This preliminary retrospective study evaluates the diagnostic value of cone-beam computed tomography (CBCT), as a potential standard preoperative procedure, in assessing mandibular invasion by oral squamous cell carcinoma (OSCC) compared with conventional preoperative panoramic radiography (PR), magnetic resonance imaging (MRI) and histological examination of the resection specimen (the golden standard). Between September 2006 and September 2009, 23 patients with histology proven primary OSCC, adjacent to or fixed to the mandible were included. The tumours were classified into four groups, ranging from no bone invasion to evident bone invasion. Sensitivity and specificity for PR were 55% (95% CI [0.350;0.619]) and 92% (95% CI 0.737;0.984]), respectively, both were significantly lower than the 91% (95% CI [0.740;0.909]) and 100% (95% CI [0.845;1]), respectively, for CBCT. MRI showed 82% sensitivity (95% CI [0.608;0.941]) and 67% specificity (95% CI [0.474;0.779]). CBCT has the potential to become a new diagnostic tool in the OSCC screening procedure to predict mandibular invasion or erosion, but its value may be limited by its relatively low sensitivity. A prospective study will start on 64 patients ( α = 0.05; power 0.8; effect size 0.5) to improve these results statistically.
Oral squamous cell carcinomas (OSCC) may be adjacent or fixed to the mandible. The treatment of choice for these carcinomas is wide excision of the tumour in the soft tissue ‘en bloc’ with the involved bone and postoperative radiotherapy depending on the final histopathology results. The preoperative extent of mandibular bone involvement is often difficult to assess, because no single imaging method is able to predict mandibular invasion and the extent of invasion 100% reliably . This can pose a dilemma for the surgeon regarding choice and design of the mandibular resection. To increase the reliability of the imaging, several combinations of imaging methods have been proposed. Recently, a diagnostic algorithm has been designed for the prediction of mandibular invasion in patients with OSCC adjacent or fixed to the mandible. This is either multislice computed tomography (MSCT) or magnetic resonance imaging (MRI), followed by a bone SPECT in case the first scan is negative. This diagnostic algorithm accurately predicts mandibular invasion in 85% of patients, without yielding false-negative results. Application of this diagnostic algorithm considerably reduces the number of unnecessary mandibular resections .
Cone-beam computed tomography (CBCT) is a relatively new, promising technology compared with traditional MSCT scanning, but has not been routinely used in staging procedures in patients with an OSCC adjacent to or fixed to the mandible . CBCT uses a different type of acquisition image compared with traditional MSCT . The X-ray source produces a cone-shaped X-ray beam. This makes it possible to capture the image in one sweep, instead of capturing every individual slice separately, as in MSCT. One major advantage is that the patient is scanned in an upright position in CBCT, the soft tissues are not distorted due to gravity, which is the case when a patient is scanned in the supine position in a conventional MSCT . Other advantages are the relatively low radiation dose, reduced metal artifacts compared with CT and the relatively low costs . Disadvantages of CBCT images are the weak soft-tissue contrast, movement artifacts and limited detector size . A contrast-enhanced technique in combination with CBCT has not been described but could lead to improved imaging results with a minimized radiation dose. Although CBCT has many advantages, this approach has not led to another diagnostic algorithm . The aim of this study is to determine the value of CBCT-scanning for the evaluation of mandibular invasion by OSCC, adjacent to or fixed to the mandible by comparing CBCT with panoramic radiography (PR), MRI and histopathology of the resection specimen.
Patients and methods
Between September 2006 and September 2009, 23 patients with proven primary OSCC, adjacent to or fixed to the mandible, examined by preoperative PR, CBCT and MRI, were included in the study. Two senior surgeons (FD, MM), specializing in head and neck oncology, independently assessed the presence or absence of mandibular invasion, using the TNM classification system 1 , by clinical examination prior to surgery. The type of mandibular resection chosen was based on the pre-treatment diagnostic results of clinical examination, PR, CBCT and MRI.
Digital panoramic radiographs were taken using the ProMax (PLANMECA Oy, Helsinki, Finland). Periapical radiographs were taken tangential to the area of suspected invasion with the Oralix 65S (Philips ® , 220 V, 5 A, 50 Hz/60 Hz). The CBCT apparatus used was the Classic I-CAT scanner (Imaging Sciences International, Inc, Hatfield, PA, USA). All patients were scanned sitting, with their heads in a natural position .
MRI was performed using a 1.5 T MR whole body system (Vision, Siemens Medical Systems, Erlangen, Germany) with a CP-neck-array coil, according to a standard protocol. MRI data acquisition consisted of fast spin echo (SE) T2-weighted images in the axial and coronal plane and SE T1-weighted images obtained before (axial plane) and after (axial and coronal planes) administration of contrast agent (gadopentetate dimeglumine, 0.2 ml/kg). All images were acquired with 3 mm slice thickness.
All patients underwent neck dissection and radical removal of the tumour. The involved bone was resected in accordance with the protocol of the Dutch Cooperative Head and Neck Group ( www.nwhht.nl ), which recommends marginal resection when the tumour is adjacent to or fixed to the mandible and segmental resection in the case of mandibular invasion.
After surgery, the radiologic records of the patients, were analyzed by two senior surgeons (FD, MM) retrospectively, for the current study. The tumours were classified into four groups. Group A showed no bone involvement, B only bone erosion, C slight bone invasion and D evident bone invasion.
The resection specimens were cut into 3 mm thick bucco-lingual slices with a water-cooled engine-driven circular diamond-coated saw blade, decalcified in 10% formic acid, processed in paraffin wax, sectioned at 5 μm and stained with haematoxylin and eosin. Absence of mandibular invasion or erosion was defined as a continuous periosteal layer separating the tumour from the bone. Cortical erosion was defined as replacement of bone by an advancing tumour front, without tumour invasion into cancellous spaces, dental canal or periodontal ligament space. Mandibular invasion through the cortex into the medullar bone was defined as diffuse growth of tumour into cancellous bone, dental canal and, if present, periodontal ligament space.
Using the histopathological results of the bone resection specimens (groups A, B, C, D) as the golden standard in the last column of Table 1 , the positive predictive values and test efficiency mentioned below was calculated.
|Patient||Location||Edentulous/dentulous||Initial stage (preoperative) = l–4 *||PR = A/B/C/D **||CBCT scan = A/B/C/D **||MRI = A/B/C/D **||Bone resections I–III ***||Histological results of bone resection|
|1||Lower alveolar process right||Dentulous||1||A||A||A||II||No invasion|
|2||Retromolor area, left site||Edentulous||2||A||A||A||II||No invasion|
|3||Anterior floor of mouth, right site||Edentulous||4||D||D||B||III||Cortex|
|4||Anterior floor of mouth, right site||Edentulous||4||D||D||C||III||Medulla|
|5||Retromolorarea, right site||Edentulous||2||B||A||A||II||No invasion|
|6||Retromolorarea, right site||Edentulous||4||B||D||B||II||Medulla|
|7||Retromolararea, right site||Edentulous||2||A||B||B||II||Cortex|
|8||Retromolararea, right site||Dentulous||4||D||D||C||III||Medulla|
|9||Anterior floor of mouth, right site||Edentulous||3||A||A||A||II||No invasion|
|10||Ower alveolar process, right site||Dentulous||1||A||B||A||II||Cortex|
|11||Ower alveolar process, right site||Edentulous||3||A||A||A||II||No invasion|
|12||Ower alveolar process, right site||Edentulous||3||C||C||C||III||Medulla|
|13||Anterior floor of mouth, right site||Edentulous||2||A||A||A||II||No invasion|
|14||Ower alveolar process, right site||Edentulous||2||A||A||B||II||No invasion|
|I5||Anterior floor of mouth, right site||Dentulous||2||A||A||C||II||No invasion|
|16||Anterior floor of mouth, right site||Dentulous||2||A||A||A||II||No invasion|
|17||Lower alveolar process, left site||Dentulous||2||A||B||A||II||Cortex|
|18||Anterior floor of mouth, left site||Dentulous||4||A||A||B||II||Cortex|
|19||Anterior floor of mouth||Dentulous||4||C||C||C||III||Medulla|
|20||Retromolar area, right site||Dentulous||4||A||B||B||III||Medulla|
|21||Retromolararea, right site||Dentulous||2||A||A||A||II||No invasion|
|22||Retromolar area, right site||Edentulous||2||A||A||B||II||No invasion|
|23||Anterior floor of mouth, left site||Dentulous||2||A||A||B||II||No invasion|