Sinonasal tumors are a rare entity [ The subclassification of sinonasal tumors into esthesioneuroblastoma, sinonasal tumor with neuroendocrine differentiation, and sinonasal tumor with poor differentiation helps as the outcome differs according to the exact subtype [ In all these three subtypes surgical resection with or without adjuvant radiation remains the cornerstone of management [ The need for of systemic treatment is felt both with radiation when given in curative setting (only chemoradiation) and in adjuvant setting (chemoradiation postsurgical resection) in locally advanced tumors [
However the role of neoadjuvant chemotherapy (NACT) in locally advanced sinonasal malignancies is largely unaddressed. It is an interesting prospect considering anatomical proximity of sinonasal malignancies to vital structure and its locally aggressive behaviour both of which make gross total resection difficult. These tumors are responsive to chemotherapy in spite of the variable histologies seen at this site (esthesioneuroblastoma, sinonasal tumor of neuroendocrine differentiation, NUT midline tumors, and sinonasal tumor undifferentiated cancers).
Neoadjuvant chemotherapy before surgery may lead to regression of tumor and improvement in gross total resection rate in locally advanced tumors which in turn may improve the local control [
Sinonasal tumors are routinely discussed in a multidisciplinary clinic. Patients with the below mentioned criteria are referred for neoadjuvant chemotherapy before local treatment: Locally advanced sinonasal tumors with extension of tumor beyond nasal and paranasal sinus: Resectable: but resection would been morbid requiring extensive surgery and would have chances of incomplete gross total resection. Unresectable: frank involvement of any vital structure or surgically inaccessible site making upfront surgery not possible. ECOG PS 0–2. Without distant metastasis.
These patients were treated with neoadjuvant chemotherapy. NACT consisted of cisplatin and etoposide. Cisplatin dose of 33 mg/m2 D1 to D3 and etoposide dose of 100 mg/m2 D1 to D3 were administered intravenously. Cisplatin was replaced with carboplatin (AUC-5 or 6) if the calculated serum creatinine clearance was below 60 mL/min. The chemotherapy was administered with standard premedications and antiemetic prophylaxis. Patients were given 1 liter of 0.9% NaCl hydration with magnesium and potassium supplementation from D1 to D3. Secondary prophylaxis with G-CSF was administered for patients having febrile neutropenia in C1. Two cycles of NACT were administered.
Following 2 cycles of NACT, patients were assessed with axial radiological imaging (either CECT or PET-CT). These patients were then discussed in skull base multidisciplinary clinic. Patients who had adequate response, which would facilitate gross total resection were offered surgical resection and adjuvant chemoradiation. Patients in whom, after 2 cycles, gross total resection was still not possible were offered radical chemoradiation. Patients who had progressed after NACT were considered for radical chemoradiation or palliative radiotherapy (RT) depending upon the patient’s performance status and tumor volume. Palliative RT was delivered when tumor volumes were large and adequate tumoricidal RT doses could not be delivered without respecting the tolerance doses of nearby vital structures. These patients were followed up after treatment till death.
For this analysis, the data of these patients was acquired from a prospectively maintained head and neck cancer NACT database. Patients treated between August 2010 and August 2014 with sinonasal tumors and nonsquamous histology were selected. Data regarding baseline clinical details, staging, the indication of NACT, NACT details, response, adverse events, post-NACT local treatment details, pathological response, and outcome details were noted.
For this analysis, as NACT was given predominantly with the intention of having a gross total resection, the locoregional extent of tumor was charted. The charting was done with the following spaces being considered: involvement of cribriform plate, involvement of intracranial space with only extradural extension, involvement of intracranial space with intradural without involvement of brain, involvement of intracranial space with involvement of brain, involvement of orbit, and involvement of infratemporal fossa. This charting was done so that each of these space involvement would be analyzed as a factor predicting for achievement of resectability after NACT.
The response to NACT was noted in accordance with RECIST version 1.1. The adverse events during NACT were documented in accordance with CTCAE version 4.03. The pathological response rate was quantified as pathologic complete response (pCR) if no viable tumor was seen post-NACT and No-pCR if any viable tumor was seen post-NACT. The outcome data noted was progression free survival and overall survival. The progression free survival was calculated from date of start of NACT to date of progression (either locoregional or distant). Those patients who had not progressed were censored at their last follow-up. The overall survival was calculated from date of start of NACT to date of death. Those patients who had not died were censored at their last follow-up.
Data was censored for analysis on September 30, 2015. Descriptive statistics was performed. Fisher’s test was used to test whether there was a difference between response rate in the different histological subtypes considered. The analysis between achievement of resectability and different sites of involvement was also done by Fisher’s test. The progression free survival and overall survival for each histology were computed by Kaplan Meier survival analysis. Log rank test was used for univariate analysis of PFS and OS. Cox regression analysis was used for multivariate analysis.
Twenty-five patients of sinonasal cavity cancer were identified. The baseline details are shown in Table
Baseline details according to histological subtypes.
Variable | Esthesioneuroblastoma |
Sinonasal tumor with neuroendocrine differentiation ( |
Total ( |
---|---|---|---|
Median age | 40 years (IQR 36.5–42.75 years) | 45 years (IQR 36.5–57.0 years) | 42 years (IQR 37–47 years) |
Gender | |||
Male | 23 | 11 | 19 |
Female | 04 | 02 | 06 |
ECOG PS | |||
PS 0-1 | 12 | 13 | 25 |
PS 2 | 00 | 00 | 00 |
Grade | |||
III-IV | 10 | 13 |
08 |
The extent of locoregional spread is shown in Table
Extent of locoregional spread.
Extent | Esthesioneuroblastoma |
SN-NEC |
Total |
---|---|---|---|
Involvement of cribriform plate | 12 (100.0%) | 13 (100.0%) | 25 (100.0%) |
Intracranial extension up to extradural region | 06 (50.0%) | 08 (61.5%) | 14 (56%) |
Intradural intracranial extension but brain parenchyma uninvolved | 06 (50.0%) | 04 (30.8%) | 10 (40%) |
Intradural extension with brain parenchyma involvement | 02 (16.7%) | 03 (23.1%) | 05 (20%) |
Involvement of orbit | 06 (50.0%) | 07 (53.8%) | 13 (52%) |
Involvement of infratemporal fossa | 03 (25%) | 03 (23.1%) | 06 (24%) |
Involvement of parapharyngeal space | 01 (08.3%) | 01 (7.7%) | 02 (08%) |
Involvement of regional lymph nodes | 03 (25%) | 05 (38.5%) | 08 (32%) |
The reason for NACT was dural involvement in 2 patients, brain parenchyma involvement in 4 patients, intracranial involvement in 5 patients, intracranial extension with orbital apex involvement in 2 patients, orbital involvement (extensive) in 1 patient, and infratemporal fossa involvement in 1 patient and extensive soft tissue disease in 10 patients.
Out of 25 patients 2 cycles of NACT were completed by all 25 patients. The median number of cycles delivered was 2 (IQR 2-3). Twelve patients (48%) received more than 2 cycles before locoregional treatment. The incidence of grade 3-4 toxicity in accordance with CTCAE version 4.03 was 76%. There was no grade 5 toxicity seen. The details of adverse events are shown in Table
Adverse events in accordance with CTCAE version 4.03 observed during NACT. Numbers shown are actual patient numbers.
Toxicity | Grade 3 | Grade 4 |
---|---|---|
Anemia | 02 | 00 |
Neutropenia | 06 | 04 |
Thrombocytopenia | 00 | 00 |
Febrile neutropenia | 01 | 02 |
Nausea | 00 | 00 |
Vomiting | 00 | 00 |
Diarrhea | 01 | 00 |
Increased serum creatinine | 00 | 00 |
Transaminitis (raised SGOT/PT) | 00 | 00 |
Hyponatremia | 07 | 05 |
Hypokalemia | 00 | 00 |
Hyperkalemia | 00 | 00 |
The response was evaluable in all 25 patients after 2 cycles of NACT. The response was PR in 20 patients (80%, 95% CI 58.7%–92.4%), SD in 03 patients, and PD in 02 patients. The difference in response according to histological subtype is shown in Table
Response rates according to histological type.
Esthesioneuroblastoma |
SN-NEC |
Total | |
---|---|---|---|
CR + PR | 8 (66.7%) | 12 (92.3%) | 20 (80%) |
SD + PD | 4 (33.3%) | 01 (07.7%) | 05 (20%) |
Post-NACT 13 patients
Factors affecting achievement of resectability.
Presence of factor | Resectability achieved |
|
|
|
---|---|---|---|---|
Anatomical factors | ||||
Intracranial extension up to extradural region | Yes: 14 | 8 | 0.430 | Not included |
No: 11 | 05 | |||
Intradural intracranial extension but brain parenchyma uninvolved | Yes: 10 | 06 | 0.404 | Not included |
No: 15 | 07 | |||
Intradural extension with brain parenchyma involvement | Yes: 05 | 03 | 0.541 | Not included |
No: 20 | 10 | |||
Involvement of orbit | Yes: 13 | 07 | 0.582 | Not included |
No: 12 | 06 | |||
Involvement of infratemporal fossa | Yes: 06 | 02 | 0.281 | 0.851 |
No: 19 | 11 | |||
Involvement of parapharyngeal space | Yes: 02 | 01 | 0.741 | Not included |
No: 23 | 12 | |||
Surgical pre-NACT status | Unresectable: 11 | 01 | 0.000 | 0.003 |
Resectable: 14 | 12 | |||
|
||||
Biological factors | ||||
Pathology | E: 12 | 08 | 0.157 | Not included |
SNE: 13 | 05 | |||
Response | CR + PR: 20 | 12 | 0.136 | 0.139 |
SD + PD: 05 | 01 |
Post-NACT 13 patients were resectable but 1 patient opted for CTRT. Post-NACT treatment received was surgery followed by adjuvant treatment in 11 patients, surgery without adjuvant in 1 patient, radical chemoradiation in 9 patients, and palliative RT in 2 patients. Two patients did not take local treatment after NACT. One patient had progressive disease after NACT and was not suitable for any local treatment. The other patient had near complete response and he did not want any further treatment.
Surgery was performed in 12 patients. The type of surgery done was craniofacial resection in 06 patients, craniofacial resection with medial maxillectomy in 01 patient, medial maxillectomy in 03 patients, sinonasal resection in 01 patient, and radical maxillectomy with orbital exenteration in 01 patient. It was a gross complete resection in all 12 patients. The pathological response was pathological complete response in 03 patients. All 12 patients were offered adjuvant treatment consisting of chemoradiation. However one patient declined adjuvant treatment as there was risk of vision loss associated with RT. 11 patients received adjuvant chemoradiation. Out of these 11 patients, 07 patients were treated with IMRT and 04 patients were treated with 3DCRT technique. The median dose to tumor bed (CTV) was 6000 cGy (IQR 6000-6000 cGy). All patients completed chemoradiation. The median number of weekly chemotherapy cycles (cisplatin 30 mg/m2) received were 6 (IQR 6-6).
Radical chemoradiation was done in 09 patients. Out of these, 07 patients were treated with IMRT and 02 patients were treated with 3DCRT technique. The median dose to tumor bed (CTV) was 6000 cGy (IQR 6000–6600 cGy). All patients except one completed radical chemoradiation. This patient progressed after 10# of RT and hence his RT was stopped. The median number of chemotherapy (cisplatin 30 mg/m2) cycles received were 5 (range 4–6).
Palliative RT was delivered by conventional method with the midline dose being 5500 cGy delivered in 22# in one patient and 5000 cGy in 25# in the other patient. Both patients had symptomatic relief postpalliative RT.
The median follow-up was 1.7 years (IQR 1.0–2.2 years). There were 6 patients who had progression. The first site of progression was local in 2 patients, local with distant in 1 patient, regional in 1 patient, and distant in 2 patients. The sites of distant failures were bony metastasis in 2 patients and regional lymph nodes in 1 patient. Five patients had died at the time of analysis and all deaths were due to disease progression.
The 2-year progression free survival and overall survival were 75% and 78.5%, respectively. The influence of different factors on PFS and OS can be seen in Table
The influence of different factors on PFS and OS.
Factor | Division | Median PFS in years | 2-year PFS |
|
---|---|---|---|---|
Histological type ( |
Esthesioneuroblastoma | NR | 91.7% | 0.094 |
SNEC | NR | 57.0% | ||
Stage ( |
Kadish C | NR | 60.0% | 0.347 |
Kadish D | NR | 81.4% | ||
Response ( |
CR + PR | NR | 77.3% | 0.266 |
SD + PD | NR | 60.0% | ||
Resectability achieved ( |
Yes | NR | 92.3% | 0.015 |
No | 1.10 | 50.0% | ||
Upfront status ( |
Resectable | NR | 100% | 0.002 |
Unresectable | 1.10 | 45.5% | ||
|
||||
Factor | Division | Median OS in years | 2-year OS |
|
|
||||
Histological type ( |
Esthesioneuroblastoma | NR | 91.7% | 0.169 |
SNEC | NR | 64.5% | ||
Stage ( |
Kadish C | NR | 79.6% | 0.677 |
Kadish D | NR | 75.0% | ||
Response ( |
CR + PR | NR | 82.7% | 0.185 |
SD + PD | NR | 60.0% | ||
Resectability achieved ( |
Yes | NR | 100.0% | 0.016 |
No | NR | 58.3% | ||
Upfront status ( |
Resectable | NR | 100% | 0.008 |
Unresectable | NR | 54.5% |
The 2-year OS (Figure
Two-year OS in accordance with achievement of resectability.
Sinonasal tumors have varied histology [
All of these patients had Kadish C or D stage disease. The extent of the local tumor is highlighted in Table
The neoadjuvant chemotherapy used here was cisplatin and etoposide. This is a standard regimen for treatment of high-grade neuroendocrine carcinomas. This regimen either alone or in combination with radiation in accordance with stage is used. Different combinations of chemotherapy have been used in the literature with platinum backbone. Many of these combinations have infusional 5 FU [
Esthesioneuroblastoma was associated with high response rates and better outcomes than SNEC. This finding is similar to that reported in other studies [
Interestingly the factors which impacted both PFS and OS were the resectability status of the patients prior to both surgery and after surgery. All of these patients had extensive disease but these patients were classified in multidisciplinary clinic upfront into potentially resectable or unresectable. Unresectable patients were those whose tumors had frank invasion of vital structures making them not a candidate for surgery. However the involvement in these tumors in most occasions either were encasing vital structures (like optic nerve, orbital apex) or had frank invasion of brain making these tumors an unlikely candidate for surgery even if they showed excellent response to NACT. As opposed to these patients termed resectable implied that gross total resection may not have been possible and if attempted would have required an extensive mutilating surgery. This classification had an impact on PFS and OS very similar to that of post-NACT resectability status. Our data validates this upfront classification as only 1 out of 11 inoperable patients was considered resectable after NACT as opposed to 12 out of 14 resectable patients. This data emphasizes proper selection criteria for selecting patients for NACT when it is administered to make the tumor resectable. However we failed to provide objective assessment criteria for such classification upfront. Upfront invasion of none of the anatomical landmarks considered in the study could predict achievement of resectability after NACT.
This study though retrospective has its own strengths and limitations. This is one of the few studies reporting on nonsquamous locally advanced sinonasal tumors who have been selected with homogenous criteria (Kadish C and D) and have been treated with homogenous protocols. The treatment planning of these cases had been done in a multidisciplinary clinic and post-NACT treatment was also decided in the same clinic. The limitation of this study is its retrospective nature and short follow-up. It is known in sinonasal tumors, especially esthesioneuroblastoma, to have delayed recurrence, in some instances even 10 years after initial treatment [
Sinonasal tumors are a group chemosensitive tumors. NACT with cisplatin and etoposide can achieve response rate of 80% in nonsquamous sinonasal tumors and is well tolerated. The protocol of NACT followed by local treatment is associated with improvement in outcomes.
The authors declare that there is no conflict of interests regarding the publication of this paper.