Preoperative staging should be aimed not only at the exclusion of distant metastases but also at the assessment of local extent of disease. Cross-sectional, contrast-enhanced imaging (computed tomography [CT] and/or MRCP) is the mainstay of investigation. Multi-phasic CT of the chest, abdomen, and pelvis to include portal venous and arterial phases should be used to assess the extent of disease in the liver and porta hepatis, while also evaluating for metastatic disease.

Imaging should include thin cuts through the liver and porta hepatis to elucidate detailed relationships between the tumor and porta hepatis structures. With respect to the local assessment of disease, one study of 118 patients with gallbladder carcinoma found CT to be 79% accurate for differentiating T1 versus T2 tumors, 93% accurate for differentiating T2 versus T3 tumors, and 100% accurate for differentiating T3 versus T4 tumors [1]. Further, the overall accuracy improved from 72% to 85% when multiplanar reconstructions were added to conventional axial imaging [1].

MRCP with intravenous contrast is a valuable radiologic modality to assess the extent of the primary tumor. Specifically, analyses of MRI for the assessment of gallbladder carcinoma have shown sensitivities of 70–100% for hepatic invasion, 100% for vascular involvement, and 75% for lymph node metastases [2, 3]. That being said, it remains unclear and relatively unstudied as to whether MRI has added benefit to that of CT. The two studies are likely complimentary and should be used in such a manner. Lastly, selective use of hepatic duplex ultrasound imaging can add valuable information in patients with locally advanced tumors. Our experience has found duplex ultrasonography to act as an excellent adjunct to assess the extent of transmural tumoral invasion into hepatic parenchyma or biliary structures (87% accuracy), while simultaneously assessing for involvement of portal venous or hepatic arterial structures, which can be challenging to assess on cross-sectional imaging [4]. It should be noted that initial imaging studies should be performed prior to biliary stenting (if it is to be performed), as stenting will cause local inflammation, making assessment of tumor extent difficult.

Additionally, FDG-PET imaging may be helpful in the identification of distant disease, as depicted by retrospective reviews. In a study of 61 patients with biliary tract malignancies, PET/CT had a sensitivity of 100%, as compared to 25% for CT alone, (p < 0.001) in the identification of distant metastases [5] Moreover, PET results changed surgical management in 17% of cases [5]. In an analysis of 41 patients with gallbladder carcinoma at Memorial Sloan Kettering Cancer Center (MSKCC), preoperative PET results altered surgical management in 23% of patients (for either the initial operation or re-resection after an incidental finding following cholecystectomy) [6]. Lastly, in a recent analysis of the efficacy of PET imaging as compared to CT and MRI, PET identified occult distant metastatic disease and also proved findings suspicious on CT to be negative; therefore altering surgical decision-making in 17% of patients [7]. Although the evidence is limited to small retrospective reviews, one should consider PET imaging when evaluating newly diagnosed, locally advanced gallbladder carcinoma, largely to evaluate for distant metastatic disease or to confirm/refute questionable findings on cross-sectional imaging.

Although uncertainty remains as to whether resection offers oncologic benefit for patients with locally advanced gallbladder carcinoma, the possibility of long-term survival after complete resection has been shown. As a baseline, documented 5-year survival rates for resected gallbladder carcinoma range from approximately 20–50% [8–16]. Specifically, in a review of 410 gallbladder carcinoma patients who presented to MSKCC between 1986 and 2000, the median and 5-year survivals for resected patients were 26 months and 38% respectively, with 14% of T3 patients and 19% of T4 patients alive at 5 years [8].

Yet, until recently, the optimal extent of resection was unknown and often debated. However, an analysis of the impact of the extent of resection on disease-specific survival (DSS) in 104 patients with gallbladder cancer at MSKCC found tumor biology and stage, rather than the extent of resection, to predict DSS. Empiric major hepatic resections and bile duct excision for early-stage tumors resulted in higher morbidity, and were not associated with prolonged survival as compared to lesser hepatic resections. However, major hepatic resections, including extended hepatectomy and CBD excision, are appropriate when necessary to clear disease and achieve negative margins. These larger resections, when necessary, were associated with acceptable long-term survival [9]. Therefore, the extent of resection should be dictated by what is necessary to achieve a negative margin.

Additionally, regional (portal) lymphadenectomy should be performed for accurate staging information [17, 18]. The chance of nodal involvement increases with increasing T stage. Bartlett et al. found nodal disease was associated with 46% of resected T2 tumors and 54% of resected T3 tumors [14]. Others have found that on progression of T stage from T2 to T4, nodal and distant metastases increased from 16% to 79% and from 33% to 69%, respectively [8]. Although the impact of node dissection on survival is controversial (rare 5-year survivors with N1 disease), the diagnostic information gained regarding node positivity may help in determining adjuvant therapies [19–23].

As stated, if resection is undertaken, the goal is the achievement of a pathologically margin negative (R0) resection. In an analysis of 135 patients subjected to definitive resection following an incidentally diagnosed gallbladder carcinoma, the presence of residual disease at any site was associated with significantly worse survival [24]. The median disease-free survival (DFS) (11.2 vs. 93.4 months, p < 0.0001) and disease-specific survival (DSS) (25.2 months vs. not reached, p < 0.0001) were dramatically lower than in patients without residual disease [24]. Moreover, residual disease identified at any particular site predicted DFS (HR 3.3, 95% CI 1.9–5.7, p = 0.0003) and DSS (HR 2.4, 95% CI 1.2–4.6, p = 0.01) and was independent of all other tumor-related variables [24]. In essence, survival among patients with residual disease at any site was not significantly different than those with stage IV disease [24]. Importantly, and relevant to our discussion of resecting T3/T4 lesions, the T stage of the gallbladder specimen was the only independent predictor of residual disease (T1b = 35.7%, T2 = 48.3%, T3 = 70%; p = 0.015) [24].

Except for the uncommon patient with concomitant common bile duct stones, patients with gallbladder carcinoma presenting with obstructive jaundice have tumor involvement of the porta hepatis by either direct extension of the tumor, diffuse invasion of the porta hepatis, or extensive nodal disease. These patients are rarely resectable and have an associated poor prognosis [25]. Therefore, upfront attempts at resection are hard to justify. Although theoretically, local extension should not necessarily affect oncologic outcome, studies have found jaundice to be an indicator of advanced malignancy and portends a poor survival.

An analysis of 240 patients with gallbladder carcinoma from Memorial Sloan Kettering Cancer Center compared oncologic outcomes between those who presented with obstructive jaundice and those who did not [25]. Overall, 34% (82/240) of patients within the cohort presented with obstructive jaundice [25]. Causes of obstructive jaundice included CBD involvement in 89% (73/82), common hepatic duct involvement in 9% (7/82) and obstructing gallstones in 2% (2/82) [25]. Among the jaundiced patients, 67% (55/82) underwent operative exploration [25]. Patients were excluded from exploration for the following reasons: 14 radiographically documented liver metastasis, 6 radiographically unresectable invasion of the portal vein, and 12 with peritoneal metastases [25]. In the 55 patients who were explored, diagnostic laparoscopy was performed in 45% (25/55) and found peritoneal metastases in 68% (17/25), precluding further exploration [25]. Ultimately, exploratory laparotomy was undertaken in 37 patients. Among these 37 patients, distant peritoneal and liver metastases were found in 8 (22%) patients, and locally unresectable disease in the porta hepatis was found in an additional 10 (27%). In summary, 52% (19/37) of jaundiced patients who underwent exploratory laparotomy were resected with curative intent, however this represented only 7% (6/82) of all patients presenting with jaundice.