Kalyani B, M. Optom

Assistant Professor, Dhanalakshmi Srinivasan University,Tiruchirappalli, India

 

Retinopathy of Prematurity

Retinopathy of prematurity (ROP) is an eye ailment that can affect premature neonates. It causes abnormal blood vessels to grow in the retina, which leads to blindness. Babies born before 31 weeks or weight less than 1500 g are most at risk.(1) ROP is classified according to four anatomical zones, five stages of disease severity, and the presence of plus disease, a posterior retinal vascular biomarker indicative of disease progression. (2) Other possible risk factors of ROP include:

  • Anaemia
  • Infection
  • Transfusions
  • Breathing difficulties
  • Heart disease
  • Ethnicity

Early treatment has demonstrated significant benefits, making screening and regular monitoring essential for optimal disease management.

Artificial Intelligence

Artificial intelligence (AI) is the technology that enables machines to mimic human intelligence and perform tasks like learning, reasoning, and problem-solving. AI aims to improve efficiency, accuracy, and capabilities across various fields, making it a transformative technology in today’s world. (3)

AI in ROP

The role of AI in ROP has significantly expanded, with various tools and technologies being developed and applied in this field. (4,5) Some notable developments include:

  1. ROP tool (FocusROP): This software aids in detecting and classifying plus disease in ROP by analysing retina images.
  2. Retinal Image Multiscale Analysis (RISA): This tool provides a detailed analysis of retinal images, helping to identify and quantify abnormalities.
  3. Vessel Map: This technology maps the retinal blood vessels, crucial for diagnosing and monitoring ROP.
  4. Computer-Assisted Image Analysis of the Retina (CAIAR): CAIAR offers automated analysis of retinal images, improving the accuracy and efficiency of ROP diagnosis.
  5. Imaging and Informatics in ROP (i-ROP): i-ROP leverages machine learning algorithms to analyse retinal images and provide diagnostic support.
  6. i-ROP DL: This is a deep learning-based extension of i-ROP, enhancing the ability to detect and classify ROP more accurately.

In these developments, i-ROP DL was the deep learning system that showed significant promise in detecting disease, thereby enhancing the reliability of ROP diagnosis. Computer-based image Analysis (CBIA) systems have focused on two classification levels: plus, versus not plus but in i-ROP DL must be classified into plus, pre-plus, and normal. The i-ROP deep learning algorithm generates a continuous vascular severity score from 1 to 9. It categorises images according to severity levels: no ROP, mild ROP, type 2 ROP, pre-plus disease, and type 1 ROP which is easily diagnosed early. The i-ROP system has proven to be highly effective, achieving an impressive 95% accuracy in the diagnosis of ROP. (6)

Conclusion

AI has made it possible to create an ROP severity score that is consistent with the International Classification of ROP (ICROP) illness categorization. This development can enhance quantitative disease monitoring, improve risk prediction, and identify treatment failures and recurrences post-treatment. Comparable to human evaluations, AI Algorithm systems can significantly reduce the workload of specialised ophthalmologists, potentially reducing the screening burden by up to 80%.(7)

 

References

  1. Gensure RH, Chiang MF, Campbell JP. Artificial intelligence for retinopathy of prematurity. Current opinion in ophthalmology. 2020 Sep 1;31(5):312-7.
  2. Chawla D, Agarwal R, Deorari AK, Paul VK. Retinopathy of prematurity. The Indian Journal of Pediatrics. 2008 Jan;75:73-6.
  3. Maitra P, Shah PK, Campbell PJ, Rishi P. The scope of artificial intelligence in retinopathy of prematurity (ROP) management. Indian Journal of Ophthalmology. 2024 Jul 1;72(7):931-4.
  4. Desideri LF, Rutigliani C, Corazza P, Nastasi A, Roda M, Nicolo M, Traverso CE, Vagge A. The upcoming role of Artificial Intelligence (AI) for retinal and glaucomatous diseases. Journal of Optometry. 2022 Jan 1;15:S50-7.
  5. Shah S, Slaney E, VerHage E, Chen J, Dias R, Abdelmalik B, Weaver A, Neu J. Application of artificial intelligence in the early detection of retinopathy of prematurity: review of the literature. Neonatology. 2023 Oct 2;120(5):558-65.
  6. Scruggs BA, Chan RP, Kalpathy-Cramer J, Chiang MF, Campbell JP. Artificial intelligence in retinopathy of prematurity diagnosis. Translational vision science & technology. 2020 Jan 28;9(2):5.
  7. Eilts SK, Pfeil JM, Poschkamp B, Krohne TU, Eter N, Barth T, Guthoff R, Lagrèze W, Grundel M, Bründer MC, Busch M. Assessment of retinopathy of prematurity regression and reactivation using an artificial intelligence–based vascular severity score. JAMA Network Open. 2023 Jan 3;6(1):e2251512.