Hepatic venous congestion (HVC) has not been assessed quantitatively prior to hepatectomy and its resolving mechanism has not been fully analyzed. We devised and verified a new method to predict HVC, in which HVC was estimated from delineation of middle hepatic vein (MHV) tributaries in computed tomography (CT) images. The predicted HVC was transferred to the right hepatic lobes of 20 living donors using a paper scale, and it was compared with the actual observed HVC that occurred after parenchymal transection and arterial clamping. The evolution of HVC from its emergence to resolution was followed up with CT. Volume proportions of the predicted and observed HVC were 31.7 ± 6.3% and 31.3 ± 9.4% of right lobe volume (RLV) (P = .74), respectively, which resulted in a prediction error of 3.8 ± 3.7% of RLV. We observed the changes in the HVC area of the right lobes both in donors without MHV trunk and in recipients with MHV reconstruction. After 7 days, the HVC of 33.5 ± 7.7% of RLV was changed to a computed tomography attenuation abnormality (CTAA) of 28.4 ± 5.3% of RLV in 12 donor remnant right lobes, and the HVC of 29.1 ± 11.5% of RLV was reduced to a CTAA of 9.3 ± 3.2% of RLV in 7 recipient right lobe grafts with MHV reconstruction. There was no parenchymal regeneration of the HVC area in donor remnant livers during first 7 days. In conclusion, we believe that this CT-based method for HVC prediction deserves to be applied as an inevitable part of preoperative donor evaluation. The changes in CTAA observed in the right lobes of donors and recipients indicate that MHV reconstruction can effectively decrease the HVC area. (Liver Transpl 2004;10:763–770.)
A right lobe without reconstruction of the middle hepatic vein (MHV) trunk has been a standard liver graft for adult-to-adult living donor liver transplantation (LDLT) in many transplantation centers. Hepatic venous congestion (HVC) caused by deprivation of MHV drainage has been tolerated in many right lobe recipients, but there has been much debate on the natural resolution of HVC as well as the need for MHV reconstruction.1–4 Recently, many surgeons have agreed on the need of MHV reconstruction because of occasional massive ischemic congestion, and and have become interested in criteria for MHV reconstruction.2, 5 To determine if MHV reconstruction is indicated, quantitative assessment of HVC in the right lobe becomes the most important step. However, this is feasible only after division of MHV tributaries during transsection of the parenchyma.
In our 398 cases of adult-to-adult LDLT prior to this study, we observed that HVC of a right lobe occurred according to the anatomical distribution of MHV tributaries.5 This implied that HVC might be predicted quantitatively by using a method based on hepatic vein anatomy. A pilot study for assessment of HVC using computed tomography (CT) convinced us that a new CT-based prediction method deserved a prospective trial.
In this study, we verified our CT-based method for preoperative quantitative assessment of HVC and analyzed the HVC-induced changes in the liver to clarify the inducing and resolving mechanisms of HVC.
Pilot Study to Assess HVC Using CT Image Analysis and Volumetry
Out of 398 adult LDLT cases, we selected 6 cases of right lobe graft recipients and 10 living donors of extended left lobe graft (left lobe with MHV trunk) showing extensive areas of attenuation abnormalities (computed tomography attenuation abnormality [CTAA]) of the liver parenchyma on postoperative CT. We extracted the outlines of CTAA in the graft or remnant livers in donors and placed them over the corresponding sections on preoperative donor CT images. This process revealed that the outlines of CTAA were roughly located midway between the minute tributaries of the MHV and the right hepatic vein (Fig. 1). We also compared the outlines of CTAA with intraoperative photographs of the discolored HVC after arterial clamping, and found a close similarity between them. We simulated the delineation of HVC with sample CT scans, on the assumption that both HVC and CTAA occur according to the hepatic venous anatomy. Based on this pilot study, we developed rules to trace a MHV territory on the venous phase CT images and summarized them in Figure 2.
From September 2002 to October 2002, we performed 15 LDLT procedures for 13 adults and 2 pediatric recipients. Underlying diseases of the recipients were hepatitis B-associated liver cirrhosis in 12, hepatitis C-associated liver cirrhosis in 1, and biliary atresia in 2. The number of living donors for these 15 recipients was 22, because there were 7 cases of dual LDLT.5
For this study, we carried out preoperative assessment of HVC in 7 right lobe donors, 12 extended left lobe donors, and 1 right posterior segment donor; 2 left lateral segment donors were excluded (n = 20). Surgical procedures for donors and recipients were performed as previously described.4, 5 All donors have recovered without any minor complication. All recipients except for 1 survive to date, after 15 months.
This study consisted of 2 parts: 1) a comparative study to evaluate the accuracy and feasibility of a new CT-based method for HVC prediction and 2) an analytic study to evaluate the postoperative changes of HVC in both graft and remnant right lobes.
CT scans were taken preoperatively, at 1 week and 2 months postoperatively in donors, and once a week in recipients, while in the hospital, according to our routine protocols for LDLT. All CT scans were carried out in triphasic mode (preenhanced, arterial, and venous phases) and output CT images were reconstructed as 5-mm thick sections. The CT images were stored in Picture Archiving and Communication System (PACS, Petavision®, Hyundai Information Technology, Seoul, Korea), enabling image processing and various measurements.
Mean values with standard deviation and range were used for numeric data. The significance of differences was assessed using paired t-test. The coefficient of determination (R2) was obtained from simple linear correlation. Differences at P < .05 were considered to be statistically significant.
CT-Based Quantitative Assessment of HVC and Mapping of the Predicted HVC Area Before Surgery
We delineated the margin of the HVC on preoperative donor CT scan and carried out CT volumetry according to the rules derived from the previous pilot study (Fig. 2). For comparison of the widths and volumes of the predicted and actual emerged HVC, we made an HVC map of a 2-dimensional coordinate system (Fig. 3). We placed the zero point of the horizontal axis at the interlobar margin of the liver surface. The zero point of the vertical axis was placed at the ventral end of the gallbladder bed. The width at the liver surface of CT images between the right margin of the HVC and the interlobar margin was measured from every 5-mm-thick CT section and mapped on the coordinates (the predicted HVC map). The volume of the predicted HVC was calculated using CT volumetry as in Figure 2.
Intraoperative Transfer of a Predicted HVC Map to the Donor Liver Surface
Immediately after cholecystectomy prior to hilar dissection, the lobar glissonian cord was clamped manually to induce hemihepatic discoloration. The interlobar margin was marked with electrocautery, and the width data of a predicted HVC map were transferred to the convex liver surface with a flexible paper scale (Fig. 3). The real transection planes of donor livers were 0–2 cm from the interlobar margin, but they were exactly matched to the predetermined CT transection planes in most donors. Within 1 hour after parenchymal transection, the right hepatic artery was temporarily clamped to make HVC apparent. A clamp duration of 1 or 2 minutes was enough to induce noticeable discoloration in most donors.
CT Volumetry for the Observed HVC
The width and shape of the discolored HVC were measured by using a paper scale and transferred to the preoperative donor CT scan to measure the actual amount of HVC (the observed HVC map).2
CT Follow-Up to Analyze the Sequences of HVC in Remnant and Graft Right Lobes
We analyzed the postoperative CTAA in 12 extended left lobe donors and 7 right lobe recipients to evaluate its relation to HVC and its resolution. CTAA, which must be a consequence of HVC, was arbitrarily classified into persistent hyperattenuation, delayed hyperattenuation, and persistent hypoattenuation according to time-sequence and intensity of attenuation (Fig. 4). In this study, we calculated only the volumes of hyperattenuation and hypoattenuation on the venous phase CT images for simplified quantitative analysis.
Quantitative Comparison of the Predicted and Observed HVC in 20 Living Donors
In a total of 19 donors discoloration of HVC became apparent within 1–2 minutes of clamping the right hepatic artery, but in 1 donor (case 20) only faint discoloration occurred in spite of prolonged clamping over 5 minutes. The predicted and observed HVC of the right lobes in 20 donors were collectively demonstrated in Figure 5 and Table 1. A right posterior segment donor (case 12) demonstrated counter-discoloration of HVC from right hepatic vein occlusion.
Table 1. Profiles of Preoperative Assessment, Intraoperative Comparison, and Postoperative Sequences of Hepatic Venous Congestion (HVC) in the Remnant and Graft Right Lobes
Donor Age (Years)
RL CT Volume (mL)
LL CT Volume (mL)
Graft Weight (gm)
Predicted HVC Proportion (% RLV)
Observed HVC Proportion (% RLV)
CTAA Proportion (RL %)
Hyperattenuation Proportion (% RLV)
Hypoattenuation Proportion (% RLV)
Hypoattenuation Proportion (% CTAA)
Abbreviations: RL, right lobe; LL, left lobe; CT, computed tomogram; RLV, right lobe volume; CTAA, CT attenuation abnormality; RPS, right posterior segment; % RLV, volume % to right lobe volume; % CTAA, volume % to total volume of CTAA.
The average widths of the predicted HVC at the upper and lower thirds of the transection plane and at the ventral end of the gallbladder bed were 5.7 ± 1.4 cm (range 4–10), 6.2 ± 1.4 cm (range 4–10), and 5.0 ± 2.4 cm (range 0–9), respectively. Those of observed HVC at the same locations were 5.9 ± 1.9 cm (range 3–11), 6.2 ± 1.9 cm (range 3–11), and 4.7 ± 2.7 cm (range 0–11), respectively.
The volume proportion of the predicted HVC in these 19 donor right lobes (except for a right posterior segment donor) was 31.7 ± 6.3% (range 21–45) of the right lobe volume (RLV) and that of observed HVC was 31.3 ± 9.4% (range 19–54) of RLV, resulting in no significant difference (P = 0.74) and high correlation (R2 = 0.708, P < .05). Mean error of the predicted HVC volume was calculated as 3.8 ± 3.7% (range 0–12) of RLV. In 14 of the 20 donors (70%), the error of predicted HVC was not more than 5% of RLV.
Evolution of HVC in Remnant Right Lobes of 12 Donors Having Undergone Extended Left Lobectomy
The wedge-shaped low-density portion in preenhanced 7-day CT appeared as slightly smaller than the sum of portions revealing hyperattenuation and hypoattenuation in the venous phase, although we could not definitely delineate the area of low density in half of the donors (Fig. 4).
The regeneration rate of 12 remnant right lobes during the first 7 days was 16.8 ± 7.9% (range 8–35) of RLV. The observed HVC proportion of 33.5 ± 7.7% (range 25–54) to RLV was changed to a CTAA proportion equivalent to 28.4 ± 5.3% (range 22–40) of RLV, which resulted in a significant difference (P = .01) and crude correlation (R2 = 0.45, P < .05) (Table 1). The proportions of hyperattenuation and hypoattenuation portions were 20.9 ± 5.1% (range 15–29) and 7.5 ± 2.8% (range 3–12) of RLV, respectively. The volumes of the observed HVC and 7-day CTAA were 294 ± 106 mL (range 189–569) and 287 ± 82 mL (range 157–492), respectively, resulting in no significant difference (P = .667). In Figure 6 the evolution of HVC is illustrated for donor cases 19 and 7.
In donor case 20, the total amount of CTAA was not changed as compared with the observed HVC volume, but there was an exceptionally small proportion of hypoattenuation (13% of total CTAA). The CTAA in these 12 extended left lobe donors had almost completely disappeared in 2-month CT scans.
Evolution of HVC in 7 Right Lobe Grafts Having Undergone MHV Reconstruction
For all 7 right lobe grafts, we reconstructed all MHV tributaries more than 5 mm in diameter (modified right lobe graft).4 Their regeneration rate during the first 7 days was 23.6 ± 14.3% (range 3–42) of preoperative RLV. The volume proportions of the observed HVC, total CTAA, hyperattenuation, and hypoattenuation were 29.1 ± 11.5% (range 20–53), 9.3 ± 3.2% (range 6–13), 6.0 ± 1.6% (range 4–9), and 3.3 ± 2.4% (range 0–7) of RLV, respectively, resulting in a marked decrease of total amount of CTAA comparing with the observed HVC volume (P = .008) (Fig. 6).
The volume proportion of hypoattenuation to total CTAA was 26.8 ± 8.8% (range 9–42) in the remnant right lobes of donors and 32.3 ± 17.8% (range 0–54) in the modified right lobe grafts (P = .373). The volume of total CTAA, as well as the proportion of hypoattenuation to total CTAA had decreased gradually on the subsequent weekly CT scans, and all CTAA finally disappeared within 1–2 months (Fig. 7).
Since encountering life-threatening HVC, assessment and management of HVC has become a hot debate in LDLT using a right lobe graft.4 This controversy about whether to reconstruct MHV tributaries or not must be based on the different experiences with right lobe grafts having various extents of HVC. In fact, most of the disputes were focused only on the resultant aspects of HVC, but not on the inducing and resolving mechanisms of HVC. This might be understandable because before hepatic transection, even the extent of HVC could not be assessed exactly prior to hepatic transsection.
Demand to quantify the amount of HVC and earlier assessment of HVC led us to develop a new prediction method. This CT-based method, which with PACS usually took 1 hour to make a complete set of HVC maps, enabled us to predict the amount of HVC with a mean error of 3.8% of RLV. It changed the previous rough estimate of the amount of HVC into a definite percentage of RLV.
Achievement of such a high accuracy is based on the assumption that occurrence of HVC depends on the anatomy of hepatic veins. Hereto MHV, right hepatic vein, and portal vein branches need to be identified in every section of the liver CT scans. Especially in segment 5, these 3 structures are intermingled unless one hepatic vein is exclusively dominant, as in donor cases 5 and 19. Adjacent CT sections provide valuable information to define these structures because their longitudinal axes are differently oriented. As a result of this study, we have now developed a simplified method, in which only a number of CT sections revealing easily identifiable vein anatomy were used instead of a complete set of consecutive CT sections. This method reduced the total evaluation time to 20–30 minutes, however, with some loss of accuracy.
In this study, we confirmed that the MHV or the right hepatic vein has its own territory of venous drainage and its boundary is defined as discoloration on simultaneous blockade of venous outflow and arterial inflow. However, one donor revealing only faint discoloration of HVC, presumably due to the lack of existent venous connections or intrahepatic collateral formation.6 Contrary to our expectations, the total amount of 7-day CTAA was not significantly decreased except for an unusually small proportion of hypoattenuation, which implies that the amount of venous outflow through venous collaterals was not enough to overcome HVC completely. As sufficient collateral drainage is uncommon, we cannot expect sufficient intrahepatic venous connections at the time of operation in most donor livers.7
From the analysis of 7-day CTAA, we assumed that certain compensatory mechanisms were active in the resolution of HVC. HVC appeared as low density on the preenhanced CT in half of the remnant right lobes, which might indicate a state of decreased parenchymal perfusion. Further analysis of enhanced CT images provided detailed information. Persistent hypoattenuation must be related to lack of tissue perfusion, despite visualization of a few enhancing glissonian branches. Some mechanisms inducing hyperattenuation were described in a CT study of temporary hepatic vein occlusion, in which the portal vein served as an outflow pathway of hepatic arterial flow.8 As arterial flow after crossing the sinusoids, flowed back into the portal vein, it was stagnated, appeared as delayed hyperattenuation in triphasic CT images. This arterial regurgitation can also be identified by intraoperative Doppler ultrasonography.2 Persistent hyperattenuation might be different from delayed hyperattenuation according to its time of emergence if the outflow tract of arterial regurgitation might be too short or the arterial driving force might be too strong to induce noticeable delay in enhancement. This presumption was upheld by close observation of CT images, in which there were many more enhancing glissonian branches at the persistent hyperattenuation portion compared to the delayed hyperattenuation portion, despite their different location. The intensity of persistent hyperattenuation became more prominent in the venous phase CT than in the arterial phase CT. After all, both persistent and delayed hyperattenuation can be regarded as being caused by arterial regurgitation, however they emerged differently in the time-sequence, depending upon the anatomical location and the state of arterial perfusion. Based on our observation, we assume that most of the hyperattenuation areas are deprived of portal flow with maintenance of reversed arterial flow, and the part of the liver with persistent hypoattenuation portion lacks both arterial and portal flows.
At 7 days after donor hepatectomy, the proportional amount of CTAA in remnant right lobes decreased when compared with that of the observed HVC. Its underlying mechanisms might be disproportionate regeneration of the right anterior segment compared with that of the right posterior segment.4, 9 To be precise, the liver parenchyma, regenerated rapidly, while the area with HVC or CTAA experienced impaired regeneration or no regeneration. We could not observe any parenchymal regeneration in the area of HVC during the first 7 days.
Evolution of HVC in the right lobe grafts might be different from those of the remnant right lobes because of ischemia-reperfusion injury, the recipient's own portal hypertension, or other transplantation-related factors. However, we could not compare these objectively in this study as all right lobes grafts underwent MHV reconstruction (Fig. 6). The profiles of 7-day CTAA revealed that all 7 right lobe grafts regenerated like the remnant right lobes with HVC of less than 20% of RLV. This restoration of graft regeneration power must be the effect of MHV reconstruction. The evolution of HVC in a graft liver might be altered by some transplantation-related factors, especially during the first week. We are convinced that changes in arterial perfusion are related to the proportional changes of hyper- and hypoattenuation. Any factors inducing arterial hypoperfusion may lead to the natural resolution of HVC on CT.
From close observation of the changes of HVC in donor remnant right lobes and recipient right lobe grafts on follow-up CT scans, we noticed that HVC required several weeks for its resolution and that reconstruction of the corresponding hepatic vein tributaries reduced the amount of HVC effectively. These data imply that MHV reconstruction is indicated for borderline-sized right lobe grafts with a considerable amount of predicted HVC and for grafts of any size with huge HVC.2
In conclusion, HVC was reliably predicted using a new CT-based method. We believe that this predictive method deserves to be applied as an essential part of preoperative donor evaluation.