Investigation of detection limits for diffuse optical tomography systems: I. Theory and experiment.

We present a statistical test using simulated photon migration data and a noise model derived from the hardware of a particular diffuse optical tomography system to predict its detection limits. Our method allows us to assess the spatial distribution of the detection sensitivity of arbitrary geometries and noise without requiring phantom measurements and reconstructions. We determine the minimal detectable lesion size at selected lesion positions and compare the predicted results with phantom measurements carried out in a cup geometry.

Evaluation of higher-order time-domain perturbation theory of photon diffusion on breast-equivalent phantoms and optical mammograms.

Time-domain perturbation theory of photon diffusion up to third order was evaluated for its accuracy in deducing optical properties of breast tumors using simulated and physical phantoms and by analyzing 141 projection mammograms of 87 patients with histology-validated tumors that had been recorded by scanning time-domain optical mammography. The slightly compressed breast was modeled as (partially) homogeneous diffusely scattering infinite slab containing a scattering and absorbing spherical heterogeneity representing the tumor. Photon flux densities were calculated from densities of transmitted photons, assuming extended boundary conditions. Explicit formulas are provided for second-order changes in transmitted photon density due to the presence of absorbers or scatterers. The results on phantoms obtained by perturbation theory carried up to third order were compared with measured temporal point spread functions, with numerical finite-element method (FEM) simulations of transmitted photon flux density, with results obtained from the diffraction of diffuse photon density waves, and from Padé approximants. The breakdown of first-, second-, and third-order perturbation theory is discussed for absorbers and a general expression was derived for the convergence of the Born series in this case. Taking tumor optical properties derived by the diffraction model as reference we conclude that estimates of tumor absorption coefficients by perturbation theory agree with reference values within +/-25% in only 65% (first order), 66% (second order), and 77% (third order) of all mammograms analyzed. In the remaining cases tumor absorption is generally underestimated due to the breakdown of perturbation theory. On average the empirical Padé approximants yield tumor absorption coefficients similar to third-order perturbation theory, yet at noticeable lower computational efforts.

Protoporphyrin IX occurs naturally in colorectal cancers and their metastases.

Colorectal cancers exhibit a red fluorescence. The nature of the responsible fluorophore and its eventual diagnostic potential were investigated. Thirty-three consecutive colorectal resection specimen, 32 of which with histologically confirmed cancer, and a total of 1053 palpable mesenteric nodes were fluorimetrically characterized ex vivo. Furthermore, frozen material from 28 patients was analyzed, selected for the availability of primary tumor material and metastatic tissue, e.g., lymphatic and liver metastases from the same patient. Biochemical characterization was carried out through chemical extraction and reversed phase high-performance liquid chromatography. The fluorescence spectra of tissues, tissue extracts, and standard solutions of porphyrins were determined using a pulsed solid-state laser system for excitation and an imaging polychromator, together with an intensified CCD camera for time-delayed observation. Protoporphyrin IX (PpIX) was identified as the predominant fluorophore in primary tumors and their metastases. The fluorophore occurred in the absence of necrosis and in sterile locations. In untreated cases (n = 24), PpIX fluorescence discriminates metastatically involved lymph nodes from all other palpable nodes with a sensitivity of 62% at a specificity of 78% (P < 0.0001). After neoadjuvant treatment of rectal cancer, the PpIX fluorescence level of the primary tumors was reduced and a discrimination of lymph nodes based on PpIX-fluorescence was impossible. We conclude that colorectal cancer metastases accumulate diagnostic levels of endogenous PpIX as a result of a tumor-specific metabolic alteration.

In-vivo tissue optical properties derived by linear perturbation theory for edge-corrected time-domain mammograms.

A valuable method is described to analyze time-domain optical mammograms measured in the slab-like geometry of the slightly compressed female breast with a method based on linear perturbation theory including edge correction. Perturbations in scattering and absorption coefficients were mapped applying a computationally efficient point model.

Somatosensory evoked fast optical intensity changes detected non-invasively in the adult human head.

This work is the first to report optical intensity changes (DeltaI/I approximately 0.05%) with a latency between 60 and 160ms after electrical median nerve stimulation at 5Hz detected non-invasively through the intact adult human skull in volunteers. The signal is localised and reproducible when measuring at the same position on successive examinations. Compared to previous reports of fast optical changes in the human adult by a single group (Psychophysiology, 32 (1995) 505) the here reported changes are much smaller. They are in line with results from a photon transport calculation on a head model employing data from exposed cortical tissue. The origin of the signal found here is still unclear, however, they might be the non-invasive equivalent to the scattering changes found in exposed cortical tissue studies (J. Neurophysiol., 78 (1997) 1707).

Fluorescence as a concept in colorectal lymph node diagnosis.

Fluorescence detection may constitute an appropriate means in gastrointestinal cancers to diagnose lymphatic tumor spread as opposed to gamma-scintillation methods. Photodiagnostic tracers have been shown to localize rapidly in malignant cells and may enable sensitive detection of small cell aggregates in lymph nodes. To reach a detection depth of several millimeters, a broad banded unspecific tissue autofluorescence may be controlled by so-called background subtracting techniques, generally based either on fluorescence observation at several wavelengths or on dual-wavelength fluorescence excitation. Using such comparative fluorescence detection techniques, some tumor entities can be differentiated soley based upon autofluorescence characteristics. Introducing a further enhancement in sensitivity for longer life-time fluorophores by time delayed fluorescence detection we ran a pilot trial comprising 174 lymph nodes from colorectal cancer specimen from 9 patients. Metastatically involved lymph nodes could be differentiated from all other palpable nodes in the mesenteric fat at a specificity of 85% with a sensitivity of 65%. Specific fluorescence features may be useful to preselect tissue samples for further histological analysis.

Near-infrared fluorescent dyes for enhanced contrast in optical mammography: phantom experiments.

Optical mammography with near-infrared (NIR) light using time-domain, frequency-domain, or continuous-wave techniques is a novel imaging modality to locate human breast tumors. By investigating excised specimens of normal and diseased mamma tissue we were able to demonstrate that differences in their scattering properties are a poor predictive parameter for normal and diseased mamma tissue. This paper describes the application of a NIR dye to improve the differentiation between breast tumors and normal tissue in a rat model. The NIR dye furnished a high tumor-to-tissue contrast ratio (6:1) in fluorescence images. Furthermore, this dye was used to develop liquid scattering phantoms with absorbing and fluorescent inhomogeneities. Using frequency-domain and time-domain instrumentation these inhomogeneities were localized at sufficient contrast by their increased absorption and fluorescence. Contrast between inhomogeneities and surrounding medium could be improved by combining fluorescence and transmittance images.

Macromolecular contrast agents for optical imaging of tumors: comparison of indotricarbocyanine-labeled human serum albumin and transferrin.

Macromolecules accumulate in solid tumors and can thus be used as carriers for the delivery of attached contrast agents to tumors. We report the synthesis and use of serum protein-dye conjugates consisting of transferrin (Tf) or human serum albumin (HSA) and an indotricarbocyanine (ITCC) derivative as contrast agents for the optical imaging of tumors. The compounds were characterized with respect to their photophysical properties and tested in vitro for their ability to bind to tumor cells and in vivo for their potential to delineate experimental tumors. In contrast to HAS-ITTC, Tf-ITCC showed receptor-mediated uptake by HT29 human colon cancer cells in vitro. After intravenous injection into HT29 tumor-bearing nude mice both compounds induced increased fluorescence contrast of tumors in vivo. After 24 h the contrast between tumor and normal tissue was significantly higher for Tf-ITCC than for HAS-ITCC. Dye-induced fluorescence was found to be predominantly located in perinecrotic areas of the tumor. Furthermore, Tf-ITCC produced fluorescence of viable tumor cells, whereas HAS-ITCC fluorescence was recorded along connective tissue. We conclude that ITCC-labeled Tf and HSA can serve as macromolecular contrast agents for the optical imaging of tumors, with Tf-ITCC showing higher efficiency.

Preparation of solid phantoms with defined scattering and absorption properties for optical tomography.

We have developed diffusely scattering solid phantoms with optical (scattering) properties amenable to theoretical calculations. Monodisperse quartz glass spheres were used as scatterers embedded in polyester resin. An infrared dye was added to simulate absorption by biological tissue. Solid phantoms were tested for their macroscopic homogeneity. Several phantoms were built with well-defined spatial variations in their transport scattering and absorption coefficients to be used for optical tomography. Scattering, transport scattering, and absorption coefficients of solid, homogeneous phantoms and of aqueous suspensions of monodisperse quartz glass spheres were derived from measurements of time-integrated collimated transmittance and time-resolved diffuse transmittance. For aqueous suspensions of monodisperse quartz glass spheres at known number density scattering and transport scattering coefficients calculated by Mie theory are in quantitative agreement with experimentally derived values. In addition, diffuse reflectance and diffuse transmittance of aqueous suspensions at various number densities were measured and found to be in excellent agreement with results of Monte Carlo calculations using theoretical values for the scattering coefficients and anisotropy parameters.

Determining changes in NIR absorption using a layered model of the human head.

A theoretical approach is presented to determine absorption changes in different compartments of a layered structure from distributions of times of flight of photons. In addition resulting changes in spatial profiles of time-integrated intensity and mean time of flight are calculated. The capability of a single-distance, time-domain method to determine absorption changes with depth resolution is tested on a layered phantom. We apply this method to in vivo measurements on the human head (motor stimulation, Valsalva manoeuvre) and introduce a small-sized time-domain experimental set-up suitable for bedside monitoring.

Comparison of four magnetic resonance methods for mapping small temperature changes.

Non-invasive detection of small temperature changes (< 1 degree C) is pivotal to the further advance of regional hyperthermia as a treatment modality for deep-seated tumours. Magnetic resonance (MR) thermography methods are considered to be a promising approach. Four methods exploiting temperature-dependent parameters were evaluated in phantom experiments. The investigated temperature indicators were spin-lattice relaxation time T1, diffusion coefficient D, shift of water proton resonance frequency (water PRF) and resonance frequency shift of the methoxy group of the praseodymium complex (Pr probe). The respective pulse sequences employed to detect temperature-dependent signal changes were the multiple readout single inversion recovery (T One by Multiple Read Out Pulses; TOMROP), the pulsed gradient spin echo (PGSE), the fast low-angle shot (FLASH) with phase difference reconstruction, and the classical chemical shift imaging (CSI). Applying these sequences, experiments were performed in two separate and consecutive steps. In the first step, calibration curves were recorded for all four methods. In the second step, applying these calibration data, maps of temperature changes were generated and verified. With the equal total acquisition time of approximately 4 min for all four methods, the uncertainties of temperature changes derived from the calibration curves were less than 1 degree C (Pr probe 0.11 degrees C, water PRF 0.22 degrees C, D 0.48 degrees C and T1 0.93 degrees C). The corresponding maps of temperature changes exhibited slightly higher errors but still in the range or less than 1 degree C (0.97 degrees C, 0.41 degrees C, 0.70 degrees C, 1.06 degrees C respectively). The calibration results indicate the Pr probe method to be most sensitive and accurate. However, this advantage could only be partially transferred to the thermographic maps because of the coarse 16 x 16 matrix of the classical CSI sequence. Therefore, at present the water PRF method appears to be most suitable for MR monitoring of small temperature changes during hyperthermia treatment.

Quantitative MRS: comparison of time domain and time domain frequency domain methods using a novel test procedure.

For quantitative analysis of in vivo MR spectra, a state-of-the-art time domain method was compared with a recently reported time domain frequency domain method which uses wavelets for background characterization. The comparison was made on the basis of results for simulated test problems that were constructed by combining measured and simulated MRS data at different signal-to-noise ratios in order to simultaneously reflect real world difficulties, in particular the overlapping background problem, and to allow for quantitative judgment of a method's accuracy. Incorporating prior knowledge was also considered. The results obtained give insight into the accuracy of the methods when applied to measured MRS data. Due to the improved background characterization, the time domain frequency domain method outperformed the time domain method in some of the test cases. Both methods were also applied to serial brain MR spectra of a healthy volunteer on 10 occasions.

Non-invasive detection of fluorescence from exogenous chromophores in the adult human brain.

This is the first report on results proving that fluorescence of exogenous dyes inside the human brain can be excited and detected non-invasively at the surface of the adult head. Boli of indocyanine green (ICG) were intravenously applied to healthy volunteers, and the passage of the contrast agent in the brain was monitored by detecting the corresponding fluorescence signal following pulsed laser excitation at 780 nm. Our hypothesis that the observed fluorescence signal contains a considerable cortical fraction was corroborated by performing measurements with picosecond temporal resolution and analyzing distributions of times of arrival of photons, hence taking advantage of the well-known depth selectivity of that method. Our experimental findings are explained by Monte Carlo simulations modeling the head as a layered medium and taking into account realistic bolus kinetics within the extra- and intracerebral compartment. Although a particular non-specific dye (ICG) was used, the results clearly demonstrate that fluorescence-mediated imaging of the adult human brain is generally feasible. In particular, we will discuss how these results serve as proof of concept for non-invasive fluorescence brain imaging and may thus open the door towards optical molecular imaging of the human brain.

Bed-side assessment of cerebral perfusion in stroke patients based on optical monitoring of a dye bolus by time-resolved diffuse reflectance.

We present a minimally invasive optical method, that is, multi-channel time-domain diffuse near-infrared reflectometry of the head to assess cerebral blood perfusion that is applicable at the bed-side and repetitively at short intervals. Following intravenous injection of an ICG bolus, its transit through intra- and extracerebral tissue is monitored based on changes in moments of distributions of times of flight of photons, recorded with a 4-channel instrument simultaneously on both hemispheres. In healthy volunteers, we found that variance of distributions of times of flight of photons is well suited to assess latency and initial slope of the increase in absorption of intracerebral tissue due to the bolus. We successfully applied our method in two patients demonstrating a reversible cerebral perfusion deficit in an ischemic stroke patient who was treated by thrombolysis and in another patient with a permanent impaired unilateral perfusion due to ipsilateral internal carotid artery occlusion. In either case, we observed a difference in bolus transit time between the hemispheres. In the stroke patient, this difference resolved when re-evaluated 1 day after thrombolysis. The study demonstrates the necessity of a technique with sub-nanosecond time resolution to allow for depth discrimination if clinical perfusion monitoring of cerebrovascular diseases is addressed by optical methods.

Imaging in turbid media by photon density waves: spatial resolution and scaling relations.

Spatial resolution of transillumination imaging through highly scattering media normalized to sample thickness d depends on only the normalized wavelength ?/d of photon density waves and normalized penetration depth delta/d. This was concluded theoretically and verified experimentally by the derivation of edge-spread functions from measured time-resolved transmittance and its Fourier transform by the use of dilute milk at various concentrations as scatterer in cuvettes of d = 2 cm and d = 4 cm. In the frequency domain and the time domain, spatial resolution was found experimentally to be given by approximately 0.3 d obtained at ? approximately d or when only the first arriving 1% of all photons detected were taken into account.