| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Gastroenterology Department (LFO-O, VA-C, RM-S), Surgery Division; Medical Oncology Department (DG-R), Internal Medicine Division; and General Director (JdlG-S); Instituto Nacional de Cancerología, México D.F. México.
Correspondence: Address correspondence and reprint requests to: Luis F. Oñate-Ocaña, MD, Departamento de Gastroenterología, Instituto Nacional de Cancerología, San Fernando 22, México D.F. 14000. México; Fax: 52-56-28-04-64; E-mail: lonate{at}prodigy.net.mx
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods: A retrospective cohort study was conducted of 151 patients allocated into four stages based on laparoscopic findings. The depth of tumor invasion and the presence of metastasis based on laparoscopic findings were used to construct these stages. Laparoscopic findings were compared with histopathology.
Results: An excellent agreement of the laparoscopy-defined depth of invasion and the surgical pathology standard was found (weighted kappa 0.85). The likelihood ratios for a positive and negative laparoscopic diagnosis of metastasis were 40.4 and 0.015, respectively (98.5% sensitivity, 97.6% specificity). Those for positive and negative diagnosis of resectability were 2.6 and 0.03, respectively (98.4% sensitivity, 62% specificity). The laparoscopic stages presented significant prognostic value. Two-year survival was 93%, 69%, 60%, and 17%, respectively. Surgical resection was possible in 100%, 100%, 49%, and 12%, respectively.
Conclusions: The proposed laparoscopic staging system is a simple and reproducible way for selection of a suitable therapy. It allows for adequate stratification of the main risk factors in the setting of clinical trials evaluating preoperative treatments.
Key Words: Gastric adenocarcinoma • Stomach cancer • Laparoscopy • Preoperative staging • Resectability • Neoadjuvant treatments
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Surgery is the only known curative treatment. But often it is an insufficient treatment for most patients in stages III and IV because low curative resection rates and high recurrence rates are common.3 Recently, neoadjuvant treatments have received extensive attention as an attempt to increase the rate of complete tumor resection and to prolong survival.4–7 However, an accurate pretreatment staging system is lacking, even when it is a prerequisite for the design and interpretation of these trials.8
Endoscopy, computed tomography (CT), endoscopic ultrasonography (EUS), and laparoscopy have been reported to be useful in the selection of candidates for surgical resection or identification of occult metastatic disease.9,10 No adequate correlation of the clinical staging methods available and the definitive surgical-pathological tumor-node-metastasis (TNM) staging system has been achieved because surgical resection of the stomach and perigastric lymph nodes is required.
Laparoscopy has been used in GC as part of the pretherapeutic staging protocol and to perform palliative derivations or even gastrectomy.11 However, its role in the selection of patients for resection or for the best therapeutic option has not been definitively stated. The aim of the present study is the proposal of a simple staging system based on laparoscopic findings only (Table 1) in order to define a treatment algorithm before surgical treatment is undertaken.
|
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The preoperative staging protocol included clinical history, physical examination, chest x-rays, blood cell count, blood coagulation tests, liver function tests, blood chemistry, urinalysis, serum Ca 19–9, carcinoembryonic antigen, endoscopy, abdominal CT scan, and EUS.
Patients with tumors <4 cm or with tumor invasion limited to mucosa or submucosa by this preoperative protocol did not undergo laparoscopy, given the low probability of unsuspected metastasis, and a surgical resection trial was performed. Patients with distant metastatic disease underwent palliative or experimental treatments, and laparoscopy was not performed, nor was it performed in patients with gastric outlet obstruction because gastro-jejunal bypass was indicated.
Informed consent was obtained in all cases.
Laparoscopy Technique
The surgeon is located at the patient’s right side and the monitor is beside the patient’s left shoulder (Fig. 1). Pneumoperitoneum is developed using the Veress technique, with gas pressure of approximately 14 mm of mercury through a small infraumbilical incision. The first trocar (10 mm) is inserted through the same incision and the peritoneal cavity is entered and examined with 0° or 45° vision telescopes. The pelvis is inspected in Trendelenburg’s position, and liquid is aspirated for cytological examination or peritoneal washing with 500-ml saline is performed. Inverted Trendelenburg’s position is established and the upper abdomen is visualized. A second (5 mm) or even a third (10 mm) trocar is inserted in the right hypochondrium and the left flank, respectively, as shown in Figure 1. Inspection of pelvic cavity, liver surfaces, gastro-hepatic and gastro-colic omentum, right and left paracolic space, and inferior surface of the transverse mesocolon and mesenteric root is performed.
|
All patients with potentially resectable disease, with or without invasion to adjacent organs, are immediately converted to open surgery for a formal resection trial. Patients with peritoneal implants or distant metastasis are considered unresectable and laparotomy is not performed unless palliative resection appears feasible.
A prospective description of T and M classifications based on laparoscopic findings only is recorded at the moment of laparoscopy. The operator does not know the EUS and CT findings.
Laparoscopic Stages
The proposed staging system consisted of four stages (Table 1). The first included early GC and tumors without serosal involvement (visualization of a normal serosal surface). The second involved patients with serosal involvement and feasible surgical resection. Serosal invasion was defined as any change in the color or texture of the serosal surface of the gastric wall, together with abnormal patterns of vascularity. The third involved tumors with direct invasion of adjacent structures or organs. The fourth category encompassed patients with metastatic disease. This laparoscopic staging system was compared with the tumor-node-metastases (TNM) staging system defined by the American Joint Committee on Cancer (AJCC).13
Statistical Analysis
The accuracy of laparoscopic diagnosis of metastasis was calculated, and the gold standard was the surgical pathology demonstration of metastatic lesions (liver, ovarian, peritoneal, or superior mesenteric lymph nodes). Accuracy for the laparoscopic diagnosis of resectability was calculated, and the gold standard for resectability was the surgical resection trial. Liver, ovarian, peritoneal, and superior mesenteric lymph node metastasis, invasion to great vessels, hepatic artery, celiac trunk, aorta, mesenteric artery, or vein were considered sufficient criteria for unresectability. However, isolated liver metastasis and limited peritoneal metastasis (H1 and P1 metastatic disease) were considered relative criteria of resectability.12 Likelihood ratios and diagnostic sensitivity and specificity were calculated as a measure of accuracy.14–16
The laparoscopic T classification was correlated with the surgical-pathologic T classification and the weighted kappa index was calculated as a measure of agreement.17
Follow-up times were calculated from the first visit to the hospital to the last visit recorded in the charts or until death. Patients lost to follow-up were localized by phone. Survival curves were constructed using the Kaplan-Meier method, and differences were compared using the log-rank test18. Cox proportional hazards model was used for multivariate analysis. Risk ratios (RR) with 95% confidence intervals (CI) were calculated as a measure of association.19 The independent variables used in the analysis were TNM stage (coded as 1, 2, 3, 4); laparoscopic stage (1, 2, 3, 4); age; sex (female, 0; male, 1); Laurén classification (intestinal, 1; diffuse or mixed, 2); differentiation grade (good, 1; moderate, 2; poor, 3); basal seralbumin level; absolute lymphocyte count; peritoneal cytology (negative, 0; positive, 1); and resection (not performed, 0; performed, 1). Probability values of .05 or less were considered significant. SPSS version 8.0 (1997) software (SPSS, Inc., Chicago, IL) was used for computations.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
One hundred fifty-one laparoscopies were attempted on 151 patients (34.4% of 439). In 148 of these patients (98%), the procedure was performed successfully. In three patients (2%), it was not possible to develop pneumoperitoneum even after open visualization of the peritoneal cavity because of extensive peritoneal disease, and laparoscopy was unsuccessful. Table 2 shows the demographic characteristics of the 151 patients who constitute the database for this study.
|
Resectability Prediction
Resectability was believed possible by laparoscopy in 96 patients; however, subsequent laparotomy proved resectability in only 63 patients (33 false positives). Fifty-four patients were considered to have true negative diagnoses of resectability and false negative was considered in only one case with incorrectly suspected peritoneal implants. The likelihood ratio for a positive laparoscopic diagnosis of resectability was 2.6 and 0.03 for a negative diagnosis, with a sensitivity of 98.4% and specificity of 62%.
Categorization of T Classification
T classification by laparoscopy was compared with T classification by surgical pathology study of the resected specimen. Only 63 patients with surgical resection (palliative or curative) and complete surgical pathology study were included. Agreement was found in 53 patients and weighted kappa was 0.82 (Table 3).
|
Complications of Laparoscopy
Self-limited, low-degree fever was recorded in four patients in the 1st to 3rd day after laparoscopy (2.6%). Insufflation of the preperitoneal space was produced in three cases (2%). No other complications or mortality attributable to laparoscopy were recorded.
In two patients, progression of malignant disease in the sites of insertion of the first trocar was recorded. In both cases, that finding represented only progressive and extensive peritoneal disease. No case of recurrent tumor activity located at the sites of trocar insertion was recorded in the group of patients with radical gastrectomy performed after staging laparoscopy.
Survival by Laparoscopic Stages
Survival by laparoscopic stages is shown in Fig. 3. All 151 patients had a median survival time of 23 months (mean, 26.5 months; 95% CI, 21.9–31.2; interval from 0.1 to 48 months). They had a 56.7%, 49.7%, and 45.2% survival at 1, 2, and 3 years, respectively.
|
None of the variables tested additionally (sex, age, Laurén classification, differentiation grade, seralbumin, absolute lymphocyte count, and peritoneal cytology) presented any independent effect on prognosis.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
CT scan is an excellent method for diagnosing liver metastases20 but it is inaccurate for analytical definition of T classification. Moss and coworkers21 defined a simple and practical staging system based solely on CT-scan findings, but the practical usefulness of this has not been reported.
EUS has been extensively studied for evaluation of T classification.22–24 A staging accuracy for T classification of 67%–92% has been reported.23,24 An agreement of 92% for EUS categorization of depth of invasion compared with surgical pathology has been reported.23
However, we found by laparoscopy that 44% of 151 patients presented downstaging by clinical studies (CT scan and EUS) because of missing peritoneal or liver metastasis. Laparoscopy is probably the best diagnostic test for distant metastasis, particularly if combined with CT scan.
Our results support that T and M classification can be defined in the pretherapeutic setting with a high degree of accuracy. On the other hand, CT scan and EUS are inaccurate for evaluating nodal status.23,24 Moreover, these studies used the old 1987 TNM classification, in which nodal status can be defined with the location of positive lymph nodes.
N classification is impossible to define without lymphadenectomy with removal of at least 15 lymph nodes, according to the new 1997 TNM classification of the AJCC.13 N staging should be regarded as a postoperative prognostic factor even when CT scan, EUS, or laparoscopy may identify lymph node metastasis. Therefore, from a practical standpoint, pretherapeutic staging methods must obviate N classification.
The main advantage of staging laparoscopy over CT scan and EUS is that it may obviate nontherapeutic laparotomy.25 The former is the only study that can identify small liver or peritoneal metastasis. The entire intra-abdominal space can be observed from a small wound, including the lesser sac,26 and the clinical stage of the patient can be accurately defined. It is easy to obtain ascites or tissue samples of peritoneal implants in the abdominal wall and of suspicious lymph nodes. Laparoscopy reduces surgical stress and postoperative pain. Patient mobilization is not as limited after operation and has a positive impact on rehabilitation time, hospital stay, and return to social activities.25 However, randomized clinical trials comparing staging laparoscopy with staging laparotomy have not been performed.
Many patients with laparoscopic stage I do not even require laparoscopy because surgical or endoscopic resection is certainly required and resectability is virtually 100%. EUS is probably very useful in this setting, identifying small tumors with mucosal or submucosal involvement. Early GC was accurately differentiated from advanced GC in 90% of cases by EUS.23–24
Serosal involvement defines the second laparoscopic stage, which also entertains feasible surgical resection. However, a high probability of recurrence suggests that these patients should be included in research trials of adjuvant or even neoadjuvant treatments.5
We believe that the role of EUS is reduced in patients with the bulky tumors found in laparoscopic stage III. Advanced locoregional disease with invasion to adjacent structures or organs is associated with approximately 50% resectability (palliative resections included). Neoadjuvant multimodal treatments should be undertaken to increase complete resection rates.
The fourth laparoscopic stage includes patients with metastatic disease who should be included in research projects with new drugs or new multimodal approaches.
The proposed staging system (Table 4) is a simplification of the TNM staging system and is not intended to be a substitute. It should be regarded as a tool for the selection of the best therapeutic option for the specific patient and also for pretherapeutic stratification of risk factors in the setting of new randomized clinical trials (Table 4).
|
Laparoscopic ultrasound has been reported as useful in patients with GC. It may additionally increase the excellent specificity for the combination of laparoscopy and CT scan in the diagnosis of liver metastasis.9,28,29 However, it may add little to the prognostic information provided by the staging system described here.
Peritoneal cytology has identified as an important prognostic factor in GC.30,31 However, its value in selection of patients for specific treatment approaches remains unknown. Peritoneal cytology was not used for staging purposes in our study because an independent prognostic value was not found. The reason could be the small sample size or the selection group of patients for laparoscopic staging.
In this study, nearly 70% of patients underwent laparotomy after staging laparoscopy. This procedure had important consequences only for 30% of patients who underwent laparoscopy and only for 11% of the total group of 439 cases. However, during this 5-year period we performed a total evaluation of the laparoscopic staging protocol and most patients had undergone a laparotomy to validate laparoscopic findings. It is possible that during the next years, fewer patients with complete staging laparoscopy will undergo laparotomy at the institute.
Self-limited fever of indefinite origin was the most common complication after staging laparoscopy in GC in this study, and no major complications directly attributable to the procedure were found.
The data presented here suggest that port site metastasis after laparoscopy in GC seldom occurs and that this is merely an indicator of the progression of widespread peritoneal disease. On the other hand, the clinical course of patients with metastatic or advanced locoregional disease evident on laparoscopy is unlikely to be affected by port site tumor progression itself. Laparoscopic port site is a known potential site of metastases and risk must be weighed against potential benefit of the procedure in GC.32,33
Surgical resection was the most important prognostic factor and both the TNM staging system and the laparoscopic staging system added prognostic value to the final models. No other prognostic factors were identified. The explanation for this is the short follow up of the cohort. When a 5-year median follow-up is completed, it will be possible for the models to find the prognostic value of other variables.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In population with high incidence of early GC in which preoperative examinations may accurately stage disease, staging laparoscopy should be used infrequently.
The results shown in this study demonstrate an excellent correlation of laparoscopic findings with the final histological specimen. Pretherapeutic staging laparoscopy is a simple and reproducible stratifying method. However, a confirmatory evaluation as the best strategy for selection of suitable therapy is warranted.
|
Acknowledgments |
|---|
Received for publication November 9, 2000. Accepted for publication April 13, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. Mori, T. Suzuki, H. Uozaki, H. Nakanishi, T. Ueda, Y. Matsuno, Y. Kodera, H. Sakamoto, N. Yamamoto, M. Sasako, et al. Detection of Minimal Gastric Cancer Cells in Peritoneal Washings by Focused Microarray Analysis with Multiple Markers: Clinical Implications Ann. Surg. Oncol., May 1, 2007; 14(5): 1694 - 1702. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
