A complete point-of-care system — from a handheld breath sampler to a proprietary optical analyzer to an AI risk report — delivered in under 15 minutes. No needles. No radiation. No specialist required.
Existing tools — imaging, lab-based breath analysis — each carry barriers that reduce access, accuracy, or both. BreathDX was designed specifically to address every one of them.
CT scans and mammograms detect a physical tumor that already exists — meaning cancer is often found only when it has grown large enough to see. 85% of lung cancers are diagnosed at late stage, when survival rates drop dramatically.
Traditional breath collection uses polymer bags (Tedlar®) that suffer from VOC adsorption, permeation, and off-gassing — meaning the sample you analyze is not the sample you collected. Cold-chain requirements add logistics cost and failure points.
Gold-standard VOC analysis (GC-MS) requires large, expensive, lab-bound equipment operated by trained chemists — ruling out point-of-care use. eNose platforms sacrifice molecular specificity. Neither is deployable in a community clinic or mobile unit.
Electronic nose (eNose) devices detect broad VOC patterns with no molecular specificity — prone to interference from diet, environment, and confounders. Reported sensitivity/specificity often falls at 70–80%, insufficient for clinical decision-making.
Mammography and CT are centralized, expensive, and uncomfortable — creating barriers for rural populations, minority communities, and those with religious or cultural hesitancy about physical examination. The result: delayed diagnoses and worse outcomes.
At $100–$250 per mammogram and $200–$500+ for CT, repeat annual screening is a significant financial burden — especially for uninsured or underinsured patients. Cost is consistently cited as a top reason for skipping recommended screening.
A fully integrated point-of-care workflow — from breath collection to clinical probability score — in any clinical setting, no specialist required.
A handheld sampler captures multiple breaths to collect enough VOCs for accurate diagnosis. Disposable mouthpiece. No prep, no discomfort.
The collected sample is loaded into our compact analyzer (~15 lbs) and the target VOCs are identified and quantified at sub-ppb sensitivity.
The BreathDX AI engine analyzes the VOC panel — incorporating patient demographics and clinical risk factors — to produce a calibrated cancer probability.
A clear probability score with supporting biomarker evidence helps the clinician decide: reassure, monitor, or refer for imaging — immediately, at the point of care.
BreathDX is not just software — it is a complete hardware-to-report system, purpose-built for point-of-care cancer screening.
BreathDX gives clinicians a fast, affordable answer for the patients most at risk — the people who need screening most but are least likely to get it.
Current tools are expensive, invasive, slow, or inaccessible to primary care. Our platform closes every gap simultaneously.
| Criterion | Low-Dose CT | Bronchoscopy / Biopsy | Blood-Based Tests | Other Breath Methods | ⚡ BreathDX (AATG) |
|---|---|---|---|---|---|
| Invasiveness | Non-invasive | Highly invasive | Blood draw required | Non-invasive | ✓Exhaled breath only |
| Radiation | Yes — ionising | Minimal (fluoroscopy) | None | None | ✓Zero radiation |
| Time to Result | Days to weeks | Days to weeks | Hours to days | Hours — varies | ✓~15 min, point-of-care |
| Cost per Test | $200–$500+ | $1,500–$8,000+ | $100–$400 | Often research-only | ✓Fraction of mammography or CT cost |
| Primary Care Ready | ✗Specialist facility | ✗Hospital only | ~Lab infrastructure | ✗Research settings | ✓Any clinical setting |
| Multi-Disease | ✗Single modality | ✗Single target site | ~Separate tests needed | ✗Single-disease typical | ✓LC + BC + expandable SW |
| AI Probability Score | ✗Radiologist reads | ✗Pathologist reads | ~Positive / negative only | ~Limited AI integration | ✓AI posterior + report |
| Patient Comfort | Acceptable | Poor — sedation needed | Minor discomfort | High | ✓Breath only — no prep |
| IP Protected | Established methods | Established methods | Competitive landscape | Publication-stage | ✓Full system IP filed |
* R&D-stage platform. Cost and performance figures are projections based on published literature. Clinical validation studies are planned.
When cancer cells grow, they alter cellular metabolism — generating Volatile Organic Compounds (VOCs) that enter the bloodstream and are exhaled. Our proprietary system detects a validated panel across multiple compound classes:
* Specific biomarker identities are proprietary. Full panel available to qualified collaborators under NDA.
Our AI model uses proprietary AI inference to combine multiple factor measurements into a single calibrated probability estimate — exactly as a physician combines test results.
Each disease links to proprietary biomarker panels through a graph database extracted from large scale research. Click a filter or hover a node to explore.
Select clinical risk factors and adjust biomarker panel readings. Watch the AI model probability update in real time. Biomarker identities are anonymised (IP protected).
New conditions are added by updating the biomarker database and AI model — a software update deployable remotely to all devices simultaneously.
Primary development focus. Multi-panel biomarker system with advanced AI model. Database built from large scale researches. Full Weight of evidence (WoE) table computed and validated.
Second application in primary development focus. Separate database and Weight of evidence (WoE) established. Differential profiling built — distinguishes lung cancer vs breast cancer biomarker patterns.
Biomarker profiles differ significantly between COPD, asthma, and healthy — enabling differential respiratory diagnosis from a single sample.
Hepatic dysfunction produces distinct breath signatures. Multi-disease graph architecture supports simultaneous multi-condition screening.
Proprietary metabolic breath markers are well-established indicators of glycemic state — enabling non-invasive glucose monitoring.
Breath biomarkers elevated in renal failure enable early detection through low-cost population screening — no blood draw required.
The AI diagnostic model is fully built and validated. Investment now accelerates the hardware and clinical pathway that brings it to market.
AATG's core expertise in advanced analytical instrumentation extends to security, pharmaceutical, and defense detection systems.
A revolutionary Explosive Trace Detector using advanced spectrometry for high-sensitivity identification of explosive materials. Designed for airports, public venues, and critical infrastructure with no radioactive ionization source.
Next-generation Chemical Warfare Agent Detector offering broad-spectrum detection of CWAs and Toxic Industrial Components. Wearable, battery-powered, designed for continuous monitoring in battlefield and environmental applications.
A fast, on-site cleaning validation analyzer for pharmaceutical manufacturing. Real-time contaminant detection in under 1 minute — eliminating HPLC/MS and dramatically reducing production downtime.
We welcome inquiries from clinical researchers, potential co-founders, investors, and SBIR/STTR program officers. Whether you're interested in the breath diagnostics platform or our other detection technologies, we'd love to hear from you.