Fax: (212) 717-3218
Patterns of failure in patients with early onset (synchronous) resectable liver metastases from rectal cancer†
Article first published online: 19 APR 2012
Copyright © 2012 American Cancer Society
Volume 118, Issue 21, pages 5414–5423, 1 November 2012
How to Cite
Butte, J. M., Gonen, M., Ding, P., Goodman, K. A., Allen, P. J., Nash, G. M., Guillem, J., Paty, P. B., Saltz, L. B., Kemeny, N. E., DeMatteo, R. P., Fong, Y., Jarnagin, W. R., Weiser, M. R. and D'Angelica, M. I. (2012), Patterns of failure in patients with early onset (synchronous) resectable liver metastases from rectal cancer. Cancer, 118: 5414–5423. doi: 10.1002/cncr.27567
Presented in part at the annual American Society of Clinical Oncology Gastrointestinal Cancers Symposium; January 20-22, 2011; San Francisco, California
- Issue published online: 19 OCT 2012
- Article first published online: 19 APR 2012
- Manuscript Accepted: 16 FEB 2012
- Manuscript Revised: 14 FEB 2012
- Manuscript Received: 30 DEC 2011
- colorectal metastases;
- recurrence pattern;
- radiation therapy;
The optimal combination of available therapies for patients with resectable synchronous liver metastases from rectal cancer (SLMRC) is unknown, and the pattern of recurrence after resection has been poorly investigated. In this study, the authors examined recurrence patterns and survival after resection of SLMRC.
Consecutive patients with SLMRC (disease-free interval, ≤12 months) who underwent complete resection of the rectal primary and liver metastases between 1990 and 2008 were identified from a prospective database. Demographics, tumor-related variables, and treatment-related variables were correlated with recurrence patterns. Competing risk analysis was used to determine the risk of pelvic and extrapelvic recurrence.
In total, 185 patients underwent complete resection of rectal primary and liver metastases. One hundred eighty patients (97%) received chemotherapy during their treatment course, and 91 patients (49%) received pelvic radiation therapy either before (N = 65; 71.4%), or after (N = 26; 28.6%) rectal resection. The 5-year disease-specific survival rate was 51% for the entire cohort with a median follow-up of 44 months for survivors. One hundred thirty patients (70%) developed a recurrence: Eighteen patients (10%) had recurrences in the pelvis in combination with other sites, and 7 of these (4%) had an isolated pelvic recurrence. Recurrence pattern did not correlate with survival. Competing risk analysis demonstrated that the likelihood of a pelvic recurrence was significantly lower than that of an extrapelvic recurrence (P < .001).
Of the patients with SLMRC who developed recurrent disease, systemic sites were overwhelmingly more common than pelvic recurrences. The current results indicated that the selective exclusion of radiotherapy may be considered in patients who are diagnosed with simultaneous disease. Cancer 2012. © 2012 American Cancer Society.
Approximately 40,000 patients were diagnosed with rectal cancer in the United States in 2010.1 In general, from 20% to 30% of patients who have rectal cancer present with metastatic disease at diagnosis, and 30% of these patients have isolated liver metastases.2 The treatment of rectal cancer without metastatic disease has been well studied, and there are established treatment paradigms.3 Complete mesorectal excision has improved the rate of local control and disease-free survival (DFS).4 Preoperative chemotherapy and pelvic radiation therapy (RT) in patients with advanced rectal cancer have been associated with lower local recurrence rates and improved sphincter preservation rates in randomized trials5, 6 but have not consistently demonstrated an impact on overall survival.5-8
Complete resection of liver metastases from colorectal cancer has become standard treatment for patients with resectable disease.9-11 Hepatic resection is associated with better survival compared with chemotherapy alone and, more important, is the only treatment for which long-term survival and cure have been reported.12 However, approximately 70% of patients will recur after partial hepatectomy, most commonly in the liver and lungs.13
There are no established guidelines for treating patients with resectable, simultaneous liver metastases from rectal cancer.14 Some authors have suggested that adjuvant or neoadjuvant systemic chemotherapy, combined with resection, should be the treatment of choice. Others have used perioperative chemotherapy and pelvic radiation for locally advanced rectal tumors followed by resection of both the primary tumor and liver metastases.14 One reason why definitive treatment recommendations have been difficult to make is that patterns of recurrence after resection of the primary rectal tumor and liver metastases have not been well defined. Thus, a better delineation of recurrence and progression would help guide therapeutic strategies.
Patients who present with simultaneous hepatic metastases at primary presentation or at a short interval (<1 year) tend to have similar outcomes and generally have been considered to have synchronous metastatic disease.15 To better understand the natural history of these patients, we defined patients with resectable synchronous liver metastases from rectal cancer (SLMRC) as those who developed liver metastases within 1 year of presentation and were able to undergo complete resection. The primary objective of the current retrospective study was to evaluate the patterns of recurrence in patients who underwent complete resection of SLMRC. Secondary aims were to explore the factors associated with recurrence patterns and long-term survival.
MATERIALS AND METHODS
Patients and Data Collection
After receipt of institutional review board approval from Memorial Sloan-Kettering Cancer Center (MSKCC), all patients who underwent liver resection for colorectal liver metastases between 1990 and 2008 were identified from a prospective database. Records of patients with SLMRC were selected and analyzed retrospectively. Synchronous presentation was defined as disease-free interval (DFI) <12 months between the diagnoses of rectal cancer and liver metastases.15 Although treatment recommendations among this group of patients likely differ based on whether or not metastases were present at presentation (simultaneous), patients with a DFI <12 months generally are categorized with synchronous15 disease and likely represent similar tumor biology, allowing us to analyze the outcome and recurrence patterns in a similar group of patients. Prospectively recorded data were supplemented with a retrospective review of the medical records that included additional details on preoperative laboratory values, operative procedures, perioperative outcomes, tumor histopathology and staging, follow-up, patterns of recurrence, and survival. This study includes patients who underwent rectal resection elsewhere and subsequently underwent resection of liver metastases at MSKCC. By using a previously published system, a clinical risk score (CRS) was calculated for each patient before liver resection.15
Selection criteria for liver resection included no medical condition that would preclude a major abdominal operation and an adequate future liver remnant. Extrahepatic disease was a relative contraindication to liver resection; however, a small group of carefully selected patients with limited and resectable extrahepatic disease was included. Preoperative evaluation included physical examination and imaging studies to assess the extent of disease, a complete colonoscopy within 1 year of hepatic resection, computed tomography (CT), abdominal and pelvic CT, or magnetic resonance imaging (MRI). 18F-fluorodeoxyglucose-positron emission tomography/CT (18FDG PET-CT) was used selectively. All patients were evaluated at a weekly multidisciplinary conference. Postoperative follow-up included physical examination and cross-sectional imaging every 4 to 6 months.
Operative and Perioperative Details
The principles of surgical treatment of locally advanced rectal cancer and liver metastases have been published previously.11, 12, 16-18 Patients who underwent surgery at MSKCC received standard mesorectal excision. All specimens from resections performed at referring hospitals were reviewed by pathologists at MSKCC. Liver resection was performed as previously described. Major hepatectomy was defined as a right or left hemihepatectomy or greater. Minor hepatectomy was defined as resection encompassing less than a hemihepatectomy. Only selected patients who had known or suspected lymph node disease underwent porta hepatis lymphadenectomy. Hepatic arterial infusion pumps (HAIPs) were placed at the time of surgery for adjuvant regional chemotherapy at the discretion of the treating physicians, most often in prospective clinical trials. Data recorded included neoadjuvant and adjuvant RT and/or chemotherapy.
Follow-Up and Survival Analysis
Surgical mortality was defined as any death resulting from postoperative complication and clearly related to the operative procedure (medical or nonmedical) within 90 days of surgery. At the time of last follow-up, patient status was categorized as follows: no evidence of disease, alive with disease, dead of disease, dead of other causes, and dead of unknown causes. Disease-specific survival (DSS) was defined as the interval between the date of the last definitive operation that resulted in complete removal of all disease and the date of last follow-up or death from rectal cancer. Recurrence-free survival (RFS) was defined as the interval between the date of complete removal of disease and either the date of first recurrence or the date of last follow-up in patients without recurrence. Patients with an uncertain date of recurrence were excluded from the RFS analysis but were included in the analysis of recurrence patterns.
Recurrence was defined as any sign of recurrent disease documented clinically, radiographically, and/or pathologically. All recurrences that initially were documented clinically were confirmed by biopsy or serial radiologic follow-up. Radiologic proof of recurrence was documented with sequential imaging that demonstrated progression of disease. Salvage treatment was defined as resection and/or ablation of all recurrent disease. Subsequent recurrences after such treatments also were recorded.
Final disease staging was based on the seventh edition of the American Joint Committee on Cancer (AJCC) manual.19 Curative-intent treatment (R0 resection) was defined as complete resection without microscopic involvement of any margins, and R1/R2 resection was defined as residual microscopic or macroscopic (or gross) disease, respectively, at 1 or more margin, at either the rectal primary or the liver metastases.
Categorical variables were summarized using proportions, and continuous variables were summarized using mean ± standard deviation and median (range) values. Characteristics of patients were compared across groups using the chi-square test for categorical variables and the Kruskal-Wallis test for continuous variables. Univariate Cox proportional hazards regression was used to identify factors individually predictive of pelvic recurrence (exclusively or associated with another site of recurrence) for the entire cohort. Multivariate analysis was not used, because few events were observed (pelvic recurrence). Survival curves were constructed using the Kaplan-Meier method and were compared using the log-rank test. Competing risk analysis was used to determine the risk of pelvic and extrapelvic recurrence and its association with or without RT. All tests were 2-sided, and statistical significance was defined at P < .05. Statistical analysis was performed using the statistical software packages SAS (version 9.2; SAS Institute Inc., Cary, NC) and SPSS (version 19.0; SPSS Inc., Chicago, Ill).
Between January 1, 1990 and June 30, 2008, in total, 1899 patients underwent liver resection for colorectal metastases at MSKCC. One hundred eighty-five patients (9.7%) presented with SLMRC and were analyzed in this study. Table 1 summarizes the clinical characteristics of the study cohort. There were 115 men (62.2%), and the median patient age was 55 years (range, 18-88 years). Simultaneous diagnoses of primary and liver metastases were made in 86 of these patients (46.5%). One hundred eleven patients (60%) underwent rectal resection at MSKCC.
|Characteristic||No. of Patients (%), N = 185|
|Mean±SD||54.7 ± 12.7|
|Median [range]||55 (18-88)|
|Not known||6 (3.2)|
|Not known||7 (3.8)|
|Not known||40 (21.6)|
|Lymph node status|
|Not known||3 (1.6)|
|Resection at MSKCC|
|Received perioperative radiotherapy at MSKCC|
|DFI 0 mo||86 (46.5)|
|DFI >0 mo and ≤12 mo||99 (53.5)|
|Median [range]||1 [0-12]|
|Liver margin, status|
|Low: 0-2||93 (50.3)|
|High: 3-5||92 (49.7)|
Perioperative treatments included systemic or HAIP chemotherapy and/or RT. Overall, 91 patients (49.2%) received pelvic RT either before (N = 65; 71.4%) or after (N = 26; 28.6%) rectal resection. With only 1 exception, all patients who received RT also received concurrent chemotherapy. Overall, 180 patients (97.3%) received systemic chemotherapy at some point during their perioperative treatments.
Among the 86 patients who had a simultaneous diagnosis, treatment of the rectal tumor included chemoradiation in 35 patients (40.7%) either before (N = 30; 85.7%) or after (N = 5; 14.3%) surgery. Of these 35 patients, all received additional systemic chemotherapy (2 preoperatively, 9 postoperatively, and 24 both preoperatively and postoperatively). Of the 47 patients who did not receive chemoradiation, all received systemic chemotherapy (3 preoperatively, 10 postoperatively, and 34 both preoperatively and postoperatively). In this subset of 86 patients, 24 patients also received HAIP chemotherapy, and 3 received experimental vaccine therapy.
Among the 99 patients without a simultaneous diagnosis, treatment of the rectal tumor included chemoradiation in 56 patients (56.6%) either before (N = 35; 62.5%) or after (N = 21; 37.5%) surgery. Of these 56 patients who received chemoradiation, 23 also received systemic chemotherapy (in addition to the concurrent chemotherapy received with radiation) either before (N = 3), after (N = 11), or both before and after (N = 9) rectal resection. Of the 43 patients in this subset who did not receive radiation, 28 received systemic chemotherapy (1 before, 26 after, and 1 both before and after resection), and 1 received experimental vaccine therapy. Once liver metastases developed in these 99 patients, most (N = 92; 92.9%) received systemic chemotherapy in addition to undergoing hepatic resection (5 before, 64 after, and 23 both before and after resection). In addition, 28 of these 99 patients (28.3%) also received HAIP chemotherapy.
Recurrence-Free Survival and Recurrence Patterns
The median follow-up was 38 months (range, 1-185 months) for all patients and 44 months (range, 1-185 months) for survivors. The median overall RFS for all patients (N = 180) was 17 months (95% confidence interval, 12.7-21.3 months), and the actuarial RFS rate was 34% at 3 years and 28% at 5 years. Overall recurrence was associated with advanced tumor classification (P = .048), perineural invasion (P = .044) and positive lymph nodes in the primary (P = .003), major liver resection (P = .025), the presence of extrahepatic disease (P = .019), number (P = .01) and size (P = .043) of liver metastases, and high CRS (P < .0001); whereas systemic (P = .38) or HAIP (P = .73) chemotherapy and RT (P = 1.0) were not associated with recurrence.
Figure 1 details the recurrence patterns, salvage treatments, and subsequent recurrence patterns. One hundred thirty patients (70.3%) developed a recurrence at a median interval of 12 months (range, 1-114 months). Most recurrences involved the liver and lung. Overall, pelvic recurrence was an uncommon event (N = 18; 9.7%) that typically was associated with an extrapelvic site of recurrence (11 of 18 patients) and was greatly exceeded by extrapelvic failures. Isolated pelvic recurrence occurred in 7 patients (3.8%).
Factors associated with pelvic recurrence are presented in Table 2. In general, patients with and without pelvic recurrence had the same clinical and pathologic features, except for younger age in patients with pelvic recurrence. It is noteworthy that pelvic radiation was not associated with the rate of pelvic recurrence. This observation was made for the second, third, and fourth recurrences after salvage treatments (data not shown). The receipt of perioperative RT was associated with the absence of lymphovascular invasion in the primary tumor (N = 55 [76.4%] vs N = 17 [23.6%]; P = .03) and patients without a simultaneous diagnosis (N = 56 [61.5%] vs N = 35 [38.5%]; P = .04). No other differences in demographics, clinical presentation, or surgical treatment were observed between the patients who did and did not receive RT.
|No. of Patients (%)|
|Characteristic||No Pelvic Recurrence, N = 167||Pelvic Recurrence, N = 18||P|
|Women||66 (39.5)||4 (22.2)|
|Men||101 (60.5)||14 (77.8)|
|Median [range]||56 [18-88]||50 [26-70]|
|Tumor classification, N = 179||.5|
|T1||3 (1.8)||0 (0)|
|T2||16 (9.6)||0 (0)|
|T3||134 (80.2)||17 (94.4)|
|T4||9 (5.4)||0 (0)|
|Tumor differentiation, N = 178||.5|
|Well||3 (1.8)||1 (5.6)|
|Moderate||140 (83.8)||14 (77.8)|
|Poor||19 (11.4)||1 (5.6)|
|Lymph node status||.8|
|Positive||103 (61.7)||12 (66.7)|
|Negative||64 (38.3)||6 (33.3)|
|LVI, N = 145||.45|
|Positive||41 (24.6)||6 (33.3)|
|Negative||91 (54.5)||7 (38.9)|
|PNI, N = 132||.3|
|Positive||35 (21)||6 (33.3)|
|Negative||85 (50.9)||6 (33.3)|
|Rectal margin status, N = 182||.7|
|Positive||12 (7.2)||2 (11.1)|
|Negative||152 (91)||16 (88.9)|
|Underwent rectal resection at MSKCC||.3|
|Yes||98 (58.7)||13 (72.2)|
|Elsewhere||69 (41.3)||5 (27.8)|
|Yes||83 (49.7)||8 (44.4)|
|No||84 (50.3)||10 (55.6)|
|DFI 0 mo||74 (44.3)||12 (66.7)|
|DFI >0 mo and ≤12 mo||93 (55.7)||6 (33.3)|
|Type of resection||.3|
|Major||90 (53.9)||12 (66.7)|
|Minor||77 (46.1)||6 (33.3)|
|Positive||21 (12.6)||1 (5.6)|
|Negative||146 (87.4)||17 (94.4)|
|No||155 (92.8)||18 (100)|
|No. of tumors||.7|
|Median [range]||2 [0-11]||2.5 [1-8]|
|Largest tumor size, cm||.3|
|Median [range]||3 [0-22]||3.4 [1-6.5]|
|Preoperative CEA, ng/mL: N = 159||.3|
|Median [range]||9.8 [0-2637]||6.2 [1-136]|
|1||22 (13.2)||0 (0)|
|2||60 (35.9)||11 (61.1)|
|3||60 (35.9)||7 (38.9)|
|4||23 (13.8)||0 (0)|
|5||2 (1.2)||0 (0)|
|Low: 0-2||82 (49.1)||11 (61.1)|
|High: 3-5||85 (50.9)||7 (38.9)|
Competing risk analysis demonstrated a significantly lower risk of any pelvic recurrence (with or without extrapelvic recurrence) compared with extrapelvic recurrence (P < .001) independent of the receipt of RT. Specifically, the risk of any pelvic recurrence increased from 3% at the first year to 8% at the fifth year in patients who received RT and from 6% at the first year to 9% at the fifth year in patients who did not receive RT (P = .53). In contrast, the risk of extrapelvic recurrence increased from 33% at the first year to 63% at the fifth year in patients who received RT and from 33% at the first year to 63% at the fifth year in patients who did not receive RT (P = .91) (Fig. 2).
The median DSS for all patients (N = 185) was 62 months (95% confidence interval, 54.3-69.7 months), and the actuarial DSS rate was 73% at 3 years and 51% at 5 years. Factors that were associated significantly with a prolonged median DSS included a negative microscopic liver resection margin (67 months vs 26 months; P < .0001) and the absence of lymphovascular invasion (84 months vs 50 months; P = .009) or lymph node disease in the primary tumor (not reached vs 58 months; P = .003), and a lower CRS (1 point = not reached, 2 points = 89 months, 3 points = 60 months, 4 points = 43 months, 5 points = 6 months; P = .001). DSS was not associated with pelvic recurrence (median DSS, 51 months vs 63 months; P = .2) or the receipt of perioperative RT (median DSS, 84 months vs 58 months; P = .3).
Analysis of Patients With Simultaneous Diagnosis of Rectal Cancer and Liver Metastases
Among the 86 patients who had a simultaneous diagnosis of liver metastases, the receipt of pelvic RT was associated significantly with a lower rate of lymphovascular invasion in the primary tumor (P = .01), less liver metastases (P = .015), and less major liver resections (P = .04) compared with patients who did not receive RT. However, patients who did and did not receive RT had the same incidence of recurrence (P = .8), pelvic recurrences (combined with other sites) (P = .7), and isolated pelvic recurrences (P = 1.0) (Table 3).
|No. of Patients (%)|
|Characteristic||No RT, N = 51||Perioperative RT, N = 35||P|
|Women||18 (35.3)||13 (37.1)|
|Men||33 (64.7)||22 (62.9)|
|Median [range]||54 [18-81]||50 [20-73[|
|Tumor classification, N = 82||.5|
|T1||0 (0)||1 (2.9)|
|T2||5 (10.6)||2 (5.7)|
|T3||39 (83)||30 (85.7)|
|T4||3 (6.4)||2 (5.7)|
|Well||0 (0)||0 (0)|
|Moderate||47 (92.2)||31 (88.6)|
|Poor||4 (7.8)||4 (11.4)|
|Lymph node status||.6|
|Positive||31 (60.8)||19 (54.3)|
|Negative||20 (39.2)||16 (45.7)|
|LVI, N = 75||.01|
|Positive||22 (48.9)||6 (20)|
|Negative||23 (51.1)||24 (80)|
|PNI, N = 71||1.0|
|Positive||14 (34.1)||10 (33.3)|
|Negative||27 (65.9)||20 (66.7)|
|Rectal margin status, N = 85||.3|
|Positive||5 (9.8)||1 (2.9)|
|Negative||46 (90.2)||33 (97.1)|
|Underwent rectal resection at MSKCC||.06|
|Yes||4 (21.1)||31 (46.3)|
|Elsewhere||15 (78.9)||36 (53.7)|
|Yes||3 (5.9)||1 (2.9)|
|No||48 (94.1)||34 (97.1)|
|No. of tumors||.015|
|Median [range]||2 [1-11]||1 [1-7]|
|Largest tumor size, cm||.1|
|Median [range]||2.9 [0.7-22]||2 [0.5-11]|
|Preoperative CEA, ng/mL: N = 72||.5|
|Median [range]||8.4 [1-1387]||5.6 [0-2637]|
|Type of resection||.04|
|Major||23 (45.1)||8 (22.9)|
|Minor||28 (54.9)||27 (77.1)|
|Surgical margin status||.3|
|Positive||7 (13.7)||2 (5.7)|
|Negative||44 (86.3)||33 (94.3)|
|1||3 (5.9)||7 (20)|
|2||22 (43.1)||15 (42.9)|
|3||20 (39.2)||9 (25.7)|
|4||5 (9.8)||4 (11.4)|
|5||1 (2)||0 (0)|
|Low: 0-2||25 (49)||23 (65.7)|
|High: 3-5||26 (51)||12 (34.3)|
|Yes||34 (66.7)||22 (62.9)|
|No||17 (33.3)||13 (37.1)|
|Yes||8 (15.7)||4 (11.4)|
|No||43 (84.3)||31 (88.6)|
|Isolated pelvic recurrence||1.0|
|Yes||4 (7.8)||3 (8.6)|
|No||47 (92.2)||32 (91.4)|
In these 86 patients, overall recurrence was associated with perineural invasion (P = .034), fewer resected lymph nodes in the primary tumor (P = .019), positive liver margin (P = .046), and high CRS (P = .02). There were no factors associated with pelvic recurrence. The median DSS for this subgroup was 80 months (95% confidence interval, 48.6-111.4 months), and the actuarial DSS rate was 58% at 5 years. Improved DSS was associated with negative microscopic liver resection margin (83 months vs 46 months; P = .003) and low CRS (not reached vs 62 months; P = .02). RT (P = .3), systemic chemotherapy (P = .38), and HAIP chemotherapy (P = .58) were not associated with overall recurrence or DSS.
The liver is the most common site of metastases from colorectal cancer.9 Hepatic resection in patients with resectable disease is associated with long-term survival and cure.11, 12 However, at least 70% of these patients ultimately will develop recurrence after complete resection, and these recurrences usually involve distant sites rather than locoregional recurrences at the primary tumor site.13 The natural history and patterns of recurrence in patients who undergo complete resection of SLMRC have not been well documented. Most publications include small series of patients with colon and rectal cancer in the same analysis.20-24 Rectal cancer differs from colon cancer, because it has a greater risk of locoregional recurrence, and pelvic RT is commonly used to lower this risk in patients with locally advanced tumors. The objective of the current study was to describe the recurrence patterns of patients undergoing complete resection of SLMRC and thereby provide natural history data from which treatment strategies and clinical trials could be developed.
Two previous studies have addressed recurrence patterns in patients with resected liver metastases from rectal cancer. Assumpcao et al25 reported a higher rate of systemic recurrence compared with pelvic recurrence in a retrospective evaluation of 141 patients. Perioperative RT was received by 60% of patients, and there was a 16% rate of pelvic failure, of which 7.5% were isolated pelvic recurrences. However, this study included patients who developed liver metastases later than 1 year after primary presentation, and such patients may have a different natural history and recurrence pattern than those with synchronous disease.26, 27 In addition, competing risk was not used in the analysis of recurrence patterns, and only the first episode of recurrence was considered. Kim et al28 reported on a small cohort of 89 patients with SLMRC (30% had received perioperative RT). In that study, 25 patients (28%) developed pelvic recurrence, and this was not significantly different among patients who did and did not receive RT. Those authors also analyzed only the first episode of recurrence and did not use a competing risk analysis of recurrence patterns. In addition, not all of their patients underwent liver resection, and there was a limited median follow-up of 28 months.
To our knowledge, the current study is the largest series from a single center to date that evaluates the pattern of recurrence in patients with completely resected SLMRC. It is noteworthy that this study was limited to patients who either presented with liver metastases or developed them within 12 months of diagnosis. We analyzed all patterns of recurrence over time and their treatments, and we used an appropriate statistical method (competing risk analysis) to estimate the risk of pelvic and extrapelvic recurrence. Systemic recurrences significantly exceeded pelvic failures, and the most common sites of recurrence were the liver and/or lung. Pelvic recurrence was an uncommon event, usually associated with other sites of recurrence, and was observed as an isolated episode in <5% of patients. In fact, the small number of events (N = 18) did not permit a multivariate analysis of prognostic factors associated with pelvic recurrence. Although this may be considered a limitation of our study, it highlights the finding that local pelvic recurrence is a rare event in patients with SLMRC. Competing risk analysis was used to determine the risk of pelvic recurrence and its association with RT. This kind of analysis was necessary to account for events other than the 1 of interest, which alters the probability of experiencing the event of interest (ie, pelvic recurrence). The greatest risk of recurrence is during the first 2 years after resection, as documented in previous studies,13, 23, 25, 28 and there is a plateau in both pelvic and extrapelvic sites after that time point. However, some patients presented with new recurrences after that 2-year period, confirming that prolonged follow-up is necessary to definitively study patterns of recurrence in these patients.
The risk of extrapelvic recurrence was significantly greater than the risk of pelvic recurrence, and RT was not associated with these patterns. The relation between the patterns of recurrence and RT must be interpreted with caution, because this was a retrospective comparison with inherent selection bias. Our study did not have the power or prospective design to determine whether pelvic RT is associated with decreased pelvic recurrence. Previous prospective, randomized trials have demonstrated better local control with additional pelvic RT in patients with resected rectal cancer; however, those trials were performed in patients without metastatic disease.5-8 Despite this, in a group of patients with resected liver metastases, there were no obvious clinical or pathologic associations with pelvic recurrence. Furthermore, the only differences between patients who did and did not receive RT were a higher rate of lymphovascular invasion and a longer DFI in those who received RT. It is logical that patients with a simultaneous diagnosis of metastatic disease are less likely to receive pelvic RT. Despite the clinical heterogeneity in our study population, the outcome clearly demonstrates that distant extrapelvic recurrence is the overwhelmingly dominant issue regardless of the clinical and pathologic factors involved.
This study had other limitations that must be recognized. It was a retrospective analysis that spanned 18 years with differing treatment options over the study period. Patients received heterogeneous chemotherapy regimens both before and after hepatic resection. However, to our knowledge, it has never been demonstrated that specific chemotherapy regimens alter recurrence patterns, and we do not believe that this heterogeneity would alter the principal findings of the study. Another limitation of this study was that a significant number of patients underwent rectal resection or received RT, elsewhere and quality control and/or assessment was not possible. We also included patients without a simultaneous diagnosis of liver metastases in this analysis. The decision about the use of pelvic radiation in patients with and without simultaneous metastatic disease is clearly an important issue. Patients with a simultaneous diagnosis of metastases are less likely to receive pelvic radiation. We included the whole cohort of patients with early onset liver metastases in the current analysis, because they generally are considered to have synchronous metastases with similar tumor biology. To address this issue, we analyzed the subset of patients who were diagnosed simultaneously with rectal cancer and liver metastases. This is the group of patients in which decisions regarding treatment are the most complex. We observed that this group of patients also had a very low rate of pelvic recurrence, and this was not associated with the receipt of RT.
The principal finding of this study is that the great majority of patients with resected SLMRC recur with extrapelvic failure, and this far exceeds the rate of pelvic recurrence. Isolated pelvic recurrence was very uncommon and occurred in <5% of patients. The risk of local recurrence was not associated with RT. These natural history data provide a background from which prospective trials of adjuvant/neoadjuvant therapy in patients with rectal cancer and resectable liver metastases may be developed. We also believe that, although RT has clear roles in improving rates of sphincter preservation and in patients for whom the likelihood of achieving an R0-resection is in doubt based on recurrence patterns, its exclusion should be considered in patients without these issues who have a simultaneous diagnosis of resectable liver metastases.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 19Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. AJCC Cancer Staging Handbook. 7th ed. New York: Springer Science+Business Media; 2010.