The diagnostic value of staging laparoscopy in gallbladder cancer: a nationwide cohort study

In this retrospective database study, patients from all seven academic centres in the Netherlands diagnosed with GBC between November 1999 and May 2018 were eligible for inclusion. Eligible patients were identified through the Netherlands Cancer Registry (NCR), which contains data on all newly diagnosed malignancies, by creating a list of all patients with a GBC diagnosis between November 1999 and May 2018 per academic centre. To make groups of patients with potentially resectable pGBC and patients with iGBC planned for re-resection, hospital medical records of all eligible patients were examined. pGBC patients deemed unresectable due to DD on radiological imaging as well as iGBC patients not planned for re-resection were excluded from the study. These data include a patient identification number, treatment hospital, and date of diagnosis. The NCR is maintained by the Netherlands Comprehensive Cancer Organization (IKNL) and is notified of newly diagnosed patients by the automated pathological archive (PALGA) [15], the nationwide network and registry of histo- and cytopathology of the Netherlands, and supplemented by data from the National Archive of Hospital Discharge Diagnosis. Of eligible patients identified from the NCR, patient information was collected from hospital medical records. Two groups of patients were included; the first group contained patients with pGBC planned for resection with curative intent. pGBC was defined as GBC identified at radiological imaging, including US (ultrasound), CT, FDG PET-CT (fluorodeoxyglucose PET-computed tomography), and MRI. The second group contained patients with iGBC planned for re-resection. iGBC was defined as unsuspected gallbladder cancer found during cholecystectomy for benign gallbladder disease or postoperative histopathological analysis. Re-resection was defined as additional GBC-directed surgery with curative intent within 6 months after primary surgical exploration. The study was approved by an accredited Medical Research Ethics Committee (METC Oost-Nederland) with number 2017-3912.

In the Netherlands, diagnostic and treatment procedures are described in hospital-specific guidelines, which in turn are based on guidelines published by the Dutch National Workgroup for Gastrointestinal Tumours [16]. The first version of this national guideline was published in 2004 and a revised version was published in 2013. Recommended and performed diagnostic and treatment procedures may therefore have varied per patient and per hospital. SL was defined as a laparoscopic operation in which resectability of the gallbladder bed is assessed and in which lymph nodes, liver, peritoneum, omentum and other visible abdominal organs are examined for signs of metastatic disease. A frozen section procedure was performed of masses suspected of metastasis. Laparoscopic ultrasonography of the liver was performed in some but not all cases. Both the one-stage approach, with direct continuation to laparotomy, and two-stage approach, with scheduling of laparotomy on a future date, were described. In iGBC patients, SL was performed after initial cholecystectomy and before (possible) re-resection.

Patient characteristics such as age, gender, ASA Physical Status Classification [17], symptoms, previous gallbladder disease, organ invasion and T- and N-stage on preoperative imaging, description and results of operative procedures, tumour characteristics, and other histopathological findings and follow-up data were extracted from patient medical records. The number of days between SL and (re-)laparatomy was recorded.

The primary outcome of this study was the yield of SL, defined as the proportion of patients with the presence of DD during SL. The yield was assessed separately for pGBC and iGBC patients, as diagnostic and treatment pathway and expected yield for these groups are different.

Secondary outcomes were also determined separately for pGBC and iGBC patients. In pGBC patients, the location of DD and the resection rate was determined. Resection rate was defined as the proportion of patients that underwent laparotomy for resection with curative intent with a macroscopic radical resection. In the pGBC group, patient characteristics were compared between patients who did and did not undergo SL. For iGBC, predictive factors for DD were analysed in order to investigate which group of iGBC patients would benefit most from SL.

Baseline characteristics were described using frequencies and percentages for categorical variables, and median value and range for continuous variables. To compare SL- and no SL-groups, Chi-squared tests were used for categorical variables and the Student’s t-test was used for continuous variables. Univariable and multivariable analysis were used to identify individually predictive factors for presence of DD during laparotomy. All clinical or histopathologic findings with a significance level of P≤0.1 in univariable analysis were included into a multivariable model using backwards Wald logistic regression. Results from the regression analysis were reported as Odds ratios (ORs) with 95% confidence interval (CI). All analyses were performed with SPSS version 26.0 [18]. P-values ≤ 0.05 were considered statistically significant.

Of the 183 included patients with pGBC, 143 (78%) underwent laparotomy without SL. Resection was performed in 101 of these 143 patients (71%). In 42 patients (29%), resection was not performed due to DD, a combination of liver metastases (n = 11), peritoneal seeding (n = 10), lymph node metastases (n = 10), and local unresectability (n = 18) (Fig. 1).

Liver invasion on preoperative imaging was present more often in the SL group (20/40 vs 45/143, p = 0.030), as well as invasion into other organs (18/40 vs 34/143, p = 0.008) (Table 1).

Forty of the 183 included pGBC patients (22%) underwent SL and DD was detected in eight (20%, 8/40). Reasons for DD were the presence of (a combination of) liver metastases (n = 3), peritoneal seeding (n = 3), distant lymph node metastases (n = 1), and local unresectability (n = 5). Thirty-two patients were considered resectable after SL and subsequently underwent laparotomy. DD was detected during laparotomy in nine out of 32 patients; due to liver metastases (n =2), peritoneal seeding (n = 1), lymph node metastases (n = 2), and local unresectability (n = 4) (Fig. 1).

The resection rate after SL did not differ significantly from the resection rate in patients who underwent laparotomy without SL (23/32 vs. 101/143; p = 0.89). The number of days between SL and resection was not significantly different between patients who did and did not have DD upon laparotomy: 13 days (range 0–43) in patients with DD and 16 days (range 0–190) without DD (p = 0.52). In 16 patients, SL and resection were performed in one session.

Of the 107 included patients with iGBC, 100 (93%) underwent laparotomy without SL. Re-resection was performed in 81 of these 100 patients (81%). In 19 patients (19%), re-resection was not performed due to DD; peritoneal seeding (n = 11), lymph node metastases (n = 4), liver metastases (n = 1), a combination of lymph node and liver metastases (n = 1), and a combination of lymph node and peritoneal metastases (n = 2) (Fig. 2).

Seven of the 107 included iGBC patients (7%) underwent SL and yielded DD in one due to peritoneal metastases. Six patients were considered resectable by SL and subsequently underwent laparotomy. DD was detected during laparotomy in two out of six patients; peritoneal metastases in one, and a distant lymph node metastasis in another (Fig. 2). Overall, 22 of 107 iGBC patients had DD (21%). The median time between primary surgery and re-resection in iGBC patients was 75 days (range 0–297).

Factors predictive for the presence of DD in univariable analysis were cholecystitis, T3 disease, perineural invasion, positive lymph node status, and R1/R2 resection at primary exploration (Table 2). Multivariable logistic regression showed that cholecystitis (OR 4.25, 95% CI = 1.51–11.91; p = 0.006) and R1/R2 resection at primary surgery (OR 3.94, 95% CI = 1.39-11.19; p = 0.010) were independent predictive factors for DD at re-resection.