Chapter 5. Laparoscopic Staging and Approaches to Cancer

The role of laparoscopy in the staging of gastrointestinal malignancy has continued to evolve over the last decade. Improvements in noninvasive diagnostic modalities have led to a more selective approach being adopted. Nonetheless, minimally invasive surgical techniques for staging and palliative bypass continue to play an important role in the staging and management of patients with upper gastrointestinal malignancies.

As the multidisciplinary management of gastrointestinal cancer has evolved over the last decade, an accurate extent of disease workup has become essential to treatment planning. Staging procedures should accurately define the extent of disease, direct appropriate therapy, facilitate the use of adjuvant therapies and avoid unnecessary interventions in a safe and cost-efficient fashion.

Recent advances in radiology have provided many noninvasive tools, such as multidetector computed tomographic (CT) scanning, magnetic resonance imaging (MRI) and combined CT with positron-emission tomographic (CT/PET) scanning, that have had a considerable impact on the extent of disease workup. Unfortunately, these modalities may underestimate the extent of disease, with small-volume metastatic disease being appreciated only at open surgical exploration. For over 100 years, laparoscopy has been suggested as a means for identifying such small-volume disease. Recently, a significant amount of data has been produced to suggest that the use of laparoscopy and laparoscopic ultrasound (LUS) in the staging of gastrointestinal malignancies has an impact on overall management.1–7 The aim of laparoscopic staging (LS) is to mimic staging at open exploration while minimizing morbidity, enhancing recovery, and thus allowing for quicker administration of adjuvant therapies if indicated. Proponents believe that LS should be viewed as complementary and not as a replacement for other staging modalities such as CT scanning, MRI, or PET scanning. In simplistic terms, the advantages of laparoscopy are that it allows the surgeon to visualize the primary tumor, determine vascular involvement, identify regional nodal metastases, detect small-volume peritoneal/liver metastases, and obtain tissue for histologic diagnosis.

Laparoscopic staging can be performed immediately before a planned open procedure or at a separate occasion. We have moved to the latter approach in the main because of logistical concerns around the availability and utilization of operating time. Generally the procedure is performed as an ambulatory/outpatient procedure with excellent patient satisfaction.

Laparoscopic staging usually is performed under general anesthesia with the patient positioned supine on the operating table. A warming blanket is placed underneath the patient, who is secured appropriately to the table with padding over the pressure points.

The following operative equipment is considered necessary for the procedure:

1. A 30-degree angled laparoscope either 5 or 10 mm in diameter

2. Five-millimeter laparoscopic instruments, including a Maryland dissector, a blunt-tip dissecting forceps, a cup/biopsy forceps, atraumatic grasping forceps, a liver retractor, and scissors

3. A 5- or 10-mm suction/irrigation device

4. An LUS probe (optional)

In general, we prefer a multiport technique. Access is gained into the peritoneal cavity using a blunt port placed subumbilically by direct cutdown. By using forceps to grasp the fascial layers, retractors can be avoided and the wound size minimized. An alternative approach, particularly in patients with previous midline incisions, is to place the initial port in either the right or the left upper quadrant of the abdomen. Many surgeons prefer to use a Veress needle to achieve pneumoperitoneum prior to placing the surgical ports. In this case, care should be exercised to avoid visceral or vascular injury. Laparoscopic access using an optical trocar, which combines the advantages of the Hasson and Veress techniques, is a safe and feasible primary insertion method, which may alleviate this risk and is becoming an increasingly accepted technique.8

Pneumoperitoneum is achieved with CO2 gas. Insufflation commences at a low flow rate until peritoneal entry is confirmed. An intraperitoneal pressure of 10–12 mm Hg is considered optimal. However, in patients with cardiopulmonary compromise, a lower maximum pressure may be chosen. A 5- to 10-mm 30-degree angled telescope is preferred, and systematic examination of the peritoneal cavity is performed. Additional trocars then are inserted under direct vision. Placement depends on the site of the primary tumor (ie, colonic, gastric, pancreatic, etc) and the findings at initial inspection (ie, whether obvious metastatic disease is present). In general, ports are placed along the planned open incision line (Fig. 5-1).

Following port placement, a detailed examination of the peritoneal cavity is performed in a similar fashion to an open exploration. The primary tumor is assessed. Any extension into contiguous organs can be identified. Following an initial survey, a systematic examination of the intra-abdominal viscera is performed commencing with the liver. To facilitate hepatic examination, the patient is placed in a 20-degree reverse Trendelenberg position with 10 degrees of left lateral tilt. The anterior and posterior surfaces of the left lateral segment of the liver are examined, followed by examination of the anterior and inferior surfaces of the right lobe. Despite the absence of tactile sensation, indirect palpation of the liver surface can be achieved by using two instruments (Fig. 5-2). A blunt suction device is particularly useful in compressing the liver tissue in order to detect small metastases. Improved visualization of diaphragmatic and posterior surfaces may be achieved by placing the camera in the right upper quadrant port. Any suspicious areas can be biopsied at this point. A cup biopsy forceps is the preferred instrument for obtaining adequate biopsies for diagnostic purposes. For this, we use a 5-mm biopsy forceps with a 2-mm cup as standard. Multiple samples may be taken to increase diagnostic yield. The cup is used to breech the liver capsule and a bite is taken out of the lesion. Further scoops can then be taken from the lesion and liver parenchyma as needed. Thorough hemostasis can easily be obtained with electrocautery or use of argon beam diathermy. If electrocautery is used, it is important to avoid direct coupling or capacitance coupling, which can lead to visceral injury. Direct coupling, when current flows directly from one instrument to the other, may occur when the instruments are too close together, especially if one is just outside of the field of view. Capacitance coupling occurs when two conductors have an insulator sandwiched between them. The high frequency AC current in the active conductor generates a magnetic field, which then induces current in the second conductor. Mixing of metal and plastic instruments and ports can lead to capacitance coupling and, at least in theory, severe burns. The incidence of complications is reduced by limiting the gain of electrocautery to 30 W and possibly by using plastic rather than metallic ports.

The hilus of the liver, hepatoduodenal ligament, and foramen of Winslow then are examined. Any abnormal lymphadenopathy can be identified. The suspicious node can be either excised or biopsied using the cup forceps. As in open surgery, care must be taken not to crush the node and possibly disseminate tumor cells during this procedure. In general, the duodenum is not mobilized. However, for patients with pancreatic or common bile duct tumors, close attention is paid to the presence or absence of tumor infiltration in the angle between the duodenum and the lateral aspect of the common bile duct because this may indicate significant vascular involvement.

The patient then is repositioned into a 10-degree Trendelenberg position without lateral tilt to facilitate examination of the transverse mesocolon and retroperitoneum. The omentum is retracted toward the left upper quadrant, elevating and enabling inspection of the transverse mesocolon an d the ligament of Treitz. The mesocolon is inspected carefully with particular attention to the middle colic vein, which usually is visible. Any abnormal adenopathy or infiltration (Fig. 5-3) around the middle colic vein is noted and may be biopsied. For patients with an upper gastrointestinal primary tumor, the lesser sac is examined. To facilitate this maneuver, the patient then is returned to a supine position, the left lobe of the liver is elevated, and the gastrohepatic omentum is incised (Fig. 5-4). This exposes the caudate lobe of the liver, the inferior vena cava, and the celiac axis. If present, an aberrant left hepatic artery should be identified and preserved. Often, adhesions between the stomach and the pancreas require division to allow entry into the lesser sac. By elevating the stomach, the “gastric pillar” can be clearly identified (Fig. 5-5). This “pillar” contains the left gastric artery and vein. This structure followed down leads us to the celiac axis, and any suspicious nodal tissue can be biopsied. The hepatic artery also is identified and followed to the hepatoduodenal ligament. The anterior aspect of pancreas, hepatic artery, and left gastric artery is also seen. Any suspicious periportal, hepatic, or celiac nodes can be biopsied.

The diagnostic yield for LS may be increased by performing peritoneal lavage cytology. In general, the specimens are taken at the start of the laparoscopy to avoid potential contamination following tumor manipulation or dissection. Between 200 and 400 mL of normal saline is instilled into the peritoneal cavity. The abdomen is agitated gently before aspiration. In pancreatic cases, samples are taken from the right and left upper quadrants. An additional sample is taken from the pelvis in patients with gastric cancer. Informing the pathologist/cytologist of the procedure timing and clinical question often leads to better clinical yields and is advisable prior to undertaking the laparoscopy.

If available, LUS can be performed at this stage. Laparoscopy by its nature is a two-dimensional modality, with the result that appreciation of deep or subsurface lesions in solid organs is often suboptimal. LUS can partially overcome this deficiency. Transducers in clinical use employ either curved or linear-array technology and have a high-frequency performance with a range in the region of 6–10 MHz, allowing for high-resolution images to be obtained that can detect lesions from 0.2 cm in size. In addition, Doppler flow capability, if present, allows for accurate vessel identification and facilitates assessment of the tumor-vessel interface. The LUS probe is inserted via a 10- to 12-mm port, usually in the right upper quadrant.

The LUS is an invaluable tool for examination of the liver. Initially, the transducer is placed over the left lateral segment (Fig. 5-6), allowing assessment of segments I, II, and III. It is important that the probe is placed in direct contact with the liver surface to maximize acoustic coupling. Examination of the right lobe commences with the probe on the dome of the liver. The vena cava is visualized at the back and as the probe is moved forward slowly to identify the hepatic and portal veins. Within the liver, these can be identified by virtue of their surrounding fibrous sheath. The remaining hepatic segments (IV, V, VI, VII, and VIII) are examined by rotating the probe over the rest of the liver. Suspicious lesions can be biopsied either by fine-needle aspiration (FNA) or by percutaneously inserting core biopsy needles under LUS guidance. With the probe over segment V, the gallbladder is assessed, and with transverse placement of the probe over the hepatoduodenal ligament, the common hepatic duct, common bile duct, and hepatic arteries along with the portal vein can be identified (Fig. 5-7). The portal vein can be followed to its confluence with the splenic and superior mesenteric vein. The superior mesenteric artery also can be seen and its relationship to a pancreatic tumor, if present, determined. The pancreas can be examined, and any lesion can be identified.

Esophageal cancer is the eighth most common cancer and is the sixth leading cause of cancer death worldwide.9 Esophageal cancer was diagnosed in 16,470 new patients in the United States in 2008, with an overall annual incidence of approximately 5.4 cases per 100,000 population.10 It is estimated that more than 14,000 patients will die of this disease each year. Unfortunately, the prognosis remains poor; with an overall survival rate of approximately 5–10% in spite of the availability of new chemotherapeutic and biologic agents in both neoadjuvant and adjuvant settings. Surgical resection remains the treatment of choice for patients with localized disease. In addition, in the last few years, there has been a significant progress in palliative nonsurgical treatment options. Therefore, accurate staging for esophageal cancer is of paramount importance.7,11–14

Common diagnostic modalities are listed in Table 5-1. The results of a meta-analysis in 2008 suggest that endoscopic ultrasonography (EUS), CT, and fluorine-18-flurodeoxyglucose (FDG)-PET each play a distinctive role in the detection of metastases in esophageal cancer patients. For the detection of regional lymph node metastases, EUS is most sensitive, whereas CT and FDG-PET are more specific tests. For the evaluation of distant metastases, FDG-PET has probably a higher sensitivity than CT.15 These have been discussed in detail elsewhere in this book. Endoscopy remains the diagnostic gold standard. Biopsies can be obtained and an assessment of local disease extent made. In patients considered unsuitable for surgical resection, a number of palliative options such as endoscopic dilation, laser ablation, or placement of luminal stents exist.

Multislice CT scanning of the thorax and abdomen is the radiologic staging modality of choice. The primary tumor can be visualized and metastatic disease detected. However, while data suggest that current-generation high-resolution multislice CT scanning is of significant value, its capacity to accurately T stage the disease and predict lymphatic and peritoneal spread remains between 65% and 80%.16

EUS enables detailed imaging of the esophageal wall, local lymph nodes, and contiguous structures, making it the ideal tool for tumor node metastasis (TNM) staging.11,12,17 The shape, pattern, and demarcated borders of nodes are examined to assess metastatic potential.18,19 EUS appears superior to CT scanning for locoregional staging.13,20 Harewood and Wiersema from the Mayo Clinic compared the cost of EUS-FNA with CT-FNA and a surgical approach in staging patients with nonmetastatic esophageal cancer. They suggested that by avoiding unnecessary surgery, primarily by detecting celiac node involvement, EUS-FNA was the least costly strategy.21

It appears that combined CT scan and EUS is a better prediction of tumor resectability than CT scan alone (81% vs 65% with p < 0.05) reported by de Graaf et al.22 In a study of the impact of EUS-FNA in the management of patients with esophageal cancer, Morris et al found that EUS-FNA altered management in 28 (67%) patients and appeared to help direct patients toward appropriate treatment strategies including palliative and neoadjuvant therapies.23 In a meta-analysis and systematic review of studies that included over 2500 patients, Puli et al24 concluded that EUS performs better with advanced (T4) than early (T1) disease and that FNA substantially improves the sensitivity and specificity of EUS in evaluating N stage disease (from 84.7% [95% CI: 82.9–86.4] to 96.7% [95% CI: 92.4–98.9]). However, while most thoracic surgeons have embraced EUS-FNA as the most accurate locoregional staging modality in esophageal cancer, this attitude is not fully reflected in utilization patterns due to a lack of quality EUS services in some centers.25

Several studies have investigated the detection of the primary tumor by FDG-PET. Increased uptake of FDG was seen in 68–100% of the esophageal tumors.26–28 Undetected tumors are mostly stages T1 and T2. T1a tumors, remaining within the submucosa, are especially difficult to detect by FDG-PET.29–30 Kato et al31 found a significant relationship between the intensity of the primary tumor FDG-uptake, expressed as SUV, and the depth of the tumor invasion. However, Flamen et al32 found no correlation between SUV and pT-stage.

To determine whether FDG-PET could delineate patients with esophageal cancer who may not benefit from esophagectomy after chemoradiotherapy, Monjazeb et al reviewed 163 patients with histologically confirmed stage I to IVA esophageal cancer receiving chemoradiotherapy with or without resection with curative intent and found that patients who achieved a complete response on FDG-PET imaging may not benefit from added resection given their excellent outcomes without resection.33 These results should be validated in a prospective trial of FDG-PET-directed therapy for esophageal cancer.

It has been suggested that FDG-PET scanning has a role for the detection of metastatic disease and for restaging after neoadjuvant therapy or evaluation of recurrence. In the study by Flamen and colleagues, FDG-PET scanning had a significantly higher rate of detection of stage IV disease compared with the combination of CT scanning and EUS. It upstaged disease in 15% and downstaged disease in 7% of patients.32 Other studies have reported similar results.34–36

In relation to the role of FDG-PET/CT in tumor delineation for radiotherapy, only three studies have reported a significant positive correlation between FDG-PET-based tumor lengths and pathological findings and so the authors of a systematic review on the role of FDG-PET/CT in tumor delineation and radiotherapy planning in patients with esophageal cancer concluded that standard implementation of FDG-PET/CT into the tumor delineation process for radiation treatment seems unjustified and needs further clinical validation first.37

Despite this increasingly sophisticated diagnostic armamentarium, between 15% and 20% of patients will continue to have radiologically occult peritoneal, nodal, or liver metastases detected at surgical exploration. Laparoscopy has been suggested as a means to detect such disease and thus exclude this cohort of patients from potentially ineffective treatment regimens.

The value of LS in esophageal cancer is accurate abdominal nodal staging and detection of occult distant mestastases. The procedure also allows for more detailed assessment of the tumor looking for serosal involvement, local invasion, or peritoneal cavity, liver, and omental disease. In a comparison of LS and EUS for esophageal cancer, Kaushik et al found an overall staging accuracy of EUS compared with LS of 72%.38 Staging differences were mostly reflected in distant metastases detected at LS (17%). The yield of LS appears to be determined at least in part by the site of disease, histologic cell type, and noninvasive stage. There are several observational studies reporting the usefulness of LS in both gastric and oesophageal cancers, the largest and most recent of which includes 416 consecutive patients undergoing staging laparoscopy.39 The authors report an 88% sensitivity of laparoscopy for resectability, with avoidance of unnecessary laparotomy in 20.2% of all patients. Staging laparoscopy was most useful in patients with adenocarcinoma, distal oesophageal, and gastro-esophageal cancer, with percentage change in treatment decision of 21.9%, 17.1%, and 17.2%, respectively. No patients in this study with upper two-third lesions had their treatment decision changed by staging laparoscopy. This would be in accordance with the general trend in the literature that the more distal the tumor in the esophagus, the greater the risk and likelihood of intra-abdominal metastases40,41 and this is likely related to lymphatic anatomy.

In a well-designed study, Samee et al report that the addition of LUS in the staging of esophagogastric cancers increases the detection rate of metastasis by 8% but that there is little impact on the false-negative rate.42 In their retrospective case series of 320 patients, LUS proved most useful in detecting metastatic lymphadenopathy beyond the limits of curative resection and liver metastasis. The main benefit appears to be in the assessment of nodal disease, particularly in the celiac axis, hepatoduodenal ligament, and para-aortic area as disease in these sites accounts for more than 40% of the positive findings at laparoscopy.

The combination of endoscopic and laparoscopic ultrasonography (EUS-LUS) is accurate for resectability assessment of patients with esophageal cancer. In a series of 256 consecutive esophageal cancer patients, Mortenson et al demonstrated a statistically significant survival difference (p < 0.01) between the different TNM stages and resectability groups predicted by a EUS-LUS combination.43 The poor prognosis for the patients with irresectable or disseminated disease was accurately predicted by EUS and LUS.

The yield of LS appears to be determined at least in part by the site of disease, histologic cell type, and noninvasive stage. In an earlier review of 369 patients with carcinoma of the distal esophagus or gastric antrum, Dagnini and colleagues demonstrated occult disease in 33% at laparoscopy in patients with adenocarcinoma of either the distal esophagus or gastric cardia.44 However, LS had a minimal impact for patients with squamous cell cancers in the upper third of the esophagus, changing management in only 3.5% of cases. Stein and colleagues reported similar results. At laparoscopy following radiologic staging, they found that 25% of patients with locally advanced (T3/T4) adenocarcinoma of the distal esophagus or gastric cardia had peritoneal or liver metastases.45 Thus, for patients with squamous cell carcinoma of the esophagus, we believe that LS is not indicated in the absence of suspicious intra-abdominal imaging findings.

The overall incidence of gastric cancers is declining; however, there has been a relative increase in the incidence of tumors of the esophagogastric junction (OGJ) and gastric cardia. The peak incidence is in the seventh decade, and the disease is approximately twice as common in men as in women.46

Despite its apparent falling prevalence in the Western world, gastric cancer remains a significant public health problem and one of the leading causes of cancer death worldwide. The prognosis remains poor, with a current overall 5-year survival of 20%47 and 50–90% of patients dying of the disease within 2 years of diagnosis, even in those who have undergone a potentially “curative” resection.48–50 The poor outcome may be related in part to late presentation and inadequate staging and subsequent poor patient selection for surgery. Historically, following diagnosis and if medically fit, patients were subjected to open exploration for either resection or palliation. In a significant series of 916 patients in the mid-1990s, Pye and colleagues51 reported that 23% of the operations were exploratory alone in nature. However, with the recent development of multidisciplinary approaches to the disease, improved staging, and the establishment of less invasive palliative algorithms, the need for operative intervention has been questioned.52–54

Accurate staging is essential for patient selection. A sophisticated and complex diagnostic armamentarium exists. While upper gastrointestinal endoscopy and biopsy remain the primary diagnostic tools, with multislice contrast-enhanced CT scanning, EUS, MRI, and CT/PET scanning being used increasingly for preoperative staging, laparoscopy and LUS continue to have an important role in the staging algorithm for selected patients with gastric cancer (Fig. 5-8).

While the literature would suggest that despite currently available imaging modalities, LS will continue to detect small-volume metastatic disease in 20–30% of cases, the identification of occult nodal disease remains problematic. EUS appears somewhat better than CT in this regard. Wakelin and colleagues have reported an overall accuracy of EUS in nodal staging for proximal or orogastric junction tumors of 72%.55 If tumors that are nontraversable by endoscope are excluded, its accuracy increases by approximately 10%. Reported accuracy rates for laparoscopy and LUS vary from 60% to 90%. With LUS, direct biopsy of suspicious nodes can be obtained, which improves the utility of the modality. In distal gastric cancer, Finch and colleagues demonstrated an accuracy of 82% in T staging with the use of LUS.56 This compares favorably with other studies looking at the use of EUS (83%) or CT scanning (66%) for T staging distal tumors.57 In addition, the authors noted an accuracy rate of 89% for LUS in assessing lymph node status. In contrast, Wakelin noted that 38% of nodes were understaged. It would appear, therefore, that as with other ultrasound data, results are operator-dependent and reflect willingness or not to aggressively biopsy suspicious nodes.

While level I evidence does not exist for the use of LS in gastric cancer, a number of large single-institution studies have been carried out that allow us to make a number of conclusions regarding its role in the staging algorithm. As in esophageal cancer, laparoscopy will detect radiologically occult metastatic disease in a significant number of patients (Fig. 5-9). Muntean et al reported overall staging laparoscopy sensitivity for distant metastases of 89%, specificity 100%, and diagnostic accuracy 95.5%. The sensitivity for lymph node metastases was 54.5%, with a specificity 100% and a diagnostic accuracy 64.3%. The positive predictive value for resectability was 96% and the negative predictive value was 50%.58 Sotiropoulos et al reported that staging laparoscopy resulted in up staging 51.1% of patients, most commonly in the form of peritoneal seeding.59 As a consequence, the therapy planning was changed and laparotomy was avoided in 14 of these patients as the first operative procedure. Sensitivity of clinical staging was especially poor for stage IV tumors (5.3%) and for the majority of stage IIIB tumors (42.9%) in this particular study.

It has been suggested that, with more advanced radiological imaging, the value of staging laparoscopy will somehow diminish; however, the literature has not borne this out. Kim et al retrospectively measured the diagnostic performance of prospective computed tomographic (CT) results obtained by using 16- or 64-detector row scanners in the detection of peritoneal metastases (PMs) in patients with advanced gastric cancer.60 In 498 patients with T2 disease and above in a retrospective comparison of CT images with operative and pathological findings, a sensitivity and a specificity of 28.3% and 98.9%, respectively, were reported in scans demonstrating definite peritoneal deposits and 50.9% and 96.2% respectively in scans reported as equivocal. The authors concluded that even with modern CT techniques, the sensitivity for PM detection is limited. Similarly when evaluating preoperative local staging with 3D multidetector row CT, Chen et al reported that reconstructions yield significantly better overall accuracy than transverse images for tumor staging but not for lymph node staging.61 This highlights the need for a multimodality staging process, including EUS, LS, and LUS.

As mentioned earlier, we routinely take peritoneal washings for cytologic examination at the time of LS. Positive cytology obtained during peritoneal lavage at staging laparoscopy is information potentially available preoperatively that identifies a patient population at very high risk for early recurrence and death after curative resection of gastric cancer. Mehzir and colleagues recently reviewed a prospectively maintained database of 1241 patients with gastric cancer who underwent laparoscopy with peritoneal washings.62 Two hundred and ninety-one (23%) patients had positive cytology. A total of 48 of the 291 cytology-positive patients had repeat staging laparoscopy after chemotherapy. Compared with patients who had persistently positive cytology (n = 21), those who converted to negative cytology (n = 27) had a significant improvement in disease-specific survival (2.5 years vs 1.4 years, p = 0.0003). In an earlier publication by the same group on a lesser number of patients in this database, multivariate analysis identified preoperative T stage, preoperative N stage, site, and cytology as significant predictors of outcome. Positive cytology was the preoperative factor most predictive of death from gastric cancer (RR 2.7, p < 0.001).63 Although evaluated at laparotomy, La Torre et al reported similar results in their cohort of 64 patients.64 Eighty-six percent of patients with positive peritoneal lavage cytology had a pT3/pT4 tumor and 100% of those positive had an N-positive tumor (p < 0.001). The median survival of patients presenting with positive cytology was significantly lower than that of patients with negative peritoneal cytology (19 and 38 months respectively, p = 0.0001). Multivariate analysis of this group of patients also identified cytology as a significant predictor of outcome (p = 0.018). Looking at the value of staging laparoscopy in advanced gastric cancer, Shimitzu et al stratified 34 patients into groups according to the presence of peritoneal deposits and/or positive peritoneal lavage cytology.65 Those who were positive for both did not receive any operative intervention and were shown to do significantly worse over all, thus validating the argument for LS.

Taking the concept of intraperitoneal disease and its consequences even further, novel methods are being evaluated to increase the sensitivity of peritoneal lavage cytology. Wong et al have recently described a novel and very interesting method of detecting free peritoneal cancer cells in gastric cancer using cancer-specific Newcastle disease virus (NDV).66 The green fluorescent protein of NDV appears to specifically infect and detect peritoneal gastric cancer cells and offers a more sensitive method compared with conventional cytology. Results were particularly impressive in advanced disease. Of patients with M1 disease discovered during laparoscopy, only 50% were cytology positive. All, however, were NDV-GFP positive. Cytology was positive in 9% of patients with T3 disease, 8% with N1 disease, and 50% with N2 disease. In contrast, NDV-GFP was positive in 95% of T3 patients and 100% of patients with N1 or N2 disease. This novel modality may offer enhanced detection of intraperitoneal cancer spread and provide important prognostic information. The same group has also looked at reverse transcriptase polymer chain reaction to detect micrometastases in peritoneal washings with promising results67 and this reflects a growing area of research in gastric cancer staging today.

In this chapter, we have concentrated on the role of LS in determining unresectability. However, the advent of minimally invasive techniques applied to early gastric cancer has raised the possibility that LS may have an increasing role in treating that spectrum of the disease. It has been argued that gastric cancer is one of the most suitable targets for minimally invasive surgery (MIS) based on sentinel node status. Staging laparoscopy combined with sentinel node mapping may become a very important adjunct to laparoscopic local resection for curative treatment of sentinel-node-negative early gastric cancer.68 More work is required before the true utility of this approach is understood.

At present, surgical resection remains the most effective therapy for primary and metastatic disease of the liver. While there are no definitive criteria that define what constitutes resectable disease in part owing to differing therapeutic philosophies and surgical experience, most surgeons would consider extrahepatic disease, extensive bilobar disease, or the presence of extensive cirrhosis as the major factors that would preclude a potentially curative resection.

As with the other gastrointestinal malignancies, imaging modalities such as multidetector CT scanning, MRI, and CT/PET scanning are available for preoperative staging. Despite the use of these modalities, a significant number of patients continue to have exploration without resection.48–50,69–71 Laparoscopy, therefore, can serve to improve curative resection rates and decrease unnecessary laparotomy with its associated morbidity and quality-of-life issues.

Laparoscopic staging detects subradiologic disease in 10–60% of cases.72,73 As with other anatomical sites, variability in part relates to the completeness and quality of preoperative imaging. Jarnagin and colleagues from Memorial Sloan-Kettering Cancer Center (MSKCC) reviewed their experience with 186 patients who had either primary or secondary hepatic malignancies who underwent surgery at their institution.70 Laparoscopy was attempted in 104 patients and completed successfully in 85%. Overall, 26 (25%) of these patients were noted to have unresectable disease at the time of LS, and although nine patients had subsequent laparotomy for palliation, 17 patients were spared a laparotomy. More extensive hepatic disease, peritoneal disease, and extensive cirrhosis were the main laparoscopic findings that precluded resection. Difficulties were encountered determining the true extent of tumor vascular invasion or extensive biliary involvement. In addition, findings at laparoscopy had an impact on the type of resection performed in a further 10%. The authors also compared the patients undergoing LS with a similar nonrandomized cohort of 82 patients who did not receive LS but went directly to operation during the same time period. At open laparotomy, 28 (34%) of this group were noted to have unresectable disease. Although nine patients had a palliative procedure, 19 patients had only an exploratory procedure, which the authors suggested potentially could have been avoided with laparoscopy. Comparing the two groups, LS was associated with increased resectability rates (83% vs 63%), shorter hospital stay (8.6 vs 11.9 days), and reduced hospital charges. A subsequent study from the same institution analyzed experience with 401 patients.74 Prior surgery did not preclude staging because a complete laparoscopic examination was performed in 291 (73%) cases. Despite a false-negative rate of 22%, LS improved the overall resectability rate from 62% to 78%.

In an attempt to define the patients who would benefit from LS, the same group created a clinical risk score (CRS) based on five factors related to the primary tumor and the hepatic disease75,76 (Table 5-2). Each criterion was assigned one point. Thus, 42% of patients with a CRS score of greater than 2 had unresectable disease detected at laparoscopy versus 0% of patients with CRS scores of 0–1. Therefore, targeting laparoscopy to high-risk patients should avoid unnecessary LS in low-risk patients, whereas performing it in the high-risk group should prevent needless staging laparotomies and overall improve the yield from laparoscopy. This scoring scheme has recently been validated by a number of groups. Mann and coworkers noted that an increasing CRS correlated with the likelihood of detecting incurable disease. Management was altered in 0%, 14%, and 53% of cases if the CRS was 0–1, 2–3, or 4–5, respectively.77 Shah and colleagues reported that in patients with a CRS ≤2 laparoscopy and LUS prevented an operation in only 7% compared to 24% in those with a CRS >2.78 This data suggests that a focused use of LS is warranted.

Others have argued that improved imaging, particularly the increased availability of preoperative CT/PET, coupled with a more aggressive surgical approach have reduced the potential yield of LS in patients with colorectal metastases to the liver. However, data to support this hypothesis is lacking and in the main relies on review of the findings at open exploration, which were noted to preclude resection and potentially would have been detected by laparoscopy if it had been performed.

The history of a prior colectomy does not preclude accurate LS. Rahusen and colleagues performed laparoscopy in 50 patients with colorectal metastases, laparoscopy completing the examination in 94% and demonstrating unresectable disease in 38%.79 Similar results have been reported by others.80,81 Failure to accurately stage patients may occur due to adhesions from the prior open surgery in up to 20% of patients.80,81

Thaler and colleagues have suggested that the addition of intraoperative ultrasonography (IUS) improves the yield of LS (Fig. 5-10). In a review of 136 patients, LS/IUS changed the treatment plan in 48% of patients. Surgically untreatable disease was noted in 25% owing to PMs, nodal involvement, or diffuse hepatic disease.82 Others have also noted the added value of LUS to LS.73,79,83,84 Foroutani and colleagues reported their experience with LUS and biopsy in 310 patients with 1080 primary and metastatic liver lesions.85 Using a linear side-viewing transducer, core needle biopsies were taken using an 18-gauge spring-loaded biopsy gun. Histologic confirmation was obtained in all patients, with no bleeding complications or visceral injuries. A recent report has suggested that the combined use of laparoscopic and LUS-guided biopsies changed patient management in 27% of patients with upper gastrointestinal cancers (including primary and secondary liver tumors). LUS–guided biopsies were supplementary to laparoscopic biopsies and accounted for 44% of the clinical impact, suggesting that LUS should be an integral component of the staging procedure.86 Hartley and colleagues reported that LUS was equivalent to MRI in determining resectability, particularly for primary hepatic tumors.87 However, they and others also have noted that determining the extent of vascular and biliary involvement was problematic.83

For primary hepatocellular disease, experience is similar.88,89 Lo and colleagues performed staging laparoscopy with LUS in 91 patients with primary hepatocellular carcinoma (HCC), identifying unresectable disease in 16%, two-thirds of whom avoided any further surgical intervention and commenced nonoperative treatment earlier.71 As the use of laparoscopic resection and radiofrequency ablation for HCC has increased, it has been suggested that the role of LS has also expanded to not only identify unresectable disease but also aid selection of the optimal therapy. Lai and coworkers described a cohort of 122 patients with potentially resectable HCC who underwent LS prior to planned open laparotomy. Laparoscopic staging was performed in 119 patients, 44 of whom were noted to have unresectable disease. Overall 25% of patients had their therapy delivered laparoscopically (22 curative resection, 8 palliative ablation/resection) with a significant reduction in median hospital stay compared to those who underwent open surgery.90 In an interesting publication, Casaccia et al noted that, while laparoscopic ultrasonography accurately staged HCC in patients with advanced cirrhosis, it also allowed for laparoscopic radiofrequency ablation to be safely performed.91

The role of LS in the evaluation of noncolorectal nonneuroendocrine tumors was studied by D'Angelica and colleagues from MSKCC.92 Following preoperative staging, 30 patients considered to have resectable disease underwent laparoscopy. Staging was completed in 80% and correctly identified six patients of the nine finally found to have unresectable disease.

There has been little work specifically directed to determining the utility of LS for gallbladder cancer. Agrawal and colleagues reported on a cohort of 91 patients with apparent resectable disease who underwent staging laparoscopy. Laparoscopic findings of either locally advanced or disseminated disease avoided open exploration in 35 cases.93 Similar results were reported by Goere and coworkers from France who noted that LS detected unresectable disease in 36% of patients with potentially resectable biliary cholangiocarcinoma or gallbladder cancers. Peritoneal and liver metastases were detected but vascular and lymphatic extension was not diagnosed leading the authors to suggest that laparoscopy was more useful in extrahepatic cholangiocarcinomas and gallbladder cancer and should only be considered in selected cases with hilar cholangiocarcinomas.94

Adenocarcinoma of the pancreas remains a lethal disease.95 Despite increased awareness and improved diagnostic modalities, most patients continue to present with advanced disease at the time of diagnosis. Actual 5-year survival is between 3% and 5%, with surgical resection offering the only chance of cure. However, resection is only appropriate for a minority of patients. For the majority, the need for surgical intervention is controversial. In common with esophagogastric cancers, the notion that all patients require an operative procedure for accurate staging or palliation no longer is true. Our increased understanding of the natural history of the disease, coupled with the improvements in nonoperative palliative techniques, suggests that effective palliation does not require an open surgical procedure. Proponents of LS argue that the combination of dynamic contrast-enhanced CT scanning and/or MRI with laparoscopy remains the most effective means of staging, preventing needless open surgery for those who would not benefit, while not precluding resection for those who would benefit. Avoidance of unnecessary open procedures potentially will result in reduced perioperative morbidity and mortality, decreased hospital stay, shorter time to appropriate therapy, improved quality of life, and overall reduced treatment costs.

Laparoscopic staging for pancreatic cancer is not a new concept. In fact, the first published case in the United States of a minimally invasive approach to cancer staging was in a patient with pancreatic cancer. Bernheim in 1911 staged a patient of W. S. Halstead with presumed pancreatic cancer prior to laparotomy.96 He stated that the procedure he termed organoscopy “may reveal general metastases or a secondary nodule in the liver, thus rendering further procedures unnecessary and saving the patient a rather prolonged convalescence.” The use of laparoscopy was sporadic and not widespread until the seminal works of Alfred Cuschieri from Scotland and Andrew Warshaw from the United States.97–100 Both used the technique before the laparoscopic revolution and began to define the role it would have in the staging algorithm.

The yield of positive laparoscopy that avoids unnecessary laparotomy is highly dependent on the quality of the preoperative radiologic studies. The yield of laparoscopy cannot be assessed from studies that have not included state-of-the-art CT scans.101 Currently, the standard protocol should include a contrast-enhanced thin-cut dynamic CT scan of the pancreas. Initial reports from MSKCC concerning LS of peripancreatic malignancy reported an improvement in resectability from 50% based on standard CT scanning alone to 92% when staging laparoscopy was performed.73,102 Compared with previous reports, improvements in technology and better patient selection have reduced the benefit of laparoscopy. However, laparoscopy continues to consistently upstage approximately 15–20% of patients with radiologically resectable disease.73,97,102–107

An early study at MSKCC examined 577 patients who following contrast-enhanced CT scans were considered to have potentially resectable disease and underwent staging laparoscopy.108 Unresectability was determined at laparoscopy if histologic proof was obtained of:

1. Metastasis (hepatic, serosal, and/or peritoneal) (Fig. 5-11)

2. Extrapancreatic extension of the tumor (ie, mesocolic involvement)

3. Celiac or high portal node involvement

4. Invasion or encasement of the celiac or hepatic artery

5. Involvement by tumor of the superior mesenteric artery

Portal or superior mesenteric venous involvement was considered a relative contraindication to resection depending on the degree and extent of involvement.

In the MSKCC series, 366 patients were considered to have resectable disease after LS and subsequently underwent open exploration, with 92% (338 patients) being resected. The predominant sites for metastases were the liver and the peritoneal cavity. The resectability rate was compared with results from the decade before the introduction of LS, during which 1135 patients at MSKCC were explored but only 35% were resected. In a recent update from the same group examining 1045 patients with radiographically resectable disease who underwent LS between 1995 and 2005, it was reported that the yield of LS had decreased to 14% for patients with pancreatic adenocarcinoma. The predominant reason for unresectability at lapaproscopy was metastatic liver disease. Only seven patients were noted to have locoregional disease highlighting the improvements in CT imaging. Of the patients considered resectable at laparoscopy 99% were subsequently resected.109 A similar yield was reported by Doran and colleagues; 239 patients with suspected periampullary cancer underwent staging laparoscopy following dual-phase helical CT scanning.110 CT “resectable” disease was noted in 190 patients, of whom laparoscopy correctly identified unresectable disease in 28 patients. Overall, owing to findings at laparoscopy, 15% of patients were spared a further procedure, leading the authors to conclude that when added to CT scanning, LS provides valuable information that improves the selection of patients for surgical or nonsurgical treatment significantly. Many other authors have reported similar results (Table 5-3).

Carbohydrate Antigen 19-9

The use of tumor markers such as the carbohydrate antigen 19-9 (CA-19-9) to further select patients for staging laparoscopy has been proposed. Halloran and colleagues used a cut off value for CA-19-9 of 150 kU/L improving resectability levels and reducing nontherapeutic laparotomies.111 Using a receiver operating characteristics (ROC) curve for preoperative CA-19-9 values and tumor respectability, Maithel and coworkers demonstrated a statistically optimal cutoff of 130 U/mL. Unresectable disease was identified in 38 of 144 patients (26%) with a preoperative level ≥130 U/mL compared to 13 of 118 patients (11%) with a value <130 U/mL.112 We believe that this added information does allow for better selection of patients. Figure 5-12 details a recommended clinical algorithm. The use of such a strategy is supported by the analyses of population-based administrative databases. Mayo and colleagues reviewed the experience in Oregon and noted that the majority of patients did not undergo LS.113 However, in the subset that did undergo LS metastatic disease, which precluded resection, was noted in 27.6%. These patients were spared an unnecessary laparotomy.

Laparoscopic Ultrasonography

To further increase the added value, we use LUS. LUS has been used by a number of groups in an attempt to increase the yield of LS.114,115 John and colleagues showed that laparoscopy demonstrated unsuspected metastatic disease in 14 of 40 patients considered to have resectable disease.103 However, laparoscopy had only 50% sensitivity in predicting tumor resectability. The accuracy in predicting resectability increased to 89% with the addition of LUS. Several other studies have demonstrated that the added value of LUS to standard laparoscopy is on the order of 14–25%.73,116–119 Callery and colleagues analyzed the effect of routine implementation of laparoscopy with LUS, determining that the addition of LUS improved staging by identifying an additional 22% of patients with unresectable disease.73 Minnard and colleagues reported the benefit of LUS over laparoscopy alone in evaluating the primary tumor and the presence of vascular involvement.117 LUS findings resulted in a change in surgical treatment in 14% of patients in whom standard laparoscopic examination was equivocal. A further study by Schachter and colleagues demonstrated a change in surgical intervention in 36% of patients, with avoidance of unnecessary laparotomy in 31%.120 Catheline and colleagues reported that LUS altered therapy in 41% of cases, avoiding open exploration in 46%.121 This group reported a 90% sensitivity for assessing positive nodal disease and 100% for hepatic and peritoneal disease. Vollmer and colleagues similarly reported an improvement in resection rates using LUS (84% with LS vs 58% without)122. Merchant and colleagues concluded that the addition of LUS during LS enhances the ability of laparoscopy to determine resectability and approaches the accuracy of open exploration without increasing morbidity or mortality significantly.123 As clinical experience has developed it appears that the main utility remains in assessment of the liver or vascular involvement (Fig. 5-13). In patients with portal/superior mesenteric vein involvement, determining (i) resectability and/or if resectable (ii) the amount of venous resection required to obtain clear margins. This approach is supported by Thomson and colleagues using the widely accepted CT classification of vascular involvement, which examined the relationship between the tumor and the major vasculature and grades the involvement between A and F.124 Tumors graded A to D were generally resectable while those graded E and F were invariably unresectable.125 The authors suggest that, using these criteria, the selective use of LUS in LS is indicated. A meta-analysis examining the role of laparoscopy and LUS was performed by Hariharan and coworkers.126 They identified 29 studies in which 3305 patients underwent LS. The true yield was 25% (95% CI 24–27). The authors suggested that this represents a significant benefit to patients with potentially resectable adenocarcinoma of the pancreas in avoiding nontherapeutic laparotomy.

Peritoneal Cytology

Cytologic examination of peritoneal washing obtained at the time of laparoscopy has also been suggested to enhance the sensitivity of staging laparoscopy.127 Laparoscopy combined with peritoneal cytology is reported to upstage approximately 10% of patients.123,128 Peritoneal recurrence is a significant site of failure following a potentially curative pancreaticoduodenectomy. Leach and colleagues studied a consecutive series of patients with suspected or biopsy-proven radiologically resectable adenocarcinoma of the pancreatic head.129 Peritoneal washings were obtained at the time of staging laparoscopy and/or at subsequent laparotomy. Positive peritoneal cytology (PPC) was noted in 7% of patients, all of whom had overt metastatic disease at a median of 4.8 months. Merchant and colleagues examined 228 patients with radiographically resectable pancreatic adenocarcinoma who underwent LS.123 Peritoneal washings were taken from both upper quadrants at the beginning of laparoscopy. Overall survival was significantly higher in patients with negative peritoneal cytology. The authors determined that PPC had a positive predictive value of 94.1%, a specificity of 98.1%, and a sensitivity of 25.6% for determining unresectability. Quantitative real time-polymerase chain reaction (RT-PCR) assay was used by Dalal and coworkers to detect tumor cells in a cohort of 35 patients undergoing staging laparoscopy.130 Positive cytology was noted in eight cases and appeared to be stage-related. However, 25 patients were RT-PCR positive suggesting that this methodology could represent a more sensitive method for the detection of subclinical disease and enable improved selection for operative therapies and clinical trials. This intriguing pilot study requires further confirmation.

Despite the above, there is no consensus around the value of LS. Critics argue that confining laparoscopy to the setting of determination of resectability overestimates the usefulness of laparoscopy because it fails to account for patients who require open procedures for palliation of unresectable disease.101

Pisters and colleagues from the MD Anderson Cancer Center reported resectability rates of 80% using high-quality CT scanning alone.101 Based on these data, the authors proposed that the maximum positive yield of routine staging laparoscopy in patients with potentially resectable disease on high-quality CT scanning would be 20%, assuming a false-negative result of zero. This group did not perform routine staging laparoscopy but rather used selective laparoscopy at the time of planned laparotomy for tumor resection in patients with localized disease on CT scan and patients at high risk for occult M1 disease.131,132 This is a strategy that has been advocated by others.133 Gouma and colleagues from Amsterdam assessed the role of LS in patients with periampullary tumors compared with standard radiologic staging with helical CT scanning.134 Laparoscopic staging identified biopsy-proven unresectable disease in only 13% of 297 patients, with a detection rate of 35%. Based on the findings, the authors proposed that LS should be performed selectively. Since their practice is to recommend a surgical bypass as palliation for patients with locally advanced unresectable disease, they believe that LS only adds value in the presence of metastatic disease.

Locally Advanced Disease

Recent reports have focused on the role of LS in patients with locally advanced unresectable disease who were considered for adjuvant chemoradiotherapy. Shoup and colleagues reviewed 100 consecutive patients with locally advanced disease who underwent LS.135 Contemporary imaging studies failed to detect metastatic disease in 37% of cases. Peritoneal disease was noted in 12 cases, liver metastases in 18, and 7 patients had both. Similar results were reported by Liu and Traverso, who described their experience with 74 patients, all of whom had undergone high-quality pancreas protocol CT examination prior to LS.136 Occult tumor was found in 34% of patients. The authors reported that tumors situated in the body and tail of the gland were more likely than head lesions to have unsuspected metastases (53% vs 28%). Morak and coworkers in a prospective cohort study reported that 24 of 68 (35%) patients with locally advanced disease on CT had metastatic disease at laparoscopy.137 These studies emphasize that despite the improvements in imaging modalities, LS should be performed in patients considered to have locally advanced disease prior to the start of combined modality therapy.

The studies cited earlier focus on invasive ductal adenocarcinoma of the pancreas. For other cell types, including neuroendocrine tumors, intraductal papillary mucinous neoplasms, and cystadenocarcinomas, the data are sparse. A review of the MSKCC experience with laparoscopy in nonfunctioning islet cell tumors by Hochwald and colleagues found a high incidence of occult metastases at laparoscopy.138 CT scan followed by laparoscopy was significantly more sensitive than CT scan alone in predicting resectability (93% vs 50%; p = 0.03). This resulted from a high false-negative rate on CT scan for small-volume metastatic disease, hepatic disease being the most common site. The predictive value for tumor resectability also was much higher for CT scan followed by laparoscopy than for CT scan alone (95% vs 74%). Brooks and colleagues examined the role of LS in 144 patients with ampullary, duodenal, and distal bile duct tumors.139 Patients with distal bile duct tumors also appeared to benefit from LS in terms of both determining resectability and avoiding unnecessary surgery. In contrast, patients with known duodenal or ampullary tumors gained little added value from LS.

In experienced hands, the procedure is safe and well tolerated as a day case procedure. Complications are low with few specific reports in the literature. Of those, only one identifies a series of complications directly attributable to the LS procedure.140 In their series published in 2003, Rodgers et al report a complication rate of 2.8% (3/106 patients), one of which was unrecognized at laparoscopy and which eventually contributed directly to the patients' death. In general, major morbidity such as hemorrhage, visceral perforation, and intra-abdominal infection may occur in 1–2% of cases.

As the use of laparoscopy in malignant disease increased, concern was expressed regarding the potential risk of disseminating disease at the time of pneumoperitoneum. An initial case report in 1978 by Dobronte and colleagues described a “port site” tumor implant in a patient with malignant ascites 2 weeks following laparoscopy.141 A number of similar reports followed, again involving patients who had disseminated disease at the time of their laparoscopic examination. Nieveen van Dijkum and colleagues from Amsterdam demonstrated an overall 2% port-site recurrence, with all cases having advanced peritoneal disease.142

Clinical experience over the last two decades appears to support the hypothesis that LS is safe from the oncologic standpoint. Pearlstone and colleagues from the MD Anderson Cancer Center described their experience with laparoscopy in 533 patients with nongynecologic intra-abdominal cancer, 339 of whom had laparoscopic procedures for upper gastrointestinal malignancies.143 They reported port-site recurrences in four patients (0.88%), three of whom had advanced disease at the time of initial laparoscopy. Similar results were noted in a report from MSKCC, which reviewed a prospective database of 1650 diagnostic laparoscopic procedures performed in 1548 patients with upper gastrointestinal malignancies, in which a total of 4299 trocars were inserted.144 The most frequent diagnosis was pancreatic cancer (51.2%). At a median follow-up of 18 months, a port-site recurrence was noted in 13 patients (0.8%). An open operation was performed in 1040 patients, of whom 9 (0.9%) developed a wound recurrence. This latter figure is similar to the 0.8% incisional recurrence rate noted by Hughes and colleagues in a review of 1600 open laparotomies for colon cancer.145 Median time for the development of the port-site recurrence in the MSKCC study was 8.2 months. Eight occurred in patients with documented metastatic disease at the time of laparoscopy, and the remaining five had local or distant disease at the time of diagnosis of the port-site implant, and therefore, the recurrence did not appear to be an isolated event but rather a marker for more advanced disease. The authors concluded that LS appeared safe from an oncologic standpoint. This is further supported by a retrospective review of 235 patients who had laparoscopy to stage pancreatic cancer. This study demonstrated a port-site recurrence rate of 3% versus a 3.9% incisional recurrence rate in those patients who had an exploratory laparotomy alone.146

A number of hypotheses have been suggested to explain port-site implantation. Tumor seeding has been associated with carbon dioxide pneumoperitoneum in animal studies; however, reports that tumor growth is established more easily after open laparotomy would appear to refute this theory.147–149 Other mechanisms, such as tissue manipulation, direct wound contamination, poor surgical technique, or immunologic effects such as changes in host immune responses, also have been suggested.150 It appears so far, however, in most studies that port-site implantation is uncommon, differs little from open surgical incision recurrence, and is more likely to reflect the underlying biologic behavior of the disease rather than the type of surgery.

Since the majority of patients with pancreatic cancer have unresectable disease at the time of presentation, palliation to minimize symptoms and maximize quality of life has a major role in the care of these patients. Palliation most commonly is required for one of three problems: biliary obstruction, gastric outlet obstruction (GOO), and relief of pain.

While both cholecystoenteric and choledochoenteric bypasses have been performed laparoscopically, the latter is much more difficult technically, requiring a high level of laparoscopic skills. A sufficient length of common duct needs to be exposed, and a difficult intracorporeal anastomosis between the small bowel and the common duct must be performed. Cholecystojejunostomy is the more commonly performed laparoscopic procedure (Fig. 5-14). Patient selection is critical. A low insertion of the cystic duct into the common bile duct or tumor impingement within 1 cm of the duct is a predictor of early technical failure. The anastomosis can be performed with either a stapled or hand-sewn technique. In patients who have experienced a prior cholecystectomy or who have a diseased gallbladder, blocked cystic duct, low insertion of the cystic duct, or tumor encroachment on the cystic duct or gallbladder, a cholecystojejunostomy is not possible; therefore, either a laparoscopic cholodochojejunostomy is performed, or the procedure is converted to open and a standard surgical bypass is performed.

Rhodes and colleagues presented in 1995 one of the first series of patients who underwent laparoscopic palliation for advanced pancreatic carcinoma. From the 16 patients, 7 underwent laparoscopic cholecystojejunostomy, 5 had laparoscopic gastroenterostomy, 3 had both procedures, and in 1 patient laparoscopic palliation failed. The median operating time was 75 minutes, the hospital stay was 4 days, the morbidity was 13%, and the median survival in 10 patients was 201 days, with the rest of the patients remaining alive at the time of the publication.151 In 1999, Rothlin and colleagues published a case-controlled study of 28 patients with pancreatic cancer divided in two groups; in one group, laparoscopic palliation was performed, and the other group underwent conventional surgical palliation.152 Of the 14 patients in the laparoscopic group, 7 had laparoscopic gastroenterostomy, 3 had gastoenterostomy and hepaticojejunostomy, and 4 had staging laparoscopy only.

Postoperative morbidity was 7% for the laparoscopic group compared with 43% for the open palliation group. There were no deaths in the laparoscopic group versus 29% mortality in the open group. Average postoperative hospital stay was 9 days for the laparoscopic group versus 21 days for the open group. Finally, the laparoscopic group required significantly less analgesia postoperatively. Choi presented a series of 78 gastrojejunostomies, 45 open and 33 laparoscopic, performed for palliation of gastric outflow obstruction caused by advanced gastric, duodenal, ampullary, and pancreatic cancers.153 In the laparoscopic group, there was less suppression of immune function, lower morbidity, and earlier recovery of bowel function. A randomized trial reported by Navarra and colleagues demonstrated that patients undergoing a laparoscopic gastrojejunostomy had significantly less intraoperative blood loss and resumed oral intake sooner than those patients undergoing an open palliative antecolic gastrojejunostomy.154

The technique for a transumbilical single-incision laparoscopic gastrojejunostomy has recently been reported.155 While this is technically feasible, the benefits compared to the conventional laparoscopic approach remain to be determined.

The true incidence of symptomatic GOO in pancreatic cancer remains unclear. Historically, it was considered that more than 25% of patients would develop GOO during the course of their illness, and therefore, prophylaxic gastric bypass was recommended at the time of exploratory laparotomy. However, as the need for open exploration for staging purposes has decreased, the need for prophylaxic bypass for the majority of patients has been questioned. GOO is a late complication of advanced pancreatic cancer affecting 10–20% of patients who survive more than 15 months.156–158 However, fewer than 3% of the patients who develop GOO require surgical bypass.156,159,160 Most important, 60% of patients with advanced pancreatic cancer have delayed gastric emptying with no evidence of gastric or duodenal invasion. This may be explained by tumor infiltration of the celiac plexus causing gastric stasis, nausea, and vomiting.161

Espat and colleagues examined in a prospective but nonrandom study of 155 patients undergoing LS.156 Following laparoscopy, 40 patients had locally advanced unresectable disease, and the remainder had metastatic disease. In follow-up, only 3% of patients required a subsequent open operation for biliary drainage or GOO. A subsequent update of this experience has confirmed the results, with over 90% of patients dead of disease. This low incidence of patients requiring operation for symptomatic GOO is consistent with the data seen from the nonoperative control groups in randomized trials of endoscopic biliary drainage versus surgery.

A laparoscopic gastroenterostomy is a relatively straightforward procedure. Nagy and colleagues reported a series of laparoscopic gastrojejunostomies.162 Nine of 10 patients in this series had GOO from pancreatic malignancy. The laparoscopic method was successful in 90%. There was no postoperative morbidity or mortality associated with the surgical technique.

Surgical Technique for Biliary and Gastric Bypass

The patient is placed supine on the operating table in 10 degrees of reverse Tredelenberg position with 10 degrees of left lateral tilt. The placement of trocars is similar to that for a standard staging procedure. However, in order to accommodate a linear stapler, the right upper quadrant 10-mm trocar is converted into a 12- to 15-mm size. Following exploration, the ligament of Trietz is identified, and a loop of jejunum approximately 30 cm distal to the ligament of Treitz is brought in an antecolic position to the gallbladder (Fig. 5-15). Using an intracorporeal suturing technique, the jejunum is approximated to the gallbladder by two 3-0 coated, braided lactomer sutures (Polysorb, US Surgical, Norwalk, CT). The distended gallbladder may be decompressed using a Veress needle attached to a suction device. There is usually minimal biliary spillage owing to the raised intra-abdominal pressure consequent on the pneumoperitoneum. Small enterotomy incisions (10 mm) are made in the gallbladder and jejunum using either scissors or a device such as the ultrasonic shears (Fig. 5-16). Hemostasis is achieved with electrocautery. Any spillage can be dealt with by suction device placed through the left upper quadrant port. An endoscopic 30-mm linear stapler using 3.5-mm staples is introduced through the right upper quadrant port, and the “jaws” are manipulated into the gallbladder and jejunum in a standard fashion. Often, this is difficult because of the proximity of the port site to the gallbladder. A reticulating stapler facilitates this maneuver. The stapler heads are approximated, and the instrument is fired (Fig. 5-17). After removing the stapler, the anastomosis is inspected, hemostasis is confirmed, and the gallbladder interior is aspirated and irrigated with saline.

The resulting enterotomy can be closed by using either a completely intracorporeal or laparoscopically assisted approach. Using an intracorporeal technique, the defect is closed with a continuous seromuscular 3-0 coated, braided lactomer suture, with knots tied using an intracorporeal technique (Fig. 5-18).

An alternative method is to create a completely hand-sewn anastomosis using 3-0 coated, braided lactomer suture. If a running suture is used, the assistant should maintain tension on the suture with an atraumatic grasping forceps following placement of each stitch. Knots can be tied either using an intracorporeal or extracorporeal technique.

A laparoscopically assisted method is suitable in thin patients. Two stay sutures are placed on either side of the anastomotic defect. These sutures are cut long. The 12-mm trocar is removed, and the incision is enlarged to 20 mm. Using retraction on the stay sutures, the newly created biliary-enteric anastomosis can be exteriorized and the enterotomy closed in a standard fashion. When this is completed, the bowel is returned to the abdominal cavity, and the wound is closed. The abdomen is reinsufflated and the anastomosis inspected. This technique allows for the construction of a 2.5-cm cholecystojejunal anastomosis without any bowel narrowing. No intra-abdominal drains are used.

The technique for fashioning a gastrojejunostomy is similar. In this case, a proximal loop of jejunum is brought in an antecolic position to the stomach. The left upper quadrant 5-mm laparoscopic trocar is converted to a 12-mm trocar. Two 3-0 coated, braided lactomer sutures (Polysorb, US Surgical) are used to approximate the jejunum to the stomach. Enterotomies are made in both stomach and jejunum. In cases in which there has been a significant period of gastric obstruction, the gastric wall may be hypertrophied, making creation of the gastrotomy difficult. Confirmation that one is inside the stomach is required before placement of the stapler. When this is achieved, a 30-mm linear stapler is inserted through the 12-mm left upper quadrant port and manipulated into both enterotomies. The instrument is positioned and fired. The stapler is removed and reloaded, returned into the anastomosis, and refired. This creates an anastomosis approximately 5 cm in length. The anterior defect can be closed in a fashion similar to the cholecystojejunostomy (Fig. 5-19). Any defects in the anastomosis can be repaired with individual 3-0 sutures.

The ideal palliative procedure for biliary or gastric obstruction should be effective in relieving jaundice or GOO, have minimal morbidity, be associated with a short hospital stay, have a low symptomatic recurrence, and maintain quality of life. Laparoscopic procedures have the potential to achieve these goals, although data do not support prophylactic bypass procedures in patients who do not otherwise require surgery.

Laparoscopy is no longer a tool of limited use and now has widespread indications within surgical oncologic practice. Despite improvements in noninvasive imaging, there is still an added value to use LS in selected patients with upper gastrointestinal cancers. In the future, the combination of NOTES technology and MIS techniques offers further exciting potential to enhance staging of these patients.