Salvage liver transplantation (LT) has been performed for recurrence of hepatocellular carcinoma (HCC) or for deterioration of liver function after primary liver resection for HCC. Questions arise, however, regarding the technical feasibility of salvage LT in patients who have undergone prior liver surgery. Because initial liver resections in these patients are usually minor in extent, performing deceased donor whole liver graft implantation is not regarded as a contraindication. In contrast, prior performance of a major liver resection, such as right lobectomy, makes subsequent recipient hepatectomy technically difficult. To date, only a small number of salvage LT operations after major liver resection have been reported.1, 2 Living donor liver transplantation (LDLT) can also be performed for salvage. Although prior minor hepatectomy is acceptable for salvage LDLT, there have been no reports to date of salvage LDLT after major hepatectomy. Moreover, the extent of recurrent HCC eligible for salvage LT is not known.
To assess the technical feasibility and indication of salvage LDLT for recurrent HCC, we retrospectively assessed our experience performing these surgical procedures over 10 years. We especially analyzed the outcome of various combinations of prior major hepatectomy and graft types in patients undergoing LDLT for recurrent HCC after prior liver resection.
From February 1997 to March 2006, a total of 1,049 cases of adult LDLT have been performed in our institution,3, 4 with 342 recipients (32.6%) diagnosed as having HCC. Among the latter, 17 (5%) had undergone salvage LDLT after prior hepatectomy, including 5 who had undergone major liver resection (right or left lobectomy) and 12 who had undergone minor liver resection (sectionectomy, segmentectomy, or subsegmentectomy) (Tables 1 and 2).
Table 1. Clinical Profiles of Patients Who Underwent Salvage and Primary Living Donor LT for Hepatocellular Carcinoma
Types of LT
Types of prior liver resection before LT
Primary (n = 200)
Salvage (n = 17)
Major resection (n = 5)
Minor resection (n = 12)
Abbreviations: LT, liver transplantation; MELD, Model for End-Stage Liver Disease; GRWR, graft-recipient weight ratio.
Fifteen of 17 patients experienced recurrence of HCC; 4 underwent major resection and 11 underwent minor resection.
In our institution, the criteria for salvage LDLT were basically similar to those for primary LDLT. In most patients, the extent of recurrent HCC was within the Milan criteria, but a few selected patients had more advanced recurrent tumors. The tumor staging at the time of initial hepatectomy was not taken into account. All patients underwent LDLT within 2 months after LT; this was considered a rescue treatment for recurrent HCC. After the decision was made to perform LDLT, the recurrent tumors were not further treated.
HCC workup routinely included positron emission tomography, radioisotope bone scan, magnetic resonance imaging, and chest computed tomography, with special attention to extrahepatic metastases.5 Recipient hepatic vascular structures were thoroughly evaluated through dynamic computed tomography with 3-dimensional reconstruction. During pretransplant evaluation, no patient was excluded as a result of prior liver resection per se.
When deciding on the type of partial liver graft to use for salvage LDLT, graft size-to-recipient matching was considered critically important, with the goal being a graft-recipient weight ratio of >0.8. The extent of prior liver resection did not change the type of liver graft selected, but dual-graft implantation was not attempted after major liver resection.
Patients who had undergone prior minor liver resection underwent routine surgical procedures for adult LDLT without technical modifications. For patients who had undergone either right or left lobectomy, however, special attention was paid during mobilization of the native remnant liver for dissection of the previous liver cut surface. After fully severing the cut surface adhesions, the remaining mobilization procedures were the same as those for ordinary LDLT. Hepatic hilar adhesions were usually not severe because the hepatic hilum had not been dissected extensively during prior hepatectomy. To accommodate a sizable liver graft, either the right or left subphrenic space was extensively dissected, depending on the graft type. For secure hepatic venous reconstruction, the recipient inferior vena cava was fully isolated. Methods of graft vascular and biliary reconstructions were prudently selected, depending on the availability of recipient hepatic vein stump, portal vein, hepatic artery and bile duct, on a case-by-case basis.6
Tumor Staging and Statistical Analysis
The modified tumor, node, metastasis system (pTNM) classification was used to assess the extent of primary and recurrent HCC lesions.7, 8 As a control group for comparison of patient survival and HCC recurrence, we used clinical data of the first 200 patients who underwent primary LDLT for HCC (Table 1).
Mean values with standard deviation and median values with range were used for numeric data. The significance of differences was assessed by χ2, Fisher's exact test, and t test. Survival curves were estimated by the Kaplan-Meier method and compared by the log rank test. Statistical significance was set at P < 0.05.
Patient profiles are summarized in Tables 1 and 2. In 15 of 17 patients, the main indication for salvage LDLT was treatment failure, either recurrent HCC or ineffective treatment of tumor recurrence. Two patients who had undergone prior left lobectomy received right lobe grafts, whereas of the 3 patients who had undergone prior right lobectomy, 2 received right lobe grafts and 1 received a left lobe graft. The mean interval between initial liver resection and LDLT surgery was 37 ± SD 26 months (range, 1-82 months).
Operation Profiles and Posttransplant Complications
Operation profiles and posttransplant complications are summarized in Table 1. Operation time was not greatly prolonged in the salvage compared with the primary LDLT group. The incidence rates of various posttransplant complications, other than bleeding, were similar in the salvage and primary LDLT groups. The incidence of bleeding complications was higher in the salvage than in the primary LDLT group (P = 0.016), but it did not differ between patients who had previously undergone major and minor liver resections (P = 0.528).
Patient Survival and HCC Recurrence
One patient (patient 2) with prior left lobectomy died of multiorgan failure that was initiated by postoperative bleeding, but there were no other cases of perioperative mortality. Overall survival did not differ significantly between the salvage and primary LDLT groups (P = 0.247, Fig. 1). After confining analysis to patients who met the Milan criteria, the survival rates between the salvage and primary LDLT groups became very close (P = 0.938, Fig. 2); in these patients, there was no incidence of HCC recurrence after salvage LDLT.
In our institution, HCC was detected in about 30% of the explanted livers after adult LT, and 5% of these patients underwent salvage LT. Although HCC in patients with cirrhotic livers has been often managed by nonsurgical treatments, the number of HCC resections is gradually increasing in our program. The number of LDLTs for HCC patients is also increasing because LT has become widely accepted as a definite treatment to cure HCC and liver cirrhosis simultaneously. If a living donor is available, a nonnegligible proportion of HCC patients with advanced liver cirrhosis prefer LDLT as initial treatment or after failure of nonsurgical HCC treatments.9, 10 It is likely that the number of patients undergoing salvage LDLT will also increase.
Our results indicate that LDLT is technically feasible, even after prior major liver resection. Compared with patients undergoing primary LDLT, there was little additional perioperative risk from prior liver resection. Only 1 patient, the first who had undergone prior major liver resection and the 14th adult LDLT recipient in our institution, died from surgical complications. Our subsequent accumulation of experience in surgical techniques and postoperative management has reduced the surgical risk from prior liver resection such that the operation time and the volume of blood transfused do not remain serious drawbacks. Because the operation profiles and complication rates were similar in patients who had undergone prior major and minor liver resections, these 2 patient groups may have the same risks from surgery.
Two technical concerns arise when performing recipient hepatectomy, especially right hepatectomy, after prior liver resection. The first is in regard to recipient liver isolation with tolerable amount of bleeding. Heavy adhesions are often encountered after prior liver resection, and minute collaterals penetrate into such adhesions in patients with portal hypertension. Inattentive dissection of such perihepatic adhesions could result in many uncontrollable sites of pinpoint bleeding at the dissection surface. This technical difficulty may be one reason many transplant surgeons are unwilling to perform salvage LT after major hepatectomy. Similar difficulties have been encountered during late retransplantation after LDLT. Our experience with salvage LDLT and late retransplantation indicates that steady, meticulous sharp dissection is the only method to successfully determine the adequate dissection plane. Once the cut surface adhesions are fully severed, other mobilization procedures are similar to those for primary LDLT. The second technical concern is in regard to secure reconstruction of the hepatic vein outflow, which depends on the anatomical characteristics of the partial liver graft and hepatic vein stump at the recipient inferior vena cava. We have found that any method of hepatic vein reconstruction can be used after full dissection of the recipient inferior vena cava.6, 11, 12
We have shown here that various combinations of prior recipient hepatectomy and partial liver graft are possible in practice. Although we did not use one combination, left lobe graft implantation after right lobectomy, our results suggest that it may be technically feasible to use all 4 major combinations of salvage LDLT: right or left lobe graft and prior right or left lobectomy. We also found that HCC recurrence and patient survival after salvage LDLT were similar to those after primary LDLT, especially within the Milan criteria. No general consensus exists, however, regarding LT for recurrent HCC after liver resection. The outcome of salvage LT has been reported to be equivalent to that of primary LT, although primary resection and subsequent salvage LT has also been found to result in unfavorable outcomes compared with primary LT.1, 2 LDLT, however, differs from deceased donor LT in that the latter usually includes intervening waiting periods, with primary resection performed to avoid tumor progression. In situations in which there is an extreme shortage of deceased donors and limited availability of living donors, HCC resection is performed only for curative intent, without considering bridges to later salvage LT. Thus, indications for salvage LDLT may differ from those for salvage deceased donor LT.
Indications for salvage LT have been reported to differ, especially regarding the acceptable extent of recurrent HCC. Liver failure after liver resection can be an indication for LT, with even minor liver resection found to induce posthepatectomy hepatic failure in patients with advanced liver cirrhosis. Such early deterioration of liver function is vulnerable to infection. We do not regard superimposed infection as a contraindication for salvage LT unless it dramatically progresses to overt sepsis. Gradual progression of liver cirrhosis after liver resection is also an indication for salvage LT.
Recurrent HCC after liver resection is the main indication for salvage LT, but there is no consensus on its eligibility criteria. We used as the indication recurrent HCC meeting or slightly exceeding the Milan criteria. Although we had assessed the tumor extent at the time of prior liver resection and the response to the nonsurgical treatments, only the current extent of recurrent HCC per se was taken into account in practice. Because many patients in this series underwent repeated treatments of tumor recurrence, there may be some downstaging effect. Our results suggest that the survival outcome of salvage LDLT, based on explant tumor staging, was equivalent to that of primary LDLT. Thus, candidates for salvage LDLT can be selected without consideration of downstaging effects from preceding adjuvant therapies, and it may be reasonable that primary and salvage LDLT share HCC indication criteria.1, 2, 13–16 In our institution, the criteria for primary LDLT for HCC have been recently established as maximum tumor diameter of 5 cm or less, 6 or fewer tumors, and the absence of gross vascular invasion after detailed survival analyses.
For salvage LDLT, determination of the timing of surgery is of critical practical importance. We performed LDLT as early as the situation permitted in order to avoid tumor progression. In this series, the timing of LDLT was influenced by the timing of referral because 11 of the 15 patients with recurrent HCC had been referred from other institutions. Because most patients had undergone repeated episodes of transarterial chemoembolization or radiofrequency ablation therapy, the interval between initial recurrence after prior liver resection and LDLT operation was prolonged. For patients followed up at our institution, we usually recommend that LDLT be performed before the recurrent tumor exceeds the Milan criteria.
It is unclear whether patients who had early HCC recurrence after liver resection are candidates for salvage LDLT. If an early recurrence is associated with a well-advanced tumor or unfavorable pathology, the patient may not be indicated for salvage LT because there is a high probability of posttransplant HCC recurrence. In contrast, if early recurrence is related to incomplete local control of small HCC lesions, the patient may be indicated for salvage LT. Patients with a longer interval between prior liver resection and salvage LDLT showed more favorable survival than those with a shorter interval. A thorough pretransplant HCC workup should be performed because extrahepatic recurrence was not uncommon in patients with recurrent HCC after resection.5, 16
In conclusion, our results suggest that every combination of recipient prior hepatectomy and living donor liver graft is feasible for salvage LDLT and that the acceptable extent of HCC for salvage LDLT is equivalent to that for primary LDLT.