Portable, AI-Assisted Oral Cancer Detection Device
A handheld optical coherence tomography probe with an embedded AI lesion classifier moved from an undetermined regulatory path to two international grant awards in seven months.
- Origin
- ICMR-affiliated cancer research centre; spin-out of two oncologists and one applied AI engineer
- Founded
- 2021
- Technology Readiness
- TRL 4 at engagement: laboratory-validated, no clinical trial initiated
- Core Innovation
- Handheld optical coherence tomography probe with embedded AI lesion classifier; sensitivity 91%+ in controlled testing.
- Capital Outcome
- ₹3.20 Crore: NIH SBIR Phase I (USD ~$150,000) and Wellcome Trust Innovations for Global Health award
Diagnostic findings at engagement
- The patent was granted in India only; no freedom-to-operate analysis had been conducted against prior art registered in the United States or European Union, the two jurisdictions targeted for grant funding.
- No CDSCO regulatory classification had been determined for the device. The founders could not state whether the product would be Class B, C, or D Software as a Medical Device under the 2017 Medical Device Rules.
- A prior NIH SBIR application had been rejected; the rejection letter cited absence of a commercialisation and translation plan, a standard SBIR Phase I review criterion.
- No cost or pricing model existed. Manufacturing cost could be stated only as an estimate, with no breakdown between bill-of-materials cost, assembly cost, and target gross margin.
- Market sizing in prior materials consisted of a single figure: approximately 100,000 new oral cancer cases annually in India, with no segmentation by care setting, procurement channel, or buyer type.
Service interventions applied
Regulatory Strategy
Determined CDSCO classification as Class C Software as a Medical Device under the Medical Device Rules 2017. Identified the FDA De Novo pathway as the relevant U.S. regulatory anchor required for SBIR eligibility, and prepared a written regulatory roadmap submitted as an appendix to the grant application.
Grant Narrative: NIH SBIR
Rewrote the Specific Aims page in full. The original framing described the technology; the revised framing opened with the unmet clinical need and positioned the device against NIH's PA-21-262 medical device funding announcement, with an explicit commercialisation and translation section addressing the prior rejection reason directly.
Grant Narrative: Wellcome Trust
Positioned the application under the Innovations for Global Health funding strand. Quantified projected impact in disability-adjusted life years averted per 1,000 screenings, calculated against Wellcome's published cost-effectiveness benchmarks for early cancer detection interventions.
Financial Model Construction
Built bottom-up unit economics: bill-of-materials cost of ₹18,400 per unit; target average selling price of ₹85,000 for institutional B2B sale and ₹42,000 for government procurement. Modelled a three-year revenue scenario tied to projected deployment across Primary Health Centres under the National Health Mission.
Intellectual Property Strategy
Commissioned a freedom-to-operate analysis covering the United States, European Union, and United Kingdom. Filed a PCT application to extend patent coverage into these jurisdictions, a contractual requirement under the NIH SBIR commercialisation clause. Two areas of claim overlap with prior art were identified and resolved through claim narrowing prior to filing.
Market Research
Re-sized the addressable market by procurement channel: 156,000 Primary Health Centres and Sub-Centres under the National Health Mission, 28,000 registered dental clinics, and 4,200 dedicated cancer treatment hospitals, each with distinct procurement cycles and pricing tolerance documented separately.
Observation: Narrative Construction
The original NIH application opened: “We have developed an AI-powered device for oral cancer detection.” The revised application opened: “Oral cancer kills more Indian men annually than cervical cancer kills Indian women, yet no portable, affordable, community-deployable screening device exists at scale anywhere in South Asia.” The distinction between describing a product and quantifying the consequence of its absence is the most consistent pattern across the grant-funded engagements in this portfolio.
Position at engagement vs. position at close
Swipe the table sideways to compare.
| Dimension | At engagement | At close |
|---|---|---|
| Grant Applications | 2 submitted, both rejected | 4 submitted, 2 awarded |
| Regulatory Classification | Undetermined | CDSCO Class B SaMD; FDA De Novo pathway mapped |
| IP Jurisdiction Coverage | India only | PCT filed: India, United States, European Union, United Kingdom |
| Financial Model | None | Three-year P&L with channel-specific unit economics |
| Capital Secured | ₹0 | ₹3.20 Crore |