Quantification of liver perfusion by dynamic magnetic resonance imaging: Experimental evaluation and clinical pilot study

Abstract

Changes in liver microcirculation are considered essential in assessing ischemia-reperfusion injury, which in turn has an impact on liver graft function and outcome following liver transplantation (LTx). The aim of this study was to introduce dynamic magnetic resonance imaging (dMRI) as a new technique for overall quantification of hepatic microcirculation and compare it to perfusion measured by laser Doppler flowmetry (LDF; hepatic artery/portal vein) and thermal diffusion (TD). The study included 3 groups, measuring hepatic blood flow and microcirculation with the help of TD, LDF, and dMRI. In group I (9 landrace pigs; 26 ± 5 kg), the native liver before and after partial portal occlusion was studied; in group II (6 landrace pigs; 25.5 ± 4.4 kg), the liver 24 hours after LTx was studied; and in group III (14 patients), the liver on days 4 to 7 following LTx was studied. A close correlation was found between dMRI measurements and TD (r = 0.7–0.9, P < 0.01) in 4 defined regions of interest. Portal blood flow and partial occlusion of the portal vein were accurately detected by LDF flowmetry and correlated well with dMRI (r = 0.95, P < 0.01). In the clinical setting, representative TD measurements in segment 4b of the transplanted liver correlated well with dMRI analysis in other segments. Quantification of the portal blood flow and imaging of the whole liver could be performed simultaneously by dMRI. In conclusion, dMRI has been proved to be a sensitive modality for the quantification of liver microcirculation and hepatic blood flow in experimental and clinical LTx. It allows for a synchronous, noninvasive assessment of macrocirculation and microcirculation of the liver and could become a valuable diagnostic tool in advanced liver surgery and transplantation. Liver Transpl 15:693–700, 2009. © 2009 AASLD.

Liver microcirculation has proved to be a valid parameter for the quantification of ischemia-reperfusion injury and allows the diagnosis of early graft dysfunction or rejection.1–4 Therefore, any tool that allows the measurement of microcirculation can help in diagnosing posttransplant problems. In clinical liver transplantation (LTx), few methods are available for direct quantification of tissue perfusion.5 During an operation, laser Doppler flowmetry (LDF)6 is available for direct quantification of liver blood flow in the liver hilum. Postoperatively, color-coded duplex sonography is routinely used to assess hepatic arterial and portal venous blood flow. An analysis of liver microcirculation is dependent on indirect parameters, including liver function tests and markers of hepatocellular integrity or sinusoidal cell injury.7 In our center, thermal diffusion (TD) has been experimentally evaluated2 and has been introduced into the clinical routine for minimally invasive monitoring of liver microcirculation early after LTx.2, 3 A noninvasive technique for the perfusion measurement of the liver parenchyma is not yet available for clinical purposes.

Contrast-enhanced dynamic magnetic resonance imaging (dMRI) is a new method measuring perfusion and has the potential advantage of combining perfusion analysis and imaging. Rapid-sequence scanning can be used to determine blood flow noninvasively following a constant intravenous injection of a nondiffusable contrast material. The principles of the indicator dilution technique in an open 2-compartment model can be applied to convert signal-time courses into concentration-time courses. This can be analyzed within the framework of pharmacokinetic modeling and can be expressed as perfusion parameters. This was first done in mammography scans and imaging of the central nervous system8, 9 and has now been modified for application in liver surgery. It was the aim of this study to evaluate dMRI for the quantification of liver perfusion in a porcine model (validation via Doppler flowmetry and TD) and to start the first clinical application.

A, amplitude; CM, contrast medium; dMRI, dynamic magnetic resonance imaging; dMRT, dynamic magnetic resonance tomography; Gd-DPTA, gadolinium diethylenetriaminepentaacetic acid; k_p, perfusion rate; LDF, laser Doppler flowmetry; LTx, liver transplantation; MRI, magnetic resonance imaging; PPO, partial portal occlusion; PVF, portal venous flow; ROI, region of interest; SD, standard deviation; SRTF, saturation recovery turbo FLASH; TD, thermal diffusion.

PATIENTS AND METHODS

All experiments were approved by the Committee on Animal Care of the Regierungspräsidium of Karlsruhe, Germany and performed in accordance with German legislation on the protection of animals.

Experimental Protocol

Perfusion Measurements of Native Pig Livers Before and After Partial Portal Occlusion (PPO)

A median laparotomy and preparation of the hilar region of the liver were performed under general anesthesia (9 landrace pigs; 26.5 ± 4 kg). A recovery period of 30 minutes was allowed. A TD probe was inserted at a reference point centered in the left medial liver lobe, and LDF probes were positioned around the hepatic artery and portal vein. Continuous monitoring was performed, and the assessment of liver perfusion was done before and after dMRI. In the first phase, baseline measurements of all parameters were obtained. In the second phase, partial occlusion of the portal vein (60% of portal blood flow) was performed in order to imitate microvascular dysfunction with a reduced parenchymal blood flow. This was achieved with an inflatable catheter positioned around the portal vein and controlled by continuous monitoring of the portal venous blood flow. Four standardized regions of interest (ROIs) were chosen for magnetic resonance imaging (MRI) perfusion analysis: ventral, dorsal, and medial ones at the level of the portal vein and a global one over the whole liver surface. Additionally, the portal venous blood flow was calculated with dMRI in a separate ROI and compared directly to LDF (Fig. 1).

Figure 1. Color-coded image analysis of homogeneous liver perfusion (A) before and (B) after partial portal occlusion.

Perfusion Measurements of Transplanted Pig Livers

Standardized orthotopic LTx was performed in 6 pigs (landrace pigs; 25.5 ± 4.4 kg). With the experience of >300 porcine LTx procedures, anesthesia induction, monitoring, and the technique of hepatectomy and transplantation are well standardized in our laboratory and have been described before.10–12 All animals received their grafts after a cold ischemia time of 6 hours so that the ischemia-reperfusion injury was comparable within the group. Twenty-four hours following LTx, dMRI was conducted under general anesthesia in cardiorespiratory-stable animals. Laparotomy was performed, and TD and LDF probes were positioned. Continuous monitoring was applied as described later. As for the analysis of the perfusion pattern, early after transplantation, 1 ROI was defined in the center of each liver lobe (right lateral, right medial, left medial, and left lateral lobes), and TD was assessed at analogous points before and after dMRI examination. Additionally, a global ROI was assessed and compared to the mean of all TD measurements.

At the end of the experiments, all animals were sacrificed by an intravenous injection of potassium chloride.

Perfusion Measurements of Transplanted Human Livers (Clinical Pilot Study)

The study was approved by the Ethics Committee of the Medical Faculty of Heidelberg and complied with the Declaration of Helsinki. Written patient consent was obtained prior to probe implantation and dMRI examination. Standardized conventional LTx was performed in 14 patients (10 male and 4 female, 43 ± 12 years). Underlying diseases were chronic hepatitis B virus/hepatitis C virus hepatitis (n = 4), alcoholic cirrhosis (n = 3), fulminant hepatic failure (n = 3), primary biliary cirrhosis (n = 1), Wilson's disease (n = 1), amyloidosis (n = 1), and idiopathic liver cirrhosis (n = 1). TD-probe placement was done at the end of transplantation in segment 4b of the liver as previously described.1–4 Microcirculation TD was recorded twice daily for 1 week after transplantation. Between days 4 and 7, dMRI was performed in all 14 patients, and in analogy with the experimental protocol, TD measurements were taken before and after dMRI. Quantification with dMRI was done in each liver segment and correlated with the TD measurement.

Quantification of Arterial and Portal Venous Liver Blood Flow and TD Microperfusion

Liver blood flow was quantified by means of LDF probes (Transonic System, Inc., Ithaca, NY) positioned around the hepatic artery and portal vein. The vessel diameter determined the probe size (hepatic artery, 6 mm; portal vein, 12 mm). The blood flow was recorded in milliliters per minute and expressed as the mean ± standard deviation of recordings over 5 minutes. The TD technique was performed according to the technique established in our institution.1–4 Briefly, the TD probe (Thermal Technologies, Inc., Cambridge, MA) has a self-heated thermal transducer at its tip (diameter, 0.9 mm), which is inserted into the tissue and heated to a fixed temperature increment of 2°C above the surrounding tissue. The heating energy required to maintain the temperature elevation is related to heat transfer caused by tissue perfusion and thermal conduction. These data are recorded continuously and stored on a personal computer and can be followed online and used for detailed analysis offline.12 Microperfusion is given as the mean value over a 5-minute sampling period [frequency of continuous measurements, 1 Hz (mL/100 g/minute)].

dMRI (for Microperfusion Assessment)

MRI was conducted on a 1.0-T superconduction unit (Magnetom Impact Expert, Siemens AG, Erlangen, Germany). The animals were fixed in a supine position, and a phased array body coil was centered over the liver. The dynamic measurements were performed with a saturation recovery turbo FLASH (SRTF) sequence (recovery time = 125 ms, repetition time = 12.8 ms, echo time = 6 ms, field of view = 300 × 300 mm², matrix = 128 × 256, number of excitations = 1, slice thickness = 7 mm, image frame rate = 2 seconds). An intravenous contrast infusion [0.2 mmol/kg gadolinium diethylenetriaminepentaacetic acid (Gd-DTPA); Magnevist, Schering AG, Berlin, Germany] over 30 seconds was applied with a magnetic resonance–compatible injector (Spectris magnetic resonance injection system, SOM200E, Medrad, Inc., Pittsburgh, PA). A total of 120 images were acquired over a 4-minute period.

Perfusion Analysis and Pharmacokinetic Modeling

Automated perfusion analysis in the liver is difficult because of the dual blood supply through the aorta and portal vein.13 After injection, the arrival of contrast material in the portal vein is significantly delayed in comparison with the aorta. Therefore, a dynamic contrast-enhanced MRI technique originally developed for brain examinations was optimized to ensure an adequate analysis of the magnetic resonance signal enhancement in the liver. The fast kinetics of the tissue response were resolved with a strongly T1-weighted SRTF sequence, which made it possible to acquire an image series with a high sampling rate. After the application of a contrast material (Gd-DTPA) was continued over a 30-second interval instead of a bolus injection, the measured signal-time courses could be converted into concentration-time courses and thus be analyzed within the framework of pharmacokinetic modeling in a 2-compartment model (Fig. 2). There are 2 important parameters: the amplitude (A) is connected to the maximal signal enhancement, whereas the exchange parameter (k_p; ie, the perfusion rate) describes the velocity of the signal increase and is correlated with tissue perfusion (Fig. 3). On the basis of a series of SRTF images, the aforementioned parameters were fitted iteratively pixel by pixel with the Marquart algorithm for least-squares estimation of nonlinear parameters in each ROI. We received 2 computed parameter images (A and k_p) providing the functional tissue information as well as anatomic images. This information was encoded in a 2-dimensional color table6 and presented as color-coded parameter images containing both sets of information (Fig. 4A,B). Data were processed with a personal computer using an Interactive Data Language (Research Systems, Inc., Boulder, CO)–based custom-built program.

Figure 2. Pharmacokinetic model. The contrast medium (CM) is injected within 30 seconds into the intravascular compartment of the body [ie, the infusion rate (K_inf)]. The elimination and distribution of the CM in this compartment are described by the elimination factor (k_el). After a delay (t_lag), CM appears in the intravascular space (V_IV) of the liver. The regional blood flow of the liver (F) is dependent on the blood distribution volume of the whole body (V_B) and the liver (V_IV) and the concentrations of the CM in these compartments (C₁ and C₂).

Figure 3. Signal-time curve for the calculation of microcirculation via dynamic magnetic resonance tomography. k_p is the perfusion rate.

Figure 4. (A) Legend of colors describing perfusion. The perfusion equivalent (k_p) describes how much contrast medium reaches a certain spot, whereas the amplitude (A) describes how fast the contrast medium reaches a certain spot. (B) Color-coded image analysis of homogeneous liver perfusion before portal occlusion.

Data Analysis

The portal blood flow was calculated with phase contrast angiography (L/minute).14 As for the data analysis of liver microcirculation by dMRI, 4 standardized ROIs were defined: ventral, dorsal, medial, and global. Contrast enhancement in each ROI was analyzed and described as the mean ± standard deviation of k_p (minute⁻¹). Statistical analysis was performed with the Pearson correlation, paired T test, Spearman correlation, and paired Wilcoxon test. P values < 0.05 were considered statistically significant. In order to compare 2 techniques measuring the same thing (eg, liver microcirculation), a statistical method for assessing agreement between 2 methods of clinical measurement was suggested by Bland and Altman.15 This is a plot of the differences between the 2 methods against their means. The so-called limits of agreement are a range of the mean ± 2 standard deviations. Agreement between methods is considered good if 95% of the measurements are positioned within these limits.

RESULTS

Perfusion Measurements of the Native Pig Liver Before and After PPO

Before and after PPO, the heart rate (before PPO, 103 ± 22 minute⁻¹; after PPO, 106 ± 26 minute⁻¹), central venous blood pressure (before PPO, 7.6 ± 0.9 mm Hg; after PPO, 7.7 ± 1 mm Hg), and oxygenation parameters remained stable. The mean arterial blood pressure decreased from 67 ± 12 to 52 ± 14 mm Hg (P = 0.011). PPO reduced portal blood flow from 817 ± 346 to 279 ± 167 mL/minute (measured by LDF; P < 0.05) and from 801 ± 279 to 407 ± 206 mL/minute (measured by dMRI; P < 0.05). The correlation between dMRI and LDF reached r = 0.95 (P < 0.01; Fig. 5). TD recorded a significant decrease in liver microcirculation from 77.8 ± 6 mL/100 g/minute before PPO to 47.9 ± 19.5 mL/100 g/minute after PPO (P < 0.05). Microperfusion measurements by dMRI before and after PPO showed a significant reduction in all ROIs (Table 1). Moreover, a close correlation between TD and dMRI could be shown, and this was also confirmed by an agreement of 100% (Fig. 6).

Figure 5. Correlation of the portal venous flow (PVF) measured by dynamic magnetic resonance tomography (dMRT) and laser Doppler flowmetry (LDF) in the native pig liver (r = 0.95, P < 0.01).

Table 1. k_p Values of the 4 ROIs Before and After PPO

	TD (mL/100 g/minute)	kp (minute^−1)
		Ventral ROI	Medial ROI	Dorsal ROI	Global ROI
Abbreviations:kp, perfusion rate; PPO, partial portal occlusion; r, Spearman correlation; ROI, region of interest; TD, thermal diffusion. * P < 0.05 (before vs. after PPO). † P < 0.01.
Before PPO	77.8 ± 6.0	13.8 ± 7.9	18.9 ± 20.6	19.7 ± 24.4	22.7 ± 22.0
After PPO	47.9 ± 19.5*	2.9 ± 1.6*	3.2 ± 1.2*	4.0 ± 0.96*	3.4 ± 1.3*
r for TD versus kp		0.89†	0.90†	0.77†	0.70†

Figure 6. Native pig liver experiment: the amount of agreement between the thermal diffusion (TD) and perfusion rate (k_p). The mean (M) values ± 2 standard deviations (SDs) of TD-k_p are shown (limits of agreement). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Perfusion Measurements Following LTx in the Pig

Monitoring showed a stable physical condition with respect to the heart rate, mean arterial blood pressure, central venous pressure, blood gases, and body temperature throughout the transplantation and postoperative course. Arterial blood flow was reduced 1 hour versus 24 hours after reperfusion (1 hour versus 24 hours post-LTx, 346 ± 112 to 150 ± 119 mL/minute; P < 0.05). Furthermore, in 2 of 6 transplanted livers, arterial thromboses occurred, and those animals were excluded from the evaluation. Portal venous flow increased from 755 ± 128 to 903 ± 255 mL/minute (1 hour versus 24 hours post-LTx; not significant), whereas liver microcirculation remained unchanged during the observation period (74.8 ± 9.8 versus 75.8 ± 13.1 mL/100 g/minute; not significant). Twenty-four hours after reperfusion, TD and dMRI showed homogeneous liver microcirculation. Similar to measurements in the native liver, the correlation between both measurements was significant (Table 2). Consequently, agreement between the 2 methods was high as all values were found within 2 standard deviations (Fig. 7).

Table 2. TD-Probe k_p Measurements by dMRI in the 4 Lobes of the Pig Liver 24 Hours After Liver Transplantation

	ROI				Average Overall
	Right Posterior Lobe	Right Medial Lobe	Left Medial Lobe	Left Lateral Lobe
NOTE: A correlation was made to kp measured by dynamic magnetic resonance tomography in the respective regions of the liver. Abbreviations: dMRI, dynamic magnetic resonance imaging; kp, perfusion rate; r, Spearman correlation; ROI, region of interest; TD, thermal diffusion. * P < 0.01.
TD (mL/100 g/minute)	77.3 ± 7.2	77.4 ± 11.8	66.3 ± 19	82.1 ± 14.3	75.8 ± 13.1
kp (minute^−1)	10.2 ± 6.3	10.8 ± 6.5	10 ± 6.4	10.8 ± 7.4	10.6 ± 6.3
r	0.66*	0.67*	0.89*	0.75*	0.84*

Figure 7. Transplanted pig liver 24 hours after transplantation: the amount of agreement between the thermal diffusion (TD) and perfusion rate (k_p). The mean (M) values ± 2 standard deviations (SDs) of TD-k_p are shown (limits of agreement). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]

Liver Perfusion Measurements Following Human Orthotopic LTx (Clinical Pilot Study)

Fourteen LTx patients were included in this pilot study. Liver perfusion was measured between the fourth and seventh postoperative days. At the time of the measurements, all patients showed a stable physical condition. Similarly to the animal model, a significant correlation was found between a representative TD measurement in segment 4b of the liver and the results of dMRI analysis in most liver segments. Although there was a good correlation in segments 3 to 8, dMRI analysis was inconsistent in 3 segments of the liver (segments 1, 2, and 4b). Correlation was best in the cranial (segments 4a, 7, and 8) and right liver segments (segments 5 and 6; Table 3). Therefore, agreement between the TD and perfusion rate were high, and all values were found within 2 standard deviations. The correlation of the perfusion equivalent (k_p) over all liver segments per patient and TD in segment 4b was significant (r = 0.71, P < 0.05; Fig. 8).

Table 3. k_p Measurements by dMRI (Mean ± Standard Deviation) in Designated Liver Segments (Couinaud Classification of Segments 1–8) 4 to 7 Days After Liver Transplantation and Pearson Correlation with TD Measured in Segment 4b of the Transplanted Liver

		s1	s2	s3	s4a	s4b	s5	s6	s7	s8	Total
Abbreviations: dMRI, dynamic magnetic resonance imaging; kp, perfusion rate; r, Pearson correlation with thermal diffusion; s, perfusion equivalent of each liver segment (minute^−1); TD, thermal diffusion (mL/100 g/minute). * P < 0.05. † P < 0.01.
dMRI	81 ± 18	60 ± 12	49 ± 22	52 ± 16	50 ± 20	54 ± 15	52 ± 16	50 ± 16	53 ± 20	51 ± 19	52 ± 16
r		0.54	0.64	0.65*	0.8†	0.57*	0.71†	0.73†	0.9†	0.8†	0.7†

Figure 8. Correlation graph of the perfusion equivalent [k_p; average (mean) values over all liver segments per patient] and thermal diffusion (TD) measurements in segment 4b for each patient following liver transplantation.

DISCUSSION

The assessment of functional hemodynamic parameters on the basis of rapid-sequence tomography techniques (computed tomography as well as MRI) is increasing.16–19 The indicator dilution technique was first used by Axel20 for the quantification of cerebral blood flow. Microcirculatory changes via MRI can be sensitively detected in the brain,19, 21 the female breast,10, 22 and the lung.20 Recently, dMRI has been used to assess liver macrocirculation and microcirculation.22–25 Because of the small size, variable anatomy, and location of the hepatic artery, the measurement of liver macrocirculation is limited to hepatic portal flow. Nevertheless, we could show here that the correlation of portal flow between LDF and dMRI is high. Furthermore, up to now, dMRI of the upper abdomen has been impeded by the fact that microvascular characteristics require acquisitions longer than 1 breath-hold.22, 24 Concerning the evaluation of liver perfusion, contrast-enhanced dMRI shows a close correlation with validated methods in the assessment of liver microcirculation with TD and portal blood flow with LDF. Changes in liver perfusion can be detected accurately in the native liver and following LTx in a porcine model.

For the assessment of microcirculation, for TD, a borderline value (60 mL/100 g/minute) has been established in porcine experiments and clinically.3 We could show that dMRI and TD in the native liver, after PPO, and in the transplanted liver showed similar perfusion. Surprisingly, in the livers with hepatic arterial occlusion due to thrombosis, the perfusion correlation was comparable, whereas absolute values were lower because of the delayed inflow via the portal system. This implies that dMRI largely depends on the arterial curve, and complications here might impede parenchymatous perfusion. On the other hand, should a delay be found in parenchymatous perfusion, an arterial problem can be assumed in reverse. In previous experiments, we could clearly show a perfusion gradient in the pig liver from cranial, higher perfused areas to caudal, lower perfused segments.12 Moreover, fatty degeneration or cirrhotic changes are known for their heterogeneous perfusion,26, 27 and respiratory movements will interfere with segmental allocation.22, 24 We took these possibilities into account by analyzing various ROIs within the liver.

Arterial inflow to the liver is essential for the evaluation of the dMRI color picture. Arterial complications such as thrombosis and stenosis are frequently found following LTx.11 It has been shown in our study that the portal venous contrast inflow will result in an adequate microcirculatory distribution, whereas values will be remarkably lower because of a delayed contrast medium increase, as if arterial perfusion were not impeded. Therefore, arterial complications have to be verified (eg, magnetic resonance angiography), and the interpretation of microcirculation has to be adapted accordingly. Furthermore, large vessels may lead to false high values and have to be excluded when the various ROIs in the liver are being assessed. Limits of the method are that the k_p value allows only for an estimation of microcirculatory flow. Although changes in parenchymatous perfusion can be sensitively detected, the absolute values in microcirculation were found to be higher by a factor of 4 to 5, as we could show with TD. This multiplication factor, required for the real microcirculation flow, depends on specific properties of the tissue and is not yet specified. Nevertheless, the excellent correlation of TD and dMRI suggests that in the future, with technical advances taken into account, absolute values of microperfusion in the liver will be able to be determined.

In our clinical pilot study on 14 patients, when dMRI was compared to the TD measurements in segment 4b, only the cranial and right segments showed a significant correlation. We assign the dissatisfactory results in segments 1 and 2 to liver anatomy and increased movement of the liver caused by the heart and diaphragm as it did not occur to that extent in the intubated pigs (breath-hold technique). Measurement in segment 4b was most likely impeded by the TD probe itself. Gd-DTPA–enhanced dMRI is an innovative method for quantifying blood flow and tissue perfusion of the liver. In this study, we evaluated the ability of this method by comparing it to TD and LDF in experimental and clinical settings. We found a good correlation between dMRI and TD with respect to microperfusion and also a good correlation between dMRI and LDF with respect to macroperfusion (portal flow). With dMRI, global and any regional quantification of liver perfusion is possible. In order to mimic microvascular dysfunction, we performed PPO experiments with native livers and showed that a reduction of microperfusion can be detected sensitively and accurately by dMRI. Therefore, we can potentially differentiate between mild and severe ischemia-reperfusion injury on a microcirculatory level without direct functional information about the liver graft. As this is done noninvasively, without exposure to X-rays and with the application of contrast media known for their low toxicity,12, 28 the assessment of liver perfusion remains available at any time after transplantation. Additionally, imaging can be obtained and directly correlated with the results of functional analysis. The combination of conventional imaging and assessment of hemodynamic data using dMRI is advantageous, especially for the diagnosis of complications after liver surgery and transplantation.25, 29

In conclusion, we consider dMRI to be a promising new technique for quantification and qualification of hepatic microcirculation. The distribution of microperfusion can be demonstrated, and variable adaptation with respect to the location of the perfusion level can be made. Furthermore, dMRI can reliably determine portal venous flow, and conventional MRI can be used to assess other abdominal organs as well. Advantages over TD are obvious. The 1-stop shopping approach will allow for conventional MRI, magnetic resonance cholangiopancreatography, and magnetic resonance angiography as well as the assessment of the macroperfusion and microcirculation of parenchymatous organs, including the liver. The disadvantage of dMRI is that it is a time-consuming, technically complex, and expensive procedure. At this time, we do not recommend dMRI as a routine instrument for the follow-up of LTx patients, but it could be helpful in patients with a need for detailed analysis regarding the macroperfusion and microperfusion and for graft imaging, including bile ducts as well as the remaining abdomen. It can be assumed, however, that these disadvantages can be overcome in due time, and the important characteristic of noninvasiveness will make it an excellent tool for future diagnostics in liver surgery and LTx.