World J Surg (2014) 38:3257–3264 DOI 10.1007/s00268-014-2734-5
ORIGINAL SCIENTIFIC REPORT
Prognostic Value of the Number and Size of Venous Invasions in pT3 Colorectal Cancer: A Prospective Observational Study Eiji Shinto • Hitoshi Tsuda • Hideki Ueno Hideyuki Shimazaki • Junji Yamamoto • Kazuo Hase
•
Published online: 29 August 2014 Ó Socie´te´ Internationale de Chirurgie 2014
Abstract Background Although venous invasion is reportedly a clinically useful prognostic marker for colorectal cancer, suitable grading criteria have not been established. Objective This prospective observational study aimed to investigate the prognostic value of the number and size of venous invasions in patients with pT3 colorectal cancer. Methods We recruited 139 consecutive patients with pT3 colorectal carcinomas resected between October 2001 and August 2003. We used slides of whole-tissue sections stained with Elastica van Gieson. Venous invasion was classified according to the number of veins with carcinoma infiltration per slide with most venous invasions (Vnumber classification): V(n)-low 0–3 and V(n)-high C4. Additionally, the findings were classified according to the maximal size of veins containing carcinoma infiltration (V-size classification): V(s)-low \1 mm and V(s)-high C1 mm. The grades of venous invasions were evaluated just after surgery. Results The 5-year survival rate of V(n)-low and V(n)-high were 89.9 and 59.1 %, respectively (p \ 0.0001). Comparisons between overall survival curves revealed that V-number classification (but not V-size classification) had a significant prognostic value in patients with pT3 cancer, especially in stage II patients (98.2 and 64.2 %, respectively; p \ 0.0001). Multivariate analysis revealed distant metastasis, age, and V-number classification (hazard ratio 3.1; p = 0.0071) as independent prognostic indicators. Conclusions V-number classification may be a useful prognostic system when evaluating and sub-grouping patients with pT3 colorectal cancer.
E. Shinto (&) H. Ueno J. Yamamoto K. Hase Department of Surgery, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan e-mail: [email protected] H. Tsuda Department of Pathology, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan H. Shimazaki Department of Laboratory Medicine, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan
Introduction Hematogenous metastasis is a life-threatening mode of cancer spread; venous invasion is an essential step in this metastatic process, and may adversely affect patient outcomes. Several reports have indicated that the pathological findings of venous invasion are associated with poor prognosis in patients with colorectal cancer [1, 2]. In addition, venous invasion can provide useful prognostic information for patients with Dukes’ B carcinoma [3] or
123
3258
stage IV liver metastasis [4]. Thus, venous invasion might be an important pathological finding in all cancer stages. According to the staging system developed by the American Joint Committee on Cancer (AJCC), venous invasion in colorectal cancer is classified (on the basis of the size of the cancer infiltration in the vessel) as follows: V0 no venous invasion; V1 microscopic venous invasion; and V2 macroscopic venous invasion [5]. In contrast, the Japanese classification of colorectal carcinoma is graded according to severity: v0 no invasion; v1 minimal invasion; v2 moderate invasion; and v3 severe invasion [6]. However, although venous invasion is widely accepted as a clinically useful prognostic indicator, no single objective set of criteria exists for grading venous invasion. Some studies have reported objective grading systems for prognostication in colorectal cancer. A notable procedure involves counting the number of vessels with cancer involvement [7–9]. Size [9, 10], location (extramural or intramural) [7–9], and morphologic type (floating or embolization) [9] of venous invasion have also been discussed; however, a consensus on their clinical significance has not been reached. To establish objective criteria, we retrospectively investigated the prognostic value of venous invasion using whole-tissue section slides stained with Elastica van Gieson (EVG), and determined that patients with pT3 colorectal cancer had a poor prognosis if they had four or more venous vessels containing carcinoma cells [9, 11]. These pT3 cancers invade the subserosa but remain confined to the peritoneum and can be curatively resected if there are no distant metastatic lesions; these cancers are thus good candidates for analyzing hematogenous metastasis caused by venous invasion. In that study, the size of venous invasions was not clinically significant, although we repeated the analysis using four cut-off values (0.5, 1.0, and 1.5 mm, and mean size) [9]. Further, neither the location nor the morphologic type was clinically significant. We hypothesized that the number of venous invasions may provide crucial prognostic information for pT3 colorectal cancer. To test this hypothesis, we used whole-tissue section slides stained with EVG and prospectively investigated the prognostic values of the number of venous invasions via univariate and multivariate analyses. In addition, we examined the prognostic significance of the size of venous invasion, because this has been adopted by the AJCC staging system for categorization.
Methods Patients This study was conducted after obtaining approval from the internal review board of our hospital. Signed informed
123
World J Surg (2014) 38:3257–3264
consent was obtained from each patient for the use of tissue in this study. Between October 2001 and August 2003, a total of 296 patients underwent surgery for primary carcinoma of the colon and rectum above the peritoneal reflection at the Department of Surgery of the National Defense Medical College Hospital in Japan. Of these, 139 consecutive patients with pT3 carcinoma participated in this study. The primary inclusion criterion was that the microscopic invasion depth was judged as pT3. Among the participants, 75, 50, and 14 were diagnosed as stage II, stage III, and stage IV, respectively. Although 36 patients with rectal carcinoma below the peritoneal reflection underwent surgery in this period, they were not included because neoadjuvant chemoradiotherapy had been administered, which strongly affects pathological findings. Curative operations were performed in 129 patients (92.8 %), and palliative operations were performed in ten patients (7.2 %). Four patients with stage IV disease underwent curative operations (R0), which comprised resection of colorectal cancer and partial hepatectomy for liver metastasis. Adjuvant chemotherapy was administered to a single stage II patient (1.3 %) and 34 stage III patients (68.0 %) after curative operations. The adjuvant chemotherapy regimen was either intravenous 5-fluorouracil and folinic acid or oral uracil and tegafur. All stage IV patients received systemic chemotherapy. No patients received preoperative chemotherapy or radiotherapy. The male to female ratio was 79:60 and the median age was 67 years (range 18–96). The median tumor size was 46 mm (range 15–150). The medians were used for the cut-off values. Assessment of venous invasion For each tumor, three whole-tissue sections in which venous invasions were frequently observed after hematoxylin and eosin staining were chosen and stained using EVG. This method facilitated the identification of venous vessels containing round elastic fibers (Fig. 1). We then evaluated the grade of venous invasion according to the number of veins with carcinoma infiltration per slide with most venous invasions (V-number classification): V(n)low = 0–3 venous vessels containing carcinoma cells and V(n)-high = four or more vessels containing carcinoma cells. Additionally, the length of the minor axis of the vascular vessel involved was defined as the size of venous invasion (Fig. 1c) [9]. Venous invasion was then classified into the following based on the maximal size (V-size classification): V(s)-low \1 mm and V(s)-high C1 mm. Cut-off values for V-number classification (0–3 and C4) and V-size classification (\1 mm and C1 mm) were determined as those offering the greatest difference in survival based on the database used in an earlier retrospective study [11]. The grades of venous invasion were
World J Surg (2014) 38:3257–3264
3259
Fig. 1 Characteristic appearance of venous invasion on a slide stained with (a) hematoxylin and eosin, and (b) Elastica van Gieson. In the latter, venous vessels have round elastic lamina. Black arrows indicate
venous vessel with cancer infiltration. (c) The length of the minor axis of the vascular vessel involved (arrow). (Magnification 9 200)
independently evaluated by two observers who conducted routine pathological diagnosis after surgery. Any cases with discrepant grades were discussed and re-evaluated.
using Statview 5 software (SAS Institute, Cary, NC, USA), and p \ 0.05 was considered significant.
Follow-up
Results
All 139 patients received regular follow-up at our outpatient clinic. Physical examination, chest radiography, serum carcinoembryonic antigen levels, and carbohydrate antigen 19–9 levels were monitored every 3 months; a contrast computed tomography scan was performed every 6 months; and colonoscopy was performed annually. Whenever any findings suggestive of cancer relapse did not appear during 5-year, the follow-up procedure was changed to an annual physical check without any other detailed examinations. If patients did not attend clinic, we evaluated their health annually by telephone. At the date of the last follow-up, 31 patients had died, with a median time from surgery to death of 32.6 months (range 5.5–116.9). Of these, 25 died from colorectal carcinoma recurrence and 6 died from other reasons or from unclear causes. The median follow-up period for the survivors (n = 108) was 62.1 months (range 6.1–147.0).
Relations to clinicopathological findings
Statistical analysis Comparisons between groups were performed using the v2 test or Fisher’s exact test. Kaplan–Meier survival curves were calculated for patients, and differences between curves were calculated using the log-rank test. Cox’s proportional hazards regression analysis was used for multivariate analyses. All statistical analyses were performed
Microscopically, carcinomas met the V(n)-high and V(s)high requirements in 41 (29.5 %) and 47 (33.8 %) patients, respectively. The remainder were classified into V(n)-low (98 patients, 70.5 %) and V(s)-low (92 patients, 66.2 %) groups. Table 1 shows the distribution of patients with high-grade and low-grade venous invasion according to their clinicopathological features. The medians were set as the cut-off values for age (C67 and \67 years) and tumor size (C46 and \46 mm). There were no significant differences in age, sex, or tumor location. However, the incidences of distant metastasis and nodal metastasis were higher in the V(n)-high group than in the V(n)-low group (p = 0.0026 and p = 0.012, respectively). The V-size classification grade was positively associated with tumor size (p = 0.032). Additionally, V-number classification was strongly associated with V-size classification (p \ 0.0001). Relation to survival Figures 2a, b show the Kaplan–Meier overall survival curves according to V-number and V-size classifications, respectively. The V(n)-high group had a significantly worse outcome than the V(n)-low group (5-year survival
123
3260
World J Surg (2014) 38:3257–3264
Table 1 Clinicopathological features and their correlations with the grades of venous invasion No of cases (%) Total (n = 139)
V-number classification V(n)-low (n = 98)
V(n)-high (n = 41)
p value
V-size classification V(s)-low (n = 92)
V(s)-high (n = 47)
47 (64)
26 (36)
45 (68)
21 (32)
49 (62)
30 (38)
43 (72)
17 (28)
32 (65)
17 (35)
60 (67)
30 (33)
p value
Agea (years) C67
73
54 (74)
19 (26)
\67
66
44 (67)
22 (33)
Male
79
60 (76)
19 (24)
Female
60
38 (63)
22 (37)
Right side
49
36 (73)
13 (27)
Left side
90
62 (69)
28 (31)
0.35
0.70
Sex 0.11
0.23
Tumor location 0.57
0.87
Tumor sizea C46 mm
71
49 (69)
22 (31)
\46 mm
68
49 (72)
19 (28)
0.69
131 8
95 (73) 3 (38)
36 (27) 5 (63)
0.049b
0.0026
41 (58)
30 (42)
51 (75)
17 (25)
0.032
87 (66) 5 (63)
44 (34) 3 (37)
0.99
0.18
Histologic typec Well/Mod Por/Muc Distant metastasis Positive
14
5 (36)
9 (64)
Negative
125
93 (74)
32 (26)
Positive
62
37 (60)
25 (40)
Negative
77
61 (79)
16 (21)
7 (50)
7 (50)
85 (68)
40 (32)
40 (65)
22 (35)
52 (68)
25 (32)
Nodal metastasis
a
0.012
0.71
The medians of age and tumor size were used for cutoff values
b
Fisher’s exact test
c
Well/mod indicates well-moderately differentiated; Por/Muc indicates poorly differentiated/mucinous adenocarcinoma
V(n) number of veins, V(s) size of veins
rate 59.1 vs. 89.9 %; p \ 0.0001), whereas V-size classification did not show prognostic significance (72.6 vs. 84.7 %, respectively; p = 0.080). Prognostic significance was also assessed by multivariate analysis (Table 2). Variables with prognostic significance after univariate analysis were used as variables in the multivariate analysis. V-number classification (hazard ratio [HR] 3.1; p = 0.0071) was an independent prognostic factor, as were distant metastasis (HR 5.7; p \ 0.0001) and age (HR 2.7; p = 0.014). In addition, the analysis of the disease-specific survival curves of all patients demonstrated the prognostic significance of V-number classification; the 5-year survival rates were 71.5 and 89.9 % for the V(n)-high and V(n)-low groups, respectively (p = 0.0003). Similar results were observed when the recurrence-free survival curves of patients following curative operations were compared (71.0 vs. 92.2 %; p = 0.0036).
123
Relation to postoperative recurrence Among the 129 patients undergoing curative operations (R0), 19 patients (14.7 %) experienced postoperative recurrence (Table 3). The overall recurrence rate of patients in the V(n)-high group was higher than that of patients in the V(n)-low group (p = 0.0068). With respect to primary recurrence sites, recurrence linked to hematogenous spread was more frequent in the V(n)-high group than in the V(n)-low group (p = 0.010). In contrast, V-size classification was not associated with either overall recurrence rate or recurrence rates at specific sites. Subgroup analysis of overall survival Subgroup analyses were performed (Table 4) across the various cancer stages in this study. The prognostic
World J Surg (2014) 38:3257–3264
3261
Fig. 2 a Cumulative overall survival curves for 139 patients with pT3 colorectal cancers stratified by V-number [V(n)] classification. The curves for the V(n)-low and V(n)-high groups were significantly different (p \ 0.0001). b Cumulative overall survival curves for the 139 patients with pT3 colorectal cancers stratified by V-size (V[s]) classification. The curves for the V(s)-low and V(s)-high groups were not significantly different (p = 0.080)
Table 2 Significance of clinicopathological parameters for overall survival Parameter
Multivariate analysisa
Comparison of overall survival 5-year survival rate (%)b
p valuec
HR
95 % CI
p value
73.0 vs. 88.6
0.031*
2.7
1.2–5.8
0.014*
78.9 vs. 81.4
0.94
Right vs. left side
85.7 vs. 77.7
0.31
Tumor size C46 vs. \46 mm
82.7 vs. 78.3
0.72
80.0 vs. 87.5
0.58
28.6 vs. 86.7
\0.0001*
5.7
2.4–13.7
\0.0001*
67.9 vs. 90.4
0.0087*
1.3
0.59–3.0
0.48
59.1 vs. 89.9
\0.0001*
3.1
1.4–7.0
0.0071*
72.6 vs. 84.7
0.080
Age (years) C67 vs. \67 Sex Female vs. male Tumor location
Histologic typed Well/Mod vs. Por/Muc Distant metastasis Positive vs. negative Nodal metastasis Positive vs. negative V-number classification V(n)-high vs. V(n)-low V-size classification V(s)-high vs. V(s)-low a b
Cox’s proportional hazard model Kaplan–Meier method
c
Log-rank test
d
Well/mod indicates well-moderately differentiated; Por/Muc indicates poorly differentiated/mucinous adenocarcinoma
* Statistically significant CI confidence interval, HR hazard ratio, V(n) number of veins, V(s) size of veins
significance of V-number classification was confirmed only in patients with stage II disease; the 5-year overall survival rates were 64.2 and 98.2 % for the V(n)-high and V(n)-low groups, respectively (p \ 0.0001). In contrast, adjuvant chemotherapy had a marked impact on the prognosis of
stage III patients (48.1 % [without chemotherapy] vs. 90.5 % [with chemotherapy]; p = 0.0068). Among patients with stage III V(n)-low cancers, the 5-year survival rate without adjuvant chemotherapy (n = 11) was much worse than that with adjuvant
123
3262
World J Surg (2014) 38:3257–3264
Table 3 Postoperative recurrence in patients with curative resection (R0) Number of cases (%) Total (n = 129)
V-number classification V(n)-low (n = 94)
V(n)-high (n = 35)
19 (15)
9 (10)
10 (29)
Liver
11 (9)
4 (4)
7 (20)
Lung
8 (6)
5 (5)
17 (13) 2 (2)
Overall
p value
V-size classification
p value
V(s)-low (n = 87)
V(s)-high (n = 42)
0.0068
10 (12)
9 (21)
0.14
0.0090b
6 (7)
5 (12)
0.34
3 (9)
0.68b
6 (7)
2 (5)
0.99b
8 (9)
9 (26)
0.010
10 (11)
7 (17)
0.42
1 (1)
1 (3)
0.47b
0 (0)
2 (5)
0.10b
Primary recurrence sitea Hematogenous
a
Total
Non-hematogenous Local a
Primary recurrence was not limited to a single organ. Primary recurrence was found in liver and lung for two patients
b
Fisher’s exact test
Table 4 Subgroup analysis of overall survival
Parameter (number of cases)
Comparison of overall survival 5-year survival ratea (%)
p valueb
89.7 vs. 93.6
0.20
64.2 vs. 98.2
\0.0001c
77.6 vs. 98.0
0.0011
91.4 vs. 100
–
66.0 vs. 91.6
0.040
69.0 vs. 85.1
0.27
85.7 vs. 76.0
0.27
48.1 vs. 90.5
0.0068c
12.5 vs. 50.0
0.51
33.3 vs. 20.0
0.76
28.6 vs. 28.6
0.50
75.0 vs. 10.0
0.27
Stage II patients Age (years) C67 (42) vs. \67 (33) V-number classification V(n)-high (15) vs. V(n)-low (60) V-size classification V(s)-high (24) vs. V(s)-low (51) Adjuvant chemotherapy Without chemo (74) vs. with chemo (1) Stage III patients Age (years) C67 (23) vs. \67 (27) V-number classification V(n)-high (17) vs. V(n)-low (33) V-size classification V(s)-high (16) vs. V(s)-low (34) Adjuvant chemotherapy Without chemo (16) vs. with chemo (34) Stage IV patients Age (years) C 67 (8) vs. \ 67 (6) V-number classification V(n)-high (9) vs. V(n)-low (5) a
Kaplan–Meier method
b
Log-rank test
c
Independent predictors disclosed by multivariate analysis
123
V-size classification V(s)-high (7) vs. V(s)-low (7) Residual tumor R0 (4) vs. R1 or 2 (10)
World J Surg (2014) 38:3257–3264
chemotherapy (n = 22) (56.3 % [without chemotherapy] vs. 95.0 % [with chemotherapy]; p = 0.020). For reference, the 5-year survival rate of stage II patients with V(n)low cancers (n = 59) was 98.2 %, despite adjuvant chemotherapy not being used. Among patients with stage III V(n)-high cancers, the 5-year survival rate without adjuvant chemotherapy (n = 5) was 40.0 %, and increased to 82.5 % with the use of adjuvant chemotherapy (n = 12); however, the difference was not significant (p = 0.15). Notably, the 5-year survival rate of patients with stage II V(n)-high cancers (n = 15) not receiving adjuvant chemotherapy was 64.2 %.
Discussion Van Wyk et al. [12] described a review article that summarized the current status of detection methods for components of vascular invasion and its clinical significance. They concluded that, when specific staining is used, blood vessel invasion appears to be a reliable independent prognostic factor. In our prospective observational study, we enrolled 139 patients with pT3 colorectal cancer to clarify the prognostic significance of the number of venous invasions using section slides stained with EVG. The results of both univariate and multivariate analyses indicate that V-number classification was an independent predictor of patient outcome after surgery. However, although V-size classification had a positive relation with V-number classification, it was not a suitable predictive marker for patient prognosis. Most pathologists accept that venous invasion can predict hematogenous recurrence based on the biological plausibility that cancer cells infiltrating venous vessels can be dissociated from the original tumor, thereby facilitating spread to other organs. Iinuma et al. [13] reported that the pathological findings of venous invasions significantly correlated with the presence of isolated tumor cells in blood samples. Furthermore, another study reported that detecting circulating tumor cells in the peripheral blood was as a useful tool for determining the patients at increased risk for recurrence [14]. These studies support the hypothesis that venous invasion is an essential step in the process of hematogenous metastasis, as does our prospective observational study, which demonstrates that venous invasion precedes hematogenous recurrence. In this study, we focused on the number and size of venous invasions, and identified clinical values for these grading systems, but were unable to determine why these values showed diversity. However, the V(s)-high group frequently contained cancer nodules in venous vessels, which appeared to reflect the expansive growth and
3263
cohesive property of cancer cells. The concept that only cancer stem cells leaving the original tumor site can be precursors of metastasis is widely accepted based on the cancer stem cell theory [15, 16]. According to this model, venous invasion by cancer stem cells can therefore be a powerful prognostic indicator. That said, cancer stem cells are characterized by low proliferation and loss of epithelial features [15, 17], which are unlikely to form tumor nodules with expansive growth and cohesive properties. On the other hand, Bugg et al. [18] suggested that the presence of magnetic resonance imaging (MRI)-detected venous invasion is a poor prognostic feature of rectal cancer. They reported that one-half of patients with MRI-detected invasion of the superior rectal vein had encountered distant recurrence [18]. Although our data implied that the size of venous invasion is not associated with patients’ prognosis, careful reconsideration is necessary to determine the significance of cancer involvement of named colic or rectal veins. The subgroup analyses demonstrated that V-number classification had a prognostic impact in patients with stage II disease, but not in those with stage III or IV disease. It is conceivable that venous invasion predicts hematogenous recurrence, and that there is therefore nothing to predict in patients who already have distant metastasis in stage IV disease. In patients with stage III disease, the prognostic impact of adjuvant chemotherapy was marked, which possibly reduced the prognostic impact of venous invasion. Clinically, it was problematic that 20 % of the stage II patients who were considered unsuitable for adjuvant chemotherapy were also judged to be V(n)-high. Because the 5-year survival rate was 82.5 % in patients with stage III V(n)-high cancers after adjuvant chemotherapy, the patients with stage II V(n)-high disease with a 5-year survival rate of 64.2 % might also need to be offered adjuvant chemotherapy. Data from the literature on the prognostic significance of vascular invasion in stage II patients are still insufficient. However, Betge et al. [19] disclosed that venous invasion is a powerful predictor of both disease progression and cancer-related death and increased the HR by a factor of approximately 5 [19]. Accordingly, they also advocated consideration of venous invasion when stratifying patients with stage II colorectal cancer for adjuvant treatment. There are potential limitations to this study. First, we evaluated only pT3 cancers. A larger trial including cancers with various T categories will confirm our conclusions. Second, this was a prospective observational study, and postoperative adjuvant chemotherapy and treatments for recurrence were not homogeneous. Adjuvant treatment may be affected by pathological diagnosis, physical condition, and patient preference. In this period, patients with node-positive cancers were offered adjuvant
123
3264
chemotherapy, which possibly reduced the impact of nodal metastasis on patient outcomes. Third, the relatively small number of patients with stage II V(n)-high cancers is a limitation, and further work is needed to establish the clinical utility. In the present study, we demonstrated that the number of venous invasions might provide crucial prognostic information for patients with pT3 colorectal cancer, especially at stage II. These novel findings may be clinically applicable for selecting suitable candidates for intensive adjuvant chemotherapy. Acknowledgments This work was supported in part by the National Cancer Center Research and Development Fund (23-A-11 and 23-A26) and by the Japan Society for the Promotion of Science KAKENHI (23501302, 25462074, and 26462032). Conflict of interest The authors certify that we have no associations, commercial or otherwise, that may pose a conflict of interest.
References 1. Minsky BD, Mies C, Richt TA et al (1988) Potentially curative surgery of colon cancer: the influence of blood vessel invasion. J Clin Oncol 6:119–127 2. Talbot IC, Ritchie S, Leighton M et al (1980) The clinical significance of invasion of veins by rectal cancer. Br J Surg 67:439–442 3. Petersen VC, Baxter KJ, Love SB et al (2002) Identification of objective pathological prognostic determinants and models of prognosis in Duke’s B colon cancer. Gut 51:65–69 4. Chafai N, Chan CL, Bokey EL et al (2005) What factors influence survival in patients with unresected synchronous liver metastases after resection of colorectal cancer? Colorectal Dis 7:176–181 5. Sobin LH, Gospodarowisz M, Wittekind C (2009) TNM: classification of malignant tumours, 7th edn. Wiley, New York 6. Japanese Society for Cancer of Colon and Rectum (2009) In: Japanese classification of colorectal carcinoma (2nd English edn). Kanehara & Co. Ltd, Tokyo, p. 27
123
World J Surg (2014) 38:3257–3264 7. Ouchi K, Sugawara T, Ono H et al (1996) Histologic features and clinical significance of venous invasion in colorectal carcinoma with hepatic metastasis. Cancer 78:2313–2317 8. Sternberg A, Mizrahi A, Amar M et al (2006) Detection of venous invasion in surgical specimens of colorectal carcinoma: the efficacy of various types of tissue blocks. J Clin Pathol 59:207–210 9. Sato T, Ueno H, Mochizuki H et al (2010) Objective criteria for the grading of venous invasion in colorectal cancer. Am J Surg Pathol 34:454–462 10. Krasna MJ, Flancbaum L, Cody RP et al (1988) Vascular and neural invasion in colorectal carcinoma. Cancer 61:1018–1023 11. Sato T, Shinto E, Hashiguchi Y et al (2006) Prognostic significance of lymphatic and venous invasions in pT3 colorectal carcinoma, quantified with a double staining combining CD34 immunostaining and elastica Staining. Jpn J Gastroenterol Surg 39:1571–1576 12. van Wyk HC, Roxburgh CS, Horgan PG et al (2014) The detection and role of lymphatic and blood vessel invasion in predicting survival in patients with node negative operable primary colorectal cancer. Crit Rev Oncol Hematol 90:77–90 13. Iinuma H, Watanabe T, Mimori K et al (2011) Clinical significance of circulating tumor cells, including cancer stem-like cells, in peripheral blood for recurrence and prognosis in patients with Dukes’ stage B and C colorectal cancer. J Clin Oncol 29:1547–1555 14. Iinuma H, Okinaga K, Egami H et al (2006) Usefulness and clinical significance of quantitative real-time RT-PCR to detect isolated tumor cells in the peripheral blood and tumor drainage blood of patients with colorectal cancer. Int J Oncol 28:297–306 15. Willis ND, Przyborski SA, Hutchison CJ et al (2008) Colonic and colorectal cancer stem cells: progress in the search for putative biomarkers. J Anat 213:59–65 16. Yeung TM, Mortensen NJ (2009) Colorectal cancer stem cells. Dis Colon Rectum 52:1788–1796 17. Mani SA, Guo W, Liao MJ et al (2008) The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133:704–715 18. Bugg WG, Andreou AK, Biswas D et al (2014) The prognostic significance of MRI-detected extramural venous invasion in rectal carcinoma. Clin Radiol 69(6):619–623 19. Betge J, Pollheimer MJ, Lindtner RA et al (2012) Intramural and extramural vascular invasion in colorectal cancer: prognostic significance and quality of pathology reporting. Cancer 118:628–638
ORIGINAL SCIENTIFIC REPORT
Prognostic Value of the Number and Size of Venous Invasions in pT3 Colorectal Cancer: A Prospective Observational Study Eiji Shinto • Hitoshi Tsuda • Hideki Ueno Hideyuki Shimazaki • Junji Yamamoto • Kazuo Hase
•
Published online: 29 August 2014 Ó Socie´te´ Internationale de Chirurgie 2014
Abstract Background Although venous invasion is reportedly a clinically useful prognostic marker for colorectal cancer, suitable grading criteria have not been established. Objective This prospective observational study aimed to investigate the prognostic value of the number and size of venous invasions in patients with pT3 colorectal cancer. Methods We recruited 139 consecutive patients with pT3 colorectal carcinomas resected between October 2001 and August 2003. We used slides of whole-tissue sections stained with Elastica van Gieson. Venous invasion was classified according to the number of veins with carcinoma infiltration per slide with most venous invasions (Vnumber classification): V(n)-low 0–3 and V(n)-high C4. Additionally, the findings were classified according to the maximal size of veins containing carcinoma infiltration (V-size classification): V(s)-low \1 mm and V(s)-high C1 mm. The grades of venous invasions were evaluated just after surgery. Results The 5-year survival rate of V(n)-low and V(n)-high were 89.9 and 59.1 %, respectively (p \ 0.0001). Comparisons between overall survival curves revealed that V-number classification (but not V-size classification) had a significant prognostic value in patients with pT3 cancer, especially in stage II patients (98.2 and 64.2 %, respectively; p \ 0.0001). Multivariate analysis revealed distant metastasis, age, and V-number classification (hazard ratio 3.1; p = 0.0071) as independent prognostic indicators. Conclusions V-number classification may be a useful prognostic system when evaluating and sub-grouping patients with pT3 colorectal cancer.
E. Shinto (&) H. Ueno J. Yamamoto K. Hase Department of Surgery, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan e-mail: [email protected] H. Tsuda Department of Pathology, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan H. Shimazaki Department of Laboratory Medicine, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan
Introduction Hematogenous metastasis is a life-threatening mode of cancer spread; venous invasion is an essential step in this metastatic process, and may adversely affect patient outcomes. Several reports have indicated that the pathological findings of venous invasion are associated with poor prognosis in patients with colorectal cancer [1, 2]. In addition, venous invasion can provide useful prognostic information for patients with Dukes’ B carcinoma [3] or
123
3258
stage IV liver metastasis [4]. Thus, venous invasion might be an important pathological finding in all cancer stages. According to the staging system developed by the American Joint Committee on Cancer (AJCC), venous invasion in colorectal cancer is classified (on the basis of the size of the cancer infiltration in the vessel) as follows: V0 no venous invasion; V1 microscopic venous invasion; and V2 macroscopic venous invasion [5]. In contrast, the Japanese classification of colorectal carcinoma is graded according to severity: v0 no invasion; v1 minimal invasion; v2 moderate invasion; and v3 severe invasion [6]. However, although venous invasion is widely accepted as a clinically useful prognostic indicator, no single objective set of criteria exists for grading venous invasion. Some studies have reported objective grading systems for prognostication in colorectal cancer. A notable procedure involves counting the number of vessels with cancer involvement [7–9]. Size [9, 10], location (extramural or intramural) [7–9], and morphologic type (floating or embolization) [9] of venous invasion have also been discussed; however, a consensus on their clinical significance has not been reached. To establish objective criteria, we retrospectively investigated the prognostic value of venous invasion using whole-tissue section slides stained with Elastica van Gieson (EVG), and determined that patients with pT3 colorectal cancer had a poor prognosis if they had four or more venous vessels containing carcinoma cells [9, 11]. These pT3 cancers invade the subserosa but remain confined to the peritoneum and can be curatively resected if there are no distant metastatic lesions; these cancers are thus good candidates for analyzing hematogenous metastasis caused by venous invasion. In that study, the size of venous invasions was not clinically significant, although we repeated the analysis using four cut-off values (0.5, 1.0, and 1.5 mm, and mean size) [9]. Further, neither the location nor the morphologic type was clinically significant. We hypothesized that the number of venous invasions may provide crucial prognostic information for pT3 colorectal cancer. To test this hypothesis, we used whole-tissue section slides stained with EVG and prospectively investigated the prognostic values of the number of venous invasions via univariate and multivariate analyses. In addition, we examined the prognostic significance of the size of venous invasion, because this has been adopted by the AJCC staging system for categorization.
Methods Patients This study was conducted after obtaining approval from the internal review board of our hospital. Signed informed
123
World J Surg (2014) 38:3257–3264
consent was obtained from each patient for the use of tissue in this study. Between October 2001 and August 2003, a total of 296 patients underwent surgery for primary carcinoma of the colon and rectum above the peritoneal reflection at the Department of Surgery of the National Defense Medical College Hospital in Japan. Of these, 139 consecutive patients with pT3 carcinoma participated in this study. The primary inclusion criterion was that the microscopic invasion depth was judged as pT3. Among the participants, 75, 50, and 14 were diagnosed as stage II, stage III, and stage IV, respectively. Although 36 patients with rectal carcinoma below the peritoneal reflection underwent surgery in this period, they were not included because neoadjuvant chemoradiotherapy had been administered, which strongly affects pathological findings. Curative operations were performed in 129 patients (92.8 %), and palliative operations were performed in ten patients (7.2 %). Four patients with stage IV disease underwent curative operations (R0), which comprised resection of colorectal cancer and partial hepatectomy for liver metastasis. Adjuvant chemotherapy was administered to a single stage II patient (1.3 %) and 34 stage III patients (68.0 %) after curative operations. The adjuvant chemotherapy regimen was either intravenous 5-fluorouracil and folinic acid or oral uracil and tegafur. All stage IV patients received systemic chemotherapy. No patients received preoperative chemotherapy or radiotherapy. The male to female ratio was 79:60 and the median age was 67 years (range 18–96). The median tumor size was 46 mm (range 15–150). The medians were used for the cut-off values. Assessment of venous invasion For each tumor, three whole-tissue sections in which venous invasions were frequently observed after hematoxylin and eosin staining were chosen and stained using EVG. This method facilitated the identification of venous vessels containing round elastic fibers (Fig. 1). We then evaluated the grade of venous invasion according to the number of veins with carcinoma infiltration per slide with most venous invasions (V-number classification): V(n)low = 0–3 venous vessels containing carcinoma cells and V(n)-high = four or more vessels containing carcinoma cells. Additionally, the length of the minor axis of the vascular vessel involved was defined as the size of venous invasion (Fig. 1c) [9]. Venous invasion was then classified into the following based on the maximal size (V-size classification): V(s)-low \1 mm and V(s)-high C1 mm. Cut-off values for V-number classification (0–3 and C4) and V-size classification (\1 mm and C1 mm) were determined as those offering the greatest difference in survival based on the database used in an earlier retrospective study [11]. The grades of venous invasion were
World J Surg (2014) 38:3257–3264
3259
Fig. 1 Characteristic appearance of venous invasion on a slide stained with (a) hematoxylin and eosin, and (b) Elastica van Gieson. In the latter, venous vessels have round elastic lamina. Black arrows indicate
venous vessel with cancer infiltration. (c) The length of the minor axis of the vascular vessel involved (arrow). (Magnification 9 200)
independently evaluated by two observers who conducted routine pathological diagnosis after surgery. Any cases with discrepant grades were discussed and re-evaluated.
using Statview 5 software (SAS Institute, Cary, NC, USA), and p \ 0.05 was considered significant.
Follow-up
Results
All 139 patients received regular follow-up at our outpatient clinic. Physical examination, chest radiography, serum carcinoembryonic antigen levels, and carbohydrate antigen 19–9 levels were monitored every 3 months; a contrast computed tomography scan was performed every 6 months; and colonoscopy was performed annually. Whenever any findings suggestive of cancer relapse did not appear during 5-year, the follow-up procedure was changed to an annual physical check without any other detailed examinations. If patients did not attend clinic, we evaluated their health annually by telephone. At the date of the last follow-up, 31 patients had died, with a median time from surgery to death of 32.6 months (range 5.5–116.9). Of these, 25 died from colorectal carcinoma recurrence and 6 died from other reasons or from unclear causes. The median follow-up period for the survivors (n = 108) was 62.1 months (range 6.1–147.0).
Relations to clinicopathological findings
Statistical analysis Comparisons between groups were performed using the v2 test or Fisher’s exact test. Kaplan–Meier survival curves were calculated for patients, and differences between curves were calculated using the log-rank test. Cox’s proportional hazards regression analysis was used for multivariate analyses. All statistical analyses were performed
Microscopically, carcinomas met the V(n)-high and V(s)high requirements in 41 (29.5 %) and 47 (33.8 %) patients, respectively. The remainder were classified into V(n)-low (98 patients, 70.5 %) and V(s)-low (92 patients, 66.2 %) groups. Table 1 shows the distribution of patients with high-grade and low-grade venous invasion according to their clinicopathological features. The medians were set as the cut-off values for age (C67 and \67 years) and tumor size (C46 and \46 mm). There were no significant differences in age, sex, or tumor location. However, the incidences of distant metastasis and nodal metastasis were higher in the V(n)-high group than in the V(n)-low group (p = 0.0026 and p = 0.012, respectively). The V-size classification grade was positively associated with tumor size (p = 0.032). Additionally, V-number classification was strongly associated with V-size classification (p \ 0.0001). Relation to survival Figures 2a, b show the Kaplan–Meier overall survival curves according to V-number and V-size classifications, respectively. The V(n)-high group had a significantly worse outcome than the V(n)-low group (5-year survival
123
3260
World J Surg (2014) 38:3257–3264
Table 1 Clinicopathological features and their correlations with the grades of venous invasion No of cases (%) Total (n = 139)
V-number classification V(n)-low (n = 98)
V(n)-high (n = 41)
p value
V-size classification V(s)-low (n = 92)
V(s)-high (n = 47)
47 (64)
26 (36)
45 (68)
21 (32)
49 (62)
30 (38)
43 (72)
17 (28)
32 (65)
17 (35)
60 (67)
30 (33)
p value
Agea (years) C67
73
54 (74)
19 (26)
\67
66
44 (67)
22 (33)
Male
79
60 (76)
19 (24)
Female
60
38 (63)
22 (37)
Right side
49
36 (73)
13 (27)
Left side
90
62 (69)
28 (31)
0.35
0.70
Sex 0.11
0.23
Tumor location 0.57
0.87
Tumor sizea C46 mm
71
49 (69)
22 (31)
\46 mm
68
49 (72)
19 (28)
0.69
131 8
95 (73) 3 (38)
36 (27) 5 (63)
0.049b
0.0026
41 (58)
30 (42)
51 (75)
17 (25)
0.032
87 (66) 5 (63)
44 (34) 3 (37)
0.99
0.18
Histologic typec Well/Mod Por/Muc Distant metastasis Positive
14
5 (36)
9 (64)
Negative
125
93 (74)
32 (26)
Positive
62
37 (60)
25 (40)
Negative
77
61 (79)
16 (21)
7 (50)
7 (50)
85 (68)
40 (32)
40 (65)
22 (35)
52 (68)
25 (32)
Nodal metastasis
a
0.012
0.71
The medians of age and tumor size were used for cutoff values
b
Fisher’s exact test
c
Well/mod indicates well-moderately differentiated; Por/Muc indicates poorly differentiated/mucinous adenocarcinoma
V(n) number of veins, V(s) size of veins
rate 59.1 vs. 89.9 %; p \ 0.0001), whereas V-size classification did not show prognostic significance (72.6 vs. 84.7 %, respectively; p = 0.080). Prognostic significance was also assessed by multivariate analysis (Table 2). Variables with prognostic significance after univariate analysis were used as variables in the multivariate analysis. V-number classification (hazard ratio [HR] 3.1; p = 0.0071) was an independent prognostic factor, as were distant metastasis (HR 5.7; p \ 0.0001) and age (HR 2.7; p = 0.014). In addition, the analysis of the disease-specific survival curves of all patients demonstrated the prognostic significance of V-number classification; the 5-year survival rates were 71.5 and 89.9 % for the V(n)-high and V(n)-low groups, respectively (p = 0.0003). Similar results were observed when the recurrence-free survival curves of patients following curative operations were compared (71.0 vs. 92.2 %; p = 0.0036).
123
Relation to postoperative recurrence Among the 129 patients undergoing curative operations (R0), 19 patients (14.7 %) experienced postoperative recurrence (Table 3). The overall recurrence rate of patients in the V(n)-high group was higher than that of patients in the V(n)-low group (p = 0.0068). With respect to primary recurrence sites, recurrence linked to hematogenous spread was more frequent in the V(n)-high group than in the V(n)-low group (p = 0.010). In contrast, V-size classification was not associated with either overall recurrence rate or recurrence rates at specific sites. Subgroup analysis of overall survival Subgroup analyses were performed (Table 4) across the various cancer stages in this study. The prognostic
World J Surg (2014) 38:3257–3264
3261
Fig. 2 a Cumulative overall survival curves for 139 patients with pT3 colorectal cancers stratified by V-number [V(n)] classification. The curves for the V(n)-low and V(n)-high groups were significantly different (p \ 0.0001). b Cumulative overall survival curves for the 139 patients with pT3 colorectal cancers stratified by V-size (V[s]) classification. The curves for the V(s)-low and V(s)-high groups were not significantly different (p = 0.080)
Table 2 Significance of clinicopathological parameters for overall survival Parameter
Multivariate analysisa
Comparison of overall survival 5-year survival rate (%)b
p valuec
HR
95 % CI
p value
73.0 vs. 88.6
0.031*
2.7
1.2–5.8
0.014*
78.9 vs. 81.4
0.94
Right vs. left side
85.7 vs. 77.7
0.31
Tumor size C46 vs. \46 mm
82.7 vs. 78.3
0.72
80.0 vs. 87.5
0.58
28.6 vs. 86.7
\0.0001*
5.7
2.4–13.7
\0.0001*
67.9 vs. 90.4
0.0087*
1.3
0.59–3.0
0.48
59.1 vs. 89.9
\0.0001*
3.1
1.4–7.0
0.0071*
72.6 vs. 84.7
0.080
Age (years) C67 vs. \67 Sex Female vs. male Tumor location
Histologic typed Well/Mod vs. Por/Muc Distant metastasis Positive vs. negative Nodal metastasis Positive vs. negative V-number classification V(n)-high vs. V(n)-low V-size classification V(s)-high vs. V(s)-low a b
Cox’s proportional hazard model Kaplan–Meier method
c
Log-rank test
d
Well/mod indicates well-moderately differentiated; Por/Muc indicates poorly differentiated/mucinous adenocarcinoma
* Statistically significant CI confidence interval, HR hazard ratio, V(n) number of veins, V(s) size of veins
significance of V-number classification was confirmed only in patients with stage II disease; the 5-year overall survival rates were 64.2 and 98.2 % for the V(n)-high and V(n)-low groups, respectively (p \ 0.0001). In contrast, adjuvant chemotherapy had a marked impact on the prognosis of
stage III patients (48.1 % [without chemotherapy] vs. 90.5 % [with chemotherapy]; p = 0.0068). Among patients with stage III V(n)-low cancers, the 5-year survival rate without adjuvant chemotherapy (n = 11) was much worse than that with adjuvant
123
3262
World J Surg (2014) 38:3257–3264
Table 3 Postoperative recurrence in patients with curative resection (R0) Number of cases (%) Total (n = 129)
V-number classification V(n)-low (n = 94)
V(n)-high (n = 35)
19 (15)
9 (10)
10 (29)
Liver
11 (9)
4 (4)
7 (20)
Lung
8 (6)
5 (5)
17 (13) 2 (2)
Overall
p value
V-size classification
p value
V(s)-low (n = 87)
V(s)-high (n = 42)
0.0068
10 (12)
9 (21)
0.14
0.0090b
6 (7)
5 (12)
0.34
3 (9)
0.68b
6 (7)
2 (5)
0.99b
8 (9)
9 (26)
0.010
10 (11)
7 (17)
0.42
1 (1)
1 (3)
0.47b
0 (0)
2 (5)
0.10b
Primary recurrence sitea Hematogenous
a
Total
Non-hematogenous Local a
Primary recurrence was not limited to a single organ. Primary recurrence was found in liver and lung for two patients
b
Fisher’s exact test
Table 4 Subgroup analysis of overall survival
Parameter (number of cases)
Comparison of overall survival 5-year survival ratea (%)
p valueb
89.7 vs. 93.6
0.20
64.2 vs. 98.2
\0.0001c
77.6 vs. 98.0
0.0011
91.4 vs. 100
–
66.0 vs. 91.6
0.040
69.0 vs. 85.1
0.27
85.7 vs. 76.0
0.27
48.1 vs. 90.5
0.0068c
12.5 vs. 50.0
0.51
33.3 vs. 20.0
0.76
28.6 vs. 28.6
0.50
75.0 vs. 10.0
0.27
Stage II patients Age (years) C67 (42) vs. \67 (33) V-number classification V(n)-high (15) vs. V(n)-low (60) V-size classification V(s)-high (24) vs. V(s)-low (51) Adjuvant chemotherapy Without chemo (74) vs. with chemo (1) Stage III patients Age (years) C67 (23) vs. \67 (27) V-number classification V(n)-high (17) vs. V(n)-low (33) V-size classification V(s)-high (16) vs. V(s)-low (34) Adjuvant chemotherapy Without chemo (16) vs. with chemo (34) Stage IV patients Age (years) C 67 (8) vs. \ 67 (6) V-number classification V(n)-high (9) vs. V(n)-low (5) a
Kaplan–Meier method
b
Log-rank test
c
Independent predictors disclosed by multivariate analysis
123
V-size classification V(s)-high (7) vs. V(s)-low (7) Residual tumor R0 (4) vs. R1 or 2 (10)
World J Surg (2014) 38:3257–3264
chemotherapy (n = 22) (56.3 % [without chemotherapy] vs. 95.0 % [with chemotherapy]; p = 0.020). For reference, the 5-year survival rate of stage II patients with V(n)low cancers (n = 59) was 98.2 %, despite adjuvant chemotherapy not being used. Among patients with stage III V(n)-high cancers, the 5-year survival rate without adjuvant chemotherapy (n = 5) was 40.0 %, and increased to 82.5 % with the use of adjuvant chemotherapy (n = 12); however, the difference was not significant (p = 0.15). Notably, the 5-year survival rate of patients with stage II V(n)-high cancers (n = 15) not receiving adjuvant chemotherapy was 64.2 %.
Discussion Van Wyk et al. [12] described a review article that summarized the current status of detection methods for components of vascular invasion and its clinical significance. They concluded that, when specific staining is used, blood vessel invasion appears to be a reliable independent prognostic factor. In our prospective observational study, we enrolled 139 patients with pT3 colorectal cancer to clarify the prognostic significance of the number of venous invasions using section slides stained with EVG. The results of both univariate and multivariate analyses indicate that V-number classification was an independent predictor of patient outcome after surgery. However, although V-size classification had a positive relation with V-number classification, it was not a suitable predictive marker for patient prognosis. Most pathologists accept that venous invasion can predict hematogenous recurrence based on the biological plausibility that cancer cells infiltrating venous vessels can be dissociated from the original tumor, thereby facilitating spread to other organs. Iinuma et al. [13] reported that the pathological findings of venous invasions significantly correlated with the presence of isolated tumor cells in blood samples. Furthermore, another study reported that detecting circulating tumor cells in the peripheral blood was as a useful tool for determining the patients at increased risk for recurrence [14]. These studies support the hypothesis that venous invasion is an essential step in the process of hematogenous metastasis, as does our prospective observational study, which demonstrates that venous invasion precedes hematogenous recurrence. In this study, we focused on the number and size of venous invasions, and identified clinical values for these grading systems, but were unable to determine why these values showed diversity. However, the V(s)-high group frequently contained cancer nodules in venous vessels, which appeared to reflect the expansive growth and
3263
cohesive property of cancer cells. The concept that only cancer stem cells leaving the original tumor site can be precursors of metastasis is widely accepted based on the cancer stem cell theory [15, 16]. According to this model, venous invasion by cancer stem cells can therefore be a powerful prognostic indicator. That said, cancer stem cells are characterized by low proliferation and loss of epithelial features [15, 17], which are unlikely to form tumor nodules with expansive growth and cohesive properties. On the other hand, Bugg et al. [18] suggested that the presence of magnetic resonance imaging (MRI)-detected venous invasion is a poor prognostic feature of rectal cancer. They reported that one-half of patients with MRI-detected invasion of the superior rectal vein had encountered distant recurrence [18]. Although our data implied that the size of venous invasion is not associated with patients’ prognosis, careful reconsideration is necessary to determine the significance of cancer involvement of named colic or rectal veins. The subgroup analyses demonstrated that V-number classification had a prognostic impact in patients with stage II disease, but not in those with stage III or IV disease. It is conceivable that venous invasion predicts hematogenous recurrence, and that there is therefore nothing to predict in patients who already have distant metastasis in stage IV disease. In patients with stage III disease, the prognostic impact of adjuvant chemotherapy was marked, which possibly reduced the prognostic impact of venous invasion. Clinically, it was problematic that 20 % of the stage II patients who were considered unsuitable for adjuvant chemotherapy were also judged to be V(n)-high. Because the 5-year survival rate was 82.5 % in patients with stage III V(n)-high cancers after adjuvant chemotherapy, the patients with stage II V(n)-high disease with a 5-year survival rate of 64.2 % might also need to be offered adjuvant chemotherapy. Data from the literature on the prognostic significance of vascular invasion in stage II patients are still insufficient. However, Betge et al. [19] disclosed that venous invasion is a powerful predictor of both disease progression and cancer-related death and increased the HR by a factor of approximately 5 [19]. Accordingly, they also advocated consideration of venous invasion when stratifying patients with stage II colorectal cancer for adjuvant treatment. There are potential limitations to this study. First, we evaluated only pT3 cancers. A larger trial including cancers with various T categories will confirm our conclusions. Second, this was a prospective observational study, and postoperative adjuvant chemotherapy and treatments for recurrence were not homogeneous. Adjuvant treatment may be affected by pathological diagnosis, physical condition, and patient preference. In this period, patients with node-positive cancers were offered adjuvant
123
3264
chemotherapy, which possibly reduced the impact of nodal metastasis on patient outcomes. Third, the relatively small number of patients with stage II V(n)-high cancers is a limitation, and further work is needed to establish the clinical utility. In the present study, we demonstrated that the number of venous invasions might provide crucial prognostic information for patients with pT3 colorectal cancer, especially at stage II. These novel findings may be clinically applicable for selecting suitable candidates for intensive adjuvant chemotherapy. Acknowledgments This work was supported in part by the National Cancer Center Research and Development Fund (23-A-11 and 23-A26) and by the Japan Society for the Promotion of Science KAKENHI (23501302, 25462074, and 26462032). Conflict of interest The authors certify that we have no associations, commercial or otherwise, that may pose a conflict of interest.
References 1. Minsky BD, Mies C, Richt TA et al (1988) Potentially curative surgery of colon cancer: the influence of blood vessel invasion. J Clin Oncol 6:119–127 2. Talbot IC, Ritchie S, Leighton M et al (1980) The clinical significance of invasion of veins by rectal cancer. Br J Surg 67:439–442 3. Petersen VC, Baxter KJ, Love SB et al (2002) Identification of objective pathological prognostic determinants and models of prognosis in Duke’s B colon cancer. Gut 51:65–69 4. Chafai N, Chan CL, Bokey EL et al (2005) What factors influence survival in patients with unresected synchronous liver metastases after resection of colorectal cancer? Colorectal Dis 7:176–181 5. Sobin LH, Gospodarowisz M, Wittekind C (2009) TNM: classification of malignant tumours, 7th edn. Wiley, New York 6. Japanese Society for Cancer of Colon and Rectum (2009) In: Japanese classification of colorectal carcinoma (2nd English edn). Kanehara & Co. Ltd, Tokyo, p. 27
123
World J Surg (2014) 38:3257–3264 7. Ouchi K, Sugawara T, Ono H et al (1996) Histologic features and clinical significance of venous invasion in colorectal carcinoma with hepatic metastasis. Cancer 78:2313–2317 8. Sternberg A, Mizrahi A, Amar M et al (2006) Detection of venous invasion in surgical specimens of colorectal carcinoma: the efficacy of various types of tissue blocks. J Clin Pathol 59:207–210 9. Sato T, Ueno H, Mochizuki H et al (2010) Objective criteria for the grading of venous invasion in colorectal cancer. Am J Surg Pathol 34:454–462 10. Krasna MJ, Flancbaum L, Cody RP et al (1988) Vascular and neural invasion in colorectal carcinoma. Cancer 61:1018–1023 11. Sato T, Shinto E, Hashiguchi Y et al (2006) Prognostic significance of lymphatic and venous invasions in pT3 colorectal carcinoma, quantified with a double staining combining CD34 immunostaining and elastica Staining. Jpn J Gastroenterol Surg 39:1571–1576 12. van Wyk HC, Roxburgh CS, Horgan PG et al (2014) The detection and role of lymphatic and blood vessel invasion in predicting survival in patients with node negative operable primary colorectal cancer. Crit Rev Oncol Hematol 90:77–90 13. Iinuma H, Watanabe T, Mimori K et al (2011) Clinical significance of circulating tumor cells, including cancer stem-like cells, in peripheral blood for recurrence and prognosis in patients with Dukes’ stage B and C colorectal cancer. J Clin Oncol 29:1547–1555 14. Iinuma H, Okinaga K, Egami H et al (2006) Usefulness and clinical significance of quantitative real-time RT-PCR to detect isolated tumor cells in the peripheral blood and tumor drainage blood of patients with colorectal cancer. Int J Oncol 28:297–306 15. Willis ND, Przyborski SA, Hutchison CJ et al (2008) Colonic and colorectal cancer stem cells: progress in the search for putative biomarkers. J Anat 213:59–65 16. Yeung TM, Mortensen NJ (2009) Colorectal cancer stem cells. Dis Colon Rectum 52:1788–1796 17. Mani SA, Guo W, Liao MJ et al (2008) The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133:704–715 18. Bugg WG, Andreou AK, Biswas D et al (2014) The prognostic significance of MRI-detected extramural venous invasion in rectal carcinoma. Clin Radiol 69(6):619–623 19. Betge J, Pollheimer MJ, Lindtner RA et al (2012) Intramural and extramural vascular invasion in colorectal cancer: prognostic significance and quality of pathology reporting. Cancer 118:628–638
