RESUMEN
Objective To evaluate the value of 3T magnetic resonance imaging (MRI) in the preoperative T staging of potentially resectable esophageal cancer(EC), compared with endoscopic ultrasonography (EUS). Methods Patients with resectable EC pathologically confirmed by biopsy from March 2015 to September 2016 were prospectively enrolled. All patients underwent MRI (including T2-TSE-BLADE,DWI and radial-VIBE)and EUS one week after the biopsy,and MRI were performed prior to EUS, both MRI and EUS were acquired within one week before surgery. Two readers with more than 5 years experiences in the MRI diagnosis evaluated the MR image quality using a 5-point score independently. T staging was assigned on MRI and EUS by the two MRI readers using double-blind method and one endoscopist in accordance with the 7th edition of AJCC TNM Classification for EC, and any disagreement between two MRI readers was resolved by consensus with discussion to the third senior MRI doctor. The inter-observer agreement between two MRI readers were calculated using Kappa test for image quality scores and T staging results. Considering postoperative pathological T staging results as the gold standard, the performances of MRI and EUS were evaluated based on the accuracy rate and analyzed by χ2 test. Results A total of 70 patients were enrolled in the study, the good image quality cases (≥ 3 scores) were 66 in reader 1 and 68 in reader 2. The inter-observer agreement of the image quality scores by two readers was excellent (Kappa=0.824, P<0.05). The pathological results revealed 16 cases of T1 stage, 18 cases of T2 stage, 30 cases of T3 stage, and 6 cases of T4a stage. The inter-observer agreement of the preoperative T staging of EC by two readers was excellent (Kappa=0.809, P<0.05). The accuracy rates of MRI and EUS for preoperative T staging of EC were 92.9% (65/70) and 67.1% (47/70), respectively, and the difference in accuracy rates of two techniques was statistically significant (χ2=14.5, P<0.05). Conclusions The accuracy rate of MRI for preoperative T staging of EC is significantly higher than that of EUS. MRI can be used as a noninvasive method for preoperative T staging of EC.
RESUMEN
Objective To analyze the value of conventional T2WI combining with breath-hold Cartesian VIBE sequence, and compared with Blade combining with breath-free radial VIBE sequences in pre-operative T staging of potentiallly resectable esophageal carcinoma. Methods Sixty-five cases of esophageal carcinoma were concluded prospectively. All patients had received pathological examination of gastroscope biopsy before MRI. Patients received MR examination, including T2WI, breath-hold VIBE, Blade, and breath-free radial VIBE sequences. Two radiologists with more than five years experiences in the diagnosis of chest, performed T staging in MRIby referring to the the 7th edition of UICC-AJCC TNM classification. The results of MRI T staging and the postoperative pathological T staging were analyzed byχ2 test. Results Sixty-five patients were included. Diagnostic coincidence rates of the preoperative T staging by using conventional T2WI combining with breath-hold Cartesian VIBE sequences and Blade combining with breath-free radial VIBE sequences were 51%(33/65) and 88%(57/65) ,with 32 and 8 cases overstaging or understaging respectively, and the statistical differences were significant(χ2=20.80, P<0.05). The former diagnostic accuracy of esophageal carcinoma in situ, muscularis violation and esophageal fiber membrane were 42%(8/19), 54%(14/26), 55%(11/20), and the latter were 89%(17/19), 88%(23/26), and 85%(17/20). Conclusions Diagnostic coincidence rate of the preoperative T staging by using Blade combining with breath-free radial VIBE sequences is much higher than conventional T2WI combining with breath-hold Cartesian VIBE sequences. Blade combining with breath-free radial VIBE sequences could be used as non-invasive imaging method in preoperative T staging of potentially resectable esophageal carcinoma.
RESUMEN
Objective To assess the diagnostic value of MR imaging in follow-up evaluation of patients with hepatocellular carcinomas ( HCC ) treated with radiofrequency ablation ( RFA ) and to compare it with that of multi-slice CT.Methods From December 2009 to September 2011,there were 48 patients (56 HCCs) treated with RFA after transcatheter arterial chemoembolization (TACE). MR imaging and multi-slice CT were performed for follow-up.Two radiologists independently reviewed these images,detection of residual or recurrent tumor were assessed on a five-point scale and compared with Kappa test and with the method of receiver operating characteristic (ROC) curve analysis.Sensitivity,specificity and accuracy were evaluated.Results The observer agreement rate for MR imaging was higher ( 0.925 ) than for multi-slice CT (0.701,P ＜ 0.05).The area under the ROC curve (AUC)of MR imaging( 0.987 and 0.971 by two radiologists respectively) was significantly higher than that of CT( 0.674 and 0.598 by two radiologists respectively),P ＜0.05. The sensitivity, specificity and accuracy of detection rate for MRI [100％(22/22),95.5％ (86/90) and 95.5％ ( 107/112),respectively] were significantly different with that for multi-slice CT [40.9％ ( 9/22 ),57.8％ ( 52/90 ) and 60.7％ ( 68/112 ),respectively]. Conclusion Diagnostic accuracy and detection rate of residual or recurrent tumor were found to be superior with MR imaging than with multi-slice CT.