Subjects and groups
We retrospectively enrolled 1974 cases of hospitalized patients with digestive tract tumors (esophagus,
appendix) who had undergone routine contrast enhanced chest and abdominal CT from Jan,
2011 to Sep,
2014 in our hospital.
58 cases among them were found to have UPE according to contrast enhanced CT images.
108 cases without UPE who were 1:2 matched with UPE group respectively in age,
pathological type and tumor stage were enrolled as the control group.
Match conditions included age variation within 5 years old,
same tumor site and pathological type,
same tumor stage or higher stage.
4 cases in the UPE group could not be completely matched by the control group due to their rare site and pathological type.
Exclusion criterion included: patients complicated with non-digestive tract malignancy; tumor removed or remission more than 1 year; outpatient; patient without pathological results; inadequate image quality to make diagnosis; patient with VTE history within 1 year; patient with ongoing anticoagulant therapy.
Clinical records review
Detailed clinical records of both the UPE group and the control group were reviewed by investigation forms,
Demographic characteristics: Age,
body mass index (BMI,
body surface area (BSA,
physical status (Zubrod-ECOG-WHO).
Basic characteristics of tumor: tumor site (esophagus,
pathological type (adenocarcinoma and non-adenocarcinoma),
tumor stage (TNM),
Oncological therapies: deep vein catheterization (within 3 months),
surgery (within 2 months),
chemotherapy (within 30 days),
targeted therapy (within 30 days).
Other VTE related histories: comorbidities (Charlson Comorbidity Index,
bed rest or immobilization(entirely best rest ≥75% within 2 weeks),
varicose vein of lower extremity,
history of VTE (＞1year)
Post processing and analysis of CT images
All subjects had undergone routine chest and abdominal CT scan by 64-row spiral CT with a supine position.
Two cardiovascular radiologists who were blind to the clinical records independently reviewed the CT images of all the cases,
and a decision was made by consensus.
AVA was measured on post processing station (Vitrea,
According to the methods mentioned in previous studies [10-13],
AVA was measured on contrast enhanced thin section axial images at the navel level (length of coverage: 10mm) (Fig.
CT value of adipose tissue was defined as -150 HU to -50HU,
SAV and VAV can be obtained automatically by post-processing software (Fig.
Then VAV/BMI (body matrix index),
VAV/BSA (body surface area) and VAV/SAV were computed.
Student’s t test was applied for comparison in normally distributed continuous variables,
while rank sum test was applied for non-normal distributed variables. Χ2 test was used for comparison in qualitative data.
Univariate logistic regression analysis were done to evaluate the correlation between UPE and other factors.
Conditional multivariate logistic regression analysis was applied to evaluate the correlation between TAVA,
VAV/BSA and the occurrence of UPE,
adjusting other correlated variables.
All statistical analysis were conducted by SPSS 20.0 software.
The significace level was 0.05.