Automated tissue type imaging with OCT Ton van Leeuwen DJ Faber, DM - PowerPoint PPT Presentation

Automated tissue type imaging with OCT Ton van Leeuwen DJ Faber, DM de Bruin, NM Weiss, DV Nguyen, SM Jansen, M Almasian, AL Post, L Wilk, MCG Aalders, Berrend Muller, A Swaan (AMC), X Attendu (PolyMtl), J Kalkman (TUD) Academic Medical Center, University of Amsterdam, Biomedical Engineering & Physics: T.G.vanLeeuwen@amc.uva.nl www.amc.nl/BMEP 0/46

Gustav Strijkers Ed van Bavel Henk Marquering Maurice Aalders

• Physics of – Interaction of light with tissue – Development of instrumentation introduction • Application of light dependent scattering – Therapeutic – Diagnostic multiple scattering • Monitoring / function • Imaging / morphology conclusion Optical imaging and sensing of functioning biological systems (cells/organs) Relation between optical properties ( µ a , µ s , g) and status of the tissue or tissue type

OCT is the “Harlem Oil” of Biomedical Optics introduction dependent scattering multiple scattering conclusion See http://www.haarlemmerolie.nl

One drop of OCT solves all the problems… introduction dependent scattering multiple scattering conclusion

What kind of tissue are we looking at? introduction dependent scattering multiple scattering conclusion Velocity (mm/sec)

One drop of OCT solves all the problems… introduction dependent scattering multiple scattering conclusion

Single scattering approximation • OCT signal dependent on: – Depth of focus introduction – Position of the focus dependent – Amount of back scattering scattering – Attenuation coefficient multiple ì ü æ ö scattering 2 zn ( ) ( ) = h g Ä µ - µ ç ÷ i z Re med h ( z ) P P exp 2 z í ý det ref sample b , NA s c è ø î þ conclusion DJ Faber, et al.Optics Express 2004, 12: 4353-4365

• OCT signal dependent on: – Depth of focus introduction – Position of the focus dependent – Amount of back scattering (amplitude of the OCT signal) scattering – Attenuation coefficient multiple ì ü æ ö scattering 2 zn ( ) ( ) = h g Ä µ - µ ç ÷ i z Re med h ( z ) P P exp 2 z í ý det ref sample b , NA s c è ø î þ conclusion 800 nm OCT of rat aorta

• OCT signal dependent on: – Depth of focus introduction – Position of the focus dependent – Amount of back scattering scattering – Attenuation coefficient (‘slope’ of the OCT signal) multiple scattering ì ü æ ö 2 zn ( ) ( ) = h g Ä µ - µ i z Re ç med ÷ h ( z ) P P exp 2 z í ý det ref sample b , NA s è c ø î þ conclusion Kodach et al, Biomedical Optics express 2010, Optics Express 2011

Cardiovascular tissues application of µ t extraction Velocity (mm/sec) Van der Meer et al, IEEE TMI, 2005, 24, 1369-1376 Lasers Med Science 2005, 20, 45-51

• OCT signal dependent on: – Depth of focus introduction – Position of the focus dependent – Amount of back scattering scattering – Attenuation coefficient multiple ì ü æ ö scattering 2 zn ( ) ( ) = h g Ä µ - µ ç ÷ i z Re med h ( z ) P P exp 2 z í ý det ref sample b , NA s c è ø î þ conclusion • Is it so simple to measure ! " (! $%& ≈ ! " )? • How to verify that (e.g. for large ! " )? – ! " ~*+,*-,./0.1+, – ! 2 ~*+,*-,./0.1+, – ! " ~! 2 3 4 – 3 5 = *+,7.0,.

All μ OCT for Intralipid • @ 600, 800, 1300 & 1600 nm (single scattering model) 30 600 nm introduction 800 nm g » 0.8 μ OCT (mm -1 ) 1300 nm 25 1600 nm dependent -1 ) scattering scattering coefficient (mm 20 g » 0.7 multiple scattering 15 conclusion Velocity 10 (mm/sec) g » 0.4 5 g » 0.3 0 0 5 10 15 20 Intralipid concentration (%) Why is the attenuation coefficient saturating for higher concentrations IL?

μ s for high concentrations Dependent scattering for all beads Corrected for dependent scattering with structure factor by Percus-Yevick model introduction 35 ! " ≁ $%&$'&()*(+%& dependent scattering Æ 1215 nm 30 multiple 25 scattering Æ 906 nm 20 conclusion Velocity -1 ) µ s (mm (mm/sec) 15 Æ 759 nm 10 5 Æ 0.376 nm 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 volume fraction Nguyen et al Optics Express 2013

OCT signal analysis: amplitude ! " ~ $%&$'&()*(+%& introduction Amplitude dependent scattering multiple scattering conclusion Amplitude M Almasian, et al.J BiomOptics, 2015, 20: 121314

OCT signal analysis: amplitude back scattering coefficient/scattering coefficient ' ( ≁ ' * introduction dependent scattering multiple scattering conclusion $ % 2 # ∅ ≈ Mitra Almasian et al, Scientific Reports 2017

Effect of multiple scattering… introduction dependent scattering multiple scattering • Picking up multiple scattered light will reduce µ OCT using conclusion Velocity the single scattering model (mm/sec) • higher scattering values, deeper into the tissue, dependent on g and NA optics • Extended Huygens Fresnel (EHF) model 1 [ ] ( ) ( ) é ù - µ - - µ 2 1 2 exp z 1 exp z W 2 ( ) ( ) [ ( ) ] µ - µ + + - - µ 2 i z exp 2 z s s 1 exp z h ê ú d s s + 2 2 2 W 1 W W W ë û h s h s D. Levitz et al., Optics.Express 2004, 12, 249-259

OCT: dependent (PY) and multiple (EHF)scattering Æ = 0.47 µm Æ = 0.70 µm g Mie+PY = 0.39-0.06 g Mie+PY = 0.71-0.56 introduction dependent scattering multiple scattering conclusion Velocity Æ = 0.91 µm Æ = 1.6 µm (mm/sec) g Mie+PY = 0.79-0.67 g Mie+PY = 0.91-0.87

Spheres, Intralipid, g>0.8, so what? • Blood, µ OCT measured by 800 nm OCT introduction Effect of chosen (fixed) g in EHF model dependent scattering multiple scattering g = 0.995 conclusion Velocity (mm/sec) g = 0.8 Faber et al, OL 2009

Forward scattering of blood • Whole blood flowing through B-scan plane static sample: flow on! phantom Flow channel Solid TiO2 phantom Whole blood flow 400 μm channel, 20 mm/sec Solid TiO2 phantom SM Jansen et al, Sensors 2018

Effect of NA (PSF) on µ OCT • Fitting both µ OCT and g in EHF: no convergence • Setting g correctly results in: introduction ! " 13%& 20%& 51%& dependent scattering % * (&& ,- ) % /01 (&& ,- ) % /01 (&& ,- ) % /01 (&& ,- ) g multiple 5.4 0.91 6 7 10 scattering 14 0.89 22 23 24 conclusion Velocity 21 0.87 40 33 55 (mm/sec) • Better model needed? M Almasian, in progress

But NA (PSF) is more or less constant.. • Fitted PSF (and Roll-of) for 4 C7 Dragonfly tm Intravascular Imaging Probe (St. Jude Medical, introduction St. Paul, Minnesota, USA) dependent scattering multiple scattering conclusion Velocity (mm/sec)

Discussion • It is not as simple as Harlem Oil – " # ≁ %&'%(')*+),&' introduction – " - ~ %&'%(')*+),&' – " # ≁ " - dependent scattering • correctly fitted PSF (and Roll-of) multiple – NA can differ between catheters scattering • If / ≾ 0.8: " 567 ≈ " # conclusion Velocity (mm/sec) – For larger g, more multiple scattered light contributes to the OCT signal • automated tissue type imaging with OCT: – be carefull – other contrast mechanisms – The eye…

What kind of tissue are we looking at? introduction dependent scattering multiple scattering conclusion Velocity (mm/sec)

What kind of tissue are we looking at? introduction dependent scattering multiple scattering conclusion Velocity (mm/sec) OCT during restauration of Rembrandt painting Marten at Rijksmuseum, see https://www.rijksmuseum.nl/nl/marten-en-oopjen

Discussion • It is not as simple as Harlem Oil – " # ≁ %&'%(')*+),&' introduction – " - ~ %&'%(')*+),&' – " # ≁ " - dependent scattering • correctly fitted PSF (and Roll-of) multiple – NA can differ between catheters scattering • If / ≾ 0.8: " 567 ≈ " # conclusion Velocity (mm/sec) – For larger g, more multiple scattered light contributes to the OCT signal • automated tissue type imaging with OCT: – be carefull – other contrast mechanisms – The eye…