Laparoscopy has evolved tremendously in the past 40 years, from a diagnostic tool to a surgical platform with nearly as many therapeutic applications as open surgery. In the most capable hands the only current primary limitation to minimally invasive approaches to most gastrointestinal (GI) cancers is the size of the incision required for removal of the specimen(s). However, it is clear that the most advanced techniques require extensive training and operative skill to perform safely and consistently on unselected patients. Given these skill sets are not realistically achievable by many surgeons, some techniques such as minimally invasive esophageal, hepatic, and pancreatic resections will remain in the purview of highly specialized practitioners. More recently, laparoscopy is being supplanted by the addition of the robotic minimally invasive platform with its promise of improved surgical exposure and increased instrument dexterity. It seems likely that with the continued advancement of technology in the field, the frontier of minimally invasive surgery will continue to expand.

While advanced laparoscopic and robotic-assisted techniques require highly specialized skill sets, diagnostic and staging laparoscopy techniques have broad clinical applicability to many GI malignancies, are easily performed, and often provide valuable information that directly impacts clinical decision making. Here we describe the use of laparoscopy in the diagnosis and treatment of common GI malignancies.

Preoperative history and physical examination are performed with particular attention to cardiopulmonary comorbidities, coagulopathy, and overall functional/nutritional reserve as with any open surgery. As laparoscopy is nearly always performed under general anesthetic and with CO₂ insufflation of the abdomen, there can be significant physiologic stress, particularly in patients with limited cardiopulmonary reserve. In most cases these comorbidities may be mitigated, allowing safe conduct of operation.

The patient is positioned supine, pressure points padded, and secured to the operating table. Arms may be extended or tucked according to surgeon preference and region(s) of the abdomen to be explored. After induction of anesthesia with medical paralysis, the peritoneum is accessed via either an open (Hasson) or closed (Veress or optical separator trocar) technique. Initial access is commonly at the umbilicus for Hasson and Veress techniques when there have been no prior operations, otherwise access may be gained through a paramedian incision (Fig. 7-1). Optical separators are best utilized in the left upper quadrant through the rectus muscle where adhesions are typically sparse. The authors prefer an optical trocar technique in most instances. CO₂ insufflation at 12 to 15 mm Hg is used except where cardiopulmonary disease prohibits full insufflation and lower pressures of 10 to 12 mm Hg may be used.

A 30-degree 5-mm scope is inserted and initial inspection commenced with attention to the presence of ascites, omental/peritoneal nodules, or liver masses. Next, additional trocars are placed in a manner that allows the region of primary interest to be examined with a head-on view and instrument ports to be aligned with the viewing angle (Fig. 7-1). The number and size of additional trocars will depend on the need for biopsies and other interventions, though a complete diagnostic laparoscopy with cup biopsies of the liver or peritoneum will generally require two additional 5-mm trocars. If peritoneal cytology is to be performed, it is done prior to manipulation or dissection of tissues by instilling 250 mL normal saline into each of the upper quadrants and aspiration into a Lugol’s trap.

Examination of the peritoneum is performed systematically, beginning in the right upper quadrant. With the patient in reverse Trendelenberg positioning and the right side tilted upward, the right lobe of the liver is gently retracted to allow examination of the surface of the diaphragm and then elevated to view the undersurface of the liver, gallbladder, and porta hepatis (Fig. 7-2). Next, with the left side elevated, the left hemidiaphragm and left lateral segment of the liver are examined. Great care is taken not to damage the spleen with inadvertent or overly aggressive manipulation. Elevation of the left lateral segment reveals the hepatogastric ligament (Fig. 7-3), which may be opened along the pars flaccida to gain access to the lesser sac and allow biopsy of hepatic, celiac, and left gastric artery lymph nodes if indicated.

With the patient in neutral positioning, the omentum is reflected upward, exposing the transverse colon, which is elevated allowing inspection of the transverse mesocolon and the ligament of Treitz—sites of potential locally advanced disease or lymphadenopathy in pancreatic cancer or lymphoma, for example (Fig. 7-3). The small intestine may be examined from ligament of Treitz to ileocecal valve at this time as well. Finally, the patient is placed in Trendelenberg position and the small bowel and colon swept out of the pelvis to facilitate examination of the peritoneum and for females, pelvic organs (Fig. 7-4).

Laparoscopic ultrasound (LUS) may be performed to examine the liver, lymph nodes (peri-portal, celiac, peri-aortic, etc.) or other solid organs, including the pancreas, though this requires opening the gastrocolic ligament along the greater curve of the stomach to gain access to the lesser sac. Liver ultrasound in particular allows identification and core-needle biopsy of parenchymal masses not apparent on visual inspection.

Upon completion of the examination of the peritoneum, the insufflation is released and the trocars are withdrawn under direct vision, taking care to inspect for bleeding. Fascial incisions of 10 mm or more are closed to prevent herniation of abdominal contents.

Esophageal/Gastric Malignancy

Squamous cell carcinoma and adenocarcinoma of the proximal and middle one-third of the esophagus is associated with a lower likelihood of liver metastases and peritoneal disease,¹ and thus diagnostic laparoscopy is seldom useful unless there is suspicion of metastatic disease or inconclusive cross-sectional imaging. For adenocarcinoma of the distal esophagus and gastroesophageal (GE) junction, the National Comprehensive Cancer Network (NCCN) guidelines currently recommend a metastatic workup including PET/CT.² PET imaging aids in the detection of lymph node metastases and metastatic disease when compared with CT and endoscopic ultrasound (EUS), though its accuracy is still modest at 70% to 82%.^3,4 Currently, staging laparoscopy is an optional part of the staging assessment on the basis that its role is poorly defined with some evidence for and some evidence against its routine use.

de Graff et al. performed staging laparoscopy in 416 patients with esophageal and GE junction tumors that were resectable by CT criteria. They found 84 (20.2%) patients had unresectable disease identified at laparoscopy (locally advanced in 17, lymph node disease in 14, and metastases in 63) that precluded resection. Twenty-seven patients who went on to have laparotomies had unresectable disease (metastases in 11 and locally advanced disease in 16), yielding a sensitivity of laparoscopy of 88%. As mentioned earlier, the yield of staging laparoscopy for proximal and mid-esophageal tumors and for squamous cell carcinoma was low (laparoscopy changed management in 0/28 and 1/33 cases, respectively).⁵ Neither CT/PET nor EUS was used for the preoperative staging evaluation in this study. Both have been shown to improve the ability to predict resectablility and are in routine use today. As a result, the utility of routine staging laparoscopy is probably overestimated in this case.

The addition of LUS to standard staging laparoscopy has been reported by Wakelin et al. Endoscopic ultrasound was better at determining T stage except where stricture prevented complete examination. However, LUS was complementary to CT for assessment of metastatic disease. They reported an accuracy of staging laparoscopy with LUS of 81% for detection of metastatic disease compared to 72% for CT.⁶ LUS was ineffective in evaluating tumors above the diaphragm. In practice, LUS is unlikely to provide additional information except in instances where EUS is unable to be performed.

Recent studies have provided important information about the negative prognostic significance of positive peritoneal cytology, which increases the importance of laparoscopy as a staging tool. Nath et al. evaluated 255 patients with esophageal (n = 82), GE junction (n = 48), and gastric (n = 125) cancers and no evidence of unresectability on preoperative EUS and CT (without PET). They found 48 patients (18.8%) with macroscopic metastatic disease and another 15 (5.9%) with positive cytology. Gastric cancer patients had radiographically occult metastatic peritoneal disease 28.8% of the time.⁷ The authors found no difference in survival between patients with macroscopic metastatic disease and only positive cytology (median survival 9 vs 13 months; p = 0.52) which led them to conclude curative resection should not be performed for patients with positive peritoneal cytology.

Nearly identical findings were reported by Convie et al. with macroscopic metastases in 22.6% of gastric adenocarcinoma patients and 11.8% of esophageal carcinoma (n = 136 adenocarcinoma, 22 squamous). Cytology was positive in an additional five gastric and six esophageal carcinoma patients. The authors found positive cytology to be an equally poor prognostic sign as macroscopic metastatic disease.⁸

Another study by Munasinghe et al. demonstrated that the location of lavage and collection of the cytology specimen is important, with the greatest yield coming from lavage and aspiration of the subphrenic region (sensitivity 90.7%), while pelvic samples have lower sensitivity (76.7%). In a study of 316 patients with esophageal, GE junction, and gastric cancers, pelvic aspiration alone understaged patients 23.3% of the time. The yield of staging laparoscopy alone was 8.9% and the addition of cytology identified another 13.6% of patients with advanced disease, for a total yield of 22.5%. Patients in this study were staged preoperatively with CT, EUS, and PET-CT providing evidence that staging laparoscopy, particularly in combination with cytology, provides meaningful information, even in the era of advanced cross-sectional imaging.⁹

Neoadjuvant chemotherapy and chemoradiation protocols are increasingly being investigated and implemented,^10,11 and though overall outcomes appear to be improved there are some patients who progress while receiving neoadjuvant therapy. The role of repeat staging laparoscopy with peritoneal cytology was investigated by Cardona et al., who found that 7% of patients had radiographically occult metastatic disease at the time of repeat laparoscopy, though cytology was only positive in the absence of macroscopic metastatic disease in one patient. They concluded that repeat staging laparoscopy was warranted but performing cytology was not, given its low yield.¹²

Primary and Secondary Hepatobiliary Malignancy

Preoperative imaging techniques have comparatively low sensitivity for detecting metastatic and locally invasive disease in hepatobiliary malignancies such as gallbladder carcinoma and hilar cholangiocarcinoma (HC). As an example, the sensitivity of PET/CT for unresectablility of gallbladder carcinoma in one recent study was only 56%,¹³ and the accuracy of predicting resectability of cholangiocarcinoma (either hilar or intrahepatic) was 72.4% for PET/CT in another study.¹⁴ A recent expert panel consensus statement on intrahepatic cholangiocarcinoma concluded in part: “…a substantial number of unresectable patients will benefit from staging laparoscopy… [and] staging laparoscopy should be routinely utilized in high-risk patients (i.e. patients with multicentric disease, high CA 19-9, questionable vascular invasion or suspicion of peritoneal disease)…”¹⁵

Goere et al. performed staging laparoscopy in 39 patients with gallbladder carcinoma, intrahepatic cholangiocarcinoma (IHC), and HC.¹⁶ All patients had triple-phase contrast CT and 90% had MRI prior to staging laparoscopy. They found metastatic disease or cirrhosis precluding resection in 14 patients (36%). Another nine patients were found to be unresectable on laparotomy, primarily because of vascular invasion or lymph node metastases. Only one patient had peritoneal metastases that were missed on laparoscopy and two patients had previously unrecognized liver metastases. The yield of laparoscopy was highest for gallbladder carcinoma (62%, accuracy 83%) while the yield for IHC (yield 36%, accuracy 67%) and HC (yield 25%, accuracy 45%) was somewhat lower. The authors recommended that staging laparoscopy be performed routinely for gallbladder carcinoma and IHC and selectively for HC.

Similarly, D’Angelica et al. studied the role of staging laparoscopy for both primary and secondary hepatobiliary malignancy in 401 patients and found a yield of 21% and an accuracy of 54.9%.¹⁷ Ninety-seven percent of study patients had preoperative CT, 45.9% had MRI, and 86.7% had two or more studies. The yield was highest for gallbladder carcinoma at nearly 50% and lowest for primary hepatocellular carcinoma (<20%) and metastatic colorectal cancer (10%). Overall accuracy for staging laparoscopy was 54.9% with most failures due to vascular invasion or lymph node metastases. The authors also performed LUS in 168 patients, and 25 (14.9%) of these yielded additional findings not seen on laparoscopy, with 8 LUS exams primarily responsible for preventing laparotomy. Finally, the authors found morbidity and hospital length of stay were significantly less in unresectable patients who were spared a laparotomy (morbidity 9.5% vs 27.5% and hospital stay 3 days vs 9 days).

In a study by Russolillo et al., 100 patients with preoperative imaging suggesting resectable gallbladder carcinoma, HC, or borderline resectable IHC (defined as a tumor larger than 10 cm or adjacent to or infiltrating the inferior vena cava, a major hepatic vein, bile duct confluence, or a first-order Glissonian pedicle) underwent staging laparoscopy combined with LUS.¹⁸ The overall yield and accuracy of laparoscopy was 18% and 60%, respectively. This was increased to a yield of 24% and accuracy of 80% with the addition of LUS. In the six patients who had false negative staging laparoscopies with LUS the reason for failure was peritoneal metastases (n = 2), lymph node metastases (n = 2), or vascular invasion (n = 2). Of note, the only factor that predicted the failure of staging laparoscopy with LUS to identify unresectable disease was preoperative biliary drainage.

Pancreatic Malignancy

A recent Cochrane review reported a meta-analysis of 15 studies including 1015 patients over a period of 26 years (1986-2012) to determine the diagnostic accuracy of laparoscopy for resectability of pancreatic or periampullary malignancy following CT scanning.¹⁹ They found the pretest probability of unresectable disease after CT scanning alone was 40.3% and the cumulative sensitivity of diagnostic laparoscopy was 68.7% (95% confidence interval [CI] 54.3%-80.2%). Thus the post-test probability of unresectable disease was 17%. The authors concluded that 23 laparotomies could be avoided for every 100 patients by performing laparoscopy in conjunction with CT scan. A subgroup analysis of only pancreatic cancer patients found similar results.

These data indicate that routine laparoscopy prior to planned laparotomy should be strongly considered given the fairly high likelihood of undetected metastatic disease. However, it is important to note that over half of the patients included in the meta-analysis were from studies published 15 years or more prior to the current study. In modern practice, the advent of multidetector helical CT scans and multiphase contrast administration protocols have increased the sensitivity of cross-sectional imaging considerably, allowing for the detection of more subtle metastatic disease as well as locally advanced disease that would preclude resection. Modern CT provides a sensitivity and specificity of 85% and 82%, respectively, for vascular involvement²⁰ and 88% and 89% for detection of liver metastases.²¹ As a result, diagnostic laparoscopy is not a compulsory prelude to laparotomy in many practitioners’ hands.

Staging laparoscopy may be performed on a more limited basis when the pretest probability of metastatic or locally advanced disease is high, as when the CA19-9 is markedly elevated. This was studied by Maithel et al., who found that serum CA19-9 values greater than 130 U/mL were predictive of unresectability, particularly for distal cancers. The most common site of unresectable disease was metastatic disease in the liver or peritoneum.²² Another more recent study found a CA19-9 value of 215.37 had a sensitivity of 72.7% and specificity of 52.3% for radiographically occult metastatic disease seen at time of laparoscopy.²³ Of note, 5% to 10% of patients do not have the Lewis antigen required to express CA19-9 and thus will not have elevated CA19-9 values, regardless of their resectability. As a result, a low CA19-9 value should not preclude diagnostic laparoscopy if preoperative imaging is equivocal. Diagnostic laparoscopy may also provide higher yield, and should be utilized when there is a question of unresectable disease on preoperative imaging.

Some metastatic disease on the posterior surface of the liver or on the retroperitoneum abutting the duodenum may not be visible at the time of standard diagnostic laparoscopy. Schnelldorfer et al. reported a series of 274 patients who underwent either initial staging laparoscopy followed by laparotomy (if laparoscopy was negative) or initial laparotomy. Both groups had radiographically occult metastatic disease 11% of the time, though only 2% of patients were found to have metastases on laparoscopy. The remaining 9% had metastases that were identified only on laparotomy. These were found on the posterior surface of the liver, paraduodenal retroperitoneum, proximal jejunal mesentery, and in the lesser sac. The authors argued that advanced laparoscopic techniques to mobilize and expose these areas are warranted to identify more subtle metastatic disease.²⁴ More extensive laparoscopic dissection such as a Kocher maneuver and mobilization of the right lobe of the liver should be reserved for cases in which there is concern preoperatively for unresectable disease, and should only be performed by experienced laparoscopists with advanced laparoscopy capabilities.

Several studies have examined the use of LUS to evaluate the liver for subcapsular/parenchymal metastases not visible on laparoscopy as well as for vascular invasion or non-regional lymph node metastases (ie, celiac/para-aortic). Piccolboni et al. performed diagnostic laparoscopy with LUS in 18 consecutive patients, four of whom had inconclusive findings of unresectable disease on preoperative imaging. LUS identified parenchymal liver metastases in two patients and vascular involvement precluding resection in another two. The authors concluded LUS was a necessary adjunct to laparoscopy in order to determine resectability for the four patients with equivocal preoperative imaging.²⁵ A larger study of 305 patients found the overall accuracy of preoperative CT for predicting resectability was 68.6%, and diagnostic laparoscopy with LUS increased this to 81%. Importantly, 4/49 patients who were deemed unresectable by CT were found to be resectable by LUS. Diagnostic laparoscopy and LUS influenced operative management 13.4% of the time.²⁶

Another study by Barabino et al. examined the role of LUS in the eras before and after the introduction of multidetector CTs. Prior to modern CT the authors performed LUS routinely and found that LUS changed the surgical strategy 30% of the time and accurately predicted resectability in 96% of patients thought to be resectable on preoperative imaging and 95% of patients who were deemed “doubtful” preoperatively. The overall yield of LUS was 45%. Following the advent of multidetector CT, the utility of LUS was curtailed significantly and the yield dropped to 1.8%.²⁷ While the authors concluded LUS should not be a routine part of the workup for pancreatic cancer, they acknowledged that in certain cases of equivocal preoperative CT imaging it may be useful to prevent an unnecessary laparotomy.