PDF
Citation
Print
INTRODUCTION
Sinonasal cancer is a rare malignant tumor, accounting for about 3% of head and neck tumors [1,2]. The most common pathology associated with sinonasal cancer is squamous cell carcinoma (SqCC) [1,3]. This cancer is often diagnosed at advanced stages because it can remain asymptomatic until the tumor invades or approaches critical areas such as the skull base, nerves, orbit, or palate [4]. Historically, the standard treatment for sinonasal cancer involved en-bloc resection of the primary tumor. However, this method can be highly invasive and may lead to unfavorable cosmetic outcomes. In cases of orbital invasion, en-bloc resection might require orbit exenteration (OE), which significantly affects a patient’s quality of life. As a result, radiotherapy (RT) and concurrent chemoradiotherapy (CCRT) have become alternative treatment options and are used in selected cases [5-9]. Additionally, evidence suggests that neoadjuvant chemotherapy (CTx) can downstage advanced sinonasal cancers and aid in preserving the eyeball during treatment [10]. However, orbital invasion is considered a poor prognostic factor, sparking debates about the role of surgery as the definitive treatment [11-14]. Moreover, there remains no consensus regarding the choice of surgical technique, particularly between tumor resection with OE and tumor resection with orbit preservation (OP) [14-17]. Typically, indications for OE include involvement of structures such as the medial rectus muscle, optic nerve, ocular bulb, or the skin near the eyelid [15,18]. However, for SqCC, some studies have indicated that preserving the orbit does not significantly increase the rate of local recurrence [19,20]. Nonetheless, some research has not conclusively linked OP with improved overall survival (OS) [21]. Especially in cases with limited invasion of orbital fat, the decision between tumor resection with OE or preservation remains highly debated due to the absence of clear differences in outcomes [19,20,22,23].
The objective of our study was to evaluate the oncological outcomes of sinonasal SqCCs with orbital invasion, considering various factors such as pathological characteristics, the extent of orbital invasion, and treatment options.
MATERIALS AND METHODS
This study protocol received approval from the Institutional Review Boards of Seoul National University Hospital and Seoul National University Bundang Hospital (No. 2210-018-1366) with a waiver of informed consent and was conducted in accordance with the principles of the Declaration of Helsinki.
Subjects
This retrospective study was conducted at two tertiary care centers, namely Seoul National University Hospital and Seoul National University Bundang Hospital. The study encompassed a ten-year period, from January 1, 2009 to December 31, 2018. The study population consisted of patients who presented with primary sinonasal SqCC with orbit involvement at the time of their initial diagnosis. Excluded from the study were patients with a follow-up period of less than 12 months for assessing survival, those with distant metastasis at presentation, individuals in a palliative care context (including those with extensive cavernous sinus involvement, encasement of the internal carotid artery, or extensive infiltration of brain parenchyma), and those with incomplete medical records, including images, surgical reports, or follow-up information. Ultimately, 68 patients met the inclusion criteria and were included as subjects for analysis. A comprehensive review of their medical records, encompassing demographics, radiological assessment of orbit invasion extent, treatment modalities, survival data, and recurrence outcomes, was conducted.
Classification of the extent of orbit invasion
To assess the radiological extent of orbit invasion, we utilized the classification system described by Turri-Zanoni et al. [24] in 2019, which is a modified version of the classification originally proposed by Iannetti et al. in 2005 [18]. In Turri-Zanoni’s classification, orbit invasion is classified as follows: grade 1 when there is erosion or destruction of the orbital bony wall (Lamina papyracea), grade 2 when there is invasion of the periorbital layer and/or focal invasion of the extraconic periorbital fat, grade 3 when there is invasion of the orbital contents (anterior 2/3 of the orbit), including extra-ocular muscles, optic nerve, ocular bulb, and the skin overlying the eyelids, and grade 4 for involvement of the orbital apex. Illustrative examples of each grade are provided in Fig. 1. In our study, grading based on Turri-Zanoni’s classification [24] for all study subjects was performed by two expert otorhinolaryngologists.
Treatment algorithm
The general treatment algorithm for advanced sinonasal SqCC is as follows. First, we conduct a comprehensive assessment of the patient’s medical history and imaging in collaboration with the Department of Radiation Oncology and the Department of Hemato-Oncology to determine the appropriateness of initiating neoadjuvant CTx. If the neoadjuvant CTx is not performed, the initial treatment involves surgery, CCRT, RT, or CTx. Subsequently, postoperative RT or CCRT is administered on a case-by-case basis. However, if the neoadjuvant CTx is performed, neoadjuvant CTx is the initial step. The subsequent course of treatment varies depending on the patient’s response. If the cancer response is still in-operable or the patient refuses the operation, CCRT or RT is initiated after neoadjuvant CTx. Conversely, if the response is stable disease (SD), progressive disease (PD), or partial response (PR) with downstaging, surgical intervention is pursued, with or without OE. Postoperative CCRT or RT is also determined on a case-specific basis. The algorithm is depicted in Fig. 2, detailing the count of patients included on a case-by-case basis.
As a result, the definitive treatment approach is primarily divided into two methods: treatments involving OE, represented as “OE” henceforth, and treatments without OE, referred to as “OP” from now on. OE unequivocally refers to surgical procedures with OE as the definitive treatment. OP encompasses not only patients who undergo surgical treatment without OE as the definitive approach (OP with surgery) but also patients managed with non-operative treatments as the primary strategy, such as CTx or CCRT (OP with other treatments). OP with surgery in grade 4 was done like debulking surgery involving optic nerve decompression. But the debulking was very wide and achievements of the functional eye were not perfect. On the other hand, OP with surgery in grades 1, 2, and 3 was done by removing all the tumor and shaving some suspicious parts of the orbits by delicate dissection from wide view or endoscopic view. The factors that determined the OP surgery and OE were mainly surgeon’s preference, and the changes in treatment strategy or trends during 10 years. The neoadjuvant CTx regimen involves a three-week cycle and includes cisplatin, docetaxel, and fluorouracil. On the other hand, the CCRT regimen consists of conventional standard fractionated RT, delivering a dose of more than 60 Gy, alongside concurrent CTx with weekly administration of cisplatin.
Statistical analysis
The definitions and methods employed in our study were as follows. OS was measured as the time elapsed from the date the patient treatment initiated to the date of their last visit or the date of their demise. Disease-free survival (DFS) was calculated as the time from the date the patient treatment initiated to the date of their last visit without a recurrence of cancer or signs thereof. We utilized the 8th American Joint Committee on Cancer guidelines for cancer staging. Our study primarily focused on the analysis of the relationships between OS, DFS, and factors such as the extent of orbit invasion, T, N stage, treatment of orbit, and neoadjuvant CTx. Survival rates were assessed using the Kaplan-Meier method, and comparisons between survival rates, including OS and DFS, were conducted using the log-rank method (Mantel-Cox). Statistical analysis was performed using SPSS Statistics for Windows version 22.0 (IBM Corp.), with statistical significance defined as a P<0.05.
RESULTS
Overall results for patients with sinonasal SqCC exhibiting orbital invasion
In this study, the average age at diagnosis among the 68 patients was 63.0±13.1 years, comprising 49 men and 19 women. According to Turri-Zanoni’s classification of orbital invasion [24], the distribution was as follows: 18 cases of grade 1, 23 of grade 2, 20 of grade 3, and 7 of grade 4. The 5-year OS (5Y OS) rate was 89.0%±4.3%, and the 5-year DFS (5Y DFS) rate was 35.1%±7.2%. The corresponding Kaplan-Meier curves are shown in Fig. 3. Staging revealed 19 cases at T3, 31 at T4a, and 18 at T4b. Additionally, nodal involvement was categorized as 53 cases at N0, 7 at N1, 3 at N2a, 2 at N2b, and 3 at N2c. Neoadjuvant CTx was administered to 37 patients. Surgery as definitive treatment was performed on 32 patients. Of these, 19 received RT and 10 underwent CCRT as adjuvant treatment, while three patients did not receive any adjuvant therapy. Eight patients underwent orbital exenteration, and 24 underwent orbital preservation surgery. A total of 36 patients received OP combined with other treatments, which included 26 CCRT, 5 RT alone, and 5 CTx (Fig. 2).
Survival outcomes of SqCC patients by treatment
We evaluated the prognosis of 68 SqCC patients who received one of three distinct types of definitive treatment: OE, OP with surgery, and OP with other treatments, which included RT, CTx, and CCRT. Eight patients underwent OE, 24 received OP with surgery, and 36 were treated with OP using other treatments. Notably, the distribution of patients across these treatment categories was even, based on orbit grade and T stage, ensuring balanced representation within each group (Supplementary Tables 1 and 2).
No significant differences in OS were observed between SqCC patients who underwent surgical treatments and those who received non-surgical definitive treatments. Similarly, no significant differences were observed among the three different definitive treatments. However, when examining DFS, significant differences emerged between surgical and non-surgical definitive treatments (overall comparison: χ2=9.057, P =0.003) and among the types of treatments (OE, OP with surgery, and OP with other treatments; overall comparison: χ2=9.071, P =0.011) as shown in Table 3 and Fig. 5. Notably, patients who underwent OP with surgery had a significantly longer DFS than those who received OP with other treatments (pairwise comparison: χ2=7.556, P = 0.006) (Supplementary Table 3).
We conducted an analysis focusing on SqCC patients with orbital invasion grades 1, 2, and 3, excluding the most severe cases (grade 4). In this subgroup, treatment distribution included 7 patients undergoing OE (5Y OS, 100%; 5Y DFS, 60.0%±21.9%), 22 patients treated with OP and surgery (5Y OS, 100%; 5Y DFS, 51.8%±15.3%), and 32 patients treated with OP and other treatments (5Y OS, 79.8%±8.2%; 5Y DFS, 23.0%±8.3%). No significant differences were observed in OS rates among these groups (overall comparison: χ2=5.296, P =0.071). However, significant differences were noted in DFS rates (overall comparison: χ2=6.725, P =0.035). Pairwise comparisons revealed that patients treated with OP and surgery had a significantly improved DFS rate compared to those treated with OP and other treatments (χ2=4.961, P =0.026). No significant differences were observed in pairwise comparisons between other definitive treatment categories. Additionally, we performed an analysis on SqCC patients with orbital invasion grades 2, 3, and 4, excluding the mildest cases (grade 1). This subset included 6 patients receiving OE (5Y OS, 100%; 5Y DFS, 40.0%±21.9%), 16 patients treated with OP and surgery (5Y OS, 93.3%±6.4%; 5Y DFS, 65.8%± 15.0%), and 28 patients treated with OP and other treatments (5Y OS, 85.5%±8.0%; 5Y DFS, 18.0%±8.0%). The analysis yielded similar results: no significant differences in OS rates among these groups (overall comparison: χ2=1.236, P =0.539) and significant differences in DFS rates (overall comparison: χ2=8.149, P =0.017). Patients treated with OP and surgery had a significantly improved DFS rate compared to those treated with OP and other treatments (χ2=7.280, P =0.007), with no significant differences observed between other definitive treatment categories in pairwise comparisons. For further clarification, we also conducted a subgroup analysis in SqCC patients with orbital invasion grades 2 and 3, excluding both severe (grade 4) and mild (grade 1) cases. Among the 43 SqCC patients in this subgroup, the treatment distribution included five patients undergoing OE (5Y OS, 100%; 5Y DFS, 50.0%±25.0%), 14 patients treated with OP and surgery (5Y OS, 100%; 5Y DFS, 61.5%±16.6%), and 24 patients treated with OP and other treatments (5Y OS, 83.5%±9.0%; 5Y DFS, 20.1%±8.9%). There were no significant differences in OS or DFS rates among the three treatments (overall comparison: χ2=2.942, P =0.230 for OS; χ2=5.473, P = 0.065 for DFS). However, the trend in DFS rates was similar to that observed in the analyses of all SqCC patients or those with grades 1–3 or grades 2–4 orbital invasion.
Finally, we analyzed SqCC patients who received neoadjuvant CTx. In this subgroup analysis, patients were divided into three categories. The first group consisted of patients who achieved PR and subsequently underwent surgery (including both OE and OP with surgery) as their definitive treatment (11 patients: 3 OE and 8 OP with surgery). The second group included patients who exhibited an SD or PD response and also received surgery as their definitive treatment (6 patients: 3 OE and 3 OP with surgery). The third group was made up of patients with an SD or PD response who received RT, CCRT, or CTx (OP with other treatments) as their definitive treatment (20 patients). No significant differences were found in OS rates among these groups. However, the DFS rates revealed that patients with an SD or PD response who underwent non-surgical treatments had significantly lower DFS rates than others (overall comparison: χ2=11.081, P =0.004) (Table 4, Supplementary Table 4). We also compared OS and DFS among the different definitive treatments (OP with other treatments, OP with surgery, and OE) following neoadjuvant CTx. The treatment distribution included five patients undergoing OE (3Y OS, 100%; 3Y DFS, 60.0%±21.9%), 12 patients treated with OP with surgery (3Y OS, 100%; 3Y DFS, 60.0%±21.9%), and 20 patients treated with OP with other treatments (3Y OS, 71.3%±12.6%; 3Y DFS, 18.3%±9.5%). While no significant differences in OS rates were observed among these groups (overall comparison: χ2=5.002, P =0.082), significant differences were noted in DFS rates (overall comparison: χ2=11.132, P =0.004). Pairwise comparisons showed that patients treated with OP with surgery had a significantly higher DFS rate than those treated with OP with other treatments (χ2= 7.777, P =0.005). Similarly, patients treated with OE also demonstrated a significantly improved DFS rate compared to those treated with OP with other treatments (χ2=5.439, P =0.020). However, no significant differences were observed in pairwise comparisons between OP with surgery and OE.
DISCUSSION
In this study, we analyzed OS and DFS among 68 patients with sinonasal SqCC that had invaded the orbit, taking various factors into account. The 5-year OS rate was 89.0%±4.3%, and the 5-year DFS rate was 35.1%±7.2%. Both OS and DFS exhibited significant differences based on the T stage, a common trend observed in many cancer types. However, no significant differences in OS and DFS were noted when considering the grade of orbital invasion as proposed by Turri-Zanoni et al. [24]. In a detailed analysis of SqCC cases, we found that choosing tumor resection with OP (OP with surgery) as the definitive treatment significantly extended DFS compared to cases where surgery was not the selected definitive treatment (OP with other treatments). Additionally, there was no significant difference in DFS between patients who underwent OE as the definitive treatment and those who underwent OP with surgery. Given the lack of significant differences in OS among the groups undergoing OE, OP with surgery, and OP with other treatments, it appears that OP with surgery may be the optimal definitive treatment option where feasible. This conclusion is supported by the fact that OP with surgery not only extends DFS but also ensures survival in SqCC cases with orbital invasion. The analysis of subgroups, excluding patients with either mild or severe orbital invasion (grades 2, 3 and 3 cases or grades 1, 2, and 3 cases), yielded consistent results. Regarding neoadjuvant CTx, surgical definitive treatments were associated with longer DFS compared to non-surgical definitive treatments.
Regarding differences in OS and DFS rates based on the extent of orbital invasion, the literature presents conflicting findings. Some studies have found no significant differences in OS, disease-specific survival, and progression-free survival rates in sinonasal cancers relative to the extent of orbital invasion, which is consistent with our results [17]. However, other research has shown that OS and DFS rates significantly vary with different grades of orbital invasion [24]. Additionally, a study focused on the pre-treatment assessment of local extension in sinonasal cancer [25]. That research assessed the diagnostic performance of CT and magnetic resonance imaging in evaluating local invasion by comparing these imaging techniques to histopathological data. It concluded that the signs of orbital invasion detected in imaging had a low positive predictive value. This variability in diagnostic accuracy could potentially explain the differing outcomes concerning the relationship between the extent of orbital invasion and the prognosis of sinonasal cancer. The imperfect diagnostic accuracy of these imaging methods may lead to diverse results.
In terms of T staging, several studies, including our own, have demonstrated significant differences in OS or DFS across T stages [24]. A key distinction of our study is that we focused exclusively on a subset of patients with SqCC. While orbital invasion is recognized as a critical prognostic factor in sinonasal cancer, our findings indicate that relying solely on the presence of orbital invasion to predict the cancer’s prognosis may be inadequate. This insufficiency is likely due to the potential for sinonasal cancer to invade other adjacent structures, including the dura, skull base, skin, and lymph nodes. Although our analysis did not extend to prognosis based on N staging due to the limited number of patients with lymph node metastasis, lymph node metastasis is also considered a significant prognostic factor. Therefore, we recommend that treatment decisions should consistently take into account the full extent of the cancer’s invasion.
Several studies have reported varying results when comparing OE to tumor resection with OP, aligning with our findings on OS and DFS rates in SqCC patients [14,17]. However, one study indicated significantly lower OS and DFS rates in the OE group compared to the OP group, especially in patients with invasion of the anterior two-thirds of the orbit [24]. This issue remains highly controversial, as another study involving patients with invasion beyond the orbital periosteum into the orbital soft tissues showed a higher OS rate in the OE group than in the OP group [26]. However, these two studies [24,26] did not focus on a specific pathology, but included cases of sinonasal cancer with various underlying pathologies. Additionally, when comparing OP treatments to OE outcomes, the OP treatments included both surgical and non-surgical approaches, such as CCRT, as the definitive treatment [24,26]. In the context of SqCC, some studies have reported that OP, whether through surgical or non-surgical means, is not significantly associated with higher local recurrence rates [19,20] or OS rates [21]. Notably, a recent meta-analysis on OP in sinonasal cancer showed a slight preference for OP in SqCC cases for better outcomes, although the difference was not statistically significant [22]. Our study has significant strengths compared to these previous studies. One key aspect is our division of orbit-preserving definitive treatments into surgical and non-surgical categories. Additionally, our study’s focus on treatment analysis for a single pathology (SqCC) is another strength, as there were differences in prognosis based on pathology. Furthermore, our study benefits from a large sample size, encompassing 68 SqCC cases, all from two connected tertiary centers.
Recent studies have highlighted the effectiveness of neoadjuvant CTx in treating advanced SqCC of the sinonasal region [10]. Our study further supports these findings, demonstrating improved outcomes with surgical definitive treatments compared to non-surgical approaches following neoadjuvant CTx. Additionally, outcomes in terms of DFS were better for patients who underwent OP with surgery and OE after receiving neoadjuvant CTx, compared to those who received other treatments. In our study, six grade 3 patients underwent OP with surgery as their definitive treatment, all following neoadjuvant CTx (Supplementary Table 5). This preoperative treatment reduced the tumor size, enabling surgeons to perform OP with less extensive removal of extraocular muscles. Instead, surgeons could focus on dissecting only the suspected parts of the muscles, achieving clear margins (R0) in three patients and near-clear margins (R1) in another three. Conversely, eight grade 2 patients received OP with surgery as their definitive treatment, with three undergoing neoadjuvant CTx beforehand. This suggests that neoadjuvant CTx is beneficial not only for improving the prognosis of cancer treatment, but also for facilitating tumor resection while preserving the orbit, rather than resorting to OE. In our opinion, this approach is particularly important for managing grade 3 patients, where the benefits of neoadjuvant CTx in facilitating less invasive surgical interventions are most pronounced.
Over the past few decades, there has been a tendency to recommend and perform OE as the extent of orbital invasion increases. However, as outlined in our study protocol, the use of neoadjuvant CTx and adjuvant therapies, including postoperative RT, CTx, or CCRT, seems to offer a more viable option for tumor resection with OP when surgery is feasible as the definitive treatment. Essentially, neoadjuvant CTx and adjuvant therapies are crucial for preserving the orbit. Based on our findings and collective experiences, we propose initiating neoadjuvant CTx as the first-line treatment for patients with grade 2, 3, or 4 orbital invasion in sinonasal cancer, whenever possible. If there is evidence of a positive treatment response, tumor resection with OP should then be considered as the primary surgical approach, if feasible. However, if the disease progresses, OE may need to be considered as a last resort.
Nonetheless, our study has several limitations. First, it included only Korean patients from two centers. Further research incorporating multiple centers and a more ethnically diverse patient population is warranted. Second, the relatively small number of patients may have influenced the results of our analysis, for example, due to potential bias from selection issues. Additionally, because our study was retrospective in nature, the included patients may not fully represent each subgroup. Moreover, the grading of orbital invasion was conducted by two expert otorhinolaryngologists. While both experts were highly skilled in interpreting rhinology-related images, the grading process may have introduced some degree of subjectivity. Finally, our additional analysis of subgroups comprised of grades (1, 2, and 3), (2, 3, and 4), and (2 and 3) revealed that for ambiguous cases, which always concern physicians regarding definitive treatments, OP with surgery appears to be an appropriate option among the three definitive treatments, though not significantly preferable. This result may stem from the fact that grade 3 encompasses a wide range, making OP with surgery practically impossible in many cases. In other words, those cases might have been included in the OE group or the OP with other treatments group. However, in grade 2 patients, all eight cases of OP with surgery were performed with R0 resection, and there were three patients who experienced recurrence (mean recurrence time: 40.33 months). Among six grade 3 patients who underwent OP with surgery, there were three patients with R0 resection and three with R1 resection, and two patients experienced recurrence (mean recurrence time: 22 months). Therefore, although further research SqCC with grades 2 and 3 orbital invasion is needed, we suggest that OP with surgery can still be considered a preferable definitive treatment choice in grade 3 patients.
Orbital invasion grading based on imaging did not accurately predict the prognosis of sinonasal cancer with orbital invasion. Surgical treatment, supported by neoadjuvant CTx, postoperative CTx, and CCRT, consistently showed improved survival outcomes for sinonasal SqCC with orbital invasion compared to non-surgical approaches. Within the surgical options, both tumor resection with OP and OE provided similar survival outcomes, independent of the orbital invasion grade.
HIGHLIGHTS
▪ Opting for tumor resection with orbit preservation as the definitive treatment has been shown to prolong disease-free survival in cases of squamous cell carcinoma, compared to non-surgical definitive treatments.
▪ No significant differences in disease-free survival were observed between patients who underwent orbit exenteration and those who received tumor resection with orbit preservation as the primary treatment for squamous cell carcinoma.
▪ The extent of orbital invasion does not significantly affect overall survival or disease-free survival rates in patients with sinonasal squamous cell carcinoma.
XML Download