Religion & Spirituality

A comparison of image quality between tissue harmonic imaging and fundamental imaging with an electronic radial scanning echoendoscope in the diagnosis of pancreatic diseases

Description
Background: The availability of an electronic radial scanning echoendoscope has facilitated the clinical use of tissue harmonic imaging. This study compares the quality of US images acquired by tissue harmonic imaging during electronic radial
Published
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  A comparison of image quality betweentissue harmonic imaging and fundamen-tal imaging with an electronic radial scan-ning echoendoscope in the diagnosis ofpancreatic diseases Hideki Ishikawa, MD, Yoshiki Hirooka, MD, Akihiro Itoh, MD,Senju Hashimoto, MD, Naoto Okada, MD, Terutomo Itoh,MD, Hiroki Kawashima, MD, Hidemi Goto, MD Background:  The availability of an electronic radialscanning echoendoscope has facilitated the clinical useof tissue harmonic imaging.This study compares thequality of US images acquired by tissue harmonic imag-ing during electronic radial scanning EUS to thoseacquired by fundamental imaging. Methods:  Electronic radial scanning EUS was performedin 108 patients with pancreatic lesions (58 cystic,50solid).US images acquired by fundamental imaging at afrequency of 7.5 MHz were compared with thoseacquired by tissue harmonic imaging by using transmit-ting and receiving frequencies of,respectively,4.0 and8.0 MHz at the same scanning plane.Cystic lesions wereevaluated for boundary/septum and nodules,and solidlesions,for boundary and internal structure.US imagesacquired by tissue harmonic imaging and fundamentalimaging during electronic radial scanning EUS werecompared,by using a Likert scale for the respectiveevaluation criteria,by two independent endoscopists. Observations:  For cystic lesions,tissue harmonicimages were significantly clearer than fundamentalimages for visualizing boundary and septum ( p  < 0.0001,both reviewers) and nodules ( p  = 0.0003,Reviewer 1; p  =0.0007,Reviewer 2).For solid lesions,tissue harmonicimages were significantly clearer than fundamentalimages for visualizing boundary ( p  = 0.0003,Reviewer 1; p  < 0.0001,Reviewer 2) and internal structures ( p  =0.0003,Reviewer 1;  p  = 0.0009,Reviewer 2). Conclusions:  US images acquired by tissue harmonicimaging appear to be clearer compared with thoseacquired by fundamental imaging. Tissue harmonic imaging (THI),a novel USimage acquisition method,is a derivative techniqueof the harmonic imaging method that was developedto increase the efficiency of visualization by using US contrast agents.Tissue harmonic imaging facili-tates acquisition of US images better than funda-mental imaging (FI),without using US contrastagents.Tissue harmonic imaging exploits the effectof nonlinear propagation on the acoustic signal as ittravels through the human body or tissue.It pro-duces US images by using second harmonic signalsgenerated by the tissue itself during this nonlinearpropagation of insonated acoustic energy.The resul-tant advantages are improved lateral resolution,reduced side lobe artifact and an increased signal-to-noise ratio. 1 Tissue harmonic imaging has been used clinical-ly during transabdominal US;there are many stud-ies of the usefulness of THI in cardiovascular medi-cine. 2-6 Moreover,many studies have found THI tobe useful for the diagnosis of liver disorders. 7-10 Useof THI during transabdominal US reportedly hasincreased the ability to visualize pancreaticlesions. 11 However,the diagnostic utility of transab-dominal US for pancreatic lesions is limited by theanatomical location of the pancreas.To compensatefor this limitation,EUS is used for closer examina-tion of pancreatic lesions. 12-21  A novel echoendo-scope for electronic radial scanning EUS (ER-EUS)was developed in 2000.The US images acquired byTHI during ER-EUS in the investigators’ hospitalappeared to be an improvement over those providedby FI.A study was,therefore,conducted in which USimages acquired by FI were compared qualitativelywith images obtained by THI during ER-EUS. PATIENTS AND METHODS  A total of 108 patients with pancreatic lesions (70 men,38 women;mean age,59 years) were enrolled between August 1,2001 and June 30,2002,and underwent ER-EUS after obtaining written informed consent.Of thesepatients,58 had cystic lesions (35 intraductal papillarymucinous tumor [IPMT],15 simple cyst,3 mucinous cys-tadenoma [MCA],3 serous cystadenoma,two miscella-neous) and 50 had solid lesions (12 ductal carcinoma,9chronic pancreatitis,8 pancreatolithiasis,6 neuroen-docrine tumor,3 solid-pseudopapillary tumor,onemetastatic tumor,one acinar cell carcinoma,one autoim-mune pancreatitis,9 miscellaneous).By surgical resec-tion,IPMT was confirmed in 20 and MCA in 3 cases.A fur-ther 3 cases of IPMT were confirmed by autopsy.Of the 12ductal carcinomas,4 were confirmed by surgical resection,4 by autopsy,and 4 by EUS-guided FNA.Three cases of solid-pseudopapillary tumor,the one metastatic tumor,and the single acinar cell carcinoma were proven by sur-gical resection.All other lesions were evaluated by ER-EUS alone every 6 months,and no growth of any tumor-ous lesion was observed during the study period. N EW  M ETHODS  & M ATERIALS VOLUME 57,NO.7,2003GASTROINTESTINAL ENDOSCOPY   931  Received October 11,2002.For revision December 12,2002. Accepted February 7,2003.Current affiliations:Second Department of Internal Medicine, Nagoya University School of Medicine,Nagoya,Japan,Departmentof Endoscopy,Nagoya University Hospital,Nagoya,Japan. Reprint requests:Yoshiki Hirooka,Second Department of Internal Medicine,Nagoya University School of Medicine,65 Tsuruma-cho, Showa-ku,Nagoya 466-8550 Japan.Copyright © 2003 by the American Society for Gastrointestinal Endoscopy0016-5107/2003/$30.00 + 0doi:10.1067/mge.2003.271  Electronic radial scanning EUS was performed with anewly developed forward-viewing,electronic radial scan-ning echoendoscope (EG-3630UR;Pentax Co.,Ltd.,Tokyo,Japan) and monitor/processing unit (EUB-6000;HitachiCo.,Ltd.,Tokyo,Japan) especially equipped with softwarefor THI.This echoendoscope has an insertion tube diameterof 12.1 mm,a visual field angle of view of 270°,a frequencybandwidth of 5 to 10 MHz,an array radius of curvature of 5.5 mmR,and 192 channels (number of elements).Tissueharmonic imaging was performed by using a transmitting frequency of 4.0 MHz and a receiving frequency of 8.0 MHz.During ER-EUS,FI also could be performed by using 3 dif-ferent frequencies (5.0,7.5,10 MHz).However,to compareFI with THI,the former was performed by using a frequen-cy of 7.5 MHz,which was within the transmitting andreceiving frequencies for THI.Both color Doppler andpower Doppler imaging are available with this instrumentsystem,so that hemodynamic information also could beobtained for pancreatic lesions.In this study,the usefulnessof B-mode US images was evaluated.With this instrument system,it is possible to changebetween THI and FI in real time by using a switch on theEUB-6000 processor.In all patients,image acquisitionwas performed by the same physician examiner by using the same scanning plane as far as possible under uniformconditions.However,the gain was adjusted in individualpatients to optimize observation.Adjustments in gainwere necessary because of the low signal intensity withTHI imaging.Acquired US images were transferred fromthe EUB-6000 processor to a personal computer via amagneto-optical disk.Subsequently,the ability to visual-ize the boundary and septum (n = 58) and nodules (n = 17)was evaluated in patients with cystic lesions.In patientswith solid lesions,the ability to visualize the boundary (n= 50) and internal structures (n = 50) was assessed.Inaddition,the ability to visualize luminal structures nearthe lesions (e.g.,bile duct,main pancreatic duct,vessels)was also appraised in some patients.The respective US images acquired by THI and FI wereevaluated independently by two physicians with extensiveexperience in EUS (respectively,1500 [Reviewer 1] and4500 [Reviewer 2] procedures) by using 5-point Likertscales.To evaluate the image quality,pairs of still imagesrandomly arranged (THI vs.FI) were viewed on a com-puter monitor screen.Image quality was graded on theLikert scale as follows:grade 1,“recognized very well”grade 5,“recognized very poorly”for each parameter. Although the two physicians grading the images did notknow the method of acquisition,the evaluation was notstrictly blinded because the physicians could speculate asto the method of acquisition based on the characteristicsof the respective images,that is,the pairs of images wererecognizably different.The two physicians did not simul-taneously examine the same US image before scoring.The Wilcoxon signed rank test was used to evaluatescores for the various evaluation criteria.Interobserver variability was assessed statistically by using the kappastatistic.A  p  value < 0.05 was considered significant.  H Ishikawa,Y Hirooka,A Itoh,et al.Tissue harmonic imaging vs fundamental imaging with electronic radial EUS 932 GASTROINTESTINAL ENDOSCOPYVOLUME 57,NO.7,2003 Table 1.Grading of cystic lesions Boundary and septumNodulen = 58n = 17Likert scaleFITHIFITHIReviewer 11280425221531117574281181512030FI vs.THI  p < 0.0001  p = 0.0003Reviewer 2101504242006318157542778159121FI vs.THI  p < 0.0001  p = 0.0007FI vs.THI by the Wilcoxon signed rank test.Likert scale:1,recognized very well;5,recognized very poorly. Table 2A.Coincidence of Likert scales in bound-ary and septumTable 2B.Coincidence of Likert scales in nodule  RESULTSCystic lesions The grading of cystic lesions (boundary/septum,nodule) by Reviewers 1 and 2 is shown in Table 1.The THI was significantly better than FI for visual-ization of the boundary and septum (  p < 0.0001,both reviewers;Wilcoxon signed rank test) and nod-ules (  p = 0.0003,Reviewer 1;  p = 0.0007,Reviewer 2;Wilcoxon signed rank test).With regard to interob-server variability for boundary and septum,thekappa is 0.365:95% CI [0.177,0.552],  p < 0.0001 forFI,and the kappa is 0.273:95% CI [0.098,0.447],  p = 0.0005 for THI,demonstrating the correlationbetween the grading by Reviewers 1 and 2 (Table2A).With respect to the interobserver variability fornodules,the kappa is 0.538:95% CI [0.198,0.878],  p = 0.003 for FI,and the kappa is 0.425:95% CI[0.098,0.752],  p = 0.005 for THI,demonstrating thecorrelation between the grading by Reviewers 1 and2 (Table 2B). Solid lesions The results of the grading of solid lesions byReviewers 1 and 2 are shown in Table 3.The THIwas significantly better than FI for visualization of the boundary (  p = 0.0003,Reviewer 1;  p < 0.0001,Reviewer 2,Wilcoxon signed rank test) and internalstructures (  p = 0.0003,Reviewer 1;  p = 0.0009,Reviewer 2;Wilcoxon signed rank test).With regardto interobserver variability for boundary,the kappais 0.380:95% CI [0.152,0.610],  p = 0.002 for FI,andthe kappa is 0.368:95% CI [0.162,0.574],  p = 0.0003for THI,demonstrating the correlation between theresults of Reviewers 1 and 2 (Table 4A).For interob-server variability with regard to internal structures,the kappa is 0.365:95% CI [0.140,0.590],  p = 0.002for FI,and the kappa is 0.381:95% CI [0.180,0.583],  p < 0.0001 for THI,demonstrating the correlationbetween the results of Reviewers 1 and 2 (Table 4B).Some data were used in multiple statistical tests,but a correction for multiple testing (e.g.,Bonferroni’smethod),would not have removed significance in anyinstance. DISCUSSION  Although nonlinear propagation of US within theliving body is well recognized,its usefulness in clin-ical practice has not been evaluated sufficiently.Tissue harmonic imaging is an important derivativetechnique of the harmonic imaging method,whichwas developed to increase the efficiency of visual-ization by using US contrast agents. Tissue harmonic imaging vs fundamental imaging with electronic radial EUSH Ishikawa,Y Hirooka,A Itoh,et al.VOLUME 57,NO.7,2003GASTROINTESTINAL ENDOSCOPY   933   Table 3.Grading of solid lesions BoundaryInternal structuren = 50n = 50Likert scaleFITHIFITHIReviewer 1107172820922327192617413413452010FI vs.THI  p < 0.0001  p = 0.0009Reviewer 210605282161733219312249411551021FI vs.THI  p < 0.0003  p = 0.0003FI vs.THI by the Wilcoxon signed rank test.Likert scale:1,recognized very well;5,recognized very poorly. Table 4A.Coincidence of Likert scales—boundaryTable 4B.Coincidence of Likert scales—internalstructures  When US is transmitted within the body,a fre-quency of receiving US is present within a frequen-cy zone that forms a gentle curve around f0,the fre-quency of transmitted US.This phenomenon isreferred to as the linearity of the receiving frequen-cy.However,harmonics are present in several fre-quency zones that peak at integrally multipliedzones of the receiving frequency (nonlinearity).During THI,US images are constructed solely byreceiving second harmonics,the largest peaks. 5,11,12  Various studies have shown THI to be useful dur-ing transabdominal US. 1,7-14,22-24 To the investiga-tors’ knowledge,however,the value of THI during EUS has not been evaluated.Therefore,this studywas conducted to evaluate the usefulness of THIduring EUS of pancreatic lesions.When various cri-teria were assessed,the US images of cystic lesionsacquired by THI were significantly better than thoseacquired by FI.This finding was consistent with theresults of other studies that demonstrated that par-ticularly clearer US images of luminal structuresand cystic lesions were acquired by THI because of reductions in side lobe and reverberation arti-facts. 11,12,24 Theoretically,because the harmoniccomponents of transmitted US are focused on thecentral axis of the main beam,the beam width of thetransmitted US,consequently,becomes narrowerthan that of basic waves.In addition,because sec-ond harmonics are proportional to a square of sound  H Ishikawa,Y Hirooka,A Itoh,et al.Tissue harmonic imaging vs fundamental imaging with electronic radial EUS 934 GASTROINTESTINAL ENDOSCOPYVOLUME 57,NO.7,2003 Figure 1. Serous cystadenoma: A, ER-EUS (FI) imageshowing multilocular mass about 20 mm in diameter in pan-creatic body. B, ER-EUS (THI) showing inner structure ofmass more clearly;coexistence of microcystic (componentseen as if solid in FI-mode) and macrocystic area character-istic of serous cystadenoma are revealed. AB Figure 2. Intraductal papillary mucinous adenoma: A, ER-EUS (FI) image showing hyperechoic mass in multilocularlesion about 40 mm in diameter in pancreatic head. B, ER-EUS (THI) image demonstrating structure of septa anddelineation of mural papillary tumor (arrows)  more clearly. AB  pressure,harmonics are rarely produced near thebody surface.Therefore,US images acquired by THIappear clearer because of the reduction of multiplereverberations.The THI may be more advantageous for cysticlesions.For example,in patients with serous cys-tadenoma,US images acquired by THI of the cysticlumen were clearer than those acquired by FI.Thus,clearer US images of microcystic and macrocysticareas were demonstrated by THI (Fig.1).In somepatients with serous cystadenoma,microcysticareas were not sufficiently visualized by FI or were visualized as a solid mass because of the influence of artifact.These features were clearly delineated byTHI,which facilitated a more accurate diagnosis of serous cystadenoma.In patients with branch duct type intraductalpapillary mucinous adenoma,FI visualized nodulesin multiple cysts localized in the head of the pan-creas.However,THI facilitated the acquisition of clearer US images of the septum and nodular mar-gin,in addition to clearer details of the nodulesthemselves.Moreover,clearer US images of the cys-tic lumen were also acquired by THI (Fig.2).Septalstructures within the cystic lumen,and the presenceor absence of mural nodules,are features that canbe used to determine whether an intraductal papil-lary mucinous tumor is likely to be benign or malig-nant. 20 By clearly demonstrating these features,THI may facilitate the diagnosis of intraductal pap-illary mucinous tumors.When the respective criteria were evaluated forsolid lesions,US images acquired by THI were alsosignificantly better than those acquired by FI.Thisfinding was consistent with the results of previousstudies in which THI enhanced the contrast of themargin and internal structure of solid tumors in theliver. 7-10 In theory,larger signal-to-noise ratios areobtained in the lateral direction of the transmittedsound field during THI.Therefore,clearer USimages may be acquired by THI because of theimprovement in tissue contrast.For example,in apatient with ductal carcinoma,US images acquiredby FI of the margin of a tumor in the head of thepancreas were slightly unclear.However,THI clear-ly demonstrated the relationships for tumor margin,bile duct,main pancreatic duct,and superior mesen-teric vein,as well as the lumens of these structures(Fig.3).In this patient,THI facilitated more accu-rate staging of the tumor by demonstrating infiltra-tion into the duodenum.In the present study,when THI was used,theimprovement in US images of cystic lesions wasslightly better than that noted for solid lesions.Also,US images acquired by FI were better than thoseacquired by THI in some patients,probably becauseof artifact caused by air in the balloon at the tip of the echoendoscope or artifact caused by deteriora-tion in a US element in the ER-EUS probe and nottechnical problems with THI itself.There are some disadvantages with THI.WhenTHI is based on the use of a band filter,as in the sys-tem used in the present study,all harmonics pro-duced during passage of US through tissue are cut off below a certain frequency.Therefore,THI is not asuseful for visualization of superficial structures.Harmonics are attenuated in the deeper regions of tissue,with the result that axial resolution may dete- Tissue harmonic imaging vs fundamental imaging with electronic radial EUSH Ishikawa,Y Hirooka,A Itoh,et al.VOLUME 57,NO.7,2003GASTROINTESTINAL ENDOSCOPY   935 Figure 3. Ductal carcinoma: A, ER-EUS (FI) image showinghypoechoic tumor about 25 mm in diameter in pancreatichead. B, ER-EUS (THI) image showing entire tumor, dilatedbile duct, and main pancreatic duct more clearly.Imagereveals tumor extension to surrounding tissue, includinginvasion of second portion of duodenum (arrows). AB
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x