World J Surg (2015) 39:732–739 DOI 10.1007/s00268-014-2846-y
ORIGINAL SCIENTIFIC REPORT
Prognostic Significance of Lymphovascular Invasion in Node-Negative Gastric Cancer Ju-Hee Lee • Min Gyu Kim • Min-Sung Jung Sung Joon Kwon
•
Published online: 7 November 2014 Ó Socie´te´ Internationale de Chirurgie 2014
Abstract Background The aim of this study was to investigate the prognostic value of lymphovascular invasion (LVI) in gastric cancer patients without lymph node metastasis. Methods A total of 699 patients with primary tumor pT1–3/pN0-1 gastric adenocarcinoma who underwent curative resection from 2001 to 2010 were categorized into 3 groups: One-hundred and eleven patients with pN0/LVI(?), 475 with pN0/LVI(-), and 103 with pN1. Results The tumors in patients with N0/LVI(?) had more aggressive clinicopathologic features than those in patients with N0/LVI(-). However, there was no significant difference in patient characteristics between patients with pN0/LVI(?) and those with pN1, except for histologic grade. There were no significant differences in the overall survival rate in patients with pN0/LVI(?) compared to those with pN0/LVI(-) or the pN1 stage. However, the recurrence-free survival rate of the pN0/LVI(?) group was lower than that of the pN0/LVI(-) group (p \ 0.001), while no significant difference was observed between the pN0/LVI(?) and the N1 groups (p = 0.216). In multivariate analysis, LVI was identified as a poor prognostic factor related to recurrence-free survival in node-negative gastric cancer patients. pT3 stage and less than D2 lymphadenectomy were poor prognostic factors affecting recurrence-free survival, and less than D2 lymphadenectomy was an independent poor prognostic factor for overall survival in pN0/LVI(?) patients. Conclusions LVI could be an indicator of biological aggressiveness and may be a reliable prognostic factor for node-negative gastric cancer. LVI should be considered in postoperative management of gastric cancer.
J.-H. Lee M.-S. Jung S. J. Kwon (&) Department of Surgery, Hanyang University Seoul Hospital, Seoul, Republic of Korea e-mail: [email protected] J.-H. Lee e-mail: [email protected] M. G. Kim Department of Surgery, Hanyang University Guri Hospital, Seoul, Republic of Korea M. G. Kim M.-S. Jung S. J. Kwon Department of Surgery, College of Medicine, Hanyang University, 222 Wangsimri-ro, Seongdong-gu, Seoul, Republic of Korea
123
Introduction Gastric cancer is the fourth most common cancer in the world and the second most common cause of cancer-related deaths worldwide [1]. The involvement of regional lymph nodes (N stage) is the most significant prognostic indicators for gastric cancer [2–4]. For early gastric cancer (EGC), in particular, it seems that LNM is a critical factor for prognosis and choice of treatment because the overall survival rate for EGC without lymph node metastasis (LNM) is 90–95 %, but this decreases to 70–85 % in the case of patients with LNM [5, 6].
World J Surg (2015) 39:732–739
Lymphovascular invasion (LVI) is a well-known risk factor for LNM [7, 8]. The prognostic effect of LVI has been previously investigated by many authors [5, 9–16]. However, the results are not consistent. Moreover, several studies have evaluated the role of LVI in gastric cancer, regardless of the lymph node status of the gastric cancer patients, and have shown a close correlation between lymph node metastasis and LVI. Studies to the pure prognostic effect of LVI are still rare. Recently, chemotherapy was shown to be an effective adjuvant treatment for stages II and III (American Joint Committee on Cancer/Union for International Cancer Control; AJCC/UICC 6th edition) gastric cancer [17, 18], but adjuvant treatment is not recommended for EGC and node-negative advanced gastric cancer (AGC) BT3 (subserosa invasion) according to recent Japanese gastric cancer treatment guidelines, because of lack of evidence [4]. Therefore, this study focused on T1–3N0 gastric cancer patients and tried to clarify the prognostic role of LVI in those patients, in order to determine its possible value in management of the disease.
Materials and methods Patients A total of 700 patients were selected from prospectively maintained gastric cancer surgery databases according to the following criteria: (i) pathologically proven gastric adenocarcinoma, (ii) primary tumor stage (pT)1–3 and regional lymph node (pN)0–1 according to AJCC/UICC 7th edition, (iii) no history of other therapy for gastric cancer (including neoadjuvant chemotherapy), and (iv) patients who underwent curative gastrectomy between 2001 and 2010 at the Hanyang University Medical Center (patients with 15 or more than 15 retrieved lymph nodes). Of those 700 patients, eleven for whom data on LVI were not available were excluded. The resected specimens and retrieved lymph nodes were stained with hematoxylin and eosin (H&E) and examined by specialized pathologists. Lymphatic invasion was defined as the presence of neoplastic cell emboli within endothelium-lined spaces. Venous invasion was identified by the presence of additional fibrin clots and/or erythrocytes in endothelium-lined spaces without erythrocyte extravasation into the surrounding tissue or by evidence of neoplastic cells within smooth muscle cell-lined spaces. Tumors were classified into three groups: 475 pN0/LVI (-), 111 pN0/LVI (?), and 103 pN1. Clinicopathologic features studied included sex, age, tumor location, extent of lymphadenectomy, tumor size, histological grade, lymphatic invasion, venous invasion,
733
perineural invasion, and tumor stage (AJCC/UICC 7th edition). Extent of lymph node dissection was classified according to the guidelines of the Japanese Gastric Cancer Association [19]. D2 lymphadenectomy was performed in most of the patients with EGC and AGC. Less than D2 lymphadenectomy (D1 ? a or D1 ? b) was performed in some patients with clinical early gastric cancer without LNM diagnosed preoperatively (n = 30) or comorbidities related to a high complication rate (n = 9). The comorbidities were heart disease in four patients (three patients had ischemic heart disease and one had heart failure), cerebrovascular accident in three, liver cirrhosis in two, and chronic obstructive pulmonary disease in 1 (one patient had both ischemic heart disease and chronic obstructive pulmonary disease history). Postoperative follow-up took place every 6 months up to 5 years, and annually thereafter. Standard clinical practice included evaluation by physical examination, laboratory tests including tumor markers, imaging, and endoscopy. Recurrences were classified as locoregional, peritoneal, hematogenous, or lymph node metastasis. Locoregional recurrence was defined as the presence of recurrent tumors in the gastrectomy bed, anastomosis, or adjacent organ. Peritoneal seeding and Krukenberg tumors were classified as peritoneal recurrence. Hematogenous recurrence included recurrence in the liver, lung, bone, brain, or other distant sites. Metastases in regional and/or retroperitoneal lymph nodes or mesenteric/ para-aortic lymph nodes were categorized as lymph node metastasis. The median period of follow-up was 74.9 months (range 2.9–158.4 months). None of the patients without LNM received adjuvant chemotherapy. Of 103 patients with pN1 stage, 12 received adjuvant chemotherapy. Systemic chemotherapy was primarily recommended in all patients with recurrent disease. Sometimes, palliative bypass or curative resection was performed to alleviate symptoms related to recurrence such as obstruction or bleeding. Best supportive care was recommended when the patient had progressive disease which did not respond to chemotherapy any more or could not continue to receive the chemotherapy due to side effects of that. Overall survival was calculated from the day of surgical resection to the time of death, and recurrence-free survival from the day of surgical resection to the time of recurrence. Statistical analysis Statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). Descriptive data were presented as means ± standard deviations (SDs). The Chi– square test and independent Student’s t test were used for comparisons between groups. Overall and recurrence-free survival probabilities were estimated using the Kaplan– Meier method and were compared with the log-rank test.
123
734
World J Surg (2015) 39:732–739
To minimize errors caused by many censored cases due to too long follow-up period, the time of last follow-up was limited to 84 months when we performed survival analysis. Cox proportional hazards regression was used to adjust the results for other factors of prognostic importance. Data were considered statistically significant at p B 0.05.
Results Comparison of clinicopathologic characteristics (Table 1) LVI was observed in 18.9 % (111/586) of patients without LNM. Of these, 99 had lymphatic vessel invasion, 7 had
venous invasion, and 5 had both. Tumor size C4 cm, presence of perineural invasion, and deeper depth of tumor invasion were significantly associated with LVI. Of onehundred and three patients with pN1 stage, 85 had lymphatic vessel invasion, three had both lymphatic and venous invasion, and 12 did not have both. The mean number of harvested lymph nodes was 33.4 ± 10 for pN0/ LVI(-) group, 35.2 ± 11.0 for pN0/LVI(?) group, and 34.8 ± 12.0 for pN1 group (there was no statistical difference between the study groups). When patients with pN0/LVI(?) were compared with those with pN1, there was no significant difference in clinicopathologic characteristics, except for histologic grade. Areas of undifferentiated tumor were more common in node-positive patients than in node-negative and LVI(?) patients.
Table 1 Comparison of clinicopathologic features between study groups
Age (years ± SD)
N0/LVI(-) n (%)
N0/LVI(?) n (%)
p*
N1 n (%)
p
57.19 ± 11.69
58.98 ± 11.39
0.144
58.39 ± 11.82
0.709
324 (68.2)
74 (66.7)
Sex Male Female Tumor size (size ± SD)
0.754 151 (31.8)
37 (33.3)
2.88 ± 2.02
4.10 ± 2.33
Tumor location L or/and M U or UM
0.803 67 (65.0) 36 (35.0)
\0.001
4.36 ± 2.19
0.082 436 (91.8) 39 (8.2)
96 (86.5) 15 (13.5)
\D2
19 (4.0)
9 (8.1)
D2
456 (96.0)
102 (91.9)
LND
91 (88.3) 12 (11.7) 0.068
Histologic grade
0.518 11 (10.7) 92 (89.3)
0.442
Differentiated
272 (57.3)
68 (61.3)
Undifferentiated
203 (42.7)
43 (38.7)
Lymphatic
0.01 45 (43.7) 58 (56.3)
NA
0.17
Negative
–
7 (6.3)
12 (11.7)
Positive
–
104 (93.7)
91 (88.3)
Negative
–
99 (89.2)
Positive
–
12 (10.8)
Venous invasion
NA
0.189 97 (94.2) 6 (5.8)
\0.001
Perineural invasion Negative Positive
447 (94.1) 25 (5.3)
72 (64.9) 34 (30.6)
Unknown
3 (0.6)
5 (4.5)
0.294 69 (67.0) 33 (32.0) 1 (1.0)
\0.001
pT stage
0.399 0.682
0.123
T1
411 (86.5)
60 (54.1)
42 (40.8)
T2
29 (6.1)
24 (21.6)
25 (24.3)
T3
35 (7.4)
27 (24.3)
36 (35.0)
SD standard deviation, LVI lymphatic and/or vascular invasion, LND lymph node dissection, LN lymph node, pT primary tumor, NA not available, L lower 1/3, M middle 1/3, U upper 1/3 * p value between N0/LVI(-) versus N0/LVI(?); p value between N0/LVI(?) versus N1
123
World J Surg (2015) 39:732–739
Prognostic value of LVI in gastric cancer Tumor recurrence occurred in 11 patients (2.3 %) from the pN0/LVI(-) group: local in 2 patients (18.2 %), lymph nodes in 2 (18.2 %), hematogenous in 5 (45.5 %), and peritoneal in 5 (45.5 %) (Two patients had mixed pattern of recurrence. One patient had metastatic lymph nodes of porta hepatis and para-aortic area and liver metastasis. In another patient, the tumor recurred in liver, gastrectomy bed, and portocaval and para-aortic lymph nodes). In the pN0/LVI(?) group, 11 patients (9.9 %) had recurrence: local in 1 patients (9.1 %), lymph nodes in 4 (36.4 %), hematogenous in 3 (27.3 %), and peritoneal in 6 (54.5 %) (Two patients had mixed pattern of recurrence. One patient had metastatic tumor on liver and para-aortic lymph nodes and another patient on peritoneum, liver, and retropancreatic, superior mesentery, and small bowel mesentery lymph nodes). Sixteen patients (15.1 %) from pN1 group experienced tumor recurrence: local in 3 patients (18.8 %), lymph nodes in 7 (43.8 %), hematogenous in 3 (18.8 %), and peritoneal in 5 (31.3 %) (One patient had recurred tumor on aortocaval lymph nodes, peritoneum and lung). There were no statistically significant differences in the pattern of recurrences between the groups (p = 0.559 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.566 between patients with pN0/LVI(?) and those with pN1).
735
Overall survival rates (p = 0.022) and recurrence-free survival rates (p \ 0.001) were significantly different between the pN0/LVI(-) and the pN1 groups. No significant differences were observed in the overall survival rates between patients with pN0/LVI(?) and those with pN0/ LVI(-) or pN1 (5-year survival rate, 91.6 vs. 92.7 vs. 85.7 %, respectively; p = 0.130 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.537 between patients with pN0/LVI(?) and those with pN1) (Fig. 1a). On the other hand, there was a significant difference in recurrence-free survival between patients with pN0/LVI(-) and those with pN0/LVI(?) (5-year recurrence-free survival rate, 97.6 vs. 91.6 %, p \ 0.001) and no significant difference between the pN0/LVI(?) and pN1 groups (5-year recurrence-free survival rate, 91.6 vs. 86.6 % p = 0.216) (Fig. 1b). After the factors related to recurrence-free survival were identified by univariate analysis, multivariate analysis was performed to adjust the results for the factors with prognostic impact. Among node-negative patients with or without LVI, the presence of LVI invasion was an independent prognostic factor related to recurrence-free survival, along with the extent of lymphadenectomy and pT stage (Table 2). In patients with pN0/LVI(?) or pN1, the extent of lymphadenectomy and pT stage had independent effects on recurrence-free survival. Lymph node status (N0 vs. N1) did not influence recurrence-free survival in the multivariate analysis (Table 3).
Fig. 1 a Overall survival: p = 0.130 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.537 between patients with pN0/ LVI(?) and those with pN1; p = 0.022 between patients with pN0/LVI(-) and those with pN1. b Recurrence-free survival: p \ 0.001 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.216 between patients with pN0/LVI(?) and those with pN1; p \ 0.001 between patients with pN0/LVI(-) and those with pN1
123
736
World J Surg (2015) 39:732–739
Table 2 Univariate analysis and Multivariate analysis of recurrence-free survival in node-negative patients
Univariate analysis
Multivariate analysis
5-RFS (%)
Hazard ratio
LND
1.253-12.490
0.019
0.212-2.335
0.565
3.322
1.312-8.412
0.011
1.000 4.316
1.573-11.844
0.005
0.001 1.000
\D2
84.6
3.958 0.007
Negative
97.0
Positive
91.4
LVI
5-RFS 5-year recurrence-free survival rate, LND lymphnode dissection, LVI lymphovascular invasion
pT
1.000 0.703 0.001
No
97.6
Yes
91.6
1.000 \0.001
T1-2 T3
Table 3 Univariate analysis and Multivariate analysis of recurrence-free survival in N0/ LVI(?) and N1 gastric cancer groups
97.4 88.3
Univariate analysis
Multivariate analysis
5-RFS (%)
Hazard ratio
p
95 % CI
p
2.11413.300
\0.001
\0.001
LND D2
91.1
1
\D2
59.7
5.303 \0.001
pT T12
93
1.000
T3
80.4
3.515
Prognostic factors in patients with N0/LVI(?) gastric cancer Multivariate analysis of factors related to recurrence-free survival in patients with N0/LVI(?) gastric cancer identified less than D2 lymphadenectomy and pT3 stage as prognostic factors for poor recurrence-free survival (Table 4). Less than D2 lymphadenectomy had an independent negative effect on the overall survival in those patients (Table 5).
Discussion Although the survival rate has improved due to early detection, radical lymphadenectomy, and modified therapeutic modalities, gastric cancer is still a major cause of cancer death [20]. Therefore, adjuvant chemotherapy is widely accepted as a standard treatment for gastric cancers with a high risk of recurrence, since there is concrete evidence to support its use in stage II and III gastric cancer [17, 18]. Many authors reported that LVI was an
123
p
97.0
We also evaluated age, sex, tumor location, tumor size,and histologic grade and found no statistical differences
5-RFS 5 year recurrence free survival rate, LND lymphnode dissection
95 % CI
D2 Perineural invasion
We also evaluated age, sex, perineural invasion, tumor location, tumor size, histologic grade and N stage[N0/LVI(?) vs. N1] and found no statistical differences
p
1.613-7.657
0.002
independent prognostic factor for LNM and an independent prognostic factor in patients with gastric cancer [5, 9–14]. In the present study, the prognostic effect of LVI was evaluated in patients with pT1–3N0 tumors, for which adjuvant chemotherapy is not recommended, in order to develop a more effective treatment plan for patients with pN0/LVI(?). LVI has been shown to be a risk factor for LNM in many studies. The current study focused on the prognostic pure impact of LVI, in order to determine whether LVI can be viewed in the same way as LNM. pN0/LVI(?) tumors had more aggressive clincopathologic characteristics than pN0/ LVI(-) tumors. In patients with pN0 stage tumors, multivariate analysis showed that LVI was an independent prognostic factor related to recurrence-free survival, while N stage (N0 vs. N1) lost significance as a prognostic factor for recurrence-free survival in patients with pN0/LVI(?) or pN1 tumors. (We focused more on the recurrence-free survival than overall survival because death rate unrelated to gastric cancer or that of unknown origin was higher than expected. It was probably because relatively early stage gastric cancer was included in this study.) This means that
World J Surg (2015) 39:732–739 Table 4 Prognostic factors related to recurrence-free survival in patients with N0/ LVI(?) gastric cancer
737
5-RFS (%) Age, years 91
[65
93.3
Sex 94.3
Male
90.2
D2
94
\D2
57.1 91.7 90.2
L or/and M
92.6
U
83.3
0.234-4.442
0.714
0.765
0.182-3.207
0.016
8.179
1.472-45.423
0.451
0.536
0.106-2.714
0.394
0.339
0.028-4.070
0.143
0.283
0.052-1.535
0.837
1.27
0.131-12.337
0.733
1.576
0.115-21.542
0.796
1.426
0.097-20.890
0.014
8.212
1.520-44.376
0.703
B4
90.4
[4
94.1
Histology
0.407
Differentiated
94.8
Undifferentiated
87.7
Lauren
Logistic regression
1.019
0.41
Tumor size, cm
0.98
0.92
Tumor location
Chi-square test
95 % CI
\0.001
Perineural invasion
*
Hazard ratio
0.848
Female LND
5-RFS 5-year recurrence-free survival rate, LND lymph-node dissection, L lower 1/3, M middle 1/3, U upper 1/3, pT primary tumor, 95 % CI 95 % confidence interval
p-value (multivariate )
0.95
B65
Negative Positive
p-value (univariate*)
0.806
Diffuse
89.2
Intestinal
94.6
Mixed
85.7
pT T1-2
0.013 93.8
T3
85.2
even patients with N0 stage tumors may have the similar prognosis as those with N1 stage if they have LVI. However, the overall survival and recurrence-free survival curve of the pN0/LVI(?) group was roughly intermediate between the pN0/LVI(-) and pN1 groups on Fig. 1. A further large-scale study is needed to confirm this result. In our study, there was no significant difference in clinicopathologic features except for histologic grade between patients with pN0/LVI(?) and those with pN1 stage, pointing to similar aggressiveness of the tumors. Despite the difference in pN stage, there was no statistically significant difference in overall survival rate and recurrence-free survival rate between the two groups. Liu et al. [9] also found no significant difference between patients with T1N0M0 gastric cancer with lymphatic vessel invasion and those with T1N1M0 gastric cancer when their cumulative survival was compared. However, these findings should be interpreted with caution because some patients (n = 12) with pN1 stage received adjuvant
chemotherapy in our study. They need to be confirmed in an adequately designed prospective study. In our study, the most common pattern of recurrence was peritoneal recurrence in pN0/LVI(?) and followed by recurrence on lymph nodes. On the other hand, recurrence on lymph nodes was the most common pattern of recurrence in patients with pN1 and followed by peritoneal recurrence. Kunisaki et al. [12] and Hyung et al. [13] reported that the recurrence pattern of node-negative AGC with LVI was mainly hematogenous. This difference might be due to different patient population in each study. In addition, previous studies did not investigate the rate of lymph node recurrence separately. In our study, the rate of recurrence on lymph nodes was more common in pN0/ LVI(?) than in pN0/LVI(-) (36.4 vs. 18.2 %), although there was no statistical difference. Moreover, the second most common pattern of recurrence was hematogenous in pN0/LVI(-) group and lymph node recurrence in pN0/ LVI(?) aside from peritoneal seeding, suggesting different
123
738 Table 5 Prognostic factors related to overall survival in patients with N0/LVI(?) gastric cancer
World J Surg (2015) 39:732–739
5-OS (%) Age, years 91.9
[65
91.2
Gender 89
Male
93.1
D2
95.2
\D2
38.1
Perineural invasion 91.3 91
Logistic regression
B4
90.4
[4
94.4
0.665
0.182-2.428
0.002
9.829
2.242-43.099
0.076
0.232
0.049-3.910
0.461
0.439
0.049-3.910
0.768
1.21
0.340-4.309
0.122
0.204
0.027-1.533
0.531
2.216
0.184-26.632
0.892
0.857
0.092-7.985
0.076
3.43
0.880-13.375
0.087
Differentiated
87.9
Undifferentiated
97.6 0.627
Diffuse
93.5
Intestinal
88.8 100
pT T1-2
0.113 92.7
T3
88.3
pattern of recurrence between the two groups. After more data collection, these results should be validated again. In patients with pN0/LVI(?), pT3 stage and performance of less than D2 lymphadenectomy were identified as poor prognostic factors. Thus, patients with stage pT3N0 gastric cancer, or node-negative patients who receive less than D2 dissection and who have LVI on permanent pathology results, may benefit from adjuvant treatment, such as chemotherapy. The effectiveness of adjuvant treatment in those patients should be addressed in future studies. Also, D2 dissection should be performed if gastric cancer patients without LNM have LVI. However, pathologic examination results using endoscopic biopsy are unsatisfactory for accurate preoperative identification of LVI due to the inadequate size of the specimens. Therefore, it is not easy to apply this finding clinically. Gastric cancer patients with LVI following endoscopic resection are often referred to the surgical department for gastrectomy. D2 lymphadenectomy should be performed in such cases even though the patients have no clinical LNM.
123
0.537
0.749
Lauren
0.257-2.887
90.3 100
Histology
* Chi-square test
0.862
0.399
Tumor size, cm
Mixed
0.809
0.51
Tumor location
5-OS 5-year overall survival rate, LND lymph-node dissection, L lower 1/3, M middle 1/3, U upper 1/3, pT primary tumor, 95 % CI 95 % confidence interval
95 % CI
\0.001
LND
U
Hazard ratio
0.688
Female
L or/and M
p-value (multivariate )
0.399
B65
Negative Positive
p-value (univariate*)
There have been reports of immunohistochemical staining combining cytokeratin and vascular markers including CD31, CD34, and D2-40 being sensitive in detecting LVI than standardized H&E alone [21, 22]. However, LVI was diagnosed by H&E staining only in our study because the immunohistochemical staining was not routinely performed for all patients. This situation is probably similar in other centers. Moreover, the issue of cost-effectiveness should also be addressed if such an approach were to be used in routine diagnostic work. Therefore, our results can be generally applied to clinical practice because of these points. In conclusion, it is anticipated that individualized and comprehensive treatment plans will be in future development for patients with gastric cancer. Analysis of the prognostic effect of LVI and the related prognostic factors for N0/LVI(?) gastric cancer showed that LVI was an independent negative prognostic factor for node-negative patients and that the prognosis was similar with that of the N1 group. In addition, less than D2 lymphadenectomy and
World J Surg (2015) 39:732–739
pT3 stage were identified as poor prognostic factors in N0/ LVI(?) patients. Clinicians should take these factors into account in assessing prognosis and developing the optimum treatment plan for patients with N0/LVI(?).
Conflict of interest
We have no conflicts of interest.
References 1. Jemal A, Bray F, Center MM et al (2011) Global cancer statistics. CA Cancer J Clin 61(2):69–90 2. Kim JP, Lee JH, Kim SJ et al (1998) Clinicopathologic characteristics and prognostic factors in 10783 patients with gastric cancer. Gastric Cancer 1(2):125–133 3. Edge SB, Byrd DR, Compton CC et al (2010) American Joint Committee on Cancer, American Cancer Society: AJCC cancer staging manual, 7th edn. Springer, New York 4. Japanese Gastric Cancer Association (2011) Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric Cancer 14(2):113–123 5. Borie F, Millat B, Fingerhut A et al (2000) Lymphatic involvement in early gastric cancer: prevalence and prognosis in France. Arch Surg 135(10):1218–1223 6. Chen R, He Q, Cui J et al (2014) Lymph node metastasis in early gastric cancer. Chin Med J (Engl) 127(3):560–567 7. Gotoda T, Yanagisawa A, Sasako M et al (2000) Incidence of lymph node metastasis from early gastric cancer: estimation with a large number of cases at two large centers. Gastric Cancer 3(4):219–225 8. An JY, Baik YH, Choi MG et al (2007) Predictive factors for lymph node metastasis in early gastric cancer with submucosal invasion: analysis of a single institutional experience. Ann Surg 246(5):749–753 9. Liu C, Zhang R, Lu Y et al (2010) Prognostic role of lymphatic vessel invasion in early gastric cancer: a retrospective study of 188 cases. Surg Oncol 19(1):4–10 10. Ichikawa D, Kubota T, Kikuchi S et al (2009) Prognostic impact of lymphatic invasion in patients with node-negative gastric cancer. J Surg Oncol 100(2):111–114
739 11. Liu E, Zhong M, Xu F et al (2011) Impact of lymphatic vessel invasion on survival in curative resected gastric cancer. J Gastrointest Surg 15(9):1526–1531 12. Kunisaki C, Makino H, Kimura J et al (2010) Impact of lymphovascular invasion in patients with stage I gastric cancer. Surgery 147(2):204–211 13. Hyung WJ, Lee JH, Choi SH et al (2002) Prognostic impact of lymphatic and/or blood vessel invasion in patients with nodenegative advanced gastric cancer. Ann Surg Oncol 9(6):562–567 14. Bu Z, Zheng Z, Li Z et al (2013) Lymphatic vascular invasion is an independent correlated factor for lymph node metastasis and the prognosis of resectable T2 gastric cancer patients. Tumour Biol 34(2):1005–1012 15. Yokota T, Kunii Y, Teshima S, et al (1999) Significant prognostic factors in patients with node-negative gastric cancer. Int Surg 84(4):331–336 16. Lee IS, Yook JH, Kim TH et al (2013) Prognostic factors and recurrence pattern in node-negative advanced gastric cancer. Eur J Surg Oncol 39(2):136–140 17. Sakuramoto S, Sasako M, Yamaguchi T et al (2007) Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine. N Engl J Med 357(18):1810–1820 18. Bang YJ, Kim YW, Yang HK et al (2012) Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial. Lancet 379(9813):315–321 19. Japanese Gastric Cancer Association (1998) Japanese classification of gastric carcinoma—2nd english edition. Gastric Cancer 1:10–24 20. Shen L, Shan YS, Hu HM et al (2013) Management of gastric cancer in Asia: resource-stratified guidelines. Lancet Oncol 14(12):e535–e547 21. Lim CS, Alexander-Sefre F, Allam M et al (2008) Clinical value of immunohistochemically detected lymphovascular space invasion in early stage cervical carcinoma. Ann Surg Oncol 15:2581–2588 22. O’Donnell RK, Feldman M, Mick R et al (2008) Immunohistochemical method identifies lymphovascular invasion in a majority of oral squamous cell carcinomas and discriminates between blood and lymphatic vessel invasion. J Histochem Cytochem 56:803–810
123
ORIGINAL SCIENTIFIC REPORT
Prognostic Significance of Lymphovascular Invasion in Node-Negative Gastric Cancer Ju-Hee Lee • Min Gyu Kim • Min-Sung Jung Sung Joon Kwon
•
Published online: 7 November 2014 Ó Socie´te´ Internationale de Chirurgie 2014
Abstract Background The aim of this study was to investigate the prognostic value of lymphovascular invasion (LVI) in gastric cancer patients without lymph node metastasis. Methods A total of 699 patients with primary tumor pT1–3/pN0-1 gastric adenocarcinoma who underwent curative resection from 2001 to 2010 were categorized into 3 groups: One-hundred and eleven patients with pN0/LVI(?), 475 with pN0/LVI(-), and 103 with pN1. Results The tumors in patients with N0/LVI(?) had more aggressive clinicopathologic features than those in patients with N0/LVI(-). However, there was no significant difference in patient characteristics between patients with pN0/LVI(?) and those with pN1, except for histologic grade. There were no significant differences in the overall survival rate in patients with pN0/LVI(?) compared to those with pN0/LVI(-) or the pN1 stage. However, the recurrence-free survival rate of the pN0/LVI(?) group was lower than that of the pN0/LVI(-) group (p \ 0.001), while no significant difference was observed between the pN0/LVI(?) and the N1 groups (p = 0.216). In multivariate analysis, LVI was identified as a poor prognostic factor related to recurrence-free survival in node-negative gastric cancer patients. pT3 stage and less than D2 lymphadenectomy were poor prognostic factors affecting recurrence-free survival, and less than D2 lymphadenectomy was an independent poor prognostic factor for overall survival in pN0/LVI(?) patients. Conclusions LVI could be an indicator of biological aggressiveness and may be a reliable prognostic factor for node-negative gastric cancer. LVI should be considered in postoperative management of gastric cancer.
J.-H. Lee M.-S. Jung S. J. Kwon (&) Department of Surgery, Hanyang University Seoul Hospital, Seoul, Republic of Korea e-mail: [email protected] J.-H. Lee e-mail: [email protected] M. G. Kim Department of Surgery, Hanyang University Guri Hospital, Seoul, Republic of Korea M. G. Kim M.-S. Jung S. J. Kwon Department of Surgery, College of Medicine, Hanyang University, 222 Wangsimri-ro, Seongdong-gu, Seoul, Republic of Korea
123
Introduction Gastric cancer is the fourth most common cancer in the world and the second most common cause of cancer-related deaths worldwide [1]. The involvement of regional lymph nodes (N stage) is the most significant prognostic indicators for gastric cancer [2–4]. For early gastric cancer (EGC), in particular, it seems that LNM is a critical factor for prognosis and choice of treatment because the overall survival rate for EGC without lymph node metastasis (LNM) is 90–95 %, but this decreases to 70–85 % in the case of patients with LNM [5, 6].
World J Surg (2015) 39:732–739
Lymphovascular invasion (LVI) is a well-known risk factor for LNM [7, 8]. The prognostic effect of LVI has been previously investigated by many authors [5, 9–16]. However, the results are not consistent. Moreover, several studies have evaluated the role of LVI in gastric cancer, regardless of the lymph node status of the gastric cancer patients, and have shown a close correlation between lymph node metastasis and LVI. Studies to the pure prognostic effect of LVI are still rare. Recently, chemotherapy was shown to be an effective adjuvant treatment for stages II and III (American Joint Committee on Cancer/Union for International Cancer Control; AJCC/UICC 6th edition) gastric cancer [17, 18], but adjuvant treatment is not recommended for EGC and node-negative advanced gastric cancer (AGC) BT3 (subserosa invasion) according to recent Japanese gastric cancer treatment guidelines, because of lack of evidence [4]. Therefore, this study focused on T1–3N0 gastric cancer patients and tried to clarify the prognostic role of LVI in those patients, in order to determine its possible value in management of the disease.
Materials and methods Patients A total of 700 patients were selected from prospectively maintained gastric cancer surgery databases according to the following criteria: (i) pathologically proven gastric adenocarcinoma, (ii) primary tumor stage (pT)1–3 and regional lymph node (pN)0–1 according to AJCC/UICC 7th edition, (iii) no history of other therapy for gastric cancer (including neoadjuvant chemotherapy), and (iv) patients who underwent curative gastrectomy between 2001 and 2010 at the Hanyang University Medical Center (patients with 15 or more than 15 retrieved lymph nodes). Of those 700 patients, eleven for whom data on LVI were not available were excluded. The resected specimens and retrieved lymph nodes were stained with hematoxylin and eosin (H&E) and examined by specialized pathologists. Lymphatic invasion was defined as the presence of neoplastic cell emboli within endothelium-lined spaces. Venous invasion was identified by the presence of additional fibrin clots and/or erythrocytes in endothelium-lined spaces without erythrocyte extravasation into the surrounding tissue or by evidence of neoplastic cells within smooth muscle cell-lined spaces. Tumors were classified into three groups: 475 pN0/LVI (-), 111 pN0/LVI (?), and 103 pN1. Clinicopathologic features studied included sex, age, tumor location, extent of lymphadenectomy, tumor size, histological grade, lymphatic invasion, venous invasion,
733
perineural invasion, and tumor stage (AJCC/UICC 7th edition). Extent of lymph node dissection was classified according to the guidelines of the Japanese Gastric Cancer Association [19]. D2 lymphadenectomy was performed in most of the patients with EGC and AGC. Less than D2 lymphadenectomy (D1 ? a or D1 ? b) was performed in some patients with clinical early gastric cancer without LNM diagnosed preoperatively (n = 30) or comorbidities related to a high complication rate (n = 9). The comorbidities were heart disease in four patients (three patients had ischemic heart disease and one had heart failure), cerebrovascular accident in three, liver cirrhosis in two, and chronic obstructive pulmonary disease in 1 (one patient had both ischemic heart disease and chronic obstructive pulmonary disease history). Postoperative follow-up took place every 6 months up to 5 years, and annually thereafter. Standard clinical practice included evaluation by physical examination, laboratory tests including tumor markers, imaging, and endoscopy. Recurrences were classified as locoregional, peritoneal, hematogenous, or lymph node metastasis. Locoregional recurrence was defined as the presence of recurrent tumors in the gastrectomy bed, anastomosis, or adjacent organ. Peritoneal seeding and Krukenberg tumors were classified as peritoneal recurrence. Hematogenous recurrence included recurrence in the liver, lung, bone, brain, or other distant sites. Metastases in regional and/or retroperitoneal lymph nodes or mesenteric/ para-aortic lymph nodes were categorized as lymph node metastasis. The median period of follow-up was 74.9 months (range 2.9–158.4 months). None of the patients without LNM received adjuvant chemotherapy. Of 103 patients with pN1 stage, 12 received adjuvant chemotherapy. Systemic chemotherapy was primarily recommended in all patients with recurrent disease. Sometimes, palliative bypass or curative resection was performed to alleviate symptoms related to recurrence such as obstruction or bleeding. Best supportive care was recommended when the patient had progressive disease which did not respond to chemotherapy any more or could not continue to receive the chemotherapy due to side effects of that. Overall survival was calculated from the day of surgical resection to the time of death, and recurrence-free survival from the day of surgical resection to the time of recurrence. Statistical analysis Statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). Descriptive data were presented as means ± standard deviations (SDs). The Chi– square test and independent Student’s t test were used for comparisons between groups. Overall and recurrence-free survival probabilities were estimated using the Kaplan– Meier method and were compared with the log-rank test.
123
734
World J Surg (2015) 39:732–739
To minimize errors caused by many censored cases due to too long follow-up period, the time of last follow-up was limited to 84 months when we performed survival analysis. Cox proportional hazards regression was used to adjust the results for other factors of prognostic importance. Data were considered statistically significant at p B 0.05.
Results Comparison of clinicopathologic characteristics (Table 1) LVI was observed in 18.9 % (111/586) of patients without LNM. Of these, 99 had lymphatic vessel invasion, 7 had
venous invasion, and 5 had both. Tumor size C4 cm, presence of perineural invasion, and deeper depth of tumor invasion were significantly associated with LVI. Of onehundred and three patients with pN1 stage, 85 had lymphatic vessel invasion, three had both lymphatic and venous invasion, and 12 did not have both. The mean number of harvested lymph nodes was 33.4 ± 10 for pN0/ LVI(-) group, 35.2 ± 11.0 for pN0/LVI(?) group, and 34.8 ± 12.0 for pN1 group (there was no statistical difference between the study groups). When patients with pN0/LVI(?) were compared with those with pN1, there was no significant difference in clinicopathologic characteristics, except for histologic grade. Areas of undifferentiated tumor were more common in node-positive patients than in node-negative and LVI(?) patients.
Table 1 Comparison of clinicopathologic features between study groups
Age (years ± SD)
N0/LVI(-) n (%)
N0/LVI(?) n (%)
p*
N1 n (%)
p
57.19 ± 11.69
58.98 ± 11.39
0.144
58.39 ± 11.82
0.709
324 (68.2)
74 (66.7)
Sex Male Female Tumor size (size ± SD)
0.754 151 (31.8)
37 (33.3)
2.88 ± 2.02
4.10 ± 2.33
Tumor location L or/and M U or UM
0.803 67 (65.0) 36 (35.0)
\0.001
4.36 ± 2.19
0.082 436 (91.8) 39 (8.2)
96 (86.5) 15 (13.5)
\D2
19 (4.0)
9 (8.1)
D2
456 (96.0)
102 (91.9)
LND
91 (88.3) 12 (11.7) 0.068
Histologic grade
0.518 11 (10.7) 92 (89.3)
0.442
Differentiated
272 (57.3)
68 (61.3)
Undifferentiated
203 (42.7)
43 (38.7)
Lymphatic
0.01 45 (43.7) 58 (56.3)
NA
0.17
Negative
–
7 (6.3)
12 (11.7)
Positive
–
104 (93.7)
91 (88.3)
Negative
–
99 (89.2)
Positive
–
12 (10.8)
Venous invasion
NA
0.189 97 (94.2) 6 (5.8)
\0.001
Perineural invasion Negative Positive
447 (94.1) 25 (5.3)
72 (64.9) 34 (30.6)
Unknown
3 (0.6)
5 (4.5)
0.294 69 (67.0) 33 (32.0) 1 (1.0)
\0.001
pT stage
0.399 0.682
0.123
T1
411 (86.5)
60 (54.1)
42 (40.8)
T2
29 (6.1)
24 (21.6)
25 (24.3)
T3
35 (7.4)
27 (24.3)
36 (35.0)
SD standard deviation, LVI lymphatic and/or vascular invasion, LND lymph node dissection, LN lymph node, pT primary tumor, NA not available, L lower 1/3, M middle 1/3, U upper 1/3 * p value between N0/LVI(-) versus N0/LVI(?); p value between N0/LVI(?) versus N1
123
World J Surg (2015) 39:732–739
Prognostic value of LVI in gastric cancer Tumor recurrence occurred in 11 patients (2.3 %) from the pN0/LVI(-) group: local in 2 patients (18.2 %), lymph nodes in 2 (18.2 %), hematogenous in 5 (45.5 %), and peritoneal in 5 (45.5 %) (Two patients had mixed pattern of recurrence. One patient had metastatic lymph nodes of porta hepatis and para-aortic area and liver metastasis. In another patient, the tumor recurred in liver, gastrectomy bed, and portocaval and para-aortic lymph nodes). In the pN0/LVI(?) group, 11 patients (9.9 %) had recurrence: local in 1 patients (9.1 %), lymph nodes in 4 (36.4 %), hematogenous in 3 (27.3 %), and peritoneal in 6 (54.5 %) (Two patients had mixed pattern of recurrence. One patient had metastatic tumor on liver and para-aortic lymph nodes and another patient on peritoneum, liver, and retropancreatic, superior mesentery, and small bowel mesentery lymph nodes). Sixteen patients (15.1 %) from pN1 group experienced tumor recurrence: local in 3 patients (18.8 %), lymph nodes in 7 (43.8 %), hematogenous in 3 (18.8 %), and peritoneal in 5 (31.3 %) (One patient had recurred tumor on aortocaval lymph nodes, peritoneum and lung). There were no statistically significant differences in the pattern of recurrences between the groups (p = 0.559 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.566 between patients with pN0/LVI(?) and those with pN1).
735
Overall survival rates (p = 0.022) and recurrence-free survival rates (p \ 0.001) were significantly different between the pN0/LVI(-) and the pN1 groups. No significant differences were observed in the overall survival rates between patients with pN0/LVI(?) and those with pN0/ LVI(-) or pN1 (5-year survival rate, 91.6 vs. 92.7 vs. 85.7 %, respectively; p = 0.130 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.537 between patients with pN0/LVI(?) and those with pN1) (Fig. 1a). On the other hand, there was a significant difference in recurrence-free survival between patients with pN0/LVI(-) and those with pN0/LVI(?) (5-year recurrence-free survival rate, 97.6 vs. 91.6 %, p \ 0.001) and no significant difference between the pN0/LVI(?) and pN1 groups (5-year recurrence-free survival rate, 91.6 vs. 86.6 % p = 0.216) (Fig. 1b). After the factors related to recurrence-free survival were identified by univariate analysis, multivariate analysis was performed to adjust the results for the factors with prognostic impact. Among node-negative patients with or without LVI, the presence of LVI invasion was an independent prognostic factor related to recurrence-free survival, along with the extent of lymphadenectomy and pT stage (Table 2). In patients with pN0/LVI(?) or pN1, the extent of lymphadenectomy and pT stage had independent effects on recurrence-free survival. Lymph node status (N0 vs. N1) did not influence recurrence-free survival in the multivariate analysis (Table 3).
Fig. 1 a Overall survival: p = 0.130 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.537 between patients with pN0/ LVI(?) and those with pN1; p = 0.022 between patients with pN0/LVI(-) and those with pN1. b Recurrence-free survival: p \ 0.001 between patients with pN0/LVI(-) and those with pN0/LVI(?); p = 0.216 between patients with pN0/LVI(?) and those with pN1; p \ 0.001 between patients with pN0/LVI(-) and those with pN1
123
736
World J Surg (2015) 39:732–739
Table 2 Univariate analysis and Multivariate analysis of recurrence-free survival in node-negative patients
Univariate analysis
Multivariate analysis
5-RFS (%)
Hazard ratio
LND
1.253-12.490
0.019
0.212-2.335
0.565
3.322
1.312-8.412
0.011
1.000 4.316
1.573-11.844
0.005
0.001 1.000
\D2
84.6
3.958 0.007
Negative
97.0
Positive
91.4
LVI
5-RFS 5-year recurrence-free survival rate, LND lymphnode dissection, LVI lymphovascular invasion
pT
1.000 0.703 0.001
No
97.6
Yes
91.6
1.000 \0.001
T1-2 T3
Table 3 Univariate analysis and Multivariate analysis of recurrence-free survival in N0/ LVI(?) and N1 gastric cancer groups
97.4 88.3
Univariate analysis
Multivariate analysis
5-RFS (%)
Hazard ratio
p
95 % CI
p
2.11413.300
\0.001
\0.001
LND D2
91.1
1
\D2
59.7
5.303 \0.001
pT T12
93
1.000
T3
80.4
3.515
Prognostic factors in patients with N0/LVI(?) gastric cancer Multivariate analysis of factors related to recurrence-free survival in patients with N0/LVI(?) gastric cancer identified less than D2 lymphadenectomy and pT3 stage as prognostic factors for poor recurrence-free survival (Table 4). Less than D2 lymphadenectomy had an independent negative effect on the overall survival in those patients (Table 5).
Discussion Although the survival rate has improved due to early detection, radical lymphadenectomy, and modified therapeutic modalities, gastric cancer is still a major cause of cancer death [20]. Therefore, adjuvant chemotherapy is widely accepted as a standard treatment for gastric cancers with a high risk of recurrence, since there is concrete evidence to support its use in stage II and III gastric cancer [17, 18]. Many authors reported that LVI was an
123
p
97.0
We also evaluated age, sex, tumor location, tumor size,and histologic grade and found no statistical differences
5-RFS 5 year recurrence free survival rate, LND lymphnode dissection
95 % CI
D2 Perineural invasion
We also evaluated age, sex, perineural invasion, tumor location, tumor size, histologic grade and N stage[N0/LVI(?) vs. N1] and found no statistical differences
p
1.613-7.657
0.002
independent prognostic factor for LNM and an independent prognostic factor in patients with gastric cancer [5, 9–14]. In the present study, the prognostic effect of LVI was evaluated in patients with pT1–3N0 tumors, for which adjuvant chemotherapy is not recommended, in order to develop a more effective treatment plan for patients with pN0/LVI(?). LVI has been shown to be a risk factor for LNM in many studies. The current study focused on the prognostic pure impact of LVI, in order to determine whether LVI can be viewed in the same way as LNM. pN0/LVI(?) tumors had more aggressive clincopathologic characteristics than pN0/ LVI(-) tumors. In patients with pN0 stage tumors, multivariate analysis showed that LVI was an independent prognostic factor related to recurrence-free survival, while N stage (N0 vs. N1) lost significance as a prognostic factor for recurrence-free survival in patients with pN0/LVI(?) or pN1 tumors. (We focused more on the recurrence-free survival than overall survival because death rate unrelated to gastric cancer or that of unknown origin was higher than expected. It was probably because relatively early stage gastric cancer was included in this study.) This means that
World J Surg (2015) 39:732–739 Table 4 Prognostic factors related to recurrence-free survival in patients with N0/ LVI(?) gastric cancer
737
5-RFS (%) Age, years 91
[65
93.3
Sex 94.3
Male
90.2
D2
94
\D2
57.1 91.7 90.2
L or/and M
92.6
U
83.3
0.234-4.442
0.714
0.765
0.182-3.207
0.016
8.179
1.472-45.423
0.451
0.536
0.106-2.714
0.394
0.339
0.028-4.070
0.143
0.283
0.052-1.535
0.837
1.27
0.131-12.337
0.733
1.576
0.115-21.542
0.796
1.426
0.097-20.890
0.014
8.212
1.520-44.376
0.703
B4
90.4
[4
94.1
Histology
0.407
Differentiated
94.8
Undifferentiated
87.7
Lauren
Logistic regression
1.019
0.41
Tumor size, cm
0.98
0.92
Tumor location
Chi-square test
95 % CI
\0.001
Perineural invasion
*
Hazard ratio
0.848
Female LND
5-RFS 5-year recurrence-free survival rate, LND lymph-node dissection, L lower 1/3, M middle 1/3, U upper 1/3, pT primary tumor, 95 % CI 95 % confidence interval
p-value (multivariate )
0.95
B65
Negative Positive
p-value (univariate*)
0.806
Diffuse
89.2
Intestinal
94.6
Mixed
85.7
pT T1-2
0.013 93.8
T3
85.2
even patients with N0 stage tumors may have the similar prognosis as those with N1 stage if they have LVI. However, the overall survival and recurrence-free survival curve of the pN0/LVI(?) group was roughly intermediate between the pN0/LVI(-) and pN1 groups on Fig. 1. A further large-scale study is needed to confirm this result. In our study, there was no significant difference in clinicopathologic features except for histologic grade between patients with pN0/LVI(?) and those with pN1 stage, pointing to similar aggressiveness of the tumors. Despite the difference in pN stage, there was no statistically significant difference in overall survival rate and recurrence-free survival rate between the two groups. Liu et al. [9] also found no significant difference between patients with T1N0M0 gastric cancer with lymphatic vessel invasion and those with T1N1M0 gastric cancer when their cumulative survival was compared. However, these findings should be interpreted with caution because some patients (n = 12) with pN1 stage received adjuvant
chemotherapy in our study. They need to be confirmed in an adequately designed prospective study. In our study, the most common pattern of recurrence was peritoneal recurrence in pN0/LVI(?) and followed by recurrence on lymph nodes. On the other hand, recurrence on lymph nodes was the most common pattern of recurrence in patients with pN1 and followed by peritoneal recurrence. Kunisaki et al. [12] and Hyung et al. [13] reported that the recurrence pattern of node-negative AGC with LVI was mainly hematogenous. This difference might be due to different patient population in each study. In addition, previous studies did not investigate the rate of lymph node recurrence separately. In our study, the rate of recurrence on lymph nodes was more common in pN0/ LVI(?) than in pN0/LVI(-) (36.4 vs. 18.2 %), although there was no statistical difference. Moreover, the second most common pattern of recurrence was hematogenous in pN0/LVI(-) group and lymph node recurrence in pN0/ LVI(?) aside from peritoneal seeding, suggesting different
123
738 Table 5 Prognostic factors related to overall survival in patients with N0/LVI(?) gastric cancer
World J Surg (2015) 39:732–739
5-OS (%) Age, years 91.9
[65
91.2
Gender 89
Male
93.1
D2
95.2
\D2
38.1
Perineural invasion 91.3 91
Logistic regression
B4
90.4
[4
94.4
0.665
0.182-2.428
0.002
9.829
2.242-43.099
0.076
0.232
0.049-3.910
0.461
0.439
0.049-3.910
0.768
1.21
0.340-4.309
0.122
0.204
0.027-1.533
0.531
2.216
0.184-26.632
0.892
0.857
0.092-7.985
0.076
3.43
0.880-13.375
0.087
Differentiated
87.9
Undifferentiated
97.6 0.627
Diffuse
93.5
Intestinal
88.8 100
pT T1-2
0.113 92.7
T3
88.3
pattern of recurrence between the two groups. After more data collection, these results should be validated again. In patients with pN0/LVI(?), pT3 stage and performance of less than D2 lymphadenectomy were identified as poor prognostic factors. Thus, patients with stage pT3N0 gastric cancer, or node-negative patients who receive less than D2 dissection and who have LVI on permanent pathology results, may benefit from adjuvant treatment, such as chemotherapy. The effectiveness of adjuvant treatment in those patients should be addressed in future studies. Also, D2 dissection should be performed if gastric cancer patients without LNM have LVI. However, pathologic examination results using endoscopic biopsy are unsatisfactory for accurate preoperative identification of LVI due to the inadequate size of the specimens. Therefore, it is not easy to apply this finding clinically. Gastric cancer patients with LVI following endoscopic resection are often referred to the surgical department for gastrectomy. D2 lymphadenectomy should be performed in such cases even though the patients have no clinical LNM.
123
0.537
0.749
Lauren
0.257-2.887
90.3 100
Histology
* Chi-square test
0.862
0.399
Tumor size, cm
Mixed
0.809
0.51
Tumor location
5-OS 5-year overall survival rate, LND lymph-node dissection, L lower 1/3, M middle 1/3, U upper 1/3, pT primary tumor, 95 % CI 95 % confidence interval
95 % CI
\0.001
LND
U
Hazard ratio
0.688
Female
L or/and M
p-value (multivariate )
0.399
B65
Negative Positive
p-value (univariate*)
There have been reports of immunohistochemical staining combining cytokeratin and vascular markers including CD31, CD34, and D2-40 being sensitive in detecting LVI than standardized H&E alone [21, 22]. However, LVI was diagnosed by H&E staining only in our study because the immunohistochemical staining was not routinely performed for all patients. This situation is probably similar in other centers. Moreover, the issue of cost-effectiveness should also be addressed if such an approach were to be used in routine diagnostic work. Therefore, our results can be generally applied to clinical practice because of these points. In conclusion, it is anticipated that individualized and comprehensive treatment plans will be in future development for patients with gastric cancer. Analysis of the prognostic effect of LVI and the related prognostic factors for N0/LVI(?) gastric cancer showed that LVI was an independent negative prognostic factor for node-negative patients and that the prognosis was similar with that of the N1 group. In addition, less than D2 lymphadenectomy and
World J Surg (2015) 39:732–739
pT3 stage were identified as poor prognostic factors in N0/ LVI(?) patients. Clinicians should take these factors into account in assessing prognosis and developing the optimum treatment plan for patients with N0/LVI(?).
Conflict of interest
We have no conflicts of interest.
References 1. Jemal A, Bray F, Center MM et al (2011) Global cancer statistics. CA Cancer J Clin 61(2):69–90 2. Kim JP, Lee JH, Kim SJ et al (1998) Clinicopathologic characteristics and prognostic factors in 10783 patients with gastric cancer. Gastric Cancer 1(2):125–133 3. Edge SB, Byrd DR, Compton CC et al (2010) American Joint Committee on Cancer, American Cancer Society: AJCC cancer staging manual, 7th edn. Springer, New York 4. Japanese Gastric Cancer Association (2011) Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric Cancer 14(2):113–123 5. Borie F, Millat B, Fingerhut A et al (2000) Lymphatic involvement in early gastric cancer: prevalence and prognosis in France. Arch Surg 135(10):1218–1223 6. Chen R, He Q, Cui J et al (2014) Lymph node metastasis in early gastric cancer. Chin Med J (Engl) 127(3):560–567 7. Gotoda T, Yanagisawa A, Sasako M et al (2000) Incidence of lymph node metastasis from early gastric cancer: estimation with a large number of cases at two large centers. Gastric Cancer 3(4):219–225 8. An JY, Baik YH, Choi MG et al (2007) Predictive factors for lymph node metastasis in early gastric cancer with submucosal invasion: analysis of a single institutional experience. Ann Surg 246(5):749–753 9. Liu C, Zhang R, Lu Y et al (2010) Prognostic role of lymphatic vessel invasion in early gastric cancer: a retrospective study of 188 cases. Surg Oncol 19(1):4–10 10. Ichikawa D, Kubota T, Kikuchi S et al (2009) Prognostic impact of lymphatic invasion in patients with node-negative gastric cancer. J Surg Oncol 100(2):111–114
739 11. Liu E, Zhong M, Xu F et al (2011) Impact of lymphatic vessel invasion on survival in curative resected gastric cancer. J Gastrointest Surg 15(9):1526–1531 12. Kunisaki C, Makino H, Kimura J et al (2010) Impact of lymphovascular invasion in patients with stage I gastric cancer. Surgery 147(2):204–211 13. Hyung WJ, Lee JH, Choi SH et al (2002) Prognostic impact of lymphatic and/or blood vessel invasion in patients with nodenegative advanced gastric cancer. Ann Surg Oncol 9(6):562–567 14. Bu Z, Zheng Z, Li Z et al (2013) Lymphatic vascular invasion is an independent correlated factor for lymph node metastasis and the prognosis of resectable T2 gastric cancer patients. Tumour Biol 34(2):1005–1012 15. Yokota T, Kunii Y, Teshima S, et al (1999) Significant prognostic factors in patients with node-negative gastric cancer. Int Surg 84(4):331–336 16. Lee IS, Yook JH, Kim TH et al (2013) Prognostic factors and recurrence pattern in node-negative advanced gastric cancer. Eur J Surg Oncol 39(2):136–140 17. Sakuramoto S, Sasako M, Yamaguchi T et al (2007) Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine. N Engl J Med 357(18):1810–1820 18. Bang YJ, Kim YW, Yang HK et al (2012) Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial. Lancet 379(9813):315–321 19. Japanese Gastric Cancer Association (1998) Japanese classification of gastric carcinoma—2nd english edition. Gastric Cancer 1:10–24 20. Shen L, Shan YS, Hu HM et al (2013) Management of gastric cancer in Asia: resource-stratified guidelines. Lancet Oncol 14(12):e535–e547 21. Lim CS, Alexander-Sefre F, Allam M et al (2008) Clinical value of immunohistochemically detected lymphovascular space invasion in early stage cervical carcinoma. Ann Surg Oncol 15:2581–2588 22. O’Donnell RK, Feldman M, Mick R et al (2008) Immunohistochemical method identifies lymphovascular invasion in a majority of oral squamous cell carcinomas and discriminates between blood and lymphatic vessel invasion. J Histochem Cytochem 56:803–810
123
