Sources and References

Primary texts — Wang, L.V. and Wu, H.-I., Biomedical Optics: Principles and Imaging, Wiley, 2007; Vo-Dinh, T. (ed.), Biomedical Photonics Handbook, 2nd ed., CRC Press, 2014.

Supplementary texts — Tuchin, V.V., Tissue Optics: Light Scattering Methods and Instruments for Medical Diagnosis, 3rd ed., SPIE Press, 2015; Prasad, P.N., Introduction to Biophotonics, Wiley, 2003; Born, M. and Wolf, E., Principles of Optics, 7th ed., Cambridge, 1999.

Online resources — MIT OCW HST.563J “Biomedical Optics”; Stanford “Biomedical Optics” course notes by Bowden & Pauly; Oregon Medical Laser Center online tissue optics tutorials (OMLC); OSA/Optica open-access tutorial papers; ANSI Z136.1 laser safety standard (publicly summarized); IEC 60825-1 laser classification.

---

Chapter 1: Light-Tissue Interactions

Biomedical optics operates at the intersection of electromagnetic theory and the heterogeneous, turbid environment of living tissue. Unlike clear optical media, tissue scatters strongly, absorbs selectively, and responds dynamically to irradiation. Every imaging or spectroscopic modality rests on understanding how a photon’s fate—absorption, elastic scattering, inelastic scattering, fluorescence, or transmission—depends on tissue composition and wavelength.

1.1 Optical Properties of Tissue

Four bulk parameters describe light propagation: the absorption coefficient \( \mu_a \) [cm\(^{-1}\)], scattering coefficient \( \mu_s \), anisotropy factor \( g = \langle \cos\theta \rangle \), and refractive index \( n \). The reduced scattering coefficient

\[ \mu_s' = (1-g)\mu_s \]

captures the effective randomization of direction after many scatters.

Typical soft tissue at 700 nm: \( \mu_a \sim 0.05 \) cm\(^{-1}\), \( \mu_s' \sim 10 \) cm\(^{-1}\), \( g \sim 0.9 \), \( n \sim 1.4 \). The “therapeutic window” (600–1000 nm) is where hemoglobin and water absorption are minimal, allowing centimeter-scale penetration.

1.2 Absorption Spectroscopy of Endogenous Chromophores

Key tissue absorbers:

• Oxyhemoglobin (HbO\(_2\)) and deoxyhemoglobin (Hb) — dominate 500–700 nm; their difference spectrum enables pulse oximetry and fNIRS.
• Water — absorbs strongly beyond 950 nm and in mid-IR.
• Melanin — broadband visible absorber.
• Lipids, collagen, NADH, FAD — provide metabolic and structural contrast.

The Beer-Lambert law for a mixture:

\[ A(\lambda) = \sum_i \varepsilon_i(\lambda) c_i L, \]

with path length \( L \). In scattering media the effective path length exceeds the geometric distance, a correction captured by the differential pathlength factor (DPF) in near-infrared spectroscopy.

1.3 Elastic Scattering

Mie and Rayleigh theories predict scattering from discrete particles. In tissue, organelles (mitochondria, nuclei) contribute most scattering; their size distribution (0.1–5 µm) gives the characteristic \( \mu_s \sim \lambda^{-0.5} \) to \( \lambda^{-1.5} \) dependence. Scatterer morphology encodes diagnostic information exploited in angle-resolved low-coherence interferometry and light-scattering spectroscopy of pre-cancerous changes in epithelia.

1.4 Fluorescence and Raman

A fluorophore absorbs a photon and re-emits at longer wavelength after vibrational relaxation. The quantum yield \( \Phi \), lifetime \( \tau \), and spectrum identify the molecular species. Autofluorescence (NADH 460 nm, FAD 525 nm, collagen) enables label-free tissue diagnostics.

Raman scattering (one in \( 10^6 \)–\( 10^8 \) photons) produces vibrational fingerprints; surface-enhanced Raman (SERS) boosts cross-sections by \( 10^6 \)–\( 10^{10} \).

Chapter 2: Transport Theory

2.1 The Radiative Transfer Equation

Photon flux in tissue obeys the RTE:

\[ \hat{s}\cdot\nabla L(\mathbf{r}, \hat{s}) = -(\mu_a + \mu_s) L + \mu_s \!\!\int_{4\pi}\!\! p(\hat{s}\cdot\hat{s}') L(\mathbf{r}, \hat{s}') d\Omega' + S(\mathbf{r}, \hat{s}), \]

with phase function \( p \) often modeled by the Henyey–Greenstein form

\[ p(\cos\theta) = \frac{1}{4\pi}\frac{1 - g^2}{(1 + g^2 - 2g\cos\theta)^{3/2}}. \]

2.2 Diffusion Approximation

When \( \mu_s' \gg \mu_a \), expanding \( L \) to first order in direction yields the photon diffusion equation:

\[ \frac{1}{c}\frac{\partial \phi}{\partial t} - D \nabla^2 \phi + \mu_a \phi = S, \]

with diffusion coefficient \( D = 1/[3(\mu_a + \mu_s')] \). Green’s functions for infinite, semi-infinite, and slab geometries underlie most diffuse optical imaging reconstruction.

2.3 Monte Carlo Simulation

For thin layers, heterogeneous media, or near-source regions, Monte Carlo photon propagation (Wang & Jacques’ MCML) is the gold standard: photons are launched, step lengths drawn from \( -\ln\xi/(\mu_a + \mu_s) \), and scattering directions sampled from the phase function. GPU implementations run \( 10^9 \) photons per minute.

Chapter 3: Spectroscopy Techniques

3.1 Pulse Oximetry

Ratio-of-ratios processing of red (660 nm) and infrared (940 nm) transmission pulsatile signals yields arterial oxygen saturation:

\[ SpO_2 = A - B \cdot R, \quad R = \frac{AC_{red}/DC_{red}}{AC_{IR}/DC_{IR}}. \]

Coefficients \( A \), \( B \) are obtained by empirical calibration against arterial blood gas, a clinically mandated approach per ISO 80601-2-61.

3.2 Near-Infrared Spectroscopy and fNIRS

Continuous-wave, frequency-domain, and time-domain NIRS each yield different biomarkers. Functional NIRS maps cortical hemodynamics analogous to BOLD-fMRI but with portability and millisecond temporal resolution. Depth sensitivity scales approximately as half the source–detector separation.

3.3 Diffuse Optical Tomography (DOT)

Source–detector pairs on the surface reconstruct volumetric \( \mu_a(\mathbf{r}) \) via a forward model (diffusion equation) and inverse problem regularized by Tikhonov or Bayesian priors. Breast DOT and functional brain imaging are prominent applications.

3.4 Fluorescence Lifetime Imaging (FLIM)

Lifetime is intrinsic to the molecule and insensitive to concentration, photobleaching, or excitation intensity. FLIM distinguishes free from protein-bound NADH, enabling metabolic imaging and tumor margin assessment.

Chapter 4: Optical Imaging Modalities

4.1 Optical Coherence Tomography

OCT is low-coherence interferometry in reflection. Axial resolution is set by the source coherence length

\[ l_c = \frac{2\ln 2}{\pi}\frac{\lambda_0^2}{\Delta \lambda}, \]

giving 1–10 µm. Spectral-domain OCT (SD-OCT) samples the interference spectrum; Fourier inversion gives the depth profile. OCT revolutionized retinal imaging; intravascular OCT visualizes coronary plaques.

4.2 Confocal and Multiphoton Microscopy

A confocal pinhole rejects out-of-focus light, giving optical sections of \( \sim 1 \) µm. Two-photon excitation (\( I^2 \) dependence) confines emission to the focal volume and uses near-infrared light, enabling deeper imaging (\( \sim 1 \) mm) in scattering tissue—the workhorse of in vivo neuroimaging.

4.3 Photoacoustic Imaging

Pulsed light absorbed by chromophores produces transient thermoelastic expansion and ultrasonic emission. Pressure rise \( p_0 = \Gamma \mu_a F \) with Grüneisen parameter \( \Gamma \). Photoacoustic microscopy and tomography combine optical contrast with ultrasonic resolution, overcoming the optical diffusion limit at depth.

Chapter 5: Therapeutic and Diagnostic System Design

5.1 Laser-Tissue Interaction Regimes

Exposure duration and irradiance determine the dominant mechanism:

• \( > 1 \) s at moderate irradiance: photochemical (PDT, UV).
• 1 µs–1 s: photothermal (coagulation, ablation).
• ns–µs: photomechanical (plasma, shock waves, photoacoustic).
• fs–ps: photodisruption, ultraprecise ablation.

The Arrhenius damage integral \( \Omega = \int A e^{-E_a/RT} dt \) quantifies thermal injury; \( \Omega = 1 \) defines the damage threshold.

5.2 Photodynamic Therapy

A photosensitizer (porphyrin derivatives) accumulates preferentially in tumor tissue. Illumination at its Q-band generates singlet oxygen, destroying targeted cells. Treatment planning matches drug pharmacokinetics, light dose, and tissue oxygenation.

5.3 System-Level Design Considerations

Designing a clinical device requires: source selection (LED, diode laser, supercontinuum, SLD) matched to absorption/scattering targets; safe delivery (fiber, scanning head, endoscope); calibrated detection (photodiode, PMT, CCD, sCMOS); signal conditioning and noise budget; and regulatory compliance (FDA 510(k), CE, IEC 60601-1). Eye-safe irradiance limits follow ANSI Z136.1 maximum permissible exposure (MPE) values.

Chapter 6: Emerging Techniques

6.1 Wavefront Shaping and Focusing Through Tissue

Digital optical phase conjugation and iterative wavefront shaping recover diffraction-limited focusing beyond one transport mean free path by inverting the tissue’s scattering matrix. Guide stars (fluorescence, ultrasound, second harmonic) provide feedback.

6.2 Structured Illumination and Super-Resolution

SIM, STED, and single-molecule localization (PALM/STORM) push lateral resolution to \( \sim 20 \) nm. Expansion microscopy physically magnifies specimens by \( 4\times \), achieving super-resolution with conventional microscopes.

6.3 Optogenetics and Photostimulation

Channelrhodopsin-2 (ChR2), halorhodopsin, and archaerhodopsin enable millisecond-scale optical control of neural activity. Implanted fiber photometry and two-photon optogenetics combine recording and stimulation for closed-loop neuroscience.