Phone-based AI Diagnostic Solution for Retinal Diseases

Diabetic Retinopathy (DR) is a retinal disease that affects millions of people worldwide, caused by high blood sugar damaging the blood vessels of the retina, leading to irreversible blindness. With more than 500 million people diagnosed with diabetes globally, approximately 30% of them are developing DR. Unfortunately, DR is one of the most common causes of blindness, with over 90% of patients residing in developing countries. Early screening and treatment can stop the progression of DR and prevent blindness. However, patients in developing countries cannot get early screening due to a lack of affordable screening technology, limited access for patients, and a shortage of healthcare professionals to perform the screening. Founded by three engineers from developing countries, Ocellux understands the severity of this issue and aims to provide affordable screening solutions to help prevent blindness caused by DR and other retinal diseases.

Ocellux's first pilot market is India, which has over 15 million DR patients, and about 60% have no access to eye screening.

Ocellux's first product (a patent-pending technology), E-Carebetics, enables healthcare workers to conduct diabetic retinopathy screenings using a smartphone-attached ophthalmoscope and a mobile app. The scope enables the smartphone to take clear retinal images using the camera on the phone, and the AI app analyzes the images and provides diagnostic data to assist medical workers’ diagnosis. E-Carebetics makes the most widely used device, the smartphone, into a low-cost and portable medical device, enabling rapid screening in and out of the office setting. The AI model has achieved 96% accuracy in detecting the disease and includes an image enhancement feature for less optimal camera. The app also includes an annotation model to mark the location of the disease and provide a detailed report for healthcare workers to make diagnostic decisions. The only step required from the medical worker is to point the camera at the patient's eye and evaluate the AI-generated report.

We serve healthcare providers in the public and private sectors, including hospitals and NGOs. Our interviews revealed that there is a significant need for a low-cost solution that enables medical workers to conduct retinal screening. The current products available are highly expensive and medical workers in developing countries cannot afford or access them.

The current method for retinal screening, which uses direct and indirect ophthalmoscopes, is not only cumbersome but also inaccurate and cannot be digitized. Moreover, retinal screening requires extensive training and only specialized eye doctors can conduct the screening. Many endocrinologists, neurologists, and nurses expressed their need in having a device that could assist them with diagnosis, given their lack of training in retinal disease diagnosis. Otherwise, they would have to refer patients to eye doctors, which increases the barrier for patients to access the screening.

Ocellux was founded by three engineers, Bingda Li, Devansh Parikh, and Mukunthan Tharmakulasingam, who bring a wealth of experience and knowledge to the table. Bingda Li is a dual Master of Science degree holder in Biomedical Engineering and Engineering & Technology Innovation Management. He is currently pursuing a Ph.D. in Biomedical Engineering at Carnegie Mellon University, where he has gained extensive experience in medical device design, research and development, and business strategy. With his deep understanding of the technical and commercial aspects of the healthcare industry, Bingda plays a critical role in driving Ocellux's growth and success.

Devansh Parikh holds a bachelor's degree in Electrical Engineering and specializes in product design and development. Based in India, Devansh manages the local operation of Ocellux and connecting with healthcare providers, supply chains, and patients.

Mukunthan Tharmakulasingam is a Ph.D. holder in Machine Learning and has years of experience in software development, data analysis, and technology management, specifically in digital healthcare and AI.

Together, the team is well-equipped to tackle medtech innovation and business development. With a deep understanding of the healthcare industry and a passion for developing innovative solutions to address pressing healthcare needs, Ocellux is poised to make a significant impact in the fight against blindness caused by retinal diseases.

We serve all diabetic patients to help them screen and prevent the development of diabetic retinopathy. Ocellux's first pilot market is India, which has over 80 million diabetic patients. Since over 60% of the patients don't have access to eye care, we serve at least 48 million diabetic patients.

- Financial Support. We would like to raise more funding for scaling up the clinical trials and testing. 

- Technical Support. We would like to work with experts to optimize our technology to provide the best possible outcomes.

- Legal. We would like to consult experts on how to connect our two headquarters (US and India) together and solve potential tax and regulatory issues. 

- Network: We would like to connect with more healthcare providers and experts to increase our visibility and learn from the experts. 

Ocellux's first product, E-Carebetics, enables all healthcare workers, not just eye doctors, to conduct diabetic retinopathy screenings using a smartphone-attached ophthalmoscope and an AI diagnostics mobile app. The AI app analyzes images taken by the ophthalmoscope and provides diagnostic data to assist medical workers' diagnosis. The device is low-cost and portable, making it ideal for rapid screenings in and out of the office. The AI model has 96% accuracy in detecting the disease and includes an image enhancement feature and annotation model to mark the location of the disease and provide a detailed report for healthcare workers. Traditional alternatives, such as desktop fundus cameras, are costly and not portable, making them poorly adopted in developing countries. Indirect and direct ophthalmoscopes also require extensive training and experience to operate and cannot digitize the diagnosis. E-Carebetics' closest competition is Remidio, which is a phone-attached fundus camera with AI diagnostic capabilities, but E-Carebetics has a higher field of view and is more durable and compatible with more smartphones. Ocellux's E-Carebetics has three main advantages over its competitors: firstly, the AI model is trained on a proprietary dataset specific to certain ethnicities, which is not fully represented in the current publicly available datasets from Western countries. Our proprietary dataset allows our AI model to be trained continuously, resulting in the ongoing enhancement of diagnostic accuracy. Secondly, it offers two options for customers with different needs, E-Carebetics Basic and Pro. The Basic device will be provided to the customers for free upon the first software subscription. The Pro device is a premium version that offers wireless and non-mydriaticHD screening. Lastly, both devices have increased the field of view (FOV) by using our novel illumination technology based on optic fiber. This technology is under patenting, and it can significantly reduce the cost of our hardware to provide a more competitive advantage for us. The cost is critical in determining willingness to pay, as healthcare providers in developing countries are highly price sensitive.

Our impact goals for the next year and the next five years focus on expanding our reach, improving access to retinal screenings, and reducing the time to receive treatment and care for patients with diabetic retinopathy. Here's a more detailed explanation:

The first is to obtain certification for our ABS plastic non-electrical smartphone ophthalmoscope. This certification ensures the quality and safety of our product, making it ready for deployment in healthcare settings. We will specifically target Public Health Centers (PHC) and Community Health Centers (CHC) in tier 2 and tier 3 cities of India.

To achieve this goal, we doing the following:

1. Work closely with regulatory authorities to ensure compliance and obtain the necessary certifications.
2. Conduct rigorous testing and validation to demonstrate the effectiveness and reliability of our ophthalmoscope.
3. Collaborate with medical professionals, including diabetologists, ophthalmologists, neurologists, and traveling nurses, to train them on using our product for retinal screenings.
4. Establish partnerships with PHCs and CHCs, providing them with the necessary infrastructure and support to integrate our solution into their existing healthcare systems.

By deploying our ophthalmoscope in these settings, we aim to empower medical professionals in remote locations to conduct more retinal screenings and identify patients with the severity and progression of diabetic retinopathy.

Over the next five years, our impact goals are as follows:

1. Scaling Deployment: Expand the deployment of our ophthalmoscope to a wider network of healthcare facilities, including government hospitals, clinics, and private healthcare providers across India. This will enable us to reach more patients and increase the accessibility of retinal screenings.

2. Training and Education: Develop comprehensive training programs for healthcare professionals on the use of our ophthalmoscope and AI software. This will enhance their skills and knowledge in diagnosing and managing diabetic retinopathy, thereby ensuring accurate screenings and timely intervention.

3. Partnerships and Collaborations: Establish strategic partnerships with healthcare organizations, non-profit entities, and government agencies to leverage their networks and resources. By collaborating with these stakeholders, we can amplify our impact and reach underserved populations in remote and rural areas.

4. Data Analysis and Research: Continuously analyze the data collected from retinal screenings to identify patterns, trends, and risk factors associated with diabetic retinopathy. This research will contribute to improving our AI algorithms and enhancing the accuracy of our diagnostic reports, enabling better patient outcomes.

5. Advocacy and Awareness: Actively engage in advocacy efforts to raise awareness about the importance of regular retinal screenings for individuals with diabetes. By partnering with patient advocacy groups and leveraging digital platforms, we will educate the public and healthcare providers about the benefits of early detection and timely treatment.

To measure our progress toward our impact goals, we are using several specific indicators. These indicators help us track the effectiveness of our solution and its impact on improving access to retinal screenings and reducing the time to receive treatment and care for patients with diabetic retinopathy. Here are a few examples:

1. The number of ophthalmoscopes deployed: We track the number of ophthalmoscopes deployed in healthcare facilities to measure the expansion of our solution's reach. This indicator allows us to assess our penetration into target regions and populations.

2. The number of retinal screenings conducted: We measure the number of retinal screenings conducted using our ophthalmoscope and AI software. This indicator helps us evaluate the uptake of our technology and the extent of its utilization in identifying cases of diabetic retinopathy.

3. Percentage increase in early detection: We track the percentage increase in the early detection of diabetic retinopathy cases using our solution compared to traditional methods. This indicator assesses the effectiveness of our technology in identifying retinal diseases at an early stage.

4. Time to treatment: We measure the time it takes for patients identified with diabetic retinopathy to receive treatment and care. This indicator helps us evaluate the impact of our solution in reducing the time gap between diagnosis and intervention, leading to better patient outcomes.

5. Patient satisfaction surveys: We collect feedback from patients who have undergone retinal screenings using our solution to assess their satisfaction levels and overall experience. This qualitative indicator helps us understand the user perspective and make improvements to enhance patient outcomes and satisfaction.

6. Partnerships and collaborations: We track the number of partnerships established with healthcare organizations, non-profit entities, and government agencies. This indicator reflects our ability to leverage external networks and resources to amplify our impact and reach underserved populations.

By continuously monitoring and analyzing these indicators, we can assess our progress toward our impact goals, identify areas for improvement, and make data-driven decisions to maximize our impact in preventing blindness caused by diabetic retinopathy and improving access to retinal screenings.

Our theory of change is based on the belief that by providing an innovative AI-based diagnostic solution for retinal screenings, we can have a significant impact on improving access to eye care and reducing the burden of diabetic retinopathy. The logical framework of our theory of change is as follows:

Activities:

1. Develop and deploy our smartphone-attached ophthalmoscope and AI software.
2. Conduct rigorous testing, validation, and certification processes to ensure quality and safety.
3. Train healthcare professionals on the use of our technology for retinal screenings.
4. Establish partnerships with healthcare centers and organizations for implementation.

Outputs:

1. Availability of a low-cost, user-friendly ophthalmoscope that attaches to smartphones.
2. Accurate and automated AI software for retinal image analysis.
3. Trained healthcare professionals equipped with the necessary skills to conduct retinal screenings.
4. Integration of our solution into healthcare systems and facilities.

Outcomes:

1. Increased accessibility: By using our smartphone-attached ophthalmoscope, healthcare workers, including non-eye specialists, can conduct retinal screenings in various settings, including remote and underserved areas.
2. Early detection: Our AI software analyzes retinal images and produces comprehensive diagnostic reports, enabling early detection of diabetic retinopathy and other retinal diseases.
3. Timely intervention: By identifying cases of diabetic retinopathy early, healthcare professionals can provide timely treatment and care, preventing the progression of the disease and reducing the risk of vision loss.
4. Cost-effectiveness: Our solution utilizes existing technology (smartphones) and offers a low-cost alternative to traditional retinal screening tools, making it more affordable and accessible for healthcare providers and patients.

Evidence supporting our theory of change includes:

1. Existing research highlights the efficacy of retinal screenings for early detection of diabetic retinopathy.
2. Studies demonstrating the accuracy and reliability of AI-based image analysis for diagnosing retinal diseases.
3. Interviews and feedback from healthcare professionals indicate the need for accessible and cost-effective retinal screening solutions.

By linking our activities to immediate outputs and longer-term outcomes, our theory of change provides a logical framework for how we expect our solution to have a positive impact on the problem of limited access to retinal screenings and the burden of diabetic retinopathy.

The core technology that powers our solution is a combination of hardware and software, leveraging advancements in AI and deep learning algorithms. Our solution utilizes cutting-edge technologies to address the problem of diagnosing retinal defects and improve accessibility to AI-based diagnosis.

1. Robust Hardware: Our solution incorporates a smartphone-attached ophthalmoscope, which serves as the hardware component. This portable device attaches to a smartphone's camera, allowing healthcare professionals to capture high-quality retinal images. The ophthalmoscope is designed to be lightweight, easy to handle, and compatible with various smartphone models, ensuring convenience and robust usage in harsh environments.

2. AI Software: At the heart of our solution is our AI software, which employs a Deep Learning-based Convolutional Neural Network (CNN) model. This AI algorithm has been specifically trained to analyze retinal images and detect defects and abnormalities in the retina. By leveraging deep learning techniques, the software can recognize patterns and features indicative of retinal diseases, including diabetic retinopathy.

3. Explainable AI: Our AI software incorporates the LIME (Local Interpretable Model-Agnostic Explanations) technique. This feature provides explanations for the detected retinal defects and abnormalities, highlighting the specific regions within the retinal image that contribute to the diagnostic outcome. LIME explanations enhance transparency and interpretability, enabling healthcare professionals to understand and trust AI-generated diagnoses.

4. Cloud-Based and Mobile-Native Offline AI Software: Our solution offers both cloud-based and mobile-native offline AI software capabilities. The cloud-based infrastructure provides access to powerful computing resources, enabling efficient analysis of retinal images and generating diagnostic reports. Additionally, the mobile-native offline capability allows healthcare professionals to perform AI-based diagnosis directly on their smartphones, even in areas with limited internet connectivity. This ensures greater accessibility and convenience, particularly for remote populations.

By combining hardware innovation with advanced AI software, our solution provides healthcare professionals with a user-friendly and accessible tool for retinal diagnosis. The deep learning algorithms, specifically the CNN model, enable accurate detection of retinal defects and abnormalities. The incorporation of LIME explanation enhances transparency and trust in the diagnostic process. Furthermore, the cloud-based and mobile-native offline capabilities ensure that individuals in remote areas have access to AI-based diagnosis, overcoming geographical and connectivity barriers.

Our solution demonstrates how the application of technology, including AI, software development, and smartphone attachments, can address the challenges of retinal diagnosis. By harnessing the power of modern technology, we strive to benefit underserved communities by improving access to accurate and timely retinal screening, ultimately leading to better healthcare outcomes for individuals at risk of retinal diseases.

We have developed the prototype and tested it with the doctors in the clinical setting in India. The device can take clear images, and our AI model is able to analyze the image. 

Ocellux was founded by a team of three engineers who come from economically disadvantaged backgrounds and identify as people of color. As a company, we recognize the significance of integrating diversity, equity, and inclusivity in our work, and are dedicated to developing innovative solutions that promote these values in the healthcare industry. With a diverse leadership team, we plan to create inclusive hiring practices, provide training and education, encourage open communication, and establish partnerships with external organizations.

India, like many developing countries, has a predominantly rural population of about 70% with low-budget healthcare providers, and a relatively well-funded urban population of around 30%. Ocellux has differentiated its product, E-Carebetics, to serve both markets by offering a basic version and a premium version.

The basic version of E-Carebetics scope is provided at no cost to customers who subscribe to our AI software for the first time. This version enables customers to access our AI analytical tool by taking clear retinal images with their smartphone. The basic scope is cost-effective and easy to handle, but it requires mydriatic screening, meaning that the patient's eyes must be dilated before taking the image. Without damage, this scope can be used for many years since it is highly durable.

Our premium scope, E-Carebetics Pro, can be connected to a smartphone or controlled wirelessly. It offers non-mydriatic screening, eliminating the need for dilation, and has its own HD infrared camera for clear images. This device provides us with an almost 70% margin while being 90% less expensive than competing products.

By offering both basic and pro versions of E-Carebetics, Ocellux can serve customers with different needs. Our subscription-based AI software ranges from $200-$400 per year, depending on the package and usage.

Medical device procurement in India takes place through various channels, with the government procuring for public hospitals and health facilities through tenders and contracts, private hospitals and clinics purchasing through distributors or manufacturers, and individuals purchasing for personal use at retail stores or online. The market is highly regulated by the Central Drugs Standard Control Organization (CDSCO). The bidding process for procurement includes reviewing tenders, preparing and submitting bids, participating in pre-bid meetings, waiting for results, contract negotiation, and delivery and payment. Companies must comply with regulations, have a good track record, and offer competitive prices to increase their chances of winning. L1, L2, and L3 quotations refer to the ranking of bids based on price, with L1 being the lowest cost and most favorable bid, L2 being the backup option, and L3 being the secondary backup option. The final decision is based on an evaluation of all factors and the procurement committee's decision. As Ocellux’s E-Carebetics has a low cost while providing a high-tech digital solution, we have been indicated by the officials as being highly likely to be the most favorable bid (L1) along with its AI capabilities which provides a competitive advantage over the other players in the market. Eventually benefiting the healthcare providers to give their patients a better healthcare outcome.

Our target customers are mostly healthcare providers in developing countries, and India is our first target market. Our research has identified over 270,000 potential customers, which include public and private hospitals, community health centers, neighborhood clinics, and NGOs. Our estimated total annual addressable market is over 600 million USD, with approximately 480 million USD being our serviceable market.

To aid our entry into the market, we have secured a grant from the Indian government, which has facilitated access to public hospitals and healthcare providers. We have also collaborated with clinicians for product development and plan to conduct research using our device, with the goal of publishing scientific articles. We anticipate that these initiatives, along with government support, will allow us to capture at least 0.5% of the market in the first year after launching, generating 2.4 million USD in revenue.

We are in discussions to collaborate with Mohalla Clinics, a chain of government-run clinics in India, to deploy our product at their locations across the country, which will help us capture more market share quickly. With this growth strategy, we expect to capture 1% of the market in the second year, generating revenue of 4.8 million USD, and 3% in the third year, generating 14.4 million USD in revenue.

During the pre-revenue phase, our expenses will mostly come from legal support and R&D. After launching, we plan to allocate our funding mostly to sales management, large-scale manufacturing, and software maintenance. We anticipate that our gross margin will be around 60%.

Assuming that our estimates are accurate, we are confident that our financial plan will allow us to successfully enter the Indian healthcare market and establish ourselves as a leading screening solution provider.

Ocellux has been through a series of milestones, starting from winning the MIT Hack4thefuture competition in November 2020 where the idea of the company was generated. The company then completed the Regional NSF I-Corps Program in December 2021 and received a grant of $3000 for customer discovery. 

In January 2022, Ocellux was incubated at Project Olympus at Carnegie Mellon University, built a team of advisors for business and legal support, and also incubated at Atal Innovation Center GUSEC in India to increase exposure and connection in India. In March 2022, the company won a $10,000 investment at McGinnis Venture Competition, gained recognition and support.

The company then completed the LifeX Spring 2022 Accelerator program in June 2022, incubated at the Indian Institute of Public Health Gandhinagar in July 2022, completed the WF Global Accelerator program in August 2022, received a $3000 Spark Grant from Carnegie Mellon University in September 2022, won MedTech CoE Open Challenge 3.0 in October 2022. Ocellux received funding from various sources such as the GUSEC Program($4,300), the Indian Department of Science ($16,000), and the Indian Government ($120,000) through ANIC 2.0 AIM program. In February, Ocellux has been selected to participate in the final of the Duquesne New Venture Challenge and secured a $5,500 cash prize.