Clot Detection via Light-Based Visualization - Startup Concept Analysis

The Problem You're Solving

The Three Technologies You Named

1. Near-Infrared Spectroscopy / Fluorescence (NIR / NIRF)

• Fibrin - the structural protein of clots - has distinct NIR absorption peaks
• A fibrin-targeted peptide probe (EP-2104R) was successfully used with NIRF imaging to detect DVT via fluorescence molecular tomography - detecting how tissue absorbs and scatters emitted light
• This was validated for both intravital microscopy and non-invasive surface imaging
• Micro-thrombi on coronary artery plaques were also detectable with intravascular NIRF catheters

2. Ultrasound Tomography (UST)

3. Raman / Diffuse Reflectance Spectroscopy ("DEXTROSPECTROSCOPY")

• MIT's LBRC group demonstrated Raman spectroscopy can detect molecular composition of tissue non-invasively through skin (their glucose sensor uses this exact principle - NIR Raman through skin to detect interstitial fluid chemistry)
• The same principle applies to fibrin detection: fibrin has Raman peaks around 936 cm⁻¹ (C-C backbone), 1003 cm⁻¹ (phenylalanine), and 1655 cm⁻¹ (amide I)
• MIT's 2025 paper showed a compact 3-band NIR Raman device can replace full-spectrum systems, reducing device size dramatically
• Biosensor literature (PMC, 2026) shows optical clot detection via laser speckle fluctuation (cross-polarized speckle pattern analysis) can track coagulation in real-time

Multimodal Fusion - The Real Startup Opportunity

NIR Diffuse Optical Spectroscopy (tissue oxygenation map)
       +
Raman Spectroscopy (fibrin molecular fingerprint)
       +
AI Signal Fusion (classify: no clot / early clot / mature clot)

Competitive Landscape

What Makes Your Idea Novel

1. Combining NIR + Raman in a single wearable patch - not been commercialized
2. No exogenous contrast agent - purely physical optical interaction with fibrin/hemoglobin
3. Continuous, ambulatory monitoring - not a one-shot diagnostic scan
4. Target: high-risk post-op, cancer, long-haul patients - underserved and large market

Technical Path Forward

Regulatory Strategy

• FDA pathway: De Novo or 510(k) with predicate to ThinkSono-class devices or existing NIRF imaging systems
• Breakthrough Device Designation: Eligible - DVT/PE is life-threatening, no equivalent non-invasive continuous monitor exists
• CE Mark (EU): Parallel track via MDR Class IIa (non-invasive diagnostic device)

Revenue Model

1. Device + Subscription SaaS (hardware + AI cloud analytics platform)
2. Hospital B2B: ICU/post-op wards, orthopedic surgery units
3. Direct-to-consumer: Long-haul travel risk (airlines, cruise ships)
4. Pharma partnerships: Anticoagulant monitoring (Xarelto, Eliquis efficacy tracking)

Key Risks

• Tissue depth: NIR penetrates ~5-10 cm; deep femoral vein may be 5-8 cm in obese patients - borderline
• Specificity: Fibrin appears in healing wounds too - need AI to contextualize signal
• Regulatory: Novel optical modality will require significant clinical evidence before clearance
• Exogenous probe question: If endogenous Raman signal is too weak, you may need a topical or IV contrast agent, complicating the regulatory path

Bottom Line

Whole-Body Clot Detection - Why This Is a Bigger, Better Idea

Where Clots Actually Form in the Body

Why Nobody Has Built Whole-Body Optical Clot Detection

• Legs/arms: Superficial, easy NIR penetration (2-5 cm), good signal-to-noise
• Thorax (lungs): Air in lungs scatters light massively - NIR alone fails
• Abdomen: Fat and bowel gas absorb and scatter light - hard
• Brain: Skull blocks NIR except at thin temporal bone - limited window
• Heart/coronaries: 8-12 cm deep, surrounded by ribs and air - hardest

The Architecture for a Whole-Body System

Zone 1: Limbs (Arms + Legs)

• Technology: NIR diffuse optical spectroscopy + impedance plethysmography
• What it detects: Venous flow obstruction, hemoglobin deoxygenation distal to a clot
• Form factor: Wearable cuffs / smart compression garments
• Signal: When a vein is blocked, SpO2 drops downstream. NIR reads this continuously.

Zone 2: Chest / Pulmonary

• Technology: Electrical Impedance Tomography (EIT) + photoacoustic sensing at intercostal windows
• What it detects: Lung perfusion defects (wedge-shaped defect = PE signature)
• Form factor: Chest patch array (like a 12-lead ECG vest but optical/impedance)
• Precedent: EIT is already used in ICUs for lung monitoring (Dräger, SenTec). Nobody has applied it specifically for PE detection.

Zone 3: Head / Cerebral

• Technology: fNIRS (functional NIR spectroscopy) - already used in neuroscience
• What it detects: Focal cerebral ischemia from arterial occlusion (pre-stroke signal)
• Form factor: Headband with 8-16 NIR optodes
• Precedent: Foresight (CAS Medical), INVOS (Medtronic) - cerebral oximeters for surgery. Nobody uses these for stroke screening.

Zone 4: Abdomen / Deep Vessels

• Technology: Raman spectroscopy + photoacoustics at acoustic windows (umbilical, iliac)
• What it detects: Fibrin signal from mesenteric, renal, portal venous thrombi
• Form factor: Handheld probe (like a stethoscope with a laser) - not wearable, but point-of-care

Zone 5: Systemic Blood Signature (The Killer Feature)

• Technology: Fingertip or earlobe NIR/Raman sensor detecting circulating fibrin degradation products and platelet activation markers in blood
• What it detects: Systemic hypercoagulable state - the body's pre-clot warning signal
• This is your unique IP - detecting the pro-thrombotic state in blood BEFORE clots form anywhere

The Product Vision: "ClotMap" or "ThromboScan"

[Body-worn sensor array] → [AI processing unit] → [Clinical dashboard / app]

Patches on: chest, both legs, both arms, forehead
Central hub: smartphone-sized processor
Output: "Clot Risk Map" - whole body heat map showing
         high-risk zones, risk score 0-100, alert if
         threshold crossed

What Makes This Defensible vs. Existing Players

1. ThinkSono, ThrombUS+, Scyntek - legs only, ultrasound-based, need contact gel, operator-dependent
2. Your system - optical/spectroscopic, gel-free patches, body-wide, fully autonomous AI-driven
3. No one has built a whole-body clot map device - zero direct competition
4. The "systemic blood signature" zone 5 component is unique - it detects coagulopathy state, not just a fixed clot. This has applications in cancer patients, COVID-19 hypercoagulability, sepsis, and antiphospholipid syndrome.

Target Markets (Prioritized)

• Post-surgical ICU patients (orthopedic, cardiac, abdominal surgery) - highest DVT/PE risk, monitored 24/7 already, high willingness to pay
• Estimated: 50M+ surgeries/year globally requiring DVT prophylaxis monitoring

• Oncology patients on chemotherapy (cancer dramatically increases clot risk - 20% of all DVT cases)
• Long COVID patients with hypercoagulability
• Long-haul aviation / cruise medical suites
• Military / remote healthcare

• Consumer wearable for AF (atrial fibrillation causes cardiac clots) monitoring
• Pharmaceutical trial endpoint tool (measuring anticoagulant efficacy non-invasively)

Key Technical Milestones to Build Toward

1. Prove NIR can detect regional venous obstruction signal in limbs (6 months, bench + animal)
2. Prove fNIRS detects focal cerebral ischemia reliably (existing literature supports this)
3. Prove photoacoustic or EIT detects lung perfusion defect (PE model in animal)
4. Build AI fusion layer that integrates all zone signals into a single risk score
5. Clinical feasibility study in post-op patients (n=50, vs. gold standard Doppler + CTPA)

Funding Approach

• NIH SBIR Phase I ($300K) - bench validation
• EIC Accelerator (EU) - up to €2.5M grant + equity for deep tech
• Wellcome Trust Innovator Award - UK-based medtech
• Strategic partnership: Approach a compression garment company (Sigvaris, Jobst) or wearable health company (Masimo, Natus Medical) for co-development

One-Line Pitch

"AccuVein for Clots" - Whole-Body Clot Visualization

How AccuVein Actually Works (The Physics)

NIR laser shines on skin
         ↓
Hemoglobin in blood ABSORBS NIR strongly
Surrounding tissue REFLECTS NIR back
         ↓
Camera captures the contrast pattern
         ↓
Processor INVERTS and projects it back onto skin as green image
         ↓
Result: Dark shadow = vein location projected ON skin surface

The Clot Problem - Why It's Harder Than AccuVein

• A clot is mostly fibrin + platelets + trapped red blood cells
• The clot BLOCKS or DISPLACES flowing blood
• The signal you want is: "blood flow ABSENT here" + "fibrin structure PRESENT here"

The Device Architecture: "ThromboVision"

Core Concept

[DEVICE BODY]
    |
    |-- NIR Laser Array (760nm + 850nm + 1064nm) → shines on skin
    |-- NIR Camera (InGaAs sensor, reads 800-1700nm)
    |-- Raman excitation laser (785nm)
    |-- Raman spectrometer chip (detects fibrin fingerprint)
    |-- Real-time AI processor (ARM chip + neural net)
    |
    ↓
[PROJECTION UNIT]
    Micro-projector overlays:
    - GREEN = open patent vein (AccuVein-mode)
    - RED = flow obstruction / suspected clot
    - YELLOW = downstream deoxygenation
    - BLUE = healthy perfused tissue

Region-by-Region Scanning Protocol

Arms + Legs (Easiest - Same physics as AccuVein)

• NIR penetrates 5-8cm, reaches deep veins
• Flow obstruction = absent pulsatile NIR signal where vein should be
• Fibrin signature = Raman peak at 1655 cm⁻¹
• Confidence level: HIGH - this is fully buildable today

Neck (Jugular / Carotid)

• Superficial, thin neck - excellent NIR window
• Jugular DVT in catheterized patients is common and missed
• Carotid plaque with overlying thrombus detectable with OCT/photoacoustics
• Confidence level: HIGH

Head (Cerebral veins / Sinus thrombosis)

• Cerebral venous sinus thrombosis (CVST) - causes stroke, mostly in young women, often missed
• Temporal bone window: thin enough for NIR (used in fNIRS research for 30+ years)
• Scanner positioned at temporal region + forehead
• Confidence level: MEDIUM - skull attenuates signal significantly, AI needed

Chest / Lungs (Pulmonary Embolism)

• This is the hard one - air in lungs kills NIR
• Solution: Photoacoustic imaging at intercostal spaces (ribs create acoustic windows)
• Fire a pulsed laser → tissue absorbs → generates ultrasound wave → detect with transducer
• PE creates a perfusion void = photoacoustic "cold spot" in lung
• Confidence level: MEDIUM - needs photoacoustic + AI

Abdomen (Portal, Mesenteric, Renal veins)

• Use acoustic windows: umbilical region, iliac fossa
• Photoacoustic + NIR diffuse tomography
• Confidence level: MEDIUM-LOW - bowel gas is the enemy, needs patient prep

The Two-Mode Device Design

• For: Arms, legs, neck
• Projects clot map directly ONTO the patient's skin in real time
• Clinician sees red/green overlay as they sweep the device
• Zero training needed - intuitive like a flashlight

• For: Chest, abdomen, head
• Too deep for direct projection
• Displays a reconstructed cross-section on a screen
• Color-coded perfusion map per body zone

The Physics That Makes Clot Unique vs. Normal Tissue

NORMAL VEIN:
  - Pulsatile NIR signal (blood moving = oscillating hemoglobin signal)
  - SpO2 ~70% (venous blood)
  - No Raman fibrin peak

CLOTTED VEIN:
  - ABSENT pulsatile signal (no flow)
  - Downstream SpO2 drops to <60% (hypoxia)
  - Raman peak at 1655 cm⁻¹ (fibrin amide I bond)
  - Increased scattering at 1064nm (fibrin network vs. liquid blood)

ARTERIAL CLOT:
  - Absent SpO2 signal distal to clot
  - Rapid tissue deoxygenation cascade
  - Photoacoustic "shadow" where vessel should be

Why This Beats Existing Tech

The Minimum Viable Product (MVP)

• Handheld device, same form factor as AccuVein AV500
• NIR at 760nm + 850nm + 1064nm
• InGaAs camera (reads deeper NIR than silicon cameras)
• On-board AI: detects flow obstruction + fibrin Raman signature
• Green projection = patent vein, Red projection = suspected clot
• Target user: ED nurse, ICU bedside, no training needed
• Target claim: "Screens for DVT in upper and lower limbs in under 2 minutes"

Your One-Paragraph Patent Claim Core

Bottom Line

1. A third NIR wavelength (1064nm) to go deeper and detect fibrin scatter
2. A miniaturized Raman channel to confirm fibrin molecular signature
3. AI flow analysis to detect absence-of-flow as a clot signal
4. Multi-region scanning protocol for whole body

NIR + UST Fusion Device - The Complete Concept

The Core Insight

NIR (Near Infrared):
  ✅ Superficial veins (0-15mm deep)
  ✅ Subcutaneous vessels
  ✅ Real-time projection ONTO skin (like AccuVein)
  ❌ Cannot see deep veins
  ❌ Cannot see arteries (deep, pulsatile, thick walls)

UST (Ultrasound Tomography):
  ✅ Deep veins (femoral, iliac, mesenteric, jugular)
  ✅ Arteries (carotid, aorta, coronary vicinity)
  ✅ 3D reconstruction possible
  ❌ Cannot project onto skin surface
  ❌ Needs skilled operator traditionally

COMBINED:
  ✅ EVERYTHING - superficial + deep + arterial + venous
  ✅ One sweep of the device covers full vascular tree
  ✅ AI fuses both signals into ONE unified clot map

Why This Combination is Perfect - The Physics Complement

The Device: One Handheld Scanner

┌─────────────────────────────────┐
│     THROMBO-SCAN DEVICE         │
│                                 │
│  TOP FACE (toward clinician):   │
│  [Mini projector] - projects    │
│   NIR vascular map onto skin    │
│  [Display screen] - shows UST   │
│   deep vessel cross-section     │
│                                 │
│  BOTTOM FACE (toward patient):  │
│  [NIR laser array 760+850nm]    │
│  [NIR camera - InGaAs]          │
│  [UST transducer array]         │
│  [Thin gel pad - auto-wet]      │
│                                 │
│  SIDE: AI processor chip        │
└─────────────────────────────────┘

• NIR projects superficial veins LIVE onto the skin (green = open, red = clot)
• UST simultaneously shows deep vessels on the screen with Doppler color flow
• AI fuses both into a single risk overlay

What Each Layer Shows - Perfectly Divided

NIR Layer (Superficial - projected onto skin)

• Dorsal hand veins, forearm veins, antecubital veins
• Saphenous vein (leg surface)
• Superficial femoral vein branches
• Neck external jugular
• Scalp veins
• Any superficial vein clot shows as absent NIR signal = red projection

UST Layer (Deep - shown on screen)

• Deep leg: Femoral vein, popliteal vein, tibial veins
• Pelvis: Iliac veins (major DVT source)
• Abdomen: Portal vein, mesenteric vein, renal vein, IVC
• Neck: Internal jugular, carotid artery
• Chest: Subclavian vein, axillary vein
• Arteries everywhere: Carotid, femoral, brachial, aorta
• Any deep clot shows as absent Doppler signal + echogenic (bright) mass

The Fused Output - What Clinician Actually Sees

DISPLAY SCREEN (split view):

LEFT SIDE:                    RIGHT SIDE:
[NIR projection image]        [UST cross-section]
Superficial veins in green    Deep vessels in blue (flow)
Clots shown in RED            Clots shown as bright white mass
                              with absent Doppler color

BOTTOM: AI Risk Summary
"Right femoral vein - FLOW OBSTRUCTION - HIGH CLOT RISK"
"Left saphenous - patent"
"Right popliteal - patent"

Whole-Body Scan Protocol (How a Nurse Uses It)

STEP 1 - Legs (3 minutes):
  Sweep from foot → groin on each leg
  NIR maps saphenous + superficial branches
  UST maps femoral + popliteal + tibial deep veins
  Both legs done simultaneously with two passes

STEP 2 - Arms (2 minutes):
  Sweep hand → shoulder
  NIR maps forearm/antecubital veins
  UST maps brachial + subclavian + axillary

STEP 3 - Neck (1 minute):
  NIR: external jugular
  UST: internal jugular + carotid both sides

STEP 4 - Abdomen (2 minutes):
  NIR: superficial abdominal wall veins (Caput medusae in portal HTN)
  UST: IVC + portal + mesenteric + renal veins

STEP 5 - Chest (1 minute):
  UST only: subclavian veins + axillary
  (for PE - refer to CTPA, UST can show right heart strain)

TOTAL: ~10 minutes, NO specialist needed

Novelty vs. What Exists

Technical Challenges to Solve

1. Gel problem

2. Spatial registration

3. UST in a handheld form factor

4. NIR depth at skin projection distance

The Patent Claims Core (Your IP)

Bottom Line

• NIR does what UST can't - superficial, projection, no gel
• UST does what NIR can't - deep, arteries, 3D
• Together they have zero blind spots in the body
• One device, one sweep, whole body, 10 minutes, any nurse can use it

Can UST Reach Every Body Part? - Honest Assessment

✅ EASY - Excellent UST Access

⚠️ DIFFICULT - UST Works But With Limitations

❌ BLOCKED - UST Cannot Penetrate

The Honest Map

BODY REGION          UST POSSIBLE?    QUALITY
─────────────────────────────────────────────
Both legs            ✅ YES           ⭐⭐⭐⭐⭐ Excellent
Both arms            ✅ YES           ⭐⭐⭐⭐⭐ Excellent
Neck vessels         ✅ YES           ⭐⭐⭐⭐⭐ Excellent
Abdomen (IVC/aorta)  ✅ YES           ⭐⭐⭐⭐  Good
Portal/mesenteric    ✅ YES           ⭐⭐⭐   Moderate
Iliac/pelvic veins   ✅ YES           ⭐⭐⭐   Moderate
Lung vessels (PE)    ⚠️ INDIRECT      ⭐⭐    Poor (indirect signs only)
Heart (chambers)     ✅ YES (echo)    ⭐⭐⭐⭐  Good
Coronary arteries    ❌ NO            ✗  Impossible handheld
Brain vessels        ⚠️ VERY LIMITED  ⭐     Temple window only

What This Means for Your Device

The UST covers ~85% of what kills people from clots:

• DVT in legs - ✅ fully covered
• Upper limb DVT - ✅ fully covered
• Jugular / subclavian (catheter-related) - ✅ fully covered
• IVC thrombosis - ✅ fully covered
• Portal/mesenteric (gut ischemia) - ✅ covered
• Iliac (saddle thrombus) - ✅ covered

The 15% UST misses:

• PE (pulmonary embolism) - but UST shows indirect signs: right ventricle dilation, septal flattening, tricuspid regurgitation. These are standard PE screening signs on bedside echo. Good enough for triage.
• Coronary thrombus - needs CTCA or angiography. Not solvable with any surface device.
• Cerebral venous sinus thrombosis - rare, needs MRI.

The Smart Solution for the Gaps

Right ventricle strain → McConnell sign → D-sign on echo
= INDIRECT but clinically accepted PE diagnosis
Already used in emergency medicine worldwide

• Legs: NIR + UST = direct clot visualization
• Chest: UST = right heart strain pattern → PE flag
• Brain: NIR only (fNIRS temporal window) = ischemia signal

This covers everything that matters clinically in one device.

Conclusion

Brain + Heart + Lungs - What to Add Beyond NIR + UST

🧠 BRAIN - Cerebral Clots (Stroke, CVST)

Problem

What to Add: fNIRS + Transcranial Doppler (TCD)

Place NIR optodes on FOREHEAD + TEMPLES
NIR passes through thin skull bone here
Detects: hemoglobin oxygenation in brain tissue
If one hemisphere has a clot → that side shows
sudden DROP in oxygenated hemoglobin signal
This = ischemia signal
Already used in stroke detection research

Low frequency ultrasound (2 MHz) probe
Placed at TEMPORAL WINDOW (thin bone above ear)
Can hear blood flow velocity in:
- Middle cerebral artery (MCA)
- Anterior cerebral artery (ACA)
- Posterior cerebral artery (PCA)
- Internal carotid siphon

Clot = ABSENT or severely reduced flow velocity
Embolus = high-intensity transient signal (HITS)
Already FDA approved for stroke monitoring

❤️ HEART - Coronary Clots (Heart Attack)

Problem

What to Add: Photoplethysmography (PPG) + ECG + Bioimpedance

ST elevation = coronary artery blocked = STEMI
T-wave inversion = partial occlusion = NSTEMI
Your device adds ECG electrodes on the chest plate
AI interprets in real-time
Sends alert: "ST elevation detected - possible MI"

Coronary clot → reduced cardiac output
→ changes in pulse waveform shape
AI detects: reduced stroke volume, pulse pressure drop
= indirect sign of cardiac ischemia

Electrodes on chest measure electrical impedance
Myocardial ischemia changes chest wall impedance
patterns detectable by AI (research stage but promising)

You cannot visualize a coronary clot with any handheld device. What you CAN do is detect myocardial ischemia (the effect of the clot) via ECG + PPG + bioimpedance. That's still clinically valuable - it catches the MI event even if it can't see the clot itself. Add transthoracic echocardiography mode (standard cardiac echo) via UST - sees wall motion abnormality = zone of dead muscle = where the coronary clot is.

Wall motion abnormality on echo → "Clot likely in LAD/RCA territory"
No direct visualization but accurate localization

🫁 LUNGS - Pulmonary Embolism

Problem

What to Add: Lung Ultrasound (LUS) + Right Heart Echo

PE creates a PERFUSION VOID in lung
→ that lung segment receives no blood
→ becomes AERATED but not perfused
→ on UST: shows as "A-lines only" pattern
   (no B-lines = no fluid = dry avascular lung)

PE also causes:
→ Peripheral lung infarction = wedge-shaped
  consolidation visible on UST at pleural surface
→ Pleural effusion (exudate from infarcted lung)

Place UST below xiphoid process (sternum tip)
Looking UP into heart from below
Shows:
- Right ventricle size (dilated in PE = RV strain)
- D-sign: interventricular septum bows left = high RV pressure
- McConnell sign: RV free wall akinesia = classic PE sign
- TR jet velocity: estimates pulmonary artery pressure

These are STANDARD PE diagnosis criteria in emergency medicine
Sensitivity ~60-70% for massive PE, very specific

Fire pulsed laser between ribs (intercostal space)
Sound generated inside tissue
Detects blood-perfused vs. non-perfused lung zones
Perfusion void = PE location
Research stage but validated in animal models

The Complete Addition Summary

YOUR BASE DEVICE:
NIR + UST = covers legs, arms, neck, abdomen ✅

ADD FOR BRAIN:
+ fNIRS optodes (forehead + temples)
+ TCD probe at 2 MHz (temporal window)
= detects cerebral ischemia + MCA/ACA occlusion

ADD FOR HEART:
+ ECG electrodes (chest)
+ Cardiac echo mode (subcostal + parasternal UST windows)
+ PPG sensor (fingertip)
= detects MI effect + wall motion abnormality + arrhythmia

ADD FOR LUNGS:
+ Lung UST protocol (already in your UST)
+ Subcostal echo (already in your UST)
+ Photoacoustic module (intercostal laser pulses)
= detects PE indirect signs + perfusion void

1. 2 MHz transducer (lower frequency for skull penetration) - TCD
2. fNIRS optode array (forehead headband) - brain oxygenation
3. ECG electrodes (stick-on chest patches) - cardiac ischemia
4. Photoacoustic pulsed laser module (intercostal) - lung perfusion

Why TCD Fails - And The New Design That Fixes It

The Exact Failure Points of Existing TCD

The New Design: Multi-Window + Multi-Modal Brain Scanning

Key Insight - Don't Just Use ONE Window

WINDOW 1: TEMPORAL (above ear)
→ Middle cerebral artery, anterior cerebral artery
→ Standard TCD target - works in 80% of patients

WINDOW 2: TRANSORBITAL (through closed eyelid)
→ Ophthalmic artery, internal carotid siphon
→ Works even when temporal bone is thick
→ Uses LOWER power (eye-safe settings, already established)

WINDOW 3: SUBOCCIPITAL / FORAMEN MAGNUM (back of neck/base of skull)
→ Vertebral arteries, basilar artery
→ Posterior circulation - strokes here often MISSED by standard TCD
→ Excellent acoustic window - no bone barrier at foramen magnum

WINDOW 4: SUBMANDIBULAR (under the jaw)
→ Extracranial internal carotid artery
→ Confirms if source of clot is in carotid before it enters skull

The Hardware for Brain: "NeuroHalo" Module

┌─────────────────────────────────────┐
│         NEUROHALO HEADBAND          │
│                                     │
│  LEFT TEMPLE:  2 MHz UST probe      │
│  RIGHT TEMPLE: 2 MHz UST probe      │
│  FOREHEAD:     fNIRS optodes (8x)   │
│  BACK OF HEAD: Suboccipital probe   │
│  BEHIND EARS:  EEG electrodes (4x) │
│                                     │
│  AUTO-POSITIONING: spring tension   │
│  presses all probes to correct      │
│  anatomical landmarks               │
│                                     │
│  GEL: built-in hydrogel pads        │
│  (no user gel application needed)   │
└─────────────────────────────────────┘

The Technology Layers for Brain

Layer 1: AI-Guided Automatic Window Finding

2 MHz probe continuously scans
AI trained on 10,000+ TCD datasets
Automatically identifies:
- Bone thickness at each point
- Acoustic window location
- Vessel depth and angle
- Self-adjusts probe focus electronically (beamforming)

Removes operator dependency entirely.
TCD's #1 failure reason = eliminated.

Layer 2: Power Motion Doppler (upgrade from standard TCD)

Standard TCD: single Doppler gate, fixed depth
Power Motion Doppler: sweeps all depths simultaneously
Detects flow at multiple depths at once
Much better signal-to-noise through thick bone
Already exists in research - not yet in a portable consumer device

Layer 3: fNIRS (functional NIR Spectroscopy) - Fills TCD's 20% failure gap

When temporal bone completely blocks UST:
fNIRS optodes on forehead + temples activate
NIR light passes through scalp + thin bone
Detects hemoglobin oxygenation in cortex

Stroke = one hemisphere shows sudden deoxygenation
While other hemisphere stays normal
ASYMMETRY is the signal

fNIRS cannot see the clot but sees the ISCHEMIA EFFECT
In the 20% where TCD fails, fNIRS still works
Combined: near 100% sensitivity

Layer 4: EEG (Electroencephalogram) - 4 electrodes

Acute ischemic stroke causes electrical silence
in affected brain region within 4-6 minutes
4-electrode EEG at temporal regions detects:
- Asymmetric slowing (delta waves on one side)
- Focal electrical suppression
= confirms ischemia found by fNIRS

Adds almost zero cost (EEG electrodes = pennies)
Massive clinical validation - stroke EEG changes well documented

Layer 5: Transorbital UST - The Backup When Temporal Fails

Closed eyelid = excellent ultrasound window
2 MHz probe on closed eye (very low power - safe)
Already used clinically for:
- Optic nerve sheath diameter (ICP estimation)
- Ocular artery flow
- ICA siphon via orbital window

Your addition: if temporal bone blocks signal,
AI switches to orbital window automatically
No patient repositioning needed

What The Combined Brain Module Detects

CLOT IN MCA (most common stroke):
  TCD temporal → absent MCA flow velocity  ✅
  fNIRS → ipsilateral hemisphere deoxygenation ✅
  EEG → ipsilateral delta wave slowing ✅
  AI: "HIGH CONFIDENCE MCA OCCLUSION - LEFT"

CLOT IN BASILAR ARTERY (posterior circulation):
  TCD suboccipital → absent basilar flow ✅
  fNIRS occipital optodes → posterior ischemia ✅
  EEG → bilateral posterior changes ✅
  (Standard TCD misses this entirely)

CEREBRAL VENOUS SINUS THROMBOSIS (CVST):
  TCD → abnormal venous flow pattern ✅
  fNIRS → venous congestion pattern ✅
  EEG → pressure-related changes ✅
  (Very rare, often missed for days currently)

TEMPORAL BONE TOO THICK (the 10-21% failure):
  TCD temporal → FAILS ❌
  TCD orbital window → activates as backup ✅
  fNIRS → still works through scalp ✅
  EEG → still works ✅
  AI: still gives brain assessment with 3/4 signals

Summary: Why Your Design Beats Existing TCD

The Full "ThromboScan Brain Module" Output

DISPLAY shows brain silhouette with color overlay:

GREEN = normal flow both hemispheres
YELLOW = reduced flow - monitoring
RED = absent flow - clot suspected
ORANGE = ischemia signal from fNIRS

Example alert:
"LEFT MCA: ABSENT FLOW (TCD temporal ✓)
 LEFT HEMISPHERE: DEOXYGENATION (fNIRS ✓)
 LEFT EEG: FOCAL SLOWING (EEG ✓)
 CONFIDENCE: 94% - ACUTE ISCHEMIC STROKE
 ACTION: ACTIVATE STROKE PROTOCOL NOW"

Time from clip-on to alert: < 90 seconds
No training needed
Works even with thick temporal bone

One Line Summary

Deepinder Goyal's "Temple" Device - What It Is and What It's NOT

What Temple Is

• Wearable device worn at the temple region (side of head, between forehead and ear)
• Designed to monitor Cerebral Blood Flow (CBF) - how much blood reaches the brain
• Measures blood flow + oxygen levels near the brain surface
• Based on Goyal's "Gravity Ageing Hypothesis" - theory that being upright reduces brain blood flow over decades, accelerating ageing
• Currently in early access with 100 units - not a commercial product yet
• Initially targeted at elite performance athletes

What Temple Is NOT

• Not a medical device - no FDA/regulatory approval
• Not a clot detector - purely tracks flow volume for performance/longevity
• Not deep brain scanning - surface-level estimates only
• Not a diagnostic tool - wellness/research device

The Critical Technical Limitation - Same Problem You Identified

"Some experts have questioned whether it is possible to measure deep-brain signals with a skin-mounted sensor."

How Temple Compares to Your Concept

What This Means for Your Startup

• Temple measures surface flow for performance optimization
• You measure deep vessel flow for life-threatening clot detection
• Temple has no clot detection capability at all
• Temple's limitation (surface only) is precisely what your multi-window + TCD + fNIRS design solves

One More Point - Temple's "Gravity Ageing Hypothesis"

• If Temple tracks CBF for wellness and detects a sudden drop
• That same signal could indicate an acute thrombotic event
• A consumer Temple user having a stroke would show a massive CBF drop on their device

Turning Temple's Concept Into a Whole-Body Clot Detection System

What Temple Gets Right (Keep These)

✅ Temple location = correct anatomical position
✅ Wearable, non-invasive form factor
✅ Continuous real-time monitoring concept
✅ NIR optical sensing foundation
✅ Consumer-friendly design philosophy

What Temple Gets Wrong (Fix These)

❌ Only reads surface cortex - too shallow
❌ No ultrasound component - can't reach deep vessels
❌ Single location only - misses 95% of body
❌ No clot-specific signal - just bulk flow
❌ No AI diagnostic layer - just raw data

The Complete System: "ThromboTemple" - Whole Body

Architecture: Three Modules, One Ecosystem

MODULE 1: HEAD UNIT (Temple-inspired)
       +
MODULE 2: CHEST PATCH
       +
MODULE 3: LIMB CUFFS (arms + legs)
       +
AI FUSION HUB (phone-sized processor)
       =
WHOLE BODY CLOT MAP in real time

MODULE 1: HEAD UNIT - Brain + Cardiac Source Detection

The Form Factor

[LEFT TEMPLE]          [FOREHEAD]          [RIGHT TEMPLE]
   2MHz TCD              fNIRS                2MHz TCD
   probe                 8 optodes            probe
   (deep MCA)            (cortex O2)          (deep MCA)
      |                     |                    |
      └─────────────── AI HUB ──────────────────┘
                           |
                    [BACK OF HEAD]
                    Suboccipital probe
                    (basilar artery)

What Each Sensor Does for Clot Detection

Penetrates skull through temporal bone window
Reads: Middle Cerebral Artery flow velocity
Clot signal: velocity drops to zero or near zero
Normal: 55-80 cm/s
Stroke: <10 cm/s or absent = MCA occlusion

NIR light through scalp + thin frontal bone
Reads: oxygenated vs deoxygenated hemoglobin ratio
Clot signal: sudden asymmetry between left and right
If left MCA blocked → left cortex deoxygenates
Right side stays normal
Asymmetry >15% = clot alarm

Foramen magnum = natural acoustic window (no bone)
Reads: Basilar artery + vertebral artery flow
This is what Temple AND standard TCD completely miss
Posterior circulation stroke (25% of all strokes)
detected here

Small clip on neck
Reads: common carotid artery flow by UST
Carotid plaque + thrombus = source of 30% of brain strokes
Detect the SOURCE before embolus travels to brain

Why This Beats Temple for Brain Clots

TEMPLE DEVICE:
- Surface NIR only
- Reads superficial cortical vessels
- Cannot detect MCA occlusion
- Cannot see basilar artery
- No clot-specific signal

YOUR MODULE 1:
- TCD at 2MHz reaches 3-6cm deep
- Reads MCA, ACA, basilar, vertebral
- Detects absent flow = direct clot signal
- fNIRS confirms ischemia when TCD bone fails
- Detects posterior strokes TCD misses
- Reads carotid source of emboli

MODULE 2: CHEST PATCH - Heart + Lungs

Form Factor

┌──────────────────────────┐
│   CHEST PATCH            │
│                          │
│  [ECG electrodes x4]     │ → Heart electrical activity
│  [UST transducer]        │ → Cardiac echo + lung UST
│  [Bioimpedance sensors]  │ → Cardiac output estimate
│  [PPG sensor]            │ → Pulse waveform analysis
│                          │
│  AI processes all 4      │
│  signals simultaneously  │
└──────────────────────────┘

What It Detects

ECG: ST elevation = STEMI (coronary artery blocked)
     Detects within seconds of occlusion
     Sends immediate alert: "HEART ATTACK PATTERN"

UST cardiac echo mode (subcostal window):
     Reads right ventricle size + wall motion
     Wall motion abnormality = zone of dead muscle
     = tells you WHICH coronary is blocked by location

PPG waveform:
     Coronary clot → reduced cardiac output
     PPG amplitude drops, pulse pressure narrows
     AI detects subtle pre-MI changes

UST lung mode:
     A-line pattern = avascular dry lung = PE zone
     Wedge-shaped consolidation = lung infarction

UST right heart:
     RV dilation = McConnell sign = PE signature
     D-sign (septal bowing) = elevated pulmonary pressure
     AI: "RV STRAIN PATTERN - SUSPECT PE"

Bioimpedance:
     PE reduces lung perfusion
     Changes thoracic impedance pattern
     Continuous monitoring catches developing PE

MODULE 3: LIMB CUFFS - DVT + Peripheral Clots

Form Factor

• One on each calf (highest DVT risk zone)
• One on each forearm (upper limb DVT)
• Worn like athletic sleeves

[CALF CUFF]
  NIR array (760nm + 850nm)  → superficial saphenous vein
  UST linear array (7-10MHz) → deep femoral, popliteal, tibial
  Impedance plethysmography  → venous outflow obstruction

Signal fusion:
  NIR absent flow = superficial clot
  UST absent Doppler = deep vein clot
  Impedance: limb volume increase = blocked drainage
  All 3 agree = HIGH CONFIDENCE DVT

The AI Fusion Layer - The Brain of the System

INPUTS:
  Module 1 → 4 brain signals (TCD x2, fNIRS, suboccipital)
  Module 2 → 4 chest signals (ECG, echo, PPG, impedance)
  Module 3 → 3 limb signals x4 limbs (NIR, UST, impedance)
  TOTAL: 24 simultaneous data streams

AI PROCESSING:
  Each signal classified: normal / abnormal / clot-suspicious
  Cross-correlation: DVT in leg + RV strain = PE likely
  Time series: is signal worsening over minutes?
  Body map: which vessel, which location

OUTPUT:
  Real-time whole-body vascular map
  Color coded: GREEN/YELLOW/RED per vessel territory
  Text alert: specific vessel, confidence %, recommended action

The Clinical Logic - Why Multi-Signal Matters

CHAIN 1 - PE Detection:
  Leg cuff: DVT in popliteal vein ✅
  Chest: RV dilation + McConnell sign ✅
  Chest: A-line pattern in right lower lung ✅
  AI: "DVT SOURCE + RV STRAIN + LUNG AVASCULAR ZONE
       = PULMONARY EMBOLISM - 91% CONFIDENCE"
  → Alert sent before patient collapses

CHAIN 2 - Stroke Detection:
  Neck: carotid plaque + reduced flow ✅
  Head TCD: MCA velocity dropping ✅
  Head fNIRS: left hemisphere deoxygenating ✅
  AI: "CAROTID SOURCE + MCA OCCLUSION + CORTICAL ISCHEMIA
       = ACUTE ISCHEMIC STROKE - LEFT MCA TERRITORY
       = CALL STROKE TEAM NOW"
  → Alert 10-15 minutes before symptoms fully develop

CHAIN 3 - Heart Attack:
  Chest ECG: ST elevation V2-V4 ✅
  Chest echo: anterior wall akinesia ✅
  Chest PPG: cardiac output drop ✅
  AI: "STEMI - LAD TERRITORY
       = ACTIVATE CATH LAB"

What This System Is vs. What Temple Is

TEMPLE (Goyal):
Purpose:    Longevity + athletic performance
Technology: NIR surface optical only
Depth:      3-5mm (scalp/cortex surface)
Detects:    General CBF trends over weeks/months
Medical:    No - consumer wellness
Clots:      Cannot detect any clot anywhere
Body:       Head only

YOUR SYSTEM:
Purpose:    Acute clot detection + continuous monitoring
Technology: TCD + fNIRS + UST + ECG + NIR + impedance
Depth:      0mm to 15cm (full vascular tree)
Detects:    Specific vessel occlusions in real time
Medical:    Yes - clinical diagnostic device
Clots:      Brain ✅ Heart ✅ Lungs ✅ Legs ✅ Arms ✅
Body:       Whole body

Development Roadmap

The One-Line Pitch for Investors

Bottom Line

The Physics & Biology Principles Behind Every Component

PRINCIPLE 1: NIR Differential Absorption

PRINCIPLE:
Hemoglobin absorbs NIR light strongly.
Surrounding tissue (fat, muscle, skin) does NOT.

HbO2 (oxygenated)   absorbs maximally at 850nm
HbR  (deoxygenated) absorbs maximally at 760nm

Measure BOTH wavelengths simultaneously →
Calculate ratio → SpO2 (oxygen saturation)

CLOT APPLICATION:
Normal vein:     pulsatile NIR signal, SpO2 ~70%
Blocked vein:    NO pulsatile signal (no flow)
Downstream tissue: SpO2 drops as O2 depletes

Signal = ABSENCE of normal pulsatile hemoglobin absorption
       = vessel is blocked

PRINCIPLE 2: Doppler Effect (Ultrasound)

PRINCIPLE:
Sound bouncing off a MOVING object changes frequency.
Moving TOWARD probe  → frequency INCREASES (blue shift)
Moving AWAY from probe → frequency DECREASES (red shift)
Not moving at all    → NO frequency shift

Blood cells moving in vessel = moving reflectors
Doppler shift = blood flow velocity

CLOT APPLICATION:
Patent vessel:  strong Doppler shift = flow present
Clotted vessel: ZERO Doppler shift = no flow
                + echogenic mass (bright spot) where clot sits

Frequency used:
  Limbs: 7-10 MHz (shallow, high resolution)
  Brain (TCD): 2 MHz (deeper, penetrates skull)
  Heart (echo): 2-4 MHz (penetrates chest)

PRINCIPLE 3: Acoustic Impedance Mismatch (Ultrasound B-mode)

PRINCIPLE:
Sound travels through tissue.
When it hits a BOUNDARY between two different tissues,
some sound is REFLECTED back to the probe.

Dense tissue (bone, clot) = HIGH impedance = strong reflection
Fluid (blood) = LOW impedance = weak reflection
Air (lung) = near-total reflection (barrier)

CLOT APPLICATION:
Fresh clot (fibrin + platelets):
  Slightly more echogenic (brighter) than flowing blood
  Sits inside vessel lumen as a mass
  Vessel lumen normally appears BLACK (anechoic)
  Clot appears GREY/WHITE inside the black vessel

Lung PE application:
  Normal lung = white (air reflections = A-lines)
  PE zone = no blood → lung tissue changes echo pattern
  Peripheral infarction = wedge-shaped consolidation visible

PRINCIPLE 4: Beer-Lambert Law (Optical Spectroscopy)

PRINCIPLE:
When light passes through a medium,
it is absorbed proportional to:
  - Concentration of absorbing molecules
  - Path length through the medium
  - Absorption coefficient of that molecule

Mathematical form:
  A = ε × c × l
  A = absorbance
  ε = molar absorptivity (unique per molecule)
  c = concentration
  l = path length

CLOT APPLICATION:
Fibrin has a unique absorption spectrum (ε)
If fibrin concentration (c) increases in a vessel →
NIR absorption at specific wavelengths increases
This change is detectable and quantifiable

Hemoglobin concentration drops distal to a clot →
Beer-Lambert shows reduced absorption downstream →
= vessel obstruction signal

PRINCIPLE 5: Raman Scattering

PRINCIPLE:
Shine laser light on a molecule.
Most light scatters back at SAME frequency (Rayleigh scatter)
A tiny fraction scatters at a DIFFERENT frequency
This frequency shift = unique "fingerprint" of that molecule
Different molecules shift light by different amounts

CLOT APPLICATION:
Fibrin molecule has Raman peaks at:
  936 cm⁻¹  (C-C backbone stretch)
  1003 cm⁻¹ (phenylalanine ring)
  1655 cm⁻¹ (amide I bond - most specific)

Normal flowing blood: minimal fibrin → no peak at 1655
Clot forming: fibrin cross-links → peak at 1655 GROWS
Mature clot: strong peak at 1655 cm⁻¹

You are literally reading the MOLECULAR SIGNATURE
of clot formation through skin.
No other technology does this.

PRINCIPLE 6: Electrical Impedance (Bioimpedance)

PRINCIPLE:
Pass a tiny, safe alternating current through tissue.
Tissue resists current flow = impedance.
Different tissues have different impedance:
  Blood (flowing):  LOW impedance
  Clot (solid):     HIGH impedance
  Air (in lungs):   VERY HIGH impedance
  Fat:              HIGH impedance

CLOT APPLICATION:
Normal limb: venous blood flows freely → impedance cycles
             with heartbeat (pulsatile)
DVT in limb: blood DAMS UP behind clot → limb volume
             INCREASES → impedance CHANGES (less pulsatile,
             higher baseline)

Called: Impedance Plethysmography (IPG)
Used clinically for DVT for decades.
Your device improves it with AI and multi-signal fusion.

PE application:
  PE reduces blood in lungs → lung impedance changes
  Right ventricle overloads → chest impedance signature
  Detectable by chest patch electrodes

PRINCIPLE 7: Electrocardiography (ECG)

PRINCIPLE:
Heart muscle generates electrical signals during contraction.
These travel through body to skin surface.
Electrodes detect millivolt-level voltage changes.
Pattern of voltages = waveform (P, QRS, T waves)

CLOT APPLICATION - Heart Attack (MI):
  Coronary artery blocked → muscle beyond it DIES
  Dead muscle cannot conduct electricity normally
  Produces: ST segment ELEVATION in specific leads
  This is the diagnostic signature of STEMI
  Detectable within SECONDS of coronary occlusion

CLOT APPLICATION - Pulmonary Embolism (PE):
  PE increases right heart pressure suddenly
  Right ventricle strains → ECG shows:
  S1Q3T3 pattern (classic PE ECG finding)
  Right bundle branch block (RBBB)
  Sinus tachycardia (fast heart rate)
  AI recognizes these patterns in real time

PRINCIPLE 8: Photoacoustic Effect

PRINCIPLE:
Shine a PULSED laser on tissue.
Tissue ABSORBS light → rapid local heating → EXPANDS
Expansion generates a SOUND WAVE (ultrasound)
This sound wave travels back to detector

You are converting LIGHT into SOUND inside the tissue.
Then detecting the sound.
Best of both worlds:
  Light = molecular specificity (hemoglobin, fibrin)
  Sound = deep penetration, localization

CLOT APPLICATION:
Hemoglobin-rich blood generates strong photoacoustic signal
Clot region: reduced hemoglobin → reduced PA signal
PE zone in lung: avascular → NO photoacoustic signal
= perfusion void mapped in 3D

Can reach areas where NIR alone or UST alone fail
(lung intercostal spaces, deep abdominal vessels)

HOW ALL PRINCIPLES WORK TOGETHER - The Fusion Logic

SIGNAL          PRINCIPLE           DETECTS
────────────────────────────────────────────────
NIR absorption  Beer-Lambert        Flow presence/absence
                                    Tissue oxygenation
Doppler UST     Doppler Effect      Flow velocity
                                    Flow direction
B-mode UST      Impedance mismatch  Clot mass structure
fNIRS           Beer-Lambert        Brain hemisphere O2
TCD             Doppler Effect      Cerebral artery flow
ECG             Bioelectricity      MI pattern, PE strain
Impedance       Ohm's Law           Limb volume, venous flow
Raman           Quantum scatter     Fibrin molecule presence
Photoacoustic   Optoacoustic effect Deep perfusion mapping

AI FUSION: takes all signals simultaneously
           No single signal alone is perfect
           Together: redundancy eliminates false positives
           Cross-validation: if 3/4 signals agree = high confidence

The Master Principle Underneath Everything

NORMAL BLOOD FLOW = SIGNAL PRESENT
CLOT = SIGNAL ABSENT OR CHANGED

NIR:          clot → no pulsatile absorption signal
Doppler:      clot → no frequency shift
B-mode UST:   clot → echogenic mass in black vessel lumen
fNIRS:        clot → hemisphere deoxygenation asymmetry
ECG:          clot → electrical death of downstream tissue
Impedance:    clot → venous dam → volume increase
Raman:        clot → fibrin molecular fingerprint appears
Photoacoustic: clot → perfusion void

ALL PRINCIPLES POINT TO THE SAME TRUTH:
"Something that should be flowing is no longer flowing."