Automated Segmentation of Infarct Lesions in T1-Weighted MRI Scans Using Variational Mode Decomposition and Deep Learning.

Automated segmentation methods are critical for early detection, prompt actions, and immediate treatments in reducing disability and death risks of brain infarction. This paper aims to develop a fully automated method to segment the infarct lesions from T1-weighted brain scans. As a key novelty, the proposed method combines variational mode decomposition and deep learning-based segmentation to take advantages of both methods and provide better results. There are three main technical contributions in this paper. First, variational mode decomposition is applied as a pre-processing to discriminate the infarct lesions from unwanted non-infarct tissues. Second, overlapped patches strategy is proposed to reduce the workload of the deep-learning-based segmentation task. Finally, a three-dimensional U-Net model is developed to perform patch-wise segmentation of infarct lesions. A total of 239 brain scans from a public dataset is utilized to develop and evaluate the proposed method. Empirical results reveal that the proposed automated segmentation can provide promising performances with an average dice similarity coefficient (DSC) of 0.6684, intersection over union (IoU) of 0.5022, and average symmetric surface distance (ASSD) of 0.3932, respectively.

Microwave Ablation Using Four-Tine Antenna: Effects of Blood Flow Velocity, Vessel Location, and Total Displacement on Porous Hepatic Cancer Tissue.

This research is concerned with microwave ablation analyses using a 2.45 GHz four-tine (4T) antenna for hepatic cancer tissue. In the study, three-dimensional finite-element models were utilized to examine the tissue temperature distributions during and after MW ablation. A preliminary study was first carried out with regard to the specific absorption rates along the 4T antenna insertion depths and the temperature distributions inside the solid and porous liver models with either 3 cm-in-diameter tumor or 5 cm-in-diameter tumor. Based on the preliminary results, the porous models were further examined for the effect of varying blood flow velocities (0-200 cm/s) with a 1 cm-in-diameter blood vessel next to the antenna and also for the effect of vessel-antenna locations (0, 0.8, and 1.3 cm) with a constant blood flow velocity of 16.7 cm/s. All scenarios were simulated under temperature-controlled mode (90°C). The findings revealed that the blood flow velocity and vessel location influence the ablation effectiveness and that increased blood flow inhibits heat transfer to the vessel wall. At the nearest and farthest vessel-antenna locations (0 and 1.3 cm), approximately 90.3% and 99.55% of the cancer cells were eradicated except for the areas adjacent to the vessel. In addition, total tissue thermal displacement is 5.9 mm which is 6.59% of the total length of the overall model.

Effect of phase difference in multi-antenna microwave thermal ablation for breast cancer treatment.

It was realized that cancer in breast is one of the most health hazards threatening women around the world for many years. Thermal ablation by using microwave energy is another alternative surgical maneuver due to its minimally invasive therapeutic technique. In this research, we investigate an effect of phase difference between three adjacent opened-slot coaxial probes in a multiple antenna alignment of microwave thermal ablation system for breast cancer treatment. FEM by using COMSOL is an implementation tools to simulate for 0, 45, 90, 135 and 180 degree of phase difference. 3D Simulation results show that temperature distribution pattern, destructive volume and SAR in breast tissue are affected from those phase-shift utilization in multi-antenna system significantly.

Finite-element analysis and in vitro experiments of placement configurations using triple antennas in microwave hepatic ablation.

This study presents analyses of triple-antenna configurations and designs for microwave (MW) hepatic ablation using 3-D finite-element (FE) analyses verified by in vitro experiments. Treatment of hepatic cancer often requires removal or destruction of large volume lesions. Using multiple antennas offers a potential solution for creating ablation zones with larger dimensions, as well as varied geometrical shapes. We performed both 3-D FE analyses and in vitro experiments using three identical open-tip MW antennas simultaneously, placing them in three types of configurations-"linear array," "triangular," and "T-shaped" arrangements. We compared coagulation volumes created, as well as temperature distribution characteristics, from the three-antenna arrangements after power delivery of 50 W for 60 s. We also performed additional tests using nonidentical antennas (open tip, slot, and slot with insulating jacket) for the three configurations. The results illustrate that arranging antennas in the "T-shaped" pattern destroyed more unwanted tissues than those found when using "linear array" and "triangular" arrangements, with maximum coagulation width and depth of 46 and 81 mm, respectively, and coagulation volume of 30.7 cm(3) . In addition, using nonidentical triple antennas caused variations in coagulation zone characteristics, and thus, the technique could be applied to treatment situations where nonsymmetric coagulation zones are required.

In vitro measurement of myocardial impedivity anisotropy with a miniature rectangular tube.

Due to rapid change of fiber orientation, it is difficult to measure myocardial impedivity separately in a longitudinal or transverse fiber direction without mutual influence in the two directions. Previously published values of the longitudinal and the transverse myocardial impedivity were derived indirectly from measurements that mixed the impedivity in all directions. Those values are questionable because the derivations were based on a simplified uniform myocardial fiber model. In this paper, a miniature rectangular tube was devised to facilitate direct measurement of myocardial impedivity in a uniform fiber direction. The average transverse-to-longitudinal ratio of the measured in vitro swine myocardial impedivity was about 1.66 from 1 Hz to 1 kHz and dropped to 1.25 at 1 MHz. The result is important for accurate modeling of the electrical property of myocardium in biomedical research of radio-frequency cardiac catheter ablation.

Mechanical compliance of the endocardium.

Radio-frequency (RF) ablation is an accepted treatment for cardiac arrhythmias related to abnormal focal cardiac substrate. The penetration depth of the electrode into the endocardium affects lesion size, a critical determinant of success of RF ablation. We measured the relation between the mechanical compliance and the penetration depth of RF ablation catheter electrode at frequently ablated areas of the endocardium and examined the influence of time after death on mechanical properties of the tissue. We measured force versus time for eight insertion depths of the catheter electrode into full-thickness endocardial samples derived from the mitral valve annulus, the left ventricular free wall and the tricuspid valve annulus. We varied the time after death at 15, 40 min, 3, 8, and 18 h and repeated our measurements. At 15 min after death, the first 0.5mm penetration depth caused the fastest relaxation at 55 s. Force decay decreased dramatically at 15 min after death as the penetration depth increased from 0.5 to 4mm. We used the force data sampled at 60s after insertion to approximate the elasticity. We observed the relations between the force versus the insertion depth. The force increased by a factor of 5 for the mitral valve annulus and 8 for the left free wall from 15 min to 18 h. We derived coefficients of a second-order polynomial equation relating the force data to insertion depth with R(2)>0.99.

Finite-element analysis of hepatic multiple probe radio-frequency ablation.

Radio-frequency (RF) ablation is an important means of treatment of nonresectable primary and metastatic liver tumors. RF ablation, unlike cryoablation (a method of tumor destruction that utilizes cold rather than heat), must be performed with a single probe placed serially. The ablation of any but the smallest tumor requires the use of multiple overlapping treatment zones. We evaluated the performance of a configuration incorporating two hooked probes (RITA model 30). The probes were lined up along the same axis in parallel 20 mm apart. Three different modes applied voltage to the probes. The first mode applied energy in monopolar mode (current flows from both probes to a dispersive electrode). The second mode applied the energy to the probes in bipolar mode (current flows from one probe to the other). The third method applied the energy sequentially in monopolar mode (in 2-s intervals switched between the probes). We used the finite-element method (FEM) and analyzed the electric potential profile and the temperature distribution at the end of simulation of a 12-min ablation. The alternating monopolar mode allowed precise independent control of the amount of energy deposited at each probe. The bipolar mode created the highest temperature in the area between the probes in the configuration we examined. The monopolar mode showed the worst performance since the two probes in close vicinity create a disadvantageous electric field configuration. We, thus, conclude that alternating monopolar RF ablation is superior to the other two methods.

In-vivo measurement of swine myocardial resistivity.

We used a four-terminal plunge probe to measure myocardial resistivity in two directions at three sites from the epicardial surface of eight open-chest pigs in-vivo at eight frequencies ranging from 1 Hz to 1 MHz. We calibrated the plunge probe to minimize the error due to stray capacitance between the measured subject and ground. We calibrated the probe in saline solutions contained in a metal cup situated near the heart that had an electrical connection to the pig's heart. The mean of the measured myocardial resistivity was 319 ohm x cm at 1 Hz down to 166 ohm x cm at 1 MHz. Statistical analysis showed the measured myocardial resistivity of two out of eight pigs was significantly different from that of other pigs. The myocardial resistivity measured with the resistivity probe oriented along and across the epicardial fiber direction was significantly different at only one out of the eight frequencies. There was no significant difference in the myocardial resistivity measured at different sites.

Error analysis of tissue resistivity measurement.

We identified the error sources in a system for measuring tissue resistivity at eight frequencies from 1 Hz to 1 MHz using the four-terminal method. We expressed the measured resistivity with an analytical formula containing all error terms. We conducted practical error measurements with in-vivo and bench-top experiments. We averaged errors at all frequencies for all measurements. The standard deviations of error of the quantization error of the 8-bit digital oscilloscope with voltage averaging, the nonideality of the circuit, the in-vivo motion artifact and electrical interference combined to yield an error of +/- 1.19%. The dimension error in measuring the syringe tube for measuring the reference saline resistivity added +/- 1.32% error. The estimation of the working probe constant by interpolating a set of probe constants measured in reference saline solutions added +/- 0.48% error. The difference in the current magnitudes used during the probe calibration and that during the tissue resistivity measurement caused +/- 0.14% error. Variation of the electrode spacing, alignment, and electrode surface property due to the insertion of electrodes into the tissue caused +/- 0.61% error. We combined the above errors to yield an overall standard deviation error of the measured tissue resistivity of +/- 1.96%.

Using electrical impedance to predict catheter-endocardial contact during RF cardiac ablation.

During radio-frequency (RF) cardiac catheter ablation, there is little information to estimate the contact between the catheter tip electrode and endocardium because only the metal electrode shows up under fluoroscopy. We present a method that utilizes the electrical impedance between the catheter electrode and the dispersive electrode to predict the catheter tip electrode insertion depth into the endocardium. Since the resistivity of blood differs from the resistivity of the endocardium, the impedance increases as the catheter tip lodges deeper in the endocardium. In vitro measurements yielded the impedance-depth relations at 1, 10, 100, and 500 kHz. We predict the depth by spline curve interpolation using the obtained calibration curve. This impedance method gives reasonably accurate predicted depth. We also evaluated alternative methods, such as impedance difference and impedance ratio.

Modeling bipolar phase-shifted multielectrode catheter ablation.

Atrial fibrillation (AFIB) is a common clinical problem affecting approximately 0.5-1% of the United States population. Radio-frequency (RF) multielectrode catheter (MEC) ablation has successes in curing AFIB. We utilized finite-element method analysis to determine the myocardial temperature distribution after 30 s, 80 degrees C temperature-controlled unipolar ablation using three 7F 12.5-mm electrodes with 2-mm interelectrode spacing MEC. Numerical results demonstrated that cold spots occurred at the edges of the middle electrode and hot spots at the side electrodes. We introduced the bipolar phase-shifted technique for RF energy delivery of MEC ablation. We determined the optimal phase-shift (phi) between the two sinusoidal voltage sources of a simplified two-dimensional finite-element model. At the optimal phi, we can achieve a temperature distribution that minimizes the difference between temperatures at electrode edges. We also studied the effects of myocardial electric conductivity (sigma), thermal conductivity (k), and the electrode spacing on the optimal phi. When we varied sigma and kappa from 50% to 150%, optimal phi ranged from 29.5 degrees to 23.5 degrees, and in the vicinity of 26.5 degrees, respectively. The optimal phi for 3-mm spacing MEC was 30.5 degrees. We show the design of a simplified bipolar phase-shifted MEC ablation system.

Three-Dimensional finite-element analyses for radio-frequency hepatic tumor ablation.

Radio-frequency (RF) hepatic ablation, offers an alternative method for the treatment of hepatic malignancies. We employed finite-element method (FEM) analysis to determine tissue temperature distribution during RF hepatic ablation. We constructed three-dimensional (3-D) thermal-electrical FEM models consisting of a four-tine RF probe, hepatic tissue, and a large blood vessel (10-mm diameter) located at different locations. We simulated our FEM analyses under temperature-controlled (90 degrees C) 8-min ablation. We also present a preliminary result from a simplified two-dimensional (2-D) FEM model that includes a bifurcated blood vessel. Lesion shapes created by the four-tine RF probe were mushroom-like, and were limited by the blood vessel. When the distance of the blood vessel was 5 mm from the nearest distal electrode 1) in the 3-D model, the maximum tissue temperature (hot spot) appeared next to electrodes A. The location of the hot spot was adjacent to another electrode 2) on the opposite side when the blood vessel was 1 mm from electrode A. The temperature distribution in the 2-D model was highly nonuniform due to the presence of the bifurcated blood vessel. Underdosed areas might be present next to the blood vessel from which the tumor can regenerate.