Accepted Manuscript Title: Magnetic resonance imaging at one year for detection of postoperative residual cholesteatoma in children: is it too early? Author: A. Lecler M. Lenoir J. Peron F. Denoyelle E.N.Garabedian H. Ducou le Pointe J. Nevoux PII: DOI: Reference:
S0165-5876(15)00248-7 http://dx.doi.org/doi:10.1016/j.ijporl.2015.05.028 PEDOT 7604
To appear in:
International Journal of Pediatric Otorhinolaryngology
Received date: Revised date: Accepted date:
10-2-2015 16-5-2015 19-5-2015
Please cite this article as: A. Lecler, M. Lenoir, J. Peron, F. Denoyelle, H.D. Pointe, J. Nevoux, Magnetic resonance imaging at one year for detection of postoperative residual cholesteatoma in children: is it too early?, International Journal of Pediatric Otorhinolaryngology (2015), http://dx.doi.org/10.1016/j.ijporl.2015.05.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Title Page
Magnetic resonance imaging at one year for detection of postoperative residual cholesteatoma in children : is it too early ? A. Lecler1,2, M. Lenoir1, J. Peron3, F Denoyelle4, E.N.Garabedian4, H. Ducou le Pointe1*, J. Nevoux5*
cr
Hôpitaux de Paris, 26 avenue du docteur Arnold Netter, 75012 Paris, France
ip t
1. Service de Radiologie pédiatrique, Hôpital Trousseau, Université Pierre et Marie Curie, Assistance Publique
us
2. Service de Neuroradiologie diagnostique, Fondation Rothschild, 25 rue Manin, 75019 Paris, France 3. Centre anticancéreux Léon Bérard, Oncologie Médicale, 28 rue Laennec, 69008 Lyon, France
an
4. Service d’Oto-Rhino-Laryngologie pédiatrique, Hôpital Necker, Université Paris René Descartes, Assistance Publique Hôpitaux de Paris, 149 rue de Sèvres, 75015 Paris, France
M
5. Service d’Oto-Rhino-Laryngologie, INSERM U1185, Hôpital Bicêtre, Université Paris Sud, Assistance Publique Hôpitaux de Paris, 78 rue du Général Leclerc, 94270 Le Kremlin-Bicêtre, France
Corresponding author
Ac ce p
Dr Augustin Lecler
te
d
* These authors equally participate to this work
Service de Neuroradiologie diagnostique Fondation Rothschild
25 rue Manin, 75019 Paris
E-mail: [email protected] Phone : +33148036464
Fax : +33148036401
1 Page 1 of 31
Abstract Objective: To compare the residual cholesteatoma detection accuracy of diffusion-weighted (DW) and T1 delayed sequences for magnetic resonance at one year postoperative with second-look surgery in pediatric patients who have
ip t
undergone primary middle ear surgery for cholesteatoma. Methods: This was a prospective monocentric consecutive study conducted in a tertiary academic referral center.
cr
Children were referred for MR imaging (MRI) one year after surgery. A 1.5 T MRI was utilized, using nonechoplanar DW images and delayed gadolinium-enhanced T1-weighted images. Accuracy of magnetic resonance
us
imaging was assessed by two radiologists before surgery. Interobserver and intraobserver agreements were assessed using the test. Magnetic resonance imaging data were compared with surgery, which was considered as the gold
an
standard.
Results: Twenty-four consecutive unselected pediatric patients were included. Sensitivity, specificity, positive
M
predictive value, and negative predictive value for the first observer were of 40%, 86%, 67%, and 67%, respectively, and those for the second observer were 30%, 86%, 60%, and 63%, respectively. The only two cholesteatoma with a
te
d
size superior to 3 mm were diagnosed before surgery, but the majority of small cholesteatoma were not detected. Conclusions : MRI is a key examen to diagnosed the residual cholesteatoma but is limited by the size of the lesion under 3 mm. Delaying the realization of MRI during follow-up could increase sensitivity, thus avoiding
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
misdiagnosis as well as unnecessary second look surgery.
1 Page 2 of 31
Keywords Residual cholesteatoma
ip t
MRI DW imaging
us
cr
Pediatric
DW MRI: diffusion-weighted magnetic resonance imaging
M
EPI: echo planar imaging
te
PPV: positive predictive value
d
TSE: turbo spin echo CT: computed tomography
an
Abbreviations and acronyms
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
NPV: negative predictive value TR: time of repetition (ms) TE: time echo (ms)
2 Page 3 of 31
1.
INTRODUCTION
Cholesteatoma is a non-neoplastic but destructive cystic lesion of keratinizing stratified squamous epithelium, which can lead to labyrinthine, facial or cerebromeningeal complications. Surgery, mainly canal wall-up
ip t
tympanomastoidectomy, is therefore necessary to eradicate cholesteatoma of the middle ear. Besides cholesteatoma disease presents several major differences in children compared with adults. It is more aggressive in children. There
cr
is a higher residual rate after surgery: 30% in children vs 3%-15% in adults.[1] Otoscopic evaluation may be more
prevents an optimal detection of retrotympanic cholesteatomas.
us
challenging because of the frequent use of cartilaginous graft for tympanic membrane repair in children, which
Thus, the main objective of the follow-up after initial surgery is to rule out the presence of a residual cholesteatoma.
an
Because of the frequent use of cartilaginous tympanic membrane repair, “second-look” surgery is currently the gold standard to detect and treat residual disease in children. However, imaging plays a major role in the follow-up. CT
M
can be useful to detect a residual cholesteatoma; a lobulated or heterogeneous soft tissue mass associated with focal bony erosion is highly evocative. Unfortunately, CT becomes unreliable when the postoperative cavity is partially or
d
completely opacified, which is more frequent in children.[2, 3]
te
Conversely, magnetic resonance imaging (MRI) has progressively gained importance in the follow-up because of performing sequences, such as delayed gadolinium-enhanced T1-weighted images (WI) and diffusion-weighted
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
(DW) images. Diffusion-weighted magnetic resonance images (DW MRI) relies on the signal produced by the diffusion motion of water protons in biologic tissues. Stratification of keratin in residual cholesteatoma has high signal intensity, whereas other tissues that can be found in the middle ear after surgery show low or intermediate signal intensity on DW images.[4] Delayed gadolinium-enhanced T1 WI relies on the nonenhancement of residual cholesteatoma, whereas surrounding inflammatory fibrous tissue is enhanced. This allows a correct distinction between disease and reactive tissue.[5]
Despite the well documented follow-up with MRI in adults, there is too few studies evaluating MRI in pediatric follow-up, and the number of pediatric patients included is too small to provide reliable data.[6–9] Cholesteatoma residual disease presents major differences in children versus adults, preventing a reliable extrapolation from adult series.
3 Page 4 of 31
The aim of our study was to compare MRI performances to “second-look” surgery in the evaluation of postoperative residual cholesteatoma in children. Our goal was to use MRI follow-up as a reliable tool to delay or prevent second-
te
d
M
an
us
cr
ip t
look surgery.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
4 Page 5 of 31
2.
MATERIALS AND METHODS
2.1 PATIENTS This is a monocentric study in a tertiary academic referral center. All consecutive children with cholesteatoma
ip t
requiring second-look surgery and managed in our institution between June and October 2010 were retrospectively included. Decision for the second-look surgery was made by the surgeon on the basis of findings at first-stage
cr
surgery and on clinical follow-up.[10]
us
We collected information on clinical, radiological, surgical, and pathological findings. Children underwent an MR imaging before “second-look” surgery, which was programmed 12 months after the initial surgery.
an
This study was approved by the local ethical committee. Informed consent was obtained from all individual
M
participants included in the study. 2.2 IMAGING TECHNIQUE
d
MR imaging was performed with a 1.5 T MRI scanner (Achieva; Philips Medical Systems, Best, The Netherlands),
te
using spin echo DW (SE-DW) images and delayed contrast-enhanced T1-weighted images. A head coil was used. Coronal 2.5-mm-thick SE-DW sequences were obtained with the following parameters: ECG triggering; RT(repetition time)/ET (echo time), 3000/106; field of view (FOV), 230 × 230; matrix 176 × 100; b factors of 0 and
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
800 s/mm²; and acquisition time, 3 minutes 12 seconds. Axial 4.5-mm-thick dual turbo SE T2-weighted images were obtained with the following parameters RT/ET, 2293/110; FOV, 230 × 230; matrix, 256 × 168; acquisition time, 1 minute 45 seconds. Coronal 1.7-mm-thick turbo SE T2-weighted images were obtained with the following parameters: RT/ET, 3500/120; FOV, 230 × 230; matrix, 336 × 212; acquisition time, 4 minutes 26 seconds. Axial and coronal 2-mm-thick SE T1-weighted images were performed 30 to 45 minutes after intravenous contrast injection of 0.1 mmol per kilogram of body weight of gadoterate meglumine (Dotarem®; Guerbet, Roissy, France): RT/ET, 526/20; FOV, 180 × 180; matrix, 205 × 204; acquisition time, 3 minutes 36 seconds (Table A).
5 Page 6 of 31
Light sedation, consisting of 5 mg/kg of pentobarbital, was used in children above the age of five, and no sedation was used for younger children. No general anesthesia was used.
ip t
2.3 IMAGING AND SURGICAL EVALUATION
cr
Criteria for diagnosis of residual cholesteatoma was very high signal intensity in the middle ear or in the mastoid cavity on DWI sequence compared with brain tissue, corresponding to restricted diffusion on ADC maps.
us
Unenhanced intermediary to low signal mass of the soft tissue mass within the postoperative cavity on delayed contrast-enhanced T1-weighted images was considered as a residual cholesteatoma. A complete (homogeneous)
an
enhancement of the mass was conclusive for noncholesteatomatous postoperative tissue.[11] All images were analyzed by two radiologists blinded to any surgical and clinical information about patients.
M
Imaging data were classified into two groups: “residual cholesteatoma” and “no cholesteatoma.” Radiologists described precise localization and measured diameter of all detected residual cholesteatoma.
d
Second-look surgery was performed by one of 6 experienced surgeons (2 to 30 years of experience in cholesteatoma
te
surgery). Surgical results were classified as “residual cholesteatoma” and “no residual cholesteatoma.” Cholesteatoma localized in the middle ear was measured during surgery.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Histologic evaluation was systematically performed, and cholesteatoma were measured again.
2.4 STATISTICAL ANALYSIS
Nonweighted kappa statistics with the 95% confidence interval (CI) was used to test for interobserver agreement of the 3-step classification system using arbitrary interpretation by Landis and Koch (0, poor agreement; 0.00–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.80–1.00, almost perfect agreement).[12]
6 Page 7 of 31
Sensitivity, specificity, positive (PPV), and negative predictive values (NPV) were assessed. We decided to keep statistical values from the most experimented radiologist to discuss our results and to compare our statistics with those of other series.
ip t
Statistical analysis was performed using R software (http://www.r-project.org/). Exact confidence intervals were
te
d
M
an
us
cr
based on a binomial distribution.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
7 Page 8 of 31
3.
RESULTS
Twenty-four consecutive children (22 males, 2 females) with scheduled second-look surgery were treated during our study. Mean age at surgery was 10.5 years (range, 4-18 years) (Table B).
ip t
All patients underwent MRI according to our protocol. MRI to surgery delay was very short: less than a week for 15 patients (63%) and over 2 weeks for only 5 patients (21%). The mean delay was 12.8 days (Table B). Mean MRI
cr
acquisition time was 25 minutes (range, 15-38) with a median time of 23 minutes.
us
All 24 patients underwent surgery. Residual cholesteatoma was surgically confirmed in 10 patients (41.6%), whereas no residual was found in 14 patients. Thirteen residual cholesteatoma lesions were found in these 10
an
patients: 9 patients with only one lesion each, and one patient with 4 lesions (Table B).
Residual cholesteatoma localizations were as follows: epitympanum (n = 7) (Fig A), hypotympanum (n = 2),
M
mesotympanum (n = 2), and mastoid recess (n = 2) (Fig B; Table B).
Median size was 2 mm (range, 0-8 mm). Among residual cholesteatoma shapes, there were 11 pearls, 1 with a
te
d
ruptured bag and 1 with a completely empty bag. The pathological analysis confirmed theses data. First and second observer identified residual cholesteatoma in 6 and 5 patients, respectively. Among these results, there were 6 and 7 false negatives and 2 false positives for each, respectively (Fig C).
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
One residual with an empty bag, one ruptured cholesteatoma with low keratin content, and one cholesteatoma within the mastoid recess were missed by both readers (Fig D). All missed residual cholesteatoma were less than 3 mm. Epitympanic localizations were almost correctly detected by MRI. Other localizations (mastoid recess and mesotympanum) were missed by both readers. All residual cholesteatoma correctly identified by the two readers had a pearl shape, 2 were in the epitympanum and one in the hypotympanum. Their sizes were 2, 4, and 8 mm. Two residual cholesteatoma with a size superior to 3 mm were diagnosed by imaging before surgery.
8 Page 9 of 31
Sensitivity, specificity, PPV, and NPV of MRI were of 40%, 86%, 67%, and 67%, respectively, for the first observer and 30%, 86%, 60%, and 63%, respectively, for the second observer. Interobserver agreement was substantial (k = 0.647).
ip t
We extracted statistical data from the second observer, which had the longest experience reading middle ear MRI, to calculate the following data: positive likelihood ratio was 2.1, negative likelihood ratio was 0.8, and accuracy was
te
d
M
an
us
cr
0.63.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
9 Page 10 of 31
4.
DISCUSSION
The present study demonstrated that MRI in children may be useful in the detection of residual cholesteatoma above 3 mm before second-look surgery but lacks accuracy for small residuals. It is, according to our knowledge, the
ip t
largest homogeneous pediatric series published in the literature. Cholesteatoma disease presents several major differences in children compared with adults. Cholesteatoma is more
cr
aggressive in children, and there is a higher residual rate after surgery.[1] There is a clear distinction between
us
residual and recurrence of cholesteatoma.[11] All recurrences are diagnosed with otoscopy, whereas residuals are most of the time an imaging or perioperative diagnosis. Adult series often mix recurrence and residual
an
cholesteatoma, making comparison between adult and pediatric series more hazardous.[13] Otoscopic evaluation may be more challenging, often impossible, because of the frequent use of cartilaginous graft for tympanic membrane repair in children. That explains why otoscopic exam may be insufficient to confidently detect
M
retrotympanic cholesteatoma residual.[11] Moreover, children present more often than adults a completely filled middle ear, leading to a noncontributive CT. These semiology points are very important issue and explain why we
d
should not extrapolate data from adult series to children, which shows the absolute necessity to get specific data for
te
children in cholesteatoma follow-up. Yet, a huge majority of series evaluating cholesteatoma disease are mixed, including children and adults.[13][14] Dedicated pediatric series are often inhomogeneous, mixing residual, recurrences, congenital and acquired cholesteatoma without any distinction.[8, 9] Our study was based on a
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
homogeneous population of consecutive children before second-look surgery, with possible residual of acquired cholesteatoma only and with MRI performed just before second-look surgery. It provided strong and specific data, directly relevant in a current practice. Imaging to surgery interval was as short as possible, to have the better correlation between imaging and surgical findings. This interval is critical in children because cholesteatoma disease is reported to be more aggressive.[15] The present study demonstrated the feasibility of short delays between imaging and surgery. Decision of second-look surgery was made during the first surgery, according to validated predictive risk factors of residual cholesteatoma.[10, 16]This allowed to schedule early both second-look surgery and imaging. In the literature, MRI appears to be a very strong tool in detection of residual cholesteatoma in adults [4, 13, 14]. In small dedicated pediatric series as well, Rajan or Plouin found impressive results with 100% PPV, NPV, sensibility
10 Page 11 of 31
and specificity using non-EP DW MRI only or fused EP DW MR and CT.[6, 9] In our study, detection rate was much lower but we have also more patients included. We tried to identify the reasons by reviewing all cases with the surgeon and pathologist : first of all, residual cholesteatoma size was mainly small (median, 2 mm). Williams and
ip t
De Foer showed that delayed postgadolinium T1-enhanced sequences and non-EP DW MR sequences, respectively, could discriminate a 2-mm-diameter cholesteatoma mass.[5, 19] Yet, the majority of small cholesteatoma less than 3 mm are not detected, even in recent series[20] (Table C). Secondly, 2 over 13 residual were infiltrative or had an
cr
empty bag, which is hard to detect using MRI. It is now clearly identified that one of the major causes of false
us
negatives of the DW MRI in children and adults is related to the presence of dry, autoevacuating retraction pockets.[21] The autoevacuation eliminates keratin accumulation, thus removing the substrate that causes the
an
restricted diffusion signal on MRI.[22] In Ganaha series, there was only one conventional residual cholesteatoma for 3 adhesive-type cholesteatoma and 4 evacuated attic cholesteatoma, explaining a relatively low sensitivity of 69% with 18 false-negative cases. [20] Third, comparison with other pediatric series is difficult, because no strong data is
M
available for now in the litterature : no dedicated pediatric series included enough children to provide significant results, with only 18 patients for the largest series[7]. Too few residual cholesteatoma were found, with only 2
d
residual in some series[6], and residual size ranged from 3 mm to 26 mm with a mean size of 1cm in some others.[8,
te
9] (Table C) Therefore, excellent results found in literature should be interpreted with high caution. Imaging remains mandatory to perform a complete middle ear exploration because some areas are totally blind to
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
the otoscopic exam, like mastoid recess or epitympanum. In our study, epitympanic residuals were almost always correctly detected using MRI (Fig B). On the contrary, residual localized in the mastoid recess, mesotympanum, and hypotympanum were missed by both readers. Two parameters explained these results: on the one hand, DW MRI skull-base artifacts and air-bone interfaces are not completely avoided even with non-EP techniques, and on the other hand, residual cholesteatomas were particularly small in these precise locations in our study (the biggest residual in mastoid recess was 3 mm). Plouin recently used a fusion technique to combine DW MRI and CT in 10 children. This allowed a correct diagnosis and precise localization in all cases and provided precious preoperative information for the surgeon.[9] There are several limitations in our study: first, the number of included children is small, thus limiting our statistical analysis. We had only 2 residual cholesteatomas above 3 mm, correctly detected by both readers, but this number is
11 Page 12 of 31
too small to conclude that all residuals above 3 mm would have been found. Second, we did not compare DW MRI versus delayed gadolinium-enhanced T1-weighted sequences, thus preventing us to conclude that DW MRI are sufficient to detect residual. Finally, we tried to prevent selection bias by including consecutive patients in a
ip t
prospective way, but we had too many “small” residuals under 3 mm, leading to an artificially low sensitivity and specificity.
cr
We guess that MRI control should be performed later than 12 months after primary operation to avoid misdiagnosis
performed in case of high presumption of residual cholesteatoma.
us
of small residual cholesteatoma, without unwisely delaying adequate treatment. Second-look surgery is still
an
Further studies with a larger number of children should be realized to provide strong data and more accurate
te
d
M
statistical analysis.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
12 Page 13 of 31
5.
CONCLUSION
MRI may be useful in the detection of residual cholesteatoma above 3 mm before second-look surgery in children but lacks accuracy for small residuals. Delaying the realization of MRI during follow-up could increase sensitivity,
te
d
M
an
us
cr
ip t
thus avoiding misdiagnosis as well as unnecessary second look surgery.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
13 Page 14 of 31
Compliance with Ethical Standards Conflicts of interest :
ip t
None Ethical approval:
cr
All procedures performed in studies involving human participants were in accordance with the ethical standards of
us
the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
an
Informed consent:
te
d
M
Informed consent was obtained from all individual participants included in the study.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
14 Page 15 of 31
REFERENCES 1. Darrouzet V, Duclos JY, Portmann D, Bebear JP (2000) Preference for the closed technique in the management of cholesteatoma of the middle ear in children: a retrospective study of 215 consecutive patients treated over 10 years. Am J Otol 21:474–481.
ip t
2. Williams MT, Ayache D (2004) Imaging of the postoperative middle ear. Eur Radiol 14:482–495. doi: 10.1007/s00330-003-2198-8
cr
3. Lemmerling MM, De Foer B, VandeVyver V, et al. (2008) Imaging of the opacified middle ear. Eur J Radiol 66:363–371. doi: 10.1016/j.ejrad.2008.01.020
us
4. Dubrulle F, Souillard R, Chechin D, et al. (2006) Diffusion-weighted MR imaging sequence in the detection of postoperative recurrent cholesteatoma. Radiology 238:604–610. doi: 10.1148/radiol.2381041649 5. Williams MT, Ayache D, Alberti C, et al. (2003) Detection of postoperative residual cholesteatoma with delayed contrast-enhanced MR imaging: initial findings. Eur Radiol 13:169–174. doi: 10.1007/s00330-002-1423-1
an
6. Rajan GP, Ambett R, Wun L, et al. (2010) Preliminary outcomes of cholesteatoma screening in children using non-echo-planar diffusion-weighted magnetic resonance imaging. Int J Pediatr Otorhinolaryngol 74:297– 301. doi: 10.1016/j.ijporl.2009.12.011
M
7. Geoffray A, Guesmi M, Nebbia JF, et al. (2013) MRI for the diagnosis of recurrent middle ear cholesteatoma in children--can we optimize the technique? Preliminary study. Pediatr Radiol 43:464–473. doi: 10.1007/s00247-012-2502-3
te
d
8. Plouin-Gaudon I, Bossard D, Fuchsmann C, et al. (2010) Diffusion-weighted MR imaging for evaluation of pediatric recurrent cholesteatomas. Int J Pediatr Otorhinolaryngol 74:22–26. doi: 10.1016/j.ijporl.2009.09.035 9. Plouin-Gaudon I, Bossard D, Ayari-Khalfallah S, Froehlich P (2010) Fusion of MRIs and CT scans for surgical treatment of cholesteatoma of the middle ear in children. Arch Otolaryngol Head Neck Surg 136:878–883. doi: 10.1001/archoto.2010.151
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
10. Roger G, Denoyelle F, Chauvin P, et al. (1997) Predictive risk factors of residual cholesteatoma in children: a study of 256 cases. Am J Otol 18:550–558. 11. Nevoux J, Lenoir M, Roger G, et al. (2010) Childhood cholesteatoma. Eur Ann Otorhinolaryngol Head Neck Dis 127:143–150. doi: 10.1016/j.anorl.2010.07.001 12. Landis JR, Koch GG (1977) An application of hierarchical kappa-type statistics in the assessment of majority agreement among multiple observers. Biometrics 33:363–374. 13. De Foer B, Vercruysse J-P, Bernaerts A, et al. (2010) Middle ear cholesteatoma: non-echo-planar diffusionweighted MR imaging versus delayed gadolinium-enhanced T1-weighted MR imaging--value in detection. Radiology 255:866–872. doi: 10.1148/radiol.10091140 14. Vercruysse J-P, De Foer B, Pouillon M, et al. (2006) The value of diffusion-weighted MR imaging in the diagnosis of primary acquired and residual cholesteatoma: a surgical verified study of 100 patients. Eur Radiol 16:1461–1467. doi: 10.1007/s00330-006-0160-2 15. Stangerup SE, Drozdziewicz D, Tos M (1999) Cholesteatoma in children, predictors and calculation of recurrence rates. Int J Pediatr Otorhinolaryngol 49 Suppl 1:S69–73.
15 Page 16 of 31
16. Lazard DS, Roger G, Denoyelle F, et al. (2007) Congenital cholesteatoma: risk factors for residual disease and retraction pockets--a report on 117 cases. Laryngoscope 117:634–637. doi: 10.1097/mlg.0b013e318030ac8c
ip t
17. Dhepnorrarat RC, Wood B, Rajan GP (2009) Postoperative non-echo-planar diffusion-weighted magnetic resonance imaging changes after cholesteatoma surgery: implications for cholesteatoma screening. Otol Neurotol 30:54–58. doi: 10.1097/MAO.0b013e31818edf4a
cr
18. Jindal M, Riskalla A, Jiang D, et al. (2011) A systematic review of diffusion-weighted magnetic resonance imaging in the assessment of postoperative cholesteatoma. Otol Neurotol 32:1243–1249. doi: 10.1097/MAO.0b013e31822e938d
us
19. De Foer B, Vercruysse J-P, Pilet B, et al. (2006) Single-shot, turbo spin-echo, diffusion-weighted imaging versus spin-echo-planar, diffusion-weighted imaging in the detection of acquired middle ear cholesteatoma. AJNR Am J Neuroradiol 27:1480–1482. 20. Ganaha A, Outa S, Kyuuna A, et al. (2011) Efficacy of diffusion-weighted magnetic resonance imaging in the diagnosis of middle ear cholesteatoma. Auris Nasus Larynx 38:329–334. doi: 10.1016/j.anl.2010.11.004
an
21. Fitzek C, Mewes T, Fitzek S, et al. (2002) Diffusion-weighted MRI of cholesteatomas of the petrous bone. J Magn Reson Imaging 15:636–641. doi: 10.1002/jmri.10118
M
22. De Foer B, Vercruysse J-P, Bernaerts A, et al. (2007) The value of single-shot turbo spin-echo diffusionweighted MR imaging in the detection of middle ear cholesteatoma. Neuroradiology 49:841–848. doi: 10.1007/s00234-007-0268-3
d
23. Lehmann P, Saliou G, Brochart C, et al. (2009) 3T MR imaging of postoperative recurrent middle ear cholesteatomas: value of periodically rotated overlapping parallel lines with enhanced reconstruction diffusion-weighted MR imaging. AJNR Am J Neuroradiol 30:423–427. doi: 10.3174/ajnr.A1352
te
24. De Foer B, Vercruysse J-P, Bernaerts A, et al. (2008) Detection of postoperative residual cholesteatoma with non-echo-planar diffusion-weighted magnetic resonance imaging. Otol Neurotol 29:513–517. doi: 10.1097/MAO.0b013e31816c7c3b
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
25. Venail F, Bonafe A, Poirrier V, et al. (2008) Comparison of echo-planar diffusion-weighted imaging and delayed postcontrast T1-weighted MR imaging for the detection of residual cholesteatoma. AJNR Am J Neuroradiol 29:1363–1368. doi: 10.3174/ajnr.A1100 26. Stasolla A, Magliulo G, Parrotto D, et al. (2004) Detection of postoperative relapsing/residual cholesteatomas with diffusion-weighted echo-planar magnetic resonance imaging. Otol Neurotol 25:879–884. 27. Aikele P, Kittner T, Offergeld C, et al. (2003) Diffusion-weighted MR imaging of cholesteatoma in pediatric and adult patients who have undergone middle ear surgery. AJR Am J Roentgenol 181:261–265. doi: 10.2214/ajr.181.1.1810261
16 Page 17 of 31
FIGURES Fig. A: Four-mm residual cholesteatoma of the epitympanum
ip t
Fig. A1: Coronal spin echo diffusion-weighted (SE-DW) image: hypersignal (white arrow) Fig. A2 : Coronal delayed contrast-enhanced T1-weighted image: lack of enhancement in an area of fibrosis (white
cr
arrow)
arrow)
an
Fig. B: Three-mm residual cholesteatoma of the mastoid recess
us
Fig. A3 : Axial delayed contrast-enhanced T1-weighted image: lack of enhancement in an area of fibrosis (white
Fig. B1: Coronal spin echo diffusion weighted (SE-DW) image: hypersignal (white arrow)
M
Fig. B2 : Coronal delayed contrast-enhanced T1-weighted image: lack of enhancement (white arrow)
d
Fig. B3 : Axial delayed contrast-enhanced T1-weighted image: lack of enhancement (white arrow)
hypersignal (white arrow)
te
Fig.C: False positive for both readers. Coronal spin echo diffusion-weighted image: left epitympanum nodular
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Fig.D: False negative for both readers. Coronal spin echo diffusion-weighted image. At surgery, residual disease was a ruptured bag in the left mesotympanum (white arrow)
17 Page 18 of 31
Table(s)
-Table A: MRI parameters SE T2
DWI B = 0
Delayed contrast
And B = 800 s/mm²
enhanced T1 WI 180
230
230
230
Slice thickness (mm)
4.5
1.7
2.5
TR (ms)
2293
3500
3000
TE (ms)
110
120
106
Number of excitations
2
6
1
Number
12
14
Acquisition plane
Axial
Coronal
Acquisition time (min)
1.45
4.26
Matrix
256 × 168
336 × 212
526
Coronal
Axial and coronal
3.12
3.36
176 × 100
205 × 204
9
11
Ac ce p
te
d
M
slice/partition
20 2
an
of
2
us
FOV (mm)
ip t
Dual SE T2
cr
Sequence
1 Page 19 of 31
Table(s)
-Table B: Patients and residual cholesteatoma characteristics Second
look to
size (mm)
localization
= no residual
reader (0 =
MRI
detected; 1 =
no residual
(days)
residual
detected; 1 =
detected)
residual
11
M
1
2
12
M
29
3
18
M
1
4
8
M
1
5
8
M
2
6
8
M
1
7
16
M
1
8
7
M
1
9
4
M
9 9 10 11
9
Ac ce p
9
2
te
1
ip t
First reader (0
cr
(years)
Residual
us
Gender
Residual
an
surgery
Second-
Hypotympanum
M
Age at
d
Patients
detected)
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
Epitympanum
1
0
2
Epitympanum
1
0
2
Epitympanum
1
0
2
Mesotympanum
0
0
14
M
7
4
Epitympanum
1
1
5
M
13
0 (empty
Mastoid recess
0
0
0
0
0
0
12
10
M
14
13
6
M
15
bag)
1 (ruptured
Mesotympanum
bag) 14
9
M
66
0
0
15
13
M
1
0
0
1 Page 20 of 31
15
M
87
8
Epitympanum
1
1
17
6
F
7
2
Hypotympanum
0
0
18
16
M
2
0
0
19
16
M
21
1
1
20
8
F
9
2
Epitympanum
21
10
M
14
3
Mastoid recess
22
13
M
4
23
9
M
1
24
10
M
1
10.5
M
12.8
0
0
1
cr
0
0
0
1
1
us
0
Epitympanum
te
d
M
an
2
0
Ac ce p
mean
5
Ac ce p
Jindal 2011[18]
De Foer 2010[13]
63
83/87
96/56
NA
Dhepnorrarat 2009[17]
22
100/100
100/100
3
Lehmann 2009[23]
35
90-100/10
100/89-100
3
De Foer 2008[24]
19
100/96
3
Venail 2008[25]
31
80/50
5
93/100
2
60/72 100/88 (if >5 mm)
De Foer 2007[22]
21
Vercruysse 2006[14]
100 (31 children)
12.5 /100
100/84
5
Dubrulle 2006[4]
24
100/91
93/100
5-24
Stasolla 2004[26]
18
86/100
100/92
5
1 Page 22 of 31
17
77/100
75/100
5
Ac ce p
te
d
M
an
us
cr
ip t
Aikele 2003[27]
2 Page 23 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 24 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 25 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 26 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 27 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 28 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 29 of 31
Ac
ce
pt
ed
M
an
us
cr
i
Figure(s)
Page 30 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 31 of 31
S0165-5876(15)00248-7 http://dx.doi.org/doi:10.1016/j.ijporl.2015.05.028 PEDOT 7604
To appear in:
International Journal of Pediatric Otorhinolaryngology
Received date: Revised date: Accepted date:
10-2-2015 16-5-2015 19-5-2015
Please cite this article as: A. Lecler, M. Lenoir, J. Peron, F. Denoyelle, H.D. Pointe, J. Nevoux, Magnetic resonance imaging at one year for detection of postoperative residual cholesteatoma in children: is it too early?, International Journal of Pediatric Otorhinolaryngology (2015), http://dx.doi.org/10.1016/j.ijporl.2015.05.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Title Page
Magnetic resonance imaging at one year for detection of postoperative residual cholesteatoma in children : is it too early ? A. Lecler1,2, M. Lenoir1, J. Peron3, F Denoyelle4, E.N.Garabedian4, H. Ducou le Pointe1*, J. Nevoux5*
cr
Hôpitaux de Paris, 26 avenue du docteur Arnold Netter, 75012 Paris, France
ip t
1. Service de Radiologie pédiatrique, Hôpital Trousseau, Université Pierre et Marie Curie, Assistance Publique
us
2. Service de Neuroradiologie diagnostique, Fondation Rothschild, 25 rue Manin, 75019 Paris, France 3. Centre anticancéreux Léon Bérard, Oncologie Médicale, 28 rue Laennec, 69008 Lyon, France
an
4. Service d’Oto-Rhino-Laryngologie pédiatrique, Hôpital Necker, Université Paris René Descartes, Assistance Publique Hôpitaux de Paris, 149 rue de Sèvres, 75015 Paris, France
M
5. Service d’Oto-Rhino-Laryngologie, INSERM U1185, Hôpital Bicêtre, Université Paris Sud, Assistance Publique Hôpitaux de Paris, 78 rue du Général Leclerc, 94270 Le Kremlin-Bicêtre, France
Corresponding author
Ac ce p
Dr Augustin Lecler
te
d
* These authors equally participate to this work
Service de Neuroradiologie diagnostique Fondation Rothschild
25 rue Manin, 75019 Paris
E-mail: [email protected] Phone : +33148036464
Fax : +33148036401
1 Page 1 of 31
Abstract Objective: To compare the residual cholesteatoma detection accuracy of diffusion-weighted (DW) and T1 delayed sequences for magnetic resonance at one year postoperative with second-look surgery in pediatric patients who have
ip t
undergone primary middle ear surgery for cholesteatoma. Methods: This was a prospective monocentric consecutive study conducted in a tertiary academic referral center.
cr
Children were referred for MR imaging (MRI) one year after surgery. A 1.5 T MRI was utilized, using nonechoplanar DW images and delayed gadolinium-enhanced T1-weighted images. Accuracy of magnetic resonance
us
imaging was assessed by two radiologists before surgery. Interobserver and intraobserver agreements were assessed using the test. Magnetic resonance imaging data were compared with surgery, which was considered as the gold
an
standard.
Results: Twenty-four consecutive unselected pediatric patients were included. Sensitivity, specificity, positive
M
predictive value, and negative predictive value for the first observer were of 40%, 86%, 67%, and 67%, respectively, and those for the second observer were 30%, 86%, 60%, and 63%, respectively. The only two cholesteatoma with a
te
d
size superior to 3 mm were diagnosed before surgery, but the majority of small cholesteatoma were not detected. Conclusions : MRI is a key examen to diagnosed the residual cholesteatoma but is limited by the size of the lesion under 3 mm. Delaying the realization of MRI during follow-up could increase sensitivity, thus avoiding
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
misdiagnosis as well as unnecessary second look surgery.
1 Page 2 of 31
Keywords Residual cholesteatoma
ip t
MRI DW imaging
us
cr
Pediatric
DW MRI: diffusion-weighted magnetic resonance imaging
M
EPI: echo planar imaging
te
PPV: positive predictive value
d
TSE: turbo spin echo CT: computed tomography
an
Abbreviations and acronyms
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
NPV: negative predictive value TR: time of repetition (ms) TE: time echo (ms)
2 Page 3 of 31
1.
INTRODUCTION
Cholesteatoma is a non-neoplastic but destructive cystic lesion of keratinizing stratified squamous epithelium, which can lead to labyrinthine, facial or cerebromeningeal complications. Surgery, mainly canal wall-up
ip t
tympanomastoidectomy, is therefore necessary to eradicate cholesteatoma of the middle ear. Besides cholesteatoma disease presents several major differences in children compared with adults. It is more aggressive in children. There
cr
is a higher residual rate after surgery: 30% in children vs 3%-15% in adults.[1] Otoscopic evaluation may be more
prevents an optimal detection of retrotympanic cholesteatomas.
us
challenging because of the frequent use of cartilaginous graft for tympanic membrane repair in children, which
Thus, the main objective of the follow-up after initial surgery is to rule out the presence of a residual cholesteatoma.
an
Because of the frequent use of cartilaginous tympanic membrane repair, “second-look” surgery is currently the gold standard to detect and treat residual disease in children. However, imaging plays a major role in the follow-up. CT
M
can be useful to detect a residual cholesteatoma; a lobulated or heterogeneous soft tissue mass associated with focal bony erosion is highly evocative. Unfortunately, CT becomes unreliable when the postoperative cavity is partially or
d
completely opacified, which is more frequent in children.[2, 3]
te
Conversely, magnetic resonance imaging (MRI) has progressively gained importance in the follow-up because of performing sequences, such as delayed gadolinium-enhanced T1-weighted images (WI) and diffusion-weighted
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
(DW) images. Diffusion-weighted magnetic resonance images (DW MRI) relies on the signal produced by the diffusion motion of water protons in biologic tissues. Stratification of keratin in residual cholesteatoma has high signal intensity, whereas other tissues that can be found in the middle ear after surgery show low or intermediate signal intensity on DW images.[4] Delayed gadolinium-enhanced T1 WI relies on the nonenhancement of residual cholesteatoma, whereas surrounding inflammatory fibrous tissue is enhanced. This allows a correct distinction between disease and reactive tissue.[5]
Despite the well documented follow-up with MRI in adults, there is too few studies evaluating MRI in pediatric follow-up, and the number of pediatric patients included is too small to provide reliable data.[6–9] Cholesteatoma residual disease presents major differences in children versus adults, preventing a reliable extrapolation from adult series.
3 Page 4 of 31
The aim of our study was to compare MRI performances to “second-look” surgery in the evaluation of postoperative residual cholesteatoma in children. Our goal was to use MRI follow-up as a reliable tool to delay or prevent second-
te
d
M
an
us
cr
ip t
look surgery.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
4 Page 5 of 31
2.
MATERIALS AND METHODS
2.1 PATIENTS This is a monocentric study in a tertiary academic referral center. All consecutive children with cholesteatoma
ip t
requiring second-look surgery and managed in our institution between June and October 2010 were retrospectively included. Decision for the second-look surgery was made by the surgeon on the basis of findings at first-stage
cr
surgery and on clinical follow-up.[10]
us
We collected information on clinical, radiological, surgical, and pathological findings. Children underwent an MR imaging before “second-look” surgery, which was programmed 12 months after the initial surgery.
an
This study was approved by the local ethical committee. Informed consent was obtained from all individual
M
participants included in the study. 2.2 IMAGING TECHNIQUE
d
MR imaging was performed with a 1.5 T MRI scanner (Achieva; Philips Medical Systems, Best, The Netherlands),
te
using spin echo DW (SE-DW) images and delayed contrast-enhanced T1-weighted images. A head coil was used. Coronal 2.5-mm-thick SE-DW sequences were obtained with the following parameters: ECG triggering; RT(repetition time)/ET (echo time), 3000/106; field of view (FOV), 230 × 230; matrix 176 × 100; b factors of 0 and
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
800 s/mm²; and acquisition time, 3 minutes 12 seconds. Axial 4.5-mm-thick dual turbo SE T2-weighted images were obtained with the following parameters RT/ET, 2293/110; FOV, 230 × 230; matrix, 256 × 168; acquisition time, 1 minute 45 seconds. Coronal 1.7-mm-thick turbo SE T2-weighted images were obtained with the following parameters: RT/ET, 3500/120; FOV, 230 × 230; matrix, 336 × 212; acquisition time, 4 minutes 26 seconds. Axial and coronal 2-mm-thick SE T1-weighted images were performed 30 to 45 minutes after intravenous contrast injection of 0.1 mmol per kilogram of body weight of gadoterate meglumine (Dotarem®; Guerbet, Roissy, France): RT/ET, 526/20; FOV, 180 × 180; matrix, 205 × 204; acquisition time, 3 minutes 36 seconds (Table A).
5 Page 6 of 31
Light sedation, consisting of 5 mg/kg of pentobarbital, was used in children above the age of five, and no sedation was used for younger children. No general anesthesia was used.
ip t
2.3 IMAGING AND SURGICAL EVALUATION
cr
Criteria for diagnosis of residual cholesteatoma was very high signal intensity in the middle ear or in the mastoid cavity on DWI sequence compared with brain tissue, corresponding to restricted diffusion on ADC maps.
us
Unenhanced intermediary to low signal mass of the soft tissue mass within the postoperative cavity on delayed contrast-enhanced T1-weighted images was considered as a residual cholesteatoma. A complete (homogeneous)
an
enhancement of the mass was conclusive for noncholesteatomatous postoperative tissue.[11] All images were analyzed by two radiologists blinded to any surgical and clinical information about patients.
M
Imaging data were classified into two groups: “residual cholesteatoma” and “no cholesteatoma.” Radiologists described precise localization and measured diameter of all detected residual cholesteatoma.
d
Second-look surgery was performed by one of 6 experienced surgeons (2 to 30 years of experience in cholesteatoma
te
surgery). Surgical results were classified as “residual cholesteatoma” and “no residual cholesteatoma.” Cholesteatoma localized in the middle ear was measured during surgery.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Histologic evaluation was systematically performed, and cholesteatoma were measured again.
2.4 STATISTICAL ANALYSIS
Nonweighted kappa statistics with the 95% confidence interval (CI) was used to test for interobserver agreement of the 3-step classification system using arbitrary interpretation by Landis and Koch (0, poor agreement; 0.00–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.80–1.00, almost perfect agreement).[12]
6 Page 7 of 31
Sensitivity, specificity, positive (PPV), and negative predictive values (NPV) were assessed. We decided to keep statistical values from the most experimented radiologist to discuss our results and to compare our statistics with those of other series.
ip t
Statistical analysis was performed using R software (http://www.r-project.org/). Exact confidence intervals were
te
d
M
an
us
cr
based on a binomial distribution.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
7 Page 8 of 31
3.
RESULTS
Twenty-four consecutive children (22 males, 2 females) with scheduled second-look surgery were treated during our study. Mean age at surgery was 10.5 years (range, 4-18 years) (Table B).
ip t
All patients underwent MRI according to our protocol. MRI to surgery delay was very short: less than a week for 15 patients (63%) and over 2 weeks for only 5 patients (21%). The mean delay was 12.8 days (Table B). Mean MRI
cr
acquisition time was 25 minutes (range, 15-38) with a median time of 23 minutes.
us
All 24 patients underwent surgery. Residual cholesteatoma was surgically confirmed in 10 patients (41.6%), whereas no residual was found in 14 patients. Thirteen residual cholesteatoma lesions were found in these 10
an
patients: 9 patients with only one lesion each, and one patient with 4 lesions (Table B).
Residual cholesteatoma localizations were as follows: epitympanum (n = 7) (Fig A), hypotympanum (n = 2),
M
mesotympanum (n = 2), and mastoid recess (n = 2) (Fig B; Table B).
Median size was 2 mm (range, 0-8 mm). Among residual cholesteatoma shapes, there were 11 pearls, 1 with a
te
d
ruptured bag and 1 with a completely empty bag. The pathological analysis confirmed theses data. First and second observer identified residual cholesteatoma in 6 and 5 patients, respectively. Among these results, there were 6 and 7 false negatives and 2 false positives for each, respectively (Fig C).
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
One residual with an empty bag, one ruptured cholesteatoma with low keratin content, and one cholesteatoma within the mastoid recess were missed by both readers (Fig D). All missed residual cholesteatoma were less than 3 mm. Epitympanic localizations were almost correctly detected by MRI. Other localizations (mastoid recess and mesotympanum) were missed by both readers. All residual cholesteatoma correctly identified by the two readers had a pearl shape, 2 were in the epitympanum and one in the hypotympanum. Their sizes were 2, 4, and 8 mm. Two residual cholesteatoma with a size superior to 3 mm were diagnosed by imaging before surgery.
8 Page 9 of 31
Sensitivity, specificity, PPV, and NPV of MRI were of 40%, 86%, 67%, and 67%, respectively, for the first observer and 30%, 86%, 60%, and 63%, respectively, for the second observer. Interobserver agreement was substantial (k = 0.647).
ip t
We extracted statistical data from the second observer, which had the longest experience reading middle ear MRI, to calculate the following data: positive likelihood ratio was 2.1, negative likelihood ratio was 0.8, and accuracy was
te
d
M
an
us
cr
0.63.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
9 Page 10 of 31
4.
DISCUSSION
The present study demonstrated that MRI in children may be useful in the detection of residual cholesteatoma above 3 mm before second-look surgery but lacks accuracy for small residuals. It is, according to our knowledge, the
ip t
largest homogeneous pediatric series published in the literature. Cholesteatoma disease presents several major differences in children compared with adults. Cholesteatoma is more
cr
aggressive in children, and there is a higher residual rate after surgery.[1] There is a clear distinction between
us
residual and recurrence of cholesteatoma.[11] All recurrences are diagnosed with otoscopy, whereas residuals are most of the time an imaging or perioperative diagnosis. Adult series often mix recurrence and residual
an
cholesteatoma, making comparison between adult and pediatric series more hazardous.[13] Otoscopic evaluation may be more challenging, often impossible, because of the frequent use of cartilaginous graft for tympanic membrane repair in children. That explains why otoscopic exam may be insufficient to confidently detect
M
retrotympanic cholesteatoma residual.[11] Moreover, children present more often than adults a completely filled middle ear, leading to a noncontributive CT. These semiology points are very important issue and explain why we
d
should not extrapolate data from adult series to children, which shows the absolute necessity to get specific data for
te
children in cholesteatoma follow-up. Yet, a huge majority of series evaluating cholesteatoma disease are mixed, including children and adults.[13][14] Dedicated pediatric series are often inhomogeneous, mixing residual, recurrences, congenital and acquired cholesteatoma without any distinction.[8, 9] Our study was based on a
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
homogeneous population of consecutive children before second-look surgery, with possible residual of acquired cholesteatoma only and with MRI performed just before second-look surgery. It provided strong and specific data, directly relevant in a current practice. Imaging to surgery interval was as short as possible, to have the better correlation between imaging and surgical findings. This interval is critical in children because cholesteatoma disease is reported to be more aggressive.[15] The present study demonstrated the feasibility of short delays between imaging and surgery. Decision of second-look surgery was made during the first surgery, according to validated predictive risk factors of residual cholesteatoma.[10, 16]This allowed to schedule early both second-look surgery and imaging. In the literature, MRI appears to be a very strong tool in detection of residual cholesteatoma in adults [4, 13, 14]. In small dedicated pediatric series as well, Rajan or Plouin found impressive results with 100% PPV, NPV, sensibility
10 Page 11 of 31
and specificity using non-EP DW MRI only or fused EP DW MR and CT.[6, 9] In our study, detection rate was much lower but we have also more patients included. We tried to identify the reasons by reviewing all cases with the surgeon and pathologist : first of all, residual cholesteatoma size was mainly small (median, 2 mm). Williams and
ip t
De Foer showed that delayed postgadolinium T1-enhanced sequences and non-EP DW MR sequences, respectively, could discriminate a 2-mm-diameter cholesteatoma mass.[5, 19] Yet, the majority of small cholesteatoma less than 3 mm are not detected, even in recent series[20] (Table C). Secondly, 2 over 13 residual were infiltrative or had an
cr
empty bag, which is hard to detect using MRI. It is now clearly identified that one of the major causes of false
us
negatives of the DW MRI in children and adults is related to the presence of dry, autoevacuating retraction pockets.[21] The autoevacuation eliminates keratin accumulation, thus removing the substrate that causes the
an
restricted diffusion signal on MRI.[22] In Ganaha series, there was only one conventional residual cholesteatoma for 3 adhesive-type cholesteatoma and 4 evacuated attic cholesteatoma, explaining a relatively low sensitivity of 69% with 18 false-negative cases. [20] Third, comparison with other pediatric series is difficult, because no strong data is
M
available for now in the litterature : no dedicated pediatric series included enough children to provide significant results, with only 18 patients for the largest series[7]. Too few residual cholesteatoma were found, with only 2
d
residual in some series[6], and residual size ranged from 3 mm to 26 mm with a mean size of 1cm in some others.[8,
te
9] (Table C) Therefore, excellent results found in literature should be interpreted with high caution. Imaging remains mandatory to perform a complete middle ear exploration because some areas are totally blind to
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
the otoscopic exam, like mastoid recess or epitympanum. In our study, epitympanic residuals were almost always correctly detected using MRI (Fig B). On the contrary, residual localized in the mastoid recess, mesotympanum, and hypotympanum were missed by both readers. Two parameters explained these results: on the one hand, DW MRI skull-base artifacts and air-bone interfaces are not completely avoided even with non-EP techniques, and on the other hand, residual cholesteatomas were particularly small in these precise locations in our study (the biggest residual in mastoid recess was 3 mm). Plouin recently used a fusion technique to combine DW MRI and CT in 10 children. This allowed a correct diagnosis and precise localization in all cases and provided precious preoperative information for the surgeon.[9] There are several limitations in our study: first, the number of included children is small, thus limiting our statistical analysis. We had only 2 residual cholesteatomas above 3 mm, correctly detected by both readers, but this number is
11 Page 12 of 31
too small to conclude that all residuals above 3 mm would have been found. Second, we did not compare DW MRI versus delayed gadolinium-enhanced T1-weighted sequences, thus preventing us to conclude that DW MRI are sufficient to detect residual. Finally, we tried to prevent selection bias by including consecutive patients in a
ip t
prospective way, but we had too many “small” residuals under 3 mm, leading to an artificially low sensitivity and specificity.
cr
We guess that MRI control should be performed later than 12 months after primary operation to avoid misdiagnosis
performed in case of high presumption of residual cholesteatoma.
us
of small residual cholesteatoma, without unwisely delaying adequate treatment. Second-look surgery is still
an
Further studies with a larger number of children should be realized to provide strong data and more accurate
te
d
M
statistical analysis.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
12 Page 13 of 31
5.
CONCLUSION
MRI may be useful in the detection of residual cholesteatoma above 3 mm before second-look surgery in children but lacks accuracy for small residuals. Delaying the realization of MRI during follow-up could increase sensitivity,
te
d
M
an
us
cr
ip t
thus avoiding misdiagnosis as well as unnecessary second look surgery.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
13 Page 14 of 31
Compliance with Ethical Standards Conflicts of interest :
ip t
None Ethical approval:
cr
All procedures performed in studies involving human participants were in accordance with the ethical standards of
us
the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
an
Informed consent:
te
d
M
Informed consent was obtained from all individual participants included in the study.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
14 Page 15 of 31
REFERENCES 1. Darrouzet V, Duclos JY, Portmann D, Bebear JP (2000) Preference for the closed technique in the management of cholesteatoma of the middle ear in children: a retrospective study of 215 consecutive patients treated over 10 years. Am J Otol 21:474–481.
ip t
2. Williams MT, Ayache D (2004) Imaging of the postoperative middle ear. Eur Radiol 14:482–495. doi: 10.1007/s00330-003-2198-8
cr
3. Lemmerling MM, De Foer B, VandeVyver V, et al. (2008) Imaging of the opacified middle ear. Eur J Radiol 66:363–371. doi: 10.1016/j.ejrad.2008.01.020
us
4. Dubrulle F, Souillard R, Chechin D, et al. (2006) Diffusion-weighted MR imaging sequence in the detection of postoperative recurrent cholesteatoma. Radiology 238:604–610. doi: 10.1148/radiol.2381041649 5. Williams MT, Ayache D, Alberti C, et al. (2003) Detection of postoperative residual cholesteatoma with delayed contrast-enhanced MR imaging: initial findings. Eur Radiol 13:169–174. doi: 10.1007/s00330-002-1423-1
an
6. Rajan GP, Ambett R, Wun L, et al. (2010) Preliminary outcomes of cholesteatoma screening in children using non-echo-planar diffusion-weighted magnetic resonance imaging. Int J Pediatr Otorhinolaryngol 74:297– 301. doi: 10.1016/j.ijporl.2009.12.011
M
7. Geoffray A, Guesmi M, Nebbia JF, et al. (2013) MRI for the diagnosis of recurrent middle ear cholesteatoma in children--can we optimize the technique? Preliminary study. Pediatr Radiol 43:464–473. doi: 10.1007/s00247-012-2502-3
te
d
8. Plouin-Gaudon I, Bossard D, Fuchsmann C, et al. (2010) Diffusion-weighted MR imaging for evaluation of pediatric recurrent cholesteatomas. Int J Pediatr Otorhinolaryngol 74:22–26. doi: 10.1016/j.ijporl.2009.09.035 9. Plouin-Gaudon I, Bossard D, Ayari-Khalfallah S, Froehlich P (2010) Fusion of MRIs and CT scans for surgical treatment of cholesteatoma of the middle ear in children. Arch Otolaryngol Head Neck Surg 136:878–883. doi: 10.1001/archoto.2010.151
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
10. Roger G, Denoyelle F, Chauvin P, et al. (1997) Predictive risk factors of residual cholesteatoma in children: a study of 256 cases. Am J Otol 18:550–558. 11. Nevoux J, Lenoir M, Roger G, et al. (2010) Childhood cholesteatoma. Eur Ann Otorhinolaryngol Head Neck Dis 127:143–150. doi: 10.1016/j.anorl.2010.07.001 12. Landis JR, Koch GG (1977) An application of hierarchical kappa-type statistics in the assessment of majority agreement among multiple observers. Biometrics 33:363–374. 13. De Foer B, Vercruysse J-P, Bernaerts A, et al. (2010) Middle ear cholesteatoma: non-echo-planar diffusionweighted MR imaging versus delayed gadolinium-enhanced T1-weighted MR imaging--value in detection. Radiology 255:866–872. doi: 10.1148/radiol.10091140 14. Vercruysse J-P, De Foer B, Pouillon M, et al. (2006) The value of diffusion-weighted MR imaging in the diagnosis of primary acquired and residual cholesteatoma: a surgical verified study of 100 patients. Eur Radiol 16:1461–1467. doi: 10.1007/s00330-006-0160-2 15. Stangerup SE, Drozdziewicz D, Tos M (1999) Cholesteatoma in children, predictors and calculation of recurrence rates. Int J Pediatr Otorhinolaryngol 49 Suppl 1:S69–73.
15 Page 16 of 31
16. Lazard DS, Roger G, Denoyelle F, et al. (2007) Congenital cholesteatoma: risk factors for residual disease and retraction pockets--a report on 117 cases. Laryngoscope 117:634–637. doi: 10.1097/mlg.0b013e318030ac8c
ip t
17. Dhepnorrarat RC, Wood B, Rajan GP (2009) Postoperative non-echo-planar diffusion-weighted magnetic resonance imaging changes after cholesteatoma surgery: implications for cholesteatoma screening. Otol Neurotol 30:54–58. doi: 10.1097/MAO.0b013e31818edf4a
cr
18. Jindal M, Riskalla A, Jiang D, et al. (2011) A systematic review of diffusion-weighted magnetic resonance imaging in the assessment of postoperative cholesteatoma. Otol Neurotol 32:1243–1249. doi: 10.1097/MAO.0b013e31822e938d
us
19. De Foer B, Vercruysse J-P, Pilet B, et al. (2006) Single-shot, turbo spin-echo, diffusion-weighted imaging versus spin-echo-planar, diffusion-weighted imaging in the detection of acquired middle ear cholesteatoma. AJNR Am J Neuroradiol 27:1480–1482. 20. Ganaha A, Outa S, Kyuuna A, et al. (2011) Efficacy of diffusion-weighted magnetic resonance imaging in the diagnosis of middle ear cholesteatoma. Auris Nasus Larynx 38:329–334. doi: 10.1016/j.anl.2010.11.004
an
21. Fitzek C, Mewes T, Fitzek S, et al. (2002) Diffusion-weighted MRI of cholesteatomas of the petrous bone. J Magn Reson Imaging 15:636–641. doi: 10.1002/jmri.10118
M
22. De Foer B, Vercruysse J-P, Bernaerts A, et al. (2007) The value of single-shot turbo spin-echo diffusionweighted MR imaging in the detection of middle ear cholesteatoma. Neuroradiology 49:841–848. doi: 10.1007/s00234-007-0268-3
d
23. Lehmann P, Saliou G, Brochart C, et al. (2009) 3T MR imaging of postoperative recurrent middle ear cholesteatomas: value of periodically rotated overlapping parallel lines with enhanced reconstruction diffusion-weighted MR imaging. AJNR Am J Neuroradiol 30:423–427. doi: 10.3174/ajnr.A1352
te
24. De Foer B, Vercruysse J-P, Bernaerts A, et al. (2008) Detection of postoperative residual cholesteatoma with non-echo-planar diffusion-weighted magnetic resonance imaging. Otol Neurotol 29:513–517. doi: 10.1097/MAO.0b013e31816c7c3b
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
25. Venail F, Bonafe A, Poirrier V, et al. (2008) Comparison of echo-planar diffusion-weighted imaging and delayed postcontrast T1-weighted MR imaging for the detection of residual cholesteatoma. AJNR Am J Neuroradiol 29:1363–1368. doi: 10.3174/ajnr.A1100 26. Stasolla A, Magliulo G, Parrotto D, et al. (2004) Detection of postoperative relapsing/residual cholesteatomas with diffusion-weighted echo-planar magnetic resonance imaging. Otol Neurotol 25:879–884. 27. Aikele P, Kittner T, Offergeld C, et al. (2003) Diffusion-weighted MR imaging of cholesteatoma in pediatric and adult patients who have undergone middle ear surgery. AJR Am J Roentgenol 181:261–265. doi: 10.2214/ajr.181.1.1810261
16 Page 17 of 31
FIGURES Fig. A: Four-mm residual cholesteatoma of the epitympanum
ip t
Fig. A1: Coronal spin echo diffusion-weighted (SE-DW) image: hypersignal (white arrow) Fig. A2 : Coronal delayed contrast-enhanced T1-weighted image: lack of enhancement in an area of fibrosis (white
cr
arrow)
arrow)
an
Fig. B: Three-mm residual cholesteatoma of the mastoid recess
us
Fig. A3 : Axial delayed contrast-enhanced T1-weighted image: lack of enhancement in an area of fibrosis (white
Fig. B1: Coronal spin echo diffusion weighted (SE-DW) image: hypersignal (white arrow)
M
Fig. B2 : Coronal delayed contrast-enhanced T1-weighted image: lack of enhancement (white arrow)
d
Fig. B3 : Axial delayed contrast-enhanced T1-weighted image: lack of enhancement (white arrow)
hypersignal (white arrow)
te
Fig.C: False positive for both readers. Coronal spin echo diffusion-weighted image: left epitympanum nodular
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Fig.D: False negative for both readers. Coronal spin echo diffusion-weighted image. At surgery, residual disease was a ruptured bag in the left mesotympanum (white arrow)
17 Page 18 of 31
Table(s)
-Table A: MRI parameters SE T2
DWI B = 0
Delayed contrast
And B = 800 s/mm²
enhanced T1 WI 180
230
230
230
Slice thickness (mm)
4.5
1.7
2.5
TR (ms)
2293
3500
3000
TE (ms)
110
120
106
Number of excitations
2
6
1
Number
12
14
Acquisition plane
Axial
Coronal
Acquisition time (min)
1.45
4.26
Matrix
256 × 168
336 × 212
526
Coronal
Axial and coronal
3.12
3.36
176 × 100
205 × 204
9
11
Ac ce p
te
d
M
slice/partition
20 2
an
of
2
us
FOV (mm)
ip t
Dual SE T2
cr
Sequence
1 Page 19 of 31
Table(s)
-Table B: Patients and residual cholesteatoma characteristics Second
look to
size (mm)
localization
= no residual
reader (0 =
MRI
detected; 1 =
no residual
(days)
residual
detected; 1 =
detected)
residual
11
M
1
2
12
M
29
3
18
M
1
4
8
M
1
5
8
M
2
6
8
M
1
7
16
M
1
8
7
M
1
9
4
M
9 9 10 11
9
Ac ce p
9
2
te
1
ip t
First reader (0
cr
(years)
Residual
us
Gender
Residual
an
surgery
Second-
Hypotympanum
M
Age at
d
Patients
detected)
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
Epitympanum
1
0
2
Epitympanum
1
0
2
Epitympanum
1
0
2
Mesotympanum
0
0
14
M
7
4
Epitympanum
1
1
5
M
13
0 (empty
Mastoid recess
0
0
0
0
0
0
12
10
M
14
13
6
M
15
bag)
1 (ruptured
Mesotympanum
bag) 14
9
M
66
0
0
15
13
M
1
0
0
1 Page 20 of 31
15
M
87
8
Epitympanum
1
1
17
6
F
7
2
Hypotympanum
0
0
18
16
M
2
0
0
19
16
M
21
1
1
20
8
F
9
2
Epitympanum
21
10
M
14
3
Mastoid recess
22
13
M
4
23
9
M
1
24
10
M
1
10.5
M
12.8
0
0
1
cr
0
0
0
1
1
us
0
Epitympanum
te
d
M
an
2
0
Ac ce p
mean
5
Ac ce p
Jindal 2011[18]
De Foer 2010[13]
63
83/87
96/56
NA
Dhepnorrarat 2009[17]
22
100/100
100/100
3
Lehmann 2009[23]
35
90-100/10
100/89-100
3
De Foer 2008[24]
19
100/96
3
Venail 2008[25]
31
80/50
5
93/100
2
60/72 100/88 (if >5 mm)
De Foer 2007[22]
21
Vercruysse 2006[14]
100 (31 children)
12.5 /100
100/84
5
Dubrulle 2006[4]
24
100/91
93/100
5-24
Stasolla 2004[26]
18
86/100
100/92
5
1 Page 22 of 31
17
77/100
75/100
5
Ac ce p
te
d
M
an
us
cr
ip t
Aikele 2003[27]
2 Page 23 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 24 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 25 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 26 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 27 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 28 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 29 of 31
Ac
ce
pt
ed
M
an
us
cr
i
Figure(s)
Page 30 of 31
Ac ce p
te
d
M
an
us
cr
ip t
Figure(s)
Page 31 of 31
