Textbook outcome after major hepatectomy for perihilar cholangiocarcinoma — definitions and influencing factors

Patients undergoing MH in curative intent for PHC between January 2005 and August 2019 at the Department of Surgery, Campus Charité – Mitte and Campus Virchow Klinikum, Charité – Universitätsmedizin Berlin were retrospectively analyzed. This retrospective study was approved by the local ethics committee (EA2/006/16 and EA1/358/16). Variables included in the analysis were general patient characteristics such as gender, age, American Association of Anesthesiologists (ASA) score, and body mass index (BMI). Perioperative and histopathological data were recorded as well as data on overall survival (OS) and disease-free survival (DFS). Postoperative morbidity as classified by Dindo-Clavien during hospitalization, 30- and 90-day mortality, length of hospital stay, and intensive care unit stay, respectively, were recorded as well [21].

All patients who were referred to our institution for surgical treatment underwent a highly individualized and detailed workup. This routinely included computed tomography and/or magnetic resonance imaging of the chest and abdomen as well as endoscopic retrograde cholangiography (ERC). Biliary drainage with ERC and biliary stenting or percutaneous transhepatic cholangiodrainage (PTCD) was performed when necessary. Routinely, Carbohydrate antigen 19-9 (CA 19-9) was measured before resection. On patients with suspected peritoneal dissemination, diagnostic laparoscopy or laparotomy was performed.

All patients who underwent MH for PHC were included in the analysis. The surgical resection was performed either as a right or left sided major hepatectomy with extrahepatic bile duct resection as described before [2, 3]. Biliary reconstruction was performed as end-to-side hepaticojejunostomy. Patients with extrahepatic bile duct resection alone or multivisceral resections, e.g., hepatoduodenopancreatectomy (HPD) were excluded from the analysis as well as patients with intrahepatic or distant metastases or local peritoneal carcinomatosis.

In all cases, PHC was confirmed according to the histopathological reports of the resected specimen. Furthermore, data on resection and lymph node status as well as perineural sheath infiltration, microvascular infiltration, lymphangiosis carcinomatosa, and tumor differentiation were collected from pathology reports. Based on the collected data and the TNM classification valid at the time of resection, patients were assigned to the appropriate tumor stage according to the Union for International Cancer Control (UICC, 7th edition).

As TO had not yet been defined for major liver resection for PHC, a systematic literature review was conducted to evaluate common TO definitions used in HPB surgery. TO was defined based on the results of this literature review and common complications after PHC resection mentioned in the established literature. The search terms “textbook outcome” and “textbook oncologic outcome” were shortened to “textbook outcom*” and “textbook oncologic outcom*” and furthermore combined with the Boolean operator AND the following search terms: “liver surgery,” “hepatectomy,” “resection of liver,” “pancreatic surgery,” “pancreaticoduodenectomy,” “resection of pancrea*,” “hepatopancreatic surgery.” The publication period was limited to the years 2010 to 2020. Only studies in English with TO as primary endpoint were considered. Study designs such as reviews or meta-analyses were excluded. In addition, studies that investigated hepatic resections performed laparoscopically were excluded. TO was defined as the absence of 90-day mortality and major complications (i.e., > grade II according to Dindo-Clavien), no hospital readmission within 90 days after discharge, and no prolonged hospital stay (i.e., <75. percentile). The dichotomous textbook outcome was achieved when all four abovementioned individual criteria were observed in one patient after resection. Patients that could not be classified as either TO or NTO due to missing data were excluded from the analysis.

Patients were followed up in the outpatient clinic or with their general practitioner. Check-ups routinely included testing of CA 19-9 serum levels and abdominal ultrasound, CT, or MRI. Whether adjuvant chemotherapy was performed was recorded as well.

Statistical analysis was performed using IBM SPSS Statistics for Macintosh Version 25.0 (IBM Corp., Armonk, NY, USA). A p-value < 0.05 was considered statistically significant. Continuous parameters are presented as median and range and statistically compared with the non-parametric Mann-Whitney-U-test. Categorical data are displayed as counts and percentages and are compared using the chi-squared test or Fisher’s exact test, when necessary. To identify independent factors influencing TO, a binary logistic regression analysis was performed. Results are reported as odds ratio (OR) and 95% confidence interval (95% CI). Prognostic factors were included in the regression model when a significant influence on TO was detected in univariate analysis. The variables age and gender were included in multivariate analysis regardless of significance. Survival was estimated with the Kaplan-Meier method and compared between the outcome groups with the log-rank test. Patients who died within 90 days of surgery were not included in the survival analyses. A subgroup analysis was performed with all patients surviving at least 30 months.

A total of 20 records were identified through database searching. After removal of duplicates, 12 records were screened for eligibility. After excluding sources that did not meet the inclusion criteria, a total of 7 studies in HPB surgery with TO as primary endpoint were identified for the final review. Figure 1 summarizes the search algorithm, whereas Table 1 and Table 2 give a summary on included studies. All studies were retrospective and multicenter studies that were published between 2019 and 2020. The smallest and largest series contained 687 and 21234 patients, respectively. Postoperative mortality, length of stay, and hospital readmission were part of the TO definition in all seven studies. Four records [17, 19, 20, 24] used 30-day mortality and 30-day readmission, whereas three studies [13, 22, 23] used a span of 90 days instead. Histopathological criteria were used in four studies [13, 17, 19, 20] to define TO. This included tumor-free resection margins in all four studies [13, 17, 19, 20] and absence of lymph node metastases in one [20] study. Adjuvant chemotherapy and the need for postoperative transfusion were part of the TO definition in two studies [17, 20]. One study was not able to get detailed information on postoperative morbidity and used LOS instead [20]. Specific complications such as bile leakage or postoperative pancreas fistula were part of the TO definition on one study [24]. All studies investigated TO rates for different tumor entities or HPB procedures, and most studies sought to identify TO-influencing factors after surgery (Table 2).

In this study, two hundred and eighty-three patients who underwent surgical resection for PHC in curative intent between January 2005 and August 2019 and met the inclusion criteria were analyzed. The cohort comprised 171 (60%) male and 112 (40%) female patients with a mean age of 65 (33–86) years. The majority of patients was grouped in either ASA 2 (56%, n=159) or ASA 3 (38%, n=107), thus suffering from pre-existing conditions. Table 3 summarizes all patient characteristics of the total cohort. The majority of patients presented with Bismuth Type IV PHC or advanced UICC stages (IIIB–IVA: 48%, n=134). Tumor-free resection margins were achieved in 68% (n=189) of all patients, whereas 48% (n=132) patients had histopathologically confirmed local lymph node metastases. Most patients presented with a moderate tumor differentiation (G2: 68%, n=190). A total of 114 patients (40%) were diagnosed with cholangitis preoperatively. Most patients underwent biliary drainage (86%, n=243) before surgery. Portal vein embolization was part of the preoperative workup in 46% (n=129) of patients, all receiving right-sided resections. The majority of patients underwent right-sided liver resection (63%, n=178) of which 58% (n=165) were right trisectionectomies. Thirty-seven percent (n=105) underwent left-sided liver resection, of which 23% (n=64) were left trisectionectomies. En bloc portal vein resection was performed in 57% (n=162) of cases.

Most patients suffered from postoperative complications (90%, n=254), while major complications (grade IIIa–V) occurred in 66% (n=186) [21]. Common complications were infection (53%, n=150), pleural effusion (34%, n=96), bile leakage (30%, n=86), and postoperative liver failure (26%, n=74). Supplementary Table S1 gives an overview of the specific postoperative complications. After resection, the median length of hospital stay was 23 (3–213) days. The 30-day and 90-day mortality were 8% (n=22) and 15% (n=42). TO could be achieved in 24% (n=67), severe postoperative complications where the main reason not to be included in the TO group. From all patients, 22% (n=61) had to be readmitted or had a prolonged hospital stay (25%, n=72), thus not meeting the TO criteria. Some differences in patients’ characteristics were noted among patients who achieved TO versus patients who did not (Table 4). Patients with TO showed lower preoperative CA 19-9 levels (53 kU/l vs. 95 kU/l, p=0.047) and did less frequently undergo preoperative biliary drainage (73% vs. 90%, p=0.001). Left-sided hepatectomy was also associated with a higher TO rate when compared to right hepatectomy (52% vs. 32%, p=0.003). With regard to right-sided resections only side and extent of resection, TO is achieved in 18% (n=29) of right trisectionectomies versus 23% (n=3) after standard major hepatectomy (p=0.619). For left-sided resections, TO is achieved in 33% (n=21) after left trisectionectomy vs. 34% after standard major hepatectomy (n=14, p=0.888).

To detect independent factors influencing TO, variables shown in Table 5 that showed significant influence on TO in univariable analysis were included in a binary regression analysis model (Table 5). The regression analysis identified preoperative biliary drainage (OR= 0.405, 95% CI: 0.194–0.845, p=0.016) and left-sided-resection (OR= 1.899, 95% CI: 1.048–3.440, p=0.035), as independent factors influencing TO, whereas left-sided resection was associated with higher and preoperative biliary drainage with lower odds of TO. Despite significant differences in univariable analysis, tumor differentiation was not an independent predictor for TO (OR= 0.547, 95% CI: 0.263–1.137, p=0.106). The patient-related factors age, sex, and ASA score had no impact on TO (age: p=0.392, sex: p=0.456, ASA: p=0.714).

Median OS of all patients was 29 (24–35) months after resection, whereas DFS was 22 (17–26) months. After excluding patients who died within 90 days after resection, no significant difference between the two groups in either OS (p=0.280) or DFS (p=0.735) analysis could be detected. However, a trend towards better overall survival in patients with TO especially in the late course could be shown that failed to reach statistical significance. Subgroup analyses of patients surviving at least 30 months (Supplementary Table S2) after resection showed better overall survival in patients with TO compared to patients without TO (92 versus 60, p=0.039). There was no difference in DFS between patients surviving at least 30 months (p=0.270, Figure 2A–D).