Surgical management of suspected gallbladder cancer: The role of intraoperative frozen section for diagnostic confirmation

Preoperative diagnosis for suspected gallbladder cancers is challenging, with a risk of overtreating benign disease, for example, xanthogranulomatous cholecystitis, with radical cholecystectomies. We retrospectively evaluated the surgeon's intraoperative assessment alone, and with the addition of intraoperative frozen sections, for suspected gallbladder cancers from a tertiary hepatobiliary multidisciplinary team (MDT).


| INTRODUCTION
Gallbladder cancer remains an uncommon malignancy in the West, with only 1.6 per 100,000 in the UK annually. 1 However, the prognosis is poor with a mortality of 0.9 per 100,000 annually due to a combination of late presentation, aggressive biology, and limited systemic therapies. Several risk factors including old age, female gender, obesity, diabetes, geography (Eastern Europe, East Asia, and Latin America), family history, gallstones (including size, weight, duration, and cholesterol gallstones), polyps greater than 6 mm, primary sclerosis cholangitis, chronic infections (e.g., Salmonella or Helicobacter), and congenital biliary cysts have all been associated with gallbladder cancer. [2][3][4] Presenting symptoms are often vague, such as abdominal pain, weight loss, gastrointestinal disturbance, and jaundice a late feature, with most patients identified incidentally following cholecystectomy. 5 Ultrasonography (US) is often the default initial investigation for gallbladder pathology, concerning features include: gallbladderreplacing or invasive mass, irregular wall thickening, and intraluminal polypoid lesions. 6 However, for full assessment of suspected gallbladder cancers, a combination of thoracic, abdominal, and pelvic computed tomography (CT), liver magnetic resonance imaging (MRI), and endoscopic ultrasound (EUS) can be deployed. 7 While preoperative imaging is helpful, often there is only a suspicion of gallbladder cancer with the absence of tissue confirmation before surgery. Small studies utilising endoscopic techniques: ultrasound-guided fine needle aspiration (EUS-FNA), or transpapillary gallbladder drainage (ETGD) cytology, have had varied success in differentiating malignant disease from benign disease, and remain to be widely adopted. 8,9 Furthermore, due to the often subtle and perplexing imaging appearances of gallbladder cancer, its detection at a curative stage remains problematic. Managing suspected gallbladder cancer requires an experienced surgeon to balance prompt curative surgery for gallbladder cancer against unnecessary radical resection for benign disease, such as xanthogranulomatous cholecystitis.
For resectable gallbladder cancer, a radical cholecystectomy generally involves en bloc resection of neighbouring liver (≥2 cm cuff or anatomical resection of segment IVb and V) and portal lymphadenectomy with or without bile duct resection is recommended. 10 There is evidence to support the role of adjuvant therapy, especially in patients with lymph node positive or R1 disease. Whereas only a minority of patients will be suitable for curative surgery, postoperative outcomes have improved the past decades with median disease-free survival of 33.4 months in specialist centres. [11][12][13][14] The aims of this study were to retrospectively evaluate the management of suspected gallbladder cancers in our tertiary multidisciplinary team (MDT), and whether additional objective frozen section analysis would complement the subjective surgeon's assessment intraoperatively before proceeding to a radical cholecystectomy.

| METHODS
All patients with complex gallbladder disease referred to the regional

| Treatment strategy
Patients with suspected gallbladder cancer were reviewed and assessed in the out-patient clinic to undergo potential radical surgery.
Those found to have distant metastasis or unfit for surgery were reviewed by the medical oncologist for consideration of palliative systemic therapy.
The standard of care for suspected gallbladder cancers considered to have resectable disease after full staging was for the surgeon to assess the gallbladder intraoperatively, with radical resection undertaken only if there were clinical concerns regarding malignancy. If not, a simple cholecystectomy was performed with histology reviewed after. This traditional approach utilised intraoperative frozen section analysis of the cystic duct resection margin only, to assess the need for a resection of the bile duct. In our alternative approach, the entire gallbladder was submitted to the pathologist, where areas of surgical concern and the cystic duct were both sampled for frozen section analysis before proceeding to a radical cholecystectomy. The two cohorts were equally distributed from the MDT.

| Follow-up protocol
Following initial postoperative review at 1 month, patients with benign disease were discharged. Patients with gallbladder cancer were further discussed at the MDT for consideration of adjuvant therapy or routine surveillance, involving a CT scan of the thorax, abdomen, and pelvis at 6-month intervals during the first year postoperatively and annually thereafter.

| Statistical analysis
Continuous data are presented as medians with interquartile range and analysed using unpaired t test or Mann-Whitney test. Categorical data are presented as frequencies or percentages and analysed using Fisher's exact test. All statistical analyses were performed using StatsDirect statistical software 3.0 (StatsDirect Ltd.) and differences were considered significant at p ≤ 0.05.
Survival curves were estimated using the Kaplan-Meier method and compared using the log-rank test, performed on IBM SPSS Statistics 24.0 (IBM).

| The impact of radical surgery
Overall, 12 patients (25%) underwent radical surgery, 8 (66.7%) were malignant including a case of squamous cell carcinoma and melanoma, and 4 (33.3%) patients had complex benign pathology ( Table 3).     course is related to the extent of surgery rather than any underlying malignant pathology (Table S1).

| Outcomes
Minor complications, Clavien-Dindo less than or equal to 2, included electrolyte disturbance, wound related problems, and urinary tract infections. In the three patients who experienced major complications, one had a volvulus, another had an upper gastrointestinal bleed, and a third patient had a colonic anastomotic leak that required reoperation. There was no 30-day mortality.  Figure S1 and Table S2).

| DISCUSSION
Gallbladder carcinoma remains a highly aggressive and morbid disease with patients often presenting at advanced stages. Similar to many other malignancies, the outcomes of gallbladder cancer have benefited from the modern MDT approach, with specialist radiology imaging, perioperative management, and adjuvant therapy. With detailed investigation and timely radical surgery, the outcomes for patients with gallbladder cancer can be improved. We have reported our experience of managing a heterogenous group of patients with complex gallbladder disease. In our study, it was clear that the MDT process reliably identifies benign gallbladder disease to streamline their management As expected, frozen section analysis did significantly add to the operating time, but this did not translate to any difference in outcome, with differences mainly driven by the extent of surgery (Table 1).
However, perhaps more importantly both approaches allowed for the de-escalation of surgery, with all radical cholecystectomies undertaken for cancer or complex benign pathology. The extent of lymphadenectomy that should be undertaken during radical cholecystectomy is controversial, we recommend a limited approach of complete portal dissection, including the skeletonize the extrahepatic biliary tree, the hepatic artery and portal vein; and others suggesting clearance of the pericholedochal, posterosuperior pancreaticoduodenal, and the interaortocaval lymph nodes. 17,18 In our series, the median lymph node yield was 7. Over the years, there is increasing support for more aggressive lymphadenectomy of greater than 4 nodes rather than the minimal 1-3 nodes, and the American Hepato-Pancreato-Biliary Association (AHPBA) 2015 consensus statement going further recommending a minimum of six nodes. 10,19 The main limitation of this study is our small cohort, it is reasonable to speculate that in a larger cohort sampling error may affect the accuracy of frozen section analysis as is the case in any biopsy sampling.

| CONCLUSION
The MDT process is highly sensitive in identifying gallbladder cancers but lacks specificity. Thus, the surgeon's intraoperative assessment remains paramount in differentiating benign disease even among a patient cohort with indeterminate lesions or suspected gallbladder cancer, to de-escalate unnecessary radical surgery that comes with increased morbidity regardless of underlying pathology.
Intraoperative frozen section analysis was found to be a safe and viable adjunct to the surgeon's assessment, at a cost of additional resources and operative time. While the use of frozen section analysis has been recommended by others, a clinical trial can better address the diagnostic and staging difficulties for suspected early gallbladder cancers to better tailor surgical strategies. 10,14