Current role of trisectionectomy for hepatopancreatobiliary malignancies

Abstract Background Trisectionectomy is a treatment option in extensive liver malignancy, including colorectal liver metastases (CRLM). However, the reported experience of this procedure is limited. Therefore, we present our experience with right hepatic trisectionectomy (RHT) for CRLM as an example and discuss the changing role of trisectionectomy in the context of modern treatment alternatives based on a literature review. Methods Between January 1993 and December 2014 all patients undergoing RHT at a single center in the UK for CRLM were included. Patient and tumor characteristics were reviewed and a multivariate analysis was done. Based on a literature review the role of trisectionectomy in the treatment of HPB malignancies was discussed. Results A total of 211 patients undergoing RHT were included. Overall perioperative morbidity was 40.3%. Overall 90‐day mortality was 7.6% but reduced to 2.8% over time. Multivariate analysis identified additional organ resection (P = .040) and blood transfusion (P = .028) as independent risk factors for morbidity. Multiple tumors, total hepatic vascular exclusion, and R1 resection were independent risk factors for significantly decreased disease‐free and disease‐specific survival. Further surgery for recurrence after RHT significantly prolonged survival compared with palliative chemotherapy only. Conclusion With the further development of surgical and multimodal treatment strategies in CRLM the indications for trisectionectomy are decreasing. Having being formerly associated with high rates of perioperative morbidity and mortality, this single‐center experience clearly shows that these concomitant risks decrease with experience, liberal use of portal vein embolization and improved patient selection. Trisectionectomy remains relevant in selected patients.


| INTRODUC TI ON
Left hepatic trisectionectomy (LHT) was first described in detail by Starzl and colleagues as a left trisegmentectomy in 1982, and then as an extended left hepatectomy by Blumgart et al in 1993. 1,2 The designation LHT was adopted following the International Hepato-Pancreato-Biliary Association Brisbane 2000 consensus statement on the nomenclature of liver anatomy and resection.
LHT is defined as excision of Couinaud liver segments 2, 3, 4, 5 and 8, with or without segment 1. 3 Despite improvements in surgical techniques and perioperative patient management, only a few papers have reported outcomes of LHT in more than 10 patients. [4][5][6][7][8] Morbidity and mortality after LHT is higher than for other hepatectomies, and this procedure is reserved for patients with a significant tumor burden and an otherwise dismal prognosis. The high morbidity rate is attributable mainly to the aggressive nature of the disease being treated, but may also be related to the extent of liver volume resected, estimated to be as high as 80 per cent. 2 In 2005, the Leeds group reported long-term outcomes of LHT in 70 consecutive patients. 4 Morbidity rate was high, but the potential for cure supported an aggressive surgical resection policy where other treatment options had been exhausted. In 2016, the same group described changes in surgical practice over time, and analyzed the short-and long-term outcomes of LHT for hepatobiliary malignancy, in order to identify factors associated with morbidity and mortality in the modern era. 9 Right hepatic trisectionectomy (RHT) was first described by Lortat-Jacob, Robert and Henry as right lobectomy in 1952. 10 This operation has had a number of different names, but, until recently, it has been most commonly known as right trisegmentectomy. The designation RHT was adopted following the International Hepato-Pancreato-Biliary Association Brisbane 2000 consensus statement on the nomenclature of liver anatomy and resection. 3,11 This procedure requires excision of segments 4, 5, 6, 7 and 8 ± 1 and it also remains one of the most challenging major hepatectomies.
Despite improvements in surgical technique and perioperative critical management, perioperative morbidity remains high and only a few hepatobiliary centers worldwide have reported their experience. 12,13 Modifications of LHT and RHT by in-contiguity and non-anatomical extension and repeat liver resection after LHT or RHT are also rarely reported. 14,15 The role of these technically demanding and extensive resections in contemporary hepatobiliary practice is established for primary liver cancers and for those tumors with no significant neoadjuvant strategies, but it is also changing as new treatments emerge. This is particularly true for patients with colorectal liver metastases (CRLM), and it is likely that this trend will be followed for other HPB malignancies as more effective preoperative strategies are developed. Emerging data for intrahepatic cholangiocarcinoma, for example, is encouraging. [16][17][18][19] For patients with CRLM, despite the lack of compelling data for most patients, there has been a paradigm shift in the oncological assessment of patients and the use of neoadjuvant and "downstaging" strategies before resection. This has been combined with a sensible development of surgical strategies aimed at parenchymal preservation, along with new developments in liver surgery such as multistage resection as a classical two-stage approach (TSH) or associating liver partition and portal vein ligation for staged hepatectomy (ALPPS). In the classical two-stage approach, portal vein embolization (PVE) or portal vein ligation (PVL) is carried out to stimulate hypertrophy in the planned future liver remnant, along with resection of tumors from the planned future liver remnant (FLR). After an interval of 4-8 weeks, with adequate hypertrophy of the FLR, the definitive resection is carried out. 20 Besides PVL/ PVE, the first step in ALPPS includes at least a 50% transection of liver parenchyma. 21 By this modification, ALPPS seems to be able to accelerate liver growth of the FLR and to shorten the interstage interval. 22,23 A recent Scandinavian randomized controlled trial has shown the benefits of ALPPS in providing a higher resection rate compared to the classical two-stage procedure, with comparable margins, complications and short-term mortality. 24 However, besides the evolvement of these promising strategies, there remains a place for up-front major resection for many patients. In the light of this trend, we have reviewed in detail a 22-year single-center experience of RHT for CRLM and evaluated factors affecting morbidity and survival in order to provide a critical appraisal for the role of RHT for CRLM in order to add these data to our previous work on LHT.

| Study design
Patients undergoing RHT between January 1993 and December 2014 were identified from a prospectively maintained database at a single institution. Additional data from the database included radiological investigations and interventions, presence or absence of jaundice, extent of surgical resection, duration of operation, requirement for transfusion of blood or blood products, need for Pringle maneuver or total vascular exclusion, additional surgery (lymphadenectomy, extrahepatic bile duct excision with reconstruction, or vascular reconstruction), histopathological diagnosis, size and distribution of tumors, perioperative morbidity and mortality, and long-term disease-free and disease-specific survival. This work has been reported in line with the PROCESS criteria. 25 All patients undergoing liver resection were offered adjuvant chemotherapy according to guidelines unless they had received adjuvant therapy following their colonic resection within the past 12 months. However, detailed data on adjuvant chemotherapy after colorectal and hepatic surgery were not routinely collected in the database owing to the large number of patients presenting from a wide geographical area of referring hospitals using chemotherapy. In 12 patients, neoadjuvant chemotherapy was used as either a downsizing technique or as a "test of time approach."

| Preoperative evaluation
Preoperative radiological assessment in all patients included thoracic, abdominal and pelvic computed tomography (CT), and magnetic resonance imaging (MRI) of the liver. The investigations were reviewed in a multidisciplinary team meeting to discuss and define the extent of resection. In selected cases, positron emission tomography CT (PET-CT) was used. From 2007, PVE was used when the future liver remnant was estimated to be <20% and was carried out 3 to 4 weeks before scheduled liver resection, but no formal volumetry studies have been done in our center.

| Perioperative care
Techniques of RHT and extensions of RHT have been described previously. 14,26,27 Intraoperative ultrasound was carried out in all

| Morbidity and mortality
Details of complications were obtained from the database and, where necessary, from the patient notes and graded according to the validated Clavien-Dindo classification system. 28 Postoperative liver failure was defined according to the International Study Group of Liver Surgery. 29 Postoperative mortality was defined by the occurrence of death within 90 days of surgery or at any time during postoperative hospital stay.

| Histopathological evaluation
Pathological reports were reviewed to determine tumor histological grade, margin status, and histological abnormalities in the non-tumorbearing liver (NTBL). A tumor-free resection margin of less than 1 mm was classified as (R1), and 1 mm or more was classified as (R0). 30 In relation to NTBL, liver steatosis was defined as diffuse accumulation of fat droplets affecting >5% of hepatocytes. 31 Fibrosis was scored according to the Metavir score, and defined as the presence of portal fibrosis with/without septa, numerous septa, or cirrhosis. 32 Sinusoidal injury was graded and defined as the presence of centrilobular involvement beyond one-third of the lobular area. 33 These findings in NTBL were defined as parenchymal liver damage in the present study.

| Follow up
All patients were followed up regularly at the outpatient clinic at 1, 3, 6 and 12 months in the first year, 18 and 24 months in the second year, and yearly thereafter if the patient remained disease-free.

Follow up included clinical examination and assessment of tumor markers (carcinoembryonic antigen [CEA], cancer antigen [CA]19-9).
Surveillance imaging included CT scans of the chest, abdomen, and pelvis at 3, 6, 12, 18 and 24 months, annually to 5 years and again at 7 and 10 years. MRI and PET-CT were carried out if recurrence was suspected in routine follow up.

| Statistical analysis
Continuous variables were expressed as median and interquartile range. To consider changes over the study period, patients were di-

| Tumor characteristics
In the cohort, 49 patients (23%) had solitary tumors and median size of the largest tumor was 50 (range 8-410) mm. Neoadjuvant chemotherapy was given to 12 (5.6%) patients. In 13 (6%) patients, portal vein embolization was done before the actual surgery.

| Short-and long-term outcomes, morbidity and mortality
Of the whole cohort, 15 (7.1%) patients died in hospital. One other patient (0.5%) died within 90 days following surgery. Therefore, 16 patients (7.6%) died within 90 days. Among these 16 patients, main causes for mortality were as follows: seven (44%) patients died from multi-organ failure; three (19%) patients died as a result of gastrointestinal bleeding; two (13%) due to acute myocardial infarction; one (6%) as a result of pneumonia or intra-abdominal abscess (n = 1, 6%); massive abdominal bleeding (n = 1, 6%) in hospital; and unknown cause (n = 1, 6%) after discharge. The 211 included patients were further divided into three time periods where 70 patients were included in the first period, 70 patients in the second period and 71 patients in the third period, respectively. With increasing experience at our center we were able to decrease 90-day mortality from 12.8% to 7.1% and 2.8% accordingly. These differences were not significant.
Univariate analysis for morbidity showed that other organ resection (P = .017) and ARBC transfusion (P = .012) were markers for poor outcome. Both variables were also found to be independent predictors for morbidity in multivariate analysis (odds ratio

| Changes in outcomes over time and redosurgery for recurrence
To evaluate the impact of the learning curve and changes over the study period, patients were divided into three operative experience periods: first (n = 70; 33.2%), second (n = 70; 33.2%), and third (n = 71; 33.6%) period as described in Table 5 Figure 1). dysfunction. 38 Therefore, the use of PVE before trisectionectomy has been advocated to decrease postoperative morbidity and mortality and make these operations safer. 39   A limitation of the present study is the incomplete data on chemotherapy. As a result of this limitation, a detailed analysis of the role of chemotherapy in this treatment algorithm was not possible.

| D ISCUSS I ON
Only a relatively small proportion of patients received neoadjuvant therapy for downsizing, but the proportion in whom downsizing strategies failed was not captured in the data set.

| CON CLUS ION
Left hepatic trisectionectomy and RHT are technically demanding liver resections with a high risk for perioperative morbidity and, in the past, also mortality. Our data show these risks are reducing with experience, better patient selection, and the more liberal use of PVE.
LHT and RHT remain relevant for many situations but innovation in surgery and neoadjuvant treatments inevitably mean that the role of these challenging operations is decreasing.