Therapeutic Approaches and Outcomes in Patients with Larynx or Hypopharynx High-grade Neuroendocrine Carcinoma: A Single-Center Retrospective Analysis

Abstract

Background:

High-grade neuroendocrine carcinoma of the larynx (HG-NECL) is rare and aggressive with limited data regarding response to systemic therapy. We evaluated clinicopathological features, therapeutic approaches, and outcomes in patients with laryngeal or hypopharyngeal HG-NECL.

Methods:

Data were retrospectively collected through 1997 to 2020. Median disease-free (mDFS), progression-free (mPFS), and overall survival (mOS) were estimated using the Kaplan-Meier method.

Results:

Fifteen patients were identified; Most had locoregional (N=7) or metastatic disease (N=5). The main curative-intent treatment was chemoradiation concurrent with platinum-based chemotherapy; the rate of complete response was 78%. Most patients (80%) developed recurrence; the mDFS was 13.1 months. For the first-line palliative therapy, the ORR and mPFS were 50% and 3.1 months, respectively. For all patients, the mOS was 17.8 months, and 8.6 months for metastatic disease.

Conclusion:

Laryngeal HG-NEC is associated with high relapse rates and dismal prognosis for those with recurrent/metastatic disease. Novel therapeutic strategies are needed.

INTRODUCTION

Laryngeal cancer affects predominantly men and is highly associated with smoking and alcohol consumption. Squamous cell carcinoma (SCC) is the most common histologic type of laryngeal cancer, but other types such as neuroendocrine carcinoma (NEC) can arise from the larynx.

NEC is a heterogeneous histologic group with a wide spectrum of differentiation and aggressiveness. In the 2017 World Health Organization (WHO) Classification of Head and Neck Tumors, NEC is divided into three categories that differ in terms of grade, clinical behavior, and prognosis: well-differentiated (low-grade, < 2 mitoses per 10 high-power fields); moderately differentiated (intermediate-grade, 2–10 mitoses per 10 high-power fields); and poorly differentiated (high-grade, > 10 mitoses per 10 high-power fields). Poorly differentiated NEC is further classified as small cell NEC (SmCC) and large cell NEC (LCNEC) (Supplementary Figure 1)¹. LCNEC had previously been considered moderately differentiated, but it was transferred into the poorly differentiated group in the 2017 WHO Classification of Head and Neck Tumors because LCNEC’s morphologic and immunohistochemical features and clinical behavior resemble those of the other tumors in the poorly differentiated, high-grade NEC (HG-NEC) category. Indeed, LCNEC has been considered HG-NEC in the WHO Classification of Lung Tumors since 2001 ², and in 2014, a meta-analysis of NEC of the larynx showed that LCNEC and SmCC had similar prognosis, supporting that the two entities have similar clinical behavior³.

HG-NEC is the second most common neuroendocrine tumor of the larynx, after intermediate-grade NEC, but accounts for only 0.5% of all laryngeal cancers⁴. Consistent with HG-NEC arising from other primary tumor sites, such as the lung, HG-NEC of the larynx is frequently diagnosed at an advanced stage and exhibits aggressive clinical behavior with early hematogenous spread and more than half of patient dying from the disease^{3, 5–9}. The most common primary tumor site of HG-NEC is the supraglottis, and associations with paraneoplastic syndromes, such as Lambert-Eaton syndrome and inappropriate secretion of ectopic adrenocorticotropic hormone, have been reported¹⁰.

Locoregionally advanced HG-NEC is frequently treated with multimodality therapy including surgery, radiation therapy (RT), and chemotherapy; for metastatic HG-NEC, a platinum-based chemotherapy regimen is usually preferred. However, those approaches are based on small retrospective series ^{3, 5–9, 11, 12}. The aim of this study was to report clinical and pathologic features of cases of HG-NEC of the larynx, including SmCC and LCNEC, that were seen at our institution and to evaluate the therapeutic approaches, clinical course, response to treatment, and survival outcomes in patients with this rare disease.

MATERIALS AND METHODS

Study population

Patients with HG-NEC of the larynx were identified within available pathology reports from 1986 to 2020. First, we retrieved reports that met the following criteria: (1) patient with previous billing to the Head and Neck Department and (2) pathology report including “small cell” OR “neuroendocrine” OR “large cell” inside the main text. The initial search returned 10,448 reports from 6002 patients. Second, we excluded any report that contained any of a pre-specified list of words unrelated to HG-NEC (available in Supplementary Figure 2) (N = 7975). The remaining 2473 reports were then manually reviewed, and reports of patients who did not have laryngeal HG-NEC or did not have complete clinical and pathological data were excluded (N = 2458). The remaining 15 reports, corresponding to 15 patients diagnosed with laryngeal HG-NEC between 2001 and 2019 with complete clinical and pathologic data, were included for analysis. The institutional review board of The University of Texas MD Anderson Cancer Center approved this study (PA 17–0865).

Data collection

Clinical data were abstracted from electronic health records using standardized forms about pathology, demographics, treatment, and outcomes. The following variables were collected for each patient: age at diagnosis, sex, smoking history, smoking pack-years, alcohol use history, date of diagnosis, primary tumor site, TNM stage per the seventh edition of the UICC-AJCC¹³, neuroendocrine and epithelial immunohistochemistry markers (synaptophysin, chromogranin, CD56, TTF1, and p63), definitive treatment (surgery, radiotherapy, and/or chemotherapy), response to treatment, palliative systemic therapy, date and site of first recurrence, last follow-up date, and date of death. All pathology reports and slides from the 15 selected patients were reviewed by an experienced head and neck pathologist (DB).

Statistical analysis

Mean, standard deviation, median, and interquartile range were used for descriptive statistics according to variable distribution. Response was assessed per radiologic report assessment. The Kaplan-Meier method was used for time-to-event analysis of events including disease-free survival (DFS), progression-free survival (PFS), and overall survival (OS). DFS was defined as the length of time from the date of treatment completion to the date of first recurrence among patients who achieved disease-free status. PFS was defined as the length of time from the date of treatment initiation to the date of disease progression or death from any cause. OS was defined as the length of time from the date of diagnosis to the date of death from any cause, with censoring at the last date the subject was known to be alive. The log-rank test was performed to assess differences in time-to-event outcomes among groups. Comparisons between specific groups were planned before data collection on the basis of literature and clinical knowledge. The null hypothesis was rejected when alpha ≤ 0.05. All statistical analyses were performed using R version 3.6.3.

RESULTS

Patient and tumor characteristics

Fifteen patients with HG-NEC of the larynx (N=13), and hypopharynx (N=2) were included in this analysis. The median follow-up was 17.8 months. Key baseline patient and tumor characteristics are summarized in Table 1. The median age at diagnosis was 58 years (range, 23–73); eight patients (53%) were male, and nine (60%) had the supraglottis as the primary disease site. Most patients (80%; 12 of 15) had a history of smoking, of whom 11 were heavy smokers (> 10 pack-years); three patients (20%) declared a history of alcohol misuse according to National Institute on Alcohol Abuse and Alcoholism (>7 drinks per week for woman; >14 drinks per week for men) ¹⁴. The median time between symptom onset and diagnosis was 6 months (range, 0.5–23). Eleven patients (73%) were diagnosed at late stages of disease, of whom seven had locally advanced disease (stage III-IVB) and five had metastatic disease (stage IVC).

Table 1.

Patient and tumor characteristics (N = 15).

Characteristic, No of patients (%)	Data value
Age at diagnosis, median (range), y	58 (23–73)
Sex
Male	8 (53)
Female	7 (47)
Smoking status
Never smoker	3 (20)
≤10 pack-years	1 (7)
>10 pack-years	11 (73)
Alcohol consumption
Never	4 (27)
Social	7 (47)
Misusea	3 (20)
Primary tumor site
Supraglottis	12 (80)
Hypopharynx	2 (13)
Undefined	1 (7)
Stage at diagnosis b
I-II (localized)	3 (20)
III-IVB (locoregional)	7 (47)
IVC (metastatic)	5 (33)
Site of metastasis c
Distant lymph nodes	9/13 (69)
Bones	6/13 (46)
Liver	5/13 (38)
Lung	5/13 (38)
Pathologic features
Small cell neuroendocrine carcinoma	8 (53)
Large cell neuroendocrine carcinoma	3 (20)
Combined squamous cell carcinoma and small cell neuroendocrine carcinoma	2 (13)
Undefined	2 (13)

^aDefined as >7 standard alcoholic drink per week for women and >2 for men, according to National Institute on Alcohol Abuse and Alcoholism

^bAccording to the AJCC 7th-edition TNM classification

^cFirst metastasis site. Included metastasis at diagnosis and metastasis that developed during follow-up.

Eight patients had SmCC, three had LCNEC, two had combined SCC and SmCC, and two had HG-NEC without further specification. Representative histologic sections of a SmCC of the supraglottis stained for neuroendocrine markers are shown in Figure 1.

Figure 1.

Left arytenoid high-grade neuroendocrine carcinoma, small cell type.

A-C. Hematoxylin-eosin-stained sections at 2x (A), 4x (B), and 10x (C) magnification with salient cytomorphologic features, including nuclear molding, high nuclear/cytoplasm ratio, mitotic activity, apoptosis, and rosetting. D-E. Ancillary studies were supportive of the neuroendocrine phenotype and included immunoperoxidase studies with anti-synaptophysin antibodies (D) and anti-cytokeratin cocktail (E).

Treatment of localized and locoregional disease

Ten patients (67%) underwent curative-intent treatment including surgery, definitive RT, and/or chemotherapy (Table 2). Only one patient, who presented with stage II disease, underwent endoscopy vocal cord tumor excision alone as primary therapy; this patient experienced local recurrence and subsequently underwent platinum and etoposide chemotherapy, supraglottic partial resection and ipsilateral modified radical neck dissection followed by adjuvant RT. Two patients (20%), with stage IVA disease, underwent induction chemotherapy; one achieved a partial response, and one had stable disease.

Table 2.

Therapeutic approaches and outcomes in patients treated with curative intent (N = 10).

Treatment or outcome, No of patients (%)	Data value
Induction chemotherapy	2 (20)
Radiation therapy	10 (100)
Definitive RT	9 (90)
Adjuvant RT	1 (10)
Definitive RT, median dose Gy (range)	72.0 (66.0–76.4)
Concomitant chemotherapy	7 (70)
Cisplatin and etoposide	4 (57)
Carboplatin and paclitaxel	1 (14)
Cisplatin only	2 (28)
Surgery, No (%)	3 (30)
Up-front surgery	1 (33)
Salvage surgery	2 (67)
Best response to definitive (chemo)radiation
CR	7 (78)
PD	1 (11)
Unknown	1 (11)
Recurrence	8 (80)
Median DFS (95% CI), mo	13.1 (8.53-NA)

CR, complete response; DFS, disease-free survival; PD, progressive disease.

All patients received RT to the larynx and neck nodes, nine as a primary definitive treatment, and one as adjuvant therapy after a salvage surgery due to local recurrence (as mentioned above). The median definitive RT dose was 72.0Gy (range, 66.0–76.4). Seven patients received RT with concurrent chemotherapy: four patients (57%) received cisplatin and etoposide, two (28%) received single-agent cisplatin, and one (14%) received carboplatin and paclitaxel.

The best response to definitive RT with or without concurrent chemotherapy was complete response in seven patients (78%), progression of disease (local and distant) in one (11%), and unknown in one (11%).

During follow-up, only two (20%) of the 10 patients treated with curative intent remained disease free, one with a stage III SmCC and one with a stage IVA combined tumor (HG-NEC and SCC), both treated with definitive chemoradiation. The other eight patients (80%) developed recurrence after either definitive (chemo)radiation, upfront surgery and/or salvage surgery. Three patients (30%) had only local recurrence, two (20%) had both local and distant recurrence, and three (30%) had only distant recurrence. The median time to progression was 11.5 months (95% CI, 8.0-NA), and the median DFS was 13.1 months (95% CI, 8.53-NA). The 1- and 2-year DFS rates were 50% and 12.5%, respectively.

Treatment of metastatic disease

Thirteen patients presented with distant metastasis (N = 5) or developed distant metastasis during follow-up (N = 8). The most prevalent metastasis sites were distant lymph nodes (9 patients; 69%), bone (6 patients; 46%), liver (5 patients; 38%), and lung (5 patients; 38%). Among the 13 patients with metastatic disease, eight received first-line systemic therapy (Table 3), three died within a month after recurrence without exposure to palliative chemotherapy, and two were lost to follow-up. The most commonly used first-line chemotherapy was the combination of platinum and etoposide, administered to four of eight patients (50%). Two patients were treated with platinum-based chemotherapy plus a PD-1 inhibitor as first-line systemic therapy; both experienced partial responses. The overall rate of response to first-line systemic therapy was 50%, and the median PFS was 3.1 months (95% CI, 2.6-NA).

Table 3.

Therapeutic approaches and outcomes in patients treated with palliative intent (N = 8).

Treatment or outcome, No of patients (%)	Data value
First-line therapy	8 (8/13, 62)
Platinum and etoposide	4/8 (50)
Other platinum-based doublet	2/8 (25)
Chemotherapy and immune checkpoint inhibitor	2/8 (25)
Second-line therapy	1 (8)
Exposure rate
Platinum agent	8/8 (100)
Etoposide	4/8 (50)
Taxane	2/8 (25)
Immune checkpoint inhibitor	3/8 (37.5)
Best overall response
First-line therapy
CR or PR	4/8 (50)
SD	1/8 (12.5)
PD	3/8 (37.5)
Second-line therapy
PD	1 (100)

CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease

Only one patient received second-line palliative systemic therapy with anti-PD-1 combined with anti-CTLA-4 (nivolumab and ipilimumab). This patient experienced disease progression after two treatment cycles.

Survival outcomes

The disease course and therapeutic interventions for each of the 15 patients in the analysis from diagnosis to death are outlined in Figure 2. The 5-year OS rate was 13.3%. The median OS for the overall population was 17.8 months (95% CI, 12.6-NA) (Figure 3A). As expected, tumor stage as a prognostic factor: patients with stage IVC disease had shorter median (95% CI) OS than patients with stage II-IVB disease (8.6 [5.1-NA] vs. 22 [17.0-NA] months, p = 0.02) (Figure 3B). There were no significant differences in median OS by pathologic subtype (SmCC 14.5 months vs non-SmCC 18.5, p = 0.89) or smoking history (≤ 10 pack-years, 22 months vs > 10 pack-years, 17.8 months, p = 0.89).

Figure 2. Swimmer plot.

Time 0 represents the date of diagnosis.

An arrow indicates that the patient was alive at end of the graph period (60 months) and had the indicated outcome outside this range.

Figure 3. Overall survival.

Shown are Kaplan-Meier curves for the patients overall (A) and according to tumor stage (B). Shaded are indicate a 95% confidence interval. OS, overall survival.

DISCUSSION

HG-NEC of the larynx is an extremely rare condition with limited data available regarding clinical behavior and response to therapies. Consistent with this situation, there is no standard optimal care of patients with loco-regional or metastatic disease. Here we report the clinical and pathologic characteristics and outcomes of 15 patients with laryngeal and hypopharynx HG-NEC and represents the largest single-center cohort of patients with this rare condition. Study limitations include the retrospective and single-center design and the small number of patients treated over 18 years.

The clinical behavior of laryngeal HG-NEC was compatible with an aggressive malignancy: most patients were diagnosed at late disease stages, and 80% of patients who received curative-intent treatment developed early distant metastasis. The 5-year OS rate was 13%. These findings are consistent with previous studies that showed a 5-year OS rate of 14%³ and recurrence rates ranging from 58% to 81% among patients with laryngeal HG-NEC^{3, 5}. Notably, in our study, three patients had LCNEC and two patients had combined SCC and SmCC, and their clinical course and prognosis were as bad as what was observed in patients with SmCC.

It is noteworthy that in locally advanced head and neck SCC, the most common histologic type of laryngeal cancer, most recurrences are loco-regional, and the distant relapse rate is approximately 20%^15,16, much lower than what we found. The clinical behavior of laryngeal HG-NEC is distinct from that of SCC and very similar to that of small cell lung cancer (SCLC), in which micrometastasis is usually present at diagnosis and recurrences are mostly at distant sites. Similar to our findings, patients with limited-stage SCLC have a relapse rate of approximately 85%, and recurrence usually occurs at distant sites and within 14 to 15 months^17,18.

The preferred therapeutic approaches for HG-NEC of the larynx are based on the SCLC literature because of the similarities between these two entities in terms of pathologic features and clinical course. The rarity of HG-NEC precludes studies with enough power to determine the effect of adding concurrent chemotherapy to RT. Based on disease biology and high rate of local recurrence (50%), we recommend an aggressive multimodality treatment approach for patients with laryngeal HG-NEC with loco-regional disease, involving RT, systemic therapy, and surgery for achieve locoregional control. Notably, while most patients in our series received chemotherapy in the curative-intent setting concurrently with RT, similar to what is done for laryngeal SCC, neoadjuvant or adjuvant systemic therapy in addition to chemoradiation should be considered for HG-NEC given the high rate of distance recurrences. For patients with tumors with extra laryngeal extension and/or cartilage invasion (T4), or those with laryngeal dysfunction before treatment, primary total laryngectomy is advised given it has shown to prevent complications related to chronic aspiration and is associated with a survival benefit for patients with SCC ^19–22. Considering our findings in addition to previous studies that showed poor outcomes related to surgery as single modality³, our recommendation is triple modality therapy for patients with T4 lesions and/or laryngeal disfunction, and good performance status.

All patients in our series who were treated with palliative intent received platinum-based chemotherapy, most commonly platinum and etoposide. Although the overall rate of response to first-line therapy was high (50%), the duration of response was short, with a median PFS of only 3.1 months. Recently, immune checkpoint inhibitors have shown activity in the treatment of extensive-stage SCLC when combined with etoposide and platinum and have supplanted chemotherapy alone as the standard first-line treatment^{23, 24}. The activity of immune checkpoint inhibitors in SCLC is partially explained by the high immunogenicity of SCLC cells and by the high tumor mutational burden²⁵, both of which are linked to long-term tobacco exposure²⁶, which is commonly seen in patients with HG-NEC of the larynx. Even though fragile, this rationale, in addition to the pathologic and clinical similarities between the two diseases, could support the use of immune checkpoint inhibitors for the treatment of laryngeal HG-NEC, as a prospective study for this rare entity is not feasible. In our study, only two patients were treated with chemotherapy and an immune checkpoint inhibitor, and both experienced partial responses but died 5 and 8 months, respectively, after treatment started. We propose a treatment algorithm for HG-NEC of the larynx based on extrapolation data from SCLC given the similarities between both entities (Supplementary Figure 3).

CONCLUSION

Our study confirms that HG-NEC of the larynx is an aggressive malignancy with high rates of relapse and death. LCNEC and combined SCC and SmCC have the same aggressive behavior as SmCC and should be managed similarly. Although the majority of patients benefit from RT and systemic therapy, the response duration is short, and survival remains poor. Given disease biology similar to SCLC, a more common entity with evidence-based treatment approaches, new drugs with benefit in SCLC can be considered for patients with HG-NEC of the larynx.

List of Abbreviations:

DFS

Disease-free survival

HG-NEC

High-grade neuroendocrine carcinoma

LCNEC

Large cell neuroendocrine carcinoma

NEC

Neuroendocrine carcinoma

OS

Overall survival

PFS

Progression-free survival

RT

Radiation therapy

SCC

Squamous cell carcinoma

SCLC

Small cell lung cancer

SmCC

Small cell carcinoma

WHO

World Health Organization