Study population
From 2006 to 2012,
86 consecutive patients (mean age: 61 years; age range: 32-92 years ) with histologically proven endometrial carcinoma were retrospectively evaluated with DCE-MRI.
Dynamic contrast-enhanced MRI Criteria
MR imaging were performed in 9 cases with 1.5T MR unit (Symphony; Siemens,
Erlangen,
Germany or Achieva; Philips Healthcare,
Eindhoven,
The Netherlands) and in 77 cases with 3T MR unit (Signa EXCITE HD or Discovery MR750; GE Medical Systems Milwaukee,
WI,
USA).
A phased-array surface multicoil is used for signal reception.
Conventional T2-weighted images and DCE-MRI were obtained in all cases.
DCE images were obtained using three-dimensional gradient echo T1-weighted sequence (VIBE; Siemens,
THRIVE; Philips and LAVA; GE) in the sagittal plane.
These sequences were performed before and immediately after rapid IV injection of 0.2 mL/kg of contrast material at a rate of 2.5 mL/sec which was flushed with 15-20 mL saline, and then repeated at 25-30,
60,
90 and 120 seconds into the examination.
MR data interpretation was performed on the clinical viewer (EV Insite,
PSP Corporation,
Tokyo,
Japan).
Image reading was performed by two radiologists independently, with no knowledge of the histologic findings.
Semi-quantitative analysis
In each case,
a circular region of interest (ROI; at least 3 mm in diameter,
avoiding partial voluming effect and spatial misregistration) was manually drawn around the most avidly enhancing part of endometrial lesion on DCE-MRI (Fig.
1). To reduce inter-MR variation,
similarly sized ROI were also drawn around the area of piriformis muscle to provide an internal reference,
then SImax/SI (piriformis) ratios were also calculated (Fig 1 g).
The semi-quantitative enhancement parameters in the uterine tumor were quantified on DCE-imaging by using the following formula:
maximum absolute enhancement (SImax)
maximum relative enhancement (SIrel)= (SImax−SI0) /SI0×100
wash-in rate (WIR)= (SImax-S0) /SImax×100
SImax was the maximum average of tissue signal intensity among each dynamic phase.
SI0 was the average of tissue signal intensity on non-enhanced T1-weighted sequence.
As described above,
SImax/SI(piriformis) ratios were also calculated.
Statistics
Interobserver variability for the ROI measurements of the two readers were analyzed by calculating interclass correlation coefficient (ICC: 0.00-0.20 poor,
0.21-0.40 fair,
0.41-0.60 moderate,
0.61-0.80 good and 0.81-1.00 excellent correlation). The Mann-Whitney U test was used for statistical comparison and receiver operating characteristic (ROC) analysis was used to obtain optimal cut off values.
A P value of less than 0.05 was considered to be statistically significant. All statistical analyses were performed using SPSS for Windows (version 19 SPSS,
Chicago,
IL,
USA).