Bone invasion by oral squamous cell carcinoma

Introduciton

Approximately 30% of all head and neck cancers are malignant tumours of the oral cavity. Histologically, more than 90% of these are squamous cell carcinomas [1]. According to global statistics, where oral squamous cell carcinoma (OSCC) is presented together with oropharyngeal carcinoma, it is the 6th most common cancer in the world. About 300,000 of new OSCC cases are diagnosed every year and aproximately two thirds of these cases are diagnosed in developing countries from areas such as South and Southeast Asia, Eastern Europe and parts of Latin America [2]. At least 95% of OSCC cases occur in patients over 40 years of age and OSCC occurs twice as often in men as in women [3]. The main risk factors include smoking, alcohol abuse and betel chewing [4]. Heavy smokers and alcohol drinkers are up to 38-times more at risk of getting the tumour as opposed to abstainers from both products [2]. The human papillomavirus (HPV) plays an important role in oropharyngeal carcinoma. In the case of OSCC, the role of HPV is less important. Oral carcinomas, associated with HPV, occur in about 4% of the cases [5]. Approximately one third of the patients with OSCC are diagnosed in stages I and II. The five-year survival rate is 80% in stage I and 65% in stage II. Unfortunately, the majority of OSCC is diagnosed in advanced stages of the disease (III and IV), where the five-year survival rate is less than 50% and the overall survival rate is 30%, especially because local recurrence or distant metastases often occur [2,6,7].

Bone invasion as a prognostic factor

According to the close anatomical relationships in the oral cavity, carcinomas in the retromolar trigone, gingiva, hard palate, floor of the mouth, buccal mucosa or tongue may involve bone of the maxilla and/or the mandible [8]. The prevalence of jaw bone involvement by OSCC ranges from 5% to 56% [8,9]. The size and the location of the tumour are important in this point of view. It seems that larger or more deeply invading tumours in the soft tissue are more likely to invade the bone. Brown et al [10] showed that 48% of tumours with a dia­meter of < 4 cm do not invade the bone and 80% of tumours with a diameter of > 4 cm invade the bone. Pandey et al [11] described that the possibility of mandibular involvement is higher in patients where tumors are located within the distance of 1 cm from the mandible. In most cases, tumours enter the bone at the point of abutment rather than pre­ferential entry through the occlusal ­surface, neural foramen, or the periodontal membrane [10]. Frequently, this area is in the dentate and edentulous jaws at the junction of the attached and reflected mucosa [10,12,13]. According to Huntley et al [14] in tumours localized lingual to the mandible the site of entry of the tumour into the bone was usually through the alveolar crest with additional spread through the lingual cortex.

According to the TNM system, tumours that invade through cortical bone are classified as T4a, stage IVA. Although many authors consider bone invasion as an independent prognostic factor, there are studies in the literature that point out the prognostic limits of TNM classification in small (≤ 4 cm) OSCC invading the bone [15–17]. Ebrahimi et al [15] suggested modifying the T stage such that the tumours are classified as T1–T3 based on their size and then upstaged by one T stage in the presence of medullary bone invasion. Okura et al [17] recommended the same modification of the TNM classification; however, only for lower gingival OSCC. The results of Kuk et al [16] showed that in small primary OSCC (≤ 4 cm) bone invasion was not significantly associated with disease progression. However, their study pointed out a small subgroup of patients with small OSCC (2–4 cm) with both buccal and lingual bone invasion, which showed a statistically significant worse prognosis compared to the other group of patients with OSCC with or without bone invasion [16]. According to Mücke et al [18] and Petrovic et al [19] in cases of adequate resection margins, the prognosis in patients with bone invasion was not worse. Nevertheless, it has to be taken into account that many studies evaluating bone invasion as a prognostic factor focus only on mandibular invasion. However, OSCC can invade also the maxilla or palatine bone. In this regard, further research is necessary in the future.

The cellular and molecular mechanism of bone invasion

Numerous studies have shown that the destruction of bone associated with the invasion of carcinoma is mediated by osteoclasts rather than directly by malignant cells of the tumor [20–23]. Quan et al [8] described bone invasion as a highly coordinated process, spatially and temporally regulated, in which the following three phases may be distinguished: initial, resorptive and final phase. Various molecules that have different roles in individual phases are involved in the whole process. In the initial phase of bone invasion, when osteoclasts have not yet been recruited, proteases such as matrix metalloproteinases and cathepsins are involved. In the resorptive phase, when osteoclasts play the main role in resorption, cytokines of the tumor necrosis factor (TNF) family and parathyroid hormone-related peptide (PTHrP) are involved. In the final phase, there are growth factors such as epidermal growth factor, transforming growth factor and connective tissue growth factor, which are released from the reservoir within mineralized bone matrix after the destruction. These growth factors promote the growth of neoplastic cells themselves, leading to a vicious cycle that accelerate the process of bone invasion [8,24].

The interaction between osteoblasts, bone marrow stromal cells and osteoclasts is essential in bone remodelling. The balance between them is tightly regulated by many parameters such as mechanical stimulation, hormones and cytokines. Any disturbance between these effectors leads to the development of skeletal abnormalities, characterized by a decrease or an increase in bone mass [25]. It has been proposed that osteoblasts or bone marrow stromal cells are involved in the osteoclastogenesis process via the mechanism involving cell-to-cell contact with osteoclast precursors [3]. The discovery of three proteins from the TNF family that are crucial for the development and activation of osteoclasts is seen as an important finding over the past two decades in bone biology, which supports this hypothesis. Those proteins are the receptor activator of nuclear factor kappa B (RANK), its ligand (RANKL) and osteoprotegerin (OPG) [25,26]. RANKL occurs in two forms – a transmembrane form expressed by osteoblasts, stromal cells and T-lymphocytes, and a soluble form produced by activated T-lymphocytes. Furthermore, RANKL is expressed in lymph nodes, thymus, mammary tissue, lungs and many other tissues. RANK is a transmembrane protein, which is expressed by osteoclasts precursors, mature osteoclasts, dendritic cells as well as some tumour cells, e.g. in breast and prostate carcinoma. OPG is secreted by osteoblasts, stromal cells and a number of other cells in the heart, kidneys, liver and spleen [26]. Soluble and transmembrane forms of RANKL expressed by osteoblasts exert activity through binding to their RANK receptor on osteoclasts [25]. The interaction between RANKL and RANK activates the intercellular signaling cascade, which leads to the differentiation of osteoclast precursors to the osteoclasts called osteoclastogenesis, to bone resorption and the survival of osteoclasts. The OPG acts as a soluble receptor antagonist for RANKL, preventing the binding to the RANK, thereby preventing its activation [27]. Thus, OPG represents a bone protecting molecule [25]. It is believed that the balance of RANKL and OPG expression in osteoblasts and stromal cells is critical for regulating osteoclast differentiation and the function under physiological and pathological conditions [21].

Bone resorption by osteoclast is an important step in the process of bone invasion, progression and metastasis in several malignancies, such as breast cancer, prostate cancer, lung carcinoma, colorectal carcinoma, multiple myeloma and OSCC [20,21,27]. Sato et al [22] showed that RANKL produced by osteoblasts, stromal cells and carcinoma cells is involved in osteoclastic bone resorption during bone destruction in OSCC. This is also in line with the results of Zhang et al [23] who described that OSCC cells express different levels of both forms of RANKL, transmembrane and so­luble and that RANKL directly supports the osteoclastic loss of bone. Tada et al [21] showed that OSCC induced by the suppressed expression of OPG in osteoblasts is involved in osteoclastogenesis. According to another study by Tada et al [20], the differentiation and regulation of osteoclast function is attributed to the modulation of the ratio of RANKL vs. OPG expression by the microenvironment of the tumor and bone, rather than just RANKL expression by OSCC cells. The results of Russmueller et al [27] showed a significantly higher expression of the OPG by OSCC cells in patients with bone invasion compared to patients without bone invasion. Ishikuro et al [28] described that osteoclastic resorption of bone during the invasion of a tumour does not require direct contact of the tumour cells with the osteoclasts of the adjacent bone. He thereby pointed out the significant role of fibrous stroma between the tumour cells and the resorbing bone in bone invasion through RANKL dependent pathway [28]. Despite the significance of the OSCC cells during the osteoclastogenesis, the role of its effects on the function of the osteoclasts is not completely clear. Most of the authors concur that further research is needed in this regard.

Histological patterns of bone invasion

Once the tumour affects the bone, there are three histological patterns of bone invasion: erosive, mixed and infiltrative [12,29,30]. The histological image of an erosive pattern of bone invasion (Fig. 1) is characterized by a broad, expansive tumor front with a sharp interface between the tumour and the bone. It is further characterized by osteoclastic bone resorption and fibrosis along the front of the tumour and the absence of bone islands within the tumour mass. The infiltrative pattern of bone invasion (Fig. 2) is composed of nests of tumor cells with fingerlike projections along an irregular tumour front. There are also residual bone islands within the tumour mass and haversian system penetration. Cases exhibiting features of both patterns, are designated as a mixed pattern [1,3]. Brown et al [10,13] showed that the size and width of the tumor increase the risk of bone invasion and also that the growing size of the tumour and the depth of the invasion into the soft tissue increase the chance of the infiltrative pattern of bone invasion. Thus, the autor assumes that the progression of the tumour into the bone develops the pattern of bone invasion from erosive to mixed and infiltrative. Ishikuro et al [28] also supposes that the erosive and infiltrative pattern of bone invasion represents different stages of the tumour development. Brown et al [10] further stated that neither the pattern of bone invasion nor the fact of whether or not the jaws were dentate or edentolous are not related to the localization of the invasion of the tumour into the bone. Wong et al [1] studied whether the infiltrative pattern of bone invasion is related to a worse prognosis in comparison to the erosive pattern. In their work a three-year disease-free survival in the infiltrative pattern was 30% and 73% in the erosive pattern. They also describe that the infiltrative pattern of bone invasion shows more aggressive behaviour with a higher probability of positive resection margins at tumour excisions and a higher probability of the disease recurrence [1]. Ishikuro et al [28] also agree with this conclusion. In their study, they also showed that the infiltrative pattern of bone invasion shows a higher number of osteoclasts on the surface of the resorpting bone lining the tumour in comparison with the erosive pattern [28]. The possibility of determining the pattern of invasion by intraoperative frozen section is a question of clinical importance. Due to the disruption of the bone's architecture and the difficulty of preparation of frozen section of the bone, this determination is not possible. However, postoperative determination of pattern of bone invasion provides important prognostic information. Therefore this information should be noted on every bone specimen with squamous cell invasion [1].

Detecion of bone invasion

Since the majority of tumours in the oral cavity are visible to the naked eye, the first step in examination of a patient suspected of suffering from OSCC is inspection and palpation. A final diagnosis is based on a histological examination of sample taken by biopsy. However, the correct disease diagnosis and tumor staging, which affect treatment planning, requires the use of complementary imaging techniques capable of offering additional information [31]. The preoperative detection of the presence of bone invasion affecting the facial skeleton in patients with OSCC is critical for planning of surgery and postoperative treatment. According to Zupi et al [32] a clinical examination in evaluating the presence of bone invasion is highly sensitive (82.6%) and is thereby considered to be an essential process. However, for low specificity (44%), a clinical examination is only the first step in further examination. When evaluating the bone invasion, it is possible to use various imaging techniques such as X-ray (Fig. 3) or more reliable techniques such as CT (Fig. 4), MRI or PET/CT. However, every technique has its limitations. According to Rao et al [33] a routine X-ray, including orthopantomogram, are not able to capture the initial bone invasion until 30% of bone minerals are lost. This finding explains the false negative results in evaluating bone invasion of carcinomas using ­X-rays. More­over evaluating the area of the mandibular symphysis in orthopantomogram is difficult for summation of the anatomical structures. The most commonly used imaging techniques for examining patients with OSCC include CT and MRI. In terms of limitations of the individual techniques, CT is asso­ciated with a higher number of artefacts caused by metal dental materials, while MRI may be burdened with low image quality due to movements of the tongue and swallowing [34]. Many studies deal with the diagnostic accuracy of these imaging techniques when evaluating the presence of bone invasion. However, which one of these is better still remains controversial. The results of Vidiri et al [35] showed that MRI was more sensitive (93%) when evaluating the presence of mandibular bone invasion in comparison with CT (70%), with the same specificity of both techniques (82%), even though the difference was not statistically significant, probably due to a small group of patients. Similarly, Wiener et al [36] described a higher sensitivity of MRI (100%) compared to CT (71%) with the same specificity (93% and 96% respectively), the differences were also not statistically significant. On the other hand, when evaluating bone invasion into mandible, Imaizumi et al [37] pointed out the potential pitfalls of MRI, which represents a significant number of false positive cases. In his study, the MRI sensitivity and specificity was 96% and 54% and the CT sensitivity and specificity was 100% and 88%, respectively. Silva et al [38] also reported the low specificity of MRI in comparison with CT as a crucial difference between these imaging techniques. Abd El-Hafez et al [39] compared PET/CT and MRI. In his study, the sensitivity and the specificitity for ­­PET/CT was 78% and 83%, and those for MRI were 97% and 61%, respectively. Due to the higher specificity, PET/CT may complement the role of MRI in the diagnosing of bone invasion by oral carcinoma. The author suggested that in dentate patients with positive MRI findings, a nega­tive PET/CT may be useful to rule out bone marrow invasion [39]. Hakim et al [40] studied the possibility of using CBCT when evaluating bone invasion. In their study, CBCT showed a higher sensitivity in comparison with the CT. According to the authors, CBCT represents an alternative imaging technique, which could be combined with another imaging technique of soft tissues such as MRI [40].

Conclusion

Bone invasion by oral squamous cell carcinoma is most commonly present in tumours close to the bone or in bigger and more advanced tumours and it is considered to be an adverse prognostic factor. Activated osteoclasts play a crucial role in the bone invasion process. From the viewpoint of histology, there are three patterns of bone invasion – erosive, mixed and infiltrative. The preoperative evaluation of the presence of bone involvement is important for planning the surgery. Despite all advances in the diagnosis and treatment, bone invasion which is caused by oral squamous cell carcinoma represents a diagnostic and therapeutic problem.

Roles of authors: All authors have been actively involved in the planning, preparation, analysis and interpretation of the findings, enactment and processing of the article with the same contribution.

Conflict of interests: The authors state that there are no conflicts of interest regarding the publication of this article. The authors declare that this study has received no financial support.

David Král, MD

Department of Oral and Maxillofacial Surgery

Faculty of Medicine

Palacký University Olomouc and University Hospital Olomouc

I.P.Pavlova 185/6

779 00 Olomouc

Czech Republic

e-mail: david.kral@fnol.cz

Submitted: 12.6.2021

Accepted: 30.7.2021