
Although the availability of a vast number of treatment options is good news for the ophthalmologist and the patient, it also makes the ophthalmologist’s task of finding the optimal treatment modality more challenging. This is especially due to the fact that the choice of treatment modality depends on the patient’s visual and demographic parameters, which do not always suggest the same therapy. Although different parameters are indicative of different treatment options, in general, we know that higher Kmax, thinner cornea, higher refractive error, and lower visual acuity are indications for more aggressive therapy.
The software uses a data-based algorithm that takes into account the various parameters of the patient in order to decide the best treatment options for him/her, thus simplifying the ophthalmologist’s task. A simple user-friendly interface lets the ophthalmologist to enter the patient’s information into the software-information that can be stored long term with the patient’s record and made accessible at future visits.
The software can also detect errors and caution the ophthalmologist to them. For instance, if the ophthalmologist forgets to input a necessary parameter, such as disease stability, the software generates a warning to enter the missing information. Similarly, if there is a logical discrepancy in the entered data, the software picks up on the error and intimates the doctor to correct it. For example, if the ophthalmologist enters a high refractive error into the software but also a high UDVA, the software generates an error message asking the physician to re-check the patient.
An additional advantage of the software, as demonstrated in the second case study, is that it provides the ophthalmologist with general guidelines and topographic patterns suitable for the use of either a single-treatment modality, such as the Myoring or ICRs, or for combination therapy such as with CXL and TG-PRK. Thus, once the ophthalmologist has chosen a preferred treatment option, he or she can consult the software for information on how best to execute the treatment.
Furthermore, this software is easy to install and suitable for use with any topographer. It can also be continually upgraded as new information becomes available and can even be improved to give measurements, locations, and types of ICRs and PIOLs.
The software is in a sense similar to consulting with a colleague, with the added advantage of the software being available at the ophthalmologist’s fingertips. Therefore, the software can help ophthalmologists navigate through the vast amount of information, indications, contraindications, and guidelines for the different keratoconus treatments available in order to choose the best treatment option for their patients based on their patients’ unique needs.
Mazen M. Sinjab, MD, PhD, is an assistant professor at Damascus University, Damascus, Syria. Dr. Sinjab did not indicate a financial interest in the products or companies mentioned. The software is not approved in the United States. Readers may contact Dr. Sinjab via e-mail at email@example.com.