Biomedical subjects

Tomas Svensson

Publications and source records attributed to Tomas Svensson.

4 recordsLinked to original sources

PubMed2005

Least-squares support vector machines modelization for time-resolved spectroscopy.

By use of time-resolved spectroscopy it is possible to separate light scattering effects from chemical absorption effects in samples. In the study of propagation of short light pulses in turbid samples the reduced scattering coefficient and the absorption coefficient are usually obtained by fitting diffusion or Monte Carlo models to the measured data by use of numerical optimization techniques. In this study we propose a prediction model obtained with a semiparametric modeling technique: the least-squares support vector machines. The main advantage of this technique is that it uses theoretical time dispersion curves during the calibration step. Predictions can then be performed by use of data measured on different kinds of sample, such as apples.

Algorithms↗

PubMed2005

Characterization of normal breast tissue heterogeneity using time-resolved near-infrared spectroscopy.

In recent years, extensive efforts have been made in developing near-infrared optical techniques to be used in detection and diagnosis of breast cancer. Variations in optical properties of normal breast tissue set limits to the performance of such techniques and must therefore be thoroughly examined. In this paper, we present intra- and intersubject as well as contralateral variations of optical and physiological properties in breast tissue as measured by using four-wavelength time-resolved spectroscopy (at 660, 786, 916 and 974 nm). In total, 36 volunteers were examined at five regions at each breast. Optical properties (absorption, mu(a), and reduced scattering, mu'(s)) are derived by employing diffusion theory. The use of four wavelengths enables determination of main tissue chromophores (haemoglobin, water and lipids) as well as haemoglobin oxygenation. Variations in all evaluated properties seen over the entire breast are approximately twice those for small-scale heterogeneity (millimetre scale). Intrasubject variations in optical properties are almost in all cases below 20% for mu'(s), and 40% for mu(a). Overall variations in water, lipid and haemoglobin concentrations are all in the order of 20%. Oxygenation is the least variable of the quantities evaluated, overall intrasubject variations being 6% on average. Extracted physiological properties confirm differences between pre- and post-menopausal breast tissue. Results do not indicate systematic differences between left and right breasts.

Adult↗

PubMed2005

Performance assessment of photon migration instruments: the MEDPHOT protocol.

We propose a comprehensive protocol for the performance assessment of photon migration instruments. The protocol has been developed within the European Thematic Network MEDPHOT (optical methods for medical diagnosis and monitoring of diseases) and is based on five criteria: accuracy, linearity, noise, stability, and reproducibility. This protocol was applied to a total of 8 instruments with a set of 32 phantoms, covering a wide range of optical properties.

Benchmarking↗

PubMed2005

Scatter correction of transmission near-infrared spectra by photon migration data: quantitative analysis of solids.

The scope of this work is a new methodology to correct conventional near-infrared (NIR) data for scattering effects. The technique aims at measuring the absorption coefficient of the samples rather than the total attenuation measured in conventional NIR spectroscopy. The main advantage of this is that the absorption coefficient is independent of the path length of the light inside the sample and therefore independent of the scattering effects. The method is based on time-resolved spectroscopy and modeling of light transport by diffusion theory. This provides an independent measure of the scattering properties of the samples and therefore of the path length of light. This yields a clear advantage over other preprocessing techniques, where scattering effects are estimated and corrected for by using the shape of the measured spectrum only. Partial least squares (PLS) calibration models show that, by using the proposed evaluation scheme, the predictive ability is improved by 50% as compared to a model based on conventional NIR data alone. The method also makes it possible to predict the concentration of active substance in samples with other physical properties than the samples included in the calibration model.

Algorithms↗