Extramural venous invasion detected by MDCT as an adverse imaging feature for predicting synchronous metastases in T4 gastric cancer

Abstract

Background

Extramural venous invasion (EMVI) is defined histologically as the active invasion of tumor cells to the lumens of mesenteric vessels beyond the muscularis propria in advanced gastrointestinal cancer, resulting in distant metastases.

Purpose

To determine the association between synchronous metastatic disease in patients with T4 gastric cancer and EMVI detected on contrast-enhanced multiple-row detector computed tomography (MDCT).

Material and Methods

A total of 152 patients with T4 gastric carcinoma were retrospectively reviewed and divided into EMVI-positive and EMVI-negative groups where EMVI, as detected on MDCT, was defined as a tubular or nodular soft tissue thickening extending from the tumor along the vessels of the mesentery. Synchronous metastases were detected by MDCT and/or confirmed by postoperative diagnosis. Logistic regression analyses were performed to analyze the predictive factors of synchronous metastases in gastric cancer.

Results

Synchronous metastases were found in 47 of 152 (30.9%) patients with T4 gastric cancer. Thirty-one of 77 (40.3%) patients in the EMVI-positive group had evidence of metastases compared to 16 (21.3%) of 75 patients in the EMVI-negative group (P = 0.019). Synchronous metastases were significantly associated with EMVI with an odds ratio (OR) of 2.250 (95% CI, 1.072–4.724).

Conclusion

EMVI-positive tumors, as an adverse imaging feature, were significantly associated with synchronous metastases in patients with T4 gastric cancer.

Keywords

Multiple-row detector computed tomography (MDCT)contrast enhancement extramural venous invasion (EMVI)gastric cancer synchronous metastases

Introduction

Extramural venous invasion (EMVI) is defined as the presence of tumor cells within endothelium-lined vessels beyond the muscularis propria of the gastrointestinal tract, the detection of which can serve as an adverse imaging feature to better inform cancer prognosis (1). It is accepted that EMVI allows tumor cells to embolize vessels via portal circulation in gastric cancer, resulting in distant metastases through hematogenous spread (2,3). Previous histopathological studies have shown that compared to intramural venous invasions, lymphatic invasions, and extramural lymphatic invasions, EMVI is an independent prognostic variable in predicting prognostic events in gastrointestinal cancer (4). Multiple-row detector computed tomography (MDCT) imaging with contrast enhancement can be used as an imaging tool to identify EMVI as a tubular or nodular soft tissue thickening extending from the tumor along the vessels of the mesentery (5). EMVI as an adverse imaging feature detected on rectal magnetic resonance imaging (MRI) can be used to identify patients at increased risk for synchronous metastatic disease before surgery is attempted (6). Regarding gastric cancer, however, no study has attempted to stratify patients with a high risk of synchronous distant metastases based on preoperative EMVI status.

The aim of this study was to determine the difference in incidence of synchronous metastases between EMVI-positive and EMVI-negative (defined by MDCT) patients with T4 gastric cancer. Furthermore, we attempted to determine the factors associated with synchronous metastases in patients with T4 gastric cancer.

Material and Methods

Patients

Our institutional review board approved this retrospective case-control study and waived the requirement for informed consent. Between January 2009 and December 2013, our institution’s electronic gastrointestinal cancer registries were searched to identify patients with primary gastric cancer confirmed by biopsy. MDCT scans of the thorax, abdomen, and pelvis are routinely conducted as part of baseline staging before neoadjuvant chemotherapy and/or curative surgery at our institution. Inclusion criteria were adult patients, who had been staged as having T4 gastric cancer on abdominal MDCT scans in accordance with the criteria of the 7th AJCC. Patients were excluded for having other synchronous malignant tumors and MDCT imaging with unsatisfactory imaging quality.

MDCT technique and image acquisition

CT images were acquired on 64-slice MDCT scanners (Light Speed Volume CT; GE Healthcare, Waukesha, WI, USA). After an 8-h fast and 5 min prior to the acquisition of MDCT images, all patients were asked to drink 600–800 mL of water in order to adequately distend the stomach. MDCT images were acquired using the following parameters: 120 kV, 240–260 mAs, collimations of 64–0.6, and slice thickness and increments of 5 mm. The late arterial and portal venous phase CT data acquisition started at 10 s and 45 s after the trigger threshold (100 Hounsfield units [HU] on the abdominal aorta) was reached. Intravenous iodinated contrast was administrated (100 mL iopromide 370 mg J/mL; Bayer Schering Pharma, Berlin, Germany) with a power injector (Missouri XD2001, Ulrich GmbH & Co., Buchbrunnenweg, Ulm, Germany) at a rate of 2.5 mL/s through an antecubital vein. Imaging coverage ranged from the inlet of the thorax to the symphysis pubis. Axial, sagittal, and coronal reconstructions with 1.25 mm slices were performed on a workstation (Advantage Workstation 4.3; GE Healthcare).

Definitions of CT imaging characteristics

Two board-certified radiologists (10 and 13 years of experience in abdominal radiology, respectively) reviewed the MDCT images independently blinded to clinical outcomes and pathological records. A third board-certified radiologist (20 years of experience in abdominal radiology) reviewed the images and made the final decision if a consensus between the two observers could not be reached.

T category detected with MDCT (ctT) was defined according to the criteria of the 7th AJCC. ctT4 category was defined as an irregular or nodular outer margin of the outer layer with (T4b) or without (T4a) direct invasion of the adjacent organs (7).

N category detected with MDCT (ctN) was defined as N-positive or N-negative. A lymph node was defined as metastatic, or N-positive, if the short diameter of the regional lymph node was ≥ 8 mm (8). No evidence of a metastatic lymph node was defined as N-negative.

EMVI category detected on MDCT (ctEMVI) was defined as ctEMVI-positive and ctEMVI-negative based on the imaging characteristics on MDCT. ctEMVI-positive was defined as contiguous tubular or nodular soft tissue extending from the tumor and exhibiting filling defects within distended veins along the vessels of the mesentery on MDCT (5). No evidence of ctEMVI-positive was defined as ctEMVI-negative.

Tumor location/growth pattern was stratified according to the following three individual classifications: proximal non-diffuse (the bulk of tumor was located in the gastric cardia, which may have extended up to the gastroesophageal junction and a small portion of the distal esophagus), distal non-diffuse (the bulk of the tumor was usually located in the distal stomach and may have extended up to the mid-body of the stomach or down to the pylorus), and diffuse (the tumor location may have been anywhere in the stomach) (9).

Tumor size was measured as the longest diameter of the mass on axial MDCT imaging.

Synchronous metastases were defined as metastatic lesions detected at the time of the diagnosis of the initial primary gastric cancer. Metastatic lesions were detected using MDCT scans of the thorax, abdomen, and pelvis, and the lesions were confirmed by follow-up MDCT or by postoperative histopathological diagnosis.

Histopathology technique and evaluation

Each specimen of radical surgery or biopsy examination was fixed in formalin for 24 h. Hematoxylin and eosin (H&E) stained slides were reviewed using a microscope (Olympus BX51; Olympus, Tokyo, Japan) to evaluate the histological type and differentiation of each tumor.

Histological type/tumor differentiation was classified as well differentiated adenocarcinoma, moderately differentiated adenocarcinoma, poorly differentiated adenocarcinoma, signet-ring cell carcinoma, and mucinous adenocarcinoma.

Statistical analysis

Inter-observer agreements regarding the presence or absence of ctEMVI, as well as the ctT/N/M categories, were calculated using Kappa statistical analyses shown as weighted κ values. According to Landis and Loch, κ values < 0.4 indicate poor agreement; 0.4–0.6, moderate agreement; 0.6–0.8, good agreement; and values > 0.8, indicate excellent agreement (10).

Categorical variables were described as number (n) and percentage (%). For quantitative variables that were not normally distributed, the median and interquartile ranges (IQR) were reported.

Associations between ctEMVI and conventional prognostic factors such as ctN status, tumor location/growth pattern, histology type/tumor differentiation, and tumor size were analyzed using Chi-squared tests.

Chi-squared tests were used in univariate analyses to assess associations between synchronous metastases and covariates including age, gender, ctN status, tumor location/growth pattern, histology type/tumor differentiation, tumor size, and ctEMVI status. Covariates with P < 0.1 were chosen to build a multivariate logistic regression model, with a stepwise selection procedure. The odds ratios (OR) with 95% confidence intervals (CI) were calculated.

All statistical analyses were performed using MedCalc computer software, Version 11.2 (Mariakierke, Belgium). All reported P values were two-sided with a significance level set at α < 0.05.

Results

Demographics

A total of 152 patients (Fig. 1) with ctT4 gastric cancer as defined on MDCT were included in this study; all had imaging available for review and further analyses. The median age was 65.0 years (IQR, 56.5–73.0 years). There were 52 female patients (34.2%) and 100 male patients (65.8%). Excellent and good agreements were obtained between the two observers regarding the identification of ctM, ctEMVI, and ctN categories with κ values of 1.000, 0.711, and 0.720, respectively.

Fig. 1.

Flow chart of patient inclusion.

Synchronous metastases found in patients with ctT4 gastric carcinoma

Synchronous metastases were detected in a total of 47 (30.9%, 47/152) patients (Fig. 2). Synchronous metastases were detected in 38 (80.9%, 38/47) patients by MDCT and confirmed by follow-up MDCT, including peritoneal/great omental metastases in 20 cases, liver metastases in nine cases, lung metastases in three cases, and ovarian metastases in one case. Three patients had synchronous metastases in both the peritoneum and liver. In addition, two patients had both peritoneal metastases and ovarian metastases. However, MDCT missed the diagnosis of eight patients (17.0%, 8/47) who had peritoneal metastases and one patient (2.1%, 1/47) who had ovarian metastases, all of whom were diagnosed based on pathological analyses after curative surgery.

Fig. 2.

Composition of synchronous metastatic patients.

Analysis on the association between the ctEMVI and the covariates

Among the entire study population, 50.7% (77/152) patients were ctEMVI-positive (Fig. 3). There was no statistically significant relationship between ctEMVI and the covariates including age, gender, ctN status, histological type/tumor differentiation, and tumor size. There was a statistically significant relationship between tumor location/growth pattern and ctEMVI (P = 0.026, Table 1).

Fig. 3.

Extramural venous invasion (EMVI) was detected by MDCT in gastric carcinoma. (a) A transverse CT image with contrast shows the tubular soft tissue filling defects within the vessel (straight arrow), which had irregular contours and an enlarged caliber compared to the normal vessel (curved arrow). The metastatic lymph nodes (blank arrow) were detected between the stomach and peritoneal vessels. (b) A transverse CT image with contrast shows the nodular soft tissue within the adjacent vessel (straight arrow) and shows that the caliber of the vessel was slightly expanded. A liver metastasis (black arrow) was also detected on the caudate lobe. (c) A transverse CT image with contrast shows that gastric cancer infiltrated an extramural vein resulting in irregular contours and nodular expansion of the vessel (straight arrows).

Table 1.

Cohort demographics by the absence and presence of extramural venous invasion detected on MDCT (ctEMVI).

	ctEMVI Absent		ctEMVI Present		Total	χ²	P value
	n	%	n	%	Total	χ²	P value
Age (years)
≤65	43	56.6	33	43.4	76
>65	32	42.1	44	57.9	76	2.632	0.105
Gender
Female	27	51.9	25	48.1	52
Male	48	48.0	52	52.0	100	0.083	0.773
ctN status
Negative	31	58.5	22	41.5	53
Positive	44	44.4	55	55.6	99	2.192	0.139
Location/growth pattern
Distal non-diffuse	54	56.8	41	43.2	95
Proximal non-diffuse/diffuse	21	36.8	36	63.2	57	4.929	0.026
Histology
Well/moderate	9	36.0	16	64.0	25
Poor/signet/mucin	66	52.0	61	48.0	127	1.540	0.215
Size (longest diameter) (cm)
<5.65*	43	56.6	33	43.4	76
≥5.65	32	32.9	44	57.9	76	2.632	0.105

Median of longest tumor diameter is 5.65 cm.

Analysis on the association between synchronous metastases and covariates

ctEMVI category and ctN were associated with synchronous metastases in patients with ctT4 gastric cancer (Table 2). In the ctEMVI-positive group, 40.3% (31/77) of patients developed synchronous metastases; in the ctEMVI-negative group, only 21.3% (16/75) of patients developed synchronous metastases (Table 2). Among ctT4 gastric cancer patients, the odds of synchronous metastases in ctEMVI-positive patients were 2.250 times higher compared to ctEMVI-negative patients (OR = 2.250, 95% CI, 1.072–4.724, P = 0.032, Table 3). Patients who were N-positive were 4.148 times more likely to have synchronous metastases compared to those who were N-negative (OR = 4.148, 95% CI, 1.684–10.218, P = 0.002, Table 3).

Table 2.

Association analysis of synchronous metastases in ctT4 gastric cancer patients.

	No synchronous metastases		With synchronous metastases		Total	χ²	P value
	n	%	n	%	Total	χ²	P value
Age (years)
≤65	50	65.8	26	34.2	76
>65	55	72.4	21	27.6	76	0.493	0.483
Gender
Female	31	40.8	21	27.6	52
Male	74	74.0	26	26.0	100	2.675	0.102
ctN status
Negative	46	86.8	7	13.2	53
Positive	59	59.6	40	40.4	99	10.714	0.001
Location/growth pattern
Distal non-diffuse	64	68.1	30	31.9	94
Proximal non-diffuse/diffuse	41	70.7	17	29.3	58	0.025	0.875
Histology
Well/moderate	18	72.0	7	28.0	25
Poor/signet/mucin	87	68.5	40	31.5	127	0.012	0.913
Size (longest diameter) (cm)
<5.65	57	75.0	19	25.0	76
≥5.65	48	63.2	28	36.8	76	0.197	0.160
EMVI
Negative	59	78.7	16	21.3	75
Positive	46	59.7	31	40.3	77	5.516	0.019

Table 3.

Logistic regression model of synchronous metastases and ctN/ctEMVI status in T4 gastric cancer.

	OR	95% CI lower	95% CI upper	P value
ctN	4.148	1.684	10.218	0.002
ctEMVI	2.250	1.072	4.724	0.032

CI, confidence interval; OR, odds ratio.

Discussion

In this study, there was a significant difference in the incidence of synchronous metastases between the ctEMVI-positive group (40.3%) and the ctEMVI-negative group (21.3%). ctEMVI-positive and ctN-positive patients with ctT4 gastric cancer were significantly more likely to develop synchronous metastases relative to ctEMVI-negative and ctN-negative patients.

Previous clinical and histopathological studies have demonstrated that the presence of venous invasion is correlated with the depth of tumor invasion and tumor differentiation, and is an adverse feature for distant metastases in gastric cancer (3,11). However, those studies did not distinguish between intra- and extramural venous invasion. In contrast, EMVI was demonstrated to be an independent prognostic variable compared to intramural venous invasion in a previous histopathological study on colorectal cancer (4). Additionally, EMVI was demonstrated as an adverse imaging feature for synchronous distant metastases in rectal cancer as reported by Hunter et al. (6). In the present study, we included patients who had advanced T4 category gastric cancer detected on MDCT. Most patients exhibited poor differentiation (127/152) as confirmed by pathological analyses. In such groups, the primary gastric tumor could invade the serosa and/or penetrate into outside structures and neighboring organs, inducing a high frequency of invasion into surrounding vasculature. Although we included patients with advanced T4 primary tumors and poor differentiation, a difference in incidence of synchronous distant metastases was demonstrated between the ctEMVI-positive and ctEMVI-negative groups.

Consistent with a prior study, ctN status was demonstrated as another independent factor associated with synchronous metastases (12). However, the wide 95% CI of 1.684–10.218 obtained in the current study indicates an OR with a low level of precision, in which case, the association between ctN category and synchronous metastases may deviate from the calculated OR of 4.148. Metastatic lymph nodes are not always enlarged, although the majority of studies incorporate a short diameter of ≥ 8 mm for lymph node involvement (8). The ability to define lymph node category preoperatively in gastric cancer patients remains poor. Accuracy of identifying ctN category varies widely, in the range of 46.7–75.5% according to previous studies (8). Therefore, an additional imaging biomarker, such as ctEMVI, combined with suspected metastatic lymph nodes may be more accurate to stratify patients who have potential distant metastases before operation.

It is important to identify distant metastases in patients with gastric cancer before surgery. CT is the recommended preoperative imaging modality for detecting synchronous distant metastases (13); however, CT has a wide range of accuracy (40.8–98%) for categorizing metastases preoperatively (8). The sensitivity of CT imaging in identifying peritoneal metastases has been reported as low as only 28.3% (14). As indicated in this study, CT imaging missed peritoneal metastases in eight (8/47, 17.0%) patients, which were confirmed by pathological analyses after surgery. Therefore, a new imaging biomarker like ctEMVI is needed to predict potential metastases and identify high-risk patients who may need more aggressive preoperative staging strategies, including fludeoxyglucose-positron emission tomography (FDG-PET) CT and liver magnetic resonance imaging (MRI), or even the laparoscopic exploration. Once gastric cancer with metastases is confirmed, the therapeutic strategy may be changed to palliative resection, systemic chemotherapy, surgical bypass procedures, endoscopic interventions, and investigational therapy (15).

Despite its promising potential, further studies are needed to address some of the limitations of this current study. Based on this retrospective study, the accuracy of ctEMVI was not validated against histopathology. In addition, synchronous metastasis could be detected more sensitively using PET/CT compared to MDCT (6,8). A future prospective study is needed where synchronous metastases could be defined more critically and confirmed using histopathological examination.

In conclusion, ctEMVI status, as an adverse imaging feature, showed a statistically significant association with synchronous metastases in patients with ctT4 gastric cancer. Patients who are ctEMVI-positive may need a more aggressive imaging strategy to screen for potential synchronous metastases.

Footnotes

Acknowledgements

The authors would like to express their sincere thanks to Samantha Schoeneman (Manager of Medical Editing and Writing, iCoreMed Technology and Service LLC) and Kelly Bauer (Editor, iCoreMed Technology and Service LLC) who conducted a linguistic revision of this manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Taylor

Swift

Blomqvist

. A systematic approach to the interpretation of preoperative staging MRI for rectal cancer. Am J Roentgenol 2008; 191: 1827–1835.

Cheng

. The prognostic significance of extramural venous invasion detected by multiple-row detector computed tomography in stage III gastric cancer. Abdom Radiol 2016; 41: 1219–1226.

Maehara

Kabashima

Koga

. Vascular invasion and potential for tumor angiogenesis and metastasis in gastric carcinoma. Surgery 2000; 128: 408–416.

Betge

Pollheimer

Lindtner

. Intramural and extramural vascular invasion in colorectal cancer: prognostic significance and quality of pathology reporting. Cancer 2012; 118: 628–638.

Tan

Vikram

Boonsirikamchai

. Extramural venous invasion by gastrointestinal malignancies: CT appearances. Abdom Imaging 2011; 36: 491–502.

Hunter

Garant

Vuong

. Adverse features on rectal MRI identify a high-risk group that may benefit from more intensive preoperative staging and treatment. Ann Surg Oncol 2012; 19: 1199–1205.

Kim

Shin

Heo

. Diagnostic performance of 64-section CT using CT gastrography in preoperative T staging of gastric cancer according to 7th edition of AJCC cancer staging manual. Eur Radiol 2012; 22: 654–662.

Seevaratnam

Cardoso

McGregor

. How useful is preoperative imaging for tumor, node, metastasis (TNM) staging of gastric cancer? A meta-analysis. Gastric Cancer 2012; 15(Suppl.1): S3–S18.

Shah

Khanin

Tang

. Molecular classification of gastric cancer: a new paradigm. Clin Cancer Res 2011; 17: 2693–2701.

10.

Landis

Koch

. The measurement of observer agreement for categorical data. Biometrics 1977; 33: 159–174.

11.

Inada

Shimokawa

Ikeda

. The clinical significance of venous invasion in cancer of the stomach. Jpn J Surg 1990; 20: 545–552.

12.

Tanaka

Kumagai

Shimizu

. Peritoneal metastasis in gastric cancer with particular reference to lymphatic advancement; extranodal invasion is a significant risk factor for peritoneal metastasis. J Surg Oncol 2000; 75: 165–171.

13.

Allum

Blazeby

Griffin

. Guidelines for the management of oesophageal and gastric cancer. Gut 2011; 60: 1449–1472.

14.

Kwee RMKwee

. Modern imaging techniques for preoperative detection of distant metastases in gastric cancer. World J Gastroenterol 2015; 21: 10502–10509.

15.

Lee

Kim

Jung

. Clinical practice guidelines for gastric cancer in Korea: an evidence-based approach. J Gastric Cancer 2014; 14: 87–104.