rus
Issue #5/2018
I. A. Raznitsyna, P. A. Kulikova, A. A. Glazkov, D. A. Rogatkin
Laser nondestructive technologies for verification of organs and tissues in low-invasive and robot-assisted surgery
Laser nondestructive technologies for verification of organs and tissues in low-invasive and robot-assisted surgery
The need to develop methods for organs and tissues differentiation in surgery is discussed. A multifactor analysis of fluorescence and backscattering spectra was carried out for different tissues. Eight out of eleven types of tissues were verified with the accuracy of the order of 80%. The obtained results allow us to assume that optical technologies could form the basis for the method of navigation in robot-assisted and low-invasive surgery in the future.
Теги: backscattered spectroscopy laser fluorescence spectroscopy optical technologies surgical navigation technology лазерная флюоресцентная спектроскопия оптические технологии спектроскопия обратного рассеяния хирургическая навигация
Any surgical intervention, even a minimal one, poses potential danger, which is not always possible to predict [1]. A dramatically increasing number of adverse effects of treatment, including surgical treatment, gave rise to a new term, "jatroepidemic" [2]. If at the preoperative stage the doctor can apply a large variety of instrumental and diagnostic approaches that help to take a decision and identify treatment strategy, at the time of surgery, the surgeon often has to rely on his / her subjective evaluation of the operating field and the data obtained at the preoperative stage. In other words, the surgeon is often forced to take quick decisions, focusing on visual and tactile evaluation of tissues, and sometimes intuitively. In terms of laparoscopic and robot-assisted surgery, intensively developing today, the doctor is almost devoid of tactile sensations, and color reproduction of the image may be slightly altered. With normal anatomy, in most cases an experienced surgeon can easily differentiate organs and tissues in such a situation, however, in the case of anomalies or variations in the organ development, the surgeon may have difficulty in verifying certain structures [3]. It is known that 5% of defects in the provision of surgical care, which led to complaints and forensic medical examinations, are connected with the organ development anomaly [1]. Also, the proximity of anatomical structures, pronounced peritoneal adhesions increase the risk of making a mistake and incorrectly differentiate organs and tissues, which can lead to damage to healthy tissues.
Thus, the necessity to create a convenient, widely available device that allows the surgeon to perform an intraoperative rapid identification of anatomical structures and their verification without damaging the tissue is apparent. To date, a number of devices for robot-assisted surgery, intended for neurosurgical operations (e. g., Spine Assist, Mazor Surgical Technologies, Caesarea, Israel), are equipped with a CT-control system, but its use is associated with radiation burden for the body. The use of ultrasound techniques that are safe for human in robot-assisted surgery is difficult today due to the need to provide impedance matching, which is not possible for open wounds. The use of a magnetic field in surgery conditions is also problematic.
There is another approach to solving this problem. It is known, that different types of tissue have different backscattering spectra of electromagnetic radiation in the visible region. Also, the difference in the content of such natural (endogenous) fluorophores, as collagen, elastin, keratin, porphyrin, etc., is also reflected in the fluorescence spectra obtained in vivo. One can record these differences by recording the spectrum secondary radiation after the object is exposed to narrow-band (laser) and broadband (lamp) sources of low-intensity visible light. The advantage of this approach is, firstly, the safety of using low-power radiation for tissues, at mode not destroying them. Secondly, most surgical robots are already equipped with a visualization system made in the form of endoscopic probes, and the introduction of additional channels for the delivery and registration of diagnostic radiation will not be of great technical difficulty. Thirdly, optical radiation allows you to receive information from the object in real time, which facilitates the instantaneous transfer of information to the surgeon.
The main obstacle standing in the way of implementation of this technology into practice, is the lack of well-proved methods and numerical criteria to define the type of tissue by its optical spectrum. There are many researches devoted to the analysis for the optical spectra of backscattering of various tissues and finding specific quantitative criteria for their differentiation [4]. There are also researches attempting to verify biological tissues using fluorescence analysis [5]. Nevertheless, none of these methods has not become universally recognized and has not been introduced into common practice for a variety of reasons, primarily because of low accuracy and specificity [6]. However, there is reason to believe that simultaneous evaluation of both fluorescence spectra and diffuse reflectance spectra of tissues can, nevertheless, help in solving the tasks faced by practicing surgeons.
From the technical point of view, the principle of operation of the majority of optical spectroscopic systems for in vivo medical diagnostics is similar: low-intensity optical radiation (narrow-band and / or broadband) interacting with the biological tissue is approached to the surface of the object under study using a fiber-optic probe. Due to differences in the level of the blood supply, biochemical composition and morphology, the biological tissues possess different absorbing, scattering and fluorescing properties. Light, penetrating into the biological tissue, undergoes various linear (elastic scattering, absorption) and nonlinear (fluorescence) interactions inside, therefore the part of radiation, leaving the tissue due to multiple scattering, that enters the detection fiber, carries comprehensive information about the tissue composition and structure. However, this information still needs to be decoded, since, for example, the detected fluorescence spectra are heavily influenced by absorbing and scattering properties of the biological tissues. Therefore, the use of combined methods of diagnosis is required to solve this problem.
STUDY OF BACKSCATTERING AND FLUORESCENCE SPECTRA ILLUSTRATED BY LABORATORY ANIMALS’ TISSUES
In the laboratory of medical and physical researches of M. F. Vladimirskiy Moscow Regional Research Clinical Institute, an experiment to evaluate secondary emission spectra after exposure of various types of tissues to narrow-band and broadband (white light) sources of optical radiation was conducted. The optical properties of various tissues of laboratory rats were studied. For analysis, the animals were sacrificed by administering a lethal dose of anesthetic (Zoletil 200 mg / kg), various fragments of different tissues and organs were taken: a fragment of lung (n = 6), cross-striated muscle tissue (n = 8), liver (n = 6), kidney (n = 7), omentum (n = 7), cardiac muscle (n = 7), a fragment of testicle (n = 6), bone (n = 4), nerve tissue (n = 6), a fragment of esophagus (n = 5), skin (n = 6). The organs and fragments of organs were washed in physiological solution and placed on a substrate made of a light-absorbing material. The measurements of backscattering spectra and tissue fluorescence were performed as follows: the distal end of the fiber optic probe of the diagnostic system was approached to organ surfaces, installed perpendicular to the surface area studied. The tissue studied was illuminated with laser radiation with a wavelength, e. g., λe = 365 nm. Further, by the user’s command, the endogenous fluorescence spectrum was registered. An example of a registered spectrum is shown in Fig. 1. Here several maxima are clearly observed: the backscattering peak at the excitation wavelength Ibs(λe) and the maximum at the fluorescence wavelengths If(λf) of endogenous fluorophores. The real value of Ibs(λe) is approximately in β = 1 000 times larger than that shown in the figure. In order to visually compare fluorescence intensities on the personal computer (PC) monitor, a system of optical filters (Fig. 2), reducing the magnitude of the backscattering peak, is used in the instrument.
Upon completion of the collection of the fluorescence signal, multiple pulses from a white light source (a xenon lamp) were automatically fed to the tissue, followed by a registration of backscattering broadband spectrum. For our study, the number of pulses was four, since it was shown empirically that with this number of pulses, it is possible to obtain the largest detectable signal, not exceeding the maximum possible for instrumental registration without saturation.
For the analysis of fluorescence spectra, we used an array of fluorescence contrast coefficients [7]:
.
For tissue backscattering spectrum that depends only on the properties of the tissue itself, normalization of the measurement results was performed with a white light source by the spectra taken from the material having a known backscattering coefficient r – fluoroplastic (r = 0,85).
For each tissue sample, 1500 values of fluorescence intensity in the wavelength range 400–730 nm, normalized by backscattering peak (excitation wavelength λe = 365 nm) and 1 600 values of backscattering intensity in the range 390–740 nm were analyzed when the tissue was exposed to a white light source.
To reduce the number of variables, the principal component analysis with rotation (varimax) was applied. For the fluorescence spectrum, 98.3% of the dispersion was explained by five principal components. For the spectrum of the intensity of diffuse scattering, 6 principal components allowed to explain 92.2% of the dispersion. The 11 principal components obtained were included in the discriminant analysis. Statistical processing of data was carried out by IBM SPSS Statistics v23 (IBM corp., USA).
The study was carried out on a new diagnostic instrument "Multicom", developed by LLC " Center for Research and Development EOS-Medica" combining the principles of laser fluorescence spectroscopy and backscattering spectroscopy [8]. Schematic diagram of the instrument is shown in Fig. 2.
Control and processing system may generate, receive and process two basic control commands: "observation" and "measurement". The "observation" command activates the selected laser source continuously and continuously registers fluorescence spectrum of the secondary radiation by the spectrometer. The "measurement" command stores the last measured fluorescence spectrum in the device memory, the laser source is switched off, the white light source is turned on for a short time and the spectrometer registers the reflection spectrum in white light, after which all the measured spectra are transferred to the control and processing system. It should be mentioned that the arrangement of the lighting fibers is uniform, around the receiving fiber, thus achieving the equality of diagnostic volumes for two successive measurements.
RESULTS OBTAINED
The results of discriminant analysis using 11 principal components from fluorescence intensity spectra and backscattering intensity are presented in Table. As you can see, 88.2% of the selected initial grouped observations are classified correctly. Thus, the proposed physical and mathematical approaches allow to differentiate the examined organs and tissues with relatively high accuracy.
CONCLUSION
The instrument allowing for rapid verification of different tissues and organs, may be useful for intraoperative navigation and in the future may allow to reduce the risk of jatrogenic injuries in surgery. One possible approach for creating such an instrument can be a combined application of different optical methods, namely, laser fluorescence spectroscopy and backscattering spectroscopy. The results of our research have shown that a multifactor analysis of fluorescence and backscattering spectra from biological tissues can successfully differentiate organs and tissues. Thus, nine out of 11 dedicated contact tissue types can be determined with an accuracy of approximately 80%. Such a result is a good basis for carrying out a larger experiment with a large training sample and a large number of tissue types. Since the applied optical methods are available, they do not damage the tissue and give real-time information with user-friendly interface, they can become a reliable additional navigation method during surgery.
Thus, the necessity to create a convenient, widely available device that allows the surgeon to perform an intraoperative rapid identification of anatomical structures and their verification without damaging the tissue is apparent. To date, a number of devices for robot-assisted surgery, intended for neurosurgical operations (e. g., Spine Assist, Mazor Surgical Technologies, Caesarea, Israel), are equipped with a CT-control system, but its use is associated with radiation burden for the body. The use of ultrasound techniques that are safe for human in robot-assisted surgery is difficult today due to the need to provide impedance matching, which is not possible for open wounds. The use of a magnetic field in surgery conditions is also problematic.
There is another approach to solving this problem. It is known, that different types of tissue have different backscattering spectra of electromagnetic radiation in the visible region. Also, the difference in the content of such natural (endogenous) fluorophores, as collagen, elastin, keratin, porphyrin, etc., is also reflected in the fluorescence spectra obtained in vivo. One can record these differences by recording the spectrum secondary radiation after the object is exposed to narrow-band (laser) and broadband (lamp) sources of low-intensity visible light. The advantage of this approach is, firstly, the safety of using low-power radiation for tissues, at mode not destroying them. Secondly, most surgical robots are already equipped with a visualization system made in the form of endoscopic probes, and the introduction of additional channels for the delivery and registration of diagnostic radiation will not be of great technical difficulty. Thirdly, optical radiation allows you to receive information from the object in real time, which facilitates the instantaneous transfer of information to the surgeon.
The main obstacle standing in the way of implementation of this technology into practice, is the lack of well-proved methods and numerical criteria to define the type of tissue by its optical spectrum. There are many researches devoted to the analysis for the optical spectra of backscattering of various tissues and finding specific quantitative criteria for their differentiation [4]. There are also researches attempting to verify biological tissues using fluorescence analysis [5]. Nevertheless, none of these methods has not become universally recognized and has not been introduced into common practice for a variety of reasons, primarily because of low accuracy and specificity [6]. However, there is reason to believe that simultaneous evaluation of both fluorescence spectra and diffuse reflectance spectra of tissues can, nevertheless, help in solving the tasks faced by practicing surgeons.
From the technical point of view, the principle of operation of the majority of optical spectroscopic systems for in vivo medical diagnostics is similar: low-intensity optical radiation (narrow-band and / or broadband) interacting with the biological tissue is approached to the surface of the object under study using a fiber-optic probe. Due to differences in the level of the blood supply, biochemical composition and morphology, the biological tissues possess different absorbing, scattering and fluorescing properties. Light, penetrating into the biological tissue, undergoes various linear (elastic scattering, absorption) and nonlinear (fluorescence) interactions inside, therefore the part of radiation, leaving the tissue due to multiple scattering, that enters the detection fiber, carries comprehensive information about the tissue composition and structure. However, this information still needs to be decoded, since, for example, the detected fluorescence spectra are heavily influenced by absorbing and scattering properties of the biological tissues. Therefore, the use of combined methods of diagnosis is required to solve this problem.
STUDY OF BACKSCATTERING AND FLUORESCENCE SPECTRA ILLUSTRATED BY LABORATORY ANIMALS’ TISSUES
In the laboratory of medical and physical researches of M. F. Vladimirskiy Moscow Regional Research Clinical Institute, an experiment to evaluate secondary emission spectra after exposure of various types of tissues to narrow-band and broadband (white light) sources of optical radiation was conducted. The optical properties of various tissues of laboratory rats were studied. For analysis, the animals were sacrificed by administering a lethal dose of anesthetic (Zoletil 200 mg / kg), various fragments of different tissues and organs were taken: a fragment of lung (n = 6), cross-striated muscle tissue (n = 8), liver (n = 6), kidney (n = 7), omentum (n = 7), cardiac muscle (n = 7), a fragment of testicle (n = 6), bone (n = 4), nerve tissue (n = 6), a fragment of esophagus (n = 5), skin (n = 6). The organs and fragments of organs were washed in physiological solution and placed on a substrate made of a light-absorbing material. The measurements of backscattering spectra and tissue fluorescence were performed as follows: the distal end of the fiber optic probe of the diagnostic system was approached to organ surfaces, installed perpendicular to the surface area studied. The tissue studied was illuminated with laser radiation with a wavelength, e. g., λe = 365 nm. Further, by the user’s command, the endogenous fluorescence spectrum was registered. An example of a registered spectrum is shown in Fig. 1. Here several maxima are clearly observed: the backscattering peak at the excitation wavelength Ibs(λe) and the maximum at the fluorescence wavelengths If(λf) of endogenous fluorophores. The real value of Ibs(λe) is approximately in β = 1 000 times larger than that shown in the figure. In order to visually compare fluorescence intensities on the personal computer (PC) monitor, a system of optical filters (Fig. 2), reducing the magnitude of the backscattering peak, is used in the instrument.
Upon completion of the collection of the fluorescence signal, multiple pulses from a white light source (a xenon lamp) were automatically fed to the tissue, followed by a registration of backscattering broadband spectrum. For our study, the number of pulses was four, since it was shown empirically that with this number of pulses, it is possible to obtain the largest detectable signal, not exceeding the maximum possible for instrumental registration without saturation.
For the analysis of fluorescence spectra, we used an array of fluorescence contrast coefficients [7]:
.
For tissue backscattering spectrum that depends only on the properties of the tissue itself, normalization of the measurement results was performed with a white light source by the spectra taken from the material having a known backscattering coefficient r – fluoroplastic (r = 0,85).
For each tissue sample, 1500 values of fluorescence intensity in the wavelength range 400–730 nm, normalized by backscattering peak (excitation wavelength λe = 365 nm) and 1 600 values of backscattering intensity in the range 390–740 nm were analyzed when the tissue was exposed to a white light source.
To reduce the number of variables, the principal component analysis with rotation (varimax) was applied. For the fluorescence spectrum, 98.3% of the dispersion was explained by five principal components. For the spectrum of the intensity of diffuse scattering, 6 principal components allowed to explain 92.2% of the dispersion. The 11 principal components obtained were included in the discriminant analysis. Statistical processing of data was carried out by IBM SPSS Statistics v23 (IBM corp., USA).
The study was carried out on a new diagnostic instrument "Multicom", developed by LLC " Center for Research and Development EOS-Medica" combining the principles of laser fluorescence spectroscopy and backscattering spectroscopy [8]. Schematic diagram of the instrument is shown in Fig. 2.
Control and processing system may generate, receive and process two basic control commands: "observation" and "measurement". The "observation" command activates the selected laser source continuously and continuously registers fluorescence spectrum of the secondary radiation by the spectrometer. The "measurement" command stores the last measured fluorescence spectrum in the device memory, the laser source is switched off, the white light source is turned on for a short time and the spectrometer registers the reflection spectrum in white light, after which all the measured spectra are transferred to the control and processing system. It should be mentioned that the arrangement of the lighting fibers is uniform, around the receiving fiber, thus achieving the equality of diagnostic volumes for two successive measurements.
RESULTS OBTAINED
The results of discriminant analysis using 11 principal components from fluorescence intensity spectra and backscattering intensity are presented in Table. As you can see, 88.2% of the selected initial grouped observations are classified correctly. Thus, the proposed physical and mathematical approaches allow to differentiate the examined organs and tissues with relatively high accuracy.
CONCLUSION
The instrument allowing for rapid verification of different tissues and organs, may be useful for intraoperative navigation and in the future may allow to reduce the risk of jatrogenic injuries in surgery. One possible approach for creating such an instrument can be a combined application of different optical methods, namely, laser fluorescence spectroscopy and backscattering spectroscopy. The results of our research have shown that a multifactor analysis of fluorescence and backscattering spectra from biological tissues can successfully differentiate organs and tissues. Thus, nine out of 11 dedicated contact tissue types can be determined with an accuracy of approximately 80%. Such a result is a good basis for carrying out a larger experiment with a large training sample and a large number of tissue types. Since the applied optical methods are available, they do not damage the tissue and give real-time information with user-friendly interface, they can become a reliable additional navigation method during surgery.
Readers feedback
