1. INTRODUCTION

Squamous Cell Carcinoma (SCC) is the most common cancer of the oral cavity [1]. Cancers in advanced stages usually involve nonepithelial structures. According to the literature, the incidence of mandibular involvement in oral cavity cancers ranges from 12% to 56% [2–4]. These bony deformities involve the enclosing soft tissues and musculature, resulting in difficult and problematic mastication, speech, and lip closure. As such, mandibular involvement affects the quality of life of patients, functionally as well as aesthetically [5].

Squamous cell carcinoma is the second most common malignancy in both men and women in Pakistan [6]. The age-standardized incidence rate per 100,000 people in Karachi South (Sindh province) is 21.3 in males and 19.3 in females [7]. A local study reported that among oral cancers, tongue and lips were involved in about half of the cases diagnosed [8].

The role of definitive surgery depends on the extent of bone resection combined with radicalness of tumor resection. Therefore, whether it is necessary to predict the bony involvement of mandible by taking imaging scans prior to the surgery is an important issue. Multiple Detector Computed Tomography (MDCT) is most commonly used to detect mandibular involvement in the oral cavity [9,10]. However, various studies conducted on the accuracy of MDCT provided inconclusive results. In their study, Zupi et al. [11] reported a sensitivity of 91.3%, a specificity of 96.3%, a Positive Predictive Value (PPV) of 92%, and a Negative Predictive Value (NPV) of 97%; however, these validity indices were not supported by other studies. Moreover, the validation study conducted by Lane et al. [12] reported 50% sensitivity and 61% NPV, suggesting that MDCT is an imprecise method for detecting mandibular involvement of the lower jaw. On the contrary, another study conducted by Kumar et al. [13] reported higher sensitivity (83%) and specificity (90%). Meanwhile, Bouhir et al. [14] reported lower sensitivity (70%) and specificity (71%) for detecting mandibular invasion among patients with head and neck cancer. Similarly, a systematic review consisting of 30 studies with the objective of detecting diagnostic efficacy of Computed Tomography (CT) in distinguishing mandibular invasion concluded that CT findings for mandibular medullar invasion had a pooled sensitivity of 0.72 and specificity of 0.90 [15].

Surgeons and radiologists play a very conclusive role in determining mandibular invasion in patients diagnosed with oral cavity SCC, because it also focuses on the extension of bony invasion, as this will influence the surgical operative plan. It is helpful for a surgeon to plan for either a marginal or segmental mandibulectomy, hemimandibulectomy, partial maxillectomy, and total maxillectomy [1].

The literature regarding the validity indices of CT have provided mixed results. Moreover, as local data regarding the sensitivity and specificity are not available, our study would help us in determining the diagnostic accuracy of CT in detecting mandibular involvement in oral cavity carcinoma. We conducted this study to determine the diagnostic accuracy of MDCT in detecting mandibular invasion among patients with oral cavity SCC, using histopathological detection (considered the gold standard) of mandibular involvement.

2. MATERIALS AND METHODS

This validation cross-sectional study was conducted in the radiology department of one of the biggest tertiary care hospitals in Karachi, Pakistan (Abbasi Shaheed Hospital) from July 2017 to June 2018. Sample size was calculated using the sample size calculator for sensitivity and specificity studies described by Dr Lin Naing @ Mohd. Ayub Sadiq (School of Dental Sciences, Universiti Sains Malaysia, Penang, Malaysia). Confidence level (1–α) was taken as 95%, with desired precision (d) of 10%. The incidence of mandibular invasion in oral cavity SCC was taken as 53.3% [4]. We used the expected sensitivity and specificity of CT in detecting mandibular invasion as 82.6% and 86.9%, respectively, from the parent study [16]. The largest sample size calculated was 75. We used this sample size in our study. For the selection of study participants, nonprobability consecutive sampling was used to recruit the patients. The case was considered squamous carcinoma if all of the following features were seen during light microscopy examination: cell showing keratinization, vesicular nuclei, prominent nucleoli, and mitotic figures. The finding on MDCT consisted of visualization of soft tissue enhancing mass lesion with infiltration into the mandible, which we labeled as mandibular invasion; meanwhile, the histopathology (gold standard) finding showed infiltration of nests of squamous cells with intercellular bridges along with keratin pearl. The inclusion criteria included diagnosis of oral carcinoma on biopsy as described above and presentation within 1 year of diagnosis. Patients who showed recurrence after stopping treatment and who were previously operated for oral cancer were excluded. Patients who had undergone MDCT only (n = 4) and those who did not undergo histopathological examination for possible bony involvement of SCC (n = 19) were also excluded. Those who showed presence of metallic artifacts in the images that could interfere with radiological interpretation (n = 7) were likewise not considered in our study. Tumors ranged in size from 4 to 32 mm and arose from the floor of the mouth (n = 20), gingiva (n = 11), retromolar trigone (n = 8), alveolar ridge (n = 5), and lip (n = 4).

After institutional approval for the study was obtained, MDCT was performed in patients with oral cavity cancers. All patients then underwent subsequent histopathology. MDCT results were compared with pathological results (gold standard reference). Next, a structured questionnaire about sociodemographics was required to be filled out by the data collector. Confidentiality of the data was maintained, and anonymity with voluntary participation was ensured. The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki.

All patients with oral cavity SCC who were referred to the radiology unit of Abbasi Shaheed Hospital for CT suspected for mandibular invasion and carcinoma diagnosed on biopsy were included. The purpose of the study was explained to the participants including the risks and benefits after informed consent was obtained; proforma was required to be filled for name, sex, age, and patient ID number. In our center, CT was performed using the Toshiba Asteion multislice CT scanner with before and after I/V contrast administration. Scanning protocols were as follows: 3-mm section thickness; collimation, 3 mm; reconstruction interval, 3 mm; scan delay, 80 s; 450 mA; 100–150 kV.

3. RESULTS

Table 1 shows the sociodemographic characteristics of patients who were enrolled (n = 75) in this study to assess the diagnostic accuracy of CT. The mean age of the patients was 54.47 ± 7.57 years. About one-third of the patients (25; 33.3%) were in the age group of 40–49 years; about two-thirds (49; 65.3%) were males, and 34.7% (26) had not attained any education. Urdu-speaking was the most common ethnicity in our study population (36; 48%). Forty-two (56%) of the participants had T4 staging, followed by 13 (17.3%) who had T3 staging.

Multiple detector computed tomography correctly diagnosed 46 of 48 cases as mandibular invasion, and correctly excluded mandibular invasion in 26 of 27 cases without invasion (Table 2).

The sensitivity of MDCT in detecting mandibular invasion in oral cavity SCC was 95.83% and its specificity was 96.30% (Table 3). This showed that MDCT’s ability to detect mandibular invasion in oral cavity SCC was 95.83%. This also showed that MDCT’s ability to detect the absence of mandibular invasion in individuals was 96.30%. The PPV of MDCT in detecting mandibular invasion in oral cavity SCC was 97.87%, whereas the NPV was 92.86% (Table 3).

4. DISCUSSION

To the best of our knowledge, this is the first study to evaluate the diagnostic accuracy of MDCT in detecting mandibular involvement among patients with oral SCC in Pakistan. We found out that MDCT has very good ability to detect mandibular invasion compared with histopathology (gold standard). It is vital for surgeons to know about mandibular involvement in oral SCC of head region, which will eventually help in decision making for surgical resection. There has been assessment of mandibular invasion by clinical assessment as well in which flexibility of the tumor is assessed in relation to the mandible bimanually.

Mandibular involvement in oral cell carcinoma is considered a very important issue in head and neck surgery because removal of the mandible is very challenging. These tumors usually receive postoperative radiation; as a result, vascularized bone grafts are needed for mandibular reconstruction. These procedures usually take a lot of time, and they are performed during mandibular resection. Therefore, it is imperative for head and neck surgeons to know about the exact status of mandibular involvement in oral SCC so that patients and attendants could be counseled adequately and appropriately about the operative procedure.

Preoperative assessment of mandibular involvement provides crucial information to head and neck surgeons. Mandibular involvement has been widely studied, and various imaging modalities have been used to assess whether patients with oral SCC are suitable for mandibular surgery procedures. Technical advances have made CT scan the most commonly used modality in detecting mandibular involvement. Results have been variable regarding diagnostic accuracy of preoperative CT in this context. One older study conducted in 1986 [18] reported a sensitivity of 100%, specificity of 97%, PPV of 92%, and NPV of 100%. In subsequent studies, Shaha [19] reported a diagnostic accuracy of 68% and Brown and Grew [20] reported a false negative rate of 28%, indicating that the predictability and reliability of CT were not up to standards. Later, Lane et al. [12] reported a sensitivity of 50% and NPV of 61% for MDCT, suggesting that CT was not very accurate method for evaluating the status of the mandible.

In 2001, Mukherji et al. [9] reported promising results in this regard: sensitivity of 96%, specificity of 87%, PPV of 89% and NPV of 95%. In 2006, Imaizumi et al. [21] also showed excellent results in a comparative study for detecting mandibular invasion using CT and Magnetic Resonance Imaging (MRI). They reported sensitivity of CT as 100%, specificity 88%, PPV 89%, and NPV 100%. These accuracy measures were higher than those reported in any other study. This is chiefly because Imaizumi et al. [21] had taken 5- and 1-mm-thick slices in axial section along with soft tissue and bone algorithm images. They had also added Dental CT, or Denta scan, a CT software program specifically developed to evaluate multiple panoramic and cross-sectional images of the jaw bone. This variability in results of different studies is most probably attributable to inconsistent imaging modalities; this was not optimal.

The results of our study for evaluation of mandibular invasion (sensitivity, 95.83%; specificity, 96.30%; PPV, 97.87%; NPV, 92.86%) are superior to those obtained by Brown and Grew [20], Shaha [19], and Lane et al. [12]. We believe this is because of the better imaging technique used. Brown and Grew [20] and Lane et al. [12] acquired images by using 5-mm-thick sections and did not routinely evaluate the mandible with bone algorithms. Shaha [19] did not describe the CT technique used in his study. Close et al. [18] also obtained 5-mm-thick contiguous sections; however, if there was suspicion of bony involvement, 3-mm sections were further obtained. Mukherji et al. [9] acquired contiguous 3-mm images of all patients. Both Close et al. [18] and Mukherji et al. [9] evaluated images on soft tissue and bone window settings.

There are two main patterns of mandibular involvement by SCC: erosive and infiltrative. Erosive involvement occurs when cortical bone recedes before a pushing tumor border. There is a scalloped excavation of the underlying medullary bone in erosive form. Meanwhile, in the infiltrative pattern of tumor involvement, cancer diffusely spreads throughout the cancellous, medullary bone. In our study, we were not able to identify the different types of patterns in histopathology.

Computed tomography has several limitations in diagnosing superficial lesion that are only abutting against an adjacent bone; hence, a full staging process should include clinical examination and imaging. Moreover, there is always a need for refinement in imaging techniques and modalities that can provide accurate information approaching the gold standard; in this regard, further studies can be conducted. MRI and fused positron emission tomography–CT may be useful. There is also a small chance that few cases of bone invasion could not be detected histologically or radiologically, which would reduce our reported sensitivity and specificity values; however, we used histopathological examination, which is unanimously considered the gold standard by clinicians, and tried to make sure that no case of bone invasion would be missed. Furthermore, the small sample size used in our study could be a limitation, a factor that needs to be considered in future studies.

5. CONCLUSION

The key findings of our study suggested that MDCT is highly reliable in detecting mandibular invasions in oral SCC. The technique, when used with histopathology, can give very promising results in early detection of bony invasions in the mandible. The use of this rapid, nonoperator-dependent, and highly accurate modality may decrease delays and inappropriate management, and it can also help in explaining the extent of the disease progression as well as in deciding about the appropriate surgical procedure to be applied. There may be few cases of bone invasion that could have been missed through the technique of histopathology or radiology, which would reduce our reported sensitivity and specificity values; however, we used histopathological examination, which is unanimously considered the gold standard by clinicians, and tried to make sure that no case of bone invasion would be missed. We also believe that the most accurate results can be consistently obtained only with high-quality studies consisting of thin-section imaging and reconstruction and reformation of images in coronal, sagittal, and three-dimensional images of the mandible using bone algorithm in CT.

CONFLICTS OF INTEREST

The authors declare they have no conflicts of interest.

AUTHORS’ CONTRIBUTION

AS, AAK, MSA and SMA conceived and designed the study, conducted research, provided research materials, and collected and organized data. AS and AAK analyzed and interpreted data. MSA and SMA wrote initial draft of article, and provided logistic support, whereas final draft was written by AS and AAK. All authors have critically reviewed and approved the final draft and are responsible for the content and similarity index of the manuscript.

FUNDING

The authors declare that this research received no external funding.