TY - BOOK
T1 - Bioimpedance-Based Real-Time Tissue-Sensing for Needle Guidance
AU - Halonen, Sanna
PY - 2026
Y1 - 2026
N2 - Physicians need to perform demanding injections and biopsies without visual control, even though achieving the right location is essential for proper treatment and diagnosis. Uncertainty in needle tip location during injection or sampling may increase complication risks, cause delays in diagnosis and treatment, increase puncture duration, unnecessary pain and tissue damage.
Bioimpedance may provide a feasible answer to the need of simple needle guidance. It measures tissue electrical properties and is safe, non-invasive and possible to implement in various of instruments. Prior to this thesis, the research has been focused on developing the measurement technology mainly in the laboratory environment. Robust real-time bioimpedance-based tissue-sensing measurement device and classification algorithms for needle guidance have not been satisfyingly studied and optimized application specifically in clinical practice in vivo.
The aim of this thesis was to develop and evaluate real-time bioimpedance-based needle guidance and tissue detection device for demanding clinical punctures. In this work the performance and feasibility of the method is studied and evaluated first in detecting body fluids, then in detecting soft tissues and finally in detecting tumorous tissue from surrounding non-tumorous tissue. Thesis work focuses on evaluating the method in real clinical use, which includes tissue microstructures and patient variation. The example applications studied are intra-articular injections, lumbar punctures for spinal anesthesia and liver biopsies.
Bioimpedance based tissue detecting device consists of sensor electrodes, impedance analyzer and classification algorithm. The measurement electrode utilized in this thesis is conventional looking needle which includes insulated metal wire. The measurement is performed between the needle cannula and the tip of the inner metal wire. Multiple injection needle sizes and one biopsy needle size are implemented. Real-time measurement is implemented using binary multifrequency excitation signal which has most of its power within 15 frequencies between 1 kHz and 349 kHz.
The classification algorithm utilized in this thesis is statistical, based on Bayesian clas-sifier. First the information within impedance spectra is transformed into features and then classified into tissue type. The features, classifier parameters, included tissue classes and the target tissue is selected and optimized application specifically.
The methods included finite element method simulations, in vivo animal study of one pig and three clinical in vivo studies. Real-time fluid detection was studied in intra-articular injections to 80 joints (51 patients) and in spinal anesthesia lumbar punctures to 45 patients. Tumor detection was studied in liver biopsy study with 26 patients.
Clinical in vivo studies in intra-articular injections and spinal anesthesia lumbar punctures showed that the method is very feasible for real-time fluid detection also in clinical environment with real patients. The accuracy to detect synovial fluid was 85% in 80 injections to different joint types, and accuracy to detect cerebrospinal fluid was 91% in 45 patients. Animal study of one pig showed that the device is capable to accurately classify also soft tissues which can have very small conductivity differences. Total accuracy to detect soft tissues was 94% to classify liver, spleen, muscle, fat and blood. Impedance signal varies a lot when measuring living tissue. Already small movement of the needle to another location changed the impedance signal significantly. Finite element method computer simulations showed that already 0.2 mm thin heterogeneities change the impedance values significantly. On the other hand, the simulations also showed that the measurement is spatially precise. Heterogeneities next to the needle tip do not affect the measurement. The sensitivity distribution is focused on the needle facet and targets with size 2 x 2 x 2 mm3 and larger are correctly measurable. Clinical in vivo liver biopsy study with 26 patients showed that tumor impedance spectrum is different from impedance spectrum from non-tumorous tissue. Difference was significant in impedance magnitude at frequencies below 25 kHz and in phase angle below 3 kHz and above 30 kHz. Developing accurate real-time tumor tissue detection requires more studies and larger data sets.
The studied real-time bioimpedance-based tissue detection method was proven to be spatially precise and feasible in clinical environment with real patients. Since the device is capable to detect tissue at the needle tip in real-time adequately accurately, it has potential to help performing demanding punctures. In the future the method may change the state-of-the-art in performing demanding punctures.
AB - Physicians need to perform demanding injections and biopsies without visual control, even though achieving the right location is essential for proper treatment and diagnosis. Uncertainty in needle tip location during injection or sampling may increase complication risks, cause delays in diagnosis and treatment, increase puncture duration, unnecessary pain and tissue damage.
Bioimpedance may provide a feasible answer to the need of simple needle guidance. It measures tissue electrical properties and is safe, non-invasive and possible to implement in various of instruments. Prior to this thesis, the research has been focused on developing the measurement technology mainly in the laboratory environment. Robust real-time bioimpedance-based tissue-sensing measurement device and classification algorithms for needle guidance have not been satisfyingly studied and optimized application specifically in clinical practice in vivo.
The aim of this thesis was to develop and evaluate real-time bioimpedance-based needle guidance and tissue detection device for demanding clinical punctures. In this work the performance and feasibility of the method is studied and evaluated first in detecting body fluids, then in detecting soft tissues and finally in detecting tumorous tissue from surrounding non-tumorous tissue. Thesis work focuses on evaluating the method in real clinical use, which includes tissue microstructures and patient variation. The example applications studied are intra-articular injections, lumbar punctures for spinal anesthesia and liver biopsies.
Bioimpedance based tissue detecting device consists of sensor electrodes, impedance analyzer and classification algorithm. The measurement electrode utilized in this thesis is conventional looking needle which includes insulated metal wire. The measurement is performed between the needle cannula and the tip of the inner metal wire. Multiple injection needle sizes and one biopsy needle size are implemented. Real-time measurement is implemented using binary multifrequency excitation signal which has most of its power within 15 frequencies between 1 kHz and 349 kHz.
The classification algorithm utilized in this thesis is statistical, based on Bayesian clas-sifier. First the information within impedance spectra is transformed into features and then classified into tissue type. The features, classifier parameters, included tissue classes and the target tissue is selected and optimized application specifically.
The methods included finite element method simulations, in vivo animal study of one pig and three clinical in vivo studies. Real-time fluid detection was studied in intra-articular injections to 80 joints (51 patients) and in spinal anesthesia lumbar punctures to 45 patients. Tumor detection was studied in liver biopsy study with 26 patients.
Clinical in vivo studies in intra-articular injections and spinal anesthesia lumbar punctures showed that the method is very feasible for real-time fluid detection also in clinical environment with real patients. The accuracy to detect synovial fluid was 85% in 80 injections to different joint types, and accuracy to detect cerebrospinal fluid was 91% in 45 patients. Animal study of one pig showed that the device is capable to accurately classify also soft tissues which can have very small conductivity differences. Total accuracy to detect soft tissues was 94% to classify liver, spleen, muscle, fat and blood. Impedance signal varies a lot when measuring living tissue. Already small movement of the needle to another location changed the impedance signal significantly. Finite element method computer simulations showed that already 0.2 mm thin heterogeneities change the impedance values significantly. On the other hand, the simulations also showed that the measurement is spatially precise. Heterogeneities next to the needle tip do not affect the measurement. The sensitivity distribution is focused on the needle facet and targets with size 2 x 2 x 2 mm3 and larger are correctly measurable. Clinical in vivo liver biopsy study with 26 patients showed that tumor impedance spectrum is different from impedance spectrum from non-tumorous tissue. Difference was significant in impedance magnitude at frequencies below 25 kHz and in phase angle below 3 kHz and above 30 kHz. Developing accurate real-time tumor tissue detection requires more studies and larger data sets.
The studied real-time bioimpedance-based tissue detection method was proven to be spatially precise and feasible in clinical environment with real patients. Since the device is capable to detect tissue at the needle tip in real-time adequately accurately, it has potential to help performing demanding punctures. In the future the method may change the state-of-the-art in performing demanding punctures.
M3 - Doctoral thesis
SN - 978-952-03-4244-9
T3 - Tampere University Dissertations - Tampereen yliopiston väitöskirjat
BT - Bioimpedance-Based Real-Time Tissue-Sensing for Needle Guidance
PB - Tampere University
CY - Tampere
ER -