Radical cholecystectomy for gallbladder cancer: Real-world applicability of perioperative and oncologic ‘benchmarks’

INTRODUCTION

Gallbladder cancer (GBC) is the most common biliary epithelial malignancy and ranks sixth most common gastrointestinal cancer.^[1,2] India is a high-incidence region for GBC and contributes to 10% of the global disease burden.^[3] Within India, incidence is high in North, North-East, Central, and East India.^[3,4] In women of North / North-East India, GBC is one of the three leading cancers. The incidence in North India ranges between 10–22/100,000 population, which is similar to other high-incidence countries such as Chile, Bolivia, South Korea, Japan, and China.^[3,4] As per GLOBOCAN data, global statistics for the year 2022 indicate that there were approximately 122,462 new cases of gallbladder cancers and close to 89,031 deaths.^[5] The estimated incidence of gallbladder cancer in India for the year 2024 is 36,640 new cases and 24,690 deaths.^[6]

According to the NCCN guidelines 2025, standard components of surgery for resectable GBC include cholecystectomy, lymphadenectomy, and hepatectomy to achieve margin-negative resection of the gallbladder fossa.^[7]

Radical cholecystectomy (RC) is the recommended term for radical surgery for GBC (IHPBA International HepatoPancreato-Biliary Association consensus guidelines).^[2] All medically fit patients with disease ≥ T1b confined to the locoregional lymph node and with no distant metastases are candidates for RC. The essential components of the procedure include en bloc cholecystectomy, wedge resection (WR) of gallbladder fossa or anatomic resection of segments 4b and 5 (AR), and lymphadenectomy for lymph node stations along the common hepatic artery (station 8), hepatoduodenal ligament (station 12c, 12b,12a, and 12p), and posterior superior pancreaticoduodenal (station 13).

More extended resections, such as major hepatectomy, vascular or adjacent organ resection, or bile duct excision, are classified as an extended radical cholecystectomy (ERC) and indicated only in selective patients.^[2]

With a reasonable consensus on the standardisation of the technical aspects of RC, focus has shifted to defining surgical/oncologic quality parameters of the procedure.^[8-11] Thus, high-quality surgery (HQS) that evaluated oncologic parameters has been defined to include cholecystectomy with hepatectomy, lymph node harvest of ≥6 lymph nodes, and negative margins.^[8] In this study of the NCDB database, of the 3796 patients, only 364 (9.6%) met HQS criteria. Further, this study reported HQS, centralisation, open approach, and insurance coverage as factors to achieve optimal outcomes.^[8] In another study, textbook outcomes for liver surgery (TOLS) have been identified to improve survival.^[9] The authors reported that age ≤ 70 years, absence of preoperative jaundice (total bilirubin≤3 mg/dl), T1 stage, N0 stage, WR, and no neoadjuvant therapy were independently associated with TOLS.^[9] In this study, it is thus apparent that TOLS was achieved in younger patients with an early stage disease who underwent WR.

A recent multicentre study from 13 high-volume centres evaluated 906 patients who underwent curative intent surgery from 2000–2021. Benchmark group included 245 (27%) low-risk patients (defined as those without significant co-morbidities and no need for vascular or bile duct reconstruction). Benchmark values were defined as the best possible outcomes in 8 domains of perioperative and oncologic outcomes.^[11]

With benchmark values as reference standards, at a tertiary care centre in the gallbladder cancer endemic area in India, we evaluated outcomes in our patients undergoing RC and ERC.

MATERIAL AND METHODS

The study group comprised consecutive adult patients (age ≥18 years) with respectable GBC who underwent open RC or ERC over 10 years duration between March 2013–2023. The work has been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). The ethical approval for the study and patient consent were duly waived by the Institutional Ethics Committee.

The staging investigations included multiphasic contrast-enhanced computed tomography (CT) scan of the abdomen and pelvis with a CT scan of the chest. In patients who presented with jaundice, an additional magnetic resonance cholangiopancreatography (MRCP) was also performed to evaluate biliary anatomy. A PET-CT scan was not routinely performed except for patients who presented with equivocal findings on a CECT scan of the abdomen. Routine haematological and biochemical blood tests and tumour markers, including CA 19.9 & CEA, were performed for all patients. Biopsy confirmation was only performed who were deemed unresectable after staging investigations or were planned for neoadjuvant chemotherapy. For patients presenting with incidental gallbladder cancer, biopsy of the resected gallbladder was reviewed at the authors’ institution.

Briefly, the steps of standard RC included an initial diagnostic laparoscopy to rule out peritoneal metastases and liver surface metastases. This step was omitted in patients who presented with prior laparoscopic cholecystectomy and incidental GBC. At subsequent laparotomy through right subcostal incision, mobilisation of the second part of the duodenum from its peritoneal attachments (Kocherization), intraoperative frozen section (FS) analysis of intercortical (station 16) lymph nodes was performed. In patients with positive FS (indicative of metastatic (M1) disease), surgery was not proceeded with. In patients with well-defined polyps ≥ 1cm, a simple cholecystectomy with FS was done. The remaining patients underwent en bloc resection of the gallbladder with wedge resection of the gallbladder fossa. For liver resection, the Pringle manoeuvre and Cavitron Ultrasonic Surgical Aspirator were used as adjuncts. Lymphadenectomy included excision of lymph nodes along the common hepatic artery (station 8), hepato-duodenal ligament (station 12c, 12b,12a, and 12p), and posterior superior pancreaticoduodenal (station 13). The common bile duct excision with hepaticojejunostomy was performed in patients where the cystic duct margin was positive (at frozen section or in patients with incidental GBC) or for indications such as the presence of a choledochal cyst.

The demographic data, operative parameters, postoperative outcomes, and histopathology details of each patient were accessed from the institutional electronic health record system and were recorded in a specifically prepared Microsoft Access database. The other parameters that were recorded were postoperative complications as per Clavien-Dindo grades. Minor complications included grade I & II, while severe complications included grade III, IV & V. Also recorded were re-exploration, readmission, and 30-day mortality, and procedure-specific complications such as surgical site infection and bile leak. Postoperative length of stay was recorded and did not include the day of operation, while including the day of discharge. Vega et al. used post-surgery mortality at 90-days for benchmarking.^[11] In the current study, 30-day mortality has been taken into consideration as it is standard practice in the department. Moreover, in our view, due to logistical considerations, 30-day mortality is more practical and feasible in vast countries such as India.

Re-exploration was defined as any unplanned return to operation theatre during index hospitalisation or within 30 days of RC. The relevant histopathology findings, including the type of tumour, resection margin status (R0 defined as surgical margin >1 mm free of tumour), the number of lymph nodes evaluated and involved by the tumour, were also recorded. For this study, the seventh edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual was used. This is due to the fact that information required for the classification of T2a and T2b as per the AJCC eighth edition was not available for patients who underwent surgery prior to 2017. Further, Vega et al. have also used the seventh AJCC cancer staging manual in their study.^[11]

Follow-up data was recorded from the electronic health record system and a telephonic interview.

Benchmark values were calculated in 8 domains of perioperative and oncologic outcomes for all patients. Patients were then divided into two groups according to the surgical procedure performed, i.e., standard RC and extended RC. The median values for each of the 8 parameters were recalculated in the RC and ERC groups separately.

RESULTS

Over a 10-year period, 51 patients underwent GBC surgery with curative intent. Of the 51 patients, 12 (23.5%) with peritoneal metastases were excluded from further analysis. The study group thus comprised the remaining 39 patients (10 males: 29 females) median age of 57 (range 41 – 74) years, who underwent RC or ERC. Of the 39 patients, 36 (92.3%) had GBC involving the fundus or body of the gallbladder. Nine patients (23%) had incidental GBC, and 8 (20.5%) patients underwent extended RC - right hepatectomy with common bile duct resection (01), common bile duct resection alone (06), and segmental colectomy (01). Complications of any severity occurred in 12 (30.7%) patients, whereas a major complication (Clavien-Dindo grade ≥III) occurred in 1 (2.6%) patient. This patient needed image-guided drain placement for an intra-abdominal collection. Seven (17.9%) patients had surgical site infection during the postoperative period. The readmission within 30-days was needed for 2 (5.1%) patients, while no patient needed reparative surgery. There was no 30-day mortality.

An overview of demographic data, operative procedure performed, and outcomes has been provided in Table 1.

During RC, wedge resection was the preferred method of hepatectomy. Indications for common bile duct resection (n = 6) included - incidental GBC (4), choledochal cyst with GBC (1) and positive frozen section margin (1). The operative details and post-operative complications have been summarised in Table 2.

Of the 39 patients, margin-negative (R0) resection was achieved in 37 (94.87%) patients. The median number of lymph nodes retrieved was 8 (range 1–20). Lymph nodes ≥4 and ≥6 were retrieved in 36 (92.3%) and 34 (87.2%) patients, respectively. Following histopathology examination, the final TNM staging was I (7), II (16), IIIA (5), IIIB (10), and IVA (1). Following the multidisciplinary meeting, all patients with GBC stage II or above were advised to undergo adjuvant chemotherapy. Chemoradiation was advised for the 2 patients who had RI margin status. However, of the 32 patients who were candidates for adjuvant treatment, only 23 (71.8%) patients enrolled for the treatment.

Of the 39 patients, 5 (12.8%) were lost to follow-up. The median overall survival for the entire cohort with Kaplan-Meier survival analysis was 48 months at a median follow-up duration of 36 months (6–72 months).

DISCUSSION

Radical cholecystectomy is a technically challenging, complex procedure that is associated with high morbidity and requires considerable expertise for optimal outcomes. Our results suggest that at tertiary care centres, RC can be performed with good perioperative (low overall/serious complications by ClavienDindo system) and oncologic outcomes (median number of lymph nodes retrieved 8, and R0 margin status of 95%).

With there being a reasonable consensus on the technical aspects of RC, the current focus is on standardisation of outcomes of the procedure. Recent studies, such as HQS and TOLS for GBC, have addressed this issue.^[8,9] The main limitation of these studies is that they address isolated aspects of surgical quality for GBC such as oncologic outcomes in HQS and only liver resection outcomes in the TOLS study. Further hepatectomy during RC comprises WR or AR, which, by Brisbane nomenclature, is classified as a minor hepatectomy.^[12]

For the purpose of this study, we compared our RC outcomes to benchmarks proposed in a recent multicentre study by Vega et al.^[11] Benchmarks are comprehensive, best achievable perioperative and oncologic outcomes in 8 domains in patients who are at low risk for complications. In our study, benchmark values were achieved in all 8 domains in the low-risk patients, therefore, implying wider applicability of the proposed benchmarks in the real-world single-centre setting. In the ERC group, benchmark values were achieved for oncologic parameters (lymph node number & margin status) and postoperative length of stay and blood loss. Expectedly, complications, perioperative blood transfusion rates, and duration of surgery were higher than the benchmark values in the ERC group of patients. Our study, therefore, suggests that proposed benchmarks may help standardise RC outcomes and may emerge as a meaningful tool for comparison of outcomes between centres.

Further, in the study by Vega et al., only 27% patients were included in the low-risk benchmark group.^[11] In comparison, in the current study, 64% patients were in the low-risk cohort; this could be due to the younger patients (median age 64 vs. 57 years) with fewer co-morbidities. Incidental GBC was lower in the current study (23%) in comparison to that by Vega et al. (64%).^[11] A possible explanation could be that GBC is endemic in India, and patients present with more advanced disease. Therefore, we believe that centres performing RC in endemic regions should adhere to the higher standards of oncologic surgery.

Benchmarks have been conceptualised as best achievable outcomes that may serve as a point of reference for comparison and an eventual improvement in surgery outcomes. The benchmark value cut-off is set at the 75th percentile of the medians, so that these values can be attained and are not too restrictive.^[13,14]

We believe that one of the issues arising out of the benchmark study by Vega et al. is what the quality parameter should be for lymph node number retrieved during RC. Benchmark value for lymph nodes retrieved has been proposed as ≥4; four which corresponds to the 25th percentile. However, recent studies on the subject have proposed at least 4 and recommend 6 lymph nodes to be evaluated for sufficient nodal staging.^[8,11,15.16] In our patient cohort, the median number of lymph nodes retrieved was 8, with ≥6 lymph nodes retrieved in 34/39 (87.2%) of patients. Therefore, we propose that ≥6 lymph nodes may be a more appropriate benchmark value at high-volume endemic centres.

The benchmark values for surgery for low-risk GBC were proposed following a study by 13 high-volume centres. This study, however, does not have data from India, a high incidence region for GBC. Our study from a high-volume centre in an endemic region in India is indicative of the wider applicability of benchmarks for GBC. In the present study, all patients underwent surgery by an open approach. This has been reported to be one of the factors to achieve an optimal outcome.^[8] Another strength of the current study is that all patients have been operated upon by the same team of surgeons, utilising standardised surgical procedures and contemporary surgical practices.

In our study, 12 (23.5%) patients had peritoneal metastases on diagnostic laparoscopy; these findings are comparable to another study from India, where diagnostic laparoscopy upstaged the disease in 23.2% patients.^[17] This indicates the limitations of current cross-sectional imaging modalities, where peritoneal deposits may remain undetected.

It may be argued that one of the limitations of the current study is the relatively small number of patients. This could be due to the fact that GBC is an aggressive cancer, and the vast majority of patients are found to have metastatic disease at the time of presentation. Thus, only a few patients are candidates for RC. It is noteworthy that our GBC surgery numbers compare well with those of other studies with similar time periods.^[11,18] One of the limitations of the study is that it is a single-centre study, and the wider applicability of benchmarks needs to be further examined in multicentre studies.

TAKE HOME MESSAGE

In conclusion, our study suggests that benchmarks provide a valuable tool for standardisation/evaluation of perioperative and oncologic outcomes following RC. Further, we propose that ≥6 lymph nodes may be a more appropriate benchmark value at high-volume endemic centres.