Ann Surg Oncol DOI 10.1245/s10434-014-4207-3
ORIGINAL ARTICLE – HEPATOBILIARY TUMORS
Is Complete Liver Resection Without Resection of Synchronous Lung Metastases Justified? Yoshihiro Mise, MD1, Scott Kopetz, MD2, Reza J. Mehran, MD3, Thomas A. Aloia, MD1, Claudius Conrad, MD1, Kristoffer W. Brudvik, MD1, Melissa W. Taggart, MD4, and Jean-Nicolas Vauthey, MD1 Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; 2Department of Gastrointestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; 3Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX; 4Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX 1
ABSTRACT Background. Advances in multidisciplinary care are changing the prognostic impact of colorectal lung metastases. Resection of colorectal liver metastases (CLM) may benefit patients with synchronous lung metastases even when lung metastases are not resected. The aim of this study was to investigate the survival of patients undergoing complete resection of CLM in the setting of unresected lung metastases. Patients and Methods. We compared survival among 98 patients who underwent resection of CLM with unresected lung metastases, 64 who received only chemotherapy for limited colorectal liver and lung metastases, and 41 who underwent resection of both liver and lung metastases. Prognostic factors were investigated in the patients who underwent resection of CLM only. Results. The 3-year/5-year overall survival (OS) rates of patients with CLM resection only (42.9 %/13.1 %) were better than those of patients treated with chemotherapy only (14.1 %/1.6 %; p \ 0.01) but worse than those of patients with resection of liver and lung metastases (68.9 %/56.9 %; p \ 0.01). Multivariate analysis of patients with CLM resection only revealed that KRAS mutation [hazard ratio (HR) 2.10; 95 % confidence interval (CI) 1.21–3.64; p \ 0.01] and rectal primary tumor (HR
Electronic supplementary material The online version of this article (doi:10.1245/s10434-014-4207-3) contains supplementary material, which is available to authorized users. Ó Society of Surgical Oncology 2014 First Received: 25 August 2014 J.-N. Vauthey, MD e-mail: [email protected]
1.72; 95 % CI 1.02–2.88; p = 0.04) were independent predictors of worse OS. Survival of patients without these risk factors was similar to that of patients with curative metastasectomy. Conclusions. Complete resection of metastases remains the primary goal of treatment for stage IV colorectal cancer. Resection of CLM without resection of lung metastases is associated with an intermediate survival between that of patients treated with palliative and curative intent and should be considered in selected patients.
Treatment for stage IV colorectal cancer has evolved during the past 2 decades. The presence of colorectal liver metastases (CLM) once precluded surgical treatment, resulting in a median survival of 12 months before 1990.1 However, hepatectomy is now performed in 25 % of patients with CLM and results in a 5-year overall survival (OS) rate of 58 %.2–5 Unfortunately, for the remaining 75 % of patients, medical therapy alone results in a median survival of 24 months or less.6 Thus, a current issue in the treatment of stage IV colorectal cancer is how to refine a multidisciplinary approach to rescue the patients for whom hepatectomy is not indicated. The key to expanding incorporation of hepatectomy lies in clarifying the impact of lung metastases, which are the most frequent form of extrahepatic metastasis of colorectal cancer.7 Recent series indicate that in the era of effective chemotherapy, long-term disease control can be expected in patients with unresected colorectal lung metastases, which are typically small volume.8–11 Mitry et al. reported a 3-year OS rate of 15.3 % in patients with colorectal lung metastases treated with chemotherapy alone from 1976 to 2005,5,9 whereas the 3-year OS has more than doubled
Y. Mise et al.
(35.5 %) in a series from 2000 to 2010 reported by Kim et al.10 These results suggest the hypothesis that resection of CLM, which has a significant impact on prognosis in patients with stage IV colorectal cancer,12–15 may benefit some patients with unresected lung metastases. Somatic gene mutation status as a tool for predicting tumor biology has become a focus of colorectal cancer research.16 Growing evidence demonstrates a propensity for KRAS-mutated tumors to metastasize to the lung; in patients with KRAS-mutant tumors, rates of KRAS mutation in lung metastases are high, as are rates of early recurrence in the lung after liver surgery for CLM.17–19 These findings suggest that KRAS mutation has the potential to predict tumor biology in patients with colorectal lung metastases. During the last decade, we conducted aggressive surgery for CLM in patients with synchronous lung lesions after multidisciplinary review.20,21 In this study, we identified patients whose lung lesions progressed after hepatectomy and were never resected. Their long-term survival was analyzed to investigate the prognostic impact of hepatectomy in patients with unresected lung metastases. We also evaluated prognostic factors, including KRAS mutation status, to identify patients who benefit from hepatectomy for CLM even in the setting of unresected lung metastases. METHODS The Institutional Review Board of The University of Texas MD Anderson Cancer Center approved this study protocol. A prospectively maintained database of the Department of Surgical Oncology was reviewed for 1,539 patients who underwent hepatectomy for CLM during the period from January 2000 through June 2012. Among these patients, we identified 376 patients (24.4 %) who had synchronous lung nodules when undergoing hepatectomy for CLM. Forty-four patients underwent pulmonary resection and one underwent Cyberknife treatment. In 221 of the remaining 331 patients, the synchronous lung nodules did not progress after hepatectomy. In the remaining 110 patients, the diagnosis of colorectal lung metastases was confirmed by progression, defined as a radiographic increase in the size or number of lung lesions after hepatectomy. Three patients who did not undergo planned second-stage hepatectomy were excluded from the study, leaving 107 patients. Of these 107 patients, 9 were excluded because they did not have available paraffin blocks or had insufficient DNA for genetic testing of KRAS mutation. The remaining 98 patients constituted the study population (patient selection was demonstrated in electronic supplementary Fig. S1). We reviewed the prospectively-collected database of the Department of Medical Oncology and identified 205 patients who had both liver and lung colorectal metastases
and who received systemic chemotherapy alone. From these patients, we selected patients who started chemotherapy at our institution, who had no more than 15 liver metastases and no more than 15 lung metastases on pretreatment images, and who had favorable performance status (Eastern Cooperative Oncology Group performance status 0–2). A total of 64 patients met these criteria. Finally, we selected a comparison group of patients who underwent resection of both types of metastases. Among the aforementioned 44 patients who underwent both liver and lung resection for pathologically confirmed metastases, 3 patients who did not complete planned curative metastasectomy were excluded, leaving 41 patients for analysis. Perioperative Management Before operation, all patients underwent imaging studies, including thin-slice (2.5 mm) helical computed tomography of the chest. Hepatectomy was scheduled after a short course of chemotherapy (four to six cycles). Hepatectomy was considered for patients with CLM when complete resection of CLM was feasible and limited lung metastases seemed amenable to subsequent lung resection or to long-term control with chemotherapy. Resection of lung metastases was performed as previously reported.20,21 Postoperative complications were monitored within 90 days after surgery and graded according to a standard classification.22 Major complications were defined as grade [ II. After surgery, patients were reassessed by physical, laboratory, and imaging evaluation every 3 months. KRAS Mutation Profiling KRAS mutation status was analyzed by examining codons 12, 13, and 61 of KRAS in the study population. DNA extracted from formalin-fixed, paraffin-embedded tissue blocks was quantified and analyzed with a polymerase chain reactionbased DNA sequencing method and Sequenom MassARRAY technology (Sequenom, Inc., San Diego, CA, USA).23 Long-Term Outcomes After Liver Resection The OS of the study population was compared with the OS of patients treated with chemotherapy alone and patients who underwent resection of synchronous liver and lung metastases. Predictors of OS were investigated in the study population. Statistical Analysis Categorical variables were compared using the Fisher’s exact test. Continuous variables were compared using the
Hepatectomy Without Resecting Lung Metastases TABLE 1 Baseline characteristics of patients with synchronous colorectal liver and lung metastases who underwent resection of liver metastases only, chemotherapy only, and resection of both liver and lung metastases Characteristic
Only liver Chemotherapy metastases only (n = 64) resected (n = 98)
Liver and lung p value p value for only liver metastases metastases resected resected vs. chemotherapy (n = 41) only
Age [years; median (range)]
55 (22–84)
59 (30–73)
57 (33–73)
0.26
0.11
Male [n (%)]
60 (61)
39 (61)
26 (63)
0.96
1.00
42 (43)
26 (41)
8 (20)
0.03
0.81
Positive lymph nodes
64 (65)
31 (66)
29 (71)
0.82
1.00
KRAS mutation [n (%)]
44 (45)
Primary disease [n (%)] Rectal primary tumor
Liver metastases Disease-free interval \1 year [n (%)]a
75 (77)
53 (83)
20 (49)
\0.01
0.43
Number [median (range)]
2 (1–15)
3 (1–14)
1 (1–9)
\0.01
0.03
Diameter of the largest nodule [cm; median (range)] 3.0 (0.5–14.0) CEA level [ng/mL; median (range)]b 4.4 (1–723)
3.4 (0.4–16.7) 2.0 (0.5–10.0) 0.02 23.7 (1.2–5,302) 2.4 (0.7–383) \0.01
0.52 \0.01
Chemotherapy [n (%)]c
84 (86)
64 (100)
Oxaliplatin and/or irinotecan [n (%)]
78 (80)
53 (83)
38 (92) 34 (83)
Bevacizumab [n (%)]
59 (60)
25 (38)
25 (61)
Cetuximab or panitumumab [n (%)]
9 (9)
2 (3)
8 (18)
3 (1–14)
Lung metastases Number [median (range)]
3 (1–14)
2 (1–11)
0.02
0.32
Diameter of the largest nodule [cm; median (range)] 0.5 (0.2–1.7)
0.9 (0.2–3.5)
13 (3–50)
\0.01
\0.01
Bilateral distribution [n (%)]
41 (65)
8 (20)
\0.01
0.63
59 (60)
Bold values are statistically significant (p \ 0.05) CEA carcinoembryonic antigen a Disease-free interval between diagnosis of primary tumor and diagnosis of liver metastases b
The values show the preoperative level in the ‘‘Only Liver Metastases Resected’’ and ‘‘Liver and Lung Metastases Resected’’ groups, and the prechemotherapy level in the ‘‘Chemotherapy Only’’ group
c
Preoperative chemotherapy in the ‘‘Only Liver Metastases Resected’’ and ‘‘Liver and Lung Metastases Resected’’ groups
Mann–Whitney U test. OS was calculated using the Kaplan–Meier method. OS was calculated from the date of hepatectomy in the study population, from the date of diagnosis of liver and lung metastases in the patients treated with chemotherapy only, and from the data of lung resection in the patients who underwent both liver and lung resection. The log-rank test was used to compare OS. To identify prognostic factors in the study population, the clinicopathological variables were evaluated in a univariate proportional hazards model. All variables associated with survival with p B 0.1 in the univariate analysis were subsequently entered into a Cox multivariate regression model with backward elimination. p values less than 0.05 were considered statistically significant. Statistical analyses were performed using the IBM SPSS Statistics software, version 19 (IBM Corporation, Armonk, NY, USA).
RESULTS Baseline Characteristics Baseline characteristics of the study population (n = 98), patients treated with chemotherapy only (n = 64), and patients who underwent curative resection of both liver and lung metastases (n = 41) are summarized in Table 1. Of the 98 patients in the study population, 44 (45 %) had KRAS mutation detected. Perioperative Course of the Study Population In the study population, 84 patients (86 %) underwent preoperative chemotherapy before hepatectomy. Among these patients, lung tumor progression was observed during preoperative chemotherapy in 16 patients (19 %).
Y. Mise et al. FIG. 1 Comparison of overall survival among patients with synchronous colorectal liver and lung metastases
Liver resecon only n = 98 p
ORIGINAL ARTICLE – HEPATOBILIARY TUMORS
Is Complete Liver Resection Without Resection of Synchronous Lung Metastases Justified? Yoshihiro Mise, MD1, Scott Kopetz, MD2, Reza J. Mehran, MD3, Thomas A. Aloia, MD1, Claudius Conrad, MD1, Kristoffer W. Brudvik, MD1, Melissa W. Taggart, MD4, and Jean-Nicolas Vauthey, MD1 Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; 2Department of Gastrointestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; 3Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX; 4Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX 1
ABSTRACT Background. Advances in multidisciplinary care are changing the prognostic impact of colorectal lung metastases. Resection of colorectal liver metastases (CLM) may benefit patients with synchronous lung metastases even when lung metastases are not resected. The aim of this study was to investigate the survival of patients undergoing complete resection of CLM in the setting of unresected lung metastases. Patients and Methods. We compared survival among 98 patients who underwent resection of CLM with unresected lung metastases, 64 who received only chemotherapy for limited colorectal liver and lung metastases, and 41 who underwent resection of both liver and lung metastases. Prognostic factors were investigated in the patients who underwent resection of CLM only. Results. The 3-year/5-year overall survival (OS) rates of patients with CLM resection only (42.9 %/13.1 %) were better than those of patients treated with chemotherapy only (14.1 %/1.6 %; p \ 0.01) but worse than those of patients with resection of liver and lung metastases (68.9 %/56.9 %; p \ 0.01). Multivariate analysis of patients with CLM resection only revealed that KRAS mutation [hazard ratio (HR) 2.10; 95 % confidence interval (CI) 1.21–3.64; p \ 0.01] and rectal primary tumor (HR
Electronic supplementary material The online version of this article (doi:10.1245/s10434-014-4207-3) contains supplementary material, which is available to authorized users. Ó Society of Surgical Oncology 2014 First Received: 25 August 2014 J.-N. Vauthey, MD e-mail: [email protected]
1.72; 95 % CI 1.02–2.88; p = 0.04) were independent predictors of worse OS. Survival of patients without these risk factors was similar to that of patients with curative metastasectomy. Conclusions. Complete resection of metastases remains the primary goal of treatment for stage IV colorectal cancer. Resection of CLM without resection of lung metastases is associated with an intermediate survival between that of patients treated with palliative and curative intent and should be considered in selected patients.
Treatment for stage IV colorectal cancer has evolved during the past 2 decades. The presence of colorectal liver metastases (CLM) once precluded surgical treatment, resulting in a median survival of 12 months before 1990.1 However, hepatectomy is now performed in 25 % of patients with CLM and results in a 5-year overall survival (OS) rate of 58 %.2–5 Unfortunately, for the remaining 75 % of patients, medical therapy alone results in a median survival of 24 months or less.6 Thus, a current issue in the treatment of stage IV colorectal cancer is how to refine a multidisciplinary approach to rescue the patients for whom hepatectomy is not indicated. The key to expanding incorporation of hepatectomy lies in clarifying the impact of lung metastases, which are the most frequent form of extrahepatic metastasis of colorectal cancer.7 Recent series indicate that in the era of effective chemotherapy, long-term disease control can be expected in patients with unresected colorectal lung metastases, which are typically small volume.8–11 Mitry et al. reported a 3-year OS rate of 15.3 % in patients with colorectal lung metastases treated with chemotherapy alone from 1976 to 2005,5,9 whereas the 3-year OS has more than doubled
Y. Mise et al.
(35.5 %) in a series from 2000 to 2010 reported by Kim et al.10 These results suggest the hypothesis that resection of CLM, which has a significant impact on prognosis in patients with stage IV colorectal cancer,12–15 may benefit some patients with unresected lung metastases. Somatic gene mutation status as a tool for predicting tumor biology has become a focus of colorectal cancer research.16 Growing evidence demonstrates a propensity for KRAS-mutated tumors to metastasize to the lung; in patients with KRAS-mutant tumors, rates of KRAS mutation in lung metastases are high, as are rates of early recurrence in the lung after liver surgery for CLM.17–19 These findings suggest that KRAS mutation has the potential to predict tumor biology in patients with colorectal lung metastases. During the last decade, we conducted aggressive surgery for CLM in patients with synchronous lung lesions after multidisciplinary review.20,21 In this study, we identified patients whose lung lesions progressed after hepatectomy and were never resected. Their long-term survival was analyzed to investigate the prognostic impact of hepatectomy in patients with unresected lung metastases. We also evaluated prognostic factors, including KRAS mutation status, to identify patients who benefit from hepatectomy for CLM even in the setting of unresected lung metastases. METHODS The Institutional Review Board of The University of Texas MD Anderson Cancer Center approved this study protocol. A prospectively maintained database of the Department of Surgical Oncology was reviewed for 1,539 patients who underwent hepatectomy for CLM during the period from January 2000 through June 2012. Among these patients, we identified 376 patients (24.4 %) who had synchronous lung nodules when undergoing hepatectomy for CLM. Forty-four patients underwent pulmonary resection and one underwent Cyberknife treatment. In 221 of the remaining 331 patients, the synchronous lung nodules did not progress after hepatectomy. In the remaining 110 patients, the diagnosis of colorectal lung metastases was confirmed by progression, defined as a radiographic increase in the size or number of lung lesions after hepatectomy. Three patients who did not undergo planned second-stage hepatectomy were excluded from the study, leaving 107 patients. Of these 107 patients, 9 were excluded because they did not have available paraffin blocks or had insufficient DNA for genetic testing of KRAS mutation. The remaining 98 patients constituted the study population (patient selection was demonstrated in electronic supplementary Fig. S1). We reviewed the prospectively-collected database of the Department of Medical Oncology and identified 205 patients who had both liver and lung colorectal metastases
and who received systemic chemotherapy alone. From these patients, we selected patients who started chemotherapy at our institution, who had no more than 15 liver metastases and no more than 15 lung metastases on pretreatment images, and who had favorable performance status (Eastern Cooperative Oncology Group performance status 0–2). A total of 64 patients met these criteria. Finally, we selected a comparison group of patients who underwent resection of both types of metastases. Among the aforementioned 44 patients who underwent both liver and lung resection for pathologically confirmed metastases, 3 patients who did not complete planned curative metastasectomy were excluded, leaving 41 patients for analysis. Perioperative Management Before operation, all patients underwent imaging studies, including thin-slice (2.5 mm) helical computed tomography of the chest. Hepatectomy was scheduled after a short course of chemotherapy (four to six cycles). Hepatectomy was considered for patients with CLM when complete resection of CLM was feasible and limited lung metastases seemed amenable to subsequent lung resection or to long-term control with chemotherapy. Resection of lung metastases was performed as previously reported.20,21 Postoperative complications were monitored within 90 days after surgery and graded according to a standard classification.22 Major complications were defined as grade [ II. After surgery, patients were reassessed by physical, laboratory, and imaging evaluation every 3 months. KRAS Mutation Profiling KRAS mutation status was analyzed by examining codons 12, 13, and 61 of KRAS in the study population. DNA extracted from formalin-fixed, paraffin-embedded tissue blocks was quantified and analyzed with a polymerase chain reactionbased DNA sequencing method and Sequenom MassARRAY technology (Sequenom, Inc., San Diego, CA, USA).23 Long-Term Outcomes After Liver Resection The OS of the study population was compared with the OS of patients treated with chemotherapy alone and patients who underwent resection of synchronous liver and lung metastases. Predictors of OS were investigated in the study population. Statistical Analysis Categorical variables were compared using the Fisher’s exact test. Continuous variables were compared using the
Hepatectomy Without Resecting Lung Metastases TABLE 1 Baseline characteristics of patients with synchronous colorectal liver and lung metastases who underwent resection of liver metastases only, chemotherapy only, and resection of both liver and lung metastases Characteristic
Only liver Chemotherapy metastases only (n = 64) resected (n = 98)
Liver and lung p value p value for only liver metastases metastases resected resected vs. chemotherapy (n = 41) only
Age [years; median (range)]
55 (22–84)
59 (30–73)
57 (33–73)
0.26
0.11
Male [n (%)]
60 (61)
39 (61)
26 (63)
0.96
1.00
42 (43)
26 (41)
8 (20)
0.03
0.81
Positive lymph nodes
64 (65)
31 (66)
29 (71)
0.82
1.00
KRAS mutation [n (%)]
44 (45)
Primary disease [n (%)] Rectal primary tumor
Liver metastases Disease-free interval \1 year [n (%)]a
75 (77)
53 (83)
20 (49)
\0.01
0.43
Number [median (range)]
2 (1–15)
3 (1–14)
1 (1–9)
\0.01
0.03
Diameter of the largest nodule [cm; median (range)] 3.0 (0.5–14.0) CEA level [ng/mL; median (range)]b 4.4 (1–723)
3.4 (0.4–16.7) 2.0 (0.5–10.0) 0.02 23.7 (1.2–5,302) 2.4 (0.7–383) \0.01
0.52 \0.01
Chemotherapy [n (%)]c
84 (86)
64 (100)
Oxaliplatin and/or irinotecan [n (%)]
78 (80)
53 (83)
38 (92) 34 (83)
Bevacizumab [n (%)]
59 (60)
25 (38)
25 (61)
Cetuximab or panitumumab [n (%)]
9 (9)
2 (3)
8 (18)
3 (1–14)
Lung metastases Number [median (range)]
3 (1–14)
2 (1–11)
0.02
0.32
Diameter of the largest nodule [cm; median (range)] 0.5 (0.2–1.7)
0.9 (0.2–3.5)
13 (3–50)
\0.01
\0.01
Bilateral distribution [n (%)]
41 (65)
8 (20)
\0.01
0.63
59 (60)
Bold values are statistically significant (p \ 0.05) CEA carcinoembryonic antigen a Disease-free interval between diagnosis of primary tumor and diagnosis of liver metastases b
The values show the preoperative level in the ‘‘Only Liver Metastases Resected’’ and ‘‘Liver and Lung Metastases Resected’’ groups, and the prechemotherapy level in the ‘‘Chemotherapy Only’’ group
c
Preoperative chemotherapy in the ‘‘Only Liver Metastases Resected’’ and ‘‘Liver and Lung Metastases Resected’’ groups
Mann–Whitney U test. OS was calculated using the Kaplan–Meier method. OS was calculated from the date of hepatectomy in the study population, from the date of diagnosis of liver and lung metastases in the patients treated with chemotherapy only, and from the data of lung resection in the patients who underwent both liver and lung resection. The log-rank test was used to compare OS. To identify prognostic factors in the study population, the clinicopathological variables were evaluated in a univariate proportional hazards model. All variables associated with survival with p B 0.1 in the univariate analysis were subsequently entered into a Cox multivariate regression model with backward elimination. p values less than 0.05 were considered statistically significant. Statistical analyses were performed using the IBM SPSS Statistics software, version 19 (IBM Corporation, Armonk, NY, USA).
RESULTS Baseline Characteristics Baseline characteristics of the study population (n = 98), patients treated with chemotherapy only (n = 64), and patients who underwent curative resection of both liver and lung metastases (n = 41) are summarized in Table 1. Of the 98 patients in the study population, 44 (45 %) had KRAS mutation detected. Perioperative Course of the Study Population In the study population, 84 patients (86 %) underwent preoperative chemotherapy before hepatectomy. Among these patients, lung tumor progression was observed during preoperative chemotherapy in 16 patients (19 %).
Y. Mise et al. FIG. 1 Comparison of overall survival among patients with synchronous colorectal liver and lung metastases
Liver resecon only n = 98 p
