Professional Documents
Culture Documents
Akiyoshi AnnSurgJSCCR
Akiyoshi AnnSurgJSCCR
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Akiyoshi et al Annals of Surgery r Volume 255, Number 6, June 2012
number of LN metastasis, presence and site of distant metastasis, proportional hazards model with the backward stepwise elimination
date of the most recent follow-up, and cancer-specific prognosis. We method (significance level to stay = 0.05) was used to multivariately
excluded patients with in situ or metastatic disease, patients with ma- assess predictors of outcome. Covariates considered in this analysis
lignancies other than adenocarcinoma, patients who did not undergo were sex, age, histological type, depth of tumor, lymphatic invasion,
a curative (R0) tumor resection, and patients in whom the region or venous invasion, LPLD, number of LNs removed, perioperative radia-
number of LN metastasis was unknown. The total number of patients tion, adjuvant chemotherapy, and status of LN metastasis. The results
after all exclusions was 11,567. of Cox model analysis are reported using hazard ratios (HRs) and
95% confidence intervals (CIs).
Division of the Lateral Pelvic Area
The lateral pelvic area was classified into 6 regions according RESULTS
to the JSCCR classification5 : internal iliac, obturator, external iliac, Table 1 summarizes the characteristics of the 11,567 patients
common iliac, aortic bifurcation, and median sacral regions (Fig. 1). with low rectal cancer who underwent curative resection from 1978
The registry database included the number of dissected and metastatic to 1998. Male prevalence was higher and the median age was 61 years
LNs in each of the 6 regions, and patients with 1 or more dissected (IQR [interquartile range], 53–69 years). Sixty-two percent of the pa-
LNs in the lateral pelvic area were classified as having undertaken tients had T3 or T4 tumors in depth. LPLD was performed in almost
LPLD. 50% of the patients. Mesenteric LN metastasis without lateral pelvic
Patients were divided into 4 groups according to the extent LN metastasis (mesorectal-LN) was present in 33.7% of the patients.
of their LN metastasis. Patients without LN metastasis were clas- Lateral pelvic LN metastasis (internal- and external-LPLN) was iden-
sified as the N0 group, whereas those with mesenteric metastatic tified in 5.7% (655/11,567) of the total patients, in 7.8% (558/7123)
LNs (mesorectal nodes or nodes adjacent to the inferior mesenteric of the patients who had T3 or T4 tumors, in 11.3% (655/5789) of the
artery) without metastatic lateral pelvic LNs were classified as the patients who underwent LPLD, and in 14.6% (558/3832) of the pa-
mesorectal-LN group. Patients who had metastatic internal iliac LNs tients with T3 or T4 tumors who underwent LPLD. Of the 655 patients
without metastatic LNs in the other 5 lateral pelvic regions were with lateral pelvic LN metastasis, metastasis in the lateral pelvic area
classified as the internal-lateral pelvic lymph nodes (LPLN) group,
whereas those with metastatic lateral pelvic LNs in the 5 lateral
pelvic regions (obturator, external iliac, common iliac, aortic bifurca-
tion, and median sacral), with or without the internal iliac area, were TABLE 1. Characteristics of Patients with Low Rectal
classified as the external-LPLN group. Cancer Who Underwent Curative Resection
1130 | www.annalsofsurgery.com
C 2012 Lippincott Williams & Wilkins
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 255, Number 6, June 2012 Lateral Lymph Node in Low Rectal Cancer
was limited to the internal iliac area in 411 patients (62.7%, internal- = 0.0004), histology other than well/moderately differentiated ade-
LPLN). nocarcinoma (P < 0.0001), T3/T4 tumor in depth (P < 0.0001),
Table 2 summarizes the characteristics of patients with and of presence of lymphatic invasion (P = 0.0069), presence of venous
those without LPLD. LPLD was significantly associated with a higher invasion (P = 0.0263), and presence of mesenteric LN metastasis
percentage of age less than 60 years (P < 0.0001), histology other (P < 0.0001) were independently associated with lateral pelvic LN
than well/moderately differentiated adenocarcinoma (P = 0.0364), metastasis.
T3/T4 tumor in depth (P < 0.0001), presence of lymphatic (P < Follow-up data were available in 8933 patients who were reg-
0.0001) or venous (P < 0.0001) invasion, number of LNs removed istered from 1978 to 1994. The overall median follow-up time was 67
12 or more (P < 0.0001), and performing perioperative radiation (P months (IQR, 32–82 months). Both overall survival (OS) and CSS
< 0.0001) or adjuvant chemotherapy (P < 0.0001). Table 3 summa- were significantly different according to the status of LN metastasis
rizes the risk factors for lateral pelvic LN metastasis in 5789 patients (Fig. 2). Five-year OSs of the N0, mesorectal-LN, internal-LPLN, and
with LPLD. Logistic regression analysis showed that female sex (P external-LPLN groups were 80% (95% CI, 79%–81%), 55% (53%–
56%), 45% (39%–50%), and 29% (23%–37%), respectively. Five-
year CSSs of the N0, mesorectal-LN, internal-LPLN, and external-
LPLN groups were 88% (95% CI, 87%–89%), 61% (59%–62%),
TABLE 2. Characteristics of Patients With and Without 49% (43%–55%), and 34% (27%–42%), respectively.
Lateral Pelvic Lymph Node Dissection To determine which variables had an independent effect on
LPLD No LPLD postoperative survival, a multivariate Cox regression analysis with the
Variables n (%) n (%) P backward stepwise elimination technique was performed using the 11
variables shown in Table 1. The results are summarized in Table 4.
Sex Independent prognostic factors associated with both OS and CSS were
Male 3659 (63.2) 3570 (61.8) 0.1153
age, histological type, depth of tumor, lymphatic invasion, LPLD,
Female 2130 (36.8) 2208 (38.2)
Age number of LNs removed, and status of LN metastasis. Gender was
<60 2792 (48.3) 2466 (42.7) <0.0001 independently associated with OS, and venous invasion and adjuvant
≥60 2993 (51.7) 3304 (57.3) chemotherapy were independently associated with CSS.
Histological type
Well/moderate 5396 (93.4) 5432 (94.3) 0.0364
Other 383 (6.6) 327 (5.7)
A 1
Depth of tumor
0.9 80%
T1/T2 1927 (33.5) 2461 (42.8) <0.0001
Probability of survival
0.8 N0
T3/T4 3832 (66.5) 3291 (57.2)
0.7
Lymphatic invasion 55%
0.6
Yes 3708 (66.6) 3209 (59.1) <0.0001 45% Mesorect al-LN
0.5
No 1863 (33.4) 2217 (40.9)
Venous invasion 0.4 29% Internal -LPLN
Yes 2877 (52.3) 2212 (41.3) <0.0001 0.3
External-LPLN
No 2619 (47.7) 3150 (58.7) 0.2 P < .0001
Number of LNs removed 0.1
1–11 762 (13.2) 3087 (53.4) <0.0001 0
0 1 2 3 4 5 6 7
≥12 5027 (86.8) 2691 (46.6) Years from surgery
Perioperative radiation No. at risk
Yes 458 (8.0) 271 (4.7) <0.0001
N0 5391 5005 4623 4296 3968 3575 2710 1388
No 5249 (92.0) 5446 (95.3)
Mesorect al-LN 3053 2760 2316 1922 1631 1373 1016 517
Adjuvant chemotherapy Internal -LPLN 316 291 232 172 140 117 100 67
Yes 3613 (63.3) 3113 (54.5) <0.0001 External -LPLN 173 149 105 79 55 46 29 13
No 2094 (36.7) 2604 (45.5)
Status of LN metastasis
N0 3335 (57.6) 3672 (63.6) <0.0001
Mesorectal-LN 1799 (31.1) 2106 (36.4) B 1 88%
Internal-LPLN 411 (7.1) 0 0.9 N0
Probability of survival
Depth of tumor (T3/4) 2.19 (1.71–2.84) <0.0001 Internal -LPLN 316 291 232 172 140 117 100 67
External -LPLN 173 149 105 79 55 46 29 13
Lymphatic invasion (yes) 1.39 (1.10–1.76) 0.0069
Venous invasion (yes) 1.24 (1.03–1.51) 0.0263
FIGURE 2. Overall survival (A) and cancer-specific survival (B)
Mesenteric LN metastasis (yes) 4.64 (3.77–5.74) <0.0001
for 8933 patients.
C 2012 Lippincott Williams & Wilkins www.annalsofsurgery.com | 1131
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Akiyoshi et al Annals of Surgery r Volume 255, Number 6, June 2012
Finally, we tried to determine whether lateral pelvic LN metas- to the internal iliac area, but it was comparable with that of N2b and
tasis is regional or distant disease. We subdivided the mesorectal-LN better than that of stage IV patients who underwent curative resection.
group into 4 categories on the basis of the AJCC Cancer Staging Our findings suggest that lateral pelvic LNs other than internal iliac
Manual (N1a: 1 regional LN metastasis; N1b: 2–3 metastases; N2a: LNs can also be treated by LPLD as regional LNs in low rectal
4–6 metastases; N2b: ≥7 metastases). Five-year OSs of the N1a, cancer.
N1b, N2a, and N2b groups were 68% (95% CI, 65%–70%), 54% With regard to the effectiveness of LPLD for rectal cancer,
(51%–57%), 45% (41%–49%), and 32% (28%–37%), respectively. Georgiou and colleagues1 demonstrated by meta-analysis that LPLD
Five-year CSSs of the N1a, N1b, N2a, and N2b groups were 74% did not confer a significant overall cancer-specific advantage. How-
(95% CI, 72%–77%), 60% (56%–63%), 51% (46%–55%), and 37% ever, there are potential limitations to this previous study by Geor-
(32%–42%), respectively. Both OS and CSS were not significantly giou and colleagues1 that might have led to misleading conclusions.21
different between the internal-LPLN and N2a groups (P = 0.9585 First, there was a significant selection bias between the LPLD and
for OS and 0.5742 for CSS), or between the external-LPLN and N2b non-LPLD groups. LPLD is generally indicated for advanced (T3 and
groups (P = 0.3342 for OS and 0.4347 for CSS) (Fig. 3). In ad- T4) rectal cancer that extends below the peritoneal reflection in Japan,
dition, the survival of 260 patients with stage IV low rectal cancer based on the fact that the rectum below the peritoneal reflection has
who underwent R0 resection from 1978 to 1994 was also analyzed. lymphatic channels that extend laterally.3 However, the final indica-
Five-year OSs and CSSs of the stage IV group were 24% (95% CI, tion is at the surgeon’s discretion, and consequently there is variability
19%–29%) and 27% (21%–33%), respectively. Both OS and CSS of in the LPLD indications among institutions. A recent large multicen-
the external-LPLN group were significantly better than those of the ter study reported by Kobayashi and colleagues11 analyzed 1272 pa-
stage IV group (P = 0.0240 for OS and 0.0117 for CSS) (Fig. 3). tients with low rectal cancer and showed that LPLD was more likely
to be performed in high-risk patients for lateral pelvic LN metastasis,
DISCUSSION that is, larger tumors, a higher stage, and more aggressive pathology.
The effectiveness of LPLD is controversial because lateral Up until the current study, Kobayashi and colleagues11 had performed
pelvic LN metastasis may represent systemic disease not amenable to the largest comparative study examining LPLD versus non-LPLD for
surgical cure.1,3 In an analysis of 11,567 patients with low rectal can- low rectal cancer, and it was included in the meta-analysis reported
cer registered in a nationwide multi-institutional database in Japan, by Georgiou and colleagues.1 Therefore, considering that more ad-
we showed that lateral pelvic LN metastasis could be considered as vanced tumors are prevalent in LPLD patients compared with those
regional disease. We also showed that LPLD conferred a significant who do not have LPLD, the finding of a similar survival and local
cancer-specific advantage when analyzed using the multivariate Cox recurrence rate between the 2 groups of patients suggests that LPLD
proportional hazards model. To date, the present study is the largest is beneficial. Second, the time period of the meta-analysis was very
study to analyze the prognosis of lateral pelvic LN metastasis in low long (1984 to 2007). Recently, Japanese surgeons have become more
rectal cancer.1–3,11–20 selective with regard to the indications for LPLD because of previous
The survival of patients with lateral pelvic LN metastasis was analyses of large studies and using modern imaging techniques12,22 ;
significantly poorer than that of those with LN metastasis limited to therefore, the effectiveness of LPLD has not been the same over time.
the mesenteric area, suggesting that lateral pelvic LN metastasis is a Given that the prognosis of patients with lateral pelvic LN metastasis
poor prognostic indicator. However, among the patients with lateral who underwent LPLD was similar to N2, we can conclude that se-
pelvic LN metastasis, the survival of patients with LN metastasis lected patients with lateral pelvic LN metastasis would benefit from
limited to the internal iliac area was comparable with that of N2a and LPLD. In contrast, it is uncertain whether LPLD can confer a survival
better than that of N2b. This result indicates that internal iliac LNs in benefit for patients without obvious lateral pelvic LN metastasis. In
low rectal cancer represent regional LNs that can be cured by LPLD, an attempt to determine this issue, the Japanese Clinical Oncology
and it supports the AJCC definition in which internal iliac LNs are Group has started a phase 3 randomized controlled trial comparing
classified as regional LNs in rectal cancer. On the contrary, survival TME with LPLD and that without LPLD for stage II-III low rec-
of patients with lateral pelvic LN metastasis extending beyond the tal cancer without clinical imaging diagnosis of lateral pelvic LN
internal iliac area was poorer than that of those with metastasis limited metastasis.
1132 | www.annalsofsurgery.com
C 2012 Lippincott Williams & Wilkins
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 255, Number 6, June 2012 Lateral Lymph Node in Low Rectal Cancer
A 1 therapy, instead of LPLD for patients with low rectal cancer without
0.9 lateral pelvic LN metastasis.
Probability of survival
0.8
0.7
68% In contrast, it is unknown whether neoadjuvant chemoradio-
0.6 54%
N1a therapy without LPLD is an adequate therapy for lateral pelvic LN
0.5 45% N1b metastasis. Liang25 demonstrated that the rate of lateral pelvic LN
45% N2a
0.4
32% Internal -LPLN
P = 0.9585 metastasis was 71.1% in 45 LPLDs when imaging showed persis-
0.3 29% N2b
P = 0.3342 tent lateral pelvic LNs after neoadjuvant chemoradiotherapy. Kim
0.2 External -LPLN
0.1
24%
StageI V P = 0.0240 and colleagues26 demonstrated that lateral pelvic recurrence is the
0 most common pattern of local recurrence after neoadjuvant chemora-
0 1 2 3 4 5 6 7 diotherapy plus TME alone, and that it is strongly associated with
Years from surgery
No. at risk the preoperative radiologic size of lateral pelvic LN. These results
N1a 1129 1043 927 806 719 629 473 248 suggest that patients with preoperative evidence of lateral pelvic
N1b 974 887 750 634 519 428 311 153 lymphadenopathy could benefit from LPLD, even after neoadjuvant
N2a 563 508 402 310 256 211 154 77
N2b 387 326 239 175 140 109 79 39
chemoradiotherapy.
Internal-LPLN 316 291 232 172 140 117 100 67 There are several potential limitations of this study. The first
External-LPLN 173 149 105 79 55 46 29 13 potential limitation involves the accuracy of the data, because there
Stage IV 260 198 122 85 65 54 36 16
may have been coding errors. However, the validity of the JSCCR
database, including the records of various clinicopathological vari-
B 1 ables and outcomes, has been well established.7–10 The second poten-
0.9 tial limitation is that data on the morbidity associated with LPLD and
Probability of survival
0.8 74% the recurrence pattern were not available in the JSCCR database. The
0.7 N1a
60% third potential limitation is that our data included only the Japanese
0.6
0.5
51% N1b population, and it is not clear whether the frequency of lateral pelvic
49% N2a
0.4 37% Internal -LPLN
P = 0.5742 LN metastasis is similar between Western countries and Japan. There
0.3 34% N2b
P = 0.4347 are some previous studies reporting the results of rectal cancer treat-
External -LPLN
0.2 27%
StageI V P = 0.0117 ment by radical resection and LPLD from Western countries, but the
0.1 lateral pelvic LN metastatic status was not reported.19,20 Despite these
0
0 1 2 3 4 5 6 7 potential limitations, we believe that the results of this study could
No. at risk
Years from surgery serve as a basis for management of lateral pelvic LN metastasis in
N1a 1129 1043 927 806 719 629 473 248
low rectal cancer.
N1b 974 887 750 634 519 428 311 153 In conclusion, internal iliac LNs are considered to be regional
N2a 563 508 402 310 256 211 154 77 LNs in low rectal cancer, as defined by the AJCC. Patients with
N2b 387 326 239 175 140 109 79 39
Internal-LPLN 316 291 232 172 140 117 100 67
metastasis to the lateral pelvic LNs other than the internal iliac LN
External-LPLN 173 149 105 79 55 46 29 13 have a poorer prognosis, but the prognosis is similar to that of N2b,
Stage IV 260 198 122 85 65 54 36 16
and better than that of stage IV. In light of the findings that lateral
FIGURE 3. Overall survival (A) and cancer-specific survival (B) pelvic LN metastasis can be regarded as regional disease in a large
for 3542 patients with lymph node metastasis and 260 patients nationwide multi-institutional study, selected patients with low rectal
with stage IV who underwent curative resection. cancer could benefit from LPLD.
ACKNOWLEDGMENTS
The authors thank all members and staff in the JSCCR mem-
In Western countries, neoadjuvant chemoradiotherapy is ber institutions for collecting data for Japanese colorectal cancer
widely used as an adjunct to TME for advanced low rectal cancer. registration.
There have only been a few studies comparing the efficacy of LPLD
and radiotherapy. Recently, Kim and colleagues18 demonstrated that
REFERENCES
LPLD patients showed a 2.2-fold increase in the local recurrence
1. Georgiou P, Tan E, Gouvas N, et al. Extended lymphadenectomy versus conven-
rate compared with chemoradiotherapy patients among stage III rec- tional surgery for rectal cancer: a meta-analysis. Lancet Oncol. 2009;10:1053–
tal cancer. However, this previous study had potential limitations 1062.
to conclude that LPLD shows a higher local recurrence rate than 2. Kusters M, Beets GL, van de Velde CJ, et al. A comparison between the
chemoradiotherapy.23 One limitation is the different indications for treatment of low rectal cancer in Japan and the Netherlands, focusing on the
LPLD and chemoradiotherapy. LPLD is indicated for advanced rectal patterns of local recurrence. Ann Surg. 2009;249:229–235.
cancer that extends below the peritoneal reflection, which corresponds 3. Yano H, Moran BJ. The incidence of lateral pelvic side-wall nodal involve-
ment in low rectal cancer may be similar in Japan and the West. Br J Surg.
to 6 to 9 cm from the anal verge.3 However, chemoradiotherapy was 2008;95:33–49.
indicated for rectal cancer located within 15 cm from the anal verge 4. Japanese Society for Cancer of the Colon and Rectum. Japanese Classification
in the study by Kim and colleagues,18 and thus the LPLD group of Colorectal Carcinoma. 2nd English ed. Tokyo, Japan: Kanehara & Co; 2009.
would have included lower rectal cancers than the chemoradiother- 5. Japanese Society for Cancer of the Colon and Rectum. Japanese Classification
apy group. The second limitation is that the study did not refer to of Colorectal Carcinoma. 1st English ed. Tokyo, Japan: Kanehara & Co; 1997.
adjuvant chemotherapy, which would have affected the postoperative 6. Edge SB, Byrd DR, Compton CC, et al. AJCC Cancer Staging Manual. 7th
outcomes of rectal cancer patients.24 No conclusion regarding the ed. New York, NY: Springer; 2010.
effectiveness of LPLD compared with neoadjuvant chemoradiother- 7. Konishi T, Watanabe T, Kishimoto J, et al. Prognosis and risk factors of
metastasis in colorectal carcinoids: results of a nationwide registry over 15
apy can be drawn at present. However, Watanabe and colleagues17 years. Gut. 2007;56:863–868.
reported no significant survival difference between patients who had 8. Watanabe T, Konishi T, Kishimoto J, et al. Ulcerative colitis-associated col-
neoadjuvant radiotherapy with TME and those who had TME plus orectal cancer shows a poorer survival than sporadic colorectal cancer: a na-
LPLD, suggesting that neoadjuvant radiotherapy can be an alternative tionwide Japanese study. Inflamm Bowel Dis. 2011;17:802–808.
C 2012 Lippincott Williams & Wilkins www.annalsofsurgery.com | 1133
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Akiyoshi et al Annals of Surgery r Volume 255, Number 6, June 2012
9. Muto T, Kotake K, Koyama Y. Colorectal cancer statistics in Japan: data from total mesorectal excision in rectal cancer. Ann Surg. 2007;246:754–
JSCCR registration, 1974–1993. Int J Clin Oncol. 2001;6:171–176. 762.
10. Kotake K, Honjo S, Sugihara K, et al. Changes in colorectal cancer during a 19. Enker WE, Pilipshen SJ, Heilweil ML, et al. En bloc pelvic lymphadenectomy
20-year period: an extended report from the multi-institutional registry of large and sphincter preservation in the surgical management of rectal cancer. Ann
bowel cancer, Japan. Dis Colon Rectum. 2003;46:S32–S43. Surg. 1986;203:426–433.
11. Kobayashi H, Mochizuki H, Kato T, et al. Outcomes of surgery alone for lower 20. Michelassi F, Block GE, Vannucci L, et al. A 5- to 21-year follow-up and
rectal cancer with and without pelvic sidewall dissection. Dis Colon Rectum. analysis of 250 patients with rectal adenocarcinoma. Ann Surg. 1988;208:379–
2009;52:567–576. 389.
12. Sugihara K, Kobayashi H, Kato T, et al. Indication and benefit of pelvic sidewall 21. Yano H, Moran BJ, Watanabe T, et al. Lateral pelvic lymph-node dissection:
dissection for rectal cancer. Dis Colon Rectum. 2006;49:1663–1672. still an option for cure. Lancet Oncol. 2010;11:114.
13. Ueno H, Mochizuki H, Hashiguchi Y, et al. Potential prognostic benefit of 22. Yano H, Saito Y, Takeshita E, et al. Prediction of lateral pelvic node involve-
lateral pelvic node dissection for rectal cancer located below the peritoneal ment in low rectal cancer by conventional computed tomography. Br J Surg.
reflection. Ann Surg. 2007;245:80–87. 2007;94:1014–1019.
14. Hida J, Yasutomi M, Fujimoto K, et al. Does lateral lymph node dissection 23. Watanabe T, Matsuda K, Nozawa K, et al. Lateral pelvic lymph node dissec-
improve survival in rectal carcinoma? Examination of node metastases by the tion or chemoradiotherapy: which is the procedure of choice to reduce local
clearing method. J Am Coll Surg. 1997;184:475–480. recurrence rate in lower rectal cancer? Ann Surg. 2008;248:342–343.
15. Ueno M, Oya M, Azekura K, et al. Incidence and prognostic significance of 24. Akasu T, Moriya Y, Ohashi Y, et al. Adjuvant chemotherapy with uracil-
lateral lymph node metastasis in patients with advanced low rectal cancer. Br tegafur for pathological stage III rectal cancer after mesorectal excision with
J Surg. 2005;92:756–763. selective lateral pelvic lymphadenectomy: a multicenter randomized controlled
16. Moriya Y, Hojo K, Sawada T, et al. Significance of lateral node dissection trial. Jpn J Clin Oncol. 2006;36:237–244.
for advanced rectal carcinoma at or below the peritoneal reflection. Dis Colon 25. Liang JT. Technical feasibility of laparoscopic lateral pelvic lymph node dis-
Rectum. 1989;32:307–315. section for patients with low rectal cancer after concurrent chemoradiation
17. Watanabe T, Tsurita G, Muto T, et al. Extended lymphadenectomy and preop- therapy. Ann Surg Oncol. 2011;18:153–159.
erative radiotherapy for lower rectal cancers. Surgery. 2002;132:27–33. 26. Kim TH, Jeong SY, Choi DH, et al. Lateral lymph node metastasis is a major
18. Kim JC, Takahashi K, Yu CS, et al. Comparative outcome between cause of locoregional recurrence in rectal cancer treated with preoperative
chemoradiotherapy and lateral pelvic lymph node dissection following chemoradiotherapy and curative resection. Ann Surg Oncol. 2008;15:729–737.
1134 | www.annalsofsurgery.com
C 2012 Lippincott Williams & Wilkins
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.