[雙語翻譯]--醫(yī)學(xué)外文翻譯--高b值彌散加權(quán)mri在結(jié)直腸癌中的應(yīng)用（英文）

1、AJR:187, July 2006 181AJR 2006; 187:181–1840361–803X/06/1871–181© American Roentgen Ray SocietyMEDICAL I MAGINGACENT U RYOFMEDICAL I MAGINGACENT U RYOFIchikawa et al. High-B-Value Diffusion- Weighted MRI in Color

2、ectal CancerGastrointestina l Imag ing ? Technical InnovationHigh-B-Value Diffusion-Weighted MRI in Colorectal CancerTomoaki Ichikawa1Sukru Mehmet Erturk2,3Utarou Motosugi1Hironobu Sou1Hiroshi Iino4Tsutomu Araki1Hideki

3、Fujii4Ichikawa T, Erturk SM, Motosugi U, et al.Keywords: cancer, colon, diffusion-weighted MRI, MRIDOI:10.2214/AJR.05.1005Received June 11, 2005; accepted after revision September 13, 2005.1Department of Radiology, Yama

4、nashi University, Shimokato, Japan.2Brigham and Women’s Hospital, Harvard Medical School, Radiology Suite, c/o One Brigham Circle, 1620 Tremont St., Boston, MA 02120. Address correspondence to S. M. Erturk (mehmetert

5、urk@superonline.com).3Department of Radiology, Sisli Etfal Training and Research Hospital, Istanbul, Turkey.4First Department of Surgery, University of Yamanashi, Shimokato, Japan.OBJECTIVE. The purpose of this article

6、 is to evaluate the usefulness of high-b-value dif- fusion-weighted MRI (DW-MRI) in the detection of colorectal adenocarcinoma. CONCLUSION. High-b-value DW-MRI allows detection of colorectal adenocarcinoma with a high se

7、nsitivity and specificity.iffusion-weighted MRI (DW- MRI) is becoming increasingly important in the assessment of ma- lignant tumors [1, 2]. It is generally accepted that DW-MRI enables noninvasive characterization of

8、biologic tissues on the basis of their water diffusion properties [3]; it pro- vides information about the biophysical prop- erties of tissues such as cell organization and density, microstructure, and microcirculation [

9、4]. DW-MRI is widely used in neuroimaging [5], but its application within the abdomen is hindered by the presence of bulk physiologic motion such as respiration, peristalsis, and blood flow, in which orders of magnitude

10、are greater in amplitude than that of diffusion [6]. Takahara et al. [7] proposed a DW-MRI technique that might provide images with im- proved signal-to-noise ratios (SNRs); reversal of the contrast of these images resul

11、ted in black-and-white images with contrast charac- teristics closely resembling those of PET. We hypothesized that high-b-value DW-MRI im- ages could be directly used for tumor detec- tion because of the different cellu

12、lar struc- tures of healthy and neoplastic tissues. We decided to study colorectal adenocarcinoma because of the general challenges of colonic MRI, including the nonsolid nature of the or- gan, peristalsis, and movement

13、of the intralu- minal contents. Thus, our aim in this prelimi- nary study was to evaluate the usefulness of high-b-value DW-MRI in the detection of co- lorectal adenocarcinoma.Materials and Methods PatientsDuring a perio

14、d of 6 months between August2004 and February 2005, 33 consecutive patients(mean age, 59; range, 33–69 years; 15 women, 18men) with 33 endoscopic colonoscopically provencolorectal cancers ranging from 20 to 70 mm(mean, 3

15、3 mm) were found in our institution andtwo related hospitals and were included in thisstudy. The lesions were located in the rectum (n =14), sigmoid colon (n = 8), transverse colon (n = 2),ascending colon (n = 8), and ce

16、cum (n = 1). An-other 15 patients who underwent endoscopiccolonoscopy during the same period with negativeresults were included as controls. All patients withcolorectal cancer finally underwent surgical resec-tion and th

17、e diagnoses were confirmed. Contrast-enhanced CT was performed in all patients and neg-ative cases in the control group before MR exami-nations. Informed consent was obtained from allpatients before participation in the

18、study, whichwas approved by our institutional review board.MRI Protocol and Parameters MRI used a combination of a commercially avail-able 1.5-T superconducting MR unit and a body coil(Signa EchoSpeed, GE Healthcare). Fi

19、rst, breath-hold, coronal T1-weighted MR images with gradi-ent-echo sequences were obtained to confirm the op-timal scan range. Then, axial T1-weighted and respi-ratory-triggered, fast spin-echo T2-weighted MRI(TR/TE, 2,

20、000–4,000/80), and high-b-value DW-MRI were performed in all patients including thecontrol group. The patients did not undergo anypreparation such as bowel cleansing before the ex-aminations. High-b-value DW-MR images we

21、re ob-tained without breath-holding during the acquisition.Detailed parameters for high-b-value DW-MRI weresequence: single-shot spin-echo echo-planar (SE-EPI); fat-suppression technique, chemical shift se-lective techni

22、que; scan direction, axial, b value, zeroand 1,000 s/mm2; TR/TE/inversion time (TI),8,000–10,000/73.2–73.4/70; matrix, 128 × 64; sliceDDownloaded from www.ajronline.org by 59.47.43.89 on 05/15/13 from IP address 59.

23、47.43.89. Copyright ARRS. For personal use only; all rights reserved High-B-Value Diffusion-Weighted MRI in Colorectal CancerAJR:187, July 2006 183without referring to any other MR images. MIP im-ages were evaluated with

24、 the rotational cine mode to-gether with the axial source images in different win-dows on diagnostic monitors. Each reviewer gradedthe presence (or absence) of lesions on a 5-pointconfidence scale based on the strength a

25、nd the ap-pearance of dark signals on high-b-value DW-MRimages as follows: 1 = definitely absent (no signal);2 = probably absent (nonlocalized, mild to moderatesignal); 3 = undetermined (localized, mild to moder-ate sign

26、al); 4 = probably present (localized, strongsignal with no definite margins); 5 = definitelypresent (localized, strong signal with definite mar-gins). If a lesion was considered to be present on ahigh-b-value DW-MR image

27、, the lesion location wasrecorded. Only lesions recorded at the correct loca-tion determined by the study coordinators were ac-cepted as true-positive.Statistical AnalysisReceiver operating characteristic (ROC) curveswer

28、e used to represent the performance of individ-ual radiologists for tumor detection. The diagnosticaccuracy for each radiologist was determined bycalculating the area under the ROC curve (Az).Grades 4 and 5 were accepted

29、 as positive for thepresence of colorectal adenocarcinoma, and thesensitivity and specificity were calculated with95% confidence intervals (CIs).The interobserver agreement among reviewersfor tumor detection was calculat

30、ed with the linear-weighted kappa statistics. A kappa statistic greaterFig. 2—44-year-old man with rectal adenocarcinoma (arrow); maximum- intensity-projection (MIP)-reconstructed sagittal high-b-value diffusion-we

31、ighted MR image.Fig. 3—54-year-old man with rectal adenocarcinoma (arrows); maximum- intensity-projection (MIP)-reconstructed coronal high-b-value diffusion-weighted MR image; two metastatic lymph nodes are also

32、 visualized (arrowheads).than 0.75 was considered excellent agreement be-yond chance; 0.4–0.75, fair to good agreement; andless than 0.4, poor agreement.Results ROC analysis yielded Az values of 0.96, 0.96, and 0.97 for

33、 the three radiologists, respec- tively. The sensitivities for the three radiolo- gists were 90.9% (30/33; 95% CI, 74.7–100%), 87.9% (29/33; 95% CI, 70.9–96%), and 90.9% (30/33; 95% CI, 74.7–100%), respectively. For each

34、 radiologist, the specificity was 100% (15/15; 95% CI, 74.7–100%). The mean sensi- tivity and specificity of high-b-value DW-MRI for detection of colorectal adenocarcinoma were 90.9% (30/33; 95% CI, 74.5%–97.6%) and 100%

35、 (15/15; 95% CI, 74.6–100%), re- spectively. Figures 1, 2, and 3 show representa- tive patients with colorectal adenocarcinomas. All kappa values indicating interobserver agreement were in the category of excellent (rang

36、e, 0.83–0.89).Discussion To investigate the reason why only colonic adenocarcinomas and not the healthy colon showed strong signal intensity on the high-b- value DW-MR images is a challenging issue that warrants further

37、studies. However, a the- oretical explanation might be constructed by means of a well-described general hypothesis of DW-MRI. It is well known that diffusion is caused by random translational molecular motion, also known

38、 as Brownian motion. DW-MRI is the only imaging method that can evaluate the diffusion process in vivo. The speed of diffusion of water molecules is dif- ferent in the extracellular and intracellular components of the ti

39、ssues [3]. In the intracel- lular component, the diffusion is relatively slow because of the presence of cellular mem- branes. Thus, apparent diffusion coefficients (ADCs), which are quantitative expressions of diffusion

40、 characteristics of tissues, are re- lated to the proportion of extracellular and in- tracellular components. They tend to decrease with increased tissue cellularity or cell den- sity [8]. Conversely, the cell density ma

41、y be indicative of tumor aggressiveness; Lyng et al. [9] reported an increased metastatic capac- ity of tumors with high cellularity. Moreover, in addition to the cellular membranes, the in- tracellular cytoskeleton, org

42、anelles, matrix fi- bers, and soluble macromolecules contribute to diffusion restrictions in tumors. Thus, dif- fusion curves that decay quickly or large ADC values may be typical for healthy tis- sues or benign patholog