Essays & Theses

Myometrial invasion in endometrial cancer: diagnostic performance of diffusion-weighted MR imaging at 1.5-T

Myometrial invasion in endometrial cancer: diagnostic performance of diffusion-weighted MR imaging at 1.5-T
of 9
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.
  Eur RadiolDOI 10.1007/s00330-009-1597-x  UROGENITAL Gilda RechichiStefania GalimbertiMauro SignorelliPatrizia PeregoMaria Grazia ValsecchiSandro Sironi Received: 29 March 2009Revised: 27 May 2009Accepted: 6 July 2009 # European Society of Radiology 2009 Myometrial invasion in endometrial cancer:diagnostic performance of diffusion-weightedMR imaging at 1.5-T Abstract  Objective:  To determinethe diagnostic accuracy of diffusion-weighted (DW) magnetic resonance(MR) imaging in the preoperativeassessment of myometrial invasion byendometrial cancer.  Materials and methods:  In this prospective study,47 patients with histologically con-firmed endometrial cancer underwent  preoperative MR imaging and totalhysterectomy. The MR protocol in-cluded spin-echo multishot T2-weighted, dynamic T1-weighted andDW images acquired with b-values of 0 and 500 s/mm 2 . Myometrial tumour spread was classified as superficial(<50%) or deep ( ≥ 50% myometrialthickness). Postoperative histopatho-logical findings served as a referencestandard. Indices of diagnostic per-formance were assessed for eachsequence.  Results:  At histopatho-logical examination, superficial myo-metrial invasion was found in 34 patients and deep myometrial inva-sion in 13. In the assessment of tumour invasion, sensitivity, specific-ity, positive and negative predictivevalues of T2-weighted images were92.3%, 76.5%, 60.0% and 96.3%,respectively. The corresponding va-lues for dynamic images were 69.2%,61.8%, 40.9% and 84.0%, and for DW images 84.6%, 70.6%, 52.4%and 92.3%. T2-weighted and DWimaging proved to be the most accurate techniques for tumour spreaddetermination.  Conclusion:  DW im-aging proved to be accurate in asses-sing myometrial invasion, and it could replace dynamic imaging as anadjunct to routine T2-weighted imag-ing for preoperative evaluation of endometrial cancer. Keywords  Endometrial carcinoma .Staging .Magnetic resonanceimaging .Diffusion-weightedimaging Introduction Endometrial carcinoma is the third most common gynae-cological cancer in women [1]. Most patients present withabnormal genital bleeding, and because of its low cost,relatively non-invasive nature and high sensitivity in thedetection of uterine abnormality, transvaginal sonographyhas been proposed as the first-line diagnostic tool [2, 3]. For a definitive diagnosis, in patients with postmenopausal bleeding and endometrial thickening on transvaginalsonography, biopsy or dilation and curettage proceduresare currently performed. The prognosis of endometrialcancer mainly depends on three factors: histologicalsubtype and grade, tumour stage at diagnosis, including G. Rechichi .S. Galimberti .M. Signorelli .P. Perego .M. G. Valsecchi .S. SironiSchool of Medicine,University of Milano-Bicocca,Milan, ItalyG. Rechichi .S. Sironi ( * )Department of Diagnostic Radiology,H S. Gerardo Monza,Via Pergolesi 33,20052 Milan, Italye-mail: Tel.: +39-039-2339536Fax: +39-039-2333463 S. Galimberti .M. G. ValsecchiDepartment of Clinical Medicine andPrevention  —  Centre of Biostatistics andClinical Epidemiology,University of Milano-Bicocca,Milan, ItalyM. SignorelliDepartment of Gynaecology andObstetrics, H S. Gerardo Monza,Milan, ItalyP. PeregoDepartment of Pathology,H S. Gerardo Monza,Milan, Italy  the depth of myometrial invasion, and the presence of lymph node metastases [4]. In particular, histologicaltumour grade and depth of myometrial invasion stronglycorrelate with the presence of lymph node metastases andoverall patient survival [5]. Patients with 50% or greater myometrial invasion have a six- to seven-fold higher  prevalence of pelvic and para-aortic lymph node metastasiscompared with patients in whom myometrial invasion isabsent or less than 50% [6]. Therefore, preoperativeknowledge of these factors represents a critical step intreatment planning and tailoring the surgical approach [7].The grade of the tumour can be diagnosed with endometrialsampling; conversely, information on myometrial tumour invasion can only be obtained preoperatively by imagingtechniques. Among them, magnetic resonance (MR)imaging has proved to be the most accurate [8, 9]. According to previous reports [10  –  14], MR T2-weightedand dynamic contrast-enhanced T1-weighted images areconsidered most useful for assessing the extent of tumour spread, although their evaluation of the depth of myome-trial invasion may be affected by several pitfalls [12  –  14].Diffusion-weighted MR imaging (DWI) is a recentlyavailable technique used to show tissue characteristics based on the diffusion motion of water molecules. DWimaging applied to abdominal and pelvic organs has beenknown to depict malignant tumours with high conspicuity[15, 16]. Also, some authors [1, 4, 17, 18, 19] have recently reported the potential value of DW imaging in small seriesof patients with endometrial cancer. Tamai et al. [4] showedthat, on DW imaging, endometrial cancer is clearlydepicted as an area of increased signal intensity relativeto the surrounding hypointense myometrium.The purpose of this study was to determine thediagnostic accuracy of DW imaging in the preoperativeassessment of myometrial tumour invasion and compare it with those of routinely used MR techniques, using postoperative histopathological findings as the referencestandard. Materials and methods Study populationThis prospective study, which was carried out at a singleinstitution during the period September 2006  –  October 2008, consisted of a total of 62 patients who underwent  preoperative MR imaging examination and subsequent total hysterectomy. They all had histopathologicallyconfirmed endometrial cancer by means of endometrial biopsy under hysteroscopy that took place between 5 and37 days before MR examination. The mean patient age was63 years (range 36  –  84), and 85% of the patient populationwas post-menopausal. Excluded from the study were 15 patients: 11 had suboptimal MR imaging examinations because of they were unable to cooperate or wereclaustrophobic; two patients had incomplete MR imagingexamination because of technical failure; the remainingtwo patients refused having dynamic imaging performed.The study was approved by our Institutional ReviewBoard, and all women included in the study gave writteninformed consent to participation.MR imagingMR examinations were performed with a 1.5-T system(Achieva Plus; Philips, The Netherlands). The MR imaging protocol included standard T1-weighted, T2-weightedturbo spin echo multishot and gadolinium-enhancedT1-weighted sequences. In addition, diffusion-weightedsequences (DW imaging) were performed. Before theMR examination, in order to reduce peristalsis artefacts,intramuscular administration of butyl-scopolamine (20 mg)was performed.T1-weighted images were obtained in the axial plane(mean field of view: 375 mm; TE: 4.6 ms; TR: 4.4 ms;number of slices: 25; slice thickness: 7 mm; slice gap:1 mm; matrix: 256×143); T2-weighted turbo spin-echoimages were obtained in the sagittal and oblique axial planes (mean field of view: 220 mm; TE: 85 ms; TR:5,000 ms; number of slices: 32; slice thickness: 3 mm; slicegap: 0.3 mm; matrix: 368×215). Furthermore, T2-weighted images were also obtained in the oblique coronal plane (field of view: 180 mm, matrix: 256×196). Theoblique axial and coronal planes were determined inrelation to the major and minor axes of the uterine body.As for dynamic imaging, contrast-enhanced T1-weighted fat-suppressed 3D gradient-echo images (fieldof view: 395 mm; number of slices: 50; slice thickness:2 mm TE: 2.1 ms; TR: 4.3 ms; matrix 256×196) wereacquired in the axial plane after the injection of gadopentetate dimeglumine (Magnevist; Schering, Berlin,Germany) at a dose of 0.2 ml/kg body weight and aninjection flow of 2 ml/s. Images were obtained at 0, 30, 60and 120 s after injection of contrast agent.DW images with echo-planar technique (field of view:400 mm; number of slices: 12; slice thickness: 6 mm; TE:49 ms; TR: 1,900 ms; b-value of 0 and 500 s/mm 2 ; matrix:144×75) were acquired in the axial plane [17]. DWimaging with a b value of 500 provided the best referenceimages concerning anatomic landmarks.Image analysisVisual evaluation of the images was performed indepen-dently by two radiologists with 5 and 10 years ’  experience,respectively, who were aware of the histopathologicaldiagnosis and were blinded to the results of postoperativehistopathological examination. At reading sessions, theMR images were presented in random order; also, when a  sequence was evaluated, the reviewer was blinded to theresults of the other sequences.For each patient, the following parameters were assessedon T2-weighted, dynamic gadolinium-enhanced and DWimaging with a b-value of 500 s/mm 2 : (1) presence of thetumour; (2) localisation of the myometrial region showingthe deepest tumour invasion; (3) depth of myometrialinvasion. The presence of the tumour was diagnosed oneach of the sequences employed according to previouslyreported criteria [20, 21, 22]. In particular, endometrial cancer was depicted on DW images as an increasedintensity area relative to surrounding hypointense myome-trium [4]. For each sequence used, the exact localisation of the deepest myometrial invasion was identified (referringto: anterior wall of the uterus; posterior wall; left-side wall;right-side wall and fundus), and in that site the extension of the tumour into the myometrium and the myometrialthickness was measured. Myometrial invasion was calcu-lated asthe ratio of the two measurementsand was classifiedinto two categories  —  no or superficial and deep  —  inagreement with previous reports [14, 23]. No or superficial myometrial invasion was defined as tumour limited to theendometrium or invading less than one half of the thicknessof the myometrium; deep invasion was defined as tumour invasion reaching one half or more of the myometrialthickness [1].Surgical procedureAll the patients included in the study underwent evaluationof peritoneal washing, total hysterectomy and bilateralsalpingo-oophorectomy 5 to 35 days after MRI examina-tion (average wait: 18 days). The surgical procedure was performed in a dedicated gynaecological oncology unit bytwo surgeons with more than 10 years ’  experience in thefield. Thirty-eight patients (81%) had video-laparoscopicsurgery, and the remaining nine patients (19%) hadstandard laparotomies.Histopathological analysisSurgical specimens were sectioned along the longitudinal plane of the uterus. A pathologist, with 15 years ’  experi-ence in gynaecological oncology, who was blinded to theimaging results, assessed the portion of the myometrialwall in which the tumour invasion was deepest andlocalised it in the same way as for MR image evaluation. Inthat site the extension of the tumour into the myometriumand the total myometrial thickness were measured, and thedepth of myometrial invasion was classified in the sameway as for the imaging analysis. In each case, thehistological type and tumour grade were established;whether or not there was a concurrent uterine benign pathological condition was also specified.Statistical analysisMean, quartiles and standard deviation were used for descriptive purposes. Cohen ’ s kappa index was used toassess the agreement between the two radiologists. In the presence of strong agreement, results of the most experienced radiologist were reported, whereas for aweaker level of agreement the results of both radiologistswere reported. For each MR technique, sensitivity,specificity and predictive values in assessing the depthof myometrial tumour invasion were determined, with pathological results as the reference standard. Ninety-fiveconfidence intervals (CI) for proportions were calculated by standard methods or according to Wilson (cited inBrown et al. [24]) if estimates were near the upper  boundary ( ≥ 75%). Because of the paired nature of thedesign, McNemar  ’ s exact test was considered in compar-ing two sequences in terms of sensitivity and specificity(two-sided test,  α  =0.05). For each combination of sequences, the ratio of sensitivities and specificities(with 95% CI) was considered as a relative measure of  performance. For all the MR techniques used, measure-ment accuracy was quantified by the difference betweenthe measurement made by the radiologist and the onemade by the pathologist, this latter serving as thereference value. Results Histopathological findingsThe postoperative histo-pathological examination docu-mented the presence of endometrioid adenocarcinoma in 44cases, adenocarcinoma with mucinous differentiation in 1caseand2casesofserouscarcinoma.Thehistologicalgradeof differentiation was G1 in 13 cases (28%), G2 in 23 cases(49%) and G3 in the remaining 11 cases (23%). In additionto endometrial cancer, 14 patients had coexisting leiomyo-mas, and 2 had adenomyosis. Histological examinationshowedthatthe tumourwasconfinedtotheendometriumor involved the inner half of the myometrium (superficialinvadingtumour) in 34 cases(72%) andtheouter halfof themyometrium (deep invading tumour) in the remaining 13cases (28%).Relevant histopathological findings are reported inTable 1.Interobserver variabilityThe analysis of interobserver agreement between the tworadiologists (Table 2) showed a substantial agreement inassessing the depth of myometrial invasion for both T2-weighted and DWimaging ( κ =0.91 and 0.74, respectively)and a moderate agreement for dynamic gadolinium-  enhanced imaging ( κ =0.45). For this latter sequence, theresults of both reviewers are reported below.Diagnostic performance of T2-weighted, dynamicand DW imagingAt MR interpretation, in all cases the presence of thetumour was identified with each pulse sequence used. At T2-weighted imaging, 27 cases (57%) were classified assuperficial invading tumour (Figs. 1, 2) and the remaining 20 (43%) as deep invading (Fig. 3); with the dynamicimaging, the same categories respectively included 25(53%) and 22 (47%) cases. DWimaging classified 26 cases(55%) as superficial and 21 cases (45%) as deep tumours.By comparing individual measurements on MR imagesobtained with the three pulse sequences with thoseobtained on pathological specimens, a systematic over-estimation of the percentage of tumour spread was found.This was mainly because of an average underestimationequal to 5.0, 4.9 and 7.0 mm in the measurement of myometrial thickness with the use of T2-weighted,dynamic and DW imaging, respectively (Fig. 4). Tumour invasion also tended to be underestimated with DWand dynamic gadolinium-enhanced imaging, with thelatter showing less precision. Table 3 shows that incorrect evaluation of tumour spread was mainlycaused by the prevalence of false-positive results. Inthese cases the quantification of myometrial invasionsuffered from the combined effect of wall thicknessunderestimation and tumour extension overestimation,with all the pulse sequences used.In determining myometrial invasion, the sensitivity,specificity, positive and negative predicted values of T2-weighted imaging were: 92.3%, 76.5%, 60.0% and 96.3%.The corresponding values for dynamic gadolinium-en-hanced imaging were: 69.2%, 61.8%, 40.9% and 84.0%; of importance, specificity and positive predictive value for thesecond reviewer were significantly different (52.9 and36.0%, respectively). For DW imaging, values were84.6%, 70.6%, 52.4% and 92.3% (Table 3). As shown inTable 4, dynamic and DW imaging detected 25% and 8%fewer deep invasive tumours relative to T2-weightedimaging, whereas DW imaging detected 22% more deepinvasive tumours relative to dynamic imaging. Suchdifferences in sensitivity, however, did not reach statisticalsignificance because of a low prevalence of postoperativedeep invasions. Pairwise differences are less marked andnot significant in the ability to detect non-invasive tumours. Discussion For clinical decision-making, MR imaging is considered[25] the most important tool among preoperative examina-tions, T2-weighted and dynamic gadolinium-enhanced T1-weighted images being part of the standard protocolcurrently used for preoperative local staging of endometrialcarcinoma. Most authors [20, 26] classify myometrial invasion by tumour as deep or superficial, instead of considering three categories of invasion (no myometrialinvasion; invasion of the inner half of the myometrium;invasion of the outer half of the myometrium) as thedistinction between no tumour invasion and superficialinvasion proved to have limited clinical relevance in termsof prognosis and choice of treatment.DW imaging is a recently prevailing technique that enables visualisation of random microscopic motion of thewater molecules, thereby providing a tissue contrast different from that of conventional T1- and T2-weightedimages [17]. Pathological processes such as inflammationand neoplasia tend to alter structural organisation mainly Table 1  Relevant histopathological findingsParameter (n=47) Mean 1st quartile median 3rd quartile Standard deviationTumour size (mm) 35.1 18.0 30.0 45.0 25.5Tumour myometrial invasion (mm) 6.0 2.0 5.0 9.0 5.1Myometrial thickness (mm) 18.4 14.0 17.0 22.0 6.9 Table 2  Interobserver agreement in evaluating myometrial tumour invasion as assessed by Cohen ’ s  κ  and 95% confidence intervals (CI)Imaging sequence (n=47) % Agreement   κ  (95% CI)Overall Positive NegativeT2W 96 95 96 0.91 (0.79  –  1.00)Dynamic T1W 72 72 72 0.45 (0.20  –  0.71)DW 87 86 88 0.74 (0.55  –  0.93)   by modifications in cellularity [27]. These changes canaffect proton mobility and diffusivity, which can beobserved on DW imaging where regions of increasedcellularity and restricted water diffusion are displayed asareas of high signal intensity [28].In our study, we have determined the diagnosticaccuracy of DW images in the preoperative assessment of myometrial invasion by endometrial carcinoma and thencompared it with those of conventional T2-weighted anddynamic MR images. Our results support the consensusthat MR imaging can be used to discriminate betweensuperficial and deep myometrial invasion by endometrialcancer with high accuracy.In particular, this study confirms that a lack of deepmyometrial invasion on preoperative MR imaging allows it to be ruled out reliably [14]. Such knowledge has criticalimportance in selecting the most appropriate surgicaltreatment in these patients, that is, whether to perform pelvic and/or para-aortic lymphadenectomy. Our resultsalso indicate that T2-weighted imaging performs better than either DW or dynamic MR imaging. DW imaging,however, appeared to be more reliable and to have higher diagnostic performance than dynamic gadolinium-en-hanced imaging.As for the performance of T2-weighted imaging, weobtained a specificity of 76.5%, which is similar to thosefound in previous studies where a range of specificityvalues of 73  –  91% has been reported [11, 18, 20, 26] and a sensitivity of 92.3%, which is higher than in most of the previous works. Reasons for errors were the presence of an Fig. 1  A 71-year-old woman with histopathologically confirmedsuperficial myometrial invasion by endometrial carcinoma.  a Oblique axial T2-weighted turbo spin-echo image shows tumour tissue as a layer of intermediate intensity in the posterior andanterior portions of the uterine cavity, with irregularity of the junctional zone ( arrow ) and invasion of the inner half of themyometrium (<50%).  b  Axial gadolinium-enhanced T1-weightedimage (venous phase) shows irregularity of the early subendometrialenhancement in the anterior wall ( arrow ); the extent of the tumour into the myometrium however is not well defined.  c  Axial diffusion-weighted image (b=500) shows endometrial carcinoma as a well-defined area of high signal intensity that invades less than half of thelow signal intensity myometrium Fig. 2  Histopathologically confirmed superficial myometrial inva-sion by endometrial carcinoma in a 56-year-old woman. In this patient, all sequences overestimated the extent of myometrialinvasion (false-positive findings).  a  Oblique axial T2-weightedturbo spin-echo image shows discontinuity of the low-signal-intensity junctional zone because of intermediate intensity tumour spread that seems to extend into the outer half of the myometrium( arrow ).  b  Axial gadolinium-enhanced T1-weighted image (venous phase) shows endometrial cancer as a hypovascular area within theenhancing myometrium, with apparent extension into the outer half of the myometrium ( arrow ).  c  Axial diffusion-weighted image (b=500) shows the tumour as a high-signal-intensity mass apparentlyinvading more than 50% of the right wall of the myometrium
Similar documents
View more...
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

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!