SindiColpo

Cervical cancer ranks as the second most prevalent cancer among females worldwide, with around 600,000 new cases and 340,000 fatalities documented in 2020 alone. In India, where cervical cancer remains a pressing health issue, reports from Globocan reveal a concerning figure of 120,000 new cases. Tragically, the toll of this disease is starkly felt, with an Indian woman succumbing to its grasp every 9th minute, although it could have been prevented if detected early.

Despite the gravity of this situation, less than 2% of women in India have undergone cervical cancer screening in the past five years, as reported by the WHO country profile for 2021. This can be attributed to a lack of awareness regarding screening and the limited availability of screening solutions in the field. This significant gap underscores the urgent need for accessible and scalable screening solutions.

Traditionally, colposcopy has served as a primary screening tool for detecting cervical cancer and invasive lesions, followed by resource-intensive cytology (Pap smear) and HPV DNA tests. However, its dependence on highly skilled manpower, subjectivity (suspicious legions which are not visible through naked human eyes) and specific infrastructure, along with its cost and bulkiness, hinder its widespread use for population-based screening efforts. Therefore, there is a pressing need for a non-invasive, cost-effective, easy-to-use, portable screening solution that can be deployed in the field for instant diagnosis and reporting of suspicious cases at an early stage.

Introducing SinDiColpo, our patented low-cost, AI-enabled cervical cancer screening device, poised to revolutionize the landscape of cervical cancer diagnosis. Designed for both diagnosis physicians and primary healthcare workers, SinDiColpo aims to provide instant diagnosis, empowering healthcare providers to conduct screenings effectively.

Our most crucial breakthrough lies in our ability to detect and mark cervical cancer lesions from images that may not be visible to the naked eye. While SinDiColpo currently relies on user input for image classification and generating templated reports, we're advancing our technology by integrating state-of-the-art data augmentation techniques, such as Generative Adversarial Networks. This will enable us to develop a robust classifier and utilize feature extraction techniques to identify specific regions within images that contribute most to the classification process. Additionally, to enhance user upskilling, we're implementing a Large Language Model (LLM)-driven agent capable of engaging in conversations with users, addressing specific questions related to diagnosis. To mitigate the risk of hallucinations by the LLMs, these conversations will be conducted within a controlled Retrieval Augmented Generation setting.

The novelty of our proposal lies in the integration of clinical data collection, image processing algorithms for diagnosis, and the introduction of explainability driven by LLMs. This approach offers a threefold advantage over traditional colposcopy methods: higher accuracy of classification, identification of key regions within images contributing to classification, and the promotion of uniformity in healthcare workers' skills through conversational LLM agents.

SinDiColpo targets two primary populations:

1. B2B: Women's Hospitals, Gynaecologists, and General Physicians operating in low-resource settings.
2. Expectant mothers and their families, particularly in rural and tribal areas of India and Africa, where awareness of cervical cancer screening is notably low. This population, often residing in lower-income settings, is particularly vulnerable to the disease.

Our immediate focus is on the developed world and developing regions with reasonable access to healthcare facilities. Since visual inspection and assessment are essential for determining the primary course of treatment for cervical cancer, we collaborate with obstetricians, surgeons, and cancer specialists to understand the decision-making process and tailor our solution to meet the unmet clinical need.

In the longer term, we aim to extend the benefits of SinDiColpo to women with access to midwives or low-resource health facilities, such as those in remote regions of developing countries or rural areas of developed nations. These patients and their caregivers can benefit from our solution to make informed decisions about continuing relevant interventions.

To reach these target populations, we leverage our existing network of over 1000 rural hospitals across 12 states and partner with more than 35 NGOs serving low-income communities in tribal and remote regions. Additionally, our partnership with the National Health Mission (NHM) project implementation plan for FY 2023-24 further strengthens our outreach efforts in India.

India alone boasts a vast network of over 100,000 active gynaecologists, who are pivotal in reaching the vulnerable population. Furthermore, with over 1 million general physicians and 2 million frontline healthcare workers, SinDiColpo can be utilized for on-field screening and remote connectivity with gynaecologists. Government entities, with over 150,000 major public healthcare facilities, also represent a significant customer base for our solution.

Team:

Founder, Mr. Aditya Kulkarni, has left out of PhD & high paying jobs and built the company since last 7+ years just because of a high level of motivation for social impact. Dr. Anupama joined CareNX in 2021 with her innovation so as to take it to the market and commercialise it through CareNX.

Rest is an interdisciplinary team of 3 product engineers, 2 biomedical & 2 public health experts and 4 sales/marketing personal. Team is mentored by professors from KEM Hospital Mumbai and IIT Bombay. The team has experience of developing and commercialising technology products related to maternal and child health since last 7+ years and has built a sustainable business with annual revenue of over USD 2 million covering for over 1000+ healthcare facilities, 500000+ pregnancies and 2500+ gynaecologists so far.

Learnings so far:

1. Partnerships is a key in delivering care. We have had worked with 35+ big inter country NGOs/Governments partners and learned that no single organisation can create end-to-end system for comprehensive care and management.
2. Repeated training and capacity building of partners is very essential.
3. Behavior change can't be underestimated for new innovations, particularly in healthcare sector.

Why we are positioned in a right manner to scale SindiColpo:

1. We have 7+ years of experience of scaling up fetosense fetal monitoring solution to 1000+ B2B hospitals including government, NGO, private and premium settings. Top private maternity chain of hospitals Cloudnine group, Sahyadri group, O&P group, Rainbow group and reputed government medical hospitals AIIMS, KEM, Nanavati, GMCH, AFMC etc. have been using our solution since 3+ years and have performed over 500000+ tests. 
2. We have established exclusive annual distribution and service partnership with a pharmaceutical companies for PAN India presence. A team of 6 regional sales managers have been activated to onboard 1000+ B2B hospitals to facilitate home monitoring for patients in FY 2023-24.
3. The product has been patented and validated with leading cancer care institute as per the standard clinical protocol in adherence with CDSCO regulatory body.

SinDiColpo has achieved significant milestones in its development and validation process:

1. Prototyping, Validation, and Patent: SinDiColpo has been prototyped, validated, and patented. A pilot study was conducted to assess its sensitivity and specificity against a gold standard hysteropathology in 2022.

2. Multi-Centric Trial: Currently, a multi-centric trial is underway involving five different hospitals. This trial, initiated to further validate the effectiveness of SinDiColpo, is scheduled to conclude in September 2024.

3. Regulatory Compliance: SinDiColpo has obtained ISO13485:2016 Quality Management System (QMS) certification and a Central Drugs Standard Control Organization (CDSCO) test license. These certifications underscore the product's adherence to international quality standards and regulatory requirements.

4. Customer Adoption and Market Testing: SinDiColpo is currently being used by over 10 paid customers. This includes healthcare institutions and providers who are actively utilizing the device for cervical cancer screening. Additionally, ongoing market testing is being conducted to assess product-market fit and refine the offering based on customer feedback.

Applying to MIT Solve offers several compelling reasons for SinDiColpo:

1. Access to Resources: MIT Solve provides access to a global network of innovators, experts, and resources that can support SinDiColpo's further development and scale-up efforts. This includes mentorship, funding opportunities, and connections to potential partners and investors.

2. Validation and Recognition: Being selected as a Solver by MIT Solve provides validation and recognition for SinDiColpo's innovative approach to addressing a pressing global challenge. This recognition can enhance the credibility and visibility of the solution, attracting attention from stakeholders and potential collaborators.

3. Collaborative Opportunities: MIT Solve fosters collaboration among Solver teams, enabling knowledge sharing and cross-pollination of ideas. By participating in MIT Solve, SinDiColpo can engage with other innovators working on complementary or related solutions, potentially leading to synergistic partnerships and collaborations.

4. Global Impact: MIT Solve focuses on solving some of the world's most pressing challenges, including healthcare disparities. By aligning with MIT Solve's mission and participating in its initiatives, SinDiColpo can amplify its impact and contribute to addressing global health inequities, particularly in underserved communities.

5. Access to Partnerships and Funding: Participation in MIT Solve can facilitate access to partnerships with relevant cross-country organizations from low- and middle-income countries (LMICs) for initial pilots and data generation. This enables SinDiColpo to test its AI in diverse contexts and mitigate bias. Additionally, access to high-risk funding for product-market fit experiments can accelerate SinDiColpo's journey towards scalability and impact.

Overall, applying to MIT Solve aligns with SinDiColpo's goals of accessing resources, gaining validation and recognition, fostering collaboration, making a meaningful global impact, and accessing partnerships and funding opportunities for further development and scale-up.

Our solution, SinDiColpo, embodies innovation not only in its technological aspects but also in its business model:

1. AI-enabled Screening: SinDiColpo harnesses the power of artificial intelligence (AI) to revolutionize cervical cancer screening. By utilizing advanced image processing algorithms, our device can detect and mark cervical cancer lesions that may not be visible to the naked eye, enhancing the accuracy and efficiency of the screening process.

2. Instant Diagnosis: Unlike traditional screening methods that may require time-consuming laboratory analysis, SinDiColpo provides instant diagnosis at the point of care. This real-time feedback enables healthcare providers to promptly identify suspicious cases and initiate appropriate follow-up measures, leading to earlier detection and intervention.

3. User-Friendly Interface: SinDiColpo is designed with usability in mind, making it accessible to a wide range of healthcare providers, including primary care physicians and midwives. Its intuitive interface and easy-to-use features empower users to conduct screenings effectively, even in low-resource settings or remote areas.

4. Data-driven Insights: Our solution goes beyond mere diagnosis by providing valuable insights derived from data analytics. By analyzing patterns and trends in screening results, SinDiColpo can offer actionable recommendations for targeted interventions and population health management, ultimately improving outcomes for patients and communities.

5. Adaptability and Scalability: SinDiColpo is designed to be adaptable to diverse healthcare settings and scalable to meet the needs of different populations. Whether deployed in urban clinics or rural health centers, our solution can seamlessly integrate into existing workflows and infrastructure, maximizing its impact on cervical cancer prevention and control.

6. Business Innovation: We plan to commercialize SinDiColpo through a per-test based subscription model instead of a fixed-cost device sale. This innovative approach eliminates the need for customers to make a capital investment and provides revenue generation opportunities at no upfront cost, ensuring accessibility and affordability for healthcare providers across different settings.

Overall, SinDiColpo represents a paradigm shift in cervical cancer screening, offering a comprehensive, efficient, and user-friendly solution that has the potential to transform healthcare delivery and save lives while introducing a novel business model to ensure accessibility and sustainability.

Our solution, SinDiColpo, uses advanced technology to help detect cervical cancer earlier and more accurately than traditional methods. Here's how it works and why it's expected to make a difference:

1. How it works: SinDiColpo uses artificial intelligence (AI) to analyze images of the cervix, looking for signs of cancer or pre-cancerous changes. These changes can be very subtle and hard to see with the naked eye, but SinDiColpo's AI can pick them up, allowing healthcare providers to catch the disease at an earlier stage.

2. Why it matters: Early detection is crucial for treating cervical cancer successfully. When caught early, the chances of curing the disease are much higher. By providing instant and accurate diagnoses, SinDiColpo helps healthcare providers start treatment sooner, potentially saving lives and reducing the need for more invasive procedures down the line.

In simple terms, SinDiColpo is like having a super-powered microscope that can spot cancer cells when they're just starting to develop. By catching the disease early, it gives patients the best possible chance of beating it. That's why we expect SinDiColpo to have a big impact on the problem of cervical cancer.

Since 2015, ‘CareNX’ program has screened 500,000+ pregnancies (12000+ high-risk) via 2500+ gynaecologists across PAN India, in rural, urban and tribal areas, with 1000+ private hospitals, 15+ Government, non-profit and corporate partners. SindiColpo will be made available to selected facilities.

Our impact goals for SinDiColpo are centered around improving cervical cancer outcomes and increasing access to screening, particularly in underserved communities. Here's how we're measuring our progress towards these goals:

1. Early Detection Rates: Our primary goal is to increase the rate of early detection of cervical cancer and pre-cancerous lesions. We aim to track the number of cases detected at an early stage using SinDiColpo compared to traditional screening methods. By monitoring this metric, we can assess the impact of our solution on improving patient outcomes and reducing mortality rates associated with cervical cancer.

2. Screening Coverage: We aim to expand access to cervical cancer screening, particularly in low-resource settings and underserved communities where screening rates are typically low. We track the number of screenings conducted using SinDiColpo in these areas and monitor changes in screening coverage over time. Increasing screening coverage is essential for early detection and prevention of cervical cancer.

3. Healthcare Provider Adoption: We measure the adoption of SinDiColpo among healthcare providers, including gynaecologists, general physicians, and frontline healthcare workers. Tracking the number of healthcare facilities and professionals using SinDiColpo provides insights into the reach and acceptance of our solution within the healthcare community.

4. Patient Outcomes: Ultimately, our goal is to improve patient outcomes, including survival rates and quality of life. We monitor patient outcomes, such as treatment adherence, recurrence rates, and long-term survival, among individuals diagnosed with cervical cancer using SinDiColpo. Improvements in patient outcomes indicate the effectiveness of our solution in facilitating early diagnosis and timely intervention.

5. Equity and Accessibility: We aim to reduce disparities in cervical cancer outcomes by ensuring equitable access to screening, diagnosis, and treatment. We track the distribution of SinDiColpo in different geographic regions and assess its impact on reducing disparities in access to cervical cancer care.

By systematically tracking these metrics and evaluating our progress against predefined impact goals, we can continuously refine and optimize SinDiColpo to maximize its effectiveness in addressing the global burden of cervical cancer.

SinDiColpo, a low-cost, AI-powered cervical cancer screening device, integrates clinical data, image processing, and explainable AI via conversational agents. This innovation aims to boost accuracy in classification of precancerous lesions, pinpoint crucial image regions for diagnosis, and standardize healthcare worker skills, overcoming barriers to widespread cervical cancer screening.

This proposal focuses on enhancing the efficacy of cervical cancer screening using image processing algorithms. The present problem with cervical cancer screening using image processing is the problem of long-tailed representation with few classes having lesser representation preventing development of robust classification algorithms. We utilize state-of-the-art algorithms such as label-distribution-aware margin LDAM loss https://arxiv.org/abs/1906.074... and Supervised Contrastive Learning https://arxiv.org/abs/2004.113... to addressed the data imbalance problem. Also, we plan to implement novel image augmentation techniques such as one-shot GANs https://arxiv.org/abs/2202.137... and DeliGAN https://arxiv.org/abs/1706.020... specially designed for low representation scenarios.

Once the data imbalance problem for classification is addressed we will develop a hotspot identification algorithm using the guided back-propagation leveraging the feature extraction capabilities of intermediate layers of a CNN to understand which parts of the image are crucial for the networks decision-making process. These hotspots would be mapped to the original image so that the diagnosing physician will be able to ask relevant questions to our LLM driven agent regarding the image features, the responses of which will be guided by Retrieval Augmented Generation RAG from a set of relevant well curated medical documents. We propose to utilize the open-source LLM, Meditron which is a medical domain fine tuned Llamma2 model.

Full time: 6

Part time: 3

Interns: 2

Consultants: 3

CareNX workforce consist of highly skilled interdisciplinary team of computer scientists, biomedical & public health experts and management. Team is mentored by professors from KEM Mumbai, IITBombay, National University of Singapore (NUS) and Stanford University. As company is Incubated in IIT, it has access to all the fabrication facilities and DSIR certified labs at IIT Bombay. Also the team has experience of developing and commercialising a product related to maternal and child health since last 7+ years Our advisors are from strong public health background, technology and start-up scaling. 

• CareNX's mission is to develop digital technologies for better women healthcare. We have a medical lead doctor who is a women for this project. There are 2 other supporting women team mates. 
• Project targets pregnant women as direct beneficiary. The geography we have selected is non-metro regions of the towns in which the access to care is limited. We will cater to beneficiaries from all ethnicities, minorities, race or religion. 
• The diversity and social inclusion is to be achieved by providing respectful women care through all the partners. The fetosense & SindiColpo technology has been developed with neutral perspective and principles of responsible AI. 

1. Customer segments: 
   1. Maternity Clinics (OBGYn)
   2. General Physician's Clinics
   3. NGOs and Governments
   4. Pharmaceutical companies
2. Value propositions: 
   1. 3-fold reduction in the price as compared to market options
   2. Auto-interpretation of results on-field
   3. Portable & ease-of-use 
   4. Digital record storage & remote monitoring
3. Revenue streams:
   1. Fixed cost model: One time instalment cost + consumables (recurring) + AMC
   2. Pay-as-you-go model: One time deposit at instalment + Per test revenue sharing

1. CareNX will fund 50% of SindiColpo model expansion through revenue of Fetosense solution for next 12 months. 
2. Rest 50% matching amount would be made available through seed investment by IIT Bombay.
3. With current projected per-test based commercial model, each deployed device shall generate break even revenue within 18 months. We will work with existing investors to achieve product market fit and then approach VCs to scale the commercial aspects.