Response to induction chemotherapy in sinonasal malignancies: A single‐institutional experience

Sinonasal malignancy (SNM) is a heterogeneous group of diseases for which induction chemotherapy (IC) may reduce tumor burden. The purpose of this study was to characterize the response to IC in SNM as a prognostic factor through its effect on survival.


| INTRODUCTION
Sinonasal malignancy is a rare and difficult-to-treat subgroup that only comprises 0.2% of malignancies in the world and 3% of all head and neck malignancies. 1 Additionally, sinonasal malignancies pose significant challenges in terms of treatment, as they generally present in advanced stages. Oncologic outcomes have not changed significantly over the past 30 years, with an overall 5-year survival of 58% for the years 2010-2017. 2 Furthermore, these tumors have heterogeneous behavior and histology, which adds to the complexity of their management. Treatment is generally guided by histologic findings. While most sinonasal malignancies require multimodal therapy, surgery is often the primary treatment strategy for resectable sinonasal malignancy. Conventional data suggest that surgery, as part of multimodal therapy, yields the best overall survival outcomes in properly selected patients. 3,4 Surgery may consist of open or endoscopic approaches, but regardless of the approach the proximity of the tumor to Sarah C. Nyirjesy and Rachel Fenberg are co-first authors.
This study was presented as an oral presentation at NASBS 2022 Annual Meeting, 18-20 February 2022 in Phoenix, Arizona. vital structures including the carotid arteries, orbits, brain, and cranial nerves may limit the degree of resection or result in significant morbidity. 5 As with malignancies of other subsites, optimal surgical outcomes for sinonasal malignancy relies upon obtaining negative margins, and survival is improved when this is achieved. [6][7][8] A review of the literature suggests that multimodal approaches including systemic therapy may be beneficial to many patients and ultimately improve outcomes. 5,[9][10][11] Induction (neoadjuvant) chemotherapy is an emerging strategy in multimodal therapy. In induction chemotherapy (IC), an agent is given prior to definitive surgery and/or chemoradiation therapy, to decrease tumor size with the goal of improving oncologic and quality of life outcomes of definitive treatments. 10 Others have suggested the benefits of IC for select patients with sinonasal malignancy, such as sinonasal undifferentiated carcinoma (SNUC) and sinonasal squamous cell carcinoma (SCC), but due to the rarity of this disease and relatively small sample sizes, it is challenging to translate these findings to large-scale recommendations. 11 The primary aim of this study was to characterize the response to IC in sinonasal malignancy as a prognostic factor through its effect on overall survival and progression-free survival. Secondarily, we sought to determine whether IC had an effect on surgical resectability, survival, and locoregional recurrence in the two most common histologic types of sinonasal malignancies at our institution, SCC and SNUC.

| METHODS
The Ohio State University Wexner Medical Center Institutional Review Board (IRB) approved this investigation, comprising a retrospective cohort analysis of all treatment-naïve patients who received IC for sinonasal malignancy at a single quaternary referral center between July 2010 and February 2019. Patients diagnosed with sinonasal malignancy during this period were identified using ICD-9 (160.0, 160.2, 160. 3, 160.4, 160.5, 160.8, 160.9, 192.0) and ICD-10 codes (C31.0, C31.1, C31.2, C31.3, C31.8, C31.9, C72.2). Retrospective electronic chart reviews were subsequently conducted to identify those who received IC to be included in the primary analysis. Patients with the histologic subtypes SNUC and SCC who did not receive IC were included for secondary analyses. Patients were excluded if they were <18 years old, had received prior cancer treatment for their sinonasal malignancy, or had known distant metastasis. Data including patient demographics, time of diagnosis, tumor staging, location, pathology, treatment regimens, recurrence, and vital status were collected.
Pre-and post-IC imaging studies were evaluated and tumor volume measured by a single investigator (MH) following RECIST1.1 guidelines. 12 Tumor response was defined as complete response (CR; no remaining target lesions and reduced lymph node size to <10 mm), partial response (PR; at least 30% reduction in sum of target lesion diameters from baseline), progressive disease (PD; absolute increase of at least 5 mm with increased sum of diameters Began induction chemotherapy (N=42)

Completed induction chemotherapy (N=36)
Did not complete induction chemotherapy (N=6) x of target lesions by 20% compared to the smallest baseline lesion, or appearance of new lesions), or stable disease (SD; inadequate change in size to qualify for PR or PD). A CR or PR to induction chemotherapy was considered favorable, while SD and PD were considered unfavorable. Primary outcomes, including 5-year overall survival and 5-year progression-free survival rates, were calculated using Kaplan-Meier analysis (Figures S1-S3, Supporting Information). Overall survival (OS) was defined as the time from the date of diagnosis to most recent follow-up, loss to follow-up, or death. Progression-free survival (PFS) was defined as the time from the beginning of treatment (date of first infusion) to the point of disease progression, recurrence, or death due to the disease. Disease-free survival (DFS) was calculated as the time from the last day of definitive treatment to recurrence or death. Categorical comparisons were made using a likelihood Chi-square test or Fisher's exact test. All analyses were conducted using Strata (Stata Statistical Software, Release 17; StataCorp LLC, College Station, TX).
Secondary outcomes sought to further characterize patients with the two most common histologic subtypes at our institution, SNUC and SCC. All SNUC patients were tested for SMARCB1-deficient, and none were identified. Patients with SNUC or SCC who did not undergo IC were compared to those treated with IC for disease characteristics, definitive treatment, and survival outcomes.

| Patient characteristics
A total of 106 patients were identified as being diagnosed with sinonasal malignancy at our institution during the study period. Of these, 43 patients were identified as having undergone IC whereas the remaining 64 underwent upfront definitive treatment ( Figure 1). One patient was excluded due to metastasis to the liver. The remaining 42 patients were included in the primary analysis. Based on recommendations from multidisciplinary tumor board discussions, individualized for each patient and based on disease burden and local invasion, the median number of completed IC cycles was 3 (min-max: 1-8 cycles). The mean age of all patients was 56.4 years, and 32 (74.4%) patients were male (Table 1). Of patients who began IC regimens, 7 (16.7%) had progressive disease, 9 (21.4%) had stable disease, 21 (50.0%) had a partial response, and 5 (11.9%) had a complete response. Two patients died prior to completion of the planned number of IC cycles, both elected for hospice care due to clinical disease progression after one infusion. General characteristics of the patients who underwent IC are found in Table 1. The most common primary site was the ethmoid sinus (N = 28). All patients had T3 or greater disease, and 39.5% had nodal disease on presentation. The most common histologic subtype was sinonasal undifferentiated carcinoma (SNUC, N = 19) followed by squamous cell carcinoma (SCC, N = 13). The most common chemotherapy regimens used were Cisplatin + Docetaxel + 5-fluorouracil (5-FU) T A B L E 2 Demographics and tumor characteristics of patients with sinonasal undifferentiated carcinoma (SNUC) comparing those who received IC to those who did not   (Figure 1). There were no differences in age, sex, tumor site, tumor staging, or chemotherapeutic regimen between favorable and unfavorable response groups. There was a statistically significant difference between the favorable and unfavorable groups for cribriform/skull base involvement, with 92.3% of patients who responded favorably having involvement compared to 68.75% of patients who had an unfavorable response ( p < 0.05).

| Induction chemotherapy response
The median time of follow-up was 29 29.2%, p < 0.001) and 5-year (66.8% vs. 9.7%, p < 0.0001) OS compared to those with an unfavorable response (Figure 2A). Similarly, those with a favorable response to IC had a higher 2-year (69.0% vs. 12.5%, p < 0.0001) and 5-year PFS (56.8% vs. 0%, p < 0.001) ( Figure 2B). Survival analyses were not used to compare CR and PR due to the small number of patients achieving a complete response.

| Sinonasal undifferentiated carcinoma
Six patients were identified in the initial chart review with sinonasal undifferentiated carcinoma (SNUC) that were not treated with IC. Patient demographics and general tumor characteristics are listed in Table 2. A large proportion of SNUC patients (72%, N = 18) received IC. For all patients with SNUC, 2-year OS was 71.5%, 5-year OS was 54.2%, 2-year DFS was 57.6%, and 5-year DFS was 51.2%.
Patients with SNUC who received IC were more likely to present with invasive disease (of the orbit, cribriform, dura, or brain parenchyma) than those who did not receive IC, but there were no differences in survival outcomes. Patients with SNUC who received IC were more likely to present with invasive disease. There was no difference in 2-year OS or DFS compared with those who did not receive IC ( Figure 3A,B). Only 11.1% of the patients with SNUC who were treated with IC underwent subsequent surgery compared to 71.4% of SNUC patients who did not receive IC.

| Sinonasal squamous cell carcinoma
Forty-five patients with sinonasal squamous cell carcinoma (SCC) that did not receive IC were identified for secondary analyses. General characteristics are shown in Table 3. In patients with SCC, 22.4% were treated with IC, and there were no differences in demographics between groups. For all patients with SCC, 2-year OS was 70.7%, 5-year OS was 54.2%, 2-year DFS was 65.8%, and 5-year DFS was 61.4%.
Although SCC patients treated with IC were more likely to have invasive disease (involving the orbit, skull base, dura, or brain parenchyma), they had no significant differences in 2-year OS and DFS compared to those who did not receive IC ( Figure 4A,B). Prior to completion of the number of planned IC cycles, 15.4% of patients with SCC receiving IC underwent surgical debulking, and 30% of patients who completed IC underwent later surgery versus definitive surgical management in 80% of patients with SCC who did not receive IC.

| DISCUSSION
IC has emerged as a promising option in the management of locally advanced sinonasal malignancy. In the present study, we examined 42 patients with sinonasal malignancy treated with IC and found that the majority (62%) had a favorable response to IC, while fewer had stable (21%) or progressive (17%) disease. A favorable response to IC was significantly associated with improved T A B L E 3 Demographics and tumor characteristics of patients with sinonasal squamous cell carcinoma (SCC) comparing those who received IC to those who did not  overall (OS) and progression-free (PFS) survival. In subset analyses of the two most common pathologies in our series, squamous cell carcinoma (SCC), and sinonasal undifferentiated carcinoma (SNUC), patients treated with IC had comparable survival to those who did not undergo IC despite having more invasive disease. The response to IC in our cohort is comparable to prior literature. At least a partial response is reported in 60%-80% of patients with advanced SCC. 11,13,14 While results are less frequently reported for other pathologies, response rates seem to fall in a similar range. [15][16][17][18] In our cohort of all sinonasal malignancy that underwent IC, there was no significant difference in IC response based on tumor histology, although this analysis is limited by very small sample sizes in rarer histologies. The benefits of IC may include decreased morbidity/toxicity of definitive treatment due to tumor response, including possible orbital preservation. 11,13,19 Of increasing interest in recent years has been the use of response to IC as a prognostic tool. Our findings support an association between a favorable response to IC and improved survival in advanced sinonasal malignancy. IC response has been found to predict survival in patients with SCC of multiple subsites of the head and neck. 10,20 Several recent studies made similar conclusions for sinonasal SCC. 11,13,14,21 For example, in a cohort study of 58 patients, Hirakawa et al. reported 93% disease-free survival in patients with a favorable response to IC versus 53% with unfavorable response. 14 As a result, a recent best practice statement was released supporting the use of tumor response to IC as an effective prognostic tool that can guide definitive treatment selection in sinonasal SCC. 22 Our findings uniquely showed a positive prognostic impact of response to IC in a cohort including all histologic types of sinonasal malignancy, which has not been previously well-established. Su et al. reported on 15 patients with esthesioneuroblastoma (ENB) and found a nonsignificant positive trend between complete response to IC and improved disease-free survival. 15 Patil et al. analyzed 25 patients with ENB or neuroendocrine tumors and found that patients with response to a resectable level had significantly improved PFS. 17 Our cohort included a relatively large number of patients with SNUC and SCC, several of which had partial or complete response to IC. Therefore, our findings may represent a prognostic value in these pathologies specifically. It is difficult to draw conclusions from our data regarding rarer entities, such as SMARCB1 deficient carcinoma, and those that rarely undergo IC, such as adenocarcinoma. Potential explanations for the prognostic impact of IC previously posed in the literature include: enhanced radiosensitivity, 10,20,23 an association between poor response to chemotherapy and poor response to radiation, 24 decreasing tumor burden to a point of being surgically resectable, 11,25,26 and administering chemotherapy while tumor blood supply is still intact. 27,28 Given the relative frequency of SCC and SNUC in our cohort, we also sought to understand whether IC is associated with a survival impact. For patients with SNUC and SCC, we did not demonstrate improvements in survival when comparing patients who received IC to those who did not. Prior reports do not provide clear insight. London  IC, several of whom had recurrences (although no statistical comparison was performed). 29 Orlandi et al. reported on a cohort of 69 patients with epithelial sionasal malignancy treated with IC and concluded their overall survival of 63 months was higher when compared to previous retrospective reports. 30 Our dataset did not show significantly worse survival in those who underwent IC despite the presence of more pre-treatment local invasion. This suggests that the utilization of IC is not worse than proceeding with upfront surgery or CRT, although conclusions are limited by our sample size. A meta-analysis or ideally a randomized controlled trial would be beneficial in confirming these findings.
Our data also highlight the different approaches currently used for differing histologic subtypes. IC is becoming a critical component for the treatment of SNUC. 10,29 Our results highlight the use of IC in SNUC, with a majority of patients with SNUC receiving IC, and a small number treated with adjuvant surgery in earlier periods when the effectiveness of IC followed by CRT had not been clearly established. This trend follows the evolution of data in treatment, as IC has now become standard of care in SNUC patients, with surgical management reserved only for the rare nonresponders. 29 In contrast, for SCC, IC was not recommended routinely. IC is typically only used for SCC in poorly differentiated cases that present at advanced stages, typically T3-T4. 5,31 In our patient population, only 10 out of 55 patients with SCC received IC. All 10 patients had T4a or T4b disease. Only 3/10 patients who received IC underwent surgery following IC. Despite having more extensive disease at presentation, patients treated with IC with SNUC and SCC had similar OS and PFS to those not treated with IC. This suggests that IC may be the least morbid treatment for those with extensive local invasion, which, although not specifically examined in this report, is consistent with single institution and national database studies. 25,29,32,33 These reports suggest that IC can result in fewer distant metastases, reduced tumor bulk, and higher likelihood of administering a complete radiation dose to local structures in sinonasal malignancy. 33,34 The major limitations of this study include those that are inherent of a single-institution retrospective analysis and the small sample size, which can be attributed to the rarity of these tumors and lack of routine use of IC for less advanced disease. A larger population, which will likely require multi-institutional data contribution, is needed to study the relationship between tumor histology and IC response, IC response by regimen, and outcomes based on definitive treatment plans.
In conclusion, IC response in sinonasal malignancy offers valuable prognostic information, as responders across all histologic subtypes have improved survival. In SNUC and SCC, patients treated with IC presented with more locally invasive disease than those not treated with IC, but 2-and 5-year survival outcomes remained similar between groups. Nevertheless, data remain limited to national databases and single institutional reviews, and a larger population is needed to fully elucidate its therapeutic role in sinonasal malignancy.