Diffuse Optical Spectroscopy and Imaging · Vioptix Odissey ~$80B/yr St. Jude C7-XR Diffuse Optics ... • Structured Light: Space and Time • Optical Pathlength Control Multiple

Diffuse Optical Spectroscopy and Imaging

Laser Microbeam and Medical Program (LAMMP) Beckman Laser Institute and Medical Clinic

Departments of Biomedical Engineering and Surgery

University of California, Irvine http://www.bli.uci.edu

Bruce J. Tromberg

BLI

Computational Biophotonics Workshop: July 2017

Disclosure

Co-founder:

Industry Collaboration:

Medical Imaging

MR, X-ray, Nuclear, US (Optical)

Mainstream Modalities

Where Does Biophotonics Fit In?

Masimo Pronto

Medtronic-Covidien Invos

Diagnostics

Imaging/spectroscopy

NOVADAQ Pinpoint

Heidelberg Engineering Spectralis

Perimed Periscan Pentax Endoscope

Syneron-Candela, Lumenis, Alma, Cutera, Cynosure, Fotona, Lutronic, etc

AMO Intralase FS

Therapeutics

Medical Lasers

Vioptix Odissey

~$80B/yr

St. Jude C7-XR


Masimo Pronto


Diagnostics


NOVADAQ Pinpoint




AMO Intralase FS

Therapeutics

Medical Lasers

Vioptix Odissey

~$80B/yr

St. Jude C7-XR


Masimo Pronto


Diagnostics


NOVADAQ Pinpoint




AMO Intralase FS

Therapeutics

Medical Lasers

Vioptix Odissey

~$80B/yr

St. Jude C7-XR

Diffuse Optics • Multi-Spectral (NIR) • Structured Light: Space and Time • Optical Pathlength Control

Multiple Light Scattering

850 nm NIR LEDs

O'Sullivan TD, et al. J Biomed Opt. 17(7):071311 (2012).

Tissue Optics Quantitative Challenge

Measuring Optical Pathlength

Molecular Absorption Loss

Absorption + Scattering Loss

A = εbC; b = 1cm

A = εbC; b = ??

Fixed pathlength

Unknown pathlength

Why Measure Optical Pathlength?

• Determine Perfusion and Metabolism at depth (Oxy/Deoxy Hb) • Determine concentration of other NIR absorbers/fluors (e.g. Water,

Lipid, exogenous dyes/particles ) • Correct fluorescence for tissue optical property distortions

1) Separate Absorption from Scattering 2) Localize Information in 3D (Tomography)

Physics

Physiology

Controlling Pathlength Three Ways to Control Pathlength

blue

green

red

Scatter Dominated

labs ~ lscat labs >> lscat

1) Wavelength:

Controlling Pathlength

1) Wavelength: blue

green

red

100% StO2

50% StO2

Sensitive to small absorption changes

Three Ways to Control Pathlength

400 labs ~ lscat labs >> lscat


2) Space: labs ~ 10 cm; lscat ~ 20-40 µm;

D-1 D-2

2 10 20 mm

Reflectance vs D

µa = 0.01 mm-1

µs’ = 0.1 mm-1

Scatter dominated

Scatter + Absorption



f = 0 mm-1 f = 0 mm-1


f = 0 mm-1 f = 0.2 mm-1


3) Time: τabs ~ 0.5 ns, τscat ~ 0.20 ps Intensity vs. Time (2 cm s-d)

Laser pulse

0.2 1.0 0.6 ns

T-1

T-2

Scatter dominated

Absorption dominated


2 cm s-d

Measuring Optical Path Length

Measuring Optical Path Length FT

(t)

FT-1

(ω)

FT(t)

FT-1

(ω)

Measuring Optical Path Length FT

(t)

FT-1

(ω)

FT(t)

FT-1

(ω)

~10 cm depth, ~ cm resolution ~1 cm in depth, ~mm resolution

Tromberg, et al., Appl Opt., (1993) Cuccia et. al., Opt Lett, (2005)

source detector (reflection)

scattering tissue

ρ

inte

nsity

time

source light

detected light

),( ωrΘ

),( ωrA

Temporal Frequency Domain Photon Migration (FDPM)


),(),(),()(),( trStrtrrDt

trcn

a =+∇⋅∇−∂

∂φμφ

φ

Light source

Loss: Light Absorption (f/λ)

Build up: Light Scattering

)]'(3/[1 saD μμ += Photon Diffusion Coefficient

In Scatter-Dominated Region: Diffusion Equation

Light Tissue Distribution

Fluence rate: Space, time

1/labs 1/ltr T. O’Sullivan et al., JBO, 2012

Diffusion equation (time dependent)

=−−=+−=Φ ])[exp()exp(4

)exp(4

),( rktirkDr

StikrDr

Srt imgrealACAC

AC ωπ

ωπ

frequency domain → scalar photon density wave

)exp()()(),( rate fluence modulated tirrrt ACDC ωΦ+Φ=Φ⇒

damped wavecDi

Dk a ωμ

+=2

Infinite medium:

Boundary Conditions: Haskell, Tromberg et al, JOSA-A (1994)

0200400600800

10001200

0 0.02 0.04 0.06 0.08 0.1

k (1/mm)

Mod

ulat

ion

Freq

uenc

y (M

hz)

Photon density wave

kreal

μa = 0.006 mm-1

μs’ = 1 mm-1

n=1.4

photon density wavelength = 2π/kimg ≅ 10 cm @ 200 MHz

(if no scattering, in air ≅ 9 m @ 200 MHz )

phase velocity Vp = ω/kimg ={ ω << cμa ⇒ Vp = 2(D/cμa )1/2 ≈16 mm/ns ω >> cμa ⇒ Vp = (2Dωc)1/2

1/τl

δDC = 1/μeff = (D/μa)1/2 = 7.4 mm

kimag

0.05 0.1 0.15 0.2 0.25 cμa= 200 MHz

δAC = 1/kreal

(independent of ω)

(dependent on ω: dispersion)

200 400 600 800 10000

1

2

3

4

5

x 10-4

Am

plitu

de (

a.u.

)

Frequency (MHz)

DataModel Fit

200 400 600 800 10000

50

100

150

200

250

300

Pha

se (

deg.

)Frequency (MHz)

DataModel Fit

700 800 900 10000

0.005

0.01

0.015

0.02

Abs

orpt

ion

Coe

ffic

ient

(m

m-1

)Wavelength(nm)

FDPM

700 800 900 10000.6

0.7

0.8

0.9

1

Red

uced

Sca

tter

ing

Coe

ff.

(mm

-1)

Wavelength(nm)

FDPM

700 800 900 10000

0.005

0.01

0.015

0.02

Abs

orpt

ion

Coe

ffic

ient

(m

m-1

)Wavelength(nm)

FDPM

700 800 900 10000.6

0.7

0.8

0.9

1

Red

uced

Sca

tter

ing

Coe

ff.

(mm

-1)

Wavelength(nm)

FDPMPower Law Fit

700 800 900 10000.6

0.7

0.8

0.9

1

Red

uced

Sca

tter

ing

Coe

ff.

(mm

-1)

Wavelength(nm)

FDPMPower Law Fit

700 800 900 10000

0.005

0.01

0.015

0.02

Abs

orpt

ion

Coe

ffic

ient

(m

m-1

)Wavelength(nm)

FDPM

700 800 900 10000

0.2

0.4

0.6

0.8

1

1.2R

efle

ctan

ce (

a.u.

)

Wavelength(nm)

700 800 900 10000

0.005

0.01

0.015

0.02

Abs

orpt

ion

Coe

ffic

ient

(m

m-1

)Wavelength(nm)

FDPMSSFDPM

700 800 900 10000.6

0.7

0.8

0.9

1

Red

uced

Sca

tter

ing

Coe

ff.

(mm

-1)

Wavelength(nm)

FDPMPower Law Fit

700 800 900 10000

0.005

0.01

0.015

0.02

Abs

orpt

ion

Coe

ffic

ient

(m

m-1

)Wavelength(nm)

FDPMSSFDPMChromophore Fit

Oxyhemoglobin = 12.7 µM Deoxyhemoglobin = 4.1 µM

Water = 21.5% Lipid = 79.6%

LcII

ε=⎟⎟⎠

⎞⎜⎜⎝

⎛0log

650 700 750 800 850 900 950 10000.0

0.2

0.4

0.6

0.8

1.0

Abs

orpt

ion

(mm

-1m

M-1)

Wavelength(nm)

Tissue NIR absorbers

HHb

O2Hb

BULK LIPID

H2O

Oxygen Metabolism

Arteriole 90 mmHg

Tissue 17 < 47 mmHg

Cell Mitochondria 2<30 mmHg

Venule 35

mmHg 10-30 mmHg

Oxygen demand

Oxygen Supply

Cell-vascular coupling

Predicting Clinical Outcome

35 M. Herringlake, et al., Anesthesiology, 114, 58 (2011)

Total Population

High Risk Group

Optical endpoint: Pre-Surgical StO2 = HbO2/Hbtot Clinical endpoint: 30 day and 1 yr survival

n = 1200, Cardiac Bypass Surgery

Covidien INVOS

Predicting Clinical Outcome

36 M. Herringlake, et al., Anesthesiology, 114, 58 (2011)

≤50% StO2 = ~25% greater chance of death at 1 year ≤50% StO2 = 45% SURVIVAL at 1 year (high risk group)

Total Population

High Risk Group

Covidien INVOS Optical endpoint: Pre-Surgical StO2 = HbO2/Hbtot Clinical endpoint: 30 day and 1 yr survival

Cerebral hemodynamics during anesthesia Phenylephrine Ephedrine

Lingzhong Meng

L. Meng, et al. Brit J. Anes, 107, 209 (2011)

ISS Oxiplex (Frequency Domain)


Lingzhong Meng



<65% ~3-4 minutes/bolus


Lingzhong Meng




Cardiac Output Drop


Lingzhong Meng




Cardiac Output Drop

Phenylephrine: Rapid Rise in MAP: drop in StO2

Cognitive impairment, survival risk

Tissue Oxygenation during Exercise

G. Ganesan et al. Ped Ex Sci, 2015


Drop in ET CO2

Rise in ET CO2 Blood Acidosis




Ganesan G, et al. Med Sci Sports Ex. (2014)

Ganesan G, et al. Med Sci Sports Ex. (2014)

Pre-frontal Cortex: Paced Breathing

OxyHb and DeoxyHb phase delay

OxyHb and Khine strain sensor DeoxyHb and Khine strain sensor

2 2.5 3 2 2.5 3

minutes minutes

Peak inhale Peak inhale

Obesity and Metabolic Syndrome

Subcutaneous Adipose Tissue (AT) Promotes Insulin, Leptin Resistance and Metabolic Syndrome:

-Shrink Adipocytes and AT burden

•  Does DOSI detect structural and functional changes in AT? (n=10)

•  Can DOSI provide early feedback on diet and other

interventions in metabolic disease? (12 weeks)

Dr. Shaista Malik, Director, Samueli Ctr., UCI Dept. of Cardiology

Non-invasive Abdominal Fat: Weight Loss Patient

12 week Calorie Restricted Diet

0.005

0.01

0.015

0.02

0.025

650 700 750 800 850 900 950 1000

µ a (m

m-1

)

Wavelength (nm)

Absorption Coefficient – M3

Baseline Post - Weight Loss (6 weeks) 0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

650 700 750 800 850 900 950 1000

µ s' (

mm

-1)

Wavelength (nm)

Reduced Scattering Coefficient –M3 (7.5 cm right of navel)

Baseline Post - Weight Loss (6 weeks)

baseline

baseline

6 weeks

6 weeks

Absorption Spectrum Lipid

Water

DeoxyHb

Scattering Spectrum

Ultrasound 2-3 cm adipose layer

G. Ganesan et al., Int J. Obesity, (2016)

Non-invasive Abdominal Fat: Weight Loss Patient 12 week Calorie Restricted Diet

t

Baseline 3 Months “A” scattering parameter


Water

Scattering

Non-invasive Abdominal Fat: Weight Loss Patient 12 week Calorie Restricted Diet


Water

Problem: Breast Cancer Detection, Diagnosis, Therapy Monitoring

Is the Tumor Malignant or Benign? Is it Responding to Chemotherapy?

Dr. Rita Mehta

Dr. Alice Police

Dr. Freddie Combs

Diffuse Optical Spectroscopic Imaging (DOSI)

Bevilacqua, F., A. J. Berger, et al. Appl Opt 39(34): 6498-507 (2000).

Scattering Absorption : :

White light

Laser diodes

Spectro- graph

APD

SS: 650-1100 nm

Frequency-Domain: ~400 MHz

Broadband temporal Frequency Domain Photon Migration (FDPM) + CW-NIRS

Tromberg, BJ, Pogue, BW, et al., Med. Phys., 35(6), 2443-2451, (2008).

500 MHz

ACRIN trial: UCI, Penn, Dartmouth, MGH, UCSF, MD Anderson, Boston U.

Oxy-Hemoglobin

Lipid

Water Deoxy-Hemoglobin

Tromberg et al. Can Res (2016)

Responder

Non Responder

Tromberg et al. Can Res (2016)

DOSI Baseline 6691-08 TOI = (HbR x H2O)/lipid

US Baseline: Depth: ~0.5-1.5 cm Size: ~2cm x 1cm

DCE-MRI Baseline

Case Study: 52 yr, peri-menoapausal, BIRADS 3, 2cm, Metaplastic carcinoma, Triple negative

DOSI 1 week 6691-08 %Δ TOI ~ -30%

US 1 Week: Depth: ~0.5-1.5 cm Size: ~2cm x 1cm

%Δ ~ 0

Case Study: 52 yr, peri-menoapausal, BIRADS 3, 2cm, Metaplastic carcinoma, Triple negative

Breast Density, Risk

Does Tamoxifen Work for You?

(e.g. CYP2D6 polymorphisms)

Should You Take it For 5 Years?

12 months Tamoxifen +10% Lipid -47% Density

%Lipid

With Gopi Meenakshisundaram and Aditi Majumder, UCI

Baseline

Conclusions

Diffuse Optics: Bedside Monitoring

•  Imaging drug/radiation therapy •  Surgical guidance •  Predicting Risk/Outcome •  Multi-modality Imaging

Continuous, Frequent Monitoring, Wearable Sensors Personalized Physiology and Medicine w/genomics

Challenge: Link optical endpoints to clinical outcome

DOSI Imaging: Anais Leproux, Rob Warren, Brian Hill, Amanda Durkin, Jesse Lam, Hossain Yazdi, Drew Reilly, Goutham Ganesan, Alex Matlock Albert Cerussi (Apple), Keunsik No (LG), Darren Roblyer (BU), Tom O’Sullivan (ND) Drs. Dan Cooper, Shlomit Aizak, Pietro Gallasetti, Shaista Malik, Alice Police, Freddy Combs, Rita Mehta

NIBIB P41 Laser Microbeam and Medical Program; NCI Chao Comprehensive Cancer Center

NIH R01CA142989, R21EB014440, R21NS078634; NCI American College of Radiology Imaging Network

AFOSR Military Photomedicine Program; Arnold and Mabel Beckman Foundation

Acknowledgements

Optical Property Mapping 2 cm

http://

www.bli.uci.edu/

http://www.virtualphotonics.org/vts/

Diffuse Optical Spectroscopy and Imaging · Vioptix Odissey ~$80B/yr St. Jude C7-XR Diffuse Optics ... • Structured Light: Space and Time • Optical Pathlength Control Multiple

Documents