Keywords
extramural venous invasion - MRI - rectal cancer
Introduction
Colorectal cancer is a common cancer affecting the Indian population, with up to 30%
being upfront metastatic.[1]
[2] The signet ring cell variety of rectal carcinoma is particularly aggressive, with
a high propensity for peritoneal metastasis and involvement of a younger age group.[3]
[4] Although the worldwide incidence of this histologic variant is 1%, Indian population
shows a higher incidence varying from 11 to 13% in different studies.[2]
[5]
Extramural venous invasion (EMVI) i.e., spread of tumor into perirectal vessels, is
an established risk factor for distant metastasis, with reduced disease-free survival
(DFS) and overall survival (OS).[6]
[7]
[8] Our literature search did not show any imaging study in the Indian population with
regards to the significance of EMVI in rectal cancer. This would be interesting to
see because of the frequently seen advanced disease at presentation and the higher
proportion of signet ring cell cases in the Indian population, increasing the confounding
factors for poor outcomes. Furthermore, while EMVI has prognostic significance, it
is possible that patients with thicker and longer extent of involvement have a worse
prognosis compared with patients with mild involvement.[8]
[9] Thus, the objective of our study was to determine the prognostic significance of
EMVI in rectal cancers in the Indian population, and to evaluate whether the length
and thickness of involvement correlated with prognosis.
Materials and Methods
Subjects
This is an Institutional Review Board approved retrospective study. We identified
166 consecutive cases of operable non-metastatic pathologically proven rectal cancer
from our colorectal disease management group database from the year 2011 to 2015.
Patients presenting with non-metastatic rectal cancer and baseline MRI available were
included in the study. MRI performed in outside institutes was included if DICOM images
were available on PACS. Exclusion criteria were suboptimal image quality, patients
that progressed during the neoadjuvant chemoradiation and did not undergo curative
treatment, and patients lost to follow-up.
Clinical and Histopathology Data
The patient demographic data, histopathology, treatment, and follow-up details were
extracted from the institutional electronic medical records. All definitive surgeries
were performed by dedicated gastrointestinal oncosurgeons at our institute. All patients
were treated with neoadjuvant chemoradiation (CTRT) followed by definitive surgery
and were followed up for a median period of 37 months (range: 2–71 months).
Imaging and Image Analysis
MR images were acquired on 1.5T (GE Healthcare, Milwaukee, United States), 1.5T (Philips
Medical Systems, Eindhoven, Netherlands), or 3T (GE Healthcare, Milwaukee, United
States) machines in our institute using the institutional rectal MRI protocol. This
included large FOV (field of view) T1W and T2W axial, T2W sagittal sequences of the
pelvis, small FOV thin section oblique axial T2W sequence perpendicular to the plane
of the rectal tumor and oblique coronal T2W sequence parallel to the plane of the
tumor. No intravenous contrast was administered as per institutional protocol. For
contrast-enhanced MRI studies performed outside, the contrast-enhanced sequences were
not assessed.
The images were retrospectively and systematically reviewed on Centricity PACS (GE
Healthcare, Milwaukee, United States) workstation by three oncoradiologists in consensus,
with 11, 9, and 4 years of experience, respectively, who were blinded to the histopathology
findings and patient outcome. EMVI was identified by expansion or irregularity of
vessels, loss of normal vascular flow void and intraluminal intermediate tumor signal
intensity, contiguous or separate from the main tumor.[10] EMVI was graded with a 0 to 4 scoring system as suggested by Jhaveri et al.[11] Score 0 denoted no vessel in vicinity of extramural tumor penetration, score 1 denoted
vessels with normal caliber and without definite tumor signal intensity, score 2 denoted
slightly expanded vessels without definite tumor signal intensity, score 3 denotes
intermediate tumor signal intensity within expanded vessels, and score 4 denotes obvious
irregular vessel contour or nodular expansion of vessel by definite tumor signal ([Figs. 1] and [2]). For the purpose of calculation, scores 0 to 2 were considered negative for EMVI,
while scores 3 and 4 were considered positive. The length and the thickness of the
involved vessel were also recorded ([Fig. 3]). Tumor within 1 mm of the mesorectal fascia was considered CRM (circumferential
resection margin) positive.
Fig. 1 (a) Coronal T2-weighted image showing rectal wall thickening (*) with no vessel in the
vicinity of the tumor, Grade 0 EMVI. (b) Coronal T2-weighted image showing rectal tumor (*) with a normal caliber vessel
in its vicinity (white arrow), Grade 1 EMVI. EMVI, extramural venous invasion.
Fig. 2 (a) Coronal T2-weighted image showing rectal tumor (*) with a prominent normal-appearing
vessel in its vicinity, Grade 2 EMVI. (b) Axial T2-weighted image showing rectal tumor (*) with an expanded vessel showing
regular contour and tumor signal intensity within its lumen (white arrow), Grade 3 EMVI. EMVI, extramural venous invasion.
Fig. 3 Illustration showing measurement of maximum length (black double arrow) and thickness (white double arrow) of EMVI. EMVI, extramural venous invasion.
Statistical Analysis
Statistical analysis was performed using SPSS (the statistical package for social
sciences), IBM Corp, released 2017, IBM SPSS Statistics for Windows, Version 25.0.
Armonk, NY: IBM Corp. Kaplan-Meier curves (95% CI) were used to determine DFS, distant-metastases
free survival (DMFS), and OS. Univariate analysis was performed by comparing groups
with log-rank test. All analysis was two sided and significance was set at p-value of 0.05.
Results
A total of 117 patients were included in the study ([Fig. 4]). 45/117 (38.5%) cases were females and 72 (61.5%) were males with a median age
of 51 years. The baseline characteristics of the patients are given in [Table 1]. EMVI was present in 34/117 (29%) cases and absent in 83 (71%) cases. Twelve patients
had grade 3 EMVI and 22 had grade 4 EMVI. Overall, 36/117 (30%) developed distant
metastasis. 15/117 (12.8%) developed local recurrence, with 8/15 of these also having
distant metastasis ([Table 2]). Three-year DFS and OS were 67 and 82%, respectively in EMVI positive cases, compared
with 74 and 82% in EMVI negative cases. However, this did not reach statistical significance
([Table 3]; [Figs. 5] and [6]).
Fig. 4 Consort diagram of case selection.
Table 1
Baseline characteristics of primary tumor determined on MRI
|
Baseline characteristics
|
Stage
|
No. of patients out of n = 117
|
|
mrT stage
|
<T3
|
6 (5%)
|
|
T3
|
91 (78%)
|
|
T4
|
20 (17%)
|
|
mrN stage
|
N0
|
25 (21%)
|
|
N1
|
51 (44%)
|
|
N2
|
41 (35%)
|
|
mrCRM status
|
Positive
|
63 (53.9%)
|
|
Negative
|
54 (46.1%)
|
Table 2
Distribution of EMVI positive and negative cases and their correlation to distant
metastases, local recurrence, and survival
|
EMVI status
|
No. of cases
|
Grade of EMVI
|
No. of cases
|
Distant metastasis Grade-wise
|
Distant metastases (DM) Total
|
Local recurrence (LR) Total
|
Deaths total
|
|
Negative
|
83/117
|
0
|
24
|
7
|
22/83 (26%)
|
10/83 (12%)
|
11/83 (13%)
|
|
|
1
|
49
|
10
|
|
|
|
|
|
2
|
10
|
5
|
|
|
|
|
Positive
|
34/117
|
3
|
12
|
2
|
14/34 (41%)
|
5/34 (14.7%)
|
7/34 (20.5%)
|
|
|
4
|
22
|
12
|
|
|
|
Table 3
Three-year DFS and OS in EMVI positive and negative cases in entire study population
and in the signet ring cell negative subset
|
Patients
|
EMVI
|
3-y DFS
|
p
|
3-y OS
|
p
|
|
All cases (n = 117)
|
Negative
|
74.2%
|
0.180
|
87%
|
0.211
|
|
Positive
|
67.2%
|
|
82%
|
|
|
Signet ring cell negative cases (n = 103)
|
Negative
|
79%
|
0.109
|
91%
|
0.22
|
|
Positive
|
72%
|
|
88.6%
|
|
Abbreviations: DFS, disease-free survival; EMVI, extramural venous invasion; OS, overall
survival.
Fig. 5 Plot of Kaplan-Meier curve (95% CI) estimating disease free survival (DFS) in EMVI
negative and positive cases in all patients (n = 117); p = 0.211. EMVI, extramural venous invasion.
Fig. 6 Plot of Kaplan-Meier curve (95% CI) of overall survival (OS) in EMVI negative and
positive cases in all patients (n = 117); p = 0.211. EMVI, extramural venous invasion.
Fourteen out of 34 (41%) EMVI-positive cases developed distant metastasis, compared
with 22/83 (26%) EMVI-negative cases ([Table 2]). The difference, however, was not statistically significant (p = 0.146). After excluding signet-ring cell positive cases (n = 14), presence of EMVI showed significant correlation with distant metastasis free
survival (p = 0.046) ([Fig. 7]), with 12/29 (41%) cases with EMVI developing metastases compared with 16/74 (22%)
patients without EMVI. However, there was no statistical difference in the overall
DFS and OS ([Table 3]).
Fig. 7 Plot of Kaplan-Meier curve (95% CI) of distance metastasis free survival (DMFS) in
EMVI negative and positive cases (in signet-ring cell negative cases, n = 103); p = 0.046. EMVI, extramural venous invasion.
The median thickness of EMVI was 6 mm and median length of EMVI was 14 mm in patients
who developed distant metastasis, compared with 5 and 11 mm, respectively, in patients
who did not develop distant metastasis. This was not statistically significant.
Discussion
MRI is the imaging modality of choice for staging and restaging of rectal cancer.
The prognosis of rectal cancer in terms of survival and likelihood of recurrence depends
on several factors such as histologic grade, T category, N category, CRM involvement,
and EMVI.[12]
[13]
[14]
[15]
Traditionally, venous invasion was a histopathological concept.[16] Many studies have shown pathological venous invasion to portend worse prognosis.[17]
[18] A systematic review by Chand et al[19] of 14 pathological studies demonstrates venous invasion and more specifically EMVI
to be associated with worse survival. The 5-year survival ranged from 20 to 33% in
EMVI positive cases to 45 to 73% in EMVI negative cases.[16]
[20]
[21]
MRI has moderate sensitivity and good specificity for detecting EMVI.[22] Multiple studies have demonstrated that mrEMVI correlates with worse prognosis.[6]
[8]
[23]
[24]
[25]
[26] A recent meta-analysis of 1,262 patients demonstrated a prevalence of mrEMVI in
about one-third patients, similar to our study.[6] They found an almost fourfold increase in the risk of developing metastases in patients
with mrEMVI present at baseline. Various studies demonstrate development of distant
metastases in 6.7 to 23% mrEMVI negative cases compared with 24.5 to 55% in mrEMVI
positive cases.[8]
[24]
[27] Our India-specific data shows a similar trend, with 26% mrEMVI negative and 41%
mrEMVI positive developing distant metastases. The percentage of mrEMVI negative patients
developing subsequent metastatic recurrence is slightly higher than other studies;
the reasons for this could include the higher incidence of signet cell tumors in the
Indian population and possible relatively advanced tumors at the time of presentation.
Gu et al[28] found a 3-year DFS of 60% for mrEMVI scores 3 and 4 and 86% with mrEMVI scores 0
to 2. Similarly, in a study of 142 patients, Smith et al[25] observed a DFS of 35% at 3 years for patients with mrEMVI scores of 3 to 4, compared
with 74% for those with mrEMVI 0 to 2. Our study results showed a 3-year DFS of 67%
for EMVI 3 to 4 and 74% for EMVI grades 0 to 2, although this was not statistically
significant.
There is mixed data with regards to correlation of mrEMVI with OS, with some studies
not finding statistically significant correlation.[28]
[29] While others demonstrating significantly worsened survival in mrEMVI positive patients,[24]
[27] our study findings showed a 3-year OS of 82% for mrEMVI positive cases, which was
lesser than the OS of 87% for mrEMVI negative patients, but not statistically significant.
India has a unique profile of rectal cancer presentation, with a high incidence of
advanced cancers and aggressive signet ring tumors. In patients with such high-risk
features at presentation, the degree of incremental worsening of prognosis due to
the presence of mrEMVI is uncertain. We observed a nonsignificant trend to worsened
prognosis in mrEMVI positive cases; the lack of significance could be due to the small
numbers. EMVI showed significant correlation with DMFS (p = 0.046) when the signet ring cell cases were excluded, but again did not correlate
with DFS or OS.
A pathological study has demonstrated strong correlation between venous invasion involving
thick-walled veins rather than thin-walled veins with respect to the incidence of
liver metastases.[18] Bugg et al and Sohn et al[8]
[9] demonstrated that involvement of larger caliber vessels (≥3 mm) by EMVI leads to
increased risk of metastasis. We similarly hypothesized that increased thickness and
length of involvement of perirectal vessels by EMVI could correlate with a higher
risk of metastases ([Fig. 8]). We did not observe a significant difference in the median length and thickness
of EMVI of patients who developed distant recurrence and those who did not, although
the median length and thickness were slightly higher in the former group. Our numbers
were small, and larger studies would be needed to evaluate this finding further.
Fig. 8 (a) Axial T2-weighted image showing grade 4 EMVI (white double arrow), measuring 9 mm in thickness. (b) Sagittal T2-weighted image showing grade 4 EMVI (black double arrow), measuring 36 mm in length. EMVI, extramural venous invasion.
Our study had a few limitations. It was a retrospective study, although the radiologists
were blinded to the histopathology and final outcome. The sample size was small with
a limited number of EMVI positive cases. There may also be a referral bias as ours
is a tertiary care cancer. Our institute protocol does not include contrast-enhanced
study for rectal cancer cases. This is as per the Society of Abdominal Radiology (SAR)
and the European Society of Gastrointestinal and Abdominal Radiology (ESGAR) guidelines.
However, some studies do suggest better detection of EMVI using gadolinium contrast
enhanced sequences.[30]
[31]
[32]
[33] Jhaveri et al found that contrast-enhanced sequences helped in re-stratification
of the equivocal Grade 2 EMVI cases which did not show overt tumor signal within a
dilated vein. EMVI was seen as a partial or complete filling defect within the contrast-filled
vessel lumen, which could then be re-classified as Grade 3 if present or Grade 1 if
absent.[11]
[34] The ESGAR panel opinion was that dynamic contrast enhanced sequence should be considered
a research tool and not adopted in routine practice. The panel acknowledged that it
may help in select cases to improve tumor conspicuity in the post-treatment setting
and for the assessment of mucinous tumors.[30]
[31]
[32]
[33]
[34]
[35]
In conclusion, we did not find a significant correlation between mrEMVI and DFS or
OS in patients with non-metastatic disease at baseline, although a higher percentage
of patients with EMVI developed distant recurrence compared with those without EMVI.
Larger multicentric studies are needed to evaluate the significance of these findings.
There was, however, a statistically significant worsening of DMFS in non-signet-ring
cell cancers.
