Ahmedabad Students’ Derma Vision: An AI Skin-Health App That Flagged 50 Cases in a 450-Woman Screening — What It Means for Community Care and AI in Dermatology

Key Highlights:
Introduction
From Classroom Prototype to Community Screening: The Derma Vision Story
How Derma Vision Works: Computer Vision Applied to Visible Skin Patterns
Why Community Screening Matters: Context and Impact
Real-World Parallels: Commercial and Research Efforts in AI Dermatology
Validation and Clinical Safeguards: From Prototype to Trustworthy Tool
Data Privacy, Consent and Governance: Non-Negotiables for Sensitive Biometrics
Addressing Bias: Skin Tone, Device Variation and Socioeconomic Factors
Regulatory and Certification Pathways: The Road from Prototype to Approved Medical Tool
Operationalising Community Screenings: Practical Considerations and Protocols
Educational Value: What This Project Means for STEM Learning
Risks and Limitations: Where Caution Is Warranted
Scaling Derma Vision: Partnerships, Funding and Priority Actions
Comparisons and What Derma Vision Adds to the Field
Practical Recommendations for Schools, NGOs and Health Authorities Interested in Similar Initiatives
The Social Dimension: Women’s Health, Access and the Role of School-Led Initiatives
A Note on Communication: Managing Expectations and Avoiding Overclaiming
What Success Looks Like: Benchmarks for Responsible Progress
FAQ

Key Highlights:

A team of Class 7–8 students from Zebar School developed Derma Vision, an AI-driven app that screened 450 women during a Women’s Day programme and identified roughly 50 participants with visible skin patterns warranting dermatological consultation.
The project won “Best Innovative Idea in Artificial Intelligence” at the Vigyantram National Championship 2026 at IIT Delhi; the students plan clinical validation with dermatologists and to strengthen ethical and clinical safeguards before wider deployment.

Introduction

A small team of middle-school students in Ahmedabad converted classroom curiosity into a practical public-health intervention. Derma Vision, a mobile application that uses computer vision to analyse visible skin patterns, served not only as a classroom project but as an instrument of community screening. When more than 450 women volunteered for scans during a Women’s Day awareness programme, the app flagged about 50 individuals with skin changes meriting professional follow-up. That result turned a classroom prototype into a case study about how early-stage AI tools can contribute to preventive care, while simultaneously exposing the technical, clinical and ethical hurdles that lie between a promising prototype and a safe, scalable health product.

Beyond the immediate impact on participants, Derma Vision raises questions that stretch from educational methodology to regulatory frameworks: How accurate are small-team-built AI tools? What does clinical validation require? How should developers design privacy and consent around sensitive biometric data? And how can community screening initiatives be structured so they identify need without creating undue alarm? Answers to these questions will determine whether Derma Vision and projects like it remain inspiring one-off successes or evolve into responsible, effective public-health tools.

The narrative that follows examines the technology and results behind Derma Vision, places the project in the context of AI in dermatology, outlines the steps required for clinical and ethical validation, and sketches a practical roadmap for scaling community-based skin-screening initiatives. Real-world parallels and implementation lessons are used to clarify where this project sits on the path from school innovation to medical device.

From Classroom Prototype to Community Screening: The Derma Vision Story

Derma Vision was built by four students—Dishen Gadhiya, Hetansh Patel, Yug Dalsania and Janmesh Darji—at Zebar School for Children in Ahmedabad. Their application applies computer vision to photographs of skin and returns plain-language results, offering basic skincare tips, preventive measures, and home remedies. It recommends dermatologist consultation only when patterns suggest potentially significant pathology.

The students tested the app in a community setting during a Women’s Day awareness programme that included more than 450 women—teachers, administrative staff and ground staff from the Udgamverse schools network. Participation was voluntary. Derma Vision’s screening indicated that roughly 50 participants displayed visible patterns that merited professional dermatological review. That finding validated two things at once: community interest in accessible screening tools, and the app’s potential to triage visible findings.

Recognition followed. Derma Vision won “Best Innovative Idea in Artificial Intelligence” at the Vigyantram National Championship 2026 at IIT Delhi, where teams from 49 schools across India competed. The prize highlights how student-led innovation can intersect with national-level science and technology fora, and it created momentum for the team to pursue clinical validation and institutional partnerships.

How Derma Vision Works: Computer Vision Applied to Visible Skin Patterns

The source article states that Derma Vision uses computer vision to analyse visible skin patterns and deliver results in simple language. A plausible technical architecture for such a tool combines three main components: image acquisition, image preprocessing and a trained model for pattern recognition.

Image acquisition: Images can be captured using a smartphone camera. For community screenings, ensuring consistent image quality means controlling lighting, distance to the camera, angle, and background. Simple aids—uniform backgrounds, fixed distance markers, or portable imaging booths—reduce variability and improve model reliability.
Preprocessing: Preprocessing corrects lighting differences, normalises colour, and segments the skin region from surrounding background. Preprocessing can also include automated detection of artefacts such as tattoos, jewellery, or clothing that could interfere with interpretation.
Pattern recognition model: Most contemporary approaches use convolutional neural networks (CNNs) or transformer-based vision models trained on labelled skin images. For triage-style apps, models are generally trained to recognise broad categories (e.g., inflammatory dermatitis, fungal patterns, pigmentary changes, suspicious lesions) rather than provide definite diagnoses. Output is usually a probability distribution across categories; app logic then translates probability into recommendations—routine care, preventive tips, or referral.

To be useful in a community setting, the interface must translate model outputs into actionable guidance for non-clinical users. Derma Vision’s reported design—simple language, clear recommendations, and an emphasis on seeing a dermatologist only when necessary—aligns with best practices for health triage tools.

Why Community Screening Matters: Context and Impact

Screening at community gatherings brings healthcare touchpoints to populations that might otherwise delay care. Community-based screening camps—ranging from vision and dental camps to mobile mammography units—have long been used in India and elsewhere to expand access to preventive services.

Skin health screening has particular relevance for several reasons:

Visibility and stigma: Skin conditions can be visible and socially stigmatizing, affecting self-esteem and social participation. Early identification can reduce psychosocial burden and open pathways to affordable treatment.
Access gaps: Dermatology services are concentrated in urban centres and tertiary hospitals. In many communities, mild-to-moderate conditions remain untreated due to cost, distance, or low perceived importance.
Preventive value: Some skin conditions are easier to treat and have better outcomes when identified early. Even non-life-threatening problems—persistent infections, chronic eczema, or certain pigmentary disorders—benefit from earlier intervention.

Derma Vision’s screening during a Women’s Day event harnessed a pre-existing gathering and a high level of engagement among participants. The result—about 11% of those screened receiving a recommendation for dermatology consultation—demonstrates that even simple triage can surface unmet needs.

Real-World Parallels: Commercial and Research Efforts in AI Dermatology

Derma Vision joins a growing field of AI tools aimed at skin assessment. Examples from the market and academia illuminate strengths and pitfalls.

SkinVision: A commercial app based in the Netherlands focused on melanoma risk assessment. It uses image analysis to provide risk levels for lesions and directs users to see a clinician when risk is elevated. SkinVision obtained regulatory clearance in some jurisdictions and pursued clinical validation studies showing certain levels of sensitivity and specificity for melanoma detection. The app’s experience emphasises two ideas: prospective validation is essential; and regulatory pathways require clinical evidence.
MoleMapper and photo-monitoring tools: Apps that allow users to track lesions over time and flag changes. These tools are valuable for monitoring but rely on user engagement and consistent imaging.
Academic studies: Several high-profile studies have shown that well-trained convolutional networks can match or exceed dermatologist-level performance in narrow tasks, such as distinguishing melanoma from benign nevi under controlled conditions. However, follow-up research highlighted challenges: dataset bias (overrepresentation of lighter skin types), variable clinical settings, differences in image acquisition, and the need for prospective clinical trials.

These examples show a route from prototype to validated product: retrospective training, prospective validation, larger diversity in data sets, integration with clinical workflows, and regulatory compliance.

Validation and Clinical Safeguards: From Prototype to Trustworthy Tool

For Derma Vision to transition from a classroom proof-of-concept to a health product, developers and partners must pursue rigorous clinical validation and embed strong ethical safeguards. The students’ plan to work with dermatologists is the correct first step. Key elements that stakeholders should follow include:

Retrospective validation with labelled datasets: Assemble a diverse dataset of skin images labelled by dermatologists. Evaluate model performance using clinically relevant metrics—sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating characteristic curve (AUC).
Prospective clinical studies: Implement prospective evaluation where the app’s recommendations are compared against dermatologist assessments in real time. Prospective studies reveal how the tool performs in operational conditions—with variable lighting, camera types and patient behaviours.
Diversity in datasets: Ensure representation across age groups, skin tones (Fitzpatrick I–VI), anatomical locations, and common skin disorders in the target population. Performance that is acceptable on lighter skin may degrade substantially on darker skin if not trained with representative data.
Clinical endpoints and triage thresholds: Define the risk thresholds that trigger a referral. For conditions where missing a diagnosis has major consequences (e.g., suspicious melanocytic lesions), adopt conservative thresholds and ensure that false negatives are minimised.
Ethical oversight and institutional review: Engage institutional review boards (IRBs) or equivalent ethics committees before conducting human-subjects research. For projects involving minors as users or participants, additional consent and child-protection measures are required.
Explainability and clinician-in-the-loop workflows: Provide clinicians with interpretable outputs—heat maps, highlighted regions, or confidence scores—so that AI augments rather than replaces clinical judgment.
Post-deployment monitoring: If the app is deployed outside a trial, continuous monitoring for performance drift, reportable adverse events, and user feedback is necessary.

Success in validation pays off beyond regulatory acceptance. Clinician buy-in, user trust and responsible scaling all depend on robust, transparent validation.

Data Privacy, Consent and Governance: Non-Negotiables for Sensitive Biometrics

Skin images are biometric data and carry privacy implications. App design must reflect that reality.

Informed consent: Users must understand what data are collected, how they will be used, whether images will be stored or shared for model training, and what measures protect their identity. Consent language should be concise, clear and available in local languages.
Data minimisation and anonymisation: Collect only data necessary for the app’s function. Where images are used for model improvement, remove metadata and consider automated face-blurring if not clinically required.
Secure storage and transmission: Encrypt data at rest and in transit. Use secure cloud providers that comply with relevant standards and regulations.
Local regulations and cross-border data flow: India and other countries have evolving frameworks for health and biometric data. Developers should be aware of national laws governing personal data protection and health information. If images or datasets cross borders, that movement may invoke additional regulatory regimes.
User control and deletion rights: Provide users with clear options to delete their images and associated data. Record retention policies should be transparent.

Handling data responsibly is as important as model performance. Breaches or opaque practices would erode community trust and jeopardise future screening efforts.

Addressing Bias: Skin Tone, Device Variation and Socioeconomic Factors

Bias in AI dermatology is well documented. It arises from skewed training data and operational disparities such as camera quality and lighting. Three practical mitigations are essential:

Inclusive datasets: Recruit images across the full range of skin tones, ages, and common local pathologies. Community partnerships and clinician-labelled image collection help achieve balance.
Device and context variability: Train with images from the range of devices likely to be used in deployment settings—basic feature phones, mid-range smartphones and high-end models. Simulate variable lighting and common environmental factors.
Local calibration: Validate models on hold-out datasets from the same community where deployment will occur. Periodic re-evaluation after deployment detects any drift induced by changing usage patterns.

Bias mitigation is not a one-time task. It is an ongoing governance requirement that must be embedded into both research protocols and commercial product lifecycles.

Regulatory and Certification Pathways: The Road from Prototype to Approved Medical Tool

Countries treat AI-based health tools differently depending on intended use. A triage app that recommends seeing a dermatologist may be classified as a medical device in many jurisdictions. Developers should consider these steps:

Determine regulatory classification: Engage with national regulatory authorities early to determine whether the app is a medical device and what class it falls into. In India, CDSCO (Central Drugs Standard Control Organization) provides guidance for medical devices and is developing frameworks for software as a medical device (SaMD).
Clinical evidence plan: Outline the evidence required for approval—often a combination of retrospective and prospective clinical data, safety data and technical documentation.
Quality management systems: Implement a quality management system (QMS) appropriate for medical device development (e.g., ISO 13485). QMS helps standardise documentation, risk management and post-market surveillance.
Transparency for users: Regulatory filings increasingly require explainability about datasets, model architecture and known limitations. Keep records of training data provenance, versioning and performance metrics.
Post-market surveillance: Regulatory bodies expect continuous monitoring of safety and performance. Plan for mechanisms to receive and act on adverse event reports.

Early regulatory engagement reduces the risk that years of development will stall due to missing documentation or unexpected requirements.

Operationalising Community Screenings: Practical Considerations and Protocols

The Derma Vision screening illustrated a practical approach: use a community event to recruit many participants rapidly. To scale such screenings while maintaining quality and safety, adopt clear protocols.

Pre-screening information and consent stations: Provide simple, multilingual information about the screening’s purpose, data usage and follow-up pathways. Use printed materials and verbal briefings so participants can ask questions.
Standardised image capture stations: Create a reproducible environment—consistent lighting, neutral background, fixed camera distance—so that image quality supports reliable analysis.
On-site clinician presence or teledermatology backup: Ideally, dermatologists should be on call to review flagged cases the same day, or teleconsultation should be available to triage urgency.
Referral networks: Establish direct referral pathways to local clinics or specialists, ideally with subsidised or low-cost options for participants who need follow-up.
Documentation and follow-up: Collect minimal contact information and secure consent for follow-up. Track outcomes to close the loop: Did participants see a clinician? What were the final diagnoses? How did the app’s triage align with clinical findings?
Risk communication: When an app flags a finding, communicate risk levels clearly and avoid alarmist language. Explain what the recommendation means and what steps participants can take next.

Real-world screening initiatives succeed when they anticipate logistical hurdles and invest in the human infrastructure—clinicians, community health workers, referral coordination—that converts screening into improved health outcomes.

Educational Value: What This Project Means for STEM Learning

Derma Vision illustrates the pedagogical power of project-based STEM education. A small group of middle-school students navigated programming, machine learning concepts, user interface design and community ethics, gaining skills far beyond textbook exercises.

Key educational takeaways:

Interdisciplinary learning: The project fused biology (skin physiology), computer science (computer vision), and social science (ethics, community engagement).
Real-world problem framing: Students worked within constraints—user comprehension, privacy concerns and community dynamics—that mirrored professional practice.
Iterative design culture: Building, testing, receiving feedback and refining the app exposed students to iterative product development, an essential skill for technology careers.
Public recognition and motivation: Winning a national award validated the students’ work and created momentum for future collaboration with clinicians, universities and potential funders.

Schools can amplify these lessons by forming partnerships with local health institutions, arranging mentorships with clinicians or data scientists, and creating processes for ethical review for student-led health research.

Risks and Limitations: Where Caution Is Warranted

AI-based screening tools offer promise but carry distinct risks. An honest appraisal clarifies boundaries and prevents harm.

False reassurance: A false-negative result could lead someone to delay seeking care for a condition that worsens without treatment. For serious or ambiguous findings, the app’s default should lean toward referral.
False positives and anxiety: Excessive false positives can generate unnecessary worry and burden clinicians with avoidable referrals. Calibration of thresholds seeks to balance sensitivity and specificity.
Overreliance by non-clinical users: Users might treat app outputs as definitive diagnoses. Clear labeling, prominent disclaimers and educational materials mitigate this risk.
Liability and ethical responsibility: Developers and partners need clarity on who is responsible for adverse outcomes. Legal frameworks and insurance can help define accountability.
Dataset privacy and misuse: Without strict governance, images could be repurposed or leaked, causing privacy harms.

A cautious, transparent approach—combined with clinician involvement and robust governance—reduces these risks significantly.

Scaling Derma Vision: Partnerships, Funding and Priority Actions

If the students and their mentors pursue broader deployment, the following roadmap provides a pragmatic sequence.

Consolidate clinical partnerships: Establish formal collaboration with dermatology departments at local medical colleges or private clinics to support validation and referral.
Build a labelled dataset: With informed consent, collect anonymised, dermatologist-annotated images that reflect the target population’s diversity.
Conduct retrospective and prospective validation studies: Publish results in peer-reviewed venues where possible. Peer review strengthens credibility.
Strengthen the app’s privacy and security measures: Implement encryption, data minimisation, clear consent flows and deletion mechanisms.
Seek regulatory guidance: Early conversations with relevant authorities avoid surprises. Determine whether the app requires registration as a medical device.
Pilot scaled screenings with evaluation: Run additional community events with improved imaging protocols and on-site or remote clinician backup. Track outcomes and health-system integration.
Secure funding for scale: Grants from science education bodies, public-health agencies, philanthropy or social-impact investors can support further development, validation and deployment.
Iterate UX and language support: Make the interface accessible—multilingual, culturally appropriate and usable by people with low digital literacy.
Create training materials for community health workers: Short training modules help local staff operate imaging stations and manage referrals.
Establish post-market surveillance: Monitor app performance and user outcomes after wider deployment.

This sequence balances scientific rigour, user safety and practical constraints, opening a realistic pathway from prototype to public-health tool.

Comparisons and What Derma Vision Adds to the Field

Derma Vision is not unique in applying computer vision to skin. What distinguishes it is the educational genesis, community deployment during a Women’s Day event, and a pivot toward conservative triage—prioritising dermatologist referrals only when necessary.

Commercial apps often focus on melanoma detection in populations with higher baseline melanoma prevalence, whereas Derma Vision’s initial deployment surfaced a broader set of skin-pattern abnormalities in a non-self-selecting community. That difference emphasises a common theme: AI tools must align with the local epidemiology and user expectations.

Derma Vision’s immediate strengths include ease of use, plain-language outputs and a screening design tailored to community events. Its priorities moving forward—clinical validation, dataset diversity and ethical safeguards—mirror the maturation path of earlier commercial and academic projects.

Practical Recommendations for Schools, NGOs and Health Authorities Interested in Similar Initiatives

Community partners aiming to replicate Derma Vision’s approach should consider these operational and ethical practices.

Start with clear objectives: Define whether the goal is awareness, triage, surveillance or research. Objectives determine study design and ethical requirements.
Partner early with clinicians: Dermatologists can guide image labelling, triage thresholds and referral pathways.
Obtain ethical approvals for research: Institutional review provides external oversight for human-subjects work.
Invest in simple imaging protocols: Affordable, portable lighting and a neutral backdrop dramatically improve data quality.
Make consent multilingual and accessible: Use verbal explanations and visual aids for participants with limited literacy.
Plan for referrals: Screenings must connect participants to low-cost or subsidised care; otherwise, identifying problems without access to treatment risks harm.
Track outcomes: Measuring whether flagged participants obtain care and what diagnoses follow is essential to evaluate impact.
Protect data and privacy: Implement encryption, limit retention and make deletion requests easy to execute.

These steps reduce the chance of unintended consequences and increase the likelihood of meaningful health outcomes.

The Social Dimension: Women’s Health, Access and the Role of School-Led Initiatives

Derma Vision’s deployment during a Women’s Day event magnified its social significance. Women in many communities face barriers to care—household responsibilities, economic constraints, and, in some contexts, stigma around seeking treatment for visible conditions. Bringing screening into a school event reduced friction and normalised proactive care.

School-led initiatives have advantages: trust in institutions, easy access to families, and an educational motive that attracts volunteerism and mentorship. That trust also creates responsibility. Schools and student teams must be careful that screenings are accompanied by clear guidance, realistic referrals and no false promises.

When designed responsibly, school projects can catalyse local health improvements, train future health-tech innovators, and bridge gaps between communities and health systems.

A Note on Communication: Managing Expectations and Avoiding Overclaiming

Public communication around early-stage AI health tools requires restraint. Headlines that frame prototypes as diagnostic breakthroughs risk misleading users and triggering inappropriate behaviours. Derma Vision’s current profile—as an awareness and triage tool developed by students—should be presented accurately: useful for preliminary screening, but not a substitute for physician diagnosis.

Clear messaging protects participants and preserves the ethical integrity of the project as it scales.

What Success Looks Like: Benchmarks for Responsible Progress

Meaningful milestones that indicate responsible progress for Derma Vision include:

Completion of dermatologist-led retrospective and prospective validation with documented performance metrics.
Demonstrated consistent performance across a representative range of skin tones and device types.
Publication of results in peer-reviewed outlets or publicly accessible technical reports.
Establishment of secure data governance policies and documented patient consent procedures.
Formal partnerships with health providers that ensure referrals lead to accessible care.
Regulatory guidance or clearance consistent with the app’s intended use.

Achieving these benchmarks positions the project for safe community deployment and fosters trust among users and clinicians.

FAQ

Q: Is Derma Vision a diagnostic tool? A: No. Derma Vision, as reported, functions as an early screening and triage tool. It analyses visible skin patterns and offers plain-language recommendations, including basic care tips and suggestions to visit a dermatologist when warranted. Definitive diagnosis requires clinical examination and, where necessary, laboratory or histopathological tests.

Q: How accurate are AI skin apps generally? A: Accuracy varies widely by task, dataset and deployment conditions. In controlled settings, models can perform well at narrow classification tasks, but accuracy often declines in real-world use due to differences in lighting, camera quality and population diversity. Rigorous retrospective and prospective validation is necessary to quantify accuracy in the specific setting where an app will be used.

Q: Should individuals trust an app’s advice instead of seeing a doctor? A: No. Apps can guide users toward care but should not replace professional medical evaluation. If an app recommends seeing a dermatologist, users should follow that advice. If symptoms persist or worsen despite app guidance, consult a clinician.

Q: Does the app store and use participant photos? A: The original coverage did not detail data-handling practices. Any deployment should include clear, informed consent that explains whether photos are stored, how they are used (for diagnosis, model training, quality control), retention periods and deletion procedures. Users should be given the option to decline storage for secondary uses.

Q: How will the students validate the app clinically? A: The students plan to collaborate with dermatologists to validate the application. Validation steps typically include assembling a dermatologist-labelled dataset, running retrospective testing, and conducting prospective clinical studies comparing app outputs with dermatologist evaluations. These studies measure sensitivity, specificity and other performance metrics and are often a prerequisite for regulatory approval.

Q: Can the app work on all skin tones? A: Proper performance across skin tones depends on the diversity of the training dataset and validation process. Historically, many dermatology datasets have underrepresented darker skin tones. Responsible validation must ensure adequate representation and report performance stratified by skin type.

Q: Who bears liability if the app misses a serious condition? A: Liability depends on legal frameworks, the app’s claims and the context of use. Developers should consult legal counsel and consider liability insurance. Clear disclaimers, conservative triage thresholds and clinician involvement reduce risk but do not eliminate legal considerations.

Q: How can other schools replicate a similar project? A: Begin by forming partnerships with local clinicians, securing ethical oversight, and defining clear objectives. Train students in data privacy, basic image acquisition protocols, and user-centred design. Start with low-risk, educational pilots and build pathways for clinical validation before wider deployment.

Q: Will Derma Vision replace dermatologists? A: No. AI tools are designed to augment clinicians, not replace them. The app’s triage function can direct people to timely care, but clinical judgment remains essential for diagnosis and management.

Q: What is the likely timeline to wider deployment? A: Timelines vary. For a responsibly developed health tool, several stages—data collection, retrospective validation, prospective trials, ethical approvals, and regulatory engagement—can take months to years. The project’s current status as a student-built prototype means additional scientific and governance work is necessary before scaling.

Q: How can community members access follow-up care if the app recommends a dermatologist? A: Effective screening programmes include clear referral pathways. Schools and organisers should establish relationships with local clinics or dermatology departments that offer subsidised care or teleconsultations. Community health workers can assist with appointment scheduling and navigating payment or travel barriers.

Q: Are there global standards for AI in healthcare? A: Several organisations provide guidance: the World Health Organization has issued recommendations for digital health technologies; regulatory bodies like the FDA (U.S.), EMA (Europe) and national agencies increasingly provide frameworks for software as a medical device. Industry standards such as ISO 13485 (quality management for medical devices) and ISO/IEC 27001 (information security) inform best practices.

Q: How can the project ensure ethical use of images for training? A: Obtain explicit consent that lists secondary uses, anonymise data, restrict access to authorised personnel, and implement governance structures such as data use agreements and ethics oversight. Consider community advisory boards to review data use plans.

Q: Can Derma Vision help with non-lesion skin problems like infections or dermatitis? A: AI models can be trained to recognise patterns consistent with common inflammatory or infectious skin conditions, but their role is primarily triage. Clinical assessment is still required to confirm diagnosis and prescribe appropriate treatment.

Q: How can someone support or collaborate with the students? A: Clinicians, researchers and institutions interested in collaboration should contact the school or programme organisers. Partnerships could provide dermatology expertise, ethical oversight, data science mentorship, or funding for validation and deployment.

Derma Vision highlights the interplay between youthful curiosity and public health utility. Its early community impact demonstrates the promise of low-cost, school-led innovation. Meaningful expansion will require clinical rigour, careful attention to privacy and bias, and institutional partnerships that turn a classroom achievement into a responsible, effective health tool.