Abstract
Background
The optimal management of gastric outlet obstruction (GOO) due to gastric cancer (GC) is unclear. We examined the relationships between clinical and management variables and outcomes in GC patients with GOO.
Methods
GOO management and clinical course were reviewed in patients with GC and GOO. Cox regression and Kaplan-Meier analyses were used to identify variables predictive of overall survival (OS).
Results
The study included 59 patients. Eleven had imaging evidence of metastasis, and 35 had pathologically confirmed peritoneal disease. Initial management included resection in 23 patients, feeding jejunostomy ± decompressive gastrostomy (JT/GT) in 25, surgical gastrojejunostomy in 5, and endoscopic intervention in 6. Seven patients with initial JT/GT underwent resection after neoadjuvant therapy. Median OS(95%CI) was 21.4 (0.0–45.1)months in the up-front resection group (median follow-up, 14.7months) and not reached in those with initial JT/GT, neoadjuvant therapy, and later resection (median follow-up, 26.5months)(P=0.18). On multivariable analysis, clinically positive nodes (hazard ratio[HR]:3.76; 95%CI:1.17–12.12; P=0.03), metastasis on CT (HR:3.97; 95%CI:1.53–10.26;P=0.01) and resection (HR:0.37; 95%CI:0.17–0.79;P=0.01) independently predicted OS.
Conclusion
In GOO due to GC, OS is similar after treatment with up-front resection compared with JT/GT, neoadjuvant therapy, and later resection. Up-front JT/GT may allow patients to tolerate chemotherapy and improve selection for gastrectomy.
Keywords: Stomach neoplasms, Gastric outlet obstruction, Gastrectomy, Palliative Care
INTRODUCTION
A multidisciplinary treatment approach is standard in the management of gastric cancer (GC). Perioperative therapy is routinely used, often in the form of chemotherapy with or without chemoradiation.[1]
Patients who present with gastric outlet obstruction (GOO) due to GC present a clinical challenge. GOO is defined as a blockage of the distal stomach or duodenum. Patients often present with symptoms related to impaired gastric emptying due to the obstruction, including nausea, vomiting, inability to tolerate oral intake, and anorexia, which in many cases prevent patients from tolerating perioperative therapy.[2] Further, patients with GOO often present with a history of significant weight loss, malnutrition, dehydration, and electrolyte abnormalities, which make patients suboptimal candidates for a complex abdominal operation.[2] Additionally, GOO due to GC is associated with advanced disease and carries a poor prognosis, raising the question of what benefit, if any, is to be gained from an up-front surgical resection.[3–5] Possible management strategies for symptom relief in patients with GOO due to GC include endoscopic stent placement, surgical resection, surgical gastrojejunostomy bypass, and decompressive gastrostomy and feeding jejunostomy tube (JT) placement.[6] Often the primary goal of symptom relief is to improve the patient’s tolerance of systemic therapy, which may increase the possibility of a resection with curative intent, which in turn may improve the overall prognosis.
The optimal management strategy for GOO due to GC is unknown. The purpose of this study was to determine the relationships between clinical and management variables and outcomes in patients with GOO due to GC.
METHODS
Patients who presented to our institution with GOO due to GC from January 1995 through April 2019 were identified through a search of a prospectively maintained departmental database. GOO was defined as biopsy-proven gastric adenocarcinoma with 1) evidence of GOO on cross-sectional imaging (gastric dilation with or without evidence of pyloric, antral, or duodenal narrowing); 2) evidence of GOO on endoscopy (gastric distention with retention of fluid and/or food particles or presence of a mass with stricturing, narrowing, or stenosis of the pyloric channel); or 3) clinical symptoms consistent with GOO (nausea, vomiting, change in oral diet, or weight loss). Clinicopathologic data were obtained from the electronic medical record. Patients were excluded if they had previously undergone cancer-specific therapy for GC.
For each patient, the primary management strategy for GOO was identified and the clinical course was evaluated through review of the electronic medical record. The primary management strategies were consolidated into 4 groups: 1) up-front resection, 2) placement of a feeding jejunostomy tube with or without a decompressive gastrostomy tube (JT/GT), 3) surgical gastrojejunostomy (GJ), and 4) endoscopic intervention, including stenting or pyloric dilation. Curative resection was defined as resection of the primary tumor with no known evidence of distant metastatic disease or peritoneal disease, performed with curative intent rather than for palliation of symptoms. Palliative resection was defined as resection of the primary tumor with the goal of symptom management in the presence of distant metastatic or peritoneal disease or incomplete resection of the primary tumor owing to invasion of adjacent structures. Serosal invasion was defined as 1) clinical T4 tumor on endoscopic ultrasonography (EUS), 2) serosal invasion on intraoperative visual inspection, as described in the operative report, or 3) serosal invasion confirmed on pathologic examination.
We retrospectively assessed symptom improvement or persistence following GOO management through review of clinical notes at the time of discharge, and at first follow-up clinic visit. Patients were considered to have symptom improvement if pre-treatment symptoms were noted to have improved in the clinical documentation.[7,8] Nutritional intake was also assessed from clinical notes at the time of discharge.
Categorical variables are reported as proportions or percentages, and continuous variables are reported as mean (standard deviation) or median (interquartile range) where appropriate. Analysis of variance and Fisher’s exact test were used to compare baseline characteristics between the primary management groups. Kaplan-Meier and Cox regression analyses were used to evaluate the impact of management and other clinicopathologic factors on overall survival (OS). Variables with P ≤ 0.05 on univariate Cox regression analysis were included in the multivariable analysis. Backwards stepwise modeling was performed to identify factors that were independently predictive of OS. OS was measured from the date of initial GOO treatment to most recent follow-up date or death. Median OS was calculated by Kaplan-Meier analysis, and differences in survival were compared by log-rank test. All tests of significance were 2-sided, and statistical significance was identified by P < 0.05.
The Institutional Review Board of The University of Texas MD Anderson Cancer Center approved this study.
RESULTS
We identified 59 patients presenting with GOO due to GC and no prior cancer-related treatment. The mean (standard deviation) age was 64 (13) years. Thirty-three patients (56%) were male, and 26 (44%) were Caucasian. The patients’ demographic and clinicopathologic data are summarized in Table 1. Twenty-three patients (39%) were unable to undergo EUS staging owing to the obstructive nature of the primary tumor and retained food particles in the stomach. Of the 36 patients who did undergo EUS, 23 (64%) had clinical T3 and 10 (28%) had clinical T4 tumors. The majority of patients had either moderately differentiated tumors (8 patients; 14%) or poorly differentiated tumors (49 patients; 83%). Thirty-three patients (56%) had signet ring cell histology, and 8 (14%) had features of linitis plastica.
Table 1.
Demographic and Clinicopathologic Characteristics of Patients Presenting with Gastric Outlet Obstruction due to Gastric Cancer.
| N (%) | ||
|---|---|---|
| Age (Mean) | 64 ±13 | |
| Gender | Male | 33 (55.9%) |
| Race | White | 26 (44.1%) |
| Black | 9 (15.3%) | |
| Asian | 7 (11.9%) | |
| Hispanic | 13 (22.0%) | |
| Native American | 4 (6.8%) | |
| Histologic Grade | Moderate | 8 (13.6%) |
| Poor/Dedifferentiated | 49 (83.1%) | |
| Unknown | 2 (3.4%) | |
| Signet ring cells | 33 (55.9%) | |
| Linitis plastica | 8 (13.6%) | |
| Clinical T stage | T2 | 2 (3.4%) |
| T3 | 23 (39.0%) | |
| T4 | 10 (16.9%) | |
| Tx | 1 (1.7%) | |
| Unable to perform | 23 (39.0%) | |
| Clinical N stage | N0 | 11 (18.6%) |
| N1 | 18 (30.5%) | |
| N2 | 1 (1.7%) | |
| N3 | 2 (3.4%) | |
| Nx | 4 (6.8%) | |
| Unable to perform | 23 (39.0%) | |
| M stage | M1 | 37 (62.7%) |
| Metastasis on CT | 11 (18.6%) | |
| Serosal invasion | 36 (61%) | |
| Carcinomatosis | 28 (47.5%) | |
| Positive cytology only | 7 (11.9%) | |
| Ascites | 12 (20.3%) | |
| GOO on imaging | 40 (67.8%) | |
| GOO on endoscopy | 42 (71.2%) | |
| GOO treatment up front | Up-front Resection | 23 (39.0%) |
| JT/GT | 25 (42.4%) | |
| Surgical GJ | 5 (8.5%) | |
| Endoscopic Intervention | 6 (10.2%) | |
| Other Treatments | none | 14 (23.7%) |
| Chemotherapy | 22 (37.3%) | |
| Chemoradiation | 6 (10.2%) | |
| Chemo/Chemoradiation | 6 (10.2%) | |
| Chemo/Radiation | 6 (10.2%) | |
| Chemo/HIPEC | 5 (8.5%) | |
Symptoms at the time of presentation included nausea and/or vomiting (49 patients; 83%), weight loss (48 patients; 81%), abdominal pain (30 patients; 51%), anorexia and/or early satiety (29 patients; 49%), gastroesophageal reflux and/or increased eructation (21 patients; 36%), symptoms of gastrointestinal bleeding (melena or hematemesis) (11 patients; 19%), and dysphagia (4 patients; 7%). Diagnosis of GOO was confirmed by imaging findings in 40 patients (68%) and on endoscopy in 42 patients (71%). Representative images from cross-sectional imaging and endoscopy demonstrating GOO are shown in Figure 1.
Figure 1.
Representative images of gastric outlet obstruction on cross-sectional imaging and endoscopy. (a) Axial slice of computed tomography imaging demonstrating thickened pylorus (arrowheads) resulting in gastric outlet obstruction, gastric distention, and retention of gastric contents (arrow). (b) Endoscopic image demonstrating large pyloric tumor resulting in narrowing of gastric outlet.
At presentation, 36 patients (61%) had evidence of serosal invasion on EUS, visual inspection, or pathologic examination, and 37 patients (63%) had metastatic disease, although only 11 (19%) had evidence of metastasis on computed tomography (CT). The most common site of metastatic disease was the peritoneum (35 patients; 59%), in the form of either carcinomatosis (28 patients; 47%) or positive peritoneal cytology (7 patients; 12%) on staging laparoscopy or at attempted resection. Liver metastases were seen in 4 patients (7%), and distant nodal disease was seen in 3 patients (5%). Univariate logistic regression analysis identified no variables that were significantly predictive of metastasis at the time of initial presentation with GOO (data not shown).
Initial management of GOO was based on physician preference and clinical presentation. It included up-front resection in 23 patients, JT/GT in 25 patients (JT and venting GT in 11; JT alone in 14), surgical GJ in 5 patients, and endoscopic intervention in 6 patients (stent placement in 5, endoscopic dilation only in 1) (Figure 2). The up-front resection group included 1 patient who had only JT/GT initially but underwent resection with curative intent during the index hospitalization after improvement of nutritional parameters; 1 patient who had a JT placed initially and underwent palliative resection 18 days later because of gastric perforation secondary to the tumor; and 1 patient who was treated initially with endoscopic dilation and underwent palliative resection 6 days later because of bleeding from the tumor. Of the 23 patients in the up-front resection group, 11 had resection with curative intent, and 12 had palliative resection. Demographic and clinicopathologic characteristics by up-front treatment group are summarized in Table 2. Patients undergoing resection were less likely to have linitis plastica, ascites, or metastatic disease, including evidence of metastasis on CT. There were no differences between the four up-front treatment groups in histologic grade, signet ring cells, clinical T or N category, or peritoneal disease.
Figure 2.
Clinical Course in Patients with Gastric Outlet Obstruction (GOO) Resulting from Gastric Cancer (GC). Chemo indicates chemotherapy with or without chemoradiation therapy; GJ, gastrojejunostomy; JT/GT, jejunostomy tube with or without decompressive gastrostomy tube; and XRT, radiation therapy alone.
Table 2.
Demographic and Clinicopathologic Characteristics of Patients Presenting with Gastric Outlet Obstruction due to Gastric Cancer, by Initial Management.
| N(%) | Resection (n=23) | JT/GT (n=25) |
GJ (n=5) |
Endoscopy (n=6) |
p | |
|---|---|---|---|---|---|---|
| Age (Mean ± SD) | 62±15 | 65 ± 13 | 63 ± 8 | 63±10 | 0.87 | |
| Gender | Male | 13 (56.5%) | 14 (56.0%) | 4 (80.0%) | 2 (33.3%) | 0.55 |
| Race | White | 8 (43.5%) | 14 (56.0%) | 1 (20.0%) | 3 (50.0%) | 0.39 |
| Black | 3 (13.0%) | 4 (16.0%) | 1 (20.0%) | 1 (16.7%) | ||
| Asian | 4 (17.4%) | 2 (8.0%) | 0 (0.0%) | 1 (16.7%) | ||
| Hispanic | 4 (17.4%) | 5 (20.0%) | 3 (60.0%) | 1 (16.7%) | ||
| Native American | 4 (17.4%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | ||
| Histologic Grade | Moderate | 3 (13.0%) | 5 (20.0%) | 0 (0.0%) | 0 (0.0%) | 0.88 |
| Poor | 19 (82.6%) | 19 (76.0%) | 5 (100%) | 6 (100%) | ||
| Unknown | 1 (4.3%) | 1 (4.0%) | 0 (0.0%) | 0 (0.0%) | ||
| Signet ring | 13 (56.5%) | 12 (48.0%) | 4 (80%) | 4 (66.7%) | 0.63 | |
| Linitis plastica | 0 (0.0%) | 4 (16.7%) | 0 (0.0%) | 4 (66.7%) | <0.01 | |
| Metastasis on CT | 0 (0.0%) | 6 (24.0%) | 2 (40.0%) | 3 (50.0%) | <0.01 | |
| Serosal invasion | 17 (73.9%) | 14 (56.0%) | 3 (60.0%) | 2 (33.3%) | 0.03 | |
| Carcinomatosis | 7 (30.4%) | 14 (56.0%) | 4 (80.0%) | 3 (50.0%) | 0.13 | |
| Positive cytology only | 2 (20.0%) | 3 (12.0%) | 0 (0.0%) | 2 (33.3%) | 0.51 | |
| Ascites | 0 (0.0%) | 7 (30.4%) | 1 (20.0%) | 4 (66.7%) | <0.01 | |
| Clinical T stage | T2 | 2 (8.7%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0.51 |
| T3 | 10 (43.5%) | 10 (40.0%) | 1 (20.0%) | 2 (33.3%) | ||
| T4 | 2 (8.7%) | 7 (28.0%) | 0 (0.0%) | 1 (16.7%) | ||
| Tx | 0 (0.0%) | 1 (4.0%) | 0 (0.0%) | 0 (0.0%) | ||
| Unable to perform | 9 (39.1%) | 7 (28.0%) | 4 (80.0%) | 3 (50.0%) | ||
| Clinical N stage | N0 | 4 (17.4%) | 5 (20.0%) | 0 (0.0%) | 2 (33.3%) | 0.60 |
| N1 | 9 (39.1%) | 8 (32.0%) | 0 (0.0%) | 1 (16.7%) | ||
| N2 | 0 (0.0%) | 1 (4.0%) | 0 (0.0%) | 0 (0.0%) | ||
| N3 | 0 (0.0%) | 2 (8.0%) | 0 (0.0%) | 0 (0.0%) | ||
| Nx | 1 (4.3%) | 2 (8.0%) | 1 (20.0%) | 0 (0.0%) | ||
| Unable to perform | 9 (39.1%) | 7 (28.0%) | 4 (80.0%) | 3 (50.0%) | ||
| M stage | M1 | 9 (39.1%) | 17 (68.0%) | 5 (100%) | 5 (83.3%) | 0.02 |
| Additional Cancer-directed treatments | No | 6 (26.1%) | 4 (16.0%) | 0 (0%) | 0 (0%) | 0.43 |
| Yes | 17 (73.9%) | 21 (84.0%) | 5 (100.0%) | 6 (100.0%) | ||
| Initial Surveillance | 7 (30.4%) | 0 (0.0%) | 1 (20.0%) | 0 (0.0%) | 0.01 | |
| Initial Chemotherapy or Chemoradiation | 10 (43.5%) | 21 (84.0%) | 4 (80.0%) | 6 (100%) | ||
Symptom improvement at discharge or at first follow-up visit was noted in 41 patients (69.5%), and specifically in 21 patients (91.3%) after resection, 9 patients (81.8%) with JT and venting GT, 3 patients (21.4%) with JT alone, 4 patients (80.0%) with surgical GJ and 4 patients (66.7%) with endoscopic interventions. Six more patients (42.9%) managed with JT alone without initial improvement, did demonstrate improvement of symptoms during the course of chemotherapy. There were two patients treated with JT alone who had persistent or worsening GOO symptoms during chemotherapy, requiring additional intervention, including resection in one, and endoscopic stenting in the other. All patients in all four groups were able to maintain their nutritional needs after upfront GOO management, either by oral intake alone, oral intake with supplemental JT feedings, or JT feedings alone.
After receiving various combinations of chemotherapy, chemoradiotherapy, radiation therapy, and/or hyperthermic intraperitoneal chemotherapy (HIPEC), 7 (28%) of the 25 patients initially treated with JT/GT and 2 (33%) of the 6 patients initially treated endoscopically ultimately underwent later resection (Figure 2). Overall, 18 patients (30%) underwent curative resection and 14 patients (24%) underwent a palliative resection.
Following initial GOO management, the majority of patients underwent cancer-directed chemotherapy, chemoradiotherapy, or cancer-specific surveillance. A total of 10 patients (16.9%) including 6 in the upfront resection group (26.1%) and 4 in the JT/GT group (16.0%) (p=0.43, Table 2) did not undergo cancer-directed treatment, either due to patient choice (n=5, 8.5%), poor postoperative recovery or complication (n=3, 5.1%), or transition to hospice (n=2, 3.4%). Six of 7 patients in the resection group (30.4%) who initially underwent surveillance, were eventually treated with chemotherapy at a later time when imaging demonstrated progression.
The median follow-up time was 11.4 months for the entire cohort and 14.7 months in surviving patients. The median OS for the entire cohort was 11.9 months (95% CI, 8.7–15.1). The median OS was 21.4 months (95% CI, 0.0–45.1) in the up-front resection group, 9.3 months (95% CI, 4.7–13.9) in the JT/GT group, 7.5 months (95% CI, 7.0–8.0) in the surgical GJ group, and 15.5 months (95% CI, 10.1–20.9) in the endoscopic intervention group (P = 0.04).
We performed a subset analysis to assess for differences in OS between 1) patients who underwent up-front resection, 2) patients who underwent JT/GT first followed by neoadjuvant therapy and then later resection, and 3) patients who underwent JT/GT first but did not undergo later resection. The demographic and clinicopathologic characteristics of these groups are summarized in Table 3. Notably, the groups differed in rates of metastasis on CT scan, peritoneal disease, and ascites at presentation. No patient in the up-front JT/GT group who had evidence of metastasis on CT scan at presentation underwent later resection. Further, the majority of patients in the JT/GT group with positive cytology or carcinomatosis did not undergo later resection. Four patients with peritoneal disease in the JT/GT group underwent later resection after chemotherapy and HIPEC. All patients in the two resection groups underwent subtotal gastrectomy.
Table 3.
Demographic and Clinicopathologic Characteristics of Patients with Gastric Outlet Obstruction due to Gastric Cancer by Resection Status.
| Up-front Resection (n=23) | Later Resection (n=7) | No resection (n=18) | p | ||
|---|---|---|---|---|---|
| Age (Mean ± SD) | 62 ± 15 | 63 ± 11 | 66 ± 14 | 0.81 | |
| Male | 13 (56.5%) | 6 (85.7%) | 8 (44.4%) | 0.19 | |
| Race | White | 8 (34.8%) | 4 (57.1%) | 10 (55.6%) | 0.42 |
| Black | 13 (13.0%) | 0 (0.0%) | 4 (22.2%) | ||
| Asian | 4 (17.4%) | 1 (14.3%) | 1 (5.6%) | ||
| Hispanic | 4 (17.4%) | 2 (28.6%) | 3 (16.6%) | ||
| Native American | 4 (17.4%) | 0 (0.0%) | 0 (0.0%) | ||
| Tumor Grade | Moderate | 3 (13.0%) | 3 (42.9%) | 2 (11.1%) | 0.35 |
| Poor | 19 (82.6%) | 4 (57.1%) | 15 (83.3%) | ||
| unknown | 1 (4.3%) | 0 (0.0%) | 0 (0.0%) | ||
| Signet ring cells | 13 (56.5%) | 3 (42.9%) | 9 (50.0%) | 0.79 | |
| Linitis plastica | 0 (0.0%) | 1 (14.3%) | 3 (17.6%) | 0.07 | |
| Clinical T stage | T2 | 2 (8.7%) | 0 (0.0%) | 0 (0.0%) | 0.54 |
| T3 | 10 (43.5%) | 3 (42.9%) | 7 (38.9%) | ||
| T4 | 2 (8.7%) | 3 (42.9%) | 4 (22.2%) | ||
| Tx | 0 (0.0%) | 0 (0.0%) | 1 (5.6%) | ||
| No EUS | 9 (39.1%) | 1 (14.3%) | 6 (33.3%) | ||
| Clinical N stage | N0 | 4 (17.4%) | 3 (42.9%) | 2 (11.1%) | 0.43 |
| N1 | 9 (39.1%) | 1 (14.3%) | 7 (38.9%) | ||
| N2 | 0 (0.0%) | 0 (0.0%) | 1 (5.6%) | ||
| N3 | 0 (0.0%) | 1 (14.3%) | 1 (5.6%) | ||
| Nx | 1 (4.3%) | 1 (14.3%) | 1 (5.6%) | ||
| No EUS | 9 (39.1%) | 1 (14.3%) | 6 (33.3%) | ||
| M stage | M1 | 10 (43.4%) | 4 (57.1%) | 13 (72.2%) | 0.11 |
| Metastasis on CT | 0 (0.0%) | 0 (0.0%) | 6 (33.3%) | <0.01 | |
| Livers mets | 1 (4.3%) | 0 (0.0%) | 2 (11.1%) | 0.74 | |
| Carcinomatosis | 7 (30.4%) | 1 (14.3%) | 13 (72.2%) | <0.01 | |
| Positive cytology only | 2 (8.7%) | 3 (42.9%) | 0 (0.0%) | <0.01 | |
| Ascites | 0 (0.0%) | 1 (14.3%) | 6 (33.3%) | <0.01 | |
| Additional Treatments | None | 7 (30.4%) | 0 (0.0%) | 5 (27.8%) | 0.27 |
| Chemo only | 9 (39.1%) | 1 (14.3%) | 6 (33.3%) | ||
| Combo Chemo/Radiation | 7 (30.4%) | 2 (28.6%) | 7 (38.9%) | ||
| Chemo/HIPEC | 0 (0.0%) | 4 (57.1%) | 0 (0.0%) | ||
Figure 3 depicts the survival curves of these patient groups. The median follow-up times were 11.9 months for the full subgroup of patients initially treated with resection or JT/GT, 14.7 months for the patients with up-front resection, 26.5 months for the patients with initial JT/GT and later resection, and 6.8 months for the patients with initial JT/GT without later resection. The median OS was 21.4 months (95% CI, 0.0–45.1) for the patients with up-front resection, not reached for the patients with initial JT/GT with later resection, and 7.03 months (95% CI, 4.1–10.0) for the patients with initial JT/GT without later resection. The 1-, 2-, and 3-year OS rates were 60.9%, 45.7%, and 39.1%, respectively, for the patients with up-front resection; 80.0%, 80.0%, and 60.0%, respectively, for the patients with initial JT/GT with later resection (P = 0.177); and 23.5%, 0.0%, and 0.0%, respectively, for the patients with initial JT/GT without later resection (P = 0.003).
Figure 3.
Kaplan-Meier Estimates of Overall Survival in Patients with Gastric Outlet Obstruction Due to Gastric Cancer by Management Strategy. UR indicates up-front resection; LR, up-front jejunostomy tube and/or gastrostomy tube (JT/GT) placement with later resection; and NR, up-front JT/GT placement without later resection. P = 0.177, UR vs LR; P = 0.003, UR vs NR; P = 0.002, LR vs NR.
Univariate Cox regression analysis identified clinically positive lymph nodes (HR: 3.69 95% CI: 1.24–10.96; P = 0.02), CT evidence of metastasis (HR: 3.89 95% CI: 1.86–8.14; P < 0.01), up-front management with surgical GJ (HR: 4.27 95% CI: 1.46–12.46; P < 0.01), liver metastasis at diagnosis (HR: 4.58 95% CI: 1.56–13.48; P < 0.01), carcinomatosis (HR: 2.88 95% CI: 1.47–5.65; P < 0.01), and presence of ascites (HR: 2.81 95% CI: 1.30–6.11; P < 0.01) as predictive of worse OS and resection at any point in the treatment course as predictive of improved OS (HR: 0.23 95% CI: 0.12–0.47; P < 0.01) (Table 4). Other than surgical GJ, up-front treatment strategies (resection, JT/GT, and endoscopic intervention) were not predictive of OS. Variables with a P-value less than 0.05 were included in the multivariable analysis and backwards stepwise modeling was performed. In the final multivariable analysis, clinically positive lymph nodes (HR: 3.76; 95% CI: 1.17–12.12; P = 0.03), CT evidence of metastasis (HR: 3.97; 95% CI: 1.53–10.26; P < 0.01) and resection at any time in the treatment course (HR: 0.37; 95% CI: 0.17–0.79; P = 0.01) were found to be independently predictive of OS (Table 4).
Table 4.
Univariate and Multivariable Cox Regression Analysis of Association between Demographic and Clinicopathologic Variables and Overall Survival in Patients Presenting with Gastric Outlet Obstruction due to Gastric Cancer.
| Univariate | Multivariable | ||||||
|---|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | ||
| Age | 1.02 | 0.996–1.05 | 0.10 | ||||
| Gender | Female | 1.70 | 0.93–3.09 | 0.08 | |||
| Race | White | Ref | 0.64 | ||||
| Black | 1.36 | 0.54–3.43 | 0.51 | ||||
| Asian | 0.60 | 0.20–1.76 | 0.35 | ||||
| Hispanic | 0.95 | 0.44–2.05 | 0.89 | ||||
| Native American | 0.58 | 0.19–1.75 | 0.34 | ||||
| Tumor Differentiation | Moderate | Ref | |||||
| Poor | 3.21 | 0.98–10.49 | 0.05 | ||||
| Signet Ring Cells | 1.04 | 0.57–1.90 | 0.90 | ||||
| Linitis Plastica | 1.95 | 0.81–4.73 | 0.14 | ||||
| Clinical T stage | T2 | Ref | |||||
| T3 | 1.08 | 0.25–4.74 | 0.92 | ||||
| T4 | 0.88 | 0.18–4.30 | 0.88 | ||||
| Clinically positive LNs | no | Ref | 0.06 | Ref | 0.07 | ||
| yes | 3.69 | 1.24–10.96 | 0.02 | 3.76 | 1.17–12.12 | 0.03 | |
| EUS not done | 3.27 | 1.11–9.64 | 0.03 | 2.16 | 0.71–6.62 | 0.18 | |
| CT metastasis | 3.89 | 1.86–8.14 | <0.01 | 3.97 | 1.53–10.26 | <0.01 | |
| Up-front Treatment | Resection | Ref | 0.07 | 0.50 | |||
| JT/GT | 1.65 | 0.82–3.34 | 0.16 | 0.70 | 0.19–2.59 | 0.60 | |
| GJ | 4.27 | 1.46–12.46 | <0.01 | 1.03 | 0.19–5.47 | 0.97 | |
| endoscopic | 1.49 | 0.48–4.57 | 0.49 | 0.34 | 0.06–1.87 | 0.21 | |
| Resection at any time | 0.23 | 0.12–0.47 | <0.01 | 0.37 | 0.17–0.79 | 0.01 | |
| Chemotherapy | 0.79 | 0.36–1.73 | 0.55 | ||||
| Serosal invasion | 1.16 | 0.55–2.44 | 0.69 | ||||
| Metastasis at diagnosis | 1.60 | 0.85–3.00 | 0.14 | ||||
| Liver metastasis | 4.58 | 1.56–13.48 | <0.01 | 0.99 | 0.17–5.65 | 0.99 | |
| Carcinomatosis | 2.88 | 1.47–5.65 | <0.01 | 2.13 | 0.99–4.60 | 0.05 | |
| Ascites | 2.81 | 1.30–6.11 | <0.01 | 1.09 | 0.18–6.67 | 0.93 | |
| Year of Diagnosis/Treatment | Before 2010 | Ref | |||||
| 2010 or later | 0.817 | 0.44–1.52 | 0.52 | ||||
Given the long study period, we analyzed changes in upfront treatment management strategy based on year of presentation. Patients treated for GOO prior to the year 2010 (n=29, 49.2%) were most commonly managed with upfront resection (n=20; 69.0%) followed by JT/GT (n=6, 20.7%) compared to those treated in 2010 or later (n=30, 50.8%), who were most commonly treated with upfront JT/GT (n=19, 63.3%) or endoscopic interventions (n=6, 20.0%) (p<0.001). The time period of treatment was not significantly associated with OS (Table 4).
DISCUSSION
Our study demonstrates that GC patients with GOO at initial presentation often have advanced disease, with high rates of serosal invasion and lymph node metastases as well as distant and peritoneal metastasis. Ours are similar to previously published findings, which have demonstrated that compared to patients without GOO, patients with GOO have more advanced stage with higher rates of serosal invasion, more aggressive pathologic features, and higher rates of lymph nodes metastases.[3–5, 9, 10] The rates of metastasis to the liver (7%) and the peritoneum (59%) in our study were also similar to those reported in previous studies (15% and 34%−40%, respectively).[5, 10]
It is interesting to note the limitations of EUS in this patient population for staging purposes. Thirty-nine percent of the patients in our cohort were unable to undergo EUS because of the obstructing nature of the tumor and inability to pass the scope or because of retention of large volumes of ingested gastric contents. Ikoma et al. previously reported the limitations of EUS in GC clinical staging with an accuracy of only 66% for identifying both T category and N category.[11] Our study suggests that there are additional limitations of EUS in patients presenting with GOO, which limits the staging information available when considering management options. While we did not specifically evaluate the accuracy, sensitivity, and specificity of EUS, this may be an interesting area for future research.
The initial management strategies in our patient cohort included endoscopic interventions, feeding JT/decompressive GT placement, surgical gastrojejunostomy, and resection with palliative or curative intent. The initial management strategy appeared to be influenced by multiple factors. Unsurprisingly, patients in our cohort who underwent up-front resection had lower rates of metastasis on imaging, carcinomatosis, and ascites than patients who had another initial management strategy. Collins et al. reported similar findings with a lower rate of distant metastasis in their gastrectomy group than in their non-resection groups.[12]
Standard practice in patients with locally advanced, potentially resectable GC is preoperative therapy; however, patients presenting with GOO may be unable to tolerate these treatments given the anatomic blockage and resulting malnutrition.[13] Selection of the optimal management strategy for these patients often comes down to a decision between up-front resection or another intervention to relieve the GOO symptoms so that the patient can ultimately tolerate systemic therapy and potentially undergo future resection. Our study found no significant difference in OS between patients who underwent up-front resection and those who were first managed with JT/GT and then underwent resection after neoadjuvant therapy. Approximately a quarter of patients in our study managed with upfront resection underwent no cancer-directed treatment after resection, and approximately a third were managed with surveillance only initially. While the number of patients who did not undergo cancer-directed treatment after resection did not differ significantly from that in the JT/GT group, previous studies have shown that as many as half of patients undergoing gastrectomy will not tolerate postoperative chemotherapy.[13–15] Alternatively, a majority of patients managed with initial JT/GT did not ultimately undergo resection, likely due in part to a high rate of metastasis but also due in part to disease progression during the course of systemic treatment. It seems likely that patients who are treated with initial JT/GT followed by systemic therapy will self-select for later resection by virtue of the nature of their disease and response to therapy. Further, neoadjuvant therapy may allow for tumor shrinkage and ultimately improve chances of a complete resection in patients with locally advanced tumors. This approach may allow patients at high risk for progression to avoid a major operation, while also potentially improving prognosis for patients who are ultimately eligible for delayed resection. While 100% of patients treated with surgical GJ or endoscopic interventions were treated with subsequent chemotherapy, the sample size in each of these groups was small, and only two patients underwent later resection, which limits analysis or conclusions regarding these two management options.
Our study also found that a significant number of patients who presented with GOO due to GC had peritoneal disease in the form of either carcinomatosis or positive cytology. Some of these patients underwent up-front palliative resection for relief of their symptoms despite their peritoneal disease. As regional therapies for peritoneal disease, particularly hyperthermic intraperitoneal chemoperfusion (HIPEC), continue to improve, treatment options available to these patients presenting with advanced disease will continue to evolve, with the potential for improving prognosis. Patients with peritoneal disease who undergo up-front resection may have delay of systemic or regionally targeted therapies, ultimately impacting overall prognosis. Management with nonsurgical interventions may improve symptoms and allow patients to tolerate systemic and regional therapies as a bridge to resection.
An additional consideration in choosing a management strategy for GOO is optimization of quality of life (QOL), particularly in patients with metastatic or unresectable disease. Few studies have evaluated QOL or patient-reported outcome measures following different GOO management strategies. Fujitani et al. evaluated changes in QOL measures after gastrectomy and surgical bypass for patients with metastatic or unresectable GC and GOO.[16] Global QOL measures were not improved after the surgical interventions, but gastric-specific symptoms, including dysphagia, reflux, eating restrictions, pain, anxiety, and dry mouth, were improved up to 3 months after surgery.[16] Schmidt et al. evaluated QOL changes after palliative interventions for patients with GOO secondary to any type of cancer and, similarly, showed improvement in gastric-specific symptoms after both endoscopic stenting and surgical bypass. Improvement was also seen in global health status and nausea/vomiting after stenting.[17] Importantly, both studies cited limitations in interpretation of the results due to participant drop-out secondary to high mortality and patient refusal to complete follow-up questionnaires, presumably due to high symptom burden. Schmidt et al. also asked patients to rank the reasons for their intervention choice and found that the biggest motivating factors in patient decision-making were surgeon recommendation and desire to eat and drink by mouth.[17] While QOL studies in palliative surgery patient populations are challenging because of short survival intervals and high symptom burden, understanding the factors that motivate patient decision-making may inform treatment discussions and recommendations for such patients.
Given the poor long-term prognosis in patients presenting with metastasis, the best management strategy may be the one that provides the most durable symptom relief with the lowest risk profile. Multiple studies comparing symptom relief and re-obstruction rates after endoscopic stenting and surgical bypass in metastatic or unresectable GC and GOO due to other malignancies have shown better short-term outcomes with stenting but better long-term outcomes after GJ bypass.[6, 18–22] As a result, some authors have recommended stenting for patients with anticipated short survival while indicating that bypass may be preferred for patients with better performance status and longer anticipated survival times.[6, 18] In our patient cohort, symptom improvement was achieved in 69.4% of patients. Fewer patients with JT alone demonstrated improvement than in other groups, including those with JT and GT. However, there is the potential for symptom improvement in these patients during the course of chemotherapy. Additionally, it is notable that all patients were able to tolerate adequate nutritional intake, either by mouth or by JT. The interpretation of these results is limited by the retrospective assessment of symptoms and nutritional intake via clinical chart review. This type of symptom assessment, while feasible, is difficult and imperfect, and is limited in its ability to fully capture quality or extent of symptom improvement.[7,8] Thus, we are unable to draw strong conclusions from the results of our study regarding the best management option for GOO symptomatology. This represents an important area of future research.
Our study has multiple limitations. By virtue of being a retrospective study, there is an inherent selection bias, with patients considered to have less advanced disease and better performance status more likely to be offered and to undergo an operation. A randomized trial evaluating outcomes of patients undergoing up-front resection or up-front JT/GT followed by neoadjuvant therapy and later resection would improve understanding of the best options for these patients. However, it would be difficult to justify a randomized trial like this as the nuances of each patient’s case warrant individualized treatment approaches based on patient presentation and patient goals of care. Secondly, our analysis cannot account for advances in systemic treatment regimens, imaging, or practice patterns during this period. The long time period included in the study also introduces additional selection bias for certain GOO management strategies in recent years, due to the increasing use of preoperative chemotherapy as a result of clinical trial data demonstrating improved survival with perioperative therapy.[13–15] As surgeons further integrate the use of preoperative therapy into their practice, our results can inform decision-making regarding upfront management in this challenging patient population. Our analysis was limited by the heterogeneity of the patient cohort in terms of management strategy and by the small number of patients. We also identified only a small number of patients treated with endoscopic interventions; this may be due to patients with advanced disease being referred directly to gastroenterologists without a surgical consultation, which would prevent inclusion in the departmental database used to identify patients for this study.
CONCLUSIONS
In conclusion, GC patients presenting with GOO represent a group with high rates of locally advanced disease with high frequency of lymph node and distant metastases and an overall poor prognosis. It is difficult to know which patients will benefit from up-front resection and which patients will have recurrence and progression within a short period of time. Up-front JT/GT may allow patients to better tolerate systemic therapies while improving nutritional status and may also facilitate selection of the patients who are the best candidates for gastrectomy or for regional therapies in the case of peritoneal metastasis. Optimization of QOL should also be a consideration in choosing a management strategy, and future research evaluating QOL and patient-reported outcomes as a treatment endpoint is necessary.
SYNOPSIS.
This retrospective review demonstrates high rates of advanced disease and metastasis in gastric cancer patients presenting with outlet obstruction. Survival outcomes were similar between patients treated with up-front resection and patients treated with up-front JT/GT, neoadjuvant therapy, and later resection.
Acknowledgements:
The authors would like to thank Stephanie Deming, in Scientific Publications, Research Medical Library at MD Anderson Cancer Center, for editing assistance with this manuscript.
Funding: This work was supported by National Institutes of Health grants T32CA009599 and P30CA016672, which supports the Clinical Trials Office; a No Stomach for Cancer Award for Gastric Cancer Research; and the Holly Clegg Gastric Cancer Research Fund at MD Anderson Cancer Center.
Footnotes
Portions of this work have been accepted for e-poster presentation at the Society of Surgical Oncology’s International Conference on Surgical Cancer Care Virtual Meeting, August 2020.
Commercial interests: The authors have no commercial interests in the subject of the work.
References
- 1.Badgwell B Multimodality Therapy of Localized Gastric Adenocarcinoma. JNCCN Journal of the National Comprehensive Cancer Network. 2016;14(10):1321–1326. [DOI] [PubMed] [Google Scholar]
- 2.Khullar S, DiSario J. Gastric outlet obstruction. Gastrointest Endosc Clin N Am. 1996;6(3):585–603. [PubMed] [Google Scholar]
- 3.Park SH, Mok YJ, Kim JH, Park SS, Kim SJ, Kim CS. Clinical significance of gastric outlet obstruction on the oncologic and surgical outcomes of radical surgery for carcinoma of the distal stomach. J Surg Oncol. 2009;100(3):215–221. [DOI] [PubMed] [Google Scholar]
- 4.Chen JH, Wu CW, Lo SS, et al. Outcome of distal gastric cancer with pyloric stenosis after curative resection. Eur J Surg Oncol. 2007;33(5):556–560. [DOI] [PubMed] [Google Scholar]
- 5.Watanabe A, Maehara Y, Okuyama T, Kakeji Y, Korenaga D, Sugimachi K. Gastric carcinoma with pyloric stenosis. Surgery. 1998;123(3):330–334. [PubMed] [Google Scholar]
- 6.Tringali A, Giannetti A, Adler DG. Endoscopic management of gastric outlet obstruction disease. Ann Gastroenterol. 2019;32(4):330–337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Pujara D, Chiang YJ, Cormier JN, Bruera E, Badgwell B. Selective Approach for Patients with Advanced Malignancy and Gastrointestinal Obstruction. J Am Coll Surg. 2017;225(1):53–59. [DOI] [PubMed] [Google Scholar]
- 8.Badgwell BD, Aloia TA, Garrett J, Chedister G, Miner T, Krouse R. Indicators of symptom improvement and survival in inpatients with advanced cancer undergoing palliative surgical consultation. J Surg Oncol. 2013;107(4):367–371. [DOI] [PubMed] [Google Scholar]
- 9.Mizutani S, Shioya T, Maejima K, et al. Significance of gastrectomy as palliative surgery for gastric carcinoma with pyloric stenosis. J Nippon Med Sch. 2007;74(3):241–245. [DOI] [PubMed] [Google Scholar]
- 10.Okumura Y, Yamashita H, Aikou S, et al. Palliative distal gastrectomy offers no survival benefit over gastrojejunostomy for gastric cancer with outlet obstruction: retrospective analysis of an 11-year experience. World J Surg Oncol. 2014;12:364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ikoma N, Lee JH, Bhutani MS, et al. Preoperative accuracy of gastric cancer staging in patient selection for preoperative therapy: race may affect accuracy of endoscopic ultrasonography. J Gastrointest Oncol. 2017;8(6):1009–1017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Collins A, Hatzaras I, Schmidt C, et al. Gastrectomy in advanced gastric cancer effectively palliates symptoms and may improve survival in select patients. J Gastrointest Surg. 2014;18(3):491–496. [DOI] [PubMed] [Google Scholar]
- 13.Badgwell B, Das P, Ajani J. Treatment of localized gastric and gastroesophageal adenocarcinoma : the role of accurate staging and preoperative therapy. J Hematol Oncol. 2017;10(149):1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med. 2006;355(1):11–20. [DOI] [PubMed] [Google Scholar]
- 15.Ychou M, Boige V, Pignon JP, et al. Perioperative chemotherapy compared with surgery alone for resectable gastroesophageal adenocarcinoma: an FNCLCC and FFCD multicenter phase III trial. J Clin Oncol. 2011;29(13):1715–1721. [DOI] [PubMed] [Google Scholar]
- 16.Fujitani K, Ando M, Sakamaki K, et al. Multicentre observational study of quality of life after surgical palliation of malignant gastric outlet obstruction for gastric cancer. BJS Open. 2017;1(6):165–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schmidt C, Gerdes H, Hawkins W, et al. A prospective observational study examining quality of life in patients with malignant gastric outlet obstruction. Am J Surg. 2009;198(1):92–99. [DOI] [PubMed] [Google Scholar]
- 18.Jeurnink SM, van Eijck CH, Steyerberg EW, Kuipers EJ, Siersema PD. Stent versus gastrojejunostomy for the palliation of gastric outlet obstruction: a systematic review. BMC Gastroenterol. 2007;7:18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ly J, O’Grady G, Mittal A, Plank L, Windsor JA. A systematic review of methods to palliate malignant gastric outlet obstruction. Surg Endosc. 2010;24(2):290–297. [DOI] [PubMed] [Google Scholar]
- 20.Park CH, Park JC, Kim EH, et al. Impact of carcinomatosis and ascites status on long-term outcomes of palliative treatment for patients with gastric outlet obstruction caused by unresectable gastric cancer: stent placement versus palliative gastrojejunostomy. Gastrointest Endosc. 2015;81(2):321–332. [DOI] [PubMed] [Google Scholar]
- 21.Jang SH, Lee H, Min BH, et al. Palliative gastrojejunostomy versus endoscopic stent placement for gastric outlet obstruction in patients with unresectable gastric cancer: a propensity score-matched analysis. Surg Endosc. 2017;31(10):4217–4223. [DOI] [PubMed] [Google Scholar]
- 22.No JH, Kim SW, Lim CH, et al. Long-term outcome of palliative therapy for gastric outlet obstruction caused by unresectable gastric cancer in patients with good performance status: endoscopic stenting versus surgery. Gastrointest Endosc. 2013;78(1):55–62. [DOI] [PubMed] [Google Scholar]