Oral cavity carcinoma is the eighth most common malignancy worldwide with an estimated incidence of 300,000 cases per year [1] . The unusual habit of chewing tobacco contributes to its significantly greater prevalence in Southeast Asian nations, particularly India. Males experience this condition more frequently than females, and the number of younger individuals is also increasing [2] . The extent of cancer spread is associated with prognosis and survival. The 5-year survival rate is 80% when the tumour is limited to the mucosa, but it decreases to 40% when the mandible is involved [3] [4] .

The AJCC (American joint committee on cancers) 8^th edition states that mandibular cortex involvement may upstage the tumour to IVa in squamous cell carcinoma of the mandibular alveolus, floor of mouth, gingivobuccal sulcus, buccal mucosa, or retromolar trigone [5] . Resection of the mandible is required in such cases, which not only increases mortality but also significantly hampers quality of life. Consequently, it is crucial to identify mandibular invasion prior to surgery to plan for surgical resection. This incursion can be infiltrative or erosive [6] [7] . There is currently no single definition for these concepts. On the other hand, the histological erosive pattern shows a broad pushing front with osteoclastic activity, while the infiltrative pattern has an uneven tumour front with projections that resemble fingers into the mandible and no discernible osteoclastic activity [8] [9] .

Clinical examination is reliable in the majority of cases but has low sensitivity and specificity in detecting tumour invasion [10] . Currently, no one study has been able to predict mandibular invasion with any degree of precision. The following imaging techniques have been employed to identify bony involvement: orthopantomography, computed tomography, magnetic resonance imaging, bone scintigraphy, and Single photon emission computed tomography (SPECT) [11] . Thus, a combination of clinical and radiological assessments is required for accurate analysis of bone invasion. This study aimed to establish a correlation between histological assessment and clinical examination and computed tomography results in patients with mandibular involvement and oral cavity squamous cell carcinoma.

Squamous cell carcinoma of the oral cavity is one of the leading causes of morbidity and mortality worldwide. Cancers of the buccal mucosa, floor of mouth, lower alveolus, and retromolar trigonal region are often linked to direct extension of the mandible in an erosive or invasive way (the mandibular involvement pattern has a prognostic influence) [12] . The AJCC 8^th edition states that an erosive pattern of involvement has a prognosis and survival comparable to that of no involvement, whereas an invasive pattern of involvement upstages the tumor. Thus, detection of mandibular bone invasion is of utmost importance in preoperative treatment planning for appropriate resection of the mandible.

Clinical examination alone is not sufficient to predict bony involvement and always requires imaging in the form of orthopantomogram, CECT, magnetic resonance imaging (MRI), SPECT, cone beam CT (CBCT) and bone scintigraphy. The choice depends upon a number of factors: surgeon preference, institutional availability, patient affordability and dental artifacts [13] . In a recent systematic review and meta-analysis, there was no discernible advantage between MRI and CECT for predicting mandibular bone invasion [14] . In 65.1% of our patients with oral squamous cell carcinoma, computed tomographic scans obtained before surgery were able to predict bone involvement. All mandibles exhibiting profound cortical invasion underwent segmental resection, whereas tumours adhering to the cortex, abutting the surface, or exhibiting only superficial cortical invasion underwent marginal mandibulectomy [15] . The remaining mandibular height was taken into consideration and converted to segmental resection when the height was less than 1 cm. When a patient has advanced paramandibular disease, the operating surgeon prefers segmental resection since there is a greater risk of bone invasion because many tumor entry sites may go undetected on CECT imaging [16] [17] . In such cases, the risk of extra resection is greater than the risk of tumor recurrence.

De Angelis et al. reported that MRI had greater sensitivity and specificity (87.1% and 80.5%, respectively) than CECT in their radiological assessment of mandibular bone invasion. CECT had a sensitivity of 69.0% and a specificity of 79.6%. Their histopathological examination revealed more patients with mandibular invasion than detected by imaging while our study showed contrasting results. In our study, HPE assessment was positive for mandibular invasion in only 12 out of 28 resected specimens. Since the majority of patients who came to us are thought to be in an advanced stage with large mandibular invasion, these individuals should be more easily detected by CECT. However, mild mandibular cortical erosion can be difficult to detect in patients with modest tumor sizes. Our study has constraints as we did not include instances with early stage oral SCC to assess CECT sensitivity.

Rajan S. Patel et al. examined the effects of bone invasion and the extent of bone resection patients with oral carcinoma. They discovered that the most significant predictors of disease-specific survival were the extent of nodal disease (pathological nodal status and extracapsular spread), positive margins, pathologic T stage, and bone invasion; soft tissue margins in the presence of bone invasion appeared to be the single most significant predictor of outcome. Patients with gross mandibular invasion tend to have extensive disease (skin involvement, maxillary invasion, extracapsular lymph node extension etc.) whose reconstruction is difficult at times and the poor prognosis in such patients is due to advanced local disease rather than mandibular invasion alone.

In a comprehensive analysis of various imaging modalities used in the detection of head and neck cancer, the sensitivity of conventional CECT was shown to be 73% for mandibular invasion and 85% for bone marrow invasion. A minimum of marginal resection must be performed even on the smallest suspicion since early cortical erosion may be overlooked on CECT scans [18] [19] . Consequently, intraoperative clinical judgment is superior to imaging alone. For the individuals in our study, mandibular resection appeared unnecessary because over half of the resected tissues did not exhibit bony invasion. However, in patients in which there is a clinical mandibular abutment, we advise minimal resection rather than no resection.

As mandibular resection is highly dependent upon preoperative imaging, newer imaging modalities, such as cone beam CT and SPECT have gained wide popularity in the last decade. Among the available imaging modalities, CBCT has the highest diagnostic efficacy and SPECT has the highest sensitivity. Hakim et al. discovered that conventional CT had a superior negative predictive value and a greater sensitivity for cortical invasion when identifying mandible invasion in oral SCC than CBCT and SPECT. Additionally, with significantly lower radiation exposure in CBCT, this invasion can be effectively ruled out preventing over resection of the mandibles [20] . Chauker et al. reported that Multi-detector CT had an accuracy of 83.3% in predicting mandibular invasion as well as the extent of invasion, with a mean underestimation of approximately one mm. The occlusal route appeared to be the predominant route of tumor entry in gingiva-buccal complex cancers in dentate patients. They suggested that in advanced tumours with paramandibular disease, mandibular conservation is contraindicated because of multiple routes of tumor entry involving larger portions of the cortex and medulla. Therefore, segmental resection seems to be the preferred approach for treating T3-4 tumours of the mandible, and the extent of resection should correspond to the adjacent soft tissue cut margins, thus avoiding unnecessary hemi mandibulectomies [21] .

Compared to SPECT and bone scintigraphy, CBCT is becoming the preferred option due to its lower cost and radiation exposure [22] [23] . It is also more frequently used in oral surgery and dentistry, making it easier for doctors to interpret and explain the results. Additionally, the latter have higher false positive rates, which might lead to inappropriate resection of the mandible. However, CT and MRI show the highest specificity among all modalities, and MRI has greater sensitivity than CT but has the disadvantage of high false positivity due to periodontal inflammation.

It was unexpected that only 42.8% of patients had invasion on postoperative histology despite preoperative CT evidence of bone involvement, which raised questions about needless mandibular resection. Histology is the recognized gold standard; however, sampling mistakes can cause false positives from imaging, which could be false negatives from histology. The lack of serial sectioning and examination of an inadequate number of slices beneath the tumor could explain the negative results observed by histology. False positives from imaging might arise due to adjacent periodontal inflammation and prior biopsy. The current study was unable to determine the wide variation observed in the results and requires more research [24] . We need more accurate preoperative imaging modalities that show greater positive predictive value, and histological analysis can be improved by meticulous assessment of serial cut sections of the mandible by an experienced head and neck pathologist.

There is an urgent need to review the usefulness of radiology in the treatment of SCC of oral cavity warranting mandibular resection. The combination of clinical and radiological examination increases both the sensitivity and specificity.

This study has several limitations. First, MRI is required to determine the extent of soft tissue invasion, which limits the use of CECT for assessing bone invasion during treatment planning. Second, there was no comparison between the various subsites for mandibular invasion. However, a large sample size study will be required to effectively conclude an association between radiological and histopathological features.