Skeletal Radiol (1996) 25:31–36 © International Skeletal Society 1996

&roles:R. Hermans E. Fossion C. Ioannides W. Van den Bogaert J. Ghekiere A. L. Baert

R. Hermans (✉) · J. Ghekiere · A. L. Baert Department of Radiology, University Hospitals, Catholic University Leuven, Herestraat 49, B-3000 Leuven, Belgium E. Fossion · C. Ioannides1 Department of Head and Neck Surgery, University Hospitals, Catholic University Leuven, Herestraat 49, B-3000 Leuven, Belgium W. Van den Bogaert Department of Radiation Oncology, University Hospitals, Catholic University Leuven, Herestraat 49, B-3000 Leuven, Belgium Present address: 1 Department of Plastic and Reconstructive Surgery, University College Hospital, London, UK&/fn-block:

A RT I C L E

CT findings in osteoradionecrosis of the mandible

&p.1:Abstract CT scans of ten patients in whom the diagnosis of mandibular osteoradionecrosis was proven pathologically or by clinical followup were reviewed. All ten patients had bony abnormalities (cortical interruptions and loss of spongiosa trabeculation) on the symptomatic side. These were predominantly seen in the body of the mandible (premolar and molar region, eight patients), in some of these cases extending into the retromolar triangle (two patients) or mandibular angle (two patients). In the remaining two patients the abnormalities were in the ramus and angle. The two patients treated with iridium implantation showed localized lingual-sided cortical destruction. Three patients had a pathological fracture. The cortical destruction was buccal-sided in two and both

Introduction Osteoradionecrosis of the mandible is a rare complication of radiation therapy of head and neck tumors. Patients with such lesions are referred for CT, either to estimate the extent of bone destruction before reconstructive surgery, or because tumor recurrence is suspected. The CT differential diagnosis between tumor recurrence and mandibular osteoradionecrosis can be difficult. A retrospective study was untertaken to determine the CT findings in osteoradionecrosis of the mandible.

Material and methods Ten patients were studied, nine men and one woman between 44 and 79 years (Table 1). All patients had been previously irradiated

buccal- and lingual-sided in three of the other five patients. Contralateral bony abnormalities were present in four patients. Soft tissue thickening on the symptomatic side was seen in nine patients. As the bony abnormalities in mandibular osteoradionecrosis are often associated with a soft tissue mass, CT differentiation from tumor recurrence can be diffficult. The association with cortical defects distant from the position of the original tumor (buccal surface or opposite side of mandible) should evoke the possibility of mandibular osteoradionecrosis. &kwd:Key words Jaws, abnormalities · Jaws, CT · Bones, necrosis · Bones, effects of irradiation on · Mandible · Mandibular injuries&bdy:

for head and neck malignancies. All patients had received external beam radiotherapy, but in two patients this was combined with interstitial iridium implantation. The original tumor was located in the tongue in three patients, in the floor in the mouth in two, in the tonsillar region in three and in the parotid gland in two. The total delivered dose ranged from 65 to 80 Gy (mean 71.3 Gy). The CT examinations were undertaken between June 1990 and March 1993. All examinations were performed on a Somatom DRH or Somatom Plus scanner (Siemens, Erlangen, Germany). Adjacent transverse sections of 4 or 5 mm thickness were obtained through the mandible. The field of view was centered on the mandible. In nine patients contrast medium (Telebrix 38, Guerbet, France) was intravenously administered during the examination as a 30-ml bolus, followed by a rapid infusion to a total dose of 100–140 ml. Soft tissue and bone windows werre obtained in all patients. In five patients, examined on the Somatom Plus scanner, additional transverse and coronal 2-mm-thick high-resolution (bone detail) sections were obtained through the pathologic mandibular region.

32

Table 1 Radiation history and CT findings in ten patients with osteoradionecrosis of the mandible (L left; R right; M midline; buccal buccal cortex of mandible; lingual lingual cortex; RT retro-

molar triangle; body mandibular body (premolar/molar region); + enhancing soft tissue mass present)&/tbl.c:&

Patient no.

Interval from irradiation to CT (years)

Patient sex, age (years)

Tumor, side

Irradiation

1

M, 51

Tongue, L

2 3 4 5 6 7 8 9

M, 53 M, 57 M, 44 M, 47 M, 50 F, 79 M, 75 M, 68

Floor of mouth, L Tonsil, L Tonsil, L Floor of mouth, M Parotid, R Parotid, L Tonsil, L Tongue, L

10

M, 57

Tongue, R

External, 46 Gy Iridium, 24 Gy External, 70 Gy External, 72 Gy External, 70 Gy External, 70 Gy External, 80 Gy External, 65 Gy External, 70 Gy External, 50 Gy Iridium, 24 Gy External, 72 Gy

5 4 2 9 2 10 10 5 5 8

Bone lesions; side

Soft tissue

Lingual, body, RT; L

+

Buccal, lingual, body, angle; L>R Lingual, buccal, angle, ramus; L Lingual, buccal, body, angle; R=L Buccal, lingual, body; R Lingual, buccal, body; R Buccal, body, angle, ramus; L Lingual, buccal, body; R»L Lingual, body; R=L

+(L) + +(R) + + + +(R) −

Buccal, lingual, body, RT; L

+

&/tbl.: The examinations were performed between 2 and 10 years after the irradiation (mean 6 years). At the time of the CT examination, nearly all the patients complained of pain. Five patients had an oral ulcer (patients 1, 5, 6, 7, and 9) and two a fistula (patients 2 and 3); in one patient the major symptom was trismus (patient 4). Three patients presented with a fracture of the mandible (patients 2, 8, and 10). In nine patients, the diagnosis of osteoradionecrosis was suspected before the CT examination, either clinically or on the basis of findings on an orthopantogram or both. One patient was referred on the suspicion of tumor recurrence, but clinical reassessment (after the CT scan) by a more experienced examiner suggested osteoradionecrosis. The diagnosis of bone necrosis was pathologically confirmed in seven patients. In the other three patients, the definitive diagnosis was based on the typical clinical history and follow-up (patients 1, 4, and 9; patient 9 had 2 years earlier suffered a pathologically proven contralateral episode of osteoradionecrosis).

Results All ten patients had mandibular abnormalities on the symptomatic side. In four patients this was the side of the original tumor. One patient originally had a midline tumor of the floor of the mouth, but developed lateralized necrosis after dental surgery. In one patient both sides of the mandible were affected, although abnormalities were more pronounced on the side of the original tumor. In two patients both sides of the mandible were affected to a similar extent. In one of these patients, the surrounding soft tissues enhanced only on the symptomatic side. The other patient, showing no soft tissue enhancement, first developed necrosis on the side opposite to the original tumor, probably due to progressive caries. In two patients, the present bone abnormalities were on the contralateral side to the original tumor, but these patients had been previously treated for ipsilateral osteoradionecrosis. The bone abnormalities were predominantly seen in the body of the mandible (premolar and molar region,

eight patients) (Fig. 1), in some cases extending into the region of the angle (two patients) or retromolar triangle (two patients). Most patients showed more or less pronounced loss of trabeculation of the spongiosa. Interruptions in the cortical margins of the mandible were seen in all patients. Three patients with a fracture of the mandible had, by definition, interruption of both the lingual and buccal cortex. Of the other patients, cortical interruptions were seen on the buccal side of the mandible in two, and on both the buccal and lingual side in three (Fig. 2); in one of these last patients, the cortical interruption was predominantly on the buccal side. In two patients cortical destruction was seen only on the lingual side of the mandible; both had been treated with iridium needle implantation (Fig. 3). Bone fragmentation was seen in eight patients. In six patients, only a few small pieces of bone were seen to be separated from the remainder of the mandible; in two patients, several larger pieces of bone appeared separated (Fig. 4). These fragmented pieces of bone nearly always seemed of cortical origin. One patient, with a large buccal cortical defect, showed a predominantly sclerotic alteration of the hemimandible. Bone sclerosis was not seen in the other patients. In eight patients, gas bubbles were visible in the pathologic bone (Figs. 1, 3, 5). In six of these patients, this was associated with some form of dehiscence of the mucous membrane (fistula, ulcer, or gingival perforation). In the two other patients (both presenting with a pathologic fracture) no mention of such mucosal dehiscence was found in the clinical notes. In four patients, bone abnormalities were seen in the contralateral side of the mandible (Fig. 6). These alterations appeared similar to the symptomatic side in two patients, somewhat less extensive in one; in another the bone alterations were limited to some rarefaction of the spongiotic trabeculation.

33

Fig. 1A, B Patient 5: Axial enhanced CT scan. A Bone window, B soft tissue window. Lytic defect in the premolar region of the mandibular body. Lysis and large defect of the buccal side of the mandibular cortex (small white arrowheads); thinning and focal interruption on the lingual side (large white arrowheads). Some small intraosseous air bubbles. The associated enhancing soft tissue mass (black arrowheads) makes differential diagnosis from recurrent tumor difficult&ig.c:/f Fig. 2A–C Patient 4. Enhanced CT scan. A, B Coronal, bone window, C axial, soft tissue window. Bilateral osteolytic changes in the body of the mandible, extending into the region of the mandibular angle; both buccal (large black arrowheads) and lingual (small black arrowheads) cortical interruptions are seen. Note associated thickening of the tissues in the masticatory region (white arrowheads), due to inflammation&ig.c:/f

In nine patients, soft tissue swelling and enhancement were seen around the site of bony alterations. This soft tissue thickening was limited to a fusiform swelling around the fracture site in two patients; in the third patient with a fracture, there was also obliteration of the fatty planes in the infratemporal region and enhancement of the surrounding muscles. The extent of soft tissue swelling and enhancement was quite variable in the patients without a fracture: in two it was limited to the neighbourhood of the bony destruction, in other patients

it was quite extensive. In one patient, apart from enhancement of the medullary region, no extraosseous contrast uptake was visible. No soft tissue swelling was seen around the contralateral mandibular alterations. Orthopantograms were available for six patients. In one, no bony abnormalities were visible, whereas CT showed a lingual cortical defect, loss of spongiosa trabeculation, and an associated soft tissue mass. In the other five cases, bony alterations (usually inhomogeneous and lytic) were seen. The exact location and extension of the bone destruction could be better evaluated on CT. The soft tissue changes were not visible on orthopantomography.

Discussion The efficacy of radiotherapy in the treatment of tumors of the head and neck has been proved extensively. To control a potentially lethal tumor, some limited damage to the surrounding normal tissues must be accepted, but the integrity and viability of these tissues should not be disturbed. One of the most severe complications of radiotherapy in the head and neck region is osteoradionecrosis of the

34

Fig. 3A, B Patient 1. Enhanced axial CT scan. A Bone window, B soft tissue window. Purely lingual cortical interruption, with focal loss of spongiosal trabeculation (black arrows); bone fragmentation (white arrowhead). Associated soft tissue swelling (black arrowheads). The patient had been treated with iridium implantation&ig.c:/f Fig. 4A, B Patient 9. Axial high resolution CT. Bone window; A is a few millimeters cranial to B. A Large lingual cortical defect on the left side; some small loose fragments are visible (small arrowheads). Some loss of spongiosal trabeculation (arrow). Reconstruction of the right hemimandible after resection of a previously radionecrotic part of the mandible by a free revascularized osseous flap (iliac crest) as an onlay bone graft (large arrowheads). B Extensive bone destruction under the osseous flap due to progressive radiation necrosis. Bone fragmentation (small arrowheads) and loss of the spongiosal trabeculation (arrows)&ig.c:/f

Fig. 5 Patient 6. Axial high-resolution CT scan. Bone window. Patient presented with pain and mucosal dehiscence. Extensive resorption of spongiosa in right mandibular body (compare to opposite side). Destruction of both lingual and buccal cortex with pathologic fracture (arrowheads). Note intraosseous air bubbles; pathological analysis revealed necrotic bone with signs of osteomyelitis&ig.c:/f

Fig. 6A, B Patient 2. Axial enhanced CT scan. A High-resolution bone window, B soft tissue window. Pathological fracture on the left side, with broad interruption of the lingual and buccal cortex. There is thickening and enhancement of the surrounding soft tissues. Note similar type of bone destruction on the contralateral side (resorption of the spongiosal trabeculae and destruction of the buccal cortex); there was no soft tissue enhancement on the right side

35

mandible. The reported incidence varies between 0 and 37.5% [1]. Respiration, deglutition, and speech can be severely compromised by necrosis of the mandible [2]. The mandible is a compact bone with a superficial location. Its high calcium content may lead to scattering of radiation, causing relatively more damage to the endosteal cells, blood vessels, and periosteum [1, 2]. Many risk factors for the development of osteoradionecrosis have been defined: advancing stage of the primary lesion, lesions extending to the mucosa overlying the mandible, presence of dental disease, poor oral hygiene, high doses of radiation, larger radiation fields. Neck dissection has also been mentioned as risk factor [2, 4, 5], especially with extensive deperiostation of the mandible. Preventive dental procedures before irradiation and strict oral hygiene following radiation therapy reduce the risk of necrosis substantially [2, 6]. The risk of osteonecrosis is greatest when an interstitial implant is used as the sole radiation source: a high dose of radiation is delivered (in a short time interval) to the bone when the implant is close to it. The extent of the bone necrosis is smaller, however, because of the rapid fall in radiation dose with increasing distance from the implant [2]. Ulceration or necrosis of the mucous membrane with exposure of necrotic bone is typically seen in mandibular osteoradionecrosis. This can be accompanied by pain, dysesthesia, or anesthesia when the inferior alveolar nerve is irritated or finally damaged by the process. Being a slow but actively progressive disease, osteoradionecrosis may lead to pathologic fracture, fistulization, and infection [7]. The pathogenesis of mandibular osteoradionecrosis is not fully understood. Its classic triad was defined as radiation-trauma-infection [8], in which trauma (most common sources are tooth extraction and sharp bony ridges after inadequate alveolectomies) acts as a portal of entry for bacteria in the underlying bone. As the bone has lost its ability to wall off the infection due to vascular injury and cellular damage, the infection can progress rapidly through the mandible, producing radiation osteomyelitis. According to Dambrain [5], infection is a necessary factor in the pathogenesis, preventing repair of osteoclastic damage and disturbing the homeostatic role of the osteocytes. The same author describes the bone-dissolving capacity of chemical components in saliva and pus on irradiated bone covered by a dehiscent mucous membrane. However, chronically nonhealing bone is seen in irradiated patients without extensive suppuration. Necrosis is also seen in patients without a history of trauma (“spontaneous osteoradionecrosis”). Furthermore, no micro-organisms could be cultured in the “deep” bone obtained by en bloc resection of osteoradionecrotic mandibles; organisms were only found on the surface exposed to the oral environment [9]

This supports another concept of the pathogenesis of osteoradionecrosis as proposed by Marx [9]: that irradiation produces a hypoxic, hypocellular, and hypovascular tissue which is unable to remodel tissue loss, thus leading to breakdown. This radiation tissue injury is continuous [10]. Spontaneous osteoradionecrosis is seen with high-dose radiation; it represents greater overall death of normal tissue elements, which pass after initial attempts at repair through the stages of hypovascularity and fibrosis into necrosis. This usually occurs within the first 2 years after radiation. Trauma is often associated with the appearance of osteoradionecrosis as it creates a demand for tissue repair beyond the capabilities of the irradiated tissue. Trauma-induced osteoradionecrosis shows a bimodal distribution in time. A first peak is seen in the first 3 months after radiation therapy; these cases are always related to surgical insult shortly before or during the irradiation (tooth removal or mandibulotomy for tumor resection closely followed by radiotherapy). As the risk of developing necrosis after a particular trauma increases with time, the second rise in the incidence of trauma-induced osteoradionecrosis begins 2 years after irradiation, shows a broad peak at 5 years, and persists for many years after [10]. However, most irradiated patients will never develop mandibular necrosis during their lifetime. Marx [9] believes that osteoradionecrosis is a problem of tissue homeostasis and wound healing, not of infection. After radiotherapy on the mandible, transcortical vascularization via the periosteum and muscular insertions becomes very important, as the inferior alveolar artery becomes occluded due to intimal fibrosis and thrombosis [11]. In osteoradionecrosis, the mental and angular regions of the mandible are relatively spared because of their muscular insertions; the buccal cortex seems to be more vulnerable. Buccal damage to the periosteum is most frequently brought about by dental surgery disrupting the transcortical vascularization of the mandible. Furthermore, there is evidence in dogs that the buccal cortex in the premolar, molar, and retromolar regions is the last part of the mandible to be revascularized after blockage of the inferior alveolar artery [12]. These pathologic findings are supported by the CT findings in our cases: the buccal cortex in the premolar, molar, and retromolar regions was the most frequent site of cortical destruction seen. The two patients showing only lingual cortex destruction were treated with brachytherapy. Differentiation between buccal and lingual sided bone requires an occlusal X-ray. Intraosseous gas has been described as pathognomonic of osteomyelitis [13]. Bone infection was histologically present in only two of our patients. As most of these patients have some form of soft tissue dehiscence, air could become trapped between exposed bone fragments. Cortical bone fragmentation was seen in the majority of our patients; in most cases it consisted of separation

36

of only a few small pieces of bone. Sequestra are notably unusual in mandibular osteoradionecrosis [14]. Nearly all our patients showed the presence of a surrounding inflammatory soft tissue mass. Absence of a soft tissue mass in the case of mandibular destruction favors the diagnosis of radiation necrosis [1]. However, our study shows that on cross-sectional imaging soft tissue swelling will often be visible in advanced osteoradionecrosis; this may complicate the differentiation from tumor recurrence. Mandibular invasion by a malignant squamous cell tumor usually occurs in the alveolar process or through the lingual cortex; destruction of the buccal cortex (often involved in this study) by tumors is less frequently seen in Belgium, since carcinomas of the cheek mucosa are relatively rare here compared to countries like India where betel nut and tobacco chewing are commonly practised. In some patients, there was quite extensive mandibular destruction, associated with a rather small soft tissue swelling; in tumors, usually the opposite is true. The detection of similar bone alterations in the contralateral side of the mandible raises the possibility of osteoradionecrosis. Contralateral osteoradionecrosis is seen because most patients are irradiated by opposed fields to reduce the doses on the entry sites. In some patients the bone destruction and associated soft tissue mass was situated at a site other then that of the primary tumor, making tumor recurrence less probable. In most patients there was rather a long interval between the initial tumor and the development of mandibular problems; most tumor recurrences appear within the first 2 years after treatment [15]. As the diagnosis can often be suspected on clinical grounds, radiology is used for confirmation and evalua-

tion of the extent of the bone involvement. We consider that the localization and extent of the bone destruction can be better evaluated with CT than with conventional (occlusal) or panoramic films. However, in most cases plain films will give sufficient information for patient management, and on the basis of the clinical and plain imaging findings a decision can be made to treat the patient conservatively or surgically. The treatment of osteoradionecrosis is variable. A mucoperiosteal defect with some superficial erosion of the exposed bone, representing the first stage of necrosis, is treated by removing any possible initiating trauma (e.g., adjustment of prosthesis), superficial curettage and rinsing with a 0.9% saline or an antiseptic solution; spontaneous healing may be accomplished in these cases. When cortical defects and medullary extension are present, antibiotics are administered, with thorough debridement of the wound and irrigation of the sinus tracts and open tissues. Hyperbaric oxygen, which enhances wound healing by promoting neovascularization, is recommended as soon as a surgical procedure is needed in irradiated tissue [10]. Necrotic bone needs to be removed as this is not affected by hyperbaric oxygen. Bone surgery is also indicated in the case of a pathological fracture or when there is extensive involvement of the mandible. Reconstruction of the surgical defect may require the use of bone grafts [7]. A free bone graft may be used in combination with a vascularized soft tissue flap and associated with pre- and postoperative sessions of hyperbaric oxygen treatment. Another option is to reconstruct the defect with a revascularized myo-osseous flap, in which case combination with the expensive hyperbaric oxygen treatment does not seem necessary [1, 6].

References 1. Dalinka MK, Mazzeo VP. Complications of radiation therapy. Rev Diagn Radiol 1985; 23: 235–267 2. Murray CG, Herson J, Daly TE, Zimmerman S. Radiation necrosis of the mandible: a 10 year study. Part I. Factors influencing the onset of necrosis. Int J Radiat Oncol Biol Phys 1980; 6: 543–548 3. Bünger B. Die Strahlenosteomyelitis des Unterkiefers. Laryngorhinootologie 1990; 69: 316–319 4. Rankow RM, Weissman B. Osteoradionecrosis of the mandible. Ann Otol 1971; 80: 603–610 5. Dambrain R. La pathogénie de l’ostéoradionécrose. Rev Stomatol Chir Maxillofac 1993; 94: 140–147 6. Pyykönen H, Malmström M, Oikarinen VJ, Salmo M, Vehkalahti M. Late effect of radiation therapy of tongue and floor-of-mouth cancer on the dentition, saliva secretion, mucous membranes

7.

8. 9. 10.

11.

12.

and the lower jaw. Int J Oral Maxillofac Surg 1986; 15: 401–409 Epstein JB, Wong FLW, StevensonMoore P. Osteoradionecrosis: clinical experience and a proposal for classification. J Oral Maxillofac Surg 1987; 45: 104–110 Meyer I. Infectious diseases of the jaws. J Oral Surg 1970; 28: 17–27 Marx RE. Osteoradionecrosis: a new concept of its pathophysiology. J Oral Maxillofac Surg 1983; 41: 283–288 Marx RE, Johnson RP. Studies in the radiobiology of osteoradionecrosis and their clinical significance. Oral Surg 1987; 64: 379–390 Bras J, De Jonghe HKT, Van Merkesteyn JPR. Osteoradionecrosis of the mandible: pathogenesis. Am J Otolaryngol 1990; 11: 244–250 Hellem S, Östrup LT. Normal and retrograde blood supply to the body of the mandible in the dog. II. The role

13.

14.

15.

16.

played by periosteomedullary and symphyseal anastomoses. Int J Oral Surg 1981; 10: 31–42 Merine D, Fishman EK, Magid D. CT detection of sacral osteomyelitis associated with pelvic abscesses. J Comput Assist Tomogr 1988; 12: 118–121 Tartaglino LM, Rao VM, Markiewicz DA. Imaging of radiation changes in the head and neck. Semin Roentgenol 1994; 29: 81–91 Mancuso AA, Hanafee WN. Computed tomography and magnetic resonance imaging of the head and neck. 2nd ed. Baltimore: Williams & Wilkins, 1985; 371 Fossion E, Boeckx W, Jacobs D, Ioannides C, Vrielinck L. La reconstruction microchirurgicale de la mandibule irradiée par le lambeau circonflexe iliaque profond. Ann Chir Plast Esthét 1992; 37: 246–251