PREOPERATIVE COMPUTED TOMOGRAPHY VOLUMETRY AND GRAFT WEIGHT ESTIMATION IN ADULT LIVING DONOR LIVER TRANSPLANTATION

ABSTRACT Background: Computed tomography volumetry (CTV) is a useful tool for predicting graft weights (GW) for living donor liver transplantation (LDLT). Few studies have examined the correlation between CTV and GW in normal liver parenchyma. Aim: To analyze the correlation between CTV and GW in an adult LDLT population and provide a systematic review of the existing mathematical models to calculate partial liver graft weight. Methods: Between January 2009 and January 2013, 28 consecutive donors undergoing right hepatectomy for LDLT were retrospectively reviewed. All grafts were perfused with HTK solution. Estimated graft volume was estimated by CTV and these values were compared to the actual graft weight, which was measured after liver harvesting and perfusion. Results: Median actual GW was 782.5 g, averaged 791.43±136 g and ranged from 520-1185 g. Median estimated graft volume was 927.5 ml, averaged 944.86±200.74 ml and ranged from 600-1477 ml. Linear regression of estimated graft volume and actual GW was significantly linear (GW=0.82 estimated graft volume, r2=0.98, slope=0.47, standard deviation of 0.024 and p<0.0001). Spearman Linear correlation was 0.65 with 95% CI of 0.45 - 0.99 (p<0.0001). Conclusion: The one-to-one rule did not applied in patients with normal liver parenchyma. A better estimation of graft weight could be reached by multiplying estimated graft volume by 0.82.


INTRODUCTION
T he first living donor liver transplantation (LDLT) was performed in 1989 at our Institution 15 . Since then, surgical techniques and patient´s care improvements have established LDLT as a valid option of overcoming the shortage of deceased donors.
Initially, LDLT was exclusive to pediatric patients, although over time it also became good option to adult recipients as well. One of the major concerns of this procedure is to provide an adequate graft volume while preserving a suitable hepatic remnant for the donor. Reduced grafts are associated with small for size syndrome and poor outcomes. In this particular setting, preoperative radiological evaluation of liver volume is the standard method for donor evaluation, aiming minimization of unnecessary risks.
Therefore, accurate assessment of the volume of the liver and its lobes prior to surgery is mandatory. Computed tomography volumetry (CTV) is a useful tool for predicting graft weights (GW) for LDLT 16,18 . However, for the conversion of volumetric weight, the rule of "one-to-one" does not offer good results due that this rule was based in the weight using cirrhotic livers 20 . Fw studies have examined the correlation between CTV and GW in normal liver parenchyma.
The aim of this paper was to analyze the correlation between CTV and GW in an adult LDLT population, and also provide a systematic review of the existing mathematical models to calculate partial liver graft weight or volume, based on different radiological parameters.

METHODS
Between January 2009 and January 2013, 28 consecutive donors undergoing right hepatectomy for LDLT were retrospectively reviewed. All donors were healthy adults with preoperative CTV.
Hepatectomy cutting plane was guided by ischemic line after a short period (about 2 min) of right pedicle clamping (i.e. right hepatic artery and right portal vein). Right hepatic graft, including middle hepatic vein, was used in all but two cases. Grafts were perfused via the right portal vein with cold histidine-tryptophan-ketoglutarate (HTK) solution as soon as hepatectomy was completed. The solution was drained from the liver and a pre-calibrated digital scale was used to define the actual GW.

Preoperative measurement of the liver volume
Multidetector computed tomography (CT) images were obtained with four different devices: CT Discovery 750 HD (GE Medical Systems, Milwaukee, Wisconsin, USA), CT Light Speed (GE Medical Systems, Milwaukee, Wisconsin, USA), CT IDT MX8000 (Philips Healthcare, Cleveland, Ohio, USA) and CT Brilhance 64 (Philips Healthcare, Cleveland, Ohio, USA), with a 1-mm slice thickness and 70 s after the intravenous administration of iodine contrast media (portal venous phase). These data were used for CT volumetry measurements. The scanning parameters were as follows: 120 kV and mAs appropriate to body habitus.
The right lobe graft volume (GV) was measured by tracing a line on the right of the middle hepatic vein, thus defining the virtual hepatectomy plan. The perimeters of the liver and the graft were outlined by hand tracing on each slice by an abdominal radiologist. The enclosed area was calculated with image analysis software Volume Viewer (GE Medical Systems, Milwaukee, Wisconsin, USA). The liver volume (in milliliters) was then obtained as the sum of all areas from the intervals of the serial CT slices (Figure 1). Estimated graft volume (EGV) was obtained by CTV, and these values were compared to the actual GW (AGW), which was defined after liver harvesting and perfusion.

Statistical analysis
Values are shown as mean±standard deviation. Statistical analysis was performed using R version 2.15.2 (The R Foudantion for Statistical Computing). A value of p<0.05 was considered significant. Relation between estimated GV and AGW was determined by use of linear regression analysis; furthermore, linear correlation of Spearman and its 95% confidence interval were presented.

RESULTS
The donor's demographics and perioperative parameters are summarized in Table 1. Blood products transfusion was not necessary in any of the donor's surgeries. Median graft versus recipient weight ratio was 1.12%, and ranged from 0.81 to 1.45%. Recipient´s estimated standard liver volume ranged from 948 to 1471g (median 1277g). The median relation between GW and estimated standard liver volume was 62.45 (44.23-80.57).

DISCUSSION
An accurate preoperative GV estimation is essential in selecting suitable donors for adequate recipients of LDTL. It can provide valuable information to keep a safe remnant liver volume to the donor as well as reducing the recipient´s postoperative morbidity and mortality.
LDTL recipients should receive sufficient hepatic parenchyma. Ideally relationship between weight of the graft and weight of the recipient should be about 0.8-1% to avoid small for size syndrome 1,10 . Several studies have shown that weight of the graft and weight of the recipient ratio below 0.8 are associated with higher initial postoperative mortality (3-months survival of 54%) 10 . Other factors, such as portal venous pressure over 15 mmHg 13 and excessive portal vein inflow (over than 250 ml/min/100g of liver weight) 8 are also associate with poor outcomes.
Therefore, predicting actual graft weight with preoperative radiological evaluation is essential to define if LDLT is feasible. CTV is widely used as the standard method for preoperative estimation of hepatic graft weight 9 . However, it has been shown that CTV tends to overestimate both graft volume and weight 12 . Another concern is about health liver density; usually hepatic parenchyma volumetric measurements are compared based on the assumption that density is 1.00 g/ml. However, the relation between graft volume and GW is not exact and the one-to-one rule should not be applied for all patients. Yoneyama et al. 22 have found correlation coefficients between estimated GV and GW in right lobe grafts of 0.84 and left lobe of 0.85 22 . They have emphasized that it could not be applicable in others institutions due to several biases.
Several factors can interfere in graft weight analysis, such as: 1) the hepatic transection line defined by ischemic area, and/or anatomical parameters; and 2) the timing to graft weight, since high osmotic preservation solutions can induce cellular dehydration causing significant weight reduction in liver grafts 5,7 . Recently, Satou et al. 17 have shown a reduction of up to 10% of the initial weight post-hepatectomy after back-table surgery has been completed 17 .
Accordingly, there are many strategies to estimate right graft weight and volume (Table 2). In our review, were found at least 11 different 2,3,4, 6,7,8,9,11,12,17,18,21,22 , and sometimes conflicting, mathematical formulas aiming to predict graft size. Some of these methods do not use the computed tomography volumetry, but body surface area, and the diameter of the portal vein and/or their branches. This wide variation could be explained by donor´s heterogeneity due to age, gender, regional demographics, inclusion or not of the medium hepatic vein in the graft, different types of preservation solution, among others. The main purpose of our study was to determine the correlation between EGV and AGW in our center. We found that 1.00 ml of preoperative CTV correlates with 0.82 g of graft weight. Then, we believe that use the coefficient of 0.82 can predict more accurately the right hepatic graft weight in our population.