Radiographic and Clinical Factors Associated with Improved Outcomes in Advanced Cancer Patients with Bowel Obstruction

Abstract

Background:

The purpose of this study was to identify preoperative clinical and radiographic factors relevant to treatment selection and outcomes in patients with advanced cancer presenting with bowel obstruction.

Methods:

Clinical and radiographic data were retrospectively obtained from records of inpatients with suspected bowel obstruction referred for palliative surgical consultation (2000–2006). Patients were stratified according to site of obstruction: gastric outlet obstruction (GOO), small bowel obstruction (SBO), and large bowel obstruction (LBO). We utilized the Cox proportional hazards model to identify preoperative clinical and radiologic variables associated with overall survival (OS).

Results:

Of 191 patients, the site of obstruction was classified as GOO in 41 (21%), SBO in 122 (64%), and LBO in 28 (15%). Almost half of the patients (47%) had received systemic therapy in the 6 weeks prior to evaluation. The most common sites of disease identified on imaging included abdominal visceral metastases (37%), carcinomatosis/sarcomatosis (46%), and an intact primary tumor or recurrence (31%). Most patients (62%) exhibited 2 or more sites of disease on imaging. Treatment strategies included nonoperative/nonprocedural management in 41% (n = 79), endoscopic/interventional radiology procedures in 25% (n = 48), and surgery in 34% (n = 64). Median OS for the cohort was 3.5 months (95% confidence interval [CI]: 2.7–4.6). Median OS for GOO, SBO, and LBO was 2.7 (95% CI: 1.7–4.1), 3.5 (95% CI: 2.5–4.9), and 7.0 (95% CI: 2.1–11) months, respectively (p = 0.17). Adverse prognostic factors for OS included endoscopic/interventional radiology procedures and ≥3 radiologically evident sites of disease.

Conclusions:

OS, although low, was not significantly different among GOO, SBO, and LBO. Single sites of disease identified on imaging were not associated with OS, although multiple sites of disease were associated with diminished OS.

Introduction

Bowel obstruction is one of the most common indications for general surgical evaluation. The approach to the management of patients with bowel obstruction generally follows accepted surgical principles in an attempt to minimize morbidity and mortality associated with bowel ischemia. The initial evaluation of patients with bowel obstruction involves determining the anatomic site(s) of obstruction, which are often classified as gastric outlet obstruction (GOO), small bowel obstruction (SBO), and large bowel obstruction (LBO). A critical aspect of the evaluation is to determine the degree of obstruction (i.e., partial versus complete). A thorough history and physical exam followed by abdominal imaging often provide surgeons with the needed information to develop a treatment algorithm.¹

Cancer patients may be more prone than patients in the general surgical population to developing bowel obstruction due to previous surgery, tumor invasion, or the effects of radiation. These factors, in addition to recent chemotherapy administration, steroids, and chronic pain medication can add to the complexity of treatment for bowel obstruction in oncology patients. Furthermore, palliative surgical treatment may be a reasonable approach for the alleviation of symptomatic obstruction in patients with known advanced or incurable malignancy. The frequency of palliative surgical intervention has been reported to account for 12.5% of surgical procedures and up to 1000 procedures per year at tertiary cancer centers.^2,3 However, the frequency of palliative surgical consultation is clearly much higher as it is estimated that only a quarter of all patients undergoing palliative surgical evaluation ultimately undergo surgical intervention, with bowel obstruction representing the most common indication for consultation.⁴ Given that almost half of all surgical consultations at major cancer centers are palliative in nature and that bowel obstruction is the most common indication, it is important to examine the treatment patterns and outcomes for these patients.

The majority of studies on bowel obstruction in advanced cancer patients have been limited, focusing on patients presenting with a single site of obstruction, reporting relatively small numbers of patients, and assessing outcomes in only those patients undergoing surgical intervention. The frequency with which varying treatment strategies are utilized and the corresponding outcomes for patients undergoing palliative surgical consultation for bowel obstruction is currently unknown. In addition, few studies have examined clinical and radiologic factors present at the time of initial evaluation which may be associated with outcomes.

The purpose of this study was to identify current practice patterns with respect to the frequency of operative, radiologic/endoscopic procedures, and nonoperative/nonprocedural management of advanced cancer patients presenting with bowel obstruction. In addition, we sought to identify preoperative clinical and radiographic factors relevant to treatment selection and outcomes.

Patients and Methods

Patients

The study population consists of all patients undergoing surgical consultation for possible bowel obstruction and met the criteria for palliative surgical consultation between January 2000 and September 2006 at The University of Texas M. D. Anderson Cancer Center. Surgical consultations were retrospectively identified from Current Procedural Terminology billing data as well as the departmental registry of consultations placed through the Department of Surgery's administrative offices. Palliative surgical consultation was defined as a consultation for symptoms attributable to an advanced or incurable malignancy, including symptoms attributable to treatment complications of an advanced or incurable malignancy.⁴ Patients evaluated for potentially curative surgery were excluded. The study was approved by the University of Texas M. D. Anderson Cancer Center Institutional Review Board.

Clinicopathologic variables

A detailed review of eachpatient's medical record was performed for demographic variables and malignancy type (hematologic versus solid tumor). Disease stage was determined by a two-physician review panel (BB and JC) and categorized as advanced or incurable. Incurable disease was defined as unequivocal evidence of surgically unresectable or incurable disease. Advanced disease was defined as that in patients with an aggressive malignancy who were deemed potentially curable. Clinicopathologic variables pertaining to the anatomic site of bowel obstruction, administration of chemotherapy or biochemotherapy within the previous 6 weeks, the presence of neutropenia, and radiographic extent of disease were extracted from each patient's medical record. Bowel obstruction site was categorized as GOO, SBO, or LBO. Neutropenia was defined as an absolute neutrophil count (ANC) <1000 cells/uL. To classify the radiographic extent of disease, available reports including computed tomography, magnetic resonance imaging, and/or positron emission tomography were reviewed in detail. The radiographic extent of disease was classified as: abdominal visceral metastases, carcinomatosis/sarcomatosis, intact primary tumor or local recurrence, ascites, lung metastases, abdominal/chest lymphadenopathy reported as suspicious for malignancy, bone metastases, brain metastases, and extraabdominal (subcutaneous/muscle/other soft tissue) metastases. The number of radiologically evident sites of disease was classified as being 1, 2, or ≥ 3.

Recommendations for treatment were based on multidisciplinary discussions among surgical oncology, medical oncology, and other involved services. The medical records of patients who underwent surgery were reviewed to identify postoperative morbidity and mortality rates and length of stay.

Statistical analysis

Overall survival (OS) was defined as the time from surgical evaluation to death from any cause or last follow-up. Median OS was estimated using the methods of Kaplan and Meier.⁵ Log-rank tests were performed to compare OS.⁶ The Cox proportional hazards model was used to examine the association between various clinicopathologic factors and OS.⁷ Computations were carried out using Stata for Windows, Release 10 (StataCorp LP; College Station, TX). A p value of less than or equal to 0.05 was considered statistically significant.

Results

Demographics and treatment

Two hundred and seventy-one patients were referred to the surgical consultation service for the evaluation of abdominal pain, nausea/emesis, abdominal distention, or bowel obstruction. A total of 80 patients without a clear diagnosis of bowel obstruction were excluded from the study cohort: 45 patients with abdominal pain of unclear etiology, 13 patients with combined pain and nausea/emesis, 10 patients with abdominal distention without radiologic evidence of obstruction, 6 patients with nausea/emesis only, and 6 patients with colonic pseudoobstruction. The remaining 191 patients with a diagnosis of bowel obstruction formed the study cohort, which consisted of 41 patients with GOO, 122 patients with SBO, and 28 patients with LBO.

The demographics and clinical variables for the study cohort are described in Table 1. The median age was 58 years with similar gender distribution. The majority of the patients had a diagnosis of solid organ malignancy (87%) and were classified as having incurable disease (85%). Specific cancer types included colorectal cancer [n = 49 (26%)], hepatobiliary [n = 32 (17%)], genitourinary [n = 25 (13%)], sarcoma [n = 15 (8%)], lymphoma [n = 14 (7%)], miscellaneous solid organ [n = 14 (7%)], gastroesophageal [n = 12 (6%)], unknown primary [n = 7 (4%)], leukemia [n = 6 (3%)], breast [n = 4 (2%)], melanoma [n = 3 (2%)], multiple cancer types [n = 3 (2%)], lung [n = 3 (2%)], miscellaneous hematologic [n = 2 (1%)], and gynecologic [n = 2 (1%)]. Almost half (47%) of the patients had received systemic therapy within the 6 weeks prior to evaluation, but only a minority (5%) were neutropenic at the time of evaluation. Radiographic imaging demonstrated the extent of disease to include abdominal visceral metastases (37%), carcinomatosis/sarcomatosis (46%), an intact primary tumor or local recurrence (31%), ascites (21%), and lung metastases (24%). Over half of the patients (62%) had 2 or more radiographically evident sites of tumor present at the time of evaluation.

Table 1.

Demographic and Clinical Variables for Study Cohort (n = 191)

Variable	No.	%
Median age—years (range)	57.5 (18–87)
Sex
Male	103	54
Female	88	46
Malignancy type
Hematologic	22	12
Solid	166	87
Both	3	1
Disease stage
Advanced	29	15
Incurable	162	85
Anatomic site of bowel obstruction
Gastric outlet	41	21
Small bowel	122	64
Large bowel	28	15
Chemo/biotherapy within 6 weeks
Yes	89	47
No	102	53
Neutropenic at time of evaluation
Yes	9	5
No	182	95
Radiographic extent of disease
Abdominal visceral metastases	71	37
Carcinomatosis/sarcomatosis	87	46
Intact primary tumor or local recurrence	60	31
Ascites	40	21
Lung metastases	45	24
Lymphadenopathy	35	18
Bone metastases	16	8
Brain metastases	5	3
Extraabdominal soft tissue metastases	5	3
Number of radiographic sites of disease
1	65	34
2	72	38
≥3	46	24
n/a	8	4

Surgery was performed in 44% (18/41), 25% (30/122), and 57% (16/28) of patients with GOO, SBO, and LBO, respectively. Endoscopic/interventional radiology procedures were more frequently performed in patients with GOO (34%) than in patients with SBO (24%) or LBO (18%). Percutaneous gastric tubes were utilized in 24% of patients with GOO and 23% of patients with SBO. Nonoperative/nonprocedural management was recommended most often in patients with SBO (52%). Specific surgical and procedural interventions stratified by obstruction location are outlined in Table 2. Morbidity and mortality rates for surgical intervention were 44% and 5%, respectively. Median postoperative length of stay for surgical patients was 11 days (range 4–61).

Table 2.

Treatment Rendered Stratified by Anatomic Site of Obstruction

Variable	GOO (n = 41)	SBO (n = 122)	LBO (n = 28)
Surgical treatment	18 (44%)	30 (25%)	16 (57%)
Bowel resection	–	8 (7%)	2 (7%)
Bowel resection with ostomy	–	1 (1%)	1 (4%)
Intestinal bypass	5 (12%)	5 (4%)	1 (4%)
Intestinal bypass with ostomy	–	3 (2%)	–
Intestinal bypass with feeding tube	8 (20%)	4 (3%)	–
Intestinal bypass with ostomy and feeding tube	–	–	1 (4%)
Ostomy formation	–	4 (3%)	9 (32%)
Ostomy formation with feeding tube	–	1 (1%)	1 (4%)
Feeding tube	5 (12%)	1 (1%)	1 (4%)
Lysis of adhesions	–	3 (2%)	–
Endoscopic or IR procedures	14 (34%)	29 (24%)	5 (18%)
IR percutaneous G-tube	9 (22%)	23 (19%)	1 (4%)
Endoscopic percutaneous G-tube	1 (2%)	5 (4%)	–
Endoscopic stent	4 (10%)	1 (1%)	2 (7%)
Endoscopic tumor ablation	–	–	1 (4%)
Percutaneous G-tube and stent	–	–	1 (4%)
Nonoperative/Nonprocedural management	9 (22%)	63 (52%)	7 (25%)

GOO, gastric outlet obstruction; LBO, large bowel obstruction; SBO, small bowel obstruction.

Overall survival

Median OS for the entire cohort was 3.5 months (95% CI: 2.7–4.6). Median OS for GOO, SBO, and LBO was 2.7 (95% CI: 1.7–4.1), 3.5 (95% CI: 2.5–4.9), and 7.0 (95% CI: 2.1–11) months, respectively (p = 0.17) as shown in Fig. 1. Fig. 2 represents the Kaplan-Meier survival curves stratified according to treatment strategy. The median OS for patients treated with surgery, interventional/endoscopic procedures, and nonoperative/nonprocedural management was 6.7 (95% CI: 3.5–8.5), 1.9 (95% CI: 1.2–3.1), and 2.9 (95% CI: 1.7–5.0) months, respectively (p ≤ 0.01). OS for each anatomic site of obstruction stratified by patients selected for various treatment strategies are presented in Fig. 3. Of note is that there were significant differences in OS based on treatment strategy for patients with SBO (p < 0.01) and GOO (p = 0.05) but not for patients with LBO (p = 0.86).

FIG. 1.

Kaplan-Meier estimates of overall survival probability for the 191 patients who met our criteria for bowel obstruction stratified by anatomic site of obstruction.

FIG. 2.

Kaplan-Meier estimates of overall survival probability for the 191 patients who met our criteria for bowel obstruction stratified according to treatment strategy.

FIG. 3.

Kaplan-Meier estimates of overall survival probability for patients with (A) GOO, (B) SBO, and (C) LBO stratified according to treatment strategy.

The results of the univariate analysis for OS are presented in Table 3. Surgical treatment, interventional radiologic/endoscopic treatment, lymphadenopathy on imaging, ascites, carcinomatosis/sarcomatosis, and ≥ 3 radiologic sites of disease were associated with OS. On multivariate analysis, patients selected for interventional radiologic/endoscopic treatment and patients demonstrating ≥ 3 radiologic sites of tumor at the time of presentation had poorer OS (Table 4).

Table 3.

Univariate Analysis for Factors Associated with Overall Survival

Variable	HR	95% CI	P
Age	0.99	0.99–1.0	0.7
Male sex	0.99	0.74–1.34	0.99
Solid organ malignanacy	0.92	0.57–1.49	0.75
Hematologic malignancy	1.04	0.67–1.64	0.85
Incurable malignancy	1.35	0.87–2.1	0.17
Bowel obstruction location
Gastric outlet	1.3	0.89–1.88	0.17
Large bowel	0.81	0.52–1.25	0.34
Small bowel	Referent	Referent	Referent
Treatment
Surgical procedure	0.68	0.48–0.97	0.03
Interventional/Endoscopic procedure	1.56	1.07–2.27	0.02
Nonoperative/Nonprocedural	Referent	Referent	Referent
Chemo/biotherapy within 6 weeks	1.11	0.82–1.5	0.51
Neutropenia	1.3	0.66–2.55	0.46
Imaging extent of disease
Intact primary/recurrent tumor	1.03	0.74–1.43	0.86
Abdominal visceral metastases	1.32	0.96–1.8	0.09
Abdominal/chest lymphadenopathy	0.65	0.43–0.97	0.03
Ascites	1.61	1.11–2.32	0.01
Bone metastases	1.34	0.79–2.28	0.29
Brain metastases	0.61	0.22–1.64	0.29
Lung metastases	0.9	0.63–1.29	0.56
Carcinomatosis/sarcomatosis	1.73	1.26–2.36	<0.01
Subcutaneous/muscle/soft tissue metastases	1.01	0.41–2.47	0.98
Number of disease sites on imaging
1	Referent	Referent	Referent
2	1.34	0.94–1.93	0.11
≥3	1.75	1.17–2.62	<0.01

CI, confidence interval; HR, hazard ratio.

Table 4.

Multivariate Analysis for Factors Associated with Overall Survival

Variable	HR	95% CI	P
Treatment
Surgical procedure	0.87	0.59–1.28	0.49
Interventional/Endoscopic procedure	1.72	1.15–2.56	<0.01
Nonoperative/Nonprocedural	Referent	Referent	Referent
Imaging extent of disease
Abdominal/chest lymphadenopathy	0.72	0.46–1.11	0.14
Ascites	1.2	0.8–1.8	0.37
Carcinomatosis/sarcomatosis	1.28	0.9–1.82	0.17
Number of disease sites on imaging
1	Referent	Referent	Referent
2	1.15	0.77–1.71	0.51
≥3	1.73	1.11–2.69	0.02

CI, confidence interval; HR, hazard ratio.

Discussion

In this study of advanced cancer patients presenting with bowel obstruction, almost half had received recent systemic therapy and also had evidence of carcinomatosis/sarcomatosis. The majority of patients had 2 or more radiologically evident sites of disease. Surgery was performed in 34% of patients and 25% of patients were referred for an endoscopic or IR procedure. As anticipated with an advanced cancer cohort, the median OS was short (3.5 months). There was an association noted between poor OS and an increasing number of sites of tumor at presentation and endoscopic/IR procedures.

Our findings are in alignment with several studies that have focused on patients with known advanced or incurable cancer presenting with bowel obstruction. A recent prospective observational study in patients with malignant GOO and incurable disease reported on 47 patients undergoing surgery (n = 16), stent placement (n = 24), or percutaneous endoscopic gastrostomy tube placement (n = 7).⁸ This study is notable in that all patients underwent operative or procedural intervention. The median overall survival was only 2.1 months, which is similar to the reported OS of 2.7 months in patients with GOO in the current study. A smaller study (n = 35) of patients with noncurative stage IV cancer and malignant bowel obstruction by Wright and colleagues reported a similar median survival of 2.7 months but did not detect a significant association between preoperative factors such as carcinomatosis and ascites and survival.⁹ Twelve patients (34%) in this study underwent surgical treatment, which is similar to our reported surgical intervention rate of 25% (although interventional or endoscopic procedures were not described). Several additional studies have confirmed the poor survival of advanced cancer patients with bowel obstruction, reporting median survival rates of 3 to 5 months.^10–12

Surgeons are often asked to evaluate patients with advanced or incurable malignancies with the goal of achieving palliation of symptoms. There is significant selection bias in how advanced cancer patients are selected for nonoperative, procedural, or surgical intervention that relies on a physician's prediction of life expectancy and 30-day postoperative morbidity and mortality rate. Reported postoperative mortality rates ranging from 9% to 40% and morbidity rates ranging from 9% to 90% highlight the significant risk associated with surgical intervention in similar patients.¹³ Prospective studies are needed that identify true contemporary risks, benefits (including patient-reported outcomes), and costs to guide future treatment algorithms for this commonly encountered clinical scenario. This study sought to examine a relatively large cohort of patients with advanced and incurable malignancies and to focus on the identification of pretreatment variables that may be potentially useful in selecting future patients for surgical intervention. Blair et al., in a study of patients with bowel obstruction and carcinomatosis, identified a noncolorectal cancer primary as a variable associated with inferior survival.¹⁰ Other studies have reported that ascites, presence of carcinomatosis, extent of peritoneal involvement, functional status, weight loss, laboratory variables, and the interval from cancer diagnosis to the presentation of bowel obstruction as predictors of outcome.^9,10,14,15

Our findings extend the current body of literature by identifying an additional prognostic factor, the number of disease sites on preoperative imaging, which should be considered when determining treatment for bowel obstruction in advanced cancer patients. Although the current study size is relatively large compared with similar studies, multi-institutional collaboration and prospective evaluation of homogeneous populations will be required prior to defining evidence-based treatment guidelines. Identification of patients at the time of hospital admission for bowel obstruction will be helpful in determining the true rates of interventional or endoscopic intervention. Our study is limited in that it included only patients undergoing surgical consultation, which may account for the low percentage of patients undergoing endoscopic stent placement. In addition to the refinement of study cohorts, future prospective studies are needed to identify the optimal outcomes and metrics that should be considered for advanced cancer patients with bowel obstruction. Outcomes that have been considered include the ability to tolerate oral intake, discharge disposition, durability of palliation, symptom improvement defined retrospectively, postoperative mortality and morbidity, and quality of life.^3,16,17 A recent international conference on malignant bowel obstruction proposed a primary endpoint of “good days” defined as days out of the hospital without nasogastric tube decompression or need for intravenous hydration.¹⁸ Qualitative research has been useful in identifying the complexity of these clinical scenarios, and future studies are needed to provide further insight into outcomes of importance other than physical well-being such as the social impact of symptoms and a need to maintain hope.¹⁹

In conclusion, overall survival is dismal for advanced cancer patients presenting with bowel obstruction regardless of the anatomic site of obstruction. In this series of patients, the most commonly utilized treatment for GOO and LBO was surgical exploration. For patients with SBO, however, treatment was primarily nonoperative and nonprocedural. Percutaneous G-tubes are frequently used for palliation in GOO and SBO; 24% and 23%, respectively. Multiple tumor sites on radiologic imaging was associated with diminished overall survival. With the optimal patient-reported outcome and primary endpoint unclear, future studies are needed to determine the optimal outcome measure for advanced cancer patients presenting with bowel obstruction.

Footnotes

Acknowledgments

Presented at the Sixth Annual Academic Surgical Congress in Huntington Beach, California on February 1–3, 2011.

Author Disclosure Statement

No competing financial interests exist.

References

Helton

, Fisichella

. Intestinal Obstruction. In: Souba W (ed): ACS Surgery. New York: WebMD, 2007; 514–533.

Krouse

, Nelson

, Farrell

, Grube

, Juarez

, Wagman

, Chu

. Surgical palliation at a cancer center: Incidence and outcomes. Arch Surg, 2001; 136:773–778.

Miner

, Brennan

, Jaques

. A prospective, symptom related, outcomes analysis of 1022 palliative procedures for advanced cancer. Ann Surg, 2004; 240:719–726.

Badgwell

, Smith

, Liu

, Bruera

, Curley

, Cormier

. Indicators of surgery and survival in oncology inpatients requiring surgical evaluation for palliation. Support Care Cancer, 2009; 17:727–734.

Kaplan

, Meier

. Nonparametric estimator from incomplete observations. J Am Statistical Assoc, 1958; 53:457–481.

Mantel

. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep, 1966; 50:163–170.

Cox

. Regression models and life tables (with discussion) J Royal Statistical Soc, 1972; Series B:122–187.

Schmidt

, Gerdes

, Hawkins

, Zucker

, Zhou

, Riedel

, Jaques

, Markowitz

, Coit

, Schattner

. A prospective observational study examining quality of life in patients with malignant gastric outlet obstruction. Am J Surg, 2009; 198:92–99.

Wright

, Chakraborty

, Helyer

, Moravan

, Selby

. Predictors of survival in patients with non-curative stage IV cancer and malignant bowel obstruction. J Surg Oncol, 2010; 101:425–429.

10.

Blair

, Chu

, Schwarz

. Outcome of palliative operations for malignant bowel obstruction in patients with peritoneal carcinomatosis from nongynecological cancer. Ann Surg Oncol, 2001; 8:632–637.

11.

Tang

, Davis

, Silberman

. Bowel obstruction in cancer patients. Arch Surg, 1995; 130:832–836.

12.

Turnbull

, Guerra

, Starnes

. Results of surgery for obstructing carcinomatosis of gastrointestinal, pancreatic, or biliary origin. J Clin Oncol, 1989; 7:381–386.

13.

Ripamonti

, Twycross

, Baines

, Bozzetti

, Capri

, De Conno

, Gemlo

, Hunt

, Krebs

, Mercadante

, Schaerer

, Wilkinson

. Working Group of the European Association for Palliative Care: Clinical-practice recommendations for the management of bowel obstruction in patients with end-stage cancer. Support Care Cancer, 2001; 9:223–233.

14.

Jong

, Sturgeon

, Jamieson

. Benefit of palliative surgery for bowel obstruction in advanced ovarian cancer. Can J Surg, 1995; 38:454–457.

15.

Woolfson

, Jennings

, Whalen

. Management of bowel obstruction in patients with abdominal cancer. Arch Surg, 1997; 132:1093–1097.

16.

Krouse

. Surgical management of malignant bowel obstruction. Surg Oncol Clin N Am, 2004; 13:479–490.

17.

Podnos

, Juarez

, Pameijer

, Uman

, Ferrell

, Wagman

. Surgical palliation of advanced gastrointestinal tumors. J Palliat Med, 2007; 10:871–876.

18.

Whalen

, Kutner

, Byock

, Gerard

, Stovall

, Sieverding

, Ganz

, Krouse

. Implementing palliative care studies. J Pain Symptom Manage, 2007; 34:S40–S48.

19.

Ferrell

, Chu

, Wagman

, Juarez

, Borneman

, Cullinane

, McCahill

. Online exclusive: patient and surgeon decision making regarding surgery for advanced cancer. Oncol Nurs Forum, 2003; 30:E106–E114.