Clinicopathological characteristics, treatment and prognosis of head & neck small cell carcinoma: a SEER population-based study

Small cell carcinomas (SmCCs), with small round or oval cell shapes, are regarded to be identical to poorly differentiated neuroendocrine tumors, the most aggressive type of neuroendocrine cancer [1]. This type of cancer was first described as oat cell sarcoma by Barnard in 1926 [2]. Epidemiological studies have suggested that SmCCs occur most commonly in the lung [1]. Extrapulmonary SmCCs are rare, comprising only 2.5 to 5% of all SmCCs [3]. Owing to the tendency for regional or distant spread, it is crucial to rule out other sites of the primary tumor before making a prognosis of the disease [4]. Computed tomography and/or magnetic resonance imaging of selected sites and positron emission tomography scanning are recommended as appropriate examinations to assess the original localization of the tumor [1]. Patients with SmCCs lack early specific symptoms. Therefore, the cancer has already evolved to an advanced stage when diagnosed in most cases [5].

Head and neck SmCCs (H&NSmCC) is a rare malignancy that is more likely to present at an advanced stage. This type of cancer carries a worse prognosis than squamous cell carcinoma of the head and neck, the most frequent tumor in the head and neck [6]. H&NSmCC is considered to occur in multiple sites associated with the upper aerodigestive tract, of which the larynx is the most frequent site for SmCCs as a primary tumor in the head and neck [1]. Tumors arising in the salivary glands have a relatively better prognosis than SmCCs in the larynx, lung, and most other sites [1]. Limited cases make it difficult for investigators to explore standard treatments that can be strictly followed for the disease [7]. Accepted therapies include surgery, radiotherapy and chemotherapy [1]. Currently, the combination of radiation therapy and chemotherapy is preferred for the treatment of patients with H&NSmCC [8, 9]. Surgery is mainly reserved for patients who truly have early local disease or cases where locoregional lesions cannot be controlled [8, 10].

Due to the rarity of the cancer, previous studies were usually based on case reports or small case series. There is a lack of studies stemming from a large population. Thus, the clinicopathological characteristics of H&NSmCC cannot be summarized comprehensively. In the current study, we performed a retrospective analysis of patients with H&NSmCC based on the Surveillance, Epidemiology and End Results (SEER) database, one of the most representative publicly available databases in the US, with the following aims. First, we aim to describe the clinicopathological characteristics and survival of patients with H&NSmCC. Second, to estimate the indicators that influence the survival of the disease. Finally, to establish a disease-specific nomogram model to predict the prognosis of H&NSmCC patients based on the factors associated with prognosis.

Our study, based on the SEER database, contains the largest samples in the field of SmCCs in the head and neck region. We described the clinicopathological features of the tumor and assessed the factors that influenced the survival of the cancer. We further established an H&NSmCC-specific nomogram model to predict the survival of H&NSmCC patients based on independent prognostic factors. Because of the infrequent occurrence of the tumor, previous reports discussing H&NSmCC were mainly based on case reports or the experience of a single institution, which were relatively less comprehensive and less objective than our study. The final results of this study are of significance for management and can help doctors estimate the risk of patients with H&NSmCC.

Concerning the clinicopathological characteristics of H&NSmCC, Wakasaki et al. reported that the median age was 74 years (53–91 years) based on 21 cases [11]. Consistent with this result, in this cohort, the median age of all patients was 64 years (2–96 years), and the incidence peaked during the period of 50 to 70 years old. This result suggests that the patients in our study are relatively younger. The data concerning sex also corresponded to the data reported in previous literature. A study reported that most patients with the disease (73.0%) were male [12]. In this study, the male to female ratio was 1.85:1. Our results also showed that white people were more affected by H&NSmCC (86.69%), but there was no significant difference in survival among races. Almost all patients with H&NSmCC in the study had advanced tumor stage, poorly differentiated and undifferentiated tumors totally accounting for 98% of all tumors, which further supports the highly aggressive behaviors of the cancer.

SmCCs of the salivary gland and the nasal cavity had a relatively better prognosis than SmCCs in other regions of the head and neck, while SmCCs of the thyroid gland had the worst survival. Older age, aggressive stage of the tumor, large tumor sizes, lymph node invasion and distant metastasis were correlated with poor prognosis. Consistent with that for other tumors, younger patients had relatively better survival than older patients [13]. Age was an independent prognostic indicator in one study [12]. Tumor grade also influenced the survival of H&NSmCC. Patients with the cancer of poor differentiation or undifferentiation pathological grade had lower survival, however, the results showed no statistically significant difference. Patients with T1 and T2 cancer had longer survival times than the other patients in our study. Similar results have been described in previous reports [12, 14]. These findings demonstrated the importance of tumor size for prognosis, but T stage was not an independent predictor in the nomogram model of H&NSmCC in the current study.

Lymph node metastasis and distant metastasis were identified as independent prognostic indicators in a previous study [5]. Our data further confirmed this conclusion. However, concerning lymph node metastasis, some controversial results have been reported. Walters et al. found that lymph node metastasis was closely related to poor survival, but another report suggested that nodal metastasis had no influence on survival [15, 16]. Our data supported the adverse impact of lymph node metastasis on survival in H&NSmCC. A previous study observed that patients with SmCC of the salivary gland at a distant stage had worse survival than those with limited tumors [15]. Another study containing 344 cases of parotid SmCC showed that distant metastasis was a significant prognosticator in the multivariate model [12]. The locations of distant metastasis were mainly concentrated in the brain, liver, lung and bones [14]. In the present study, 38.28 and 22.94% of patients with H&NSmCC were classified into distant stage and M stage, respectively, and M stage, as another influencing factor, was closely related to prognosis. H&NSmCC is highly sensitive to chemotherapy, however, due to extremely aggressive behaviors of the cancer, tumor cells cannot be eradicated at M1 stage and H&NSmCC is likely to relapse after chemotherapy and radiotherapy, leading to a worse prognosis for patients with distant metastases [1, 7].

Radiotherapy and chemotherapy are considered the mainstays of treatment [1]. Surgery has only a very limited role in locoregional lesions [1]. In the limited stage, surgical resection is suggested and can improve survival [17]. Radiotherapy and chemotherapy, as adjuvant treatment, can largely extend patients’ life expectancy [8, 18]. Radiotherapy is effective in palliating intrathoracic symptoms, such as chest pain, shortness of breath and other systemic symptoms due to the metastasis of tumors. Platinum-based chemotherapy regimens, the most frequent therapy, lead to better survival than earlier chemotherapy [18]. Cisplatin/etoposide was also reported to have a high response for H&NSmCC [19]. Prophylactic cranial irradiation (PCI) is still controversial as a routine therapy for H&NSmCC [20, 21]. In the small lung carcinoma, the recommended dose of PCI is 25 Gy in 2.5-Gy fractions or 30 Gy in 1.8-Gy to 2-Gy fractions for H&NSmCC patients [22, 23].

Patients receiving chemotherapy, radiotherapy and surgery had the longest survival in our study, but the comparison of the three cohorts of treatments showed no significant differences. In addition, owing to the diversity of the clinicopathological features among patients, the results could not demonstrate that the combination of surgery, chemotherapy and radiotherapy was the most appropriate therapy for all H&NSmCC patients. Both radiotherapy and chemotherapy were vital for the survival of H&NSmCC patients in our study, but only chemotherapy was an independent prognostic factor in the predictive model, which highlighted the importance of chemotherapy for the tumor. Interestingly, patients receiving chemotherapy had a worse survival than others, which might be because patients with chemotherapy mostly had advanced stage tumors, and thus, the survival was not satisfactory. Due to the low incidence, standard treatment is still to be confirmed in the future.

In the current study, we constructed a specific prognostic model for H&NSmCC patients. The model can better predict the prognosis of H&NSmCC than the AJCC staging system as the information from the SEER database, such as age, pathological grade, AJCC stage and treatment, were all considered and therefore the result of the model is closer to the actual prognosis. Of course, there are some limitations of the current investigation. First, lack of detailed comorbidity data represents one of biggest limitation of SEER database. Second, the study has shortcomings due to its retrospective nature. Third, information on H&NSmCC patients was not complete, so we failed to analyze several factors that may influence the prognosis of the cancer. For instance, patients with information on surgery and T stage only accounted for 42.3 and 45.1%, respectively. Only one patient did not receive surgery which is not adequate for survival analysis. Owing to the limited number of patients with clear T stage, the factor was not included in the nomogram model. Fourth, the nomogram should be applied to patients with caution. Because it was built based on the retrospective investigation with lower level of evidence.

In summary, our study analyzed the clinicopathological features and treatment outcomes of H&NSmCC based on the SEER database, one of the largest cancer databases in the world. H&NSmCC, with very low occurrence, had a poor prognosis. The incidence peaked during the period of 50 to 70 years old, and the most frequent location in the cohort was the salivary gland. The 5-year OS and 5-year DSS were 26.2 and 27%, respectively, for H&NSmCC. N stage, M stage AJCC stage and chemotherapy were independent prognostic indicators. The nomogram model was constructed according to the above indicators. The model can better predict the survival of H&NSmCC patients than the AJCC staging system and facilitates doctors in the assessment of patient prognosis and helps patients choose reasonable treatments.