Fax: (713) 745-1163
Chemotherapy with 5-fluorouracil and a platinum compound improves outcomes in metastatic small bowel adenocarcinoma†
Version of Record online: 29 AUG 2008
Copyright © 2008 American Cancer Society
Volume 113, Issue 8, pages 2038–2045, 15 October 2008
How to Cite
Overman, M. J., Kopetz, S., Wen, S., Hoff, P. M., Fogelman, D., Morris, J., Abbruzzese, J. L., Ajani, J. A. and Wolff, R. A. (2008), Chemotherapy with 5-fluorouracil and a platinum compound improves outcomes in metastatic small bowel adenocarcinoma. Cancer, 113: 2038–2045. doi: 10.1002/cncr.23822
Presented at the American Society of Clinical Oncology 2007 Gastrointestinal Cancers Symposium, Orlando, Florida, January 19-21, 2007 [abstract 172]. This article contains original work not previously published.
- Issue online: 3 OCT 2008
- Version of Record online: 29 AUG 2008
- Manuscript Accepted: 9 JUN 2008
- Manuscript Revised: 6 JUN 2008
- Manuscript Received: 18 APR 2008
- small bowel adenocarcinoma;
Metastatic small bowel adenocarcinoma (SBA) has a poor prognosis. Because of the rarity of SBA, only a few studies have evaluated the role of chemotherapy in the treatment of metastatic SBA; thus, the benefit, if any, of adding a platinum compound to fluorouracil (5-FU) is unknown. The objective of this retrospective study was to determine whether the addition of a platinum compound to 5-FU provided any benefit in the treatment of patients with metastatic SBA.
The authors identified 80 patients with metastatic SBA who were treated with chemotherapy at the University of Texas M. D. Anderson Cancer Center between 1978 and 2005. Response rates, progression-free survival (PFS), and overall survival (OS) were compared between patients who received 5-FU and a platinum compound and patients who received other chemotherapy combinations.
The median patient age was 53 years. The primary tumor site was the jejunum in 35 patients (43%), duodenum in 30 patients (38%), ileum in 6 patients (8%), and nonspecified small bowel in 9 patients (11%). Of all 80 patients, 29 patients (36%) received 5-FU and a platinum compound, 41 patients (51%) received 5-FU without a platinum compound, and 10 patients (13%) received non-5-FU–based treatment. Compared with other chemotherapy regimens, treatment with 5-FU and a platinum agent resulted in a higher response rate (46% vs 16% with other regimens; P = .01) and longer median PFS (8.7 months vs 3.9 months; P ≤ .01) but not better OS (14.8 months vs 12 months; P = .1). In multivariate analysis, treatment with 5-FU and a platinum compound was a significant predictor of response (odds ratio, 4.5; 95% confidence interval [CI], 1.3-15.8; P = .02) and PFS (hazard ratio. 0.49; 95% CI, 0.29-0.84; P = .01) but only reached borderline significance for OS (hazard ratio, 0.63; 95% CI, 0.37-1.07; P = .08).
To the authors' knowledge, the current analysis represents the largest number of patients with metastatic SBA treated with chemotherapy in the literature, and the results suggested that the combination of 5-FU and a platinum compound leads to a higher response rate and PFS compared with other chemotherapy regimes. The authors concluded that prospective investigation of platinum analogues in the treatment of SBA is warranted. Cancer 2008. © 2008 American Cancer Society.
Small bowel adenocarcinoma (SBA) is an aggressive neoplasm with a poor prognosis. Although it is rare, with an estimated annual incidence in the United States of about 2400, SBA is the most common histologic subtype of small intestinal cancer.1–3 Because of the difficulties in diagnosing SBA, the majority of patients with SBA are not treated until the disease has reached an advanced stage. The outlook for patients with advanced SBA is poor: According to the National Cancer Data Base, the 5-year overall survival (OS) rates for patients with stage III or stage IV SBA are 35% and 4%, respectively.4
Because of the rarity of SBA, no randomized trials have evaluated the role of chemotherapy in this disease. Recently, retrospective studies have demonstrated that treating advanced SBA patients with palliative chemotherapy has a survival benefit compared with patients who do not receive palliative chemotherapy.5–7 In a retrospective study conducted by the British Columbia Cancer Agency, patients with locally advanced or metastatic SBA who received 5-fluorouracil (5-FU)-based chemotherapy had a longer median OS (15.6 months) than patients who received no chemotherapy (7.7 months).5 Another retrospective study demonstrated that chemotherapy for patients with inoperable or metastatic SBA resulted in an improved median OS (12 months) compared with no chemotherapy (2 months).6
The choice of chemotherapy has not been studied well for the treatment of SBA. A summary of studies that evaluated various chemotherapy combinations in the treatment of metastatic SBA is provided in Table 1. In 1965, Rochlin et al reported the first retrospective study to evaluate the role of systemic chemotherapy for metastatic SBA. In their study, 11 patients with SBA who had been treated at the University of California-Los Angeles Medical Center had a 36% response rate (RR) to single-agent fluorouracil (5-FU) treatment.8 Other retrospective studies have reported RRs for 5-FU–based treatment ranging from 5% to 18%.5, 9, 10 In 2005, the Eastern Cooperative Group conducted the largest prospective study of chemotherapy for SBA; 38 patients with SBA received 5-FU, doxorubicin, and mitomycin C and demonstrated an 18% RR and a median survival of 8 months.9
|Study||Study Type||No. of Patients||Chemotherapy||RR, %||OS, mo|
|Overman 200822||Phase 2||30||CAPOX||50||20.3|
|Fishman 20067||Retrospective||44||Various agents||29||18.6|
|Locher 200514||Retrospective||20||5-FU and platinum||21||14|
|Gibson 20059||Phase 2||38||FAM||18||8|
|Enzinger 200523||Phase 1||4||5-FU, cisplatin, irinotecan||50||NR|
|Czaykowski 20045||Retrospective||37||5-FU based||5||16|
|Goetz 200324||Phase 1||5||5-FU, oxaliplatin, irinotecan||40||NR|
|Morgan & Busuttil 197725||Retrospective||7||5-FU-based||0||NR|
The benefit observed from the addition of a platinum chemotherapy agent to 5-FU–based chemotherapy in both colon adenocarcinoma and gastric adenocarcinoma has suggested a possible role for this kind of chemotherapy in the treatment of SBA.11, 12 Researchers at the Royal Marsden Hospital reported a 37% RR and a median OS of 13 months for 8 patients with SBA who were treated with epirubicin, cisplatin, and 5-FU.13 A second retrospective study by Fishman et al from the Princess Margaret Hospital reported an RR of 23% for 11 patients who received platinum-based therapy.7 Another retrospective study reported a 21% RR for 20 patients with advanced SBA who were treated with a combination of 5-FU and a platinum agent.14 In an interim analysis of 21 patients from a phase 2 study of capecitabine plus oxaliplatin in the treatment of adenocarcinoma of the small bowel and the ampulla of Vater, we observed an RR of 52% and a median OS of 16.3 months.15 To clarify the optimal chemotherapy for patients with metastatic SBA, we undertook this retrospective study to determine the benefit, if any, of adding a platinum agent to 5-FU.
MATERIALS AND METHODS
Patients and Data Collection
We identified all patients from the M. D. Anderson Tumor Registry who were treated for metastatic SBA at the M. D. Anderson Cancer Center (M. D. Anderson) between 1978 and 2005. In total, 177 patients with metastatic SBA were identified, and 80 patients who met the following criteria were included in this report. Patients were required to have tumor histology that was reviewed at M. D. Anderson and was consistent with SBA. Only patients who were undergoing first-line chemotherapy for metastatic disease and who received ≥1 cycle of chemotherapy were included. Documentation of metastatic disease, either radiographically or at the time of surgery, was required. Patients had to have undergone a radiologic assessment of tumor response after the administration of chemotherapy. Six patients who had documented peritoneal involvement at surgery but did not have radiographically measurable disease were included in the study but were excluded from the univariate and multivariate analyses for treatment response. Chemotherapy treatment in these 6 patients consisted of 5-FU (3 patients), 5-FU and carboplatin (1 patient), 5-FU and cisplatin (1 patient), and 5-FU and oxaliplatin (1 patient). Patients who had tumors that involved the ampulla of Vater were excluded. Patients who were enrolled in an ongoing, prospective, phase 2 study of capecitabine and oxaliplatin for SBA were not included in this study.
Patient medical records were reviewed for information regarding demographic data, tumor characteristics, treatment types, treatment responses, and survival. Because of the long period covered by the study and because not all radiologic images were available for our review, response evaluation was based on the treating physician's response assessment. Tumor stage was determined according to the American Joint Committee on Cancer staging system.16 Histologic grading was determined according to the World Health Organization standard grading system: poorly differentiated, moderately differentiated, well differentiated, or undifferentiated. According to this system, tumors with 2 different degrees of histiologic differentiation are recorded as having the lesser grade. This analysis was approved by the M. D. Anderson Institutional Review Board.
We created descriptive summaries for each demographic and clinical variable. The following variables were examined in univariate and multivariate analyses for RR, progression-free survival (PFS), and OS: age, sex, histologic grade, primary site, stage at presentation, initial site of metastasis, surgery for primary tumor, prior adjuvant chemotherapy, and use of platinum-based or 5-FU–based chemotherapy for metastatic disease.
The Fisher exact test was used to assess the independence between 2 categorical variables. Survival curves were calculated from the start of chemotherapy using the Kaplan-Meier method. The 2-sided log-rank test was used to test the association between variables and OS and PFS. In the multivariate analysis, a Cox proportional hazards model was used to assess the simultaneous effect of ≥2 variables on OS and PFS. To compensate for advances in supportive care, radiation therapy techniques, and surgical techniques over the period studied, we included chronological time as a covariate in all multivariate modeling. To obtain the best subset of variables in the final model, we performed stepwise model selection. P values were derived from 2-sided tests, and the statistical analyses were carried out using S-PLUS 2000 software (MathSoft, Inc., Seattle, Wash).
Table 2 shows the clinical characteristics of the 80 patients who met our study's inclusion criteria. The median age was 53 years (range, 23-76 years), and 59 patients (74%) initially presented with stage IV disease. Of the 59 patients who presented with metastatic disease, 34 patients (58%) underwent surgical resection of their primary tumor, 11 patients (19%) underwent a surgical bypass, and 14 patients (23%) underwent a biopsy. The liver and the peritoneum were the most common initial sites of metastatic disease. Elevations in carcinoembryonic antigen (CEA) were not observed commonly, and only 31% of patients had a CEA level >10 mg/dL.
|Characteristics||No. (%) of Patients|
|Primary tumor site|
|Small bowel NOS||9 (11)|
|Well differentiated||5 (6)|
|Moderately differentiated||46 (58)|
|Poorly differentiated||17 (21)|
|Adenocarcinoma NOS||12 (15)|
|Mucinous histology||12 (15)|
|Initial sites of distant metastasis|
|Distant lymph nodes||7 (9)|
|Primary tumor resected||55 (69)|
|Primary resected (stage IV only)||34 (58)|
|Adjuvant chemotherapy||8 (10)|
|CEA >10 mg/dL||16/52 (31)|
The majority of patients had a primary tumor site within the jejunum (n = 35 patients; 44%) or the duodenum (n = 30 patients; 38%). Fifty-one patients (64%) had moderately or well differentiated SBA. Twelve patients (15%) had tumors with mucinous histologic features, and 8 of those patients had well to moderately differentiated tumors.
Of the 21 patients who initially presented with stage I, II, and III disease but subsequently developed metastatic disease, 8 patients received adjuvant chemotherapy, including 3 patients who received single-agent 5-FU therapy, 3 patients who received 5-FU and radiation, 1 patient who received 5-FU and mitomycin C, and 1 patient who received 5-FU and carboplatin. Of the 8 patients who received adjuvant chemotherapy, 7 patients (88%) had lymph node involvement, whereas only 3 of the 13 patients (23%) who did not receive adjuvant chemotherapy had lymph node involvement. The median time from completion of adjuvant therapy to development of metastatic disease was 7 months (range, 3-18 months). The chemotherapy administered to these 8 patients when they developed metastatic disease was irinotecan-based therapy in 2 patients, the combination of 5-FU and a platinum agent in 1 patient, and other 5-FU–based therapies in 5 patients.
First-line Chemotherapy Treatment and Response
Forty-one patients (51%) received 5-FU–based therapy without a platinum agent; 29 patients (36%) received 5-FU–based therapy with a platinum agent; 5 patients (6%) received non-5-FU–containing, platinum-based therapy; and 5 patients (6%) received a nonplatinum, non-5-FU–based therapy (Table 3). 5-FU was delivered intravenously in 53 patients (76%) and as the oral 5-FU prodrug, capecitabine, in 17 patients (24%). Capecitabine was administered either alone (1 patient), combined with cisplatin (9 patients), combined with oxaliplatin (6 patients), or combined with carboplatin and paclitaxel (1 patient). Of the 29 patients who received 5-FU and a platinum agent, 3 patients (10%) also received a third agent (paclitaxel in 1 patient, interferon-alpha in 1 patient, and bevacizumab in 1 patient). Irinotecan combinations were given to 2 patients, and a gemcitabine combination was given to 1 patient with no observed responses.
|Chemotherapy||No. (%) of Patients||No. [& Response Rate]|
|5-FU and no platinum||41 (51)||7 |
|5-FU and mitomycin-C||3||0|
|Other 5-FU–based combinations*||3||1|
|5-FU and platinum||29 (36)||12 |
|5-FU and cisplatin||19||7|
|5-FU and carboplatin||4||1|
|5-FU and oxaliplatin||6||4|
|Non-5-FU–based treatment||10 (13)||1 |
|Platinum and no 5-FU†||5||0|
|No platinum and no 5-FU‡||5||1|
The overall RR for the 80 patients was 25%, with 19 partial responses and 1 complete response. The single complete response occurred in a patient who had duodenal adenocarcinoma with biopsy-proven peritoneal metastases. The patient had radiographic resolution of her disease after 6 months of capecitabine and cisplatin and was alive without evidence of recurrence 50 months after she completed chemotherapy. None of the 8 patients who had received prior adjuvant chemotherapy responded to chemotherapy for their metastatic disease.
Table 4 compares patients who received a platinum agent and those who did not receive a platinum agent. Platinum-based chemotherapy was given more commonly to patients who had primary tumors in the duodenum, initially presented with stage IV disease, or did not undergo resection of their primary tumor. Of the 74 patients who had metastatic disease that was measurable radiographically at the start of chemotherapy, the 26 patients who received 5-FU and a platinum agent had a higher RR (46%) compared with the 48 patients who received other chemotherapy combinations (16%; P = .01). Treatment with the combination of 5-FU and a platinum agent was the only significant predictor of response in both univariate and multivariate analyses (odds ratio, 4.5; 95% confidence interval, 1.3-15.8; P = .02). Patients who had metastatic duodenal adenocarcinoma had a slightly higher RR (34%) compared with patients who had metastatic disease form other small bowel sites (RR, 22%), but the difference was not significant (P = .29). Among the 26 patients who received 5-FU and a platinum agent, the primary tumor site did not correlate with response (RR: 47% for duodenal primaries, 44% for other small bowel sites).
|Characteristic||No. (%) of Patients||P|
|Median age, y||52||54||.51|
|Men||16 (47)||27 (59)|
|Women||18 (53)||19 (41)|
|I-III||4 (12)||17 (37)|
|IV||30 (88)||29 (63)|
|Duodenum||20 (59)||10 (22)|
|Jejunum||11 (32)||24 (52)|
|Ileum||1 (3)||5 (11)|
|Small bowel NOS||2 (6)||7 (15)|
|Well to moderately differentiated||26 (76)||25 (54)|
|Poorly differentiated||5 (15)||12 (26)|
|Adenocarcinoma NOS||3 (9)||9 (20)|
|Initial sites of distant metastasis||.9|
|Liver||16 (47)||21 (46)|
|Peritoneum||18 (53)||19 (41)|
|Primary tumor resected||16 (47)||39 (85)||<.01|
|Adjuvant chemotherapy||2 (6)||6 (13)||.46|
Survival After First-line Chemotherapy Treatment
For the 80 patients who were included in this study, the median follow-up was 47 months, the median OS was 13 months (95% confidence interval, 10.4-16.5 months), and the median PFS was 4.6 months (95% confidence interval, 3.9-7 months). Patients who received 5-FU and a platinum agent had a higher median PFS (8.7 months) than patients who received other chemotherapy combinations (3.9 months; P = .003) (Fig. 1B). In the univariate analysis of PFS, treatment with 5-FU (P = .01) and treatment with a platinum agent (P = .02) were the only factors associated with improved PFS.
There was a trend toward a longer median OS in the patients who were treated with 5-FU and a platinum agent (14.8 months) compared with other patients (12 months), but the difference did not reach statistical significance (P = .1) (Table 5, Fig. 1B). In the univariate analysis of OS, both resection of the primary tumor (P = .01) and 5-FU treatment (P = .02) were associated with improved OS. In addition, patients whose grade of adenocarcinoma was not specified had a worse OS (P = .05); however, this subset of patients was small (n = 9). Resection of the primary tumor was also analyzed in the subgroup of patients who initially presented with stage IV disease. Undergoing primary tumor resection in this subgroup was associated with an improved median OS of 14.7 months compared with patients whose primary tumor remained intact (median OS, 10.4 months; P = .048).
|Characteristic||Median Survival, mo||P|
|Well to moderately differentiated||13.7|
|Primary tumor site||.58|
|Small bowel NOS||19.2|
|Primary tumor resected||.01|
Table 6 shows the results of the multivariate analyses of potential predictors of PFS and OS. More recent chronological treatment and the lack of prior adjuvant chemotherapy both were predictors of longer PFS and OS. Treatment with 5-FU and a platinum agent was a significant predictor of PFS (P < .01) but did not reach statistical significance for OS (P = .08). Resection of the primary tumor was a significant predictor of OS (P = .019) but not PFS.
|Primary tumor resected||0.71||0.43–1.2||.21|
|Treatment with 5-FU and platinum||0.49||0.29–0.84||<.001|
|Primary tumor resected||0.53||0.31–0.9||.019|
|Treatment with 5-FU and platinum||0.63||0.37–1.07||.08|
Second- and Third-line Chemotherapy
Fifty-seven chemotherapy treatments were administered as second- or third-line chemotherapy to 38 patients. Six patients (11%) responded to treatment, including 2 patients who responded to irinotecan-based therapy (n = 16 treated patients), 2 patients who responded to platinum-based therapy (n = 6 treated patients), 1 patient who responded to mitomycin C-based therapy (n = 9 treated patients), and 1 patient who responded to gemcitabine-based therapy (n = 2 treated patients). In general, all responses were of brief duration (<4 months).
In the current study, we observed that using 5-FU and a platinum agent to treat SBA resulted in a significantly improved RR (46% vs 16%), and PFS (8.7 months vs 3.9 months) compared with other chemotherapy combinations. Treatment with 5-FU and a platinum agent remained a significant predictor of response and PFS in our multivariate analyses. With regard to OS, the use of 5-FU and a platinum agent was not a significant predictor, although the value (P = .08) was suggestive of a possible relation. It is our opinion that the results of this study support the use of combination chemotherapy with 5-FU and platinum analogue for the treatment of metastatic SBA.
The factors associated with an improved OS in multivariate analysis were more recent chronological treatment, absence of prior adjuvant chemotherapy, and resection of the primary tumor. Of the 8 patients who received adjuvant 5-FU–based therapy, none responded to chemotherapy treatment for metastatic disease. Of those 8 patients, 5 patients were retreated with 5-FU-based nonplatinum and nonirinotecan therapy. It is also possible that the patients with localized disease who received adjuvant therapy had more aggressive tumor characteristics. Whether the lower OS observed in patients who previously received adjuvant 5-FU–based therapy was because of prior exposure to 5-FU or because of a selection bias is not known. Resection of the primary tumor was significant not only for the entire 80-patient cohort but also for the subset of patients who initially presented with stage IV disease. In this subgroup, patients who underwent resection of their primary tumor had an OS of 14.7 months compared with 10.4 months for patients who did not undergo surgical removal of the primary tumor. However, this result may be confounded by the presence of a greater burden of disease in those patients who did not undergo resection.
Previous studies conducted primarily in patients with localized SBA have suggested a worse outcome for patients with adenocarcinoma of the duodenum compared with to adenocarcinoma of the ileum or jejeunm.4, 6 In the current study, we observed no correlation between the primary tumor site and survival in patients with metastatic SBA. Although more patients with duodenal tumors were treated with a platinum agent, duodenal tumors did not respond better than nonduodenal small bowel tumors to platinum-based therapy.
The histologic grade of the tumor did not correlate with RR or PFS. Tumors that were classified as adenocarcinoma not otherwise specified demonstrated a worse OS than tumors with well, moderate, or poor differentiation. Only 9 patients had adenocarcinoma not otherwise specified, and those patients tended to be patients who had been treated in the earlier decades of this study. This result contrasts with other studies, in which poorly differentiated histology has been associated with a worse survival.7, 17, 18 The most common sites of metastases were the peritoneum and the liver, similar to prior reports.19, 20 Although peritoneal disease has been associated with worse outcomes for other gastrointestinal malignancies, we observed no difference in median OS between patients with and without peritoneal disease (15 months and 12.7 months, respectively).21
There is little information in the literature on the use of second- and third-line chemotherapy for SBA. Locher et al observed no radiographic responses to primarily irinotecan-based, second-line chemotherapy administered to 13 patients with SBA.14 Fishman et al observed responses to second-line chemotherapy in 4 of 18 patients, and 3 of their patients responded to gemcitabine-based therapy.7 In our study, a total of 57 chemotherapy treatments were given as either second-line or third-line therapy, and only 6 radiographic responses were observed (RR, 11%).
We acknowledge that our report has several limitations and that our findings are hypothesis-generating. Important limitations are 1) the retrospective nature of the report; 2) despite having a large denominator, interpretation of variables is limited by the many small subgroups of patients; and 3) heterogeneous therapy over a long time (partially compensated by incorporating chronological time as a covariate in our multivariate model). In addition, the young age of our patient population suggests a possible referral bias.
The infrequency of SBA has made it difficult to study. It is extremely unlikely that a randomized trial comparing 2 chemotherapy regimens will be conducted. However, a prospective phase 2 trial conducted within a cooperative group or outside a cooperative group (but in a multicenter setting) should be encouraged.
In the current study, we observed that treatment with 5-FU and a platinum agent was highly effective, producing an RR of 41% and a median PFS of 8.7 months. Patients who receive this combination may also gain an OS benefit. The combination of 5-FU and a platinum agent appears to be superior to other chemotherapy treatments, and this combination is recommended for the frontline treatment of patients with metastatic SBA.
- 5Chemotherapy for small bowel adenocarcinoma (SBA): experience over a 10-year period at the BC Cancer Agency [abstract]. Presented at the American Society of Clinical Oncology 2004 Gastrointestinal Cancers Symposium, San Francisco, California, January 22-24, 2004. Abstract 109..
- 15A phase II study of capecitabine and oxaliplatin (CAPOX) in metastatic adenocarcinoma of the small bowel or ampulla of Vater [abstract]. Presented at the American Society of Clinical Oncology 2007 Gastrointestinal Cancers Symposium, Orlando, Florida, January 19-21, 2007. Abstract 139., , , et al.
- 16GreeneFL,PageDL,FlemingID, et al, eds. AJCC Cancer Staging Manual,6th ed. New York, NY: Springer-Verlag; 2002.
- 22Final response data of a phase II study of capecitabine and oxaliplatin (CAPOX) in advanced adenocarcinoma of the small bowel or ampulla of Vater [abstract]. J Clin Oncol. 2008; 26S. Abstract 4538., , , et al.
- 23Phase I dose-finding and pharmacologic study of cisplatin, irinotecan, and either capecitabine or infusional 5-FU in patients with advanced gastrointestinal malignancies [abstract]. Presented at the American Society of Clinical Oncology Gastrointestinal Cancers Symposium, Hollywood, Florida, January 27-29, 2005. Abstract 28., , , et al.