| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, University of North Carolina School of Medicine, 3324
Old Infirmary CD# 7510, Chapel Hill, NC 27599-7510.
2
Department of Surgery, University of North Carolina School of Medicine, Chapel
Hill, NC 27599-7510.
3
Lineberger Comprehensive Cancer Center, University of North Carolina School of
Medicine, Chapel Hill, NC 27599-7510.
4
Present address: Department of Radiology, University of Michigan, 1500 E.
Medical Center Dr., UH B2B311-0030, Ann Arbor, MI 48109-0030.
5
Department of Pathology, University of North Carolina School of Medicine,
Chapel Hill, NC 27599-7510.
Received June 27, 2000;
accepted after revision December 15, 2000.
Presented at the annual meeting of the American Roentgen Ray Society,
Washington, DC, May 2000.
Abstract
 Top Abstract IntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Forty-nine patients who had squamous cell carcinoma of the oral cavity that was clinically fixed to the mandible were treated with mandibulectomy. All patients underwent contrast-enhanced CT (contiguous 3-mm-thick sections) through the primary site before surgery. All studies were reconstructed with bone algorithm. These studies were retrospectively reviewed by a neuroradiologist for evidence of mandibular invasion. The imaging results were compared with the histologic findings in all cases.
RESULTS. CT correctly revealed 25 of 26 cases with mandibular invasion. CT correctly excluded mandibular invasion in 20 of 23 cases without invasion. The diagnostic accuracy of CT for detecting mandibular invasion was as follows: sensitivity, 96%; specificity, 87%; positive predictive value, 89%; and negative predictive value, 95%.
CONCLUSION. Thin-section (3-mm) CT reconstructed with bone algorithm is an accurate technique to detect mandibular involvement by squamous cell carcinoma of the oral cavity.
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
|
|
|
A marginal mandibulectomy refers to partial excision of the superior portion of the mandible in the vertical plane. In general, one cortical surface and a portion of the underlying medullary cavity are excised. The inferior half of the body of the mandible is intact, thereby preserving mandibular continuity and permitting the patient to chew. Such an excision is warranted for carcinomas abutting the mandible without evidence of direct bony erosion. In contradistinction, when a segmental mandibulectomy is performed, an entire through-and-through segment of mandible is resected, resulting in mandibular discontinuity. Segmental mandibulectomy requires a major reconstructive procedure for cosmetic reasons and functional purposes.
Various imaging modalities have been used to predict mandibular invasion by squamous cell carcinoma of the oral cavity when attempting to determine whether patients are candidates for mandibular-sparing procedures [5,6,7,8,9,10,11,12,13]. CT is the imaging technique that is most commonly used to determine mandibular involvement. The diagnostic accuracy of pre-operative CT to detect mandibular invasion has been evaluated with varied results [7, 8, 11,12,13]. In 1986, Close et al. [8] reported a sensitivity of 100%, specificity of 97%, positive predictive value of 92%, and negative predictive value of 100%. However, these impressive results have not been reproduced in subsequent studies [7, 11,12,13]. After the report of Close et al., Shaha [12] reported a diagnostic accuracy of 68% and concluded that clinical examination was superior to CT for assessing for mandibular involvement. Brown et al. [11] claimed a 28% false-negative rate for CT for detecting mandibular invasion and argued that the predictability and reliability of CT were "disappointing." In 2000, Lane et al. [14] reported a sensitivity of 50% with a negative predictive value of 61% and suggested that CT was an inaccurate method for detecting mandibular invasion.
This variability may be the result of previous studies having been performed with an inconsistent technique that is not optimal. The purpose of this study was to reevaluate the ability of CT to detect mandibular invasion by squamous cell carcinoma of the oral cavity using 3-mm-thick axial sections reconstructed with bone algorithms.
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
The study was limited to patients who had tumors that were revealed by clinical examination to be fixed to the mandible and who were surgically treated with either a marginal or segmental mandibulectomy. As a result, there was pathologic correlation of mandibular invasion with all CT studies. Tumors that are freely mobile to bimanual palpation have little likelihood of mandibular invasion and are treated with local excision without mandibulectomy. Patients with such tumors were not included in our study because the mandible would not be resected and there would be no pathologic correlation.
The primary sites of squamous cell carcinoma in our patients were as follows: floor of mouth (n = 25), gingival and buccal mucosa (n = 10), alveolar ridge (n = 4), and retromolar trigone (n = 10). No primary mobile (oral) tongue carcinomas were included in our study because of the low likelihood of tumors localized to this subsite being fixed to the mandible and treated with mandibulectomy. Of the 49 patients, 34 patients underwent marginal mandibulectomies and 15 patients underwent segmental mandibulectomies.
In our study, 36 of the 49 patients had no treatment before CT and subsequent mandibulectomy, whereas 13 of the 49 had undergone the following forms of treatment: radiation therapy alone (n = 4), local excision alone (n = 3), local excision and radiation therapy (n = 5), and chemotherapy and radiation therapy (n = 1). None of the patients included in our study had undergone mandibular surgery before CT.
All patients underwent axial contrast-enhanced CT (3-mm-thick contiguous sections) performed from the skull base to the thoracic inlet. CT studies were re-angled to avoid dental amalgam when necessary. In patients without dental amalgam, the study was performed from the skull base to the thoracic inlet, with the plane of the gantry angle parallel to the hard palate. In patients with dental amalgam, the study was performed in two parts. The first portion of the study was performed from the skull base to the maxillary alveolar ridge with the plane of the gantry approximating the hard palate. The gantry angle was then changed to approximate the angle of the mandible, and images were obtained from the mandibular alveolar ridge to the thoracic inlet. This technique substantially reduced the amount of streak artifact that could potentially obscure visualization of the oral cavity.
Only axial images were evaluated, because coronal images are not routinely obtained for squamous cell carcinoma of the oral cavity at our institution. All studies were reconstructed with soft-tissue and bone algorithms. The bone algorithm settings were a width of 3500 H and level of 700 H. Specialized dental software with postprocessing algorithms was not used.
The average time period between CT and surgery was 14.3 days. One patient underwent surgery 61 days after CT. With the exclusion of this patient, the range between CT and surgery was 1-36 days.
The CT studies were reviewed by a neuroradiologist who was aware of the primary site but unaware of the histologic findings. The radiologist was asked to determine whether there was evidence of mandibular erosion. The criterion for mandibular erosion was absence of the cortex of the mandible adjacent to an abnormal soft-tissue mass.
Mandibulectomy specimens were fixed in 10% formalin solution and decalcified for a period of 1-4 days. Specimens were cut and grossly evaluated for bone invasion. Microscopic sections of mandible were not taken when no definite evidence bone invasion was seen at surgery. In mandibles suspicious for invasion or with apparent gross invasion seen at surgery, representative sections of the tumor—mandible interface and the mandibular bony margins were submitted. One pathologist who was unaware of the imaging findings reviewed the histologic material in all cases for evidence of bone invasion. No histologic distinction was made between cortical erosive and infiltrative patterns [15].
All imaging studies were correlated with histologic evidence of mandibular invasion. Statistical analysis included calculation of sensitivity, specificity, and the positive and negative predictive values of the ability of CT to detect mandibular invasion.
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
|
|
|
|
There were three false-positive studies and one false-negative study (Figs. 4A,4B and 5A,5B). Of the three false-positive studies, two consisted of floor-of-mouth tumors and one arose from the alveolar ridge (Fig. 5A,5B). Both floor-of-mouth lesions were recurrent tumors that had undergone treatment previously. The patient with the alveolar ridge carcinoma presented for the first time in our study. One patient with a primary floor-of-mouth carcinoma was initially treated with combined chemotherapy and radiation therapy. This patient had a recurrent anterior floor-of-mouth tumor 12 months after completion of therapy and was treated surgically with segmental mandibulectomy. The other patient had a prior oral tongue carcinoma that was initially treated with hemiglossectomy followed by radiation therapy. This patient's tumor recurred in the floor of the mouth 8 months after completion of therapy and was treated surgically with segmental mandibulectomy.
|
|
|
|
The one false-negative study in our series was of a patient with a carcinoma arising from the lingual surface of the right inferior alveolar ridge who had not previously undergone treatment. The tumor was a superficial lesion that measured 2 cm in its largest dimension. On initial interpretation, a small cortical defect was noted but, because no adjacent soft-tissue mass was initially identified, it did not meet our criterion for mandibular erosion and the defect was attributed to periodontal disease (Fig. 4A,4B). In retrospect, there may be slightly more soft tissue in the anterior floor of the mouth associated with a cortical defect along the lingual cortex of the mandible. Thus, it is possible that the misclassification of this patient's case was the result of an interpretative mistake rather than an inherent limitation of CT. Because of the fixation to the mandible on clinical examination, the patient was treated with a marginal mandibulectomy. Histologic examination revealed a small focus of cortical invasion.
In our study, the diagnostic accuracy of CT did not appear to be hindered by dental artifact. As evidenced by our high diagnostic accuracy, the technique of angling the gantry to avoid dental amalgam appeared to successfully reduce the amount of streak artifact caused by dental amalgam. None of the misclassifications in our series could be attributed to dental artifact (Figs. 4A,4B and 5A,5B).
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
If no imaging evidence of mandibular invasion is seen with a tumor that is clinically fixed to the mandible, the surgeon usually has two options. If, during surgery, the tumor is easily separable from an intact cortical plate, the surgeon may decide to remove the tumor and completely spare the mandible. Patients who underwent this procedure were excluded from our study because no histologic correlation with mandibular invasion was possible. If the tumor is fixed to the cortical plate and unable to be easily separated, the surgeon may perform a marginal (rim or shave) mandibulectomy, which entails resecting the cortical plate of bone adjacent to the tumor. If there is evidence of bony invasion, the standard procedure is segmental mandibulectomy. Patients having this procedure require reconstructive surgery to correct the resultant cosmetic and functional deficits.
CT has been shown to be an important adjunct for evaluating patients with squamous cell carcinoma of the oral cavity. CT may provide information regarding staging of the primary site and lymph nodes [3, 16, 17]. However, the diagnostic accuracy of CT in detecting mandibular invasion is variable. Lack of reproducibility has resulted in persistent doubts being raised about the true diagnostic accuracy of CT for detecting mandibular invasion and has caused several investigators to state that CT is an inaccurate technique that is not helpful for preoperative evaluation of mandibular invasion [7, 11,12,13,14]. These discordant results have also resulted in the evaluation of other modalities to improve the ability to preoperatively detect mandibular invasion [5,6,7].
The results of our investigation compare favorably with those of Close et al. [18] and are superior to more recent studies that have cast doubt on the ability of CT to accurately predict involvement of the mandible by squamous cell carcinoma of the oral cavity [7, 8, 11,12,13,14]. We believe that we obtained improved diagnostic accuracy in our study because we used a better imaging technique. All of the patients in our study had tumors that were clinically fixed to the mandible and underwent CT using contiguous 3-mm-thick sections that were reconstructed with soft-tissue and bone algorithms. It is not known whether the patient population in previous studies evaluating the ability of CT to predict mandibular erosion had tumors that were clinically fixed to the mandible. Brown et al. [18] (false-positive rate = 28%) and Curran et al. [7] (specificity = 57%; positive predictive value = 73%) acquired their images using 4- to 5-mm-thick sections and did not routinely evaluate the mandible with bone algorithms. Shaha [12] (diagnostic accuracy = 68%) and Bahadur [13] (false-negative rate = 28%) did not describe their CT techniques. Lane et al. [14] obtained 5-mm-thick sections and did not reconstruct the studies in bone algorithm when evaluating for bone invasion. Close et al. [8] obtained 5-mm-thick contiguous sections. However, if there was suspicion of bone destruction, the studies were repeated with overlapping 5-mm-thick sections obtained every 3 mm; 1.5-mm contiguous sections were occasionally obtained. All studies in the Close et al. series were evaluated in bone and soft-tissue settings.
Our results, along with the current literature, make it clear that a high-quality imaging technique is necessary for CT to be a reliable modality for detecting mandibular invasion. We think that the slice thickness should not exceed 3 mm and that the mandible should be reconstructed using bone algorithm. We used a width of 3500 H and a level of 700 H. The diagnostic accuracy for detecting mandibular invasion may also be improved using volumetric acquisition. It is possible that data obtained at 3-mm-thick slices reconstructed at 1- to 2-mm intervals may improve the ability of CT to detect mandibular invasion. The diagnostic accuracy in detecting mandibular invasion also may be improved with acquisition of thinner slice thickness. Helical acquisition using new multidetector CT units now permits even faster acquisition of studies performed with thin sections (contiguous 1-2 mm) over a large area [18]. Multidetector units allow thin-section imaging to be performed from the skull base to the thoracic inlet with short acquisition times. The potential for improving diagnostic accuracy using these new techniques warrants future investigation.
Direct mandibular involvement by squamous cell carcinoma of the oral cavity occurs in two main patterns: erosive and infiltrative. Erosive involvement occurs when cortical bone recedes before a pushing tumor border [15]. In this form of involvement, there is often a scalloped excavation of underlying medullary bone. In the infiltrative pattern of tumor involvement, cancer diffusely spreads throughout the cancellous, medullary bone [15, 19]. We do not attempt to identify the different types of patterns in pathology or by radiology because bony invasion caused by either type is treated with some form of mandibulectomy at our institution.
Neoplastic invasion of the medullary cavity of the nonirradiated mandible often occurs through the occlusive surface of the alveolar ridge and is facilitated by cortical defects in the edentulous mandible [20, 21]. This is one of the most difficult areas to evaluate with axial CT and this potential pitfall was apparent in our data. Two of our four misclassifications (false-positive = 1, false-negative = 1) were tumors that arose from the alveolar ridge and had not previously been subject to treatment (Figs. 3A,3B and 4A,4B). When all four alveolar ridge tumors were excluded from our study, our diagnostic accuracy increased as follows: sensitivity of 100%, specificity of 91%, positive predictive value of 92%, and negative predictive value of 100%. Previous studies have suggested that intraoral occlusive views may be useful for detecting early cortical erosion along the occlusal surface of the alveolar ridge [21]. Earlier investigators have suggested, and our experience agrees, that these studies will reveal a defect as small as 2.5 mm [21]. However, other authors have reported high false-positive and false-negative rates. The high false-positive rates have been attributed to periodontal disease [8]. The high false-negative rates may be related to studies suggesting that 50-75% of bone thickness must be missing for a cancellous defect to be detected on unenhanced radiography [21]. As a result of the high false-positive and false-negative rates, this technique has not been widely accepted [8, 12, 13, 22]. It is unclear whether our results could have been further improved by the use of occlusive views. Some authors have advocated the use of specialized postprocessing dental CT software with algorithms that allow sagittal and coronal reformations of the occlusal surface [23]. Whether this technique would be valuable in detecting early cortical erosion of the alveolar ridge is a question that warrants further investigation.
The purpose of this investigation was to reevaluate the diagnostic accuracy of CT in detecting mandibular invasion by squamous cell carcinoma of the oral cavity. This study did not directly compare CT with other imaging modalities, such as MR imaging, for identifying invasion. However, our CT results compare favorably and are superior to the reported diagnostic accuracy of MR imaging and bone scanning in detecting mandibular erosion [6, 7]. Chung et al. [6] reported a sensitivity of 100%, specificity of 71%, positive predictive value of 50%, and negative predictive value of 100% using MR imaging. Curran et al. [7] reported a sensitivity of 100%, specificity of 29%, positive predictive value of 64%, and negative predictive value of 100% for bone scintigraphy using single-photon emission computed tomography (SPECT). The higher specificity and positive predictive value of CT suggest that properly performed CT is superior to MR imaging and bone scintigraphy with SPECT for excluding tumor [6, 7, 24]. The higher specificity and positive predictive value of CT compared with MR imaging and bone scintigraphy may prevent unnecessary segmental mandibulectomies and reconstructive procedures. However, a direct comparison of CT with MR imaging would be necessary to determine which study is better at detecting mandibular invasion.
Because of the retrospective nature of our study, there are limitations that should be addressed. All patients in our study underwent mandibulectomy; therefore, we were unable to determine the true false-negative rate of CT. Possibly that rate could be determined if all patients with squamous cell carcinoma of the oral cavity were included in our investigation. However, we would not have pathologic correlation on this group of patients because it is highly unlikely that a freely mobile lesion in the oral cavity would invade the mandible, and therefore the mandible would not be resected. This study specifically examined the scenario in which imaging would directly impact treatment (i.e., when tumors were fixed to the mandible on palpation). Because this was a retrospective investigation, we were unable to provide information identifying the number of times the surgical decision was altered because of the information provided by CT. This information will become even more difficult to determine in the future because the 1997 combined American Joint Committee on Cancer/International Union Against Cancer staging manual recommends that "...any diagnostic information which contributes to the overall accuracy of pre-treatment assessment should be considered in the clinical staging..." [4]. Integration of imaging into final staging will likely render the separation between "clinical staging" and "imaging staging" indistinct. Finally, only axial images were evaluated in our investigation. Coronal images would have likely been useful and may have further improved our diagnostic accuracy, especially for retromolar trigone and alveolar ridge carcinomas. Because coronal images were not routinely obtained, none were evaluated in our study.
In conclusion, the results of our investigation show that CT is a reliable technique for predicting mandibular invasion and may be a helpful adjunct when considering mandibular-sparing procedures for squamous cell carcinoma of the oral cavity. However, we believe that the most accurate and reliable results can be consistently obtained only with high-quality studies consisting of thin-section (3-mm) imaging and reconstruction of the mandible using bone algorithm.
Acknowledgments
We thank Dianne Armao for her editorial assistance.
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Imaizumi, N. Yoshino, I. Yamada, K. Nagumo, T. Amagasa, K. Omura, N. Okada, and T. Kurabayashi A Potential Pitfall of MR Imaging for Assessing Mandibular Invasion of Squamous Cell Carcinoma in the Oral Cavity AJNR Am. J. Neuroradiol., January 1, 2006; 27(1): 114 - 122. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bolzoni, J. Cappiello, C. Piazza, G. Peretti, R. Maroldi, D. Farina, and P. Nicolai Diagnostic Accuracy of Magnetic Resonance Imaging in the Assessment of Mandibular Involvement in Oral-Oropharyngeal Squamous Cell Carcinoma: A Prospective Study Arch Otolaryngol Head Neck Surg, July 1, 2004; 130(7): 837 - 843. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Brockenbrough, G. J. Petruzzelli, and L. Lomasney DentaScan as an Accurate Method of Predicting Mandibular Invasion in Patients With Squamous Cell Carcinoma of the Oral Cavity Arch Otolaryngol Head Neck Surg, January 1, 2003; 129(1): 113 - 117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. N. Myers and B. R. Gastman Clinical Commentary on Marginal vs Segmental Resection of the Mandible Arch Otolaryngol Head Neck Surg, May 1, 2002; 128(5): 605 - 606. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
