- Email: [email protected]
Heterogeneity of Tumor Sizes in Multiple Pulmonary Metastases of Colorectal Cancer as a Prognostic Factor
Heterogeneity of Tumor Sizes in Multiple Pulmonary Metastases of Colorectal Cancer as a Prognostic Factor Tomohiro Maniwa, MD, Keita Mori, PhD, Yasuhisa Ohde, MD, Takehiro Okumura, MD, Narikazu Boku, MD, PhD, Tomoyuki Hishida, MD, Yukinori Sakao, MD, Katsuo Yoshiya, MD, Ichinosuke Hyodo, MD, PhD, and Haruhiko Kondo, MD Department of Thoracic Surgery, Yao Municipal Hospital, Osaka; Division of Thoracic Surgery and Clinical Trial Coordination Office, Shizuoka Cancer Center Hospital, Shizuoka; Department of Surgery, University Hospital Mizonokuchi, Teikyo University School of Medicine, Tokyo; Division of Gastrointestinal Medical Oncology, National Cancer Center Hospital, Tokyo; Department of Thoracic Surgery, National Cancer Center Hospital East, Chiba; Department of Thoracic Surgery, Aichi Cancer Center Hospital, Aichi; Department of Chest Surgery, Niigata Cancer Center Hospital, Niigata; Division of Gastroenterology Clinical Medicine, Faculty of Medicine, University of Tsukuba University Hospital, Ibaraki; and General Thoracic Surgery, School of Medicine, Kyorin University, Tokyo, Japan
Background. The number of metastatic lesions is closely correlated with prognosis in most cancers. The aim of this study was to clarify the relationship between individual heterogeneity of metastatic tumor sizes and prognosis in patients with multiple pulmonary metastasis of colorectal cancer who received surgical treatment. Methods. Clinical data for patients who had pulmonary metastasis from colorectal cancer and underwent curative resection at 46 Japanese institutions between January 2004 and December 2008 were collected. Among 898 patients eligible considering these inclusion criteria, 247 patients had multiple metastases and were analyzed. A difference between the maximum and minimum tumor diameters (Dmax-min) on pathologic findings was used to evaluate size heterogeneity. Results. The overall survival rate was 75% at 3 years and 58% at 5 years, with a median follow-up period of 65
months (range, 0 to 112). When Dmax-min of 5 mm was set as a cutoff value, overall survival was significantly different between small (£5 mm, n [ 95) and large (>5 mm, n [ 152) tumor groups (5-year survival rates, 66.5% and 53.3%, respectively; log rank test, p [ 0.025). Multivariate analysis using a Cox proportional hazards model revealed that disease-free interval from resection of primary lesion, serum carcinoembryonic antigen level, number of pulmonary metastases, and Dmax-min were independent prognostic factors. Conclusions. The heterogeneity of metastatic tumor sizes may be an indicator for prognosis in patients with multiple pulmonary metastases of colorectal cancer who underwent resection.
M
Multiple PM of similar and differing sizes can exist. Differences in tumor size may reveal features of tumor biology. Namely, the cancer cells of PM with narrow ranges of sizes may arrive at the lungs and grow for only one period. In contrast, the cancer cells of PM with wide ranges of sizes may arrive at the lungs and grow for various periods. Here, we report the findings of a retrospective multicenter study of pulmonary metastasectomy for PM from colorectal cancer that was conducted after the
any studies have examined the resection of pulmonary metastases (PM) from colorectal cancer [1–13]. These studies have generally suggested that resection of PM can be beneficial to the patient’s prognosis. In these studies, 5-year rates of overall survival have ranged from 38% to 68%. Some of these studies have used large multicenter retrospective designs [1, 3], and others have been systematic reviews [4]; however, few studies have focused specifically on multiple PM. It has been reported that multiple PM indicate a poorer prognosis than a solitary PM [1, 2, 4–7]. Recently, new drugs [14, 15] and targeted therapies [16, 17] have been developed for colorectal cancer. Consequently, the surgical criteria for multiple PM are broadening.
Accepted for publication July 28, 2016. Address correspondence to Dr Maniwa, Department of Thoracic Surgery, Yao Municipal Hospital, 1-3-1 Ryuge-cho, Yao City, Osaka 581-0069, Japan; email: [email protected].
Ó 2016 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2016;-:-–-) Ó 2016 by The Society of Thoracic Surgeons
Dr Boku discloses a financial relationship with Taiho.
The Supplemental Table can be viewed in the online version of this article [http://doi.dx.org/10.1016/ j.athoracsur.2016.07.070] on http://www.annalsthoracic surgery.org.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2016.07.070
2
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
introduction of new active chemotherapeutic regimens [14–17]. The immediate aim of this study was to clarify the relationship between the individual heterogeneity of metastatic tumor sizes and prognosis. The ultimate objective of this analysis was to gain new evidence that could be relevant to the surgical criteria for patients who have colorectal cancer with PM.
Patients and Methods Patients The study was a large cohort trial by the Tsukuba Cancer Clinical Group based in numerous institutions in Japan. The data in our analysis were originally obtained for all patients who had PM from colorectal cancer and underwent pulmonary resection between January 2004 and December 2008 at 46 institutions throughout Japan. The inclusion criteria for this study were (1) first metastasectomy for PM; (2) metastasectomy performed with curative intent; (3) pathologic diagnosis of the resected lesions as metastases from colorectal cancer; (4) metastasectomy with pathologically confirmed complete resection; and (5) curative resection of the primary cancer. Fig 1. Study design and participants. (CONSORT ¼ Consolidated Standards of Reporting Trials; CRC ¼ colorectal cancer; PM ¼ pulmonary metastases; R0 ¼ resection for cure or complete remission.)
Ann Thorac Surg 2016;-:-–-
Patients with metastasis from cancer of the appendix or anal cancer were excluded. This study was a multicenter retrospective study and operative indications for pulmonary metastasectomy, in detail, were referred to each institution. However, indications were based on the following: metastatic lesions are confined to the lungs; all lesions are resectable; there are no extrathoracic lesions except for liver metastasis or extralesions are clinically controlled by treatment; and the surgical procedure is tolerable for the patient. Whether pulmonary resection was performed through thoracotomy or video-assisted thoracic surgery was decided according to the situation of each case and the judgment of the physician of each institution. Lymph node dissection combined with pulmonary resection was not regulated, according to the judgment of the physician of each institution. Follow-up was not uniform in this study. However, most of the patients were treated according to the treatment guideline for colorectal cancer in Japan [18]. The study protocol was approved by the Institutional Review Board of all participating hospitals, including that of the Ethics Committee, Kyorin University Graduate School of Medicine (H24-051). We conducted a large-scale multiinstitutional retrospective cohort study to inform the
Ann Thorac Surg 2016;-:-–-
decision-making process for these patients. Data were initially obtained on 1,237 consecutive patients who had PM from colorectal cancer and underwent first pulmonary resection at any of the 46 institutions. After reviewing all the data for eligibility, completeness, and consistency, 339 cases were excluded. The remaining 898 cases of PM of colorectal cancer were studied in a previous related study [19]. Among these 898 cases, 651 cases were solitary PM and 247 cases were multiple PM. These 247 cases were analyzed in the present study (Fig 1). We examined differences in tumor sizes in two steps, as follows.
Step 1, Individual Heterogeneity of Metastatic Tumor Size Patients in the study cohort had multiple PM of both similar and different sizes. Given the variation in tumor sizes within individual patients, we decided to investigate the relationship between prognosis and the individual heterogeneity of tumor sizes within each patient. To obtain the individual heterogeneity of tumors sizes, we investigated the difference between the maximum and minimum tumor diameters (Dmax-min) on pathologic findings (Supplemental Table). We classified patients into three groups: Dmax-min 0 mm to 5 mm; Dmax-min 6 mm to 10 mm, and Dmax-min 11 mm or more. We examined the relationship between individual heterogeneity of tumor sizes and risk ratios using a Cox proportional hazards model.
Step 2, Prognostic Factors of Colorectal Cancer PM and Multivariate Analyses We also analyzed prognostic factors of colorectal cancer PM. The patients’ medical records were reviewed for the following information: age; sex; location of the primary cancer; disease-free interval (DFI); synchronicity with the primary cancer; site of PM; prethoracotomy level of carcinoembryonic antigen (CEA); surgical procedure; type of lymph node dissection; pathologic maximum diameter of PM; pathologic number of PM; pathologic pulmonary lymph node metastasis; chemotherapy before discovery of PM; preoperative chemotherapy before the discovery of PM; pathologic evaluation of preoperative chemotherapy; adjuvant chemotherapy after pulmonary resection; and heterogeneity of tumor sizes. Additionally, we performed multivariate analyses of the prognostic factors of colorectal cancer PM.
Statistical Analysis The DFI was defined as the time between the date of resection of the primary tumor and the date at which PM was first detected. The DFI was set to zero for patients with a primary cancer that was found after resection of a metastatic lesion. Overall survival was defined as the time between the date of pulmonary metastasectomy and the date of death. To identify prognostic factors after pulmonary metastasectomy, survival curves and survival rates (ie, overall survival rates at 5 years) were estimated using the Kaplan-Meier method. Log rank tests were used to compare survival across the levels of categoric variables.
3
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Table 1. Patient Characteristics in Overall Study Cohort All Patients (n ¼ 247)
Characteristics Sex Male Female Age, years Synchronicity with primary cancer Synchronous Metachronous Disease-free interval, years <2 2 Location of primary tumors Colon Rectum Extrathoracic lesions Absent Present Site of PM Unilateral Bilateral Carcinoembryonic antigen p-Max diameter of PM Number of PM Mode of surgery Less than lobectomya Lobectomy or moreb p-LN metastasis Absent Present Not available CT before finding PM Yes No Not available Preoperative CT after finding PM Yes No Adjuvant CT after PM resection Yes No a Wedge resection or segmentectomy. pneumonectomy.
142 105 63 (20–85) 40 207 170 77 103 144 187 60 133 114 3.9 (0.4–140.5) 16 (5–80) 2 (2–10) 57 190 136 11 100 136 101 10 56 191 124 123 b
Lobectomy, bilobectomy,
Values are n or median (range). CT ¼ chemotherapy; pulmonary metastases;
p-LN ¼ pathologic lymph node; p-Max ¼ pathologic maximum.
PM ¼
To counteract the arbitrariness of choosing a threshold and dichotomizing, the effects of continuous variables on survival were also evaluated using Cox proportional hazards models. In multivariate analyses of categoric variables, we used Cox proportional hazards models with the stepwise selection method. All p values less than 0.05 were considered statistically significant. The statistical analyses were conducted using JMP software, version 9.0 (SAS Institute, Cary, NC).
4
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Fig 2. Overall survival rate of all patients with multiple metastases. Overall survival was 75.4% at 3 years and 58.4% at 5 years.
Results The median follow-up period was 65.1 months (range, 0 to 111.8). Table 1 shows the characteristics of the overall patient cohort. Figure 2 presents the overall survival of all patients with multiple PM. The 3-year and 5-year overall survival rates were 75.4% and 58.4%, respectively. Figure 3 shows the heterogeneity of tumor sizes and risk ratios. The Dmax-min 6 mm to 10 mm group and the Dmax-min 11 mm or more group had a trend toward poorer survival than the Dmax-min 0 mm to 5 mm group, with hazard ratios of 1.43 and 1.79, respectively. The hazard ratio increased with increasing Dmax-min. Table 2 shows the results of univariate analyses of other risk factors in patients with multiple PM. Significant prognostic factors for overall survival included DFI, CEA, number of PM, use of preoperative chemotherapy after discovery of PM, and Dmax-min even though the maximum diameter of PM was not significantly prognostic. In the present study, exploration of lymph nodes was not routinely performed combined with pulmonary metastasectomy, but sampling or systematic mediastinal lymph node dissection was performed in 59.3% of
Ann Thorac Surg 2016;-:-–-
patients (147 of 247). The incidence of lymph node metastasis confirmed pathologically was only 7.5% (11 of 147), and the lymph node metastasis was not a significant prognostic factor in our analysis. There was no significant difference between synchronous disease and metachronous metastases, although there was significant difference in DFI. Multivariate analysis using a Cox proportional hazards model revealed that DFI, CEA, number of PM, Dmax-min and were significant prognostic factors, even though preoperative chemotherapy after the discovery of PM was not significantly prognostic (Table 3). Accordingly, we classified a 5 mm or less difference of tumor sizes as “small” and a greater than 5 mm difference of tumor sizes as “large.” Ninety-five patients had small (5 mm or less) ranges of tumor sizes (group A) and 151 patients had large (greater than 5 mm) ranges of tumor sizes (group B). Figure 4 shows the overall survival rates of patients in group A (n ¼ 95) and group B (n ¼ 152). The 5-year overall survival rates in group A and group B were 65.5% and 53.2%, respectively. There was a significant difference in overall survival between the two groups (p ¼ 0.025).
Comment Although many studies have investigated pulmonary resection for PM of colorectal cancer [1–13], few have specifically investigated patients with multiple PM. Previous reports noted 5-year overall survival rates of 15.7% to 45.0% among patients with multiple PM from colorectal cancer [1, 9, 10]; this prognosis is poor. The 5-year overall survival rate was 58.4% in the present study. Survival was favorable in the present study as compared with that observed in previous studies [1, 9, 10]. The major explanation of this difference is considered due to the period of this trial. Compared with reports from the 1990s, the more selective group of patients with PM in this study had undergone pulmonary resection, owing to the emergence of helical computed tomography and positron
Fig 3. Relationship between the heterogeneity of tumor sizes and risk ratios. The difference between the maximum and minimum tumor diameters (Dmax-min) of the 6 mm to 10 mm group and the Dmax-min 11 mm or more group tended to have poorer survival than the Dmax-min 0 mm to 5 mm group, with hazard ratios of 1.43 and 1.79, respectively. (CI ¼ confidence interval.)
Ann Thorac Surg 2016;-:-–-
5
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Table 2. Univariate Analyses of Other Risk Factors in Patients With More Than One Primary Metastasis
Table 3. Cox Proportional Hazards Regression Analysis for Survival
Variables
Variables
RR
95% CI
p Value
Dmax-min, >5 mm versus 5 mm PM, 5 versus <5 CEA, high versus normal DFI, <2 years versus 2 years Preoperative CT after PM, yes versus no
1.55 2.15 1.61 1.73 1.23
1.02–2.42 1.23–3.57 1.07–2.43 1.07–2.88 0.76–1.93
0.042 0.0082 0.022 0.023 0.38
Sex Male Female Age, years >65 65 Synchronicity with primary cancer Synchronous Metachronous Location of primary tumors Colon Rectum Disease-free interval, years <2 2 Extrathoracic lesion Absent Present Site of PM Unilateral Bilateral Carcinoembryonic antigen Normal High p-Max diameter of PM 2 cm >2 cm Number of PM 2–4 5 Mode of surgery Less than lobectomya Lobectomy or moreb p-LN metastasis Absent Present Chemotherapy before finding PM Yes No Preoperative CT after PM Yes No Adjuvant CT after PM resection Yes No Dmax-min 5 mm >5 mm a
n
5-Year OS (%) p Value 0.46
142 105
58.6 57.9
139 108
56.0 61.4
40 207
64.9 57.1
103 144
59.6 57.6
170 77
51.9 72.0
77 170
58.8 57.0
133 114
63.5 53.0
139 106
66.9 45.7
167 80
59.7 55.3
222 25
61.6 33.9
190 57
56.7 64.1
136 11
59.8 66.7
136 101
55.5 59.8
56 191
43.8 62.3
124 123
59.1 57.6
95 152
66.5 53.2
0.39
0.43
0.42
0.011
0.36
0.074
0.0032
0.29
0.0025
0.30
0.76
0.27
CEA ¼ carcinoembryonic antigen; CI ¼ confidence interval; CT ¼ chemotherapy; DFI ¼ disease-free interval; Dmax-min ¼ difference maximum-minimum diameter; PM ¼ pulmonary metastases; RR ¼ risk ratio.
emission tomography. These new diagnostic modalities could exclude the patients with unresectable metastasis invisible by the old modality. Indeed, patients who underwent pulmonary resection in this study had less lymph node metastasis and smaller metastatic tumor sizes than that observed in previous studies [1, 4–6, 11, 13]. Furthermore, the median overall survival of patients with unresectable colon cancer was 5 to 8 months in previous reports; however, the emergence of new drugs, such as oxaliplatin and irinotecan, increased that to 20 months [14, 15]. Moreover, targeted therapy led to a better prognosis [16, 17]. Therefore, the prognosis even after recurrence seemed longer in this study than in the previous reports. When considering the hematogenous metastatic pathway from colorectal cancer, tumor cells reach the lungs through the liver through the portal vein or directly through the inferior vena cava. In either case, cancer cells that migrate into the blood necessarily arrive at the lungs and then spread through the body. However, if cancer cells with metastatic ability are trapped within the lungs (as a filter organ), then metastases to the lungs may be regarded as semilocal disease, despite the distant metastasis from colorectal cancer. The hypothesis that the lungs act as a filter for cancer cells has been suggested as an anatomic and mechanical function, based on experimental animal studies and autopsy cases [20, 21].
0.032
0.45
0.025
Wedge resection or segmentectomy. pneumonectomy.
b
Lobectomy, bilobectomy,
CT ¼ chemotherapy; Dmax-min ¼ difference maximum-minimum diameter; OS ¼ overall survival; p-LN ¼ pathologic lymph node; PM ¼ pulmonary metastases; p-Max ¼ pathologic maximum.
Fig 4. Overall survival rates of patients in group A (n ¼ 95 [dashed line]) and group B (n ¼ 151 [solid line]). The 5-year overall survival rates were 66.5% for group A and 53.0% for group B. Overall survival differed significantly between the groups (p ¼ 0.025).
6
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Multiple PM of similar and various sizes exist. Based on this hypothesis of a cascade, we assume in the cases with small differences of PM sizes, migration and attachment of circulating cancer cells to the pulmonary capillary wall and its extravasation at each site occur in almost the same period. This period may be the time of surgery for the primary lesion or may reflect some condition of the patients. In this case, cancer cells with metastatic ability may stay in the lungs. Therefore, PM with a small difference of tumor sizes may be a form of less aggressive disease, just as a solitary metastasis. In contrast, we believe that in the cases with PM with a large difference of tumor sizes, cancer cells arrive at the lungs and grow for various periods. It is possible that cancer cells migrate into the blood and circulate at various periods, or do so constantly. Therefore, PM with a large difference of tumor sizes may be a form of systemic disease. We thought that the patients with a larger range of tumors size would have a poorer prognosis; however, there was no significant difference in larger differences by multivariate analysis. We think that the reason was confounding factors such as size or number of multiple tumors. Moreover, the number of patients with larger differences was smaller. Consequently, there was a significant difference in 5-mm difference of tumor sizes by multivariate analysis, and we applied “5 mm” to the cutoff value in the present study. It has been unclear whether lymphadenectomy with metastasectomy results in better survival. A poor prognosis has been observed for patients with hilar and mediastinal lymph node metastasis identified in PM during or after surgery; the 5-year overall survival rate is 6.2% to 23% [11–13]. Therefore, most patients with clinical lymph node metastasis do not undergo pulmonary resection. In the present study, the lymph nodes of 100 patients (40.5%) were not evaluated pathologically. And only 11 of 147 patients confirmed pathologically had lymph node metastasis. This small number reduces the statistical power showing the significance of lymph node metastasis as a prognostic factor. In group A, lymph node metastasis developed in only 1 of 38 patients (2.6%) who underwent more than lymph sampling. Conversely, in group B, 10 of 109 patients (9.2%) had lymph node metastasis. This result may also show that multiple PM with small differences of tumor sizes could be interpreted as less aggressive disease. An increase in the number of multiple metastases is a negative predictor of survival [5, 6]. Twenty-five patients had more than five PM. The 5-year overall survival rate of patients with more than 5 PM was 33.9%. Only 1 patient in group A had more than five PM. The number of metastases and the difference of tumor sizes were both significant and independent prognostic factors, although the difference of tumor sizes was associated with the number of metastases. In the present study, the DFI was defined as the time between the date of resection of the primary tumor and the date at which PM was first detected. There was no significant difference between synchronous disease and metachronous metastases, although there was significant
Ann Thorac Surg 2016;-:-–-
difference in DFI. Consequently, we did not separate the synchronous disease from the metachronous metastases. In the present study, we identified the following prognostic factors for patients with multiple PM: number of metastases, CEA, DFI, and range of tumor sizes. Although univariate analysis showed a significant difference in overall survival for preoperative chemotherapy, this difference did not remain significant in the multivariate analysis. There are two reasons for these findings. First, patients who had the highest risk factors for recurrence—such as many PM, elevated CEA, and short DFI—received chemotherapy before pulmonary resection. Second, there were conversion cases, in which patients could not initially undergo pulmonary resection because of oncologic or patient-related factors, but subsequently received chemotherapy and showed downstaging or had tumor regression because of good response to chemotherapy. Consequently, these patients were ultimately able to undergo pulmonary resection. In either case, there was no significant difference in preoperative chemotherapy. This study has several limitations that should be considered as its results are interpreted. First, this was a retrospective multicenter cohort study; pulmonary resection was performed according to the judgment of each individual institute, following criteria that were based on protocols. Second, this is a study of surgical cases only; accordingly, the patients included in the analysis were highly selected and might not be representative of all patients with PM from colorectal cancer. Nonetheless, there have been no reports of multicenter, large studies that are specific to multiple PM from colorectal cancer. Improvements in chemotherapy have been reported, such as the development of new drugs and targeted therapy for colorectal cancer. We believe that patients who have multiple PM with a small difference in tumor sizes should undergo initial pulmonary resection— just as they have to date—because their disease may be less aggressive. However, we think that it may be better for patients with multiple PM and a large difference in tumor sizes to undergo induction chemotherapy or follow a wait-and-see approach for some period after pulmonary resection, because their disease may be systemic. In conclusion, the range of tumor sizes may be a good indicator when judging the appropriateness of surgical treatment for multiple PM in patients with colorectal cancer. The authors wish to thank all the patients and physicians at the institutions participating in the Tsukuba Cancer Trial Group for contributing to this series: Osaka Medical Center for Cancer and Cardiovascular Disease (Osaka, Japan); Kurashiki Central Hospital (Okayama, Japan); Hokkaido Cancer Center (Hokkaido, Japan); Kitasato University School of Medicine (Tokyo, Japan); Tokyo Women’s Medical University (Tokyo, Japan); Hyogo Cancer Center (Hyogo, Japan); Hiroshima City Hospital (Hiroshima, Japan); Toho University School of Medicine (Tokyo, Japan); Kyoto University Graduate School of Medicine (Kyoto, Japan); Shinshu University School of Medicine, (Nagano, Japan); Nagoya City University Graduate School of Medical Sciences (Nagoya, Japan); Tokyo Metropolitan Cancer and Infectious Diseases Center, Komagome Hospital (Tokyo, Japan); National
Ann Thorac Surg 2016;-:-–-
Kyushu Cancer Center (Fukuoka, Japan); Shikoku Cancer Center (Ehime, Japan); Niigata University Graduate School of Medical and Dental Sciences (Niigata, Japan); Gunma Prefectural Cancer Center (Gunma, Japan); Osaka University Graduate School of Medicine (Osaka, Japan); Hamamatsu University School of Medicine (Shizuoka, Japan); Yamagata Prefectural Central Hospital (Yamagata, Japan); Jikei University School of Medicine (Tokyo, Japan); Kumamoto University Hospital (Kumamoto, Japan); Fukujuji Hospital, Japan Anti-Tuberculosis Association (Tokyo, Japan); Saitama Medical Center, Saitama Medical University (Saitama, Japan); Tokai University Hachioji Hospital (Tokyo, Japan); Saitama Cardiovascular and Respiratory Center (Saitama, Japan); Saitama Medical University International Medical Center (Saitama, Japan); Kurume University School of Medicine (Fukuoka, Japan); Toyama Prefectural Central Hospital (Toyama, Japan); University of Tsukuba (Ibaraki, Japan); Mie University School of Medicine (Mie, Japan); St. Marianna University School of Medicine (Tokyo, Japan); Kanagawa Cancer Center (Kanagawa, Japan); Gunma University Graduate School of Medicine (Gunma, Japan); Nishi-Niigata Chuo National Hospital (Niigata, Japan); Fukushima Medical University (Fukushima, Japan); Okayama Saiseikai General Hospital (Okayama, Japan); Ishikawa Prefectural Central Hospital (Ishikawa, Japan); Oita Prefectural Hospital (Oita, Japan); Showa University School of Medicine (Tokyo, Japan); and Kawasaki Medical School (Okayama, Japan).
References 1. Iida T, Nomori H, Shiba M, et al, for the Metastatic Lung Tumor Study Group of Japan. Prognostic factors after pulmonary metastasectomy for colorectal cancer and rationale for determining surgical indications: a retrospective analysis. Ann Surg 2013;257:1059–64. 2. Watanabe K, Nagai K, Kobayashi A, Sugito M, Saito N. Factors influencing survival after complete resection of pulmonary metastases from colorectal cancer. Br J Surg 2009;96: 1058–65. 3. Hirosawa T, Itabashi M, Ohnuki T, Yamaguchi N, Sugihara K, Kameoka S, for the Japanese Society for Cancer of the Colon and Rectum (JSCCR) Study Group for Pulmonary Metastases from Colorectal Cancer. Prognostic factors in patients undergoing complete resection of pulmonary metastases of colorectal cancer: a multi-institutional cumulative follow-up study. Surg Today 2013;43:494–9. 4. Pfannschmidt J, Dienemann H, Hoffmann H. Surgical resection of pulmonary metastases from colorectal cancer: a systematic review of published series. Ann Thorac Surg 2007;84:324–38. 5. Welter S, Jacobs J, Krbek T, Poettgen C, Stamatis G. Prognostic impact of lymph node involvement in pulmonary metastases from colorectal cancer. Eur J Cardiothorac Surg 2007;31:167–72. 6. Kanemitsu Y, Kato T, Hirai T, Yasui K. Preoperative probability model for predicting overall survival after resection of pulmonary metastases from colorectal cancer. Br J Surg 2004;91:112–20.
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
7
7. Watanabe I, Arai T, Ono M, et al. Prognostic factors in resection of pulmonary metastasis from colorectal cancer. Br J Surg 2003;90:1436–40. 8. Onaitis MW, Petersen RP, Haney JC, et al. Prognostic factors for recurrence after pulmonary resection of colorectal cancer metastases. Ann Thorac Surg 2009;87:1684–8. 9. Lin BR, Chang TC, Lee YC, Lee PH, Chang KJ, Liang JT. Pulmonary resection for colorectal cancer metastases: duration between cancer onset and lung metastasis as an important prognostic factor. Ann Surg Oncol 2009;16: 1026–32. 10. Iizasa T, Suzuki M, Yoshida S, et al. Prediction of prognosis and surgical indications for pulmonary metastasectomy from colorectal cancer. Ann Thorac Surg 2006;82:254–60. 11. Okumura S, Kondo H, Tsuboi M, et al. Pulmonary resection for metastatic colorectal cancer: experiences with 159 patients. J Thorac Cardiovasc Surg 1996;112:867–74. 12. Saito Y, Omiya H, Kohno K, et al. Pulmonary metastasectomy for 165 patients with colorectal carcinoma: a prognostic assessment. J Thorac Cardiovasc Surg 2002;124: 1007–13. 13. B€ ol€ ukbas S, Sponholz S, Kudelin N, Eberlein M, Schirren J. Risk factors for lymph node metastases and prognosticators of survival in patients undergoing pulmonary metastasectomy for colorectal cancer. Ann Thorac Surg 2014;97: 1926–32. 14. Grothey A, Sargent D, Goldberg RM, Schmoll HJ. Survival of patients with advanced colorectal cancer improves with the availability of fluorouracil-leucovorin, irinotecan, and oxaliplatin in the course of treatment. J Clin Oncol 2004;22: 1209–14. 15. Tournigand C, Andr e T, Achille E, et al. FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal cancer: a randomized GERCOR study. J Clin Oncol 2004;22: 229–37. 16. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–42. 17. Giantonio BJ, Catalano PJ, Meropol NJ, et al, for the Eastern Cooperative Oncology Group Study E3200. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200. J Clin Oncol 2007;25:1539–44. 18. Watanabe T, Itabashi M, Shimada Y, et al, for the Japanese Society for Cancer of the Colon and Rectum. Japanese Society for Cancer of the Colon and Rectum (JSCCR) guidelines 2014 for treatment of colorectal cancer. Int J Clin Oncol 2015;20:207–39. 19. Hishida T, Okumura T, Boku N, et al. Surgical outcome for pulmonary metastasis of colorectal cancer in the modern chemotherapy era: results of a retrospective Japanese multicenter study. J Clin Oncol 2014;32(Suppl):3528. 20. Weiss L, Grundmann E, Torhorst J, et al. Haematogenous metastatic patterns in colonic carcinoma: an analysis of 1541 necropsies. J Pathol 1986;150:195–203. 21. Chambers AF, Groom AC, MacDonald IC. Dissemination and growth of cancer cells in metastatic sites. Nat Rev Cancer 2002;2:563–72.
Background. The number of metastatic lesions is closely correlated with prognosis in most cancers. The aim of this study was to clarify the relationship between individual heterogeneity of metastatic tumor sizes and prognosis in patients with multiple pulmonary metastasis of colorectal cancer who received surgical treatment. Methods. Clinical data for patients who had pulmonary metastasis from colorectal cancer and underwent curative resection at 46 Japanese institutions between January 2004 and December 2008 were collected. Among 898 patients eligible considering these inclusion criteria, 247 patients had multiple metastases and were analyzed. A difference between the maximum and minimum tumor diameters (Dmax-min) on pathologic findings was used to evaluate size heterogeneity. Results. The overall survival rate was 75% at 3 years and 58% at 5 years, with a median follow-up period of 65
months (range, 0 to 112). When Dmax-min of 5 mm was set as a cutoff value, overall survival was significantly different between small (£5 mm, n [ 95) and large (>5 mm, n [ 152) tumor groups (5-year survival rates, 66.5% and 53.3%, respectively; log rank test, p [ 0.025). Multivariate analysis using a Cox proportional hazards model revealed that disease-free interval from resection of primary lesion, serum carcinoembryonic antigen level, number of pulmonary metastases, and Dmax-min were independent prognostic factors. Conclusions. The heterogeneity of metastatic tumor sizes may be an indicator for prognosis in patients with multiple pulmonary metastases of colorectal cancer who underwent resection.
M
Multiple PM of similar and differing sizes can exist. Differences in tumor size may reveal features of tumor biology. Namely, the cancer cells of PM with narrow ranges of sizes may arrive at the lungs and grow for only one period. In contrast, the cancer cells of PM with wide ranges of sizes may arrive at the lungs and grow for various periods. Here, we report the findings of a retrospective multicenter study of pulmonary metastasectomy for PM from colorectal cancer that was conducted after the
any studies have examined the resection of pulmonary metastases (PM) from colorectal cancer [1–13]. These studies have generally suggested that resection of PM can be beneficial to the patient’s prognosis. In these studies, 5-year rates of overall survival have ranged from 38% to 68%. Some of these studies have used large multicenter retrospective designs [1, 3], and others have been systematic reviews [4]; however, few studies have focused specifically on multiple PM. It has been reported that multiple PM indicate a poorer prognosis than a solitary PM [1, 2, 4–7]. Recently, new drugs [14, 15] and targeted therapies [16, 17] have been developed for colorectal cancer. Consequently, the surgical criteria for multiple PM are broadening.
Accepted for publication July 28, 2016. Address correspondence to Dr Maniwa, Department of Thoracic Surgery, Yao Municipal Hospital, 1-3-1 Ryuge-cho, Yao City, Osaka 581-0069, Japan; email: [email protected].
Ó 2016 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2016;-:-–-) Ó 2016 by The Society of Thoracic Surgeons
Dr Boku discloses a financial relationship with Taiho.
The Supplemental Table can be viewed in the online version of this article [http://doi.dx.org/10.1016/ j.athoracsur.2016.07.070] on http://www.annalsthoracic surgery.org.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2016.07.070
2
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
introduction of new active chemotherapeutic regimens [14–17]. The immediate aim of this study was to clarify the relationship between the individual heterogeneity of metastatic tumor sizes and prognosis. The ultimate objective of this analysis was to gain new evidence that could be relevant to the surgical criteria for patients who have colorectal cancer with PM.
Patients and Methods Patients The study was a large cohort trial by the Tsukuba Cancer Clinical Group based in numerous institutions in Japan. The data in our analysis were originally obtained for all patients who had PM from colorectal cancer and underwent pulmonary resection between January 2004 and December 2008 at 46 institutions throughout Japan. The inclusion criteria for this study were (1) first metastasectomy for PM; (2) metastasectomy performed with curative intent; (3) pathologic diagnosis of the resected lesions as metastases from colorectal cancer; (4) metastasectomy with pathologically confirmed complete resection; and (5) curative resection of the primary cancer. Fig 1. Study design and participants. (CONSORT ¼ Consolidated Standards of Reporting Trials; CRC ¼ colorectal cancer; PM ¼ pulmonary metastases; R0 ¼ resection for cure or complete remission.)
Ann Thorac Surg 2016;-:-–-
Patients with metastasis from cancer of the appendix or anal cancer were excluded. This study was a multicenter retrospective study and operative indications for pulmonary metastasectomy, in detail, were referred to each institution. However, indications were based on the following: metastatic lesions are confined to the lungs; all lesions are resectable; there are no extrathoracic lesions except for liver metastasis or extralesions are clinically controlled by treatment; and the surgical procedure is tolerable for the patient. Whether pulmonary resection was performed through thoracotomy or video-assisted thoracic surgery was decided according to the situation of each case and the judgment of the physician of each institution. Lymph node dissection combined with pulmonary resection was not regulated, according to the judgment of the physician of each institution. Follow-up was not uniform in this study. However, most of the patients were treated according to the treatment guideline for colorectal cancer in Japan [18]. The study protocol was approved by the Institutional Review Board of all participating hospitals, including that of the Ethics Committee, Kyorin University Graduate School of Medicine (H24-051). We conducted a large-scale multiinstitutional retrospective cohort study to inform the
Ann Thorac Surg 2016;-:-–-
decision-making process for these patients. Data were initially obtained on 1,237 consecutive patients who had PM from colorectal cancer and underwent first pulmonary resection at any of the 46 institutions. After reviewing all the data for eligibility, completeness, and consistency, 339 cases were excluded. The remaining 898 cases of PM of colorectal cancer were studied in a previous related study [19]. Among these 898 cases, 651 cases were solitary PM and 247 cases were multiple PM. These 247 cases were analyzed in the present study (Fig 1). We examined differences in tumor sizes in two steps, as follows.
Step 1, Individual Heterogeneity of Metastatic Tumor Size Patients in the study cohort had multiple PM of both similar and different sizes. Given the variation in tumor sizes within individual patients, we decided to investigate the relationship between prognosis and the individual heterogeneity of tumor sizes within each patient. To obtain the individual heterogeneity of tumors sizes, we investigated the difference between the maximum and minimum tumor diameters (Dmax-min) on pathologic findings (Supplemental Table). We classified patients into three groups: Dmax-min 0 mm to 5 mm; Dmax-min 6 mm to 10 mm, and Dmax-min 11 mm or more. We examined the relationship between individual heterogeneity of tumor sizes and risk ratios using a Cox proportional hazards model.
Step 2, Prognostic Factors of Colorectal Cancer PM and Multivariate Analyses We also analyzed prognostic factors of colorectal cancer PM. The patients’ medical records were reviewed for the following information: age; sex; location of the primary cancer; disease-free interval (DFI); synchronicity with the primary cancer; site of PM; prethoracotomy level of carcinoembryonic antigen (CEA); surgical procedure; type of lymph node dissection; pathologic maximum diameter of PM; pathologic number of PM; pathologic pulmonary lymph node metastasis; chemotherapy before discovery of PM; preoperative chemotherapy before the discovery of PM; pathologic evaluation of preoperative chemotherapy; adjuvant chemotherapy after pulmonary resection; and heterogeneity of tumor sizes. Additionally, we performed multivariate analyses of the prognostic factors of colorectal cancer PM.
Statistical Analysis The DFI was defined as the time between the date of resection of the primary tumor and the date at which PM was first detected. The DFI was set to zero for patients with a primary cancer that was found after resection of a metastatic lesion. Overall survival was defined as the time between the date of pulmonary metastasectomy and the date of death. To identify prognostic factors after pulmonary metastasectomy, survival curves and survival rates (ie, overall survival rates at 5 years) were estimated using the Kaplan-Meier method. Log rank tests were used to compare survival across the levels of categoric variables.
3
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Table 1. Patient Characteristics in Overall Study Cohort All Patients (n ¼ 247)
Characteristics Sex Male Female Age, years Synchronicity with primary cancer Synchronous Metachronous Disease-free interval, years <2 2 Location of primary tumors Colon Rectum Extrathoracic lesions Absent Present Site of PM Unilateral Bilateral Carcinoembryonic antigen p-Max diameter of PM Number of PM Mode of surgery Less than lobectomya Lobectomy or moreb p-LN metastasis Absent Present Not available CT before finding PM Yes No Not available Preoperative CT after finding PM Yes No Adjuvant CT after PM resection Yes No a Wedge resection or segmentectomy. pneumonectomy.
142 105 63 (20–85) 40 207 170 77 103 144 187 60 133 114 3.9 (0.4–140.5) 16 (5–80) 2 (2–10) 57 190 136 11 100 136 101 10 56 191 124 123 b
Lobectomy, bilobectomy,
Values are n or median (range). CT ¼ chemotherapy; pulmonary metastases;
p-LN ¼ pathologic lymph node; p-Max ¼ pathologic maximum.
PM ¼
To counteract the arbitrariness of choosing a threshold and dichotomizing, the effects of continuous variables on survival were also evaluated using Cox proportional hazards models. In multivariate analyses of categoric variables, we used Cox proportional hazards models with the stepwise selection method. All p values less than 0.05 were considered statistically significant. The statistical analyses were conducted using JMP software, version 9.0 (SAS Institute, Cary, NC).
4
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Fig 2. Overall survival rate of all patients with multiple metastases. Overall survival was 75.4% at 3 years and 58.4% at 5 years.
Results The median follow-up period was 65.1 months (range, 0 to 111.8). Table 1 shows the characteristics of the overall patient cohort. Figure 2 presents the overall survival of all patients with multiple PM. The 3-year and 5-year overall survival rates were 75.4% and 58.4%, respectively. Figure 3 shows the heterogeneity of tumor sizes and risk ratios. The Dmax-min 6 mm to 10 mm group and the Dmax-min 11 mm or more group had a trend toward poorer survival than the Dmax-min 0 mm to 5 mm group, with hazard ratios of 1.43 and 1.79, respectively. The hazard ratio increased with increasing Dmax-min. Table 2 shows the results of univariate analyses of other risk factors in patients with multiple PM. Significant prognostic factors for overall survival included DFI, CEA, number of PM, use of preoperative chemotherapy after discovery of PM, and Dmax-min even though the maximum diameter of PM was not significantly prognostic. In the present study, exploration of lymph nodes was not routinely performed combined with pulmonary metastasectomy, but sampling or systematic mediastinal lymph node dissection was performed in 59.3% of
Ann Thorac Surg 2016;-:-–-
patients (147 of 247). The incidence of lymph node metastasis confirmed pathologically was only 7.5% (11 of 147), and the lymph node metastasis was not a significant prognostic factor in our analysis. There was no significant difference between synchronous disease and metachronous metastases, although there was significant difference in DFI. Multivariate analysis using a Cox proportional hazards model revealed that DFI, CEA, number of PM, Dmax-min and were significant prognostic factors, even though preoperative chemotherapy after the discovery of PM was not significantly prognostic (Table 3). Accordingly, we classified a 5 mm or less difference of tumor sizes as “small” and a greater than 5 mm difference of tumor sizes as “large.” Ninety-five patients had small (5 mm or less) ranges of tumor sizes (group A) and 151 patients had large (greater than 5 mm) ranges of tumor sizes (group B). Figure 4 shows the overall survival rates of patients in group A (n ¼ 95) and group B (n ¼ 152). The 5-year overall survival rates in group A and group B were 65.5% and 53.2%, respectively. There was a significant difference in overall survival between the two groups (p ¼ 0.025).
Comment Although many studies have investigated pulmonary resection for PM of colorectal cancer [1–13], few have specifically investigated patients with multiple PM. Previous reports noted 5-year overall survival rates of 15.7% to 45.0% among patients with multiple PM from colorectal cancer [1, 9, 10]; this prognosis is poor. The 5-year overall survival rate was 58.4% in the present study. Survival was favorable in the present study as compared with that observed in previous studies [1, 9, 10]. The major explanation of this difference is considered due to the period of this trial. Compared with reports from the 1990s, the more selective group of patients with PM in this study had undergone pulmonary resection, owing to the emergence of helical computed tomography and positron
Fig 3. Relationship between the heterogeneity of tumor sizes and risk ratios. The difference between the maximum and minimum tumor diameters (Dmax-min) of the 6 mm to 10 mm group and the Dmax-min 11 mm or more group tended to have poorer survival than the Dmax-min 0 mm to 5 mm group, with hazard ratios of 1.43 and 1.79, respectively. (CI ¼ confidence interval.)
Ann Thorac Surg 2016;-:-–-
5
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Table 2. Univariate Analyses of Other Risk Factors in Patients With More Than One Primary Metastasis
Table 3. Cox Proportional Hazards Regression Analysis for Survival
Variables
Variables
RR
95% CI
p Value
Dmax-min, >5 mm versus 5 mm PM, 5 versus <5 CEA, high versus normal DFI, <2 years versus 2 years Preoperative CT after PM, yes versus no
1.55 2.15 1.61 1.73 1.23
1.02–2.42 1.23–3.57 1.07–2.43 1.07–2.88 0.76–1.93
0.042 0.0082 0.022 0.023 0.38
Sex Male Female Age, years >65 65 Synchronicity with primary cancer Synchronous Metachronous Location of primary tumors Colon Rectum Disease-free interval, years <2 2 Extrathoracic lesion Absent Present Site of PM Unilateral Bilateral Carcinoembryonic antigen Normal High p-Max diameter of PM 2 cm >2 cm Number of PM 2–4 5 Mode of surgery Less than lobectomya Lobectomy or moreb p-LN metastasis Absent Present Chemotherapy before finding PM Yes No Preoperative CT after PM Yes No Adjuvant CT after PM resection Yes No Dmax-min 5 mm >5 mm a
n
5-Year OS (%) p Value 0.46
142 105
58.6 57.9
139 108
56.0 61.4
40 207
64.9 57.1
103 144
59.6 57.6
170 77
51.9 72.0
77 170
58.8 57.0
133 114
63.5 53.0
139 106
66.9 45.7
167 80
59.7 55.3
222 25
61.6 33.9
190 57
56.7 64.1
136 11
59.8 66.7
136 101
55.5 59.8
56 191
43.8 62.3
124 123
59.1 57.6
95 152
66.5 53.2
0.39
0.43
0.42
0.011
0.36
0.074
0.0032
0.29
0.0025
0.30
0.76
0.27
CEA ¼ carcinoembryonic antigen; CI ¼ confidence interval; CT ¼ chemotherapy; DFI ¼ disease-free interval; Dmax-min ¼ difference maximum-minimum diameter; PM ¼ pulmonary metastases; RR ¼ risk ratio.
emission tomography. These new diagnostic modalities could exclude the patients with unresectable metastasis invisible by the old modality. Indeed, patients who underwent pulmonary resection in this study had less lymph node metastasis and smaller metastatic tumor sizes than that observed in previous studies [1, 4–6, 11, 13]. Furthermore, the median overall survival of patients with unresectable colon cancer was 5 to 8 months in previous reports; however, the emergence of new drugs, such as oxaliplatin and irinotecan, increased that to 20 months [14, 15]. Moreover, targeted therapy led to a better prognosis [16, 17]. Therefore, the prognosis even after recurrence seemed longer in this study than in the previous reports. When considering the hematogenous metastatic pathway from colorectal cancer, tumor cells reach the lungs through the liver through the portal vein or directly through the inferior vena cava. In either case, cancer cells that migrate into the blood necessarily arrive at the lungs and then spread through the body. However, if cancer cells with metastatic ability are trapped within the lungs (as a filter organ), then metastases to the lungs may be regarded as semilocal disease, despite the distant metastasis from colorectal cancer. The hypothesis that the lungs act as a filter for cancer cells has been suggested as an anatomic and mechanical function, based on experimental animal studies and autopsy cases [20, 21].
0.032
0.45
0.025
Wedge resection or segmentectomy. pneumonectomy.
b
Lobectomy, bilobectomy,
CT ¼ chemotherapy; Dmax-min ¼ difference maximum-minimum diameter; OS ¼ overall survival; p-LN ¼ pathologic lymph node; PM ¼ pulmonary metastases; p-Max ¼ pathologic maximum.
Fig 4. Overall survival rates of patients in group A (n ¼ 95 [dashed line]) and group B (n ¼ 151 [solid line]). The 5-year overall survival rates were 66.5% for group A and 53.0% for group B. Overall survival differed significantly between the groups (p ¼ 0.025).
6
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
Multiple PM of similar and various sizes exist. Based on this hypothesis of a cascade, we assume in the cases with small differences of PM sizes, migration and attachment of circulating cancer cells to the pulmonary capillary wall and its extravasation at each site occur in almost the same period. This period may be the time of surgery for the primary lesion or may reflect some condition of the patients. In this case, cancer cells with metastatic ability may stay in the lungs. Therefore, PM with a small difference of tumor sizes may be a form of less aggressive disease, just as a solitary metastasis. In contrast, we believe that in the cases with PM with a large difference of tumor sizes, cancer cells arrive at the lungs and grow for various periods. It is possible that cancer cells migrate into the blood and circulate at various periods, or do so constantly. Therefore, PM with a large difference of tumor sizes may be a form of systemic disease. We thought that the patients with a larger range of tumors size would have a poorer prognosis; however, there was no significant difference in larger differences by multivariate analysis. We think that the reason was confounding factors such as size or number of multiple tumors. Moreover, the number of patients with larger differences was smaller. Consequently, there was a significant difference in 5-mm difference of tumor sizes by multivariate analysis, and we applied “5 mm” to the cutoff value in the present study. It has been unclear whether lymphadenectomy with metastasectomy results in better survival. A poor prognosis has been observed for patients with hilar and mediastinal lymph node metastasis identified in PM during or after surgery; the 5-year overall survival rate is 6.2% to 23% [11–13]. Therefore, most patients with clinical lymph node metastasis do not undergo pulmonary resection. In the present study, the lymph nodes of 100 patients (40.5%) were not evaluated pathologically. And only 11 of 147 patients confirmed pathologically had lymph node metastasis. This small number reduces the statistical power showing the significance of lymph node metastasis as a prognostic factor. In group A, lymph node metastasis developed in only 1 of 38 patients (2.6%) who underwent more than lymph sampling. Conversely, in group B, 10 of 109 patients (9.2%) had lymph node metastasis. This result may also show that multiple PM with small differences of tumor sizes could be interpreted as less aggressive disease. An increase in the number of multiple metastases is a negative predictor of survival [5, 6]. Twenty-five patients had more than five PM. The 5-year overall survival rate of patients with more than 5 PM was 33.9%. Only 1 patient in group A had more than five PM. The number of metastases and the difference of tumor sizes were both significant and independent prognostic factors, although the difference of tumor sizes was associated with the number of metastases. In the present study, the DFI was defined as the time between the date of resection of the primary tumor and the date at which PM was first detected. There was no significant difference between synchronous disease and metachronous metastases, although there was significant
Ann Thorac Surg 2016;-:-–-
difference in DFI. Consequently, we did not separate the synchronous disease from the metachronous metastases. In the present study, we identified the following prognostic factors for patients with multiple PM: number of metastases, CEA, DFI, and range of tumor sizes. Although univariate analysis showed a significant difference in overall survival for preoperative chemotherapy, this difference did not remain significant in the multivariate analysis. There are two reasons for these findings. First, patients who had the highest risk factors for recurrence—such as many PM, elevated CEA, and short DFI—received chemotherapy before pulmonary resection. Second, there were conversion cases, in which patients could not initially undergo pulmonary resection because of oncologic or patient-related factors, but subsequently received chemotherapy and showed downstaging or had tumor regression because of good response to chemotherapy. Consequently, these patients were ultimately able to undergo pulmonary resection. In either case, there was no significant difference in preoperative chemotherapy. This study has several limitations that should be considered as its results are interpreted. First, this was a retrospective multicenter cohort study; pulmonary resection was performed according to the judgment of each individual institute, following criteria that were based on protocols. Second, this is a study of surgical cases only; accordingly, the patients included in the analysis were highly selected and might not be representative of all patients with PM from colorectal cancer. Nonetheless, there have been no reports of multicenter, large studies that are specific to multiple PM from colorectal cancer. Improvements in chemotherapy have been reported, such as the development of new drugs and targeted therapy for colorectal cancer. We believe that patients who have multiple PM with a small difference in tumor sizes should undergo initial pulmonary resection— just as they have to date—because their disease may be less aggressive. However, we think that it may be better for patients with multiple PM and a large difference in tumor sizes to undergo induction chemotherapy or follow a wait-and-see approach for some period after pulmonary resection, because their disease may be systemic. In conclusion, the range of tumor sizes may be a good indicator when judging the appropriateness of surgical treatment for multiple PM in patients with colorectal cancer. The authors wish to thank all the patients and physicians at the institutions participating in the Tsukuba Cancer Trial Group for contributing to this series: Osaka Medical Center for Cancer and Cardiovascular Disease (Osaka, Japan); Kurashiki Central Hospital (Okayama, Japan); Hokkaido Cancer Center (Hokkaido, Japan); Kitasato University School of Medicine (Tokyo, Japan); Tokyo Women’s Medical University (Tokyo, Japan); Hyogo Cancer Center (Hyogo, Japan); Hiroshima City Hospital (Hiroshima, Japan); Toho University School of Medicine (Tokyo, Japan); Kyoto University Graduate School of Medicine (Kyoto, Japan); Shinshu University School of Medicine, (Nagano, Japan); Nagoya City University Graduate School of Medical Sciences (Nagoya, Japan); Tokyo Metropolitan Cancer and Infectious Diseases Center, Komagome Hospital (Tokyo, Japan); National
Ann Thorac Surg 2016;-:-–-
Kyushu Cancer Center (Fukuoka, Japan); Shikoku Cancer Center (Ehime, Japan); Niigata University Graduate School of Medical and Dental Sciences (Niigata, Japan); Gunma Prefectural Cancer Center (Gunma, Japan); Osaka University Graduate School of Medicine (Osaka, Japan); Hamamatsu University School of Medicine (Shizuoka, Japan); Yamagata Prefectural Central Hospital (Yamagata, Japan); Jikei University School of Medicine (Tokyo, Japan); Kumamoto University Hospital (Kumamoto, Japan); Fukujuji Hospital, Japan Anti-Tuberculosis Association (Tokyo, Japan); Saitama Medical Center, Saitama Medical University (Saitama, Japan); Tokai University Hachioji Hospital (Tokyo, Japan); Saitama Cardiovascular and Respiratory Center (Saitama, Japan); Saitama Medical University International Medical Center (Saitama, Japan); Kurume University School of Medicine (Fukuoka, Japan); Toyama Prefectural Central Hospital (Toyama, Japan); University of Tsukuba (Ibaraki, Japan); Mie University School of Medicine (Mie, Japan); St. Marianna University School of Medicine (Tokyo, Japan); Kanagawa Cancer Center (Kanagawa, Japan); Gunma University Graduate School of Medicine (Gunma, Japan); Nishi-Niigata Chuo National Hospital (Niigata, Japan); Fukushima Medical University (Fukushima, Japan); Okayama Saiseikai General Hospital (Okayama, Japan); Ishikawa Prefectural Central Hospital (Ishikawa, Japan); Oita Prefectural Hospital (Oita, Japan); Showa University School of Medicine (Tokyo, Japan); and Kawasaki Medical School (Okayama, Japan).
References 1. Iida T, Nomori H, Shiba M, et al, for the Metastatic Lung Tumor Study Group of Japan. Prognostic factors after pulmonary metastasectomy for colorectal cancer and rationale for determining surgical indications: a retrospective analysis. Ann Surg 2013;257:1059–64. 2. Watanabe K, Nagai K, Kobayashi A, Sugito M, Saito N. Factors influencing survival after complete resection of pulmonary metastases from colorectal cancer. Br J Surg 2009;96: 1058–65. 3. Hirosawa T, Itabashi M, Ohnuki T, Yamaguchi N, Sugihara K, Kameoka S, for the Japanese Society for Cancer of the Colon and Rectum (JSCCR) Study Group for Pulmonary Metastases from Colorectal Cancer. Prognostic factors in patients undergoing complete resection of pulmonary metastases of colorectal cancer: a multi-institutional cumulative follow-up study. Surg Today 2013;43:494–9. 4. Pfannschmidt J, Dienemann H, Hoffmann H. Surgical resection of pulmonary metastases from colorectal cancer: a systematic review of published series. Ann Thorac Surg 2007;84:324–38. 5. Welter S, Jacobs J, Krbek T, Poettgen C, Stamatis G. Prognostic impact of lymph node involvement in pulmonary metastases from colorectal cancer. Eur J Cardiothorac Surg 2007;31:167–72. 6. Kanemitsu Y, Kato T, Hirai T, Yasui K. Preoperative probability model for predicting overall survival after resection of pulmonary metastases from colorectal cancer. Br J Surg 2004;91:112–20.
MANIWA ET AL MULTIPLE METASTASES OF COLORECTAL CANCER
7
7. Watanabe I, Arai T, Ono M, et al. Prognostic factors in resection of pulmonary metastasis from colorectal cancer. Br J Surg 2003;90:1436–40. 8. Onaitis MW, Petersen RP, Haney JC, et al. Prognostic factors for recurrence after pulmonary resection of colorectal cancer metastases. Ann Thorac Surg 2009;87:1684–8. 9. Lin BR, Chang TC, Lee YC, Lee PH, Chang KJ, Liang JT. Pulmonary resection for colorectal cancer metastases: duration between cancer onset and lung metastasis as an important prognostic factor. Ann Surg Oncol 2009;16: 1026–32. 10. Iizasa T, Suzuki M, Yoshida S, et al. Prediction of prognosis and surgical indications for pulmonary metastasectomy from colorectal cancer. Ann Thorac Surg 2006;82:254–60. 11. Okumura S, Kondo H, Tsuboi M, et al. Pulmonary resection for metastatic colorectal cancer: experiences with 159 patients. J Thorac Cardiovasc Surg 1996;112:867–74. 12. Saito Y, Omiya H, Kohno K, et al. Pulmonary metastasectomy for 165 patients with colorectal carcinoma: a prognostic assessment. J Thorac Cardiovasc Surg 2002;124: 1007–13. 13. B€ ol€ ukbas S, Sponholz S, Kudelin N, Eberlein M, Schirren J. Risk factors for lymph node metastases and prognosticators of survival in patients undergoing pulmonary metastasectomy for colorectal cancer. Ann Thorac Surg 2014;97: 1926–32. 14. Grothey A, Sargent D, Goldberg RM, Schmoll HJ. Survival of patients with advanced colorectal cancer improves with the availability of fluorouracil-leucovorin, irinotecan, and oxaliplatin in the course of treatment. J Clin Oncol 2004;22: 1209–14. 15. Tournigand C, Andr e T, Achille E, et al. FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal cancer: a randomized GERCOR study. J Clin Oncol 2004;22: 229–37. 16. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–42. 17. Giantonio BJ, Catalano PJ, Meropol NJ, et al, for the Eastern Cooperative Oncology Group Study E3200. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200. J Clin Oncol 2007;25:1539–44. 18. Watanabe T, Itabashi M, Shimada Y, et al, for the Japanese Society for Cancer of the Colon and Rectum. Japanese Society for Cancer of the Colon and Rectum (JSCCR) guidelines 2014 for treatment of colorectal cancer. Int J Clin Oncol 2015;20:207–39. 19. Hishida T, Okumura T, Boku N, et al. Surgical outcome for pulmonary metastasis of colorectal cancer in the modern chemotherapy era: results of a retrospective Japanese multicenter study. J Clin Oncol 2014;32(Suppl):3528. 20. Weiss L, Grundmann E, Torhorst J, et al. Haematogenous metastatic patterns in colonic carcinoma: an analysis of 1541 necropsies. J Pathol 1986;150:195–203. 21. Chambers AF, Groom AC, MacDonald IC. Dissemination and growth of cancer cells in metastatic sites. Nat Rev Cancer 2002;2:563–72.