Diagnostic accuracy of laparoscopy following computed tomography (CT) scanning for assessing the resectability with curative intent in pancreatic and periampullary cancer

Abstract

Surgical resection is the only potentially curative treatment for pancreatic and periampullary cancer. A considerable proportion of patients undergo unnecessary laparotomy because of underestimation of the extent of the cancer on computed tomography (CT) scanning. Laparoscopy can detect metastases not visualised on CT scanning, enabling better assessment of the spread of cancer (staging of cancer). There has been no systematic review or meta-analysis assessing the role of diagnostic laparoscopy in assessing the resectability with curative intent in patients with pancreatic and periampullary cancer. To determine the diagnostic accuracy of diagnostic laparoscopy performed as an add-on test to CT scanning in the assessment of curative resectability in pancreatic and periampullary cancer. We searched the Cochrane Register of Diagnostic Test Accuracy Studies, the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE via PubMed, EMBASE via OvidSP (from inception to 13 September 2012), and Science Citation Index Expanded (from 1980 to 13 September 2012). We included diagnostic accuracy studies of diagnostic laparoscopy in patients with potentially resectable pancreatic and periampullary cancer on CT scan, where confirmation of liver or peritoneal involvement was by histopathological examination of suspicious (liver or peritoneal) lesions obtained at diagnostic laparoscopy or laparotomy. We accepted any criteria of resectability used in the studies. We included studies irrespective of language, publication status, or study design (prospective or retrospective). We excluded case-control studies. Two authors independently performed data extraction and quality assessment using the QUADAS-2 tool. The specificity of diagnostic laparoscopy in all studies was 1 because there were no false positives since laparoscopy and the reference standard are one and the same if histological examination after diagnostic laparoscopy is positive. Therefore, the sensitivities were meta-analysed using a univariate random-effects logistic regression model. The probability of unresectability in patients who had a negative laparoscopy (post-test probability for patients with a negative test result) was calculated using the median probability of unresectability (pre-test probability) from the included studies and the negative likelihood ratio derived from the model (specificity of 1 assumed). The difference between the pre-test and post-test probabilities gave the overall added value of diagnostic laparoscopy compared to the standard practice of CT scan staging alone. Fifteen studies with a total of 1015 patients were included in the meta-analysis. Only one study including 52 patients had a low risk of bias and low applicability concern in the patient selection domain. The median pre-test probability of unresectable disease after CT scanning across studies was 40.3% (that is 40 out of 100 patients who had resectable cancer after CT scan were found to have unresectable disease on laparotomy). The summary sensitivity of diagnostic laparoscopy was 68.7% (95% CI 54.3% to 80.2%). Assuming a pre-test probability of 40.3%, the post-test probability of unresectable disease for patients with a negative test result was 0.17 (95% CI 0.12 to 0.24). This indicates that if a patient is said to have resectable disease after diagnostic laparoscopy and CT scan, there is a 17% probability that their cancer will be unresectable compared to a 40% probability for those receiving CT alone.A subgroup analysis of patients with pancreatic cancer gave a summary sensitivity of 67.9% (95% CI 41.1% to 86.5%). The post-test probability of unresectable disease after being considered resectable on both CT and diagnostic laparoscopy was 18% compared to 40% for those receiving CT alone. Diagnostic laparoscopy may decrease the rate of unnecessary laparotomy in patients with pancreatic and periampullary cancer found to have resectable disease on CT scan. On average, using diagnostic laparoscopy with biopsy and histopathological confirmation of suspicious lesions prior to laparotomy would avoid 23 unnecessary laparotomies in 100 patients in whom resection of cancer with curative intent is planned.