- Email: [email protected]
Childhood cholesteatoma
European Annals of Otorhinolaryngology, Head and Neck diseases (2010) 127, 143—150
REVIEW ARTICLE
Childhood cholesteatoma J. Nevoux a,∗, M. Lenoir b, G. Roger a, F. Denoyelle a, H. Ducou Le Pointe b, E.-N. Garabédian a a
Inserm U587, service d’oto-rhinolaryngologie et de chirurgie cervicofaciale, hôpital d’enfants Armand-Trousseau, AP—HP, UMPC, université Paris 6, 26-28, avenue du Dr Arnold-Netter, 75571 Paris cedex 12, France b Service de radiologie, hôpital d’enfants Armand-Trousseau, 26-28, avenue du Dr Arnold-Netter, 75571 Paris cedex 12, France Available online 11 August 2010
KEYWORDS Acquired cholesteatoma; Congenital cholesteatoma; Children
Summary Although cholesteatoma was first described in 1683, its etiopathogeny remains unexplained. In children, there are two forms: acquired cholesteatoma, resembling the adult form, and congenital cholesteatoma. The acquired form has become less frequent in recent years, thanks to progress in the treatment of childhood otitic pathology. Diagnosis of congenital cholesteatoma, on the contrary, is increasing, due to improvements in information to health care professionals and in diagnostic tools. Clinical and histological evidence points to greater aggressiveness in childhood forms, although this difference cannot, at present, be precisely explained. Diagnosis is clinical, but CT and MR imaging is indispensable for preoperative assessment and postoperative follow-up. New delayed gadolinium-enhanced T1-weighted and diffusion-weighted MRI sequences have recently been developed and provide more precise radiological diagnosis. Treatment is surgical; alternatives, notably by laser, have proved unsuccessful. Complications concern involvement of neighbouring structures, and are mainly infectious; some can be life-threatening, and should be systematically screened. © 2010 Published by Elsevier Masson SAS.
Introduction Cholesteatoma is a severe middle-ear pathology affecting both adults and children. It would seem to have been first described by De Verney in France in 1683, as what he called ‘‘steatoma’’ [1]. In 1829, the French anatomopathologist Cruveilhier described it as a pearl tumor of the temporal bone [2]. The term ‘‘cholesteatoma’’, introduced in 1838 by the German physiologist Johannes Müller [3], is now well established, but in fact, etymologically faulty in
∗
Corresponding author. Tel.: +33144736114; fax: +33144736108. E-mail address: [email protected] (J. Nevoux).
as much as this benign tumor (‘‘. . .oma’’) contains neither cholesterol (‘‘chole’’) nor fat (‘‘steat’’). In 1855, Virchow categorized cholesteatoma under epidermoid carcinoma and atheroma, but only in 1861 did Von Troeltsch consider its epidermal origin. In the studies by Gruber, Wendt and Rokitansky (1855—1888), the physiopathology of cholesteatoma was described as a malpighian metaplasia of the middle-ear mucosa in response to chronic inflammation. At the end of the 19th century, Bezold and Habermann overturned this theory, showing that cholesteatoma was caused by migration of external auditory canal skin to the middle ear, induced by chronic inflammation [4]. At the present time, despite numerous investigations, the physiopathology of cholesteatoma has still not been elucidated.
1879-7296/$ – see front matter © 2010 Published by Elsevier Masson SAS. doi:10.1016/j.anorl.2010.07.001
144 In children, there are two different types of cholesteatoma: acquired, which also affects adults, and congenital, which is specific to childhood. The most recent progress has been in imaging, with the introduction of MRI in the early 2000s, followed by the recent development of novel sequences. Treatment, on the other hand, has not greatly progressed and remains mainly surgical.
J. Nevoux et al. accepted theory. Both authors demonstrated the presence of an epidermoid residue in the anterosuperior quadrant of the middle ear in the human fetus, on the lateral wall of the Eustachian tube, near the annulus. Failure to reabsorb the epidermoid residue normally found in embryos between the 10th and 33rd week of gestation has been observed [10,11].
Acquired cholesteatoma
Epidemiology Incidence is hard to determine, especially for congenital cholesteatoma. Recent figures testify to a fall in the incidence of both adult and childhood acquired cholesteatoma: in 1925, the rate of cholesteatoma in the under-16’s was one in three; it is now much less frequent, but incidence in the early 2000s was still three per 100,000 (compared to nine per 100,000 in adults) [5,6]. The incidence of congenital cholesteatoma, on the other hand, seems to be on the rise, at 0.12 per 100,000 children [7]. Some 1 to 3% of cases of childhood cholesteatoma, and 1 to 5% of cholesteatoma as a whole, are congenital [8]. The relative increase is due to the fall in the number of cases of acquired cholesteatoma and an increase in the diagnosis of congenital cholesteatoma. The latter increase is thanks to the pathology being better known by ENT specialists and pediatricians; diagnostic tools, in particular microscopy, have improved, as has otologic follow-up of children. Prevalence varies between populations: it is highest among Caucasians, followed by Africans, and very low in Asians. Ratnesar reported extremely low prevalence of cholesteatoma in Inuits [9].
Physiopathology Cholesteatoma is a non-neoplastic but destructive cystic lesion containing keratin layers from a keratinized malpighian epithelium surrounded by a matrix of epithelium that rests on a perimatrix formed of a conjunctive stroma of varying thickness. It generally occurs in the middle ear, but other locations in or around the petrosal bone are also possible [6]. The pathologic substrate of cholesteatoma is a keratinized stratified malpighian epithelium, the middleear origin of which is, however, controversial. Various etiopathogenic theories have been proposed to account for the two forms of cholesteatoma. Despite recent research, the underlying mechanisms remain unknown.
Congenital cholesteatoma Various theories have been put forward to explain the formation of congenital cholesteatoma. In 1854, Von Remak postulated that the origin of dermoid and other related tumors (such as cholesteatoma) was a skin follicle trapped during the early stages of embryogenesis [6]. Subsequently, many theories were formulated, culminating in that of epidermoid formation, close to Von Remak’s postulate, hypothesized by Teed in 1936 and then by Michaels in 1986, and which remains the most widely
During the course of the 20th century, four physiopathological theories of the formation of acquired or secondary cholesteatoma were put forward. In 1864, Von Tröltsch first proposed a squamous metaplasia theory; in 1890, Habermann and Bezold proposed the theory of migration or invasion; in 1925, Lange first reported the phenomenon underlying the basal hyperplasia theory; and finally, in 1933, Wittmaack formulated the retraction pocket or invagination theory, which is now the most widely accepted [6]. A1though there is no firmly established theory, primary and secondary acquired cholesteatoma are classically distinguished: the former evolves from a retraction pocket, while the latter is due to epithelial migration through a ruptured ear-drum or to epithelium implanted in the middle ear. Such implantation may be iatrogenic, occurring during ear surgery, or due to epithelium left in the tympanic cavity after healing of a blast injury, or else arising posttraumatically at a petrosal bone fracture site [4].
Histopathologic differences between adult and childhood cholesteatoma Childhood cholesteatoma has long been known to be more aggressive than the adult form, with poorer clinical prognosis. Several studies have sought a physiopathological explanation for this finding [12—16]. According to Quaranta, the explanation lies in the cholesteatoma perimatrix. Following up this hypothesis, he found the number of mononuclear inflammation elements (plasmocytes, lymphocytes, macrophages, granulocytes and giant cells) in the perimatrix to be greater in children than in adults [17]. Bujia’s group analyzed the expression of MIB1 (a monoclonal antibody marker of cell proliferation) in child and adult cholesteatoma and within the external auditory canal skin; the proliferation index was normal in the latter, elevated in the former, and relatively higher in children [14]. Recent studies confirmed that the perimatrix is thicker and the epithelium better contoured in child cholesteatoma [18]. Other studies, in contrast, found no significant histopathological differences between adults and children in the matrix of acquired cholesteatoma that could account for the difference in clinical aggressiveness [19].
Progress in histology There have been many studies of the cellular basis of the specificities of childhood cholesteatoma, its local aggressiveness and growth potential. However, they have yet to provide an account that could enable medical treatment to
Childhood cholesteatoma be developed. There are, however, phenomena of hyperproliferation specific to middle-ear skin. Gene expression studies confirmed growth-factor and signalling molecule upregulation. Likewise, protein expression studies demonstrated elevated rates of the molecules involved in growth, proliferation, antiapoptosis and bone resorption. Recent breakthroughs have concerned angiogenesis and metalloproteinases. These markers seem to explain the difference in aggressiveness between childhood and adult forms. In 2009, a Brazilian group demonstrated significantly elevated inflammation and metalloproteinase production in childhood cholesteatoma; perimatrix angiogenesis, in terms of blood-vessel count, was elevated in children. Metalloproteinases are proteolytic enzymes associated with pathologic extracellular matrix resorption in chronic inflammation. Inflammatory processes, such as infected cholesteatoma, stimulate production of these enzymes. Dornelles’s group demonstrated cytoplasmic hyperexpression of MMP2 (matrix metalloproteinase) and nuclear hyperexpression of MMP2 and MMP9 in pediatric cholesteatoma. There also appears to be a correlation between expression level and age: the younger the child, the higher the cytoplasmic expression of MMP2 in the cholesteatoma. Metalloproteinases are associated with the capacity to penetrate and invade neighboring tissue. Their other role is to regulate bone homeostasis. They are elevated in inflammatory osteolytic pathology. The greater aggressiveness and osteolytic capacity of childhood cholesteatoma would thus seem to derive from hyperexpression of perimatricial metalloproteinases [18]. The most recent studies have investigated the possible role of protein transcription regulators, micro-RNAs. These RNA fragments of 22 to 24 nucleotide length, strongly conserved over evolution and non-coding, inhibit protein transcription. They play an essential role in cell regulation pathways and are strongly implicated in oncogenesis. In 2008, Cioffi’s group reported significant upregulation of micro-RNA-21 (hsa-miR-21) in cholesteatoma as compared to normal skin in humans, and proposed a growth and proliferation model of cholesteatoma [20]. At the present time, the histologic and cellular phenomena underlying the development of cholesteatoma and its differential aggressiveness between childhood and adult forms remain unknown.
Clinical aspect Although cholesteatoma occurs mainly in the middle ear, other locations are also possible. Classically, extra- and intradural locations are distinguished. Extradural involvement mainly concerns the middle ear, but may lie in any part of the petrosal bone: mastoid, external auditory canal or petrous apex. Intradural involvement, better known as epidermoid cyst, has been reported at various locations, the most frequent being the pontocerebellar angle. Whatever the location, complex surgery is required, as there is at present no medical treatment [4].
Congenital cholesteatoma In 1965, Derlacki and Clemis defined congenital cholesteatoma as an embryologic residue of epithelial
145
Figure 1
Congenital left-ear cholesteatoma.
tissue behind a normal tympanic membrane, in the absence of any history of infection or ear surgery [21]. They formulated a series of diagnostic criteria: normal tympanic membrane, retrotympanic white mass, and no history of acute otitis media, otorrhea, eardrum rupture or otologic surgery (Fig. 1). In 1986, these criteria were modified by Levenson, who suggested that history of acute otitis media does not necessarily exclude diagnosis of congenital cholesteatoma, given its high frequency in young children [22]. Mean age at diagnosis is 4—5 years. Systematic recurrence of acute otitis media or otorrhea in the same ear is an alarm signal. According to Valvassori, congenital cholesteatoma can be found at four locations in the region of the petrosal bone: middle ear (middle ear cavity and mastoid), petrous apex, pontocerebellar angle, and jugular foramen [23]. It is a rare congenital lesion, making up 1% of intracranial tumors. Forty to 50% of cases are located in the pontocerebellar angle.
Acquired cholesteatoma Various classifications have been drawn up for acquired cholesteatoma, but there is not one, which is universally employed. Tos’s classification, however, basing indications on the location of the pathological process, seems to be the most frequently used. It distinguishes three types according to location: attic cholesteatoma, tympani sinus cholesteatoma with a posterior subligamentary starting point, and pars tensa cholesteatoma [24]. In children, 80% of retraction pockets and cholesteatomas develop from or depend on the pars tensa (especially the posterior subligamentary pars tensa) (Fig. 2), while in adults, they mainly develop in the pars flaccid (or Shrapnell’s membrane) (Fig. 3) [25]. To cholesteatomas must be added unstable retraction pockets, which are equivalent in as much as they have the same predictive factors as cholesteatoma. The clinical features of these pockets are their marginal location, their fixation (outside of the ossicular chain), the impossibility of complete control, and associated otorrhea and/or squamata [25].
146
J. Nevoux et al. Residual lesion and/or recurrence require long-term clinical and radiological follow-up over at least 8 years [25]. Recurrence and residual lesion rates are higher in children: for recurrence, the rate is 30%, versus 3 to 15% in adults [29,30] and, for residual lesion, about 40% versus 15 to 20% in adults [31].
Imaging assessment Computed tomography (CT)
Figure 2
Acquired childhood cholesteatoma.
Acquired cholesteatoma is more aggressive in children, for a number of reasons: it is often more extensive at diagnosis; ossicular chain status is often poorer, due to more extensive ossicle lysis; and the rates of recurrence and residual lesion are higher. This greater aggressiveness in children has been explained in terms of childhood specificities in the cholesteatoma matrix and/or persisting predisposing factors such as impaired aeration of the ear [26]. Predictive factors for recurrence are large size and extension, age < 8 years, associated otorrhea, ossicular chain status, and impaired preoperative aeration of the ear [27]. Predictive factors for residual lesion are initial posterior mid-tympanic invasion, ossicular chain interruption, lack of expertise on the part of the surgeon, and incomplete resection. The greater the association of these factors, the higher the risk of residual lesion. In case of one risk factor, secondlook surgery should systematically be performed within 12 to 18 months. In case of two or more risk factors, second-look surgery should be performed within 9 months [25]. Although cholesteatoma is more aggressive in children, the number of local (labyrinthine fistula or facial palsy) or regional (intracranial) complications is lower [28].
Temporal bone CT assesses cholesteatoma extension, exploring for complications and detecting anatomic variants so as to prepare for any difficulties in surgery. Diagnosis is normally clinical and otoscopic. Technically, CT examination should be high-resolution, non-contrast, with bone filter, with overlapping slices of subor millimetric thickness in the axial (orbitomeatal plane: i.e., the plane of the lateral semi-circular canal) and coronal planes, orthogonally through the entire pars petrosa. Multislice CT further requires submillimetric volume acquisition centered on the pars petrosa, with the patient installed so as to protect the orbits from radiation, with axial and coronal reconstruction in the plane of the lateral semi-circular canal.[32,33] The negative predictive value (NPV) of CT is excellent [34]. Cholesterol granuloma, however, which is frequently found in association with cholesteatoma, cannot be differentiated on CT; MRI enables the distinction. The main CT criteria for cholesteatoma are: nodular soft tissue mass within the tympano-mastoid cavities, with associated osteolysis (wall of the epitympanic recess, and ossicles, especially the long apophysis of the incus, tegmen tympani and bone shell of the lateral semicircular canal) [35]. In case of doubt on first examination, repeat scan performed a few months later detects growth of the residual cholesteatoma and thus provides better visualization. The main limitation of CT is fluid effusion in the middle ear, hindering diagnosis of a residual lesion. MRI can then characterize soft tissue and differentiate between cholesteatoma and effusion. One essential issue in CT is radioprotection, especially in children presenting with chronic otitis media, requiring numerous scans over their lifetime. This mainly concerns the crystalline lens, and the risk of radio-induced cancer. The crystalline lens is conserved with multislice CT, which allows volume acquisition of the pars petrosa in a slice plane excluding the orbit, with reconstruction in the classic slice plane. Irradiation is standardized in petrous bone CT, at about 8 years’ natural irradiation, whereas radioprotection standards give a risk threshold of more than 100 years for natural radiation-induced cancer. This point needs to be taken into account in the radiological follow-up of childhood cholesteatoma.
MRI Figure 3
Acquired adult cholesteatoma.
MRI has more recently been introduced in cholesteatoma assessment, and is indicated for initial extension assess-
Childhood cholesteatoma ment, in case of contact with the meninges in particular, and for postoperative follow-up in case of doubtful CT images. MRI characteristics in cholesteatoma are: isosignal with or without peripheral enhancement on post-gadolinium T1-weighted sequences, and isosignal on T2-weighted sequences. Associating delayed (30—50 minutes, or ideally 45 minutes according to certain authors) gadoliniumenhanced T1-weighted sequences improved diagnosis of residual cholesteatoma. The cholesteatoma never shows enhancement, appearing in hyposignal with respect to surrounding tissue. In the last few years, a new type of sequence based on the signal produced by displaced water molecules, known as diffusion-weighted sequence, has provided further information for residual lesion screening [34]. Cholesteatoma shows as hypersignal on diffusion-weighted sequences. The b-factor is usually b800 in children and b1,000 in adults. The MRI resolution threshold is 3 mm on standard sequences, but just under 5 mm on diffusion-weighted sequences. The other limitations of this new sequence are magnetic susceptibility artifacts and a T2 effect. These artifacts are related to the numerous air-bone interfaces due to cranial base anatomy, inducing considerable distortion. Mean sensitivity and specificity in classic diffusion-weighted sequences are respectively 81.6% (range, 77 to 86%) and 100%, and positive predictive value (PPV) and NPV respectively 100 and 75% [36]. To get around these limitations, an improved technique has been developed: turbo-spin echo (TSE) or non-echo-planar (non-EPI) instead of echo-planar (EPI) diffusion-weighted imaging [37]. This technique improves spatial resolution, reducing the detection threshold to 3 mm. PPV, NPV, sensitivity and specificity seem to be 100%. Moreover, acquisition time is shorter, at about 3—4 minutes, compared to nearly 5—8 minutes with the classical technique [34]. Performing both CT and MRI is essential, as the aim is to locate a tissue process within the temporal bone: CT provides better spatial location, and MRI better tissue characterization. The power of most MRI machines used in everyday practice is 1.5 Tesla. With the development and purchasing of new models, MRI at 3T or greater is becoming more common and available. The usual advantage is shorter acquisition time; but due to artifacts specific to this anatomic region, diffusion-weighted sequence acquisition is paradoxically longer, incurring a risk of motion artifacts, especially in children making interpretation tricky even for experienced radiologists [38,39]. The two main problems with diffusion-weighted sequences, whether classic or echoplanar, are the lack of spatial resolution and the amount of artifacts. The latter mainly concern bone/air and tissue/air interfaces. The problems are partially resolved using TSE sequences, although spatial resolution remains poor. Examination duration and its requirements (sedation and contention) are also important points. The time needed for CT scan of the petrous bone varies greatly depending on the equipment. In our experience of multislice CT, standard millimetric acquisition takes approximately 30 secondes, and submillimetric acquisition 70 secondes (such better resolution being needed notably to visualize the stapes). MRI takes
147 approximately 30 minutes. In our experience, sedation is employed for CT in children under 3 years of age, and for MRI in under-4-year-olds; older children can tolerate the examination if they and their parents have been properly informed of the importance of the examination and how it takes place. In case of difficulties of communication (such as associated neuropsychiatric pathology or profound deafness), general anesthesia is required (Table 1).
Treatment Treatment remains surgical, although other options have been tested. The prime objective is to eradicate the disease so as to restore a healthy aerated ear. No less important, the second objective is to restore or improve the ossicular chain. Since the first description of mastoidectomy by Riolanus in 1649, numerous procedures have been described. The main distinction, however, is whether or not to conserve the posterior wall of the external auditory canal, distinguishing canal wall-up tympanoplasty (with conservation) from canal wall-down tympanoplasty (non-conservation). Middle-ear surgery is hampered in children by obstruction of the Eustachian tube and chronic inflammatory hyperplasic mucosa, increasing the risk of incomplete resection and residual lesion. Fitting ventilation tubes ahead of cholesteatoma resection improves surgical preparation and postoperative course [40]. Cartilage should be used for eardrum reinforcement in children, whereas the temporal aponeurosis is preferred in adults. Posterior reinforcement is essential, given the preponderance of recurrence and retraction in this area of the tympanum. The problems encountered with cartilage are failure to control retrotympanic liquid effusion, and possible residual lesion [41,42]. Attempts at laser treatment, mainly to reduce the rate of residual lesion in second-stage surgery, have not been greatly followed up. In 2005, Hamilton reported a significant reduction in residual lesions, using the laser to control exeresis. He applied it specifically for ossicle locations and areas inaccessible to classic surgery [43].
Complications Complications are due to involvement of neighboring structures and to cholesteatoma location, and are generally of infectious origin. Classically, a distinction is drawn between intra- and extracranial complications, the latter being further divided into intra- and extratemporal complications [44].
Extratemporal complications Subperiosteal abscess Subperiosteal abscess is the most common extratemporal complication. It is caused by spread of infection from the mastoid towards the periosteal space by erosion of the mastoid cortex or, more rarely, by vascular propagation secondary to mastoid emissary vein thrombosis. It is most frequent in young children. The clinical presentation is of
148
J. Nevoux et al.
Low signal
Low signal
Low signal
Low signal
High signal
Low signal
3 mm
Low signal
Low signal
Low signal
Low signal
High signal
Low signal
5 mm 3 mm
No enhancement
3 mm
Intermediate signal
High signal Low signal
2.5 mm
Fluid
Resolution size
Intermediate signal
2.5 mm
No enhancement
No enhancement
Intermediate signal Intermediate signal
Inflammatory tissue Hyperplasic mucosa Cholesteatoma
No enhancement
Enhancement
Intermediate signal Intermediate signal Scar tissue fibrosis
Enhancement
No enhancement
Intermediate signal High signal Cholesterol granuloma
No enhancement No enhancement
Intermediate signal Intermediate signal Granulation tissue
Delayed post-gadolinium T1-weighted 30—50 minutes Enhancement Early post-gadolinium T1 weighted 10—15 minutes Enhancement T2-weighted T1-weighted
Signal according to MRI sequence and middle-ear process. Table 1
Low enhancement
TSE diffusion-weighted EPI diffusion-weighted
classical mastoiditis. Diagnosis is confirmed by CT; treatment is basically surgical, associating abscess drainage and cholesteatoma exeresis in a single step. Bezold’s abscess Bezold’s abscess is a cervical abscess, physiopathologically identical to a subperiosteal abscess, the only difference being a bone erosion area at the mastoid apex, which explains the extension to the neck below the sternocleidomastoid muscle. It is more frequent in older children and adults, due to the late pneumatization of the mastoid apex. Like subperiosteal abscesses, it is more frequent in case of acute otitis media but may also complicate cholesteatoma. Diagnosis and treatment are as in subperiosteal abscess, with one difference: the cervical incision.
Intra-temporal complications Labyrinthine fistula is also very frequent, in 7% of cholesteatomas. The lateral semicircular canal is, due to its location, the part of the labyrinth most frequently involved (in about 90% of cases), although superior and posterior semicircular canal and cochlear locations have also been reported. Otic capsule erosion may occur in two ways. Osteolysis mediators activated within the cholesteatoma matrix or the pressure exerted by the cholesteatoma itself can lead to osteolysis and uncovering of the labyrinth; or inflammation mediators may induce resorption within the otic capsule. Fistulae are not systematically seen on the CT (57 to 60%), which is not more indicative than patient interrogation and clinical examination. Treatment is surgical, associating careful resection and coverage with aponeurosis and bone pâté, performed in the same step as cholesteatoma surgery. The risk of hearing loss depends on fistula location: it is 35% in case of a promontory location and 3% when the location is in the semicircular canals. Mastoiditis Mastoiditis is the internal equivalent of subperiosteal abscess. Apicitis (infection of the petrous apex) or petrositis Apicitis (infection of the petrous apex) or petrositis is rare but dangerous, being located near to the middle and posterior cranial fossae. The petrous apex is pneumatized in 30% of patients, and such cells are in continuity with the mastoid cells and the middle ear, enabling infection to spread. The classical clinical triad, known as Gradenigo’s syndrome, associates deep retro-orbital pain (50%), otorrhea and sixth cranial nerve palsy (25%). Primary treatment is medical; in case of failure, bone necrosis or abscess, surgery is required, with various approaches (infralabyrinthine, retrolabyrinthine, supralabyrinthine or subpetrosal), using existing cells to reach the apex if hearing has been conserved. Facial palsy Facial palsy is due to infection spreading by canal dehiscence or to direct nerve compression by the cholesteatoma.
Childhood cholesteatoma
Intracranial complications Brain abscess Brain abscess is the most frequent and most dangerous intracranial complication. Location is most frequently temporal or cerebellar. Etiology is usually hematogenic dissemination secondary to thrombophlebitis, or in some cases extension from subdural empyema following tegmen dehiscence. Evolution comprises three stages: encephalitis, then a quiescent phase with subdued symptoms, with final resurgence of symptoms in the form of elevated intracranial pressure due to abscess rupture or extension. MRI is the most suitable examination. Treatment associates broad-spectrum antibiotics, corticosteroids to reduce the brain edema and anticonvulsants. Emergency neurosurgical drainage is associated to mastoidectomy if the patient is neurologically stable enough. Frequent CT checks are the rule for monitoring. Meningitis Meningitis is the second most frequent complication. Physiopathologically, there are three distinct extension pathways: hematogenic (the most frequent), contiguity via existing foramens and fissures (Hyrtl’s fissure), or bone erosion. In 50% of cases, there is an associated intracranial complication, to be explored for on contrast CT and MRI. Treatment associates broad-spectrum antibiotics, corticosteroids to limit auditory and neurological sequelae, and mastoidectomy. Lateral or sigmoid sinus thrombosis Lateral or sigmoid sinus thrombosis makes up 20% of intracranial complications. The mechanism is either bone erosion exposing the perisinus space, or spread of mastoid emissary vein thrombophlebitis. Intraluminal thrombi induce secondary risk of hydrocephalus caused by extensive sinus thrombosis and/or septic emboli (notably pulmonary). Enhanced CT shows sinus wall uptake and ‘‘empty delta’’ sign. The exact extent of thrombus can be visualized on angio-MRI. Treatment comprises at least mastoidectomy associated to sinus denudation to drain the abscess. Surgery is not performed on the thrombus itself. Control angio-MRI is performed two weeks after initiation of treatment. Anticoagulants are a controversial issue, and are not currently recommended except in case of sagittal sinus involvement or signs of elevated intracranial pressure resistant to medical treatment. Subdural abscess or empyema Subdural abscess or empyema is induced by bone dehiscence. Signs (worsening ear- and head-ache) are often rudimentary, with insidious development. Diagnosis is by CT. Treatment is surgical, mastoidectomy enabling the abscess bone cover to be excised. Otitic hydrocephalus Otitic hydrocephalus is another complication. Its physiopathology is unclear. Development is due to superior sagittal sinus inflammation or infection blocking CSF resorption by the arachnoid villi in the Pacchioni granulations, inducing intracranial pressure elevation. Diagnosis and
149 treatment are as in intracranial pressure elevation in general. Diagnosis is based on an association of intracranial pressure elevation, its clinical symptomatology and papillary edema in the absence of ventricular dilatation or signs of meningitis. Angio-MRI confirms venous sinus thrombosis. Treatment targets the ear infection and brain pressure and prevention of catastrophic impact on the optic nerve. General anticoagulation therapy is recommended only in case of superior longitudinal sinus thrombosis.
Conflict of interest The authors have not communicated any conflict of interest.
References [1] De Verney JG. Traité de l’organe de l’ouïe. Paris, France: E. Michallet; 1683. [2] Cruveillhier J. Anatomie pathologique du corps humain. Paris: Baillière; 1829. [3] Müller J. Ueber den feineren bau und die formen der krankhaften geschwulste. Berlin: G. Reimer; 1838. [4] Semaan MT, Megerian CA. The pathophysiology of cholesteatoma. Otolaryngol Clin North Am 2006;39(6):1143—59. [5] Nager FR. The cholesteatoma of the middle ear-its etiology, pathogenesis, diagnosis and therapy. Ann Otol Rhinol Laryngol 1925;34:1249—58. [6] Olszewska E, et al. Etiopathogenesis of cholesteatoma. Eur Arch Otorhinolaryngol 2004;261(1):6—24. [7] Bennett M, et al. Congenital cholesteatoma: theories, facts, and 53 patients. Otolaryngol Clin North Am 2006;39(6):1081—94. [8] Koltai PJ, et al. The natural history of congenital cholesteatoma. Arch Otolaryngol Head Neck Surg 2002;128(7):804—9. [9] Ratnesar P. Aeration: a factor in the sequels of chronic ear disease along the Labrador and northern Newfoundland Coast 1977. [10] Michaels L. An epidermoid formation in the developing middle ear: possible source of cholesteatoma. J Otolaryngol 1986;15(3):169—74. [11] Teed RW. Cholesteatoma verum tympani: its relationship to the first epibranchial placode. Arch Otolaryngol 1936;24:455—74. [12] Glasscock 3rd ME, Dickins JR, Wiet R. Cholesteatoma in children. Laryngoscope 1981;91(10):1743—53. [13] Ruah CB, et al. Mechanisms of retraction pocket formation in the pediatric tympanic membrane. Arch Otolaryngol Head Neck Surg 1992;118(12):1298—305. [14] Bujia J, et al. Immunobiological peculiarities of cholesteatoma in children: quantification of epithelial proliferation by MIB1. Laryngoscope 1996;106(7):865—8. [15] Palva A, Karma P, Karja J. Cholesteatoma in children. Arch Otolaryngol 1977;103(2):74—7. [16] Sudhoff H, Tos M. Pathogenesis of attic cholesteatoma: clinical and immunohistochemical support for combination of retraction theory and proliferation theory. Am J Otol 2000;21(6):786—92. [17] Quaranta A, Resta L, Santangelo A. Otomastoid cholesteatoma in children: histopathological findings. Int J Pediatr Otorhinolaryngol 1986;12(2):121—6. [18] Dornelles Cde C, et al. Comparison of acquired cholesteatoma between pediatric and adult patients. Eur Arch Otorhinolaryngol 2009;266(10):1553—61.
150 [19] Alves AL, et al. Analysis of histopathological aspects in acquired middle ear cholesteatoma. Braz J Otorhinolaryngol 2008;74(6):835—41. [20] Friedland DR, et al. Cholesteatoma growth and proliferation: post-transcriptional regulation by MicroRNA-21. Otol Neurotol 2009;30(9):998—1005. [21] Derlacki EL, Clemis JD. Congenital cholesteatoma of the middle ear and mastoid. Ann Otol Rhinol Laryngol 1965;74(3):706—27. [22] Levenson MJ, et al. A review of 20 congenital cholesteatomas of the middle ear in children. Otolaryngol Head Neck Surg 1986;94(5):560—7. [23] Valvassori GE. Benign tumors of the temporal bone. Radiol Clin North Am 1974;12(3):533—42. [24] Tos M. Recurrence and the condition of the cavity after surgery for cholesteatoma using various techniques. Kugler and Ghedini: 1989. [25] Roger G, et al. Predictive risk factors of residual cholesteatoma in children: a study of 256 cases. Am J Otol 1997;18(5):550—8. [26] De Corso E, et al. Aural acquired cholesteatoma in children: surgical findings, recurrence and functional results. Int J Pediatr Otorhinolaryngol 2006;70(7):1269—73. [27] Stangerup SE, Drozdziewicz D, Tos M. Cholesteatoma in children, predictors and calculation of recurrence rates. Int J Pediatr Otorhinolaryngol 1999;49(Suppl. 1):S69—73. [28] Sheehy JL, Brackman DE, Graham MD. Complications of cholesteatoma: a report on 1024 cases. In: Hattori BF, Sade J, Abramson M, editors. Cholesteatoma: first International Conference. Birmingham: Alabama: Aesculapius Publishing; 1977. p. 420—9. [29] Edelstein DR, Parisier SC, Han JC. Acquired cholesteatoma in the pediatric age group. Otolaryngol Clin North Am 1989;22(5):955—66. [30] Darrouzet V, et al. 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 2000;21(4):474—81. [31] Iino Y, et al. Risk factors for recurrent and residual cholesteatoma in children determined by second stage operation. Int J Pediatr Otorhinolaryngol 1998;46(1—2):57—65.
J. Nevoux et al. [32] Williams MT, Ayache D. Imaging of the postoperative middle ear. Eur Radiol 2004;14(3):482—95. [33] Williams MT, Ayache D. Imaging in adult chronic otitis. J Radiol 2006;87(11 Pt 2):1743—55. [34] Dhepnorrarat RC, Wood B, Rajan GP. Postoperative nonecho-planar diffusion-weighted magnetic resonance imaging changes after cholesteatoma surgery: implications for cholesteatoma screening. Otol Neurotol 2009;30(1): 54—8. [35] Manolis EN, et al. Radiologic evaluation of the ear anatomy in pediatric cholesteatoma. J Craniofac Surg 2009;20(3): 807—10. [36] Vercruysse JP, et al. 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 2006;16(7):1461—7. [37] Dubrulle F, et al. Diffusion-weighted MR imaging sequence in the detection of postoperative recurrent cholesteatoma. Radiology 2006;238(2):604—10. [38] Schmitz BL, et al. Advantages and pitfalls in 3T MR brain imaging: a pictorial review. AJNR Am J Neuroradiol 2005;26(9):2229—37. [39] Lehmann P, et al. 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 2009;30(2):423—7. [40] Mueller DT, Isaacson G. Staged tympanostomy tube placement facilitates pediatric cholesteatoma management. Int J Pediatr Otorhinolaryngol 2008;72(2):167—71. [41] Denoyelle F, et al. Myringoplasty in children: predictive factors of outcome. Laryngoscope 1999;109(1):47—51. [42] Dornhoffer JL. Cartilage tympanoplasty. Otolaryngol Clin North Am 2006;39(6):1161—76. [43] Hamilton JW. Efficacy of the KTP laser in the treatment of middle ear cholesteatoma. Otol Neurotol 2005;26(2):135—9. [44] Smith JA, Danner CJ. Complications of chronic otitis media and cholesteatoma. Otolaryngol Clin North Am 2006;39(6):1237—55.
REVIEW ARTICLE
Childhood cholesteatoma J. Nevoux a,∗, M. Lenoir b, G. Roger a, F. Denoyelle a, H. Ducou Le Pointe b, E.-N. Garabédian a a
Inserm U587, service d’oto-rhinolaryngologie et de chirurgie cervicofaciale, hôpital d’enfants Armand-Trousseau, AP—HP, UMPC, université Paris 6, 26-28, avenue du Dr Arnold-Netter, 75571 Paris cedex 12, France b Service de radiologie, hôpital d’enfants Armand-Trousseau, 26-28, avenue du Dr Arnold-Netter, 75571 Paris cedex 12, France Available online 11 August 2010
KEYWORDS Acquired cholesteatoma; Congenital cholesteatoma; Children
Summary Although cholesteatoma was first described in 1683, its etiopathogeny remains unexplained. In children, there are two forms: acquired cholesteatoma, resembling the adult form, and congenital cholesteatoma. The acquired form has become less frequent in recent years, thanks to progress in the treatment of childhood otitic pathology. Diagnosis of congenital cholesteatoma, on the contrary, is increasing, due to improvements in information to health care professionals and in diagnostic tools. Clinical and histological evidence points to greater aggressiveness in childhood forms, although this difference cannot, at present, be precisely explained. Diagnosis is clinical, but CT and MR imaging is indispensable for preoperative assessment and postoperative follow-up. New delayed gadolinium-enhanced T1-weighted and diffusion-weighted MRI sequences have recently been developed and provide more precise radiological diagnosis. Treatment is surgical; alternatives, notably by laser, have proved unsuccessful. Complications concern involvement of neighbouring structures, and are mainly infectious; some can be life-threatening, and should be systematically screened. © 2010 Published by Elsevier Masson SAS.
Introduction Cholesteatoma is a severe middle-ear pathology affecting both adults and children. It would seem to have been first described by De Verney in France in 1683, as what he called ‘‘steatoma’’ [1]. In 1829, the French anatomopathologist Cruveilhier described it as a pearl tumor of the temporal bone [2]. The term ‘‘cholesteatoma’’, introduced in 1838 by the German physiologist Johannes Müller [3], is now well established, but in fact, etymologically faulty in
∗
Corresponding author. Tel.: +33144736114; fax: +33144736108. E-mail address: [email protected] (J. Nevoux).
as much as this benign tumor (‘‘. . .oma’’) contains neither cholesterol (‘‘chole’’) nor fat (‘‘steat’’). In 1855, Virchow categorized cholesteatoma under epidermoid carcinoma and atheroma, but only in 1861 did Von Troeltsch consider its epidermal origin. In the studies by Gruber, Wendt and Rokitansky (1855—1888), the physiopathology of cholesteatoma was described as a malpighian metaplasia of the middle-ear mucosa in response to chronic inflammation. At the end of the 19th century, Bezold and Habermann overturned this theory, showing that cholesteatoma was caused by migration of external auditory canal skin to the middle ear, induced by chronic inflammation [4]. At the present time, despite numerous investigations, the physiopathology of cholesteatoma has still not been elucidated.
1879-7296/$ – see front matter © 2010 Published by Elsevier Masson SAS. doi:10.1016/j.anorl.2010.07.001
144 In children, there are two different types of cholesteatoma: acquired, which also affects adults, and congenital, which is specific to childhood. The most recent progress has been in imaging, with the introduction of MRI in the early 2000s, followed by the recent development of novel sequences. Treatment, on the other hand, has not greatly progressed and remains mainly surgical.
J. Nevoux et al. accepted theory. Both authors demonstrated the presence of an epidermoid residue in the anterosuperior quadrant of the middle ear in the human fetus, on the lateral wall of the Eustachian tube, near the annulus. Failure to reabsorb the epidermoid residue normally found in embryos between the 10th and 33rd week of gestation has been observed [10,11].
Acquired cholesteatoma
Epidemiology Incidence is hard to determine, especially for congenital cholesteatoma. Recent figures testify to a fall in the incidence of both adult and childhood acquired cholesteatoma: in 1925, the rate of cholesteatoma in the under-16’s was one in three; it is now much less frequent, but incidence in the early 2000s was still three per 100,000 (compared to nine per 100,000 in adults) [5,6]. The incidence of congenital cholesteatoma, on the other hand, seems to be on the rise, at 0.12 per 100,000 children [7]. Some 1 to 3% of cases of childhood cholesteatoma, and 1 to 5% of cholesteatoma as a whole, are congenital [8]. The relative increase is due to the fall in the number of cases of acquired cholesteatoma and an increase in the diagnosis of congenital cholesteatoma. The latter increase is thanks to the pathology being better known by ENT specialists and pediatricians; diagnostic tools, in particular microscopy, have improved, as has otologic follow-up of children. Prevalence varies between populations: it is highest among Caucasians, followed by Africans, and very low in Asians. Ratnesar reported extremely low prevalence of cholesteatoma in Inuits [9].
Physiopathology Cholesteatoma is a non-neoplastic but destructive cystic lesion containing keratin layers from a keratinized malpighian epithelium surrounded by a matrix of epithelium that rests on a perimatrix formed of a conjunctive stroma of varying thickness. It generally occurs in the middle ear, but other locations in or around the petrosal bone are also possible [6]. The pathologic substrate of cholesteatoma is a keratinized stratified malpighian epithelium, the middleear origin of which is, however, controversial. Various etiopathogenic theories have been proposed to account for the two forms of cholesteatoma. Despite recent research, the underlying mechanisms remain unknown.
Congenital cholesteatoma Various theories have been put forward to explain the formation of congenital cholesteatoma. In 1854, Von Remak postulated that the origin of dermoid and other related tumors (such as cholesteatoma) was a skin follicle trapped during the early stages of embryogenesis [6]. Subsequently, many theories were formulated, culminating in that of epidermoid formation, close to Von Remak’s postulate, hypothesized by Teed in 1936 and then by Michaels in 1986, and which remains the most widely
During the course of the 20th century, four physiopathological theories of the formation of acquired or secondary cholesteatoma were put forward. In 1864, Von Tröltsch first proposed a squamous metaplasia theory; in 1890, Habermann and Bezold proposed the theory of migration or invasion; in 1925, Lange first reported the phenomenon underlying the basal hyperplasia theory; and finally, in 1933, Wittmaack formulated the retraction pocket or invagination theory, which is now the most widely accepted [6]. A1though there is no firmly established theory, primary and secondary acquired cholesteatoma are classically distinguished: the former evolves from a retraction pocket, while the latter is due to epithelial migration through a ruptured ear-drum or to epithelium implanted in the middle ear. Such implantation may be iatrogenic, occurring during ear surgery, or due to epithelium left in the tympanic cavity after healing of a blast injury, or else arising posttraumatically at a petrosal bone fracture site [4].
Histopathologic differences between adult and childhood cholesteatoma Childhood cholesteatoma has long been known to be more aggressive than the adult form, with poorer clinical prognosis. Several studies have sought a physiopathological explanation for this finding [12—16]. According to Quaranta, the explanation lies in the cholesteatoma perimatrix. Following up this hypothesis, he found the number of mononuclear inflammation elements (plasmocytes, lymphocytes, macrophages, granulocytes and giant cells) in the perimatrix to be greater in children than in adults [17]. Bujia’s group analyzed the expression of MIB1 (a monoclonal antibody marker of cell proliferation) in child and adult cholesteatoma and within the external auditory canal skin; the proliferation index was normal in the latter, elevated in the former, and relatively higher in children [14]. Recent studies confirmed that the perimatrix is thicker and the epithelium better contoured in child cholesteatoma [18]. Other studies, in contrast, found no significant histopathological differences between adults and children in the matrix of acquired cholesteatoma that could account for the difference in clinical aggressiveness [19].
Progress in histology There have been many studies of the cellular basis of the specificities of childhood cholesteatoma, its local aggressiveness and growth potential. However, they have yet to provide an account that could enable medical treatment to
Childhood cholesteatoma be developed. There are, however, phenomena of hyperproliferation specific to middle-ear skin. Gene expression studies confirmed growth-factor and signalling molecule upregulation. Likewise, protein expression studies demonstrated elevated rates of the molecules involved in growth, proliferation, antiapoptosis and bone resorption. Recent breakthroughs have concerned angiogenesis and metalloproteinases. These markers seem to explain the difference in aggressiveness between childhood and adult forms. In 2009, a Brazilian group demonstrated significantly elevated inflammation and metalloproteinase production in childhood cholesteatoma; perimatrix angiogenesis, in terms of blood-vessel count, was elevated in children. Metalloproteinases are proteolytic enzymes associated with pathologic extracellular matrix resorption in chronic inflammation. Inflammatory processes, such as infected cholesteatoma, stimulate production of these enzymes. Dornelles’s group demonstrated cytoplasmic hyperexpression of MMP2 (matrix metalloproteinase) and nuclear hyperexpression of MMP2 and MMP9 in pediatric cholesteatoma. There also appears to be a correlation between expression level and age: the younger the child, the higher the cytoplasmic expression of MMP2 in the cholesteatoma. Metalloproteinases are associated with the capacity to penetrate and invade neighboring tissue. Their other role is to regulate bone homeostasis. They are elevated in inflammatory osteolytic pathology. The greater aggressiveness and osteolytic capacity of childhood cholesteatoma would thus seem to derive from hyperexpression of perimatricial metalloproteinases [18]. The most recent studies have investigated the possible role of protein transcription regulators, micro-RNAs. These RNA fragments of 22 to 24 nucleotide length, strongly conserved over evolution and non-coding, inhibit protein transcription. They play an essential role in cell regulation pathways and are strongly implicated in oncogenesis. In 2008, Cioffi’s group reported significant upregulation of micro-RNA-21 (hsa-miR-21) in cholesteatoma as compared to normal skin in humans, and proposed a growth and proliferation model of cholesteatoma [20]. At the present time, the histologic and cellular phenomena underlying the development of cholesteatoma and its differential aggressiveness between childhood and adult forms remain unknown.
Clinical aspect Although cholesteatoma occurs mainly in the middle ear, other locations are also possible. Classically, extra- and intradural locations are distinguished. Extradural involvement mainly concerns the middle ear, but may lie in any part of the petrosal bone: mastoid, external auditory canal or petrous apex. Intradural involvement, better known as epidermoid cyst, has been reported at various locations, the most frequent being the pontocerebellar angle. Whatever the location, complex surgery is required, as there is at present no medical treatment [4].
Congenital cholesteatoma In 1965, Derlacki and Clemis defined congenital cholesteatoma as an embryologic residue of epithelial
145
Figure 1
Congenital left-ear cholesteatoma.
tissue behind a normal tympanic membrane, in the absence of any history of infection or ear surgery [21]. They formulated a series of diagnostic criteria: normal tympanic membrane, retrotympanic white mass, and no history of acute otitis media, otorrhea, eardrum rupture or otologic surgery (Fig. 1). In 1986, these criteria were modified by Levenson, who suggested that history of acute otitis media does not necessarily exclude diagnosis of congenital cholesteatoma, given its high frequency in young children [22]. Mean age at diagnosis is 4—5 years. Systematic recurrence of acute otitis media or otorrhea in the same ear is an alarm signal. According to Valvassori, congenital cholesteatoma can be found at four locations in the region of the petrosal bone: middle ear (middle ear cavity and mastoid), petrous apex, pontocerebellar angle, and jugular foramen [23]. It is a rare congenital lesion, making up 1% of intracranial tumors. Forty to 50% of cases are located in the pontocerebellar angle.
Acquired cholesteatoma Various classifications have been drawn up for acquired cholesteatoma, but there is not one, which is universally employed. Tos’s classification, however, basing indications on the location of the pathological process, seems to be the most frequently used. It distinguishes three types according to location: attic cholesteatoma, tympani sinus cholesteatoma with a posterior subligamentary starting point, and pars tensa cholesteatoma [24]. In children, 80% of retraction pockets and cholesteatomas develop from or depend on the pars tensa (especially the posterior subligamentary pars tensa) (Fig. 2), while in adults, they mainly develop in the pars flaccid (or Shrapnell’s membrane) (Fig. 3) [25]. To cholesteatomas must be added unstable retraction pockets, which are equivalent in as much as they have the same predictive factors as cholesteatoma. The clinical features of these pockets are their marginal location, their fixation (outside of the ossicular chain), the impossibility of complete control, and associated otorrhea and/or squamata [25].
146
J. Nevoux et al. Residual lesion and/or recurrence require long-term clinical and radiological follow-up over at least 8 years [25]. Recurrence and residual lesion rates are higher in children: for recurrence, the rate is 30%, versus 3 to 15% in adults [29,30] and, for residual lesion, about 40% versus 15 to 20% in adults [31].
Imaging assessment Computed tomography (CT)
Figure 2
Acquired childhood cholesteatoma.
Acquired cholesteatoma is more aggressive in children, for a number of reasons: it is often more extensive at diagnosis; ossicular chain status is often poorer, due to more extensive ossicle lysis; and the rates of recurrence and residual lesion are higher. This greater aggressiveness in children has been explained in terms of childhood specificities in the cholesteatoma matrix and/or persisting predisposing factors such as impaired aeration of the ear [26]. Predictive factors for recurrence are large size and extension, age < 8 years, associated otorrhea, ossicular chain status, and impaired preoperative aeration of the ear [27]. Predictive factors for residual lesion are initial posterior mid-tympanic invasion, ossicular chain interruption, lack of expertise on the part of the surgeon, and incomplete resection. The greater the association of these factors, the higher the risk of residual lesion. In case of one risk factor, secondlook surgery should systematically be performed within 12 to 18 months. In case of two or more risk factors, second-look surgery should be performed within 9 months [25]. Although cholesteatoma is more aggressive in children, the number of local (labyrinthine fistula or facial palsy) or regional (intracranial) complications is lower [28].
Temporal bone CT assesses cholesteatoma extension, exploring for complications and detecting anatomic variants so as to prepare for any difficulties in surgery. Diagnosis is normally clinical and otoscopic. Technically, CT examination should be high-resolution, non-contrast, with bone filter, with overlapping slices of subor millimetric thickness in the axial (orbitomeatal plane: i.e., the plane of the lateral semi-circular canal) and coronal planes, orthogonally through the entire pars petrosa. Multislice CT further requires submillimetric volume acquisition centered on the pars petrosa, with the patient installed so as to protect the orbits from radiation, with axial and coronal reconstruction in the plane of the lateral semi-circular canal.[32,33] The negative predictive value (NPV) of CT is excellent [34]. Cholesterol granuloma, however, which is frequently found in association with cholesteatoma, cannot be differentiated on CT; MRI enables the distinction. The main CT criteria for cholesteatoma are: nodular soft tissue mass within the tympano-mastoid cavities, with associated osteolysis (wall of the epitympanic recess, and ossicles, especially the long apophysis of the incus, tegmen tympani and bone shell of the lateral semicircular canal) [35]. In case of doubt on first examination, repeat scan performed a few months later detects growth of the residual cholesteatoma and thus provides better visualization. The main limitation of CT is fluid effusion in the middle ear, hindering diagnosis of a residual lesion. MRI can then characterize soft tissue and differentiate between cholesteatoma and effusion. One essential issue in CT is radioprotection, especially in children presenting with chronic otitis media, requiring numerous scans over their lifetime. This mainly concerns the crystalline lens, and the risk of radio-induced cancer. The crystalline lens is conserved with multislice CT, which allows volume acquisition of the pars petrosa in a slice plane excluding the orbit, with reconstruction in the classic slice plane. Irradiation is standardized in petrous bone CT, at about 8 years’ natural irradiation, whereas radioprotection standards give a risk threshold of more than 100 years for natural radiation-induced cancer. This point needs to be taken into account in the radiological follow-up of childhood cholesteatoma.
MRI Figure 3
Acquired adult cholesteatoma.
MRI has more recently been introduced in cholesteatoma assessment, and is indicated for initial extension assess-
Childhood cholesteatoma ment, in case of contact with the meninges in particular, and for postoperative follow-up in case of doubtful CT images. MRI characteristics in cholesteatoma are: isosignal with or without peripheral enhancement on post-gadolinium T1-weighted sequences, and isosignal on T2-weighted sequences. Associating delayed (30—50 minutes, or ideally 45 minutes according to certain authors) gadoliniumenhanced T1-weighted sequences improved diagnosis of residual cholesteatoma. The cholesteatoma never shows enhancement, appearing in hyposignal with respect to surrounding tissue. In the last few years, a new type of sequence based on the signal produced by displaced water molecules, known as diffusion-weighted sequence, has provided further information for residual lesion screening [34]. Cholesteatoma shows as hypersignal on diffusion-weighted sequences. The b-factor is usually b800 in children and b1,000 in adults. The MRI resolution threshold is 3 mm on standard sequences, but just under 5 mm on diffusion-weighted sequences. The other limitations of this new sequence are magnetic susceptibility artifacts and a T2 effect. These artifacts are related to the numerous air-bone interfaces due to cranial base anatomy, inducing considerable distortion. Mean sensitivity and specificity in classic diffusion-weighted sequences are respectively 81.6% (range, 77 to 86%) and 100%, and positive predictive value (PPV) and NPV respectively 100 and 75% [36]. To get around these limitations, an improved technique has been developed: turbo-spin echo (TSE) or non-echo-planar (non-EPI) instead of echo-planar (EPI) diffusion-weighted imaging [37]. This technique improves spatial resolution, reducing the detection threshold to 3 mm. PPV, NPV, sensitivity and specificity seem to be 100%. Moreover, acquisition time is shorter, at about 3—4 minutes, compared to nearly 5—8 minutes with the classical technique [34]. Performing both CT and MRI is essential, as the aim is to locate a tissue process within the temporal bone: CT provides better spatial location, and MRI better tissue characterization. The power of most MRI machines used in everyday practice is 1.5 Tesla. With the development and purchasing of new models, MRI at 3T or greater is becoming more common and available. The usual advantage is shorter acquisition time; but due to artifacts specific to this anatomic region, diffusion-weighted sequence acquisition is paradoxically longer, incurring a risk of motion artifacts, especially in children making interpretation tricky even for experienced radiologists [38,39]. The two main problems with diffusion-weighted sequences, whether classic or echoplanar, are the lack of spatial resolution and the amount of artifacts. The latter mainly concern bone/air and tissue/air interfaces. The problems are partially resolved using TSE sequences, although spatial resolution remains poor. Examination duration and its requirements (sedation and contention) are also important points. The time needed for CT scan of the petrous bone varies greatly depending on the equipment. In our experience of multislice CT, standard millimetric acquisition takes approximately 30 secondes, and submillimetric acquisition 70 secondes (such better resolution being needed notably to visualize the stapes). MRI takes
147 approximately 30 minutes. In our experience, sedation is employed for CT in children under 3 years of age, and for MRI in under-4-year-olds; older children can tolerate the examination if they and their parents have been properly informed of the importance of the examination and how it takes place. In case of difficulties of communication (such as associated neuropsychiatric pathology or profound deafness), general anesthesia is required (Table 1).
Treatment Treatment remains surgical, although other options have been tested. The prime objective is to eradicate the disease so as to restore a healthy aerated ear. No less important, the second objective is to restore or improve the ossicular chain. Since the first description of mastoidectomy by Riolanus in 1649, numerous procedures have been described. The main distinction, however, is whether or not to conserve the posterior wall of the external auditory canal, distinguishing canal wall-up tympanoplasty (with conservation) from canal wall-down tympanoplasty (non-conservation). Middle-ear surgery is hampered in children by obstruction of the Eustachian tube and chronic inflammatory hyperplasic mucosa, increasing the risk of incomplete resection and residual lesion. Fitting ventilation tubes ahead of cholesteatoma resection improves surgical preparation and postoperative course [40]. Cartilage should be used for eardrum reinforcement in children, whereas the temporal aponeurosis is preferred in adults. Posterior reinforcement is essential, given the preponderance of recurrence and retraction in this area of the tympanum. The problems encountered with cartilage are failure to control retrotympanic liquid effusion, and possible residual lesion [41,42]. Attempts at laser treatment, mainly to reduce the rate of residual lesion in second-stage surgery, have not been greatly followed up. In 2005, Hamilton reported a significant reduction in residual lesions, using the laser to control exeresis. He applied it specifically for ossicle locations and areas inaccessible to classic surgery [43].
Complications Complications are due to involvement of neighboring structures and to cholesteatoma location, and are generally of infectious origin. Classically, a distinction is drawn between intra- and extracranial complications, the latter being further divided into intra- and extratemporal complications [44].
Extratemporal complications Subperiosteal abscess Subperiosteal abscess is the most common extratemporal complication. It is caused by spread of infection from the mastoid towards the periosteal space by erosion of the mastoid cortex or, more rarely, by vascular propagation secondary to mastoid emissary vein thrombosis. It is most frequent in young children. The clinical presentation is of
148
J. Nevoux et al.
Low signal
Low signal
Low signal
Low signal
High signal
Low signal
3 mm
Low signal
Low signal
Low signal
Low signal
High signal
Low signal
5 mm 3 mm
No enhancement
3 mm
Intermediate signal
High signal Low signal
2.5 mm
Fluid
Resolution size
Intermediate signal
2.5 mm
No enhancement
No enhancement
Intermediate signal Intermediate signal
Inflammatory tissue Hyperplasic mucosa Cholesteatoma
No enhancement
Enhancement
Intermediate signal Intermediate signal Scar tissue fibrosis
Enhancement
No enhancement
Intermediate signal High signal Cholesterol granuloma
No enhancement No enhancement
Intermediate signal Intermediate signal Granulation tissue
Delayed post-gadolinium T1-weighted 30—50 minutes Enhancement Early post-gadolinium T1 weighted 10—15 minutes Enhancement T2-weighted T1-weighted
Signal according to MRI sequence and middle-ear process. Table 1
Low enhancement
TSE diffusion-weighted EPI diffusion-weighted
classical mastoiditis. Diagnosis is confirmed by CT; treatment is basically surgical, associating abscess drainage and cholesteatoma exeresis in a single step. Bezold’s abscess Bezold’s abscess is a cervical abscess, physiopathologically identical to a subperiosteal abscess, the only difference being a bone erosion area at the mastoid apex, which explains the extension to the neck below the sternocleidomastoid muscle. It is more frequent in older children and adults, due to the late pneumatization of the mastoid apex. Like subperiosteal abscesses, it is more frequent in case of acute otitis media but may also complicate cholesteatoma. Diagnosis and treatment are as in subperiosteal abscess, with one difference: the cervical incision.
Intra-temporal complications Labyrinthine fistula is also very frequent, in 7% of cholesteatomas. The lateral semicircular canal is, due to its location, the part of the labyrinth most frequently involved (in about 90% of cases), although superior and posterior semicircular canal and cochlear locations have also been reported. Otic capsule erosion may occur in two ways. Osteolysis mediators activated within the cholesteatoma matrix or the pressure exerted by the cholesteatoma itself can lead to osteolysis and uncovering of the labyrinth; or inflammation mediators may induce resorption within the otic capsule. Fistulae are not systematically seen on the CT (57 to 60%), which is not more indicative than patient interrogation and clinical examination. Treatment is surgical, associating careful resection and coverage with aponeurosis and bone pâté, performed in the same step as cholesteatoma surgery. The risk of hearing loss depends on fistula location: it is 35% in case of a promontory location and 3% when the location is in the semicircular canals. Mastoiditis Mastoiditis is the internal equivalent of subperiosteal abscess. Apicitis (infection of the petrous apex) or petrositis Apicitis (infection of the petrous apex) or petrositis is rare but dangerous, being located near to the middle and posterior cranial fossae. The petrous apex is pneumatized in 30% of patients, and such cells are in continuity with the mastoid cells and the middle ear, enabling infection to spread. The classical clinical triad, known as Gradenigo’s syndrome, associates deep retro-orbital pain (50%), otorrhea and sixth cranial nerve palsy (25%). Primary treatment is medical; in case of failure, bone necrosis or abscess, surgery is required, with various approaches (infralabyrinthine, retrolabyrinthine, supralabyrinthine or subpetrosal), using existing cells to reach the apex if hearing has been conserved. Facial palsy Facial palsy is due to infection spreading by canal dehiscence or to direct nerve compression by the cholesteatoma.
Childhood cholesteatoma
Intracranial complications Brain abscess Brain abscess is the most frequent and most dangerous intracranial complication. Location is most frequently temporal or cerebellar. Etiology is usually hematogenic dissemination secondary to thrombophlebitis, or in some cases extension from subdural empyema following tegmen dehiscence. Evolution comprises three stages: encephalitis, then a quiescent phase with subdued symptoms, with final resurgence of symptoms in the form of elevated intracranial pressure due to abscess rupture or extension. MRI is the most suitable examination. Treatment associates broad-spectrum antibiotics, corticosteroids to reduce the brain edema and anticonvulsants. Emergency neurosurgical drainage is associated to mastoidectomy if the patient is neurologically stable enough. Frequent CT checks are the rule for monitoring. Meningitis Meningitis is the second most frequent complication. Physiopathologically, there are three distinct extension pathways: hematogenic (the most frequent), contiguity via existing foramens and fissures (Hyrtl’s fissure), or bone erosion. In 50% of cases, there is an associated intracranial complication, to be explored for on contrast CT and MRI. Treatment associates broad-spectrum antibiotics, corticosteroids to limit auditory and neurological sequelae, and mastoidectomy. Lateral or sigmoid sinus thrombosis Lateral or sigmoid sinus thrombosis makes up 20% of intracranial complications. The mechanism is either bone erosion exposing the perisinus space, or spread of mastoid emissary vein thrombophlebitis. Intraluminal thrombi induce secondary risk of hydrocephalus caused by extensive sinus thrombosis and/or septic emboli (notably pulmonary). Enhanced CT shows sinus wall uptake and ‘‘empty delta’’ sign. The exact extent of thrombus can be visualized on angio-MRI. Treatment comprises at least mastoidectomy associated to sinus denudation to drain the abscess. Surgery is not performed on the thrombus itself. Control angio-MRI is performed two weeks after initiation of treatment. Anticoagulants are a controversial issue, and are not currently recommended except in case of sagittal sinus involvement or signs of elevated intracranial pressure resistant to medical treatment. Subdural abscess or empyema Subdural abscess or empyema is induced by bone dehiscence. Signs (worsening ear- and head-ache) are often rudimentary, with insidious development. Diagnosis is by CT. Treatment is surgical, mastoidectomy enabling the abscess bone cover to be excised. Otitic hydrocephalus Otitic hydrocephalus is another complication. Its physiopathology is unclear. Development is due to superior sagittal sinus inflammation or infection blocking CSF resorption by the arachnoid villi in the Pacchioni granulations, inducing intracranial pressure elevation. Diagnosis and
149 treatment are as in intracranial pressure elevation in general. Diagnosis is based on an association of intracranial pressure elevation, its clinical symptomatology and papillary edema in the absence of ventricular dilatation or signs of meningitis. Angio-MRI confirms venous sinus thrombosis. Treatment targets the ear infection and brain pressure and prevention of catastrophic impact on the optic nerve. General anticoagulation therapy is recommended only in case of superior longitudinal sinus thrombosis.
Conflict of interest The authors have not communicated any conflict of interest.
References [1] De Verney JG. Traité de l’organe de l’ouïe. Paris, France: E. Michallet; 1683. [2] Cruveillhier J. Anatomie pathologique du corps humain. Paris: Baillière; 1829. [3] Müller J. Ueber den feineren bau und die formen der krankhaften geschwulste. Berlin: G. Reimer; 1838. [4] Semaan MT, Megerian CA. The pathophysiology of cholesteatoma. Otolaryngol Clin North Am 2006;39(6):1143—59. [5] Nager FR. The cholesteatoma of the middle ear-its etiology, pathogenesis, diagnosis and therapy. Ann Otol Rhinol Laryngol 1925;34:1249—58. [6] Olszewska E, et al. Etiopathogenesis of cholesteatoma. Eur Arch Otorhinolaryngol 2004;261(1):6—24. [7] Bennett M, et al. Congenital cholesteatoma: theories, facts, and 53 patients. Otolaryngol Clin North Am 2006;39(6):1081—94. [8] Koltai PJ, et al. The natural history of congenital cholesteatoma. Arch Otolaryngol Head Neck Surg 2002;128(7):804—9. [9] Ratnesar P. Aeration: a factor in the sequels of chronic ear disease along the Labrador and northern Newfoundland Coast 1977. [10] Michaels L. An epidermoid formation in the developing middle ear: possible source of cholesteatoma. J Otolaryngol 1986;15(3):169—74. [11] Teed RW. Cholesteatoma verum tympani: its relationship to the first epibranchial placode. Arch Otolaryngol 1936;24:455—74. [12] Glasscock 3rd ME, Dickins JR, Wiet R. Cholesteatoma in children. Laryngoscope 1981;91(10):1743—53. [13] Ruah CB, et al. Mechanisms of retraction pocket formation in the pediatric tympanic membrane. Arch Otolaryngol Head Neck Surg 1992;118(12):1298—305. [14] Bujia J, et al. Immunobiological peculiarities of cholesteatoma in children: quantification of epithelial proliferation by MIB1. Laryngoscope 1996;106(7):865—8. [15] Palva A, Karma P, Karja J. Cholesteatoma in children. Arch Otolaryngol 1977;103(2):74—7. [16] Sudhoff H, Tos M. Pathogenesis of attic cholesteatoma: clinical and immunohistochemical support for combination of retraction theory and proliferation theory. Am J Otol 2000;21(6):786—92. [17] Quaranta A, Resta L, Santangelo A. Otomastoid cholesteatoma in children: histopathological findings. Int J Pediatr Otorhinolaryngol 1986;12(2):121—6. [18] Dornelles Cde C, et al. Comparison of acquired cholesteatoma between pediatric and adult patients. Eur Arch Otorhinolaryngol 2009;266(10):1553—61.
150 [19] Alves AL, et al. Analysis of histopathological aspects in acquired middle ear cholesteatoma. Braz J Otorhinolaryngol 2008;74(6):835—41. [20] Friedland DR, et al. Cholesteatoma growth and proliferation: post-transcriptional regulation by MicroRNA-21. Otol Neurotol 2009;30(9):998—1005. [21] Derlacki EL, Clemis JD. Congenital cholesteatoma of the middle ear and mastoid. Ann Otol Rhinol Laryngol 1965;74(3):706—27. [22] Levenson MJ, et al. A review of 20 congenital cholesteatomas of the middle ear in children. Otolaryngol Head Neck Surg 1986;94(5):560—7. [23] Valvassori GE. Benign tumors of the temporal bone. Radiol Clin North Am 1974;12(3):533—42. [24] Tos M. Recurrence and the condition of the cavity after surgery for cholesteatoma using various techniques. Kugler and Ghedini: 1989. [25] Roger G, et al. Predictive risk factors of residual cholesteatoma in children: a study of 256 cases. Am J Otol 1997;18(5):550—8. [26] De Corso E, et al. Aural acquired cholesteatoma in children: surgical findings, recurrence and functional results. Int J Pediatr Otorhinolaryngol 2006;70(7):1269—73. [27] Stangerup SE, Drozdziewicz D, Tos M. Cholesteatoma in children, predictors and calculation of recurrence rates. Int J Pediatr Otorhinolaryngol 1999;49(Suppl. 1):S69—73. [28] Sheehy JL, Brackman DE, Graham MD. Complications of cholesteatoma: a report on 1024 cases. In: Hattori BF, Sade J, Abramson M, editors. Cholesteatoma: first International Conference. Birmingham: Alabama: Aesculapius Publishing; 1977. p. 420—9. [29] Edelstein DR, Parisier SC, Han JC. Acquired cholesteatoma in the pediatric age group. Otolaryngol Clin North Am 1989;22(5):955—66. [30] Darrouzet V, et al. 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 2000;21(4):474—81. [31] Iino Y, et al. Risk factors for recurrent and residual cholesteatoma in children determined by second stage operation. Int J Pediatr Otorhinolaryngol 1998;46(1—2):57—65.
J. Nevoux et al. [32] Williams MT, Ayache D. Imaging of the postoperative middle ear. Eur Radiol 2004;14(3):482—95. [33] Williams MT, Ayache D. Imaging in adult chronic otitis. J Radiol 2006;87(11 Pt 2):1743—55. [34] Dhepnorrarat RC, Wood B, Rajan GP. Postoperative nonecho-planar diffusion-weighted magnetic resonance imaging changes after cholesteatoma surgery: implications for cholesteatoma screening. Otol Neurotol 2009;30(1): 54—8. [35] Manolis EN, et al. Radiologic evaluation of the ear anatomy in pediatric cholesteatoma. J Craniofac Surg 2009;20(3): 807—10. [36] Vercruysse JP, et al. 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 2006;16(7):1461—7. [37] Dubrulle F, et al. Diffusion-weighted MR imaging sequence in the detection of postoperative recurrent cholesteatoma. Radiology 2006;238(2):604—10. [38] Schmitz BL, et al. Advantages and pitfalls in 3T MR brain imaging: a pictorial review. AJNR Am J Neuroradiol 2005;26(9):2229—37. [39] Lehmann P, et al. 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 2009;30(2):423—7. [40] Mueller DT, Isaacson G. Staged tympanostomy tube placement facilitates pediatric cholesteatoma management. Int J Pediatr Otorhinolaryngol 2008;72(2):167—71. [41] Denoyelle F, et al. Myringoplasty in children: predictive factors of outcome. Laryngoscope 1999;109(1):47—51. [42] Dornhoffer JL. Cartilage tympanoplasty. Otolaryngol Clin North Am 2006;39(6):1161—76. [43] Hamilton JW. Efficacy of the KTP laser in the treatment of middle ear cholesteatoma. Otol Neurotol 2005;26(2):135—9. [44] Smith JA, Danner CJ. Complications of chronic otitis media and cholesteatoma. Otolaryngol Clin North Am 2006;39(6):1237—55.