The United States Patent and Trademark Office (USPTO) has awarded iHealthScreen Inc. the full U.S. patent for its iPredict glaucoma detection model.
Let’s start with iHealthScreen.
Launched in 2015, iHealthScreen is developing novel, innovative software for eye- and systemic-disease screening and prediction services.
Ophthalmic focus: The company is targeting retinal image grading and early-stage screening and diagnosis of ocular diseases such as age-related macular degeneration (AMD), glaucoma, and diabetic retinopathy (DR).
Now the iPredict system.
The screening system was initially designed to provide fully automated screening for AMD via retinal imaging for instant results in a clinical setting.
For testing: High-resolution images of the eye with a color fundus camera are captured and submitted to the iPredict system; screening results are available via a fully automated report in under 1 minute.
According to the company, the entire test can be completed within 5 minutes.
What happens if disease is detected?
If the system detects a referable disease, the report provides an ophthalmologist visit recommendation for treatment; if no disease is detected, a 1-year follow-up visit is recommended.
And its clinical status?
The iPredict was granted CE certification in 2021 for the detection of mild and non-proliferative DR, diabetic macular edema (DME), and glaucoma suspects based on abnormal optic discs and in patients over the age of 50 for AMD.
On the global front: The system was approved in 2022 by the Australian Health Therapeutics Goods Administration (FDA equivalent).
Didn’t the company seek FDA clearance as well?
It did! Back in May 2023, iHealthScreen submitted an application for 510(k) clearance to the FDA for the system for AMD—making it the first company to seek FDA clearance in AMD screening.
See our coverage here.
Now talk about this new patent.
Officially awarded on March 26, 2024, the full U.S. patent for glaucoma detection (#11,941,809 B1) is titled “Glaucoma Detection and Early Diagnosis By Combined Machine Learning Based Risk Score Generation and Feature Optimization.”What it covers: for the iPredict to provide “a fully automated detection report for glaucoma, which can be implemented in the primary care settings for screening and detection of early-stage glaucoma,” according to the company.
iHealthScreen notes that such glaucoma detection has an overall accuracy of 94.3%.
And is this its first patent?
Nope! In 2022, the company was awarded the “Late AMD Screening and Prediction Model” patent for AMD, making it the first to have two U.S. patents for AMD and glaucoma screening/prediction.
Prior to these, the company was granted two patent applications by the USPTO:
- Patent #US11416987B2: Image-based screening system for prediction of individuals at risk of late age-related macular degeneration (AMD)
- Published: 2022
- Patent#US20190014982A1: Automated blood vessel feature detection and quantification for retinal image grading and disease screening
- Published: 2019 (status is currently “abandoned”)
So … is there any clinical data on the system?
A prospective trial (NCT05324189) evaluated the use of the screening system among the general population (n = 1,000; 18+ years of age) for accuracy, sensitivity, and specificity in early diagnosis for DR.
While the trial is slated to conclude in December 2024, results thus far have found that iPredict demonstrated 86.9% sensitivity and 94.1% specificity.
Anything else?
Yes! The iPredict is the subject of an upcoming presentation at the Association for Research in Vision and Ophthalmology (ARVO) 2024 annual meeting.
The poster session, titled Glaucoma Suspects Detection by Combined Machine Learning-Based Risk Score Generation and Feature Optimization is being presented by iHealthScreen CEO Alauddin Bhuiyan, PhD.
Its purpose: to asses the iPredict system for detecting early glaucoma suspects via retinal cup-disc ratio (CDR), disc hemorrhages, and peripapillary atrophy.
What are the findings?
Per the presentation’s abstract, the investigators found that:
- For CDR classification:
- The model categorized images into three classes with a weighted kappa of 0.785.
- Disc hemorrhage detection achieved:
- Accuracy of 93.13% (95% confidence interval [CI]: 91.35% to 94.64%), sensitivity of 71.53% (95% CI: 63.42% to 78.73%), and specificity of 96.87% (95% CI: 95.45% to 97.95%).
- Peripapillary atrophy model exhibited:
- Sensitivity of 93.18% (95% CI: 81.34% to 98.57%) and specificity of 97.67% (95% CI: 87.71% to 99.94%).
And overall?
The abstract noted that the combined model detected glaucoma with:
- An accuracy of 94.3% (95% CI: 89.05% to 97.50%)
- A sensitivity of 95.0% (95% CI: 83.08% to 99.39%)
- A specificity of 94.0% (95% CI: 87.40% to 97.77%)
Conclusion: “The high accuracy and robustness of the individual models demonstrate the potential of this system in early glaucoma detection,” according to the abstract, “which could significantly impact public health by enabling timely interventions and prevention of glaucoma.”
Click here to search for and view this abstract.