- Email: [email protected]
Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial
ARTICLE IN PRESS
JID: YMSY
[m5G;October 2, 2018;16:8]
Surgery 0 0 0 (2018) 1–7
Contents lists available at ScienceDirect
Surgery journal homepage: www.elsevier.com/locate/surg
Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial Fadi S. Dahdaleh, MD a,∗, Scott K. Sherman, MD a, Elizabeth C. Poli, MD a, Janani Vigneswaran, MD a, Blase N. Polite, MD, MPP b, Manish R Sharma, MD b, Daniel V. Catenacci, MD b, Steven B. Maron, MD b, Kiran K. Turaga, MD, MPH a a b
The University of Chicago Medicine, Section of General Surgery/Surgical Oncology, Chicago, IL; and The University of Chicago Medicine, Section of Medical Oncology, Chicago, IL
a r t i c l e
i n f o
Article history: Accepted 22 June 2018 Available online xxx
a b s t r a c t Background: Patients with colon cancer often present with obstruction. Large series have reported obstruction among the high-risk features, yet prospective data on its specific prognostic influence are lacking. We hypothesized that obstruction is an independent risk factor for poor prognosis in patients with stage III colon cancer. Methods: N0147 was a trial conducted between 2004 and 2009 that randomly assigned patients with stage III colon cancer to adjuvant regimens of folinic acid (leucovorin calcium), fluorouracil, and oxaliplatin or fluorouracil, leucovorin, and irinotecan, with or without cetuximab. Patient-level data from the control chemotherapy-only arms were obtained. Patient, tumor, and treatment characteristics were abstracted. Disease-free survival and overall survival were estimated by the Kaplan-Meier method. Proportions were compared by χ 2 and Fisher exact tests. Univariable and multivariable survival analyses were performed using Cox proportional hazards models. Results: Of 1,543 patients with stage III colon cancer, 250 (16.2%) presented with obstruction. Patients with obstruction were equally likely to complete 12 cycles of adjuvant chemotherapy (75.9% vs 77.1%, P = .6). With median follow-up time of 30.9 months among survivors, five-year overall survival and disease-free survival were worse among patients with obstruction (overall survival 67.7% vs 78.0%, P < .001; disease-free survival 53.9% vs 67.0%, P < .0 0 01). On multivariable analysis, obstruction remained significantly associated with worse survival after adjusting for T stage, N stage, performance status, age, sex, histologic grade, and body mass index (overall survival hazard ratio 1.57, 95% confidence interval 1.12–2.20, P = .001; disease-free survival 1.52, 95% confidence interval 1.18–1.95, P < .001). Conclusion: In this prospectively followed cohort of patients with stage III colon cancer treated with adjuvant chemotherapy, obstruction was associated with recurrence and worse survival. Moreover, this effect was independent of T and N stage and histologic grade. These results suggest that obstruction should be incorporated into novel risk-stratification models. © 2018 Elsevier Inc. All rights reserved.
Introduction Colon cancer is the fourth most common cancer and second leading cause of cancer death in the United States.1,2 Acute obstructive presentation complicates 8.3% to 20% of colon cancer cases and is a common cause of large bowel obstruction.3–5 In addition to the well-established risks related to emergency surgery under such circumstances, obstructive presentation has further
∗ Reprint requests: Fadi S. Dahdaleh, MD, The University of Chicago Medicine, 5841 S. Maryland Ave, Room S214, Chicago IL 60637. Tel: 847-924-2580; fax: 773702-6120. E-mail address: [email protected] (F.S. Dahdaleh).
been recognized as a high-risk feature that is predictive of inferior oncologic outcomes in several studies.6,7 More recently, endoscopic stenting as a bridge to surgery has emerged as a favorable strategy and is possibly associated with improved short-term results; however, few studies were designed with the specific intent of examining long-term outcomes.8-10 In addition to informing prognosis, cancer staging carries significant implications for selection of optimal adjuvant therapy and follow-up.11 In its most recent edition, the American Joint Commission on Cancer (AJCC) staging manual continued to include nonanatomic variables (including molecular factors) to encourage more personalized risk stratification.12,13 Indeed, large data sets supply proof of this concept, indicating how incorporating
https://doi.org/10.1016/j.surg.2018.06.044 0039-6060/© 2018 Elsevier Inc. All rights reserved.
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY 2
ARTICLE IN PRESS
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
clinicopathologic elements better discriminates overall survival compared with traditional TNM staging alone.14 The present study uses the randomized, controlled North Central Cancer Treatment Group clinical trial (N0147) to test the hypothesis that obstructive presentation represents an additional prognostic feature in stage III colon cancer. It further seeks to evaluate the relationship between obstruction and other relevant clinicopathologic parameters in node-positive colon cancer. Methods Trial design and data acquisition This study retrospectively reviewed participants enrolled in N0147. Original trial details have been reported previously.15 Deidentified data were obtained from the Project Data Sphere repository.16 Because the accessed information is both publicly available and referenced in a manner that prohibits participant identification, this met the criteria for Institutional Review Board exemption (Code of Federal Regulations, Title 45, Part 46, Subpart A). N0147 was a phase III prospective randomized controlled trial designed to test the effect of adding cetuximab to oxaliplatin plus 5-fluorouracil and leucovorin (FOLFOX) after curative resection of stage III colon cancer.15 Initially, the trial called for random assignment to either FOLFOX; fluorouracil, leucovorin, and irinotecan (FOLFIRI), or a hybrid regimen (FOLFOX followed by FOLFIRI). Results from other trials prompted the inclusion of cetuximab in September 2004 and the discontinuation of the FOLFIRI arms in 2005.17-19 Between February 10, 2004 and November 25, 2009, patients were enrolled at multiple institutions across North America. Written informed consent was obtained in the original trial.
lymph node involvement (N1: 1–3, N2: ≥4), race (black, white, or other), Kirsten Rat Sarcoma virus (KRAS) biomarker status (mutant versus wild type), body mass index (BMI) (≥25 vs <25), and ECOG PS were recorded. Histopathologic review was performed centrally before random allocation to treatment arms. The primary endpoint of the original trial was disease-free survival (DFS), defined as time from random allocation to the first of either tumor recurrence or death from any cause. Secondary trial endpoints were time to recurrence (TTR) and overall survival (OS). TTR was measured from random allocation to tumor recurrence, whereas OS was from random allocation to death from any cause. OS was censored at 8 years, whereas DFS and TTR were censored at 5 years. Patients who died without recurrence were censored for TTR at the time of death. Patients who were lost to follow-up were censored at the date of their most recent disease assessment or contact. Statistics Continuous variables are reported as means, standard errors, medians, and interquartile ranges when possible. Categorical variables are reported as frequencies and percentages. Pearson χ 2 and Student’s t tests were used for descriptive analyses of proportions and continuous variables, respectively, and to investigate the relationship between relevant clinicopathologic variables and obstructive presentation. Survival analyses used Cox proportional hazards models and log-rank tests. The Kaplan-Meier method was used to plot survival. OS and DFS were determined at 5 years for the following variables: age, gender, BMI, ECOG performance status (PS), KRAS status, race, T stage, histologic grade, nodal status, chemotherapy received, and obstruction.
Inclusion criteria
Results
N0147 included patients ≥18 years old with Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 2 who underwent complete resection of stage III (T1-4N1-2M0) colon adenocarcinoma at least 12 cm from the anal verge. En bloc resection was required for locally advanced tumors and, although tumor at the radial margins (serosal, circumferential) was allowed, the protocol specified that specimens with division across the main tumor, suggesting incomplete resection, were not eligible. Patients with documented metastasis (stage IV) or a history of prior chemotherapy, radiotherapy, or immunotherapy were excluded. Tumor-related obstruction was allowed and, per trial protocol, was defined and categorized by participating experienced surgeons. Further details on that, such as whether it was clinically or endoscopically apparent, and whether any interventions such as stenting before surgery were performed, were not specified in the available trial documents. Post-treatment follow-up occurred at six-month intervals until 5 years from random allocation, with physical examination, computed tomographic scans, and laboratory assessment. At years 1 and 4 after resection, a follow-up colonoscopy was recommended. To conduct time-to-event analyses, follow-up for all patients took place 5 years from random allocation.15
Patients
Clinicopathologic characteristics and outcomes Patient-level records, including tumor, treatment, and survival data, were abstracted for participants in the control “chemotherapy-only” arms. Participants who received cetuximab were not included in this analysis. Patients in the trial arm receiving 6 cycles of FOLFOX followed by 6 cycles of FOLFIRI (n = 95) were included. Clinical T stage, histologic status (poorly differentiated or undifferentiated versus well or moderately differentiated),
Characteristics of the 1,543 included patients are summarized in Table 1. Of these patients, the majority were between 40 and 69 years of age (1,226, 79.5%) and 814 (52.7%) were women. The majority of patients had a BMI of ≥25 (1072, 69.5%), a PS of 0 to 1 (1538, 99.7%), and were white (1329, 86.1%). KRAS mutations were identified in 446 participants (30.2%), and obstruction was the presenting symptom in 250 patients (16.2%). From a histopathologic standpoint, T4 lesions were identified in 173 (11.2%), high-grade histologic status occurred in 391 (25.3%), and N2 disease existed in 637 (41.3%). The majority of patients received FOLFOX alone (86.6%). A small number of patients randomly assigned to FOLFIRI who had not completed it before discontinuation of this arm in the original trial were permitted to cross over to FOLFOX, but the number of these patients was not specified in the study data.15 Correlation of obstructive presentation with clinicopathologic factors Obstructive presentation was not significantly associated with different age groups, gender, race, or histologic grade (P > 0.2 for all; Table 1). Importantly, patients with obstruction were equally likely to complete 12 cycles of adjuvant therapy compared with participants without obstruction (75.9% vs 77.3%, P = .6). Patients with 1 to 3 and those with 4 or more involved lymph nodes had similar rates of obstruction (P = .053). A significant association was found between BMI and obstruction, with patients with lower BMIs (<25) presenting more commonly with obstruction (43.8% of obstructed patients had lower BMI vs. 30.6% of patients overall, P < .005). Deeper T4 lesions occurred more commonly in obstructed patients than in the overall study population (18.0% vs 11.2%, P < .001).
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
ARTICLE IN PRESS
JID: YMSY
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
3
Table 1 Baseline characteristics of the patients included in the control arm. Variable Age (y)† <40 40–69 ≥70 Sex Male Female BMI categorical <25 25–30 ≥30 BMI continuous ECOG performance status 0 1 2 KRAS status Mutant Wild type Race: Black White Otherwise not specified T stage T1 or T2 T3 T4 Histology Low (well or moderately differentiated) High (poorly differentiated or undifferentiated) Adjuvant chemotherapy 12 cycles <12 cycles No. of positive LNs 1–3 ≥4 Adjuvant regimen FOLFOX FOLFIRI
Total (%) N = 1,543
Obstruction (%) N = 250
100 (6.5) 1,226 (79.5) 217 (14.1)
21 (8.4) 189 (75.6) 40 (16)
729 (47.3) 814 (52.7)
115 (46) 135 (54)
471 (30.6) 541 (35.1) 529 (34.3) NA
109 (43.8) 84 (33.7) 56 (22.5) NA
1,185 (76.8) 353 (22.9) 5 (0.3)
179 (71.6) 68 (27.2) 3 (1.2)
446 (30.2) 1,032 (69.8)
84 (35.0) 156 (65)
105 (6.8) 1329 (86.1) 109 (7.1)
23 (9.2) 209 (83.6) 18 (7.2)
219 (14.2) 1,150 (74.6) 173 (11.2)
4 (1.6) 201 (80.4) 45 (18.0)
1,184 (74.7) 391 (25.3)
72 (28.8) 178 (71.2)
1,184 (77.3) 347 (22.7)
189 (75.9) 60 (24.1)
906 (58.7) 637 (41.3)
133 (53.2) 117 (46.8)
1337 (86.6) 206 (13.4)
217 (86.8) 33 (13.2)
P∗ .2
.7
<.001
<.001 .005
.08
.3
<.001
.2
.6
.053
1
LN, lymph node; NA, not applicable. Percentages reflect patients for whom each data point was available. ∗ P values show univariate association with obstruction by χ 2 test. † Project Datasphere repository only provided categorical data here; therefore a median could not be computed.
Survival With median follow-up time of 30.9 months among survivors, 5-year OS and DFS were 76.1% and 64.7%, respectively. On univariate analysis, patients with obstruction had significantly worse long-term outcomes compared with their nonobstructed counterparts (Table 2). Recurrence was more common in patients with obstruction, with only 53.9% remaining disease free at 5 years, compared with 67.0% of nonobstructed patients. Overall survival in obstructed patients was more than 10% lower at 5 years (67.7% vs 78.0%, P < .001; Table 2, Fig. 1). A similar trend was identified with higher T stage, poorly differentiated or undifferentiated histologic types, an increase in the number of positive lymph nodes, and in patients with BMI ≥30. Performance status was significantly associated with OS but not DFS. The type of chemotherapy received (FOLFOX alone versus FOLFIRI) was not associated with DFS or OS. After adjusting for age, sex, conventional AJCC staging variables (including T and N stage), high-grade histologic status, BMI, and ECOG performance status with Cox proportional hazards models, obstruction remained significantly associated with worse OS (hazard ratio [HR] 1.57 (1.12–2.20), P = .001) and DFS (HR 1.52, 95% confidence interval [CI] 1.18–1.95, P < .001; Table 3). Factors other than obstruction independently associated with DFS on multivariable analysis included BMI, lymph node status, and T stage.
Reduced OS independently correlated with age (P = .04), male sex (HR: 1.61, P < .001), N2 status (HR: 2.28, P < .001), higher T stage (P = .005), and poorly differentiated or undifferentiated histologic type (HR: 1.55, P = .006). Repeating this analysis while including trial patients who received cetuximab yielded comparable results (N = 2,968; Supplemental Table 1). Discussion In this study, data from N0147 were used to evaluate the longterm prognostic impact of obstructive presentation in stage III colon cancer and revealed an independent detrimental effect of obstruction on long-term DFS and OS. This corroborates several previous reports that recognized obstruction as a high-risk feature in colon cancer.20,21 However, this study is unique because it considers a cohort of patients subjected to prospective random allocation and systematic follow-up for a relatively long period. Importantly, the worsened OS and DFS attributed to obstruction here were independent of known adverse prognostic features such as T stage, histologic grade, number of positive lymph nodes, and receipt of adjuvant therapy. Collectively, these findings suggest that obstruction in stage III colon cancer could represent a biologically distinct state that is not captured by conventional staging systems. Indeed, although stage III colon cancer, by definition, is a disease not detectable
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
ARTICLE IN PRESS
JID: YMSY 4
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7 Table 2 Univariate analysis of overall survival (OS) and disease-free survival (DFS) at 5 years. Variable Age <40 40–69 ≥70 Sex Male Female BMI categorical <25 25–30 ≥30 BMI continuous ECOG performance status 0 1–2 KRAS status Mutant Wild type Race Black White Other T stage T1 or T2 T3 T4 Histology Low (well or moderately differentiated) High (poorly differentiated or undifferentiated) No. of positive LNs 1–3 ≥4 Chemotherapy FOLFOX FOLFIRI Obstruction Yes No
OS at 5 y (%)
P
DFS at 5 y (%)
.004 74.2 77.2 73.1
.04 61.8 66.6 59.6
.001 70.0 83.0
.2 62.2 67.4
.04 79.4 80.1 69.0 NA
P
.048 .02
78.3 69.3
.009 68.7 68.8 56.5 NA
0.03 0.1
67.0 57.8 .2
73.4 77.6
.06 59.4 67.4
.08 69.7 76.0 84.2
.02 45.0 64.9 84.9 <.001
.001 85.63 76.49 60.9
85.4 62.9 52.0 <.001
72.7 77.3
.001 60.0 66.8
<.001 82.7 65.9
<.001 72.8 52.5
.9 73.4 76.6
.9 63.2 64.8
<.001 67.7 78.0
<.001 53.9 67.0
LN, lymph node; NA, not applicable.
Fig 1. Patients presenting with obstruction (dashed blue lines) had significantly worse disease-free (left panel) and overall survival (right panel) than patients who were not obstructed (solid red lines, P < .001 for both).
beyond regional lymph nodes, a significant proportion of patients experience disease recurrence both locally and distantly after resection, suggesting that microscopic metastasis often exists at the time of presentation.22,23 Conceptually, this has been validated by observing circulating tumor cells in the majority of patients with stage III colon cancer.24 This phenomenon may be accen-
tuated with obstruction because luminal distension and inflammation can lead to increased mucosal permeability and subsequent release of mediators known to promote distant metastasis.25,26 Clinical evidence to support this hypothesis can be gleaned from a retrospective analysis of patients with stage II and III colon cancer where preoperative obstruction correlated with distant
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
ARTICLE IN PRESS
JID: YMSY
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
5
Table 3 Multivariable analysis of overall survival (OS) and disease-free survival (DFS) expressed in hazard ratios. Variable
HR (95% CI)
P
HR (95% CI) DFS
P
.04
1.00 0.62 (0.43–0.91) 0.64 (0.41–1.00) 1.15 (0.93–1.43) 1.00 2.81 (1.71–4.62) 4.20 (2.43–7.29) 2.07 (1.67–2.56) 1.25 (0.99–1.57) 1.09 (0.86–1.39) 1.00 1.09 (0.83–1.44) 1.67 (1.29–2.17) 1.52 (1.18–1.95)
.1
OS Age (<40 y reference)∗ 40–69 ≥70 Sex (male versus female) T stage (T1–2 reference)∗ T3 T4 No. of positive LNs (≥4 vs 1–3) Histology (high versus low) ECOG PS (1–2 vs 0) BMI (<25 reference) 25–30 >30 Obstruction (yes versus no)
1.00 0.52 (0.32–0.86) 0.72 (0.41–1.27) 1.61 (1.19–2.18) 1.00 1.57 (0.89–2.74) 2.11 (1.10–4.06) 2.28 (1.69–3.06) 1.55 (1.14–2.09) 1.32 (0.96–1.81) 1.00 1.17 (0.80–1.71) 1.75 (1.22–2.52) 1.57 (1.12–2.20)
<.001 .005
<.001 .006 .1 .006
.001
.1 <.001
<.001 .07 .5 <.001
<.001
LN, lymph node. ∗ For age and T stage, P values represent significance of each variable category in the model.
(liver) (P = .019) but not with peritoneal or local recurrences.27 Further insight on the natural history of obstructed colon cancer can be learned by analyzing the Dutch Stent-In 2 trial, which compared emergency surgery with stenting for the treatment of colon cancer presenting acutely.28 Interestingly, in patients undergoing immediate resection, 10% of specimens harbored microperforations, suggesting that occult direct tumor spillage occurs more often than commonly appreciated in obstructive tumors. There, locoregional and distant recurrences developed in 6.3% and 21.9%, respectively, in the resected group as opposed to 19.2% and 30.8% in the stented group.29 The authors hypothesized that malignant cell shedding occurs with obstruction-related perforations that predispose to locoregional (peritoneal) recurrences, representing an additional mechanism of future failure in obstructive tumors. In a recent meta-analysis of long-term oncologic outcomes after stent versus emergency surgery for malignant left-sided colon cancer obstructions, Ceresoli et al10 found no significant differences in recurrence rates among either approach and further highlighted the paucity of studies that reported outcomes beyond 3 years. Specifically, only 3 randomized controlled trials met the inclusion criteria and reported median follow-up times of more than 30 months. Those trials included a combined 134 patients in total. This lack of prospective survival data is also noted when one examines exclusion criteria of relevant and recent clinical trials on resectable colon cancer where obstructive presentation commonly precluded enrollment (Supplemental Table 2). Completing recommended courses of adjuvant chemotherapy within appropriate time frames has been found to lead to improved survival in stage III colon cancer.30 It has been postulated that worse outcomes with obstruction might be due to patients suffering postoperative complications and therefore being less able to complete chemotherapy. In this study we found that participants presenting with obstruction who enrolled in the clinical trial after their resection were no less likely to complete recommended adjuvant doses of chemotherapy. High rates of chemotherapy completion in this cohort could be explained by the selection bias inherent to N0197, which required an adequate PS for random allocation after resection. However, even if this population does not represent all patients with obstruction, many of whom will not be able to receive optimal chemotherapy, the observation that patients with obstruction in N0197 had poorer survival despite receiving the same intensity of chemotherapy further emphasizes the importance of obstruction as a prognostic marker.
Several clinicopathologic factors independently influencing outcomes have been identified in stage III colon cancer; however, the AJCC has traditionally focused on anatomic extent of disease for staging because it allows for straightforward stratification and communication.31,32 In its most recent edition, the AJCC colorectal committee endorsed the use of the ACCENT-based Web calculator to help predict recurrence and OS in stage III colon cancer.14 This tool incorporates essential additional variables such as age, sex, race, BMI, PS, grade, treatment group, and tumor location; however, it does not account for obstructive presentation. The present study confirms the importance of traditional staging variables, with T and N stage correlating strongly with survival in the multivariable model. Although more advanced T and N stage were more common in patients with obstruction, even after accounting for these factors, obstruction remained independently associated with survival, suggesting that obstruction confers additional risk beyond that described by T and N stage alone in stage III colorectal cancer. Our results suggest that inclusion of obstruction could improve this tool. In this study, patients with lower BMIs were more likely to have presented with obstruction. This trend was reported in a review of 2 National Surgical Adjuvant Breast and Bowel Project trials in which patients with lower BMIs (<18.5) more commonly presented with obstruction than obese (≥35.0) patients (17.7% vs 8.0%).33 Although the recognized increased colon cancer mortality risk in obese participants may be explained through insulin-like growth factor axis disturbances from hyperinsulinemia,34 the relationship of BMI to colon cancer obstruction is less clear. One possible explanation is that patients with higher BMIs were more likely to have suffered postoperative complications after emergent operations performed for obstruction, prohibiting them from meeting trial inclusion criteria. Another possibility is that near-obstructive tumors might impair alimentation, causing weight loss and low BMI. Several limitations of this study must be noted because important granular data on preoperative carcinoembryonic antigen levels, the presence of lymphovascular invasion, KRAS and BRAF mutational status, other genomic profiling, and tumor location were absent in the available dataset. In one study on 1,877 patients with node-negative colon and rectal cancers, emergency presentation (251, 13.4%) was associated with older age (P = .023), T4 stage (P < .001), extramural vascular invasion (P = .001), and care under “nonspecialist surgeons” (P < .001), all of which could lead to worse outcomes.35 Those factors could only be partially accounted
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY 6
ARTICLE IN PRESS
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
for in this report and may have affected results. Although emergency resections were allowed, those patients who fared poorly and were not fit for chemotherapy after surgery would not have been included. Another limitation is that the trial protocol documents do not rigorously define obstruction. Obstruction was documented when the enrolling physician indicated on enrollment forms that it was present. The lack of further data regarding location, degree, preoperative interventions, and other details about the obstruction may complicate generalizability of these results to individual patients. Despite this limitation, knowing that obstruction, even if broadly defined, correlates with worse outcomes holds value, and further studies could investigate whether the impact of obstruction depends on the details of the obstruction, such as whether clinical or endoscopic, whether stents were placed, and other factors. Perhaps of most significance is the lack of tumor location information, particularly right- versus left-sidedness because it is associated both with long-term outcomes and likelihood of obstruction. Right-sided obstructions are believed to occur with more bulky tumors capable of overcoming a larger circumference, which are linked to worse oncologic outcomes.36 A more recent study, however, contested this finding and asserted no difference in rates of recurrence or OS among proximal and distal tumors presenting emergently.37 The precise effect of an obstruction’s location, therefore, remains uncertain, and this study is unfortunately unable to provide additional information on this. Finally, that random allocation occurred after resection may have precluded those patients adversely affected by comorbidities and postoperative complications from participating. Failing to include the sickest patients, however, would likely act to reduce the impact of obstruction on survival, suggesting that the influence of obstruction in the general population may be even greater. In conclusion, the long-term prognosis in patients with stage III colon cancer was adversely affected by preoperative obstruction independent of other known negative prognostic factors. We suggest that obstruction represents a clinical marker that can be used to identify patients at higher risk of recurrence and death. Such patients may be more likely to benefit from novel or more intensive therapies, and we advocate for their inclusion in future trials. Acknowledgments This publication is based on research using information obtained from www.projectdatasphere.org, which is maintained by Project Data Sphere, LLC. Neither Project Data Sphere, LLC nor the owner(s) of any information from the website have contributed to, approved or are in any way responsible for the contents of this publication. Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.surg.2018.06.044. References 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin. 2016;66:7–30. 2. DeSantis CE, Siegel RL, Sauer AG, Miller KD, Fedewa SA, Alcaraz KI, et al. Cancer statistics for African Americans, 2016: progress and opportunities in reducing racial disparities. CA Cancer J Clin. 2016;66:290–308. 3. Mulcahy HE, Skelly MM, Husain A, O’Donoghue DP. Long-term outcome following curative surgery for malignant large bowel obstruction. Br J Surg. 1996;83:46–50. 4. Cheynel N, Cortet M, Lepage C, Benoit L, Faivre J, Bouvier AM. Trends in frequency and management of obstructing colorectal cancers in a well-defined population. Dis Colon Rectum. 2007;50:1568–1575. 5. Cortet M, Grimault A, Cheynel N, Lepage C, Bouvier AM, Faivre J. Patterns of recurrence of obstructing colon cancers after surgery for cure: a population-based study. Colorectal Dis. 2013;15:1100–1106.
6. Carraro PG, Segala M, Cesana BM, Tiberio G. Obstructing colonic cancer: failure and survival patterns over a ten-year follow-up after one-stage curative surgery. Dis Colon Rectum. 2001;44:243–250. 7. Mella J, Biffin A, Radcliffe AG, Stamatakis JD, Steele RJ. Population-based audit of colorectal cancer management in two UK health regions. Colorectal Cancer Working Group, Royal College of Surgeons of England Clinical Epidemiology and Audit Unit. Br J Surg. 1997;84:1731–1736. 8. Matsuda A, Miyashita M, Matsumoto S, Matsutani T, Sakurazawa N, Takahashi G, et al. Comparison of long-term outcomes of colonic stent as “bridge to surgery” and emergency surgery for malignant large-bowel obstruction: a meta-analysis. Ann Surg Oncol. 2015;22:497–504. 9. Sabbagh C, Browet F, Diouf M, Cosse C, Brehant O, Bartoli E, et al. Is stenting as “a bridge to surgery” an oncologically safe strategy for the management of acute, left-sided, malignant, colonic obstruction? A comparative study with a propensity score analysis. Ann Surg. 2013;258:107–115. 10. Ceresoli M, Allievi N, Coccolini F, Montori G, Fugazzola P, Pisano M, et al. Long-term oncologic outcomes of stent as a bridge to surgery versus emergency surgery in malignant left side colonic obstructions: a meta-analysis. J Gastrointest Oncol. 2017;8:867–876. 11. Dienstmann R, Salazar R, Tabernero J. Personalizing colon cancer adjuvant therapy: selecting optimal treatments for individual patients. J Clin Oncol. 2015;33:1787–1796. 12. Amin MB, Greene FL, Edge SB, Compton CC, Gershenwald JE, Brookland RK, et al. The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more "personalized" approach to cancer staging. CA Cancer J Clin. 2017;67:93–99. 13. Kattan MW, Hess KR, Amin MB, Lu Y, Moons KG, Gershenwald JE, et al. American Joint Committee on Cancer acceptance criteria for inclusion of risk models for individualized prognosis in the practice of precision medicine. CA Cancer J Clin. 2016;66:370–374. 14. Renfro LA, Grothey A, Xue Y, Saltz LB, Andre T, Twelves C, et al. ACCENT-based web calculators to predict recurrence and overall survival in stage III colon cancer. J Natl Cancer Inst. 2014;106. 15. Alberts SR, Sargent DJ, Nair S, Mahoney MR, Mooney M, Thibodeau SN, et al. Effect of oxaliplatin, fluorouracil, and leucovorin with or without cetuximab on survival among patients with resected stage III colon cancer: a randomized trial. JAMA. 2012;307:1383–1393. 16. Project Data Sphere [online database] [cited 2018 Jan]. Available from https: //www.projectdatasphere.org. 17. Saltz LB, Niedzwiecki D, Hollis D, Goldberg RM, Hantel A, Thomas JP, et al. Irinotecan fluorouracil plus leucovorin is not superior to fluorouracil plus leucovorin alone as adjuvant treatment for stage III colon cancer: results of CALGB 89803. J Clin Oncol. 2007;25:3456–3461. 18. Van Cutsem E, Labianca R, Bodoky G, Barone C, Aranda E, Nordlinger B, et al. Randomized phase III trial comparing biweekly infusional fluorouracil/ leucovorin alone or with irinotecan in the adjuvant treatment of stage III colon cancer: PETACC-3. J Clin Oncol. 2009;27:3117–3125. 19. Ychou M, Raoul JL, Douillard JY, Gourgou-Bourgade S, Bugat R, Mineur L, et al. A phase III randomised trial of LV5FU2 + irinotecan versus LV5FU2 alone in adjuvant high-risk colon cancer (FNCLCC Accord02/FFCD9802). Ann Oncol. 2009;20:674–680. 20. Burdy G, Panis Y, Alves A, Nemeth J, Lavergne-Slove A, Valleur P. Identifying patients with T3-T4 node-negative colon cancer at high risk of recurrence. Dis Colon Rectum. 2001;44:1682–1688. 21. Andre T, Boni C, Navarro M, Tabernero J, Hickish T, Topham C, et al. Improved overall survival with oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment in stage II or III colon cancer in the MOSAIC trial. J Clin Oncol. 2009;27:3109–3116. 22. Pantel K, Brakenhoff RH. Dissecting the metastatic cascade. Nat Rev Cancer. 2004;4:448–456. 23. Bernards R, Weinberg RA. A progression puzzle. Nature. 2002;418:823. 24. Sotelo MJ, Sastre J, Maestro ML, Veganzones S, Vieitez JM, Alonso V, et al. Role of circulating tumor cells as prognostic marker in resected stage III colorectal cancer. Ann Oncol. 2015;26:535–541. 25. Granger DN, Kvietys PR, Mortillaro NA, Taylor AE. Effect of luminal distension on intestinal transcapillary fluid exchange. Am J Physiol. 1980;239:G516–G523. 26. Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature. 2008;454:436–444. 27. Katoh H, Yamashita K, Wang G, Sato T, Nakamura T, Watanabe M. Prognostic significance of preoperative bowel obstruction in stage III colorectal cancer. Ann Surg Oncol. 2011;18:2432–2441. 28. van Hooft JE, Bemelman WA, Oldenburg B, Marinelli AW, Lutke Holzik MF, Grubben MJ, et al. Colonic stenting versus emergency surgery for acute left-sided malignant colonic obstruction: a multicentre randomised trial. Lancet Oncol. 2011;12:344–352. 29. Sloothaak DA, van den Berg MW, Dijkgraaf MG, Fockens P, Tanis PJ, van Hooft JE, et al. Oncological outcome of malignant colonic obstruction in the Dutch Stent-In 2 trial. Br J Surg. 2014;101:1751–1757. 30. Cheung HY, Chung CC, Tsang WW, Wong JC, Yau KK, Li MK. Endolaparoscopic approach vs conventional open surgery in the treatment of obstructing left-sided colon cancer: a randomized controlled trial. Arch Surg. 2009;144:1127–1132. 31. Vaccaro CA, Bonadeo FA, Benati ML, Quintana GM, Rubinstein F, Mullen E, et al. Colorectal cancer staging: reappraisal of N/PN classification. Dis Colon Rectum. 2004;47:66–69.
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY
ARTICLE IN PRESS
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7 32. Johnson PM, Porter GA, Ricciardi R, Baxter NN. Increasing negative lymph node count is independently associated with improved long-term survival in stage IIIB and IIIC colon cancer. J Clin Oncol. 2006;24:3570–3575. 33. Dignam JJ, Polite BN, Yothers G, Raich P, Colangelo L, O’Connell MJ, et al. Body mass index and outcomes in patients who receive adjuvant chemotherapy for colon cancer. J Natl Cancer Inst. 2006;98:1647–1654. 34. Renehan AG, Frystyk J, Flyvbjerg A. Obesity and cancer risk: the role of the insulin-IGF axis. Trends Endocrinol Metab. 2006;17:328–336. 35. Oliphant R, Mansouri D, Nicholson GA, McMillan DC, Horgan PG, Morrison DS, et al. Emergency presentation of node-negative colorectal cancer treated with curative surgery is associated with poorer short and longer-term survival. Int J Colorectal Dis. 2014;29:591–598.
7
36. Wolmark N, Wieand HS, Rockette HE, Fisher B, Glass A, Lawrence W, et al. The prognostic significance of tumor location and bowel obstruction in Dukes B and C colorectal cancer. Findings from the NSABP clinical trials. Ann Surg. 1983;198:743–752. 37. Frago R, Biondo S, Millan M, Kreisler E, Golda T, Fraccalvieri D, et al. Differences between proximal and distal obstructing colonic cancer after curative surgery. Colorectal Dis. 2011;13:e116–e122.
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY
[m5G;October 2, 2018;16:8]
Surgery 0 0 0 (2018) 1–7
Contents lists available at ScienceDirect
Surgery journal homepage: www.elsevier.com/locate/surg
Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial Fadi S. Dahdaleh, MD a,∗, Scott K. Sherman, MD a, Elizabeth C. Poli, MD a, Janani Vigneswaran, MD a, Blase N. Polite, MD, MPP b, Manish R Sharma, MD b, Daniel V. Catenacci, MD b, Steven B. Maron, MD b, Kiran K. Turaga, MD, MPH a a b
The University of Chicago Medicine, Section of General Surgery/Surgical Oncology, Chicago, IL; and The University of Chicago Medicine, Section of Medical Oncology, Chicago, IL
a r t i c l e
i n f o
Article history: Accepted 22 June 2018 Available online xxx
a b s t r a c t Background: Patients with colon cancer often present with obstruction. Large series have reported obstruction among the high-risk features, yet prospective data on its specific prognostic influence are lacking. We hypothesized that obstruction is an independent risk factor for poor prognosis in patients with stage III colon cancer. Methods: N0147 was a trial conducted between 2004 and 2009 that randomly assigned patients with stage III colon cancer to adjuvant regimens of folinic acid (leucovorin calcium), fluorouracil, and oxaliplatin or fluorouracil, leucovorin, and irinotecan, with or without cetuximab. Patient-level data from the control chemotherapy-only arms were obtained. Patient, tumor, and treatment characteristics were abstracted. Disease-free survival and overall survival were estimated by the Kaplan-Meier method. Proportions were compared by χ 2 and Fisher exact tests. Univariable and multivariable survival analyses were performed using Cox proportional hazards models. Results: Of 1,543 patients with stage III colon cancer, 250 (16.2%) presented with obstruction. Patients with obstruction were equally likely to complete 12 cycles of adjuvant chemotherapy (75.9% vs 77.1%, P = .6). With median follow-up time of 30.9 months among survivors, five-year overall survival and disease-free survival were worse among patients with obstruction (overall survival 67.7% vs 78.0%, P < .001; disease-free survival 53.9% vs 67.0%, P < .0 0 01). On multivariable analysis, obstruction remained significantly associated with worse survival after adjusting for T stage, N stage, performance status, age, sex, histologic grade, and body mass index (overall survival hazard ratio 1.57, 95% confidence interval 1.12–2.20, P = .001; disease-free survival 1.52, 95% confidence interval 1.18–1.95, P < .001). Conclusion: In this prospectively followed cohort of patients with stage III colon cancer treated with adjuvant chemotherapy, obstruction was associated with recurrence and worse survival. Moreover, this effect was independent of T and N stage and histologic grade. These results suggest that obstruction should be incorporated into novel risk-stratification models. © 2018 Elsevier Inc. All rights reserved.
Introduction Colon cancer is the fourth most common cancer and second leading cause of cancer death in the United States.1,2 Acute obstructive presentation complicates 8.3% to 20% of colon cancer cases and is a common cause of large bowel obstruction.3–5 In addition to the well-established risks related to emergency surgery under such circumstances, obstructive presentation has further
∗ Reprint requests: Fadi S. Dahdaleh, MD, The University of Chicago Medicine, 5841 S. Maryland Ave, Room S214, Chicago IL 60637. Tel: 847-924-2580; fax: 773702-6120. E-mail address: [email protected] (F.S. Dahdaleh).
been recognized as a high-risk feature that is predictive of inferior oncologic outcomes in several studies.6,7 More recently, endoscopic stenting as a bridge to surgery has emerged as a favorable strategy and is possibly associated with improved short-term results; however, few studies were designed with the specific intent of examining long-term outcomes.8-10 In addition to informing prognosis, cancer staging carries significant implications for selection of optimal adjuvant therapy and follow-up.11 In its most recent edition, the American Joint Commission on Cancer (AJCC) staging manual continued to include nonanatomic variables (including molecular factors) to encourage more personalized risk stratification.12,13 Indeed, large data sets supply proof of this concept, indicating how incorporating
https://doi.org/10.1016/j.surg.2018.06.044 0039-6060/© 2018 Elsevier Inc. All rights reserved.
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY 2
ARTICLE IN PRESS
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
clinicopathologic elements better discriminates overall survival compared with traditional TNM staging alone.14 The present study uses the randomized, controlled North Central Cancer Treatment Group clinical trial (N0147) to test the hypothesis that obstructive presentation represents an additional prognostic feature in stage III colon cancer. It further seeks to evaluate the relationship between obstruction and other relevant clinicopathologic parameters in node-positive colon cancer. Methods Trial design and data acquisition This study retrospectively reviewed participants enrolled in N0147. Original trial details have been reported previously.15 Deidentified data were obtained from the Project Data Sphere repository.16 Because the accessed information is both publicly available and referenced in a manner that prohibits participant identification, this met the criteria for Institutional Review Board exemption (Code of Federal Regulations, Title 45, Part 46, Subpart A). N0147 was a phase III prospective randomized controlled trial designed to test the effect of adding cetuximab to oxaliplatin plus 5-fluorouracil and leucovorin (FOLFOX) after curative resection of stage III colon cancer.15 Initially, the trial called for random assignment to either FOLFOX; fluorouracil, leucovorin, and irinotecan (FOLFIRI), or a hybrid regimen (FOLFOX followed by FOLFIRI). Results from other trials prompted the inclusion of cetuximab in September 2004 and the discontinuation of the FOLFIRI arms in 2005.17-19 Between February 10, 2004 and November 25, 2009, patients were enrolled at multiple institutions across North America. Written informed consent was obtained in the original trial.
lymph node involvement (N1: 1–3, N2: ≥4), race (black, white, or other), Kirsten Rat Sarcoma virus (KRAS) biomarker status (mutant versus wild type), body mass index (BMI) (≥25 vs <25), and ECOG PS were recorded. Histopathologic review was performed centrally before random allocation to treatment arms. The primary endpoint of the original trial was disease-free survival (DFS), defined as time from random allocation to the first of either tumor recurrence or death from any cause. Secondary trial endpoints were time to recurrence (TTR) and overall survival (OS). TTR was measured from random allocation to tumor recurrence, whereas OS was from random allocation to death from any cause. OS was censored at 8 years, whereas DFS and TTR were censored at 5 years. Patients who died without recurrence were censored for TTR at the time of death. Patients who were lost to follow-up were censored at the date of their most recent disease assessment or contact. Statistics Continuous variables are reported as means, standard errors, medians, and interquartile ranges when possible. Categorical variables are reported as frequencies and percentages. Pearson χ 2 and Student’s t tests were used for descriptive analyses of proportions and continuous variables, respectively, and to investigate the relationship between relevant clinicopathologic variables and obstructive presentation. Survival analyses used Cox proportional hazards models and log-rank tests. The Kaplan-Meier method was used to plot survival. OS and DFS were determined at 5 years for the following variables: age, gender, BMI, ECOG performance status (PS), KRAS status, race, T stage, histologic grade, nodal status, chemotherapy received, and obstruction.
Inclusion criteria
Results
N0147 included patients ≥18 years old with Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 2 who underwent complete resection of stage III (T1-4N1-2M0) colon adenocarcinoma at least 12 cm from the anal verge. En bloc resection was required for locally advanced tumors and, although tumor at the radial margins (serosal, circumferential) was allowed, the protocol specified that specimens with division across the main tumor, suggesting incomplete resection, were not eligible. Patients with documented metastasis (stage IV) or a history of prior chemotherapy, radiotherapy, or immunotherapy were excluded. Tumor-related obstruction was allowed and, per trial protocol, was defined and categorized by participating experienced surgeons. Further details on that, such as whether it was clinically or endoscopically apparent, and whether any interventions such as stenting before surgery were performed, were not specified in the available trial documents. Post-treatment follow-up occurred at six-month intervals until 5 years from random allocation, with physical examination, computed tomographic scans, and laboratory assessment. At years 1 and 4 after resection, a follow-up colonoscopy was recommended. To conduct time-to-event analyses, follow-up for all patients took place 5 years from random allocation.15
Patients
Clinicopathologic characteristics and outcomes Patient-level records, including tumor, treatment, and survival data, were abstracted for participants in the control “chemotherapy-only” arms. Participants who received cetuximab were not included in this analysis. Patients in the trial arm receiving 6 cycles of FOLFOX followed by 6 cycles of FOLFIRI (n = 95) were included. Clinical T stage, histologic status (poorly differentiated or undifferentiated versus well or moderately differentiated),
Characteristics of the 1,543 included patients are summarized in Table 1. Of these patients, the majority were between 40 and 69 years of age (1,226, 79.5%) and 814 (52.7%) were women. The majority of patients had a BMI of ≥25 (1072, 69.5%), a PS of 0 to 1 (1538, 99.7%), and were white (1329, 86.1%). KRAS mutations were identified in 446 participants (30.2%), and obstruction was the presenting symptom in 250 patients (16.2%). From a histopathologic standpoint, T4 lesions were identified in 173 (11.2%), high-grade histologic status occurred in 391 (25.3%), and N2 disease existed in 637 (41.3%). The majority of patients received FOLFOX alone (86.6%). A small number of patients randomly assigned to FOLFIRI who had not completed it before discontinuation of this arm in the original trial were permitted to cross over to FOLFOX, but the number of these patients was not specified in the study data.15 Correlation of obstructive presentation with clinicopathologic factors Obstructive presentation was not significantly associated with different age groups, gender, race, or histologic grade (P > 0.2 for all; Table 1). Importantly, patients with obstruction were equally likely to complete 12 cycles of adjuvant therapy compared with participants without obstruction (75.9% vs 77.3%, P = .6). Patients with 1 to 3 and those with 4 or more involved lymph nodes had similar rates of obstruction (P = .053). A significant association was found between BMI and obstruction, with patients with lower BMIs (<25) presenting more commonly with obstruction (43.8% of obstructed patients had lower BMI vs. 30.6% of patients overall, P < .005). Deeper T4 lesions occurred more commonly in obstructed patients than in the overall study population (18.0% vs 11.2%, P < .001).
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
ARTICLE IN PRESS
JID: YMSY
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
3
Table 1 Baseline characteristics of the patients included in the control arm. Variable Age (y)† <40 40–69 ≥70 Sex Male Female BMI categorical <25 25–30 ≥30 BMI continuous ECOG performance status 0 1 2 KRAS status Mutant Wild type Race: Black White Otherwise not specified T stage T1 or T2 T3 T4 Histology Low (well or moderately differentiated) High (poorly differentiated or undifferentiated) Adjuvant chemotherapy 12 cycles <12 cycles No. of positive LNs 1–3 ≥4 Adjuvant regimen FOLFOX FOLFIRI
Total (%) N = 1,543
Obstruction (%) N = 250
100 (6.5) 1,226 (79.5) 217 (14.1)
21 (8.4) 189 (75.6) 40 (16)
729 (47.3) 814 (52.7)
115 (46) 135 (54)
471 (30.6) 541 (35.1) 529 (34.3) NA
109 (43.8) 84 (33.7) 56 (22.5) NA
1,185 (76.8) 353 (22.9) 5 (0.3)
179 (71.6) 68 (27.2) 3 (1.2)
446 (30.2) 1,032 (69.8)
84 (35.0) 156 (65)
105 (6.8) 1329 (86.1) 109 (7.1)
23 (9.2) 209 (83.6) 18 (7.2)
219 (14.2) 1,150 (74.6) 173 (11.2)
4 (1.6) 201 (80.4) 45 (18.0)
1,184 (74.7) 391 (25.3)
72 (28.8) 178 (71.2)
1,184 (77.3) 347 (22.7)
189 (75.9) 60 (24.1)
906 (58.7) 637 (41.3)
133 (53.2) 117 (46.8)
1337 (86.6) 206 (13.4)
217 (86.8) 33 (13.2)
P∗ .2
.7
<.001
<.001 .005
.08
.3
<.001
.2
.6
.053
1
LN, lymph node; NA, not applicable. Percentages reflect patients for whom each data point was available. ∗ P values show univariate association with obstruction by χ 2 test. † Project Datasphere repository only provided categorical data here; therefore a median could not be computed.
Survival With median follow-up time of 30.9 months among survivors, 5-year OS and DFS were 76.1% and 64.7%, respectively. On univariate analysis, patients with obstruction had significantly worse long-term outcomes compared with their nonobstructed counterparts (Table 2). Recurrence was more common in patients with obstruction, with only 53.9% remaining disease free at 5 years, compared with 67.0% of nonobstructed patients. Overall survival in obstructed patients was more than 10% lower at 5 years (67.7% vs 78.0%, P < .001; Table 2, Fig. 1). A similar trend was identified with higher T stage, poorly differentiated or undifferentiated histologic types, an increase in the number of positive lymph nodes, and in patients with BMI ≥30. Performance status was significantly associated with OS but not DFS. The type of chemotherapy received (FOLFOX alone versus FOLFIRI) was not associated with DFS or OS. After adjusting for age, sex, conventional AJCC staging variables (including T and N stage), high-grade histologic status, BMI, and ECOG performance status with Cox proportional hazards models, obstruction remained significantly associated with worse OS (hazard ratio [HR] 1.57 (1.12–2.20), P = .001) and DFS (HR 1.52, 95% confidence interval [CI] 1.18–1.95, P < .001; Table 3). Factors other than obstruction independently associated with DFS on multivariable analysis included BMI, lymph node status, and T stage.
Reduced OS independently correlated with age (P = .04), male sex (HR: 1.61, P < .001), N2 status (HR: 2.28, P < .001), higher T stage (P = .005), and poorly differentiated or undifferentiated histologic type (HR: 1.55, P = .006). Repeating this analysis while including trial patients who received cetuximab yielded comparable results (N = 2,968; Supplemental Table 1). Discussion In this study, data from N0147 were used to evaluate the longterm prognostic impact of obstructive presentation in stage III colon cancer and revealed an independent detrimental effect of obstruction on long-term DFS and OS. This corroborates several previous reports that recognized obstruction as a high-risk feature in colon cancer.20,21 However, this study is unique because it considers a cohort of patients subjected to prospective random allocation and systematic follow-up for a relatively long period. Importantly, the worsened OS and DFS attributed to obstruction here were independent of known adverse prognostic features such as T stage, histologic grade, number of positive lymph nodes, and receipt of adjuvant therapy. Collectively, these findings suggest that obstruction in stage III colon cancer could represent a biologically distinct state that is not captured by conventional staging systems. Indeed, although stage III colon cancer, by definition, is a disease not detectable
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
ARTICLE IN PRESS
JID: YMSY 4
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7 Table 2 Univariate analysis of overall survival (OS) and disease-free survival (DFS) at 5 years. Variable Age <40 40–69 ≥70 Sex Male Female BMI categorical <25 25–30 ≥30 BMI continuous ECOG performance status 0 1–2 KRAS status Mutant Wild type Race Black White Other T stage T1 or T2 T3 T4 Histology Low (well or moderately differentiated) High (poorly differentiated or undifferentiated) No. of positive LNs 1–3 ≥4 Chemotherapy FOLFOX FOLFIRI Obstruction Yes No
OS at 5 y (%)
P
DFS at 5 y (%)
.004 74.2 77.2 73.1
.04 61.8 66.6 59.6
.001 70.0 83.0
.2 62.2 67.4
.04 79.4 80.1 69.0 NA
P
.048 .02
78.3 69.3
.009 68.7 68.8 56.5 NA
0.03 0.1
67.0 57.8 .2
73.4 77.6
.06 59.4 67.4
.08 69.7 76.0 84.2
.02 45.0 64.9 84.9 <.001
.001 85.63 76.49 60.9
85.4 62.9 52.0 <.001
72.7 77.3
.001 60.0 66.8
<.001 82.7 65.9
<.001 72.8 52.5
.9 73.4 76.6
.9 63.2 64.8
<.001 67.7 78.0
<.001 53.9 67.0
LN, lymph node; NA, not applicable.
Fig 1. Patients presenting with obstruction (dashed blue lines) had significantly worse disease-free (left panel) and overall survival (right panel) than patients who were not obstructed (solid red lines, P < .001 for both).
beyond regional lymph nodes, a significant proportion of patients experience disease recurrence both locally and distantly after resection, suggesting that microscopic metastasis often exists at the time of presentation.22,23 Conceptually, this has been validated by observing circulating tumor cells in the majority of patients with stage III colon cancer.24 This phenomenon may be accen-
tuated with obstruction because luminal distension and inflammation can lead to increased mucosal permeability and subsequent release of mediators known to promote distant metastasis.25,26 Clinical evidence to support this hypothesis can be gleaned from a retrospective analysis of patients with stage II and III colon cancer where preoperative obstruction correlated with distant
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
ARTICLE IN PRESS
JID: YMSY
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
5
Table 3 Multivariable analysis of overall survival (OS) and disease-free survival (DFS) expressed in hazard ratios. Variable
HR (95% CI)
P
HR (95% CI) DFS
P
.04
1.00 0.62 (0.43–0.91) 0.64 (0.41–1.00) 1.15 (0.93–1.43) 1.00 2.81 (1.71–4.62) 4.20 (2.43–7.29) 2.07 (1.67–2.56) 1.25 (0.99–1.57) 1.09 (0.86–1.39) 1.00 1.09 (0.83–1.44) 1.67 (1.29–2.17) 1.52 (1.18–1.95)
.1
OS Age (<40 y reference)∗ 40–69 ≥70 Sex (male versus female) T stage (T1–2 reference)∗ T3 T4 No. of positive LNs (≥4 vs 1–3) Histology (high versus low) ECOG PS (1–2 vs 0) BMI (<25 reference) 25–30 >30 Obstruction (yes versus no)
1.00 0.52 (0.32–0.86) 0.72 (0.41–1.27) 1.61 (1.19–2.18) 1.00 1.57 (0.89–2.74) 2.11 (1.10–4.06) 2.28 (1.69–3.06) 1.55 (1.14–2.09) 1.32 (0.96–1.81) 1.00 1.17 (0.80–1.71) 1.75 (1.22–2.52) 1.57 (1.12–2.20)
<.001 .005
<.001 .006 .1 .006
.001
.1 <.001
<.001 .07 .5 <.001
<.001
LN, lymph node. ∗ For age and T stage, P values represent significance of each variable category in the model.
(liver) (P = .019) but not with peritoneal or local recurrences.27 Further insight on the natural history of obstructed colon cancer can be learned by analyzing the Dutch Stent-In 2 trial, which compared emergency surgery with stenting for the treatment of colon cancer presenting acutely.28 Interestingly, in patients undergoing immediate resection, 10% of specimens harbored microperforations, suggesting that occult direct tumor spillage occurs more often than commonly appreciated in obstructive tumors. There, locoregional and distant recurrences developed in 6.3% and 21.9%, respectively, in the resected group as opposed to 19.2% and 30.8% in the stented group.29 The authors hypothesized that malignant cell shedding occurs with obstruction-related perforations that predispose to locoregional (peritoneal) recurrences, representing an additional mechanism of future failure in obstructive tumors. In a recent meta-analysis of long-term oncologic outcomes after stent versus emergency surgery for malignant left-sided colon cancer obstructions, Ceresoli et al10 found no significant differences in recurrence rates among either approach and further highlighted the paucity of studies that reported outcomes beyond 3 years. Specifically, only 3 randomized controlled trials met the inclusion criteria and reported median follow-up times of more than 30 months. Those trials included a combined 134 patients in total. This lack of prospective survival data is also noted when one examines exclusion criteria of relevant and recent clinical trials on resectable colon cancer where obstructive presentation commonly precluded enrollment (Supplemental Table 2). Completing recommended courses of adjuvant chemotherapy within appropriate time frames has been found to lead to improved survival in stage III colon cancer.30 It has been postulated that worse outcomes with obstruction might be due to patients suffering postoperative complications and therefore being less able to complete chemotherapy. In this study we found that participants presenting with obstruction who enrolled in the clinical trial after their resection were no less likely to complete recommended adjuvant doses of chemotherapy. High rates of chemotherapy completion in this cohort could be explained by the selection bias inherent to N0197, which required an adequate PS for random allocation after resection. However, even if this population does not represent all patients with obstruction, many of whom will not be able to receive optimal chemotherapy, the observation that patients with obstruction in N0197 had poorer survival despite receiving the same intensity of chemotherapy further emphasizes the importance of obstruction as a prognostic marker.
Several clinicopathologic factors independently influencing outcomes have been identified in stage III colon cancer; however, the AJCC has traditionally focused on anatomic extent of disease for staging because it allows for straightforward stratification and communication.31,32 In its most recent edition, the AJCC colorectal committee endorsed the use of the ACCENT-based Web calculator to help predict recurrence and OS in stage III colon cancer.14 This tool incorporates essential additional variables such as age, sex, race, BMI, PS, grade, treatment group, and tumor location; however, it does not account for obstructive presentation. The present study confirms the importance of traditional staging variables, with T and N stage correlating strongly with survival in the multivariable model. Although more advanced T and N stage were more common in patients with obstruction, even after accounting for these factors, obstruction remained independently associated with survival, suggesting that obstruction confers additional risk beyond that described by T and N stage alone in stage III colorectal cancer. Our results suggest that inclusion of obstruction could improve this tool. In this study, patients with lower BMIs were more likely to have presented with obstruction. This trend was reported in a review of 2 National Surgical Adjuvant Breast and Bowel Project trials in which patients with lower BMIs (<18.5) more commonly presented with obstruction than obese (≥35.0) patients (17.7% vs 8.0%).33 Although the recognized increased colon cancer mortality risk in obese participants may be explained through insulin-like growth factor axis disturbances from hyperinsulinemia,34 the relationship of BMI to colon cancer obstruction is less clear. One possible explanation is that patients with higher BMIs were more likely to have suffered postoperative complications after emergent operations performed for obstruction, prohibiting them from meeting trial inclusion criteria. Another possibility is that near-obstructive tumors might impair alimentation, causing weight loss and low BMI. Several limitations of this study must be noted because important granular data on preoperative carcinoembryonic antigen levels, the presence of lymphovascular invasion, KRAS and BRAF mutational status, other genomic profiling, and tumor location were absent in the available dataset. In one study on 1,877 patients with node-negative colon and rectal cancers, emergency presentation (251, 13.4%) was associated with older age (P = .023), T4 stage (P < .001), extramural vascular invasion (P = .001), and care under “nonspecialist surgeons” (P < .001), all of which could lead to worse outcomes.35 Those factors could only be partially accounted
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY 6
ARTICLE IN PRESS
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7
for in this report and may have affected results. Although emergency resections were allowed, those patients who fared poorly and were not fit for chemotherapy after surgery would not have been included. Another limitation is that the trial protocol documents do not rigorously define obstruction. Obstruction was documented when the enrolling physician indicated on enrollment forms that it was present. The lack of further data regarding location, degree, preoperative interventions, and other details about the obstruction may complicate generalizability of these results to individual patients. Despite this limitation, knowing that obstruction, even if broadly defined, correlates with worse outcomes holds value, and further studies could investigate whether the impact of obstruction depends on the details of the obstruction, such as whether clinical or endoscopic, whether stents were placed, and other factors. Perhaps of most significance is the lack of tumor location information, particularly right- versus left-sidedness because it is associated both with long-term outcomes and likelihood of obstruction. Right-sided obstructions are believed to occur with more bulky tumors capable of overcoming a larger circumference, which are linked to worse oncologic outcomes.36 A more recent study, however, contested this finding and asserted no difference in rates of recurrence or OS among proximal and distal tumors presenting emergently.37 The precise effect of an obstruction’s location, therefore, remains uncertain, and this study is unfortunately unable to provide additional information on this. Finally, that random allocation occurred after resection may have precluded those patients adversely affected by comorbidities and postoperative complications from participating. Failing to include the sickest patients, however, would likely act to reduce the impact of obstruction on survival, suggesting that the influence of obstruction in the general population may be even greater. In conclusion, the long-term prognosis in patients with stage III colon cancer was adversely affected by preoperative obstruction independent of other known negative prognostic factors. We suggest that obstruction represents a clinical marker that can be used to identify patients at higher risk of recurrence and death. Such patients may be more likely to benefit from novel or more intensive therapies, and we advocate for their inclusion in future trials. Acknowledgments This publication is based on research using information obtained from www.projectdatasphere.org, which is maintained by Project Data Sphere, LLC. Neither Project Data Sphere, LLC nor the owner(s) of any information from the website have contributed to, approved or are in any way responsible for the contents of this publication. Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.surg.2018.06.044. References 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin. 2016;66:7–30. 2. DeSantis CE, Siegel RL, Sauer AG, Miller KD, Fedewa SA, Alcaraz KI, et al. Cancer statistics for African Americans, 2016: progress and opportunities in reducing racial disparities. CA Cancer J Clin. 2016;66:290–308. 3. Mulcahy HE, Skelly MM, Husain A, O’Donoghue DP. Long-term outcome following curative surgery for malignant large bowel obstruction. Br J Surg. 1996;83:46–50. 4. Cheynel N, Cortet M, Lepage C, Benoit L, Faivre J, Bouvier AM. Trends in frequency and management of obstructing colorectal cancers in a well-defined population. Dis Colon Rectum. 2007;50:1568–1575. 5. Cortet M, Grimault A, Cheynel N, Lepage C, Bouvier AM, Faivre J. Patterns of recurrence of obstructing colon cancers after surgery for cure: a population-based study. Colorectal Dis. 2013;15:1100–1106.
6. Carraro PG, Segala M, Cesana BM, Tiberio G. Obstructing colonic cancer: failure and survival patterns over a ten-year follow-up after one-stage curative surgery. Dis Colon Rectum. 2001;44:243–250. 7. Mella J, Biffin A, Radcliffe AG, Stamatakis JD, Steele RJ. Population-based audit of colorectal cancer management in two UK health regions. Colorectal Cancer Working Group, Royal College of Surgeons of England Clinical Epidemiology and Audit Unit. Br J Surg. 1997;84:1731–1736. 8. Matsuda A, Miyashita M, Matsumoto S, Matsutani T, Sakurazawa N, Takahashi G, et al. Comparison of long-term outcomes of colonic stent as “bridge to surgery” and emergency surgery for malignant large-bowel obstruction: a meta-analysis. Ann Surg Oncol. 2015;22:497–504. 9. Sabbagh C, Browet F, Diouf M, Cosse C, Brehant O, Bartoli E, et al. Is stenting as “a bridge to surgery” an oncologically safe strategy for the management of acute, left-sided, malignant, colonic obstruction? A comparative study with a propensity score analysis. Ann Surg. 2013;258:107–115. 10. Ceresoli M, Allievi N, Coccolini F, Montori G, Fugazzola P, Pisano M, et al. Long-term oncologic outcomes of stent as a bridge to surgery versus emergency surgery in malignant left side colonic obstructions: a meta-analysis. J Gastrointest Oncol. 2017;8:867–876. 11. Dienstmann R, Salazar R, Tabernero J. Personalizing colon cancer adjuvant therapy: selecting optimal treatments for individual patients. J Clin Oncol. 2015;33:1787–1796. 12. Amin MB, Greene FL, Edge SB, Compton CC, Gershenwald JE, Brookland RK, et al. The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more "personalized" approach to cancer staging. CA Cancer J Clin. 2017;67:93–99. 13. Kattan MW, Hess KR, Amin MB, Lu Y, Moons KG, Gershenwald JE, et al. American Joint Committee on Cancer acceptance criteria for inclusion of risk models for individualized prognosis in the practice of precision medicine. CA Cancer J Clin. 2016;66:370–374. 14. Renfro LA, Grothey A, Xue Y, Saltz LB, Andre T, Twelves C, et al. ACCENT-based web calculators to predict recurrence and overall survival in stage III colon cancer. J Natl Cancer Inst. 2014;106. 15. Alberts SR, Sargent DJ, Nair S, Mahoney MR, Mooney M, Thibodeau SN, et al. Effect of oxaliplatin, fluorouracil, and leucovorin with or without cetuximab on survival among patients with resected stage III colon cancer: a randomized trial. JAMA. 2012;307:1383–1393. 16. Project Data Sphere [online database] [cited 2018 Jan]. Available from https: //www.projectdatasphere.org. 17. Saltz LB, Niedzwiecki D, Hollis D, Goldberg RM, Hantel A, Thomas JP, et al. Irinotecan fluorouracil plus leucovorin is not superior to fluorouracil plus leucovorin alone as adjuvant treatment for stage III colon cancer: results of CALGB 89803. J Clin Oncol. 2007;25:3456–3461. 18. Van Cutsem E, Labianca R, Bodoky G, Barone C, Aranda E, Nordlinger B, et al. Randomized phase III trial comparing biweekly infusional fluorouracil/ leucovorin alone or with irinotecan in the adjuvant treatment of stage III colon cancer: PETACC-3. J Clin Oncol. 2009;27:3117–3125. 19. Ychou M, Raoul JL, Douillard JY, Gourgou-Bourgade S, Bugat R, Mineur L, et al. A phase III randomised trial of LV5FU2 + irinotecan versus LV5FU2 alone in adjuvant high-risk colon cancer (FNCLCC Accord02/FFCD9802). Ann Oncol. 2009;20:674–680. 20. Burdy G, Panis Y, Alves A, Nemeth J, Lavergne-Slove A, Valleur P. Identifying patients with T3-T4 node-negative colon cancer at high risk of recurrence. Dis Colon Rectum. 2001;44:1682–1688. 21. Andre T, Boni C, Navarro M, Tabernero J, Hickish T, Topham C, et al. Improved overall survival with oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment in stage II or III colon cancer in the MOSAIC trial. J Clin Oncol. 2009;27:3109–3116. 22. Pantel K, Brakenhoff RH. Dissecting the metastatic cascade. Nat Rev Cancer. 2004;4:448–456. 23. Bernards R, Weinberg RA. A progression puzzle. Nature. 2002;418:823. 24. Sotelo MJ, Sastre J, Maestro ML, Veganzones S, Vieitez JM, Alonso V, et al. Role of circulating tumor cells as prognostic marker in resected stage III colorectal cancer. Ann Oncol. 2015;26:535–541. 25. Granger DN, Kvietys PR, Mortillaro NA, Taylor AE. Effect of luminal distension on intestinal transcapillary fluid exchange. Am J Physiol. 1980;239:G516–G523. 26. Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature. 2008;454:436–444. 27. Katoh H, Yamashita K, Wang G, Sato T, Nakamura T, Watanabe M. Prognostic significance of preoperative bowel obstruction in stage III colorectal cancer. Ann Surg Oncol. 2011;18:2432–2441. 28. van Hooft JE, Bemelman WA, Oldenburg B, Marinelli AW, Lutke Holzik MF, Grubben MJ, et al. Colonic stenting versus emergency surgery for acute left-sided malignant colonic obstruction: a multicentre randomised trial. Lancet Oncol. 2011;12:344–352. 29. Sloothaak DA, van den Berg MW, Dijkgraaf MG, Fockens P, Tanis PJ, van Hooft JE, et al. Oncological outcome of malignant colonic obstruction in the Dutch Stent-In 2 trial. Br J Surg. 2014;101:1751–1757. 30. Cheung HY, Chung CC, Tsang WW, Wong JC, Yau KK, Li MK. Endolaparoscopic approach vs conventional open surgery in the treatment of obstructing left-sided colon cancer: a randomized controlled trial. Arch Surg. 2009;144:1127–1132. 31. Vaccaro CA, Bonadeo FA, Benati ML, Quintana GM, Rubinstein F, Mullen E, et al. Colorectal cancer staging: reappraisal of N/PN classification. Dis Colon Rectum. 2004;47:66–69.
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044
JID: YMSY
ARTICLE IN PRESS
[m5G;October 2, 2018;16:8]
F.S. Dahdaleh et al. / Surgery 000 (2018) 1–7 32. Johnson PM, Porter GA, Ricciardi R, Baxter NN. Increasing negative lymph node count is independently associated with improved long-term survival in stage IIIB and IIIC colon cancer. J Clin Oncol. 2006;24:3570–3575. 33. Dignam JJ, Polite BN, Yothers G, Raich P, Colangelo L, O’Connell MJ, et al. Body mass index and outcomes in patients who receive adjuvant chemotherapy for colon cancer. J Natl Cancer Inst. 2006;98:1647–1654. 34. Renehan AG, Frystyk J, Flyvbjerg A. Obesity and cancer risk: the role of the insulin-IGF axis. Trends Endocrinol Metab. 2006;17:328–336. 35. Oliphant R, Mansouri D, Nicholson GA, McMillan DC, Horgan PG, Morrison DS, et al. Emergency presentation of node-negative colorectal cancer treated with curative surgery is associated with poorer short and longer-term survival. Int J Colorectal Dis. 2014;29:591–598.
7
36. Wolmark N, Wieand HS, Rockette HE, Fisher B, Glass A, Lawrence W, et al. The prognostic significance of tumor location and bowel obstruction in Dukes B and C colorectal cancer. Findings from the NSABP clinical trials. Ann Surg. 1983;198:743–752. 37. Frago R, Biondo S, Millan M, Kreisler E, Golda T, Fraccalvieri D, et al. Differences between proximal and distal obstructing colonic cancer after curative surgery. Colorectal Dis. 2011;13:e116–e122.
Please cite this article as: F.S. Dahdaleh et al., Obstruction predicts worse long-term outcomes in stage III colon cancer: A secondary analysis of the N0147 trial, Surgery (2018), https://doi.org/10.1016/j.surg.2018.06.044