CerviScanner

What makes your solution innovative?

Our solution, CerviScanner, revolutionizes cervical cancer detection by leveraging AI technology to transform any smartphone into an effective screening tool. This innovative approach stands in stark contrast to costly alternatives such as telepathology, which can incur expenses up to $120M for software alone in a country like Kenya, and digital microscopes, averaging around $1,500 per unit—costs that are prohibitive for many healthcare sites within the African context.

In contrast, our app is accessible, offered for free, and integrates seamlessly with existing microscopes at no additional cost. This ensures widespread adoption is feasible and realistic across diverse healthcare settings. By utilizing smartphones, our app bypasses the need for expensive digitalizing equipment or new hardware. This not only reduces direct costs significantly compared to other methods but also speeds up the diagnosis process, eliminating the need for specialist intervention and the associated wait times.

Our app is designed to be compatible with a minimum camera resolution of 8 megapixels and works with both Android and iOS systems, ensuring broad usability. It functions effectively with conventional microscopes available in all hospital laboratories in Tunisia. CerviScanner is designed to function even when there is no internet connection. In such scenarios, the app will automatically save the microscope picture and store it locally on the device. Once an internet connection is re-established, the app will automatically upload the saved picture to the server. 

In terms of catalyzing broader positive impacts, our reach and data collection are exponential. Starting with 20 facilities in Tunisia and expanding to 80, with every facility conducting around 20 screenings/ day, we expect to conduct over 1 million tests across Africa within three years. This volume of data will refine our AI for greater precision, create cost savings for healthcare systems, and most importantly, save lives. The ripple effect of our app has the potential to set a new paradigm in healthcare efficiency and community empowerment, showcasing a sustainable model for global health initiatives.

Describe in simple terms how and why you expect your solution to have an impact on the problem.

CerviScanner leverages AI technology to enhance early detection of cervical cancer efficiently and affordably. Below is an outline of our Theory of change, detailed in the accompanying image, which describes how we plan to transform cervical cancer screening practices across Africa.

• Inputs: Our initiative leverages critical resources including funding, machine learning expertise, local healthcare partnerships, cloud infrastructure, and dedicated personnel.

• Activities: Key activities involve:
  • Refining our app’s accuracy.
  • Integrating user feedback for continuous improvement.
  • Collaborating with healthcare facilities to ensure seamless integration.
  • Forming an Expert Committee with diverse healthcare stakeholders.
  • Evaluating our impact using specific KPIs.

• Outputs: Initially, the app will be deployed in 20 facilities in Tunisia, expanding to 40 in Tunisia and Kenya by the second year, with a marked increase in screenings.

• Mid-term and ultimate outcomes: We aim to increase screening rates from their current level <30% to 50% by year one and 70% by year two. This objective resonates with the WHO’s recommendations. Ultimately, we strive to raise the five-year survival rate for cervical cancer from below 50% to at least 60%, bringing it closer to international standards. This will significantly reduce morbidity and mortality, and improve equitable access to cancer care across diverse communities.

What are your impact goals for your solution and how are you measuring your progress towards them?

Our solution’s impact goals are tightly interwoven with the UN Sustainable Development Goals, (SDG) and particularly SDG 3, underscoring our commitment to improving health outcomes through increased access to and quality of healthcare services. We will use these KPIs To measure our progress:

• Screening frequency: we monitor screenings through our app's backend, collecting real-time usage data to assess our reach and gauge the potential reduction in premature mortality.

• Diagnostic accuracy: to support universal health coverage as outlined in SDG target 3.8, diagnostic outcomes from our app are cross-referenced with medical records from healthcare facilities, ensuring accurate and reliable diagnostics.

• Early interventions: tracking early interventions in line with SDG target 3.4, we gather data from healthcare providers' records, enabling us to evaluate the effectiveness of early treatments that our app facilitates.

• Mortality rate: our contribution to SDG target 3.4 extends to studying mortality trends through public health records, which informs the long-term impact of our solution on reducing cervical cancer mortality rates.

• User satisfaction: we conduct periodic surveys with users and HPs to ensure the app delivers a satisfactory and effective user experience.

Data from these various sources allows us to closely monitor our KPIs, not only to measure our app’s direct impact but also to continually refine our approach. This ensures that as we scale, our strategies remain data-driven, user-focused, and aligned with the global and sustainable health advancements.

Describe the core technology that powers your solution.

Our model has been trained using 2 robust publicly available datasets that include 1,415 images, categorized under The Bethesda System. These images, sourced from real-world clinical settings, ensure our model is both relevant and effectively tuned for precision in medical diagnostics.

The AI model operates through a series of sophisticated steps beginning with image preprocessing, where images are resized, cropped, and formatted to optimize analysis. Feature extraction follows, with the AI identifying crucial image characteristics such as edges, colors, and textures. These features are essential for the model to recognize various stages of cervical lesions. During the training phase, the model adjusts its parameters to reduce errors and improve its ability to predict pre-cancerous and cancerous conditions accurately.

Once trained, the model applies its refined algorithms to new images, making predictions based on learned patterns. Post-processing of predictions enhances accuracy and reliability by filtering out uncertain results and refining the diagnosis. Beyond simple image classification, the model also supports object detection, image segmentation, and even image restoration—capabilities that broaden its applicability and effectiveness in medical imaging tasks.

Supporting the AI model is a robust backend infrastructure designed around a microservices architecture orchestrated by Kubernetes, which ensures scalability and reliability. This setup manages various functions from handling user requests to processing complex data efficiently. Our system employs cloud services including Kubernetes for automated deployment and scaling, PostgreSQL for reliable data storage and management, and GPU-accelerated virtual machines that provide the necessary computational power for intensive data processing and model training.

To further enhance the algorithm, we have incorporated pilot data collected from 189 samples across two hospitals in Tunisia, enabling us to refine the model with direct input from local real-world scenarios. This practical validation helps ensure that our application is not only theoretically sound but also effective in everyday clinical use.

Which of the following categories best describes your solution?

A new application of an existing technology

Please select the technologies currently used in your solution:

• Artificial Intelligence / Machine Learning

In which countries do you currently operate?

• Tunisia

Which, if any, additional countries will you be operating in within the next year?

• Kenya

How many people work on your solution team?

Our team dedicated to this solution consists of:

• Full-time staff:

  • 1 MD & Public health expert: Overseeing project alignment with objectives and providing research expertise.
  • 1 Intern and 2 volunteers: Focused on outreach, partnership establishment, and data analysis, with numbers varying based on project phase and current needs.
  • 1 Data scientist: Managing the technical development of the application.

• Part-time staff: Local project coordinators in Tunisia, 2 medical doctors who adapt and implement the project locally.

All team members are originally from Africa, providing an authentic perspective on the needs of the target regions.

How long have you been working on your solution?

Our dedicated team has been actively developing and refining CerviScanner for one year. Throughout 2023, we operated on a volunteer basis, channeling our collective expertise and passion into a successful pilot test of the app. Our commitment during this foundational year laid the groundwork for our app’s capabilities and set the stage for the subsequent scaling of our initiative.

Tell us about how you ensure that your team is diverse, minimizes barriers to opportunity for staff, and provides a welcoming and inclusive environment for all team members.

Our team at Medntech embodies diversity in professional expertise and life experiences. All team members are originally from Africa, offering a genuine understanding of the region's unique healthcare challenges. We have diverse backgrounds coming from different walks of life  such as medicine, AI, and public health. We are steadfast in our commitment to furthering gender balance, especially as our work directly impacts women's health.

We prioritize inclusivity, ensuring that all team members have a voice. Our meetings are spaces where ideas can be expressed freely and without judgment. This is vital as it reflects our roots in and dedication to the LMICs we serve, ensuring that our implementation strategies are not only effective but also culturally sensitive.

Our actions to foster a diverse and equitable workplace include:

• Expanding our recruitment drive to increase female representation, recognizing the importance of women in our mission and in leadership roles (currently we have a gender ratio of 1:1) 

• Conducting team-building activities that reinforce our commitment to diverse viewpoints and collective unity.

• Collaborating closely with local communities to align our health solutions with cultural expectations and needs.

What is your business model?

Medntech is a nonprofit organization that provides a free AI-powered cervical cancer screening app to healthcare providers in low-resource African countries. Our mission is to make critical healthcare services accessible to all, especially in regions where affordability and access to advanced technology are major challenges.We firmly believe that basic healthcare services, especially critical screenings like those for cervical cancer, should not be commercialized but should be accessible to all, particularly in regions where financial resources are major barriers. Thus, our mission is to enhance preventive healthcare, especially in regions where access to advanced technology are major challenges.

CerviScanner integrates seamlessly with existing medical infrastructure, requiring no additional equipment and working effectively even in areas with sporadic internet connectivity. This ensures that healthcare providers, including hospitals and clinics serving large populations at risk for cervical cancer, can improve diagnostic accuracy and patient outcomes without incurring extra costs.

To sustain our nonprofit operations and expand our impact, we are exploring a Software as a Service (SaaS) model in economically developed countries. This approach allows institutions with the budget for innovative health technologies to enhance their diagnostic tools while subsidizing the cost for users in lower-income regions. The revenue generated from this model supports our mission of providing the app for free where it is most needed and enables continuous improvement and expansion of the app’s capabilities.

Do you primarily provide products or services directly to individuals, to other organizations, or to the government?

What is your plan for becoming financially sustainable, and what evidence can you provide that this plan has been successful so far?

Our innovation is transitioning from a volunteer-driven model to a hybrid revenue structure, incorporating public funding, strategic partnerships, and a commercial SaaS model to ensure financial sustainability while staying true to our mission of providing high-quality healthcare technology for free in low-resource settings. Currently, we are in the process of securing funds from USAID. As we move forward, we will actively pursue grants from major global health entities such as Global Affairs Canada, and the Bill and Melinda Gates Foundation, which are essential for funding our initial rollout, research activities, and pilot program expansions. We have also prepared a list of 20 potential funders to support our development and scaling efforts

Simultaneously, we are forging strategic partnerships to enhance our reach and impact. Working with organizations like The Canada-Africa Community Health Alliance (CACHA) and the Africa Cancer Foundation (ACF) provides us with logistical support and community engagement, as well as crucial insights into cancer care in Africa. These partnerships not only help secure additional funding but also strengthen our operational capabilities in our target communities.

While we are developing a SaaS model targeted at healthcare providers in developed countries to generate a sustainable revenue stream, this revenue will support the free distribution of our app in underserved areas. This approach aligns our commercial activities with our social mission and is designed to be cost-effective for users in developed markets, providing critical support for our free services elsewhere. Additionally, we plan to rely on donations and government support to further sustain our initiatives and expand our impact