PDF
Citation
Print
I. Introduction
The retromolar trigone (RMT) area is a posterior triangular mucosal region in the oral cavity overlying the ascending ramus and extending from the third molar to its apex at the maxillary tuberosity1. Squamous cell carcinoma (SCC) of the RMT is a rare entity, accounting for 3.7% to 9.1% of all SCCs that arise primarily in the oral cavity2-4. However, owing to its unique spatial relationship with surrounding vital structures, carcinomas of this subsite are associated with an aggressive clinical course and consequent poor prognosis that differs drastically from that of similar cancers in nearby oral regions5.
Given its rare presentation, existing evidence regarding the treatment and prognosis of RMT SCC is scarce and has been documented only through case reports and small case series. Consequently, establishing adequate treatment guidelines according to the anticipated effects of tumors arising from this sublocation has been a difficult endeavor. The aim of this article is to provide a qualitative assessment of the outcomes and potential prognostic factors of RMT SCC through a systematic evaluation of the current literature.
The following PICOS (Patient population, Intervention, Comparatic controls, Outcomes, Statistical analysis) question6 was established to address the objective of this study: “In patients with biopsy-proven SCC of the RMT, which patient-, tumor-, and therapy-related variables can influence the outcomes and complications following treatment when compared one to the other, as demonstrated by randomized or non-randomized clinical trials and cohort and case-control studies?”
II. Materials and Methods
1. Literature search
This systematic review was performed in accordance with the recommended ‘Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Guidelines’7. An adequate search strategy for each database was designed with the support of a trained librarian. Preliminary trials were performed to identify the best search terms.
2. Study inclusion and exclusion criteria
The following inclusion criteria were designed a priori to select eligible studies for the systematic review. (1) Study design: All randomized, controlled clinical trials and cohort, case-control, and case series studies investigating the outcomes of various treatment modalities for RMT SCC and published in the English language were included. (2) Target patient population: All patients who underwent treatment for biopsy-proven RMT SCC and had pathological staging information available were included. (3) Interventions: Studies that used one or more of the following treatment modalities—surgical resection, radiotherapy, chemotherapy—were considered. (4) Outcome: Studies that evaluated the survival rates, complications, recurrences, or functional outcomes following treatment were considered.
Studies were excluded if they were published before 1980, if more than 50% of the subjects had SCC with its epicenter in regions other than the RMT, or if fewer than 10 cases were included in the final analysis. Case reports and grey literature such as abstracts, conference presentations, editorials, and expert opinions were excluded to limit the possibility of duplicated results.
Furthermore, among the duplicate studies with accumulating numbers of patients or increased lengths of follow-up that were published by the same institution, only the most complete reports were selected for analysis.
3. Search strategy
Electronic database searches were performed of the Cochrane Library, Embase, PubMed, and ScienceDirect. To ensure that the included studies were up-to-date, the search was limited to research published after January 1980. To achieve maximum sensitivity in the search and thus identify all relevant articles, we combined the following terms as keywords: (“Retromolar Trigone”) AND (“Cancer” OR “Carcinoma”) AND (“Outcome” OR “Prognosis” OR “Survival”). To identify other potentially relevant articles, we also conducted a manual review of the references in all the retrieved articles.
4. Study selection
For screening and data management, the retrieved citation abstracts and full text articles were uploaded to Distiller Systematic Review (DSR) software, an internet-based software program that facilitates collaboration among reviewers during the study selection process. Two authors (H.W.K. and M.Y.K.) discussed and developed screening forms and relevant questions based on the inclusion and exclusion criteria to complete the formal screening form in the DSR software.
The screening was performed in two phases. For phase I screening, two authors (H.W.K. and M.Y.K.) independently reviewed the titles, keywords, and abstracts of all references and selected articles with abstracts that met the inclusion criteria. The same authors then independently carried out the second screening phase by reviewing the full texts of the literature selected in phase I. Those that did not meet the inclusion criteria, as selected a priori, were excluded. Discrepancies were resolved through comparison and mutual agreement between the two authors. A third author (C.H.K.) resolved disagreements when the two authors failed to reach consensus.
5. Risk of bias assessment
Before the assessment, the three reviewers analyzed and discussed adequate tools for evaluating the quality of the selected studies. A modified version of the method suggested by Saltaji et al.8 was chosen to assess the risk of bias in individual articles.(Table 1) The symbol ‘✓’ and a score of 1 was given for every question answered as ‘yes.’ The answer ‘no’ was indicated by the symbol ‘×’ and did not contribute to the total score. The score that could be attributed to each study ranged from 0 to 16, and final risk of bias was classified as follows: low risk of bias if the score was 16; moderate risk of bias if the score was ≥10 and <16; high risk of bias if the score was <10. For the assessment, two reviewers (H.W.K. and M.Y.K.) independently assessed all of the included studies and then compared their results. Unresolved disagreements between the two reviewers were resolved by the third reviewer (C.H.K.).
6. Data management and extraction
A data extraction form was developed a priori via the DSR data management program. Using this worksheet, the data screening and extraction were conducted in duplicate by two independent reviewers (H.W.K. and M.Y.K.). Calibration exercises were conducted before the data extraction process to ensure consistency between the reviewers. Discrepancies were resolved through comparison and mutual agreement between the two authors, with a third author (C.H.K.) resolving disagreements when the two authors failed to reach consensus. The third reviewer (C.H.K.) also checked the final extracted data for accuracy and completeness. The data extracted were the study design, aim, location, and mean follow-up period; patient (age, sex, comorbidities) and tumor (sublocation, stage, nodal status, bone invasion) characteristics; treatment modalities; outcome measurement tools; and results. The retrieved data were synthesized by the DSR program.
III. Results
The review stages and details of identification, exclusion, and inclusion following the assessment for eligibility are portrayed in the PRISMA flowchart.(Fig. 1) The searches of the four electronic databases and other sources, such as the reference lists of the included articles, yielded 1,248 articles. After eliminating duplicates, 1,083 unique articles remained. The phase I assessment removed 968 irrelevant citations, leaving 115 full-text articles to be evaluated for eligibility. That process eliminated another 100 articles. Thus, at the end of the selection process, 15 studies remained for the final assessment1,9-22. All 15 articles were retrospective case-control studies. A total of 4,838 patients were included in the review. A summary of the overall data for all the included articles is shown in Table 2.
1. Eligibility and quality assessment
In the risk of bias assessment, most studies had a total score between 10 and 15 and were thus categorized as having a moderate risk of bias1,9-14,16-18,20-22. Two studies presented a score of 9 and were thus classified as having a high risk of bias15,19. All the analyzed studies clearly declared their study aim and population. The data measurement methods, which were also stated in all articles, were a retrospective review of medical records in all but one study, which used the nationwide Surveillance, Epidemiology, and End Results (SEER) database22. Two studies included fewer than 20 study subjects with RMT SCC, which was considered inadequate1,17. The inclusion criteria were not clearly stated in one study18, and four studies did not provide details regarding the follow-up duration for their subjects10,15,19,22. Three studies did not provide any data regarding the overall survival rates of their subjects10,13,15, and the 5-year overall survival rates were not stated in three other articles9,18,21. Among the studies, only six mentioned the number of dropouts9,10,12,13,16,21. Two studies did not specify their method for statistical analysis1,19, and confidence intervals were presented in only five11,14,15,20,22. The results of the bias assessment are summarized in Table 3.
2. Study characteristics
The 15 retrospective case-control studies included in the final analysis1,9-22 had heterogeneous populations for sample size, tumor stage, tumor characteristics, reported comorbidities, and treatment modalities. The detailed objectives and subject characteristics for each study are summarized in Table 4. A total of 4,838 patients underwent treatment for histopathologically confirmed RMT SCC from 196011 to 201915,21. In total, 69.7% of the subjects were male, and 30.3% were female. The sample sizes ranged from 1415 to 4,02222 participants, and the mean sample size was 323. The mean age at presentation for the patients analyzed ranged from 509 to 6712 years and was not declared in the study of Hitchcock et al.17. The mean follow-up time was 151 to 6012 months. Seven studies were conducted in North America10,12,13,15,17,19,22, four in Asia1,9,18,20, three in Europe14,16,21, and one in South America11.
The outcomes of tumors originating from other regions were observed and compared with the RMT SCC group in two articles13,15. Six studies documented the rate of comorbidities among their subjects1,9,10,12,15,19. Smoking was the most commonly documented comorbidity, comprising 35.5%1 to 97%19 of the various study populations.
3. Tumor characteristics
The sublocations to which the tumor extended were specified in seven studies: anterior tonsillar pillar (36.3%), soft palate (35.9%), lower gingiva (32.7%), and buccal mucosa (20.8%) were the primary sublocations. Other reported subsites were the tongue base (9.9%), mandible (5.4%), floor of mouth (5.0%), upper gingiva (4.3%), mobile tongue (4.1%), hard palate (3.6%), and tonsillar fossa (1.4%).(Tables 2, 5)
In all the studies, the tumor stage was classified using the primary tumor (T), regional node (N), and metastasis (M) classification of the American Joint Committee on Cancer Staging. Details of the considered tumor stages are provided in Table 2. Among the patients, the presence of pathologically positive lymph nodes was confirmed in 38.8% (14.5%19 to 65%21). In three studies in which the patients did not undergo surgical treatment14,16,22, only clinical lymph node metastasis (ranging from 30.2%14 to 46%16) was described. Eight articles1,10,11,13,15,19-21 reported the prevalence of histologically confirmed extranodal extension in 11.10% (0.9%11 to 43%15) of their study populations. Tumor infiltration to the underlying mandibular bone was confirmed histopathologically in 24% (14%19 to 47.6%9) of the patients in six studies9,13,18-21. The evaluation method for mandibular infiltration was not specified in two articles10,16. Only one study mentioned the invasion rate of the maxilla, 22% among 50 study subjects, which was higher than that of the mandible20.
All the included articles described the treatment modalities performed for patients. Primary surgery followed by adjuvant radiotherapy or chemoradiotherapy was the most commonly used treatment modality, delivered to 35.5% of the total study population. Differences in treatment methods according to the different T stage groups were elaborated in only one article17. According to that study, approximately half of the patients diagnosed with T stages 1 or 2 and 78.1% of those with T stage 3 or 4 underwent surgery as the primary modality of treatment. Although the other studies did not clearly specify methods of treatment according to the tumor stages, most mentioned the use of postoperative radiotherapy for patients with pathologically confirmed positive lymph nodes and margins. Among the 8 studies that described the use of chemotherapy1,15-17,20,21, only two15,17 specified their methods to be platinum based.(Tables 2, 5)
Details regarding surgical procedures and postoperative margin status were provided in ten studies1,9,11-13,15,18-21.(Tables 2, 5) Among those patients, 92% underwent aggressive surgical procedures: lip-splitting (92%), segmental or hemimandibulectomy (61.1%), radical neck dissection (44.1%) or modified radical neck dissection (23.6%), and reconstruction via free flaps (49.5%).(Table 2)
4. Recurrence
The rate of local and regional recurrence was discussed in all but two studies19,22.(Table 6) Among 770 patients, there were 180 cases of local recurrence and 65 cases of regional recurrence. In two studies17,21, the type of recurrence was not specified for 44 locoregional recurrence cases. Distant recurrence developed in 56 patients in the 12 studies that discussed distant control following treatment1,9-13,15-18,20,21.
Six studies investigated prognostic factors that affected recurrence10,15-17,19,20. Two of them found that primary surgical resection significantly increased local17,19 and locoregional control rates10,17. In two other studies, advanced tumor stage correlated with significantly decreased regional16 and local19 control rates. A significant correlation between advanced nodal stage and lower regional control rates was also detected in those studies16,19. Nodal metastasis was found to result in higher distant recurrence rates in another study10. Hitchcock reported that a higher overall prognostic stage resulted in significantly decreased locoregional control17. Another study reported that poor prognostic factors for local control were the presence of bone invasion and a second primary lesion20. In one study, however, no significant prognostic factors for recurrence could be found15.
5. Survival outcomes
Disease-free survival was reported in eight studies9-13,15,18,21 and ranged from 40%10 to 76.9%13.(Table 6) Disease-specific survival was reported in seven other studies12,14,16-18,21,22. Among those studies, six investigated the presence of significant prognostic factors for disease-specific survival10,15-17,19,20.(Table 7) Five studies found that advanced nodal stage had a negative effect on disease-free survival9-11,16,22. Similar results were found for advanced tumor stage in two studies11,22 and inferior gingival involvement11,17 in two studies. However, contrary to those findings, Bayman et al.14 reported that tumor stage or nodal status did not influence the disease-specific or overall survival rates. Treatment modality was correlated with disease-free survival in three studies10,17,22. Rizvi et al.22 mentioned other significant prognostic factors such as age, race, and overall stage.
Among the 4,633 study subjects in the nine articles that discussed 5-year overall survival1,11,12,14,16,17,19,20,22, the total 5-year overall survival rate was approximately 38.90% (26%19 to 60.6%20). Only 3-year overall survival rates of 71%9 and 59.8%18 were reported in two studies, and the evaluation period was not specified in two other articles13,21.(Tables 2, 6)
Nine studies identified prognostic factors that correlated with overall survival11,12,14,16-18,20-22.(Table 7) Nodal status was the most commonly reported prognostic factor, with five articles detecting its significant effect on overall survival11,18,20-22. Overall stage17,20,22, tumor stage11,22, treatment modality17,22, invasion to adjacent anatomic spaces11,20, margin status12, and age at diagnosis22 were also found to be significant factors affecting survival.
6. Complications and functional outcomes
Table 6 lists the rates of complications following surgery and radiotherapy documented in the analyzed articles. Of all the studies, only five investigated postoperative complications rates9,11,17,19,21, and six examined those related to radiotherapy9,10,14,16,17,19. Oro-cutaneous fistula was the most prevalent postoperative complication, with its presence confirmed in 10.6% of the 395 subjects in the five studies, followed by flap dehiscence (8.1%), wound infection (8.1%), flap necrosis (5.6%), seroma formation (3.5%), pneumonia (3.5%), percutaneous endoscopic gastrotomy (PEG) tube placement (2.5%), hematoma formation (1.0%), mandibular fracture (0.8%), and carotid rupture (0.5%).(Table 2) One study found that subjects with tumors that extended to the inferior gingiva or hard palate and those who underwent reconstruction via tongue flaps were more prone to postoperative complications than other patients11.(Table 7)
Among the 328 subjects who underwent radiotherapy in five studies10,14,16,17,19, osteoradionecrosis occurred in 11.6%. The highest prevalence of 26.6% was reported by Byers et al.19 in 1984. More recent articles reported smaller rates of osteoradionecrosis, ranging from 5.6%14 to 10%17. Details about the other complications of radiotherapy documented in the reviewed studies are specified in Table 6.
IV. Discussion
RMT SCC generally carries a worse prognosis than neoplasms originating from other regions of the oral cavity because of its unique anatomic position and relatively late diagnosis23. To the best of our knowledge, this is the first study to systematically review worldwide data for nearly 5,000 patients with RMT SCC and analyze their recurrence, survival, complications, functional outcomes, and relevant prognostic factors.
1. Significant findings
The articles retrieved for this review were all retrospective, and most of them originated from centers in North America. The patient-specific factor most commonly observed among those diagnosed with RMT SCC was smoking, followed by alcohol abuse. This is in accordance with findings from many other articles that have clearly established smoking and excessive alcohol consumption as risk factors for the development of oral SCC24-26.
A significant proportion of the reviewed subjects presented with advanced tumor stage, nodal metastasis, extranodal extension, and adjacent bone invasion. Similar results have also been reported in the previous literature. Many articles have demonstrated the strong correlation between the RMT site and advanced tumor stages27,28 and lymph node metastasis29 at the time of diagnosis. Because of its aggressive nature, one article recommended that neck dissection be performed for all cases of RMT SCC29.
It is also well known that the unique location of RMT SCCs in the posterior triangular-shaped area directly over the pterygomandibular raphe and their consequent proximity to the surrounding muscles, tonsillar pillar, maxillary tuberosity, and anterior ascending ramus of the mandible gives them a tendency to extend to other anatomic structures and erode bone, even in their early stages5. In this review, we also found that extension to several sublocations, especially the tonsillar pillar (36.3%), soft palate (35.9%), lower gum (32.7%), and buccal mucosa (20.8%), was reported in a significant number of subjects.
Mandibular invasion was pathologically confirmed in 24% of the 283 patients in six of the studies reviewed here. In 3.6% to 15.4% of those patients, signs of bony invasion could not be detected in the preoperative computed tomography (CT) scans, but the invasion was confirmed through histopathological diagnosis9,13,19. This is in accordance with the findings of Lane et al.30, who reported that the sensitivity of CT for detecting underlying bone involvement in RMT SCC was only 50%. They thus warned of the significant limitations that could exist in using negative CT scan results to determine the therapeutic approach for the overlying mandible. One limitation of the current systematic review is that most of the reviewed studies focused only on mandibular invasion; details about the maxillary infiltration rate were elaborated in only one study20. The authors of that article reported that in their series, the maxilla was more readily involved than the mandible. Because of the RMT’s proximity to the maxillary tuberosity, pterygoid plate, and soft palate, paying attention to the maxilla is also imperative when treating RMT SCCs5,20. This highlights the need to consider tumor extension to the maxilla in future studies.
The management of patients with SCC varied considerably among studies; radical resection followed by adjunctive therapy, surgery alone, radiotherapy or chemotherapy without surgery, and neo-adjuvant therapy followed by salvage surgery were all used. Not surprisingly, many of the articles analyzed by this review reported that 5-year disease-specific survival and overall survival depended on the treatment approach, with primary surgery resulting in the highest survival rates1,10,17,19,22. Due to the high rate of lymph node metastasis, mandibular infiltration, and advanced stage at the time of diagnosis, aggressive surgery is often warranted for the definitive treatment of carcinomas of the RMT region5,27-29,31. In this study, aggressive mandibular surgery, such as hemimandibulectomy (39.6%) and segmental mandibulectomy (21.5%), was undertaken for more than 60% of the reviewed populations, and either radical neck dissection (RND) (44.1%) or modified radical neck dissection (mRND) (23.6%) was selected as the neck dissection method of choice for nearly 70% of patients. Free flap reconstruction was necessary in 49.5% of patients, which suggests the extensive nature of the resection and advanced tumor stage at the time of surgery. On the contrary, for 31.90% of the study subjects, primary closure of the surgical wound was possible9,11-13,18-21. While it would be plausible to assume that most of those patients were diagnosed at early stages, neither the tumor nor overall prognostic stages for which primary closure or free flap surgery were performed were specified in any of the reviewed manuscripts. Therefore, future studies should further explore those associations and determine the precise role of reconstructive surgery in the treatment of early and advanced RMT SCCs.
In this review, despite the aggressive approach to the tumor via the lip-splitting method with or without mandibulectomy in 92% of the study population, high positive surgical margin rates of 33%12, 20%15, and 57%18 were reported. Many previous articles have demonstrated the significantly higher prevalence of positive margins in RMT tumors compared with those originating from other oral regions3,31-35. The limited surgical excision and exposure3,32-35 due to confinement from adjacent anatomic structures and the invasive nature of carcinomas arising in that location31 could explain that finding.
The total 5-year overall survival of the patients in this review was 38.9%, which is significantly lower than the 5-year survival rate of 66.2% among oral cavity cancers reported by the SEER database in 201636. This low 5-year survival rate can be partially attributed to the publication years, which date back to 198419, and the inclusion of studies that included patients who underwent radiotherapy as the primary treatment14,16. However, all but one study20 in the current review reported 5-year survival rates lower than 60%, indicating the aggressive nature of tumors arising from this subsite.
The retrieved articles also reported relatively high rates of complications following surgery and radiotherapy. Complications associated with free flaps, such as the development of oro-cutaneous fistula, flap dehiscence, wound infection, and flap necrosis, were the most prevalent. Osteoradionecrosis following radiotherapy was also commonly reported. One previous retrospective review of 1,692 patients reported the RMT subsite to be an independent factor associated with osteoradionecrosis. In another systematic review, the retromolar region, along with the tongue and floor of the mouth, was reported to be one of the tumor sites with the highest risk of developing osteoradionecrosis (ORN) owing to its proximity to the mandible and consequent direct irradiation of the bone30. Similar findings have been reported in other previous studies37-39.
2. Heterogeneity
The aim of the current systematic review was to assess all the relevant data regarding the outcomes of RMT SCC, regardless of the individual tumor and treatment characteristics of the included subjects. Therefore, the studies included in this qualitative assessment displayed significant heterogeneity in terms of baseline tumor and treatment characteristics that could not be controlled. Those methodological and clinical differences might have biased the overall results of this study.
The number of subjects per study was widely distributed from 1415 to 4,02222. Approximately 83.1% of the total study population were patients in a single study22, and thus, it is arguable that our overall results might have been overly influenced by the results of that article, resulting in biased data. Conversely, the data from some studies13,15,21 might lack reliability and statistical power because they had study population of 20 or fewer people. Thus, post-treatment complication and mortality rates might have been under-reported.
Our review included studies from different publication years and study locations. Although more than half of the analyzed studies1,9,14,15,17,18,21,22 were published in the past 10 years, two studies were published more than 20 years ago. In contrast to the latter group of studies, which used only a combination of surgery and radiotherapy, most recent publications documented the use of chemotherapy1,14-18,20,21, which could have had a positive effect on survival and locoregional control. On the other hand, no significant differences were found in the general treatment patterns of American, Asian, and European hospitals.
Also of note is the heterogeneity in treatment modality and tumor traits of interest among the individual studies. The unfavorable survival outcomes, high rates of stage T3-4 patients, and aggressive surgery might partially be due to inclusion criteria demanding advanced stage tumors9 or the need for extended operations11. It is also plausible that the results regarding treatment modality and outcomes were influenced by the studies that investigated the outcomes of only subjects who underwent radiotherapy as the primary treatment14,17.
Our objective comparison of post-treatment survival rates between the included studies was also limited by their heterogeneous methodological traits. Although survival rates were reported in all the analyzed articles, only disease-free or disease-specific survival rates were reported in three10,13,15, and among those that did provide details regarding overall survival, only 2- or 3-year overall survival was reported in three9,18,21. Therefore, only the remaining nine articles could be included in the final analysis of 5-year overall survival.
3. Other study limitations
As a systematic review, this study has several limitations: the small number of analyzed articles, search boundaries, potential subjectivity in data selection, and poor quality of evidence in the overall data.
To enhance the reliability of the current review, we sought to minimize selection bias by applying stringent inclusion criteria for the articles, and as a result, only 15 studies were included in our analysis. The language inclusion criteria could also be a limitation because relevant data from studies published in languages other than English were not retrieved. We are aware that, despite cross-checking between independent reviewers via the DSR program, only data from the articles we selected were analyzed, so the possibility of selection bias exists.
Because SCCs originating from the RMT region are relatively uncommon and because the goal of our review was to investigate the overall outcomes of all RMT SCC patients regardless of the type of treatment, all the data analyzed came from retrospective, case-controlled studies. No randomized controlled trials or cohort studies have been conducted on this topic. Therefore, the overall level of evidence for the analyzed data was poor, and that might also have introduced selection biases. Moreover, the criteria for selecting the treatment modality were not clarified in most of the reviewed studies, which could also indicate biased data. These limitations highlight the need for randomized controlled trials or cohort studies with appropriate treatment selection and adequate methods for assessing outcomes.
Despite these shortcomings, this is the first retrospective review of extensive data on the outcomes of treatment for RMT SCC, and thus it will be a useful guideline for clinicians who need to select the treatment that will yield the best outcomes for their patients suffering from this rare but dangerous disease.
XML Download