Spindle Cell Squamous Cell Carcinoma of Head and Neck Region: a Clinicopathological and Immunohistochemical Study

Abstract

Spindle cell squamous cell carcinoma (SCSCC) represents a rare, dedifferentiated form of conventional squamous cell carcinoma (SCC) with propensity for head and neck. Morphological heterogeneity of this tumor poses significant diagnostic challenge for pathologists. This study aimed to evaluate detailed morphological and immunohistochemical (IHC) characteristics of a series of SCSCC cases. All head and neck SCSCCs diagnosed over seven years were retrospectively analyzed to assess morphological features in detail with evaluation of possible differential diagnoses. Elaborate IHC panel was performed in each case. Eleven cases (8 males, 3 females) were included. Most patients were in 5th to 7th decade. Oral cavity (n = 6; 54.5%); oropharynx (n = 3; 27.3%), and larynx (n = 2; 18.9%) were involved. Tumors were polypoidal (n = 6), ulceroproliferative (n = 3), or ulceroinfiltrative (n = 2). Microscopically, epithelial ulceration was common (90.9%). Squamous differentiation was evident on light microscopy in 6 (54.5%) cases as surface dysplasia (n = 1; 9.1%), infiltrative tumor (n = 4; 36.4%), or both (n = 1; 9.1%). The spindled tumor component showed markedly heterogeneous morphology mimicking mesenchymal malignancies or benign mesenchymal proliferations. IHC for epithelial markers (CK, EMA, p40, and p63) established epithelial differentiation in 40% (2 of 5) cases where it was not evident morphologically. CK showed highest positivity followed by p40 and p63. Aberrant mesenchymal marker expression was variably seen. Morphological diagnosis of SCSCC is challenging with several histological mimics. Surface dysplasia and component of invasive SCC in a predominantly spindled tumor are important subtle morphological clues. Systematic morphological approach aided by IHC is helpful in clinching the accurate diagnosis.

Introduction

Spindle cell squamous cell carcinoma (SCSCC) is a unique, rare neoplasm proven to be a monoclonal dedifferentiated form of conventional squamous cell carcinoma (SCC) [1–3]. This tumor accounts for about 3% of all squamous carcinomas in the head and neck region. Larynx is the commonest site, though involvement of gingiva, tongue, hypopharynx, and nasal cavity has been described [2, 3]. Morphologically, SCSCC is a biphasic tumor composed of a squamous cell carcinoma (either dysplasia, in situ, and/or invasive cancer) and a malignant spindle cell component with a mesenchymal appearance, but of epithelial origin [1–3]. The spindled component usually forms bulk of the tumor, and around one-third of cases has an entirely monophasic spindled morphology with considerable morphological and immunohistochemical overlap with other benign and malignant spindle cell tumors [2–4]. Due to these reasons, SCSCC poses a significant diagnostic challenge to the pathologist. We evaluated the clinical, detailed morphological, and immunohistochemical (IHC) characteristics of a series of head and neck SCSCCs with an attempt to assess the problems in differentiating this tumor from other spindle cell tumors occurring at these sites.

Materials and Methods

We retrospectively reviewed all cases of SCSCC involving head and neck region diagnosed in the Department of Pathology in our institute over a period of 7 years. Slides were reviewed to assess morphological features and histologic differential diagnoses in detail. An IHC panel comprising CK, EMA, p40, p63, SMA, S100, CD34, and Ki67 was performed (if not previously done) on one representative paraffin block in each case using Biogenics X Matrix elite auto immunostainer following the manufacturer’s protocols. SOX10 and HMB45 were applied where indicated. Details regarding patients’ age and gender, clinical presentation, tumor site, and history of irradiation to head and neck were retrieved from medical records. Type (polypoidal/nonpolypoidal) and size of lesion were recorded. On microscopic examination, status of overlying squamous epithelium, presence of frank squamous differentiation, morphology and arrangement of the spindled tumor cells, degree of anaplasia, presence of giant cells, and mitotic activity were assessed. IHC stains were evaluated in the spindle component in all cases.

Observation and Results

Eleven cases (8 males and 3 females) of head and neck SCSCCs were reported during the study period. Male to female ratio was 2.7:1. Age at presentation ranged from 33 to 80 years (median age being 52 years) with maximum number of patients in the 5th to 7th decade (n = 8). Two patients had previous history of irradiation to head and neck region, for carcinoma tonsil in one case and for tongue cancer in the other (details of previous surgery were not available). Oral cavity was the commonest affected site (n = 6; 54.5% cases), followed by oropharynx (n = 3; 27.3%) and larynx (n = 2; 18.9%). Patients with oral cavity/oropharyngeal lesions presented with difficulty in mouth opening and/or swallowing, whereas laryngeal lesions presented with progressive hoarseness of voice. Most patients had a short history with a median duration of 2 months.

In 9 cases, biopsies from the tumor were received. Radical surgery in the form of partial glossectomy and segmental mandibulectomy with cervical lymph node dissection was performed in one case each. Grossly, the biopsy bits/resection specimens showed polypoidal (n = 6, 54.5%), ulceroproliferative (n = 3; 27.3%), or ulceroinfiltrative (n = 2; 18.2%) tumor.

Microscopic Examination

Broad spectrum of morphological features was noted. Epithelial ulceration of varying degree was common to most cases (n = 10; 90.9%), with 6 such cases (54.5%) showing near total epithelial ulceration. Six cases showed evidence of epithelial differentiation on light microscopy (54.5%) in the form of either surface epithelial dysplasia (n = 1; 9.1%) or squamous differentiation in the infiltrative tumor component (n = 4; 36.4%) or both (n = 1; 9.1%). Surface epithelial dysplasia was confined to lower 1/3 of epithelium in one case and extended through the full epithelial thickness in the other. The invasive squamous component was moderately (n = 1) to poorly (n = 4) differentiated and was either intimately admixed with the spindled component (n = 3; 60%) or abutting it (n = 2; 40%). The squamous component formed only a minor proportion of the lesion, and histological picture was dominated by the spindled component in all cases.

The spindled component revealed remarkable morphologic heterogeneity mimicking mesenchymal malignancies in some cases and reactive/benign mesenchymal proliferations in few others (Fig. 1). Towards the benign spectrum, extensive epithelial ulceration with underlying inflammatory infiltrate mimicking granulation tissue; prominent stromal hyalinization simulating reactive myofibroblastic proliferation; and florid subepithelial inflammation mimicking an inflammatory pseudotumor (1 case each) were seen [Fig. 1a-c]. However, careful search revealed scattered atypical spindle to polygonal cells and increased mitosis in the spindle cells to suggest their malignant nature. On other end of the spectrum, 3 cases showed markedly pleomorphic spindle cells arranged in storiform pattern with admixed tumor giant cells, simulating pleomorphic undifferentiated sarcoma (Fig. 1d–f, g–i). Though most tumors displayed brisk mitotic activity (n = 9), including atypical mitotic figures, mitosis was conspicuously sparse in 2 cases. Ki67 proliferation index ranged from 8 to 30%. Multinucleated giant cells, whether foreign body type or neoplastic tumor giant cells, were present in many of the cases (n = 5; 45.5%) dispersed throughout the neoplasm. Metaplastic osteoid deposition in the spindle cell component was observed in one case. Table 1 depicts the clinicopathological profile of the study group.

Fig. 1.

Many “morphologic faces” of SCSCC. Case 11: a Diffuse epithelial ulceration with underlying granulation tissue (HE; 4 ×) showing b scattered atypical polygonal cells (HE; 20x) which are c positive for p63 (IHC; 20 ×). Case 2: d Surface dysplasia with underlying tumor revealing polygonal to spindle cells (HE; 10 ×) with e marked pleomorphism, brisk mitosis, tumor giant cells (HE; 20 ×), and f focal CK positivity (IHC; 20 ×). Case 10: g Moderately pleomorphic spindle cells in fascicles (HE; 10 ×) with h CK expression (IHC; 10 ×) and i focal p40 expression (IHC;10 ×)

Table 1.

Clinicopathological profile of the study group (n = 11)

No	Age/sex	Clinical presentation	Tumor site	Nature of growth	Surface epithelium	Squamous component	Salient features of spindle component
1	42/M	Difficulty in opening mouth × 2 months	PE fold	Polypoidal	Complete ulceration	Infiltrative	Osteoid deposition
2	62/M	Difficulty in deglutition × 1 month	BOT, vallecula	Polypoidal	Full thickness dysplasia	Dysplasia + infiltrative	Marked pleomorphism, storiform pattern
3	52/M	Hoarseness of voice × 2 months	Right vallecula, epiglottis	Ulceroproliferative	Complete ulceration	Infiltrative	Bland spindle cells within abundant granulation tissue
4	60/M	Hoarseness of voice × 2 months	Left pyriform fossa	Ulceroproliferative	Complete ulceration	Infiltrative	Marked pleomorphism, storiform pattern
5	35/F	Growth in mouth × 3 months	Right inferior gbs	Ulceroinfiltrative	Hyperplastic to ulcerated	Infiltrative	Spindled to polygonal cells in sheets
6	64/F	Hoarseness of voice × 2 months	Glottis	Polypoidal	Mild dysplasia	Dysplasia only	Bland, prominent stromal hyalinization
7	44/M	Growth over tongue × 1 month	Tongue (right lateral border)	Polypoidal	Hyperplastic to ulcerated	No	Prominent giant cells
8	58/M	Growth over tongue × 2 months	Tongue (left lateral border)	Polypoidal	Complete ulceration	No	Marked pleomorphism, storiform pattern
9	45/M	Difficulty in deglutition × 2 months	Tongue (anterior 2/3)	Polypoidal	Complete ulceration	No	Prominent giant cells
10	33/F	Growth over inferior gbs × 2 months	Buccal mucosa and inferior gbs	Ulceroinfiltrative	Hyperplastic to ulcerated	No	Spindled to polygonal cells in sheets
11	80/M	Difficulty in opening mouth × 1 month	Right gbs up till right rmt	Ulceroproliferative	Complete ulceration	No	Dispersed singly; florid subepithelial inflammation

pe Pharyngoepiglottic, BOT base of tongue, gbs gingivobuccal sulcus, rmt retromolar trigone

IHC was performed in all cases, and results are detailed in Table 2. A total of 7 cases were positive for one or more epithelial markers (63.6%). CK (5 out of 11 cases) emerged as the most sensitive epithelial marker, followed by p40 and p63 (4 out of 11 cases each). EMA was the least sensitive among the epithelial markers with only single case showing weak expression. However, in 6 cases, epithelial component was identified on light microscopy, and therefore IHC did not play a diagnostic role. These 6 cases were thus disregarded for further evaluation of epithelial markers.

Table 2.

Expression of various IHC markers in the spindle cell component (n = 11)

Case No	CK	EMA	p40	p63	Vimentin	SMA	Desmin	S100	HMB 45, SOX 10	CD34	Ki67
1	-	-	+	+	+  +	+  +	-	-	nd	-	15%
2	+	-	-	-	+  +  +	-	-	+	-	-	10–15%
3	-	-	-	+	+  +  +	-	-	-	nd	-	15%
4	+	-	-	-	+  +  +	+  +	-	+	-	-	10–15%
5	+  +	-	+  +	+  +	+  +	+	-	+	-	-	8–10%
6	-	-	-	-	+  +  +	-	-	-	nd	-	5–6%
7x	-	-	-	-	+  +  +	-	-	+	-	-	30–40%
8x	-	-	-	-	+  +  +	-	-	-	nd	-	30–40%
9x	-	-	-	-	+  +  +	-	-	-	nd	-	15%
10x	+  +  +	+	+  +	-	+  +  +	-	-	-	nd	-	20–30%
11x	+  +	-	+  +	+	+  +  +	-	-	+  +	-	-	8–10%

^xNo obvious squamous differentiation on light microscopy, nd not done, + weak and focal, +  + strong and focal/or weak and diffuse, +  +  + strong and diffuse

Of the remaining 5 cases, which lacked light microscopic evidence of epithelial differentiation, IHC for epithelial markers was positive in 2 cases (40%) and was therefore instrumental in establishing the diagnosis of an epithelial malignancy. In this group, CK and p40 were positive in both cases (40%), whereas EMA and p63 were positive in 1 case each (20%). Thus, a combination of IHC with morphology yielded a diagnostic rate of 72.7% (8/11) versus 54.5% (6/11) on morphology alone. Among the mesenchymal markers, vimentin positivity was seen in all 11 cases (100%) with a diffuse, strong staining observed in most cases. Expression of other mesenchymal markers was also observed in some cases causing a diagnostic conundrum. S100 was positive in 5 cases (45.5%); staining varied from focal and weak in 4 cases to diffuse in 1 case. None of the S100 positive cases were positive for HMB45 or SOX10, thus ruling out melanocytic and neural differentiation respectively. Three cases (27.2%) revealed SMA expression, staining was focal and weak in all cases, and no diffuse strong SMA positivity was seen. Desmin and CD34 were negative in all the cases.

Three cases (case nos. 7, 8, and 9; Table 2) were negative for all epithelial markers. Out of these, two cases were also negative for all lineage-related mesenchymal markers, whereas one case showed focal S100 positivity but was HMB45 and SOX10 negative. Thus, use of an extended IHC panel comprising relevant mesenchymal markers helped to exclude other differential diagnoses. Accordingly, a final diagnosis of SCSCC was rendered in all three cases.

Patients were retrospectively followed up at the time of manuscript preparation. One patient was alive and well 84 months post primary diagnosis. Two patients expired within 1 month of diagnosis. Two patients died from complications of disease recurrence 6 months and 12 months post primary surgery respectively. One patient expired due to unrelated causes 36 months post-surgery. In one patient, who died 30 months post-surgery, the cause of death was not available. No follow-up information was available for the remaining four cases.

Discussion

SCSCC or sarcomatoid carcinoma was first described by Virchow in 1864 [3]. Though the histogenesis of this rare tumor was initially unclear, there is mounting evidence that SCSCC is a monoclonal epithelial neoplasm with a divergent (mesenchymal) differentiation, rather than a collision tumor, or biphasic derivation, or “pseudosarcoma” [3, 4]. Both the sarcomatoid and conventional squamous carcinoma (SCC) components have now been proven to arise monoclonally from a single stem cell, and evidence shows that the sarcomatoid component represents dedifferentiation and molecular progression of the conventional SCC component [5–7]. The tumor displays features of epithelial differentiation on electron microscopy, such as tonofilaments and desmosomes [8]. Glottis is the commonest site, and majority of patients present with hoarseness of voice, dysphagia, and/or airway obstruction. Male predominance is reported with a male:female ratio ranging from 3.7:1 to as high as 13:1, although this tumor is now becoming increasingly common in females [3, 4, 8]. In our study, a male:female ratio of 2.7:1 was seen. It is usually seen in 6th and 7th decades of life and commonly presents as a small (often < 2 cm) polypoidal or pedunculated mass (98.9%). Symptoms are usually of relatively short duration (< 1 year) before the patient seeks medical attention. Our study also showed similar clinical profile of patients. Although its exact cause is not known, studies have demonstrated a strong association with history of tobacco use (both cigarette smoking and tobacco chewing), alcohol abuse, and radiation exposure [3, 4, 8]. History of tobacco use could not be evaluated in our study; however, two patients had previous history of irradiation to head and neck region.

Histologically, diagnosis of SCSCC ideally requires demonstration of both malignant spindle cells and squamous cell differentiation. The latter may be seen as dysplasia of overlying epithelium, carcinoma in situ or infiltrating tumor; however, its incidence is reported to be highly variable in various studies (46.6% to 79.6%) [3, 4, 8]. We noted histological evidence of squamous differentiation in 54.5% (6/11) cases. When the squamous component is histologically evident, SCSCC can be diagnosed with confidence without need of corroboration by IHC. However, when the surface epithelium is ulcerated/denuded and/or epithelial differentiation is not morphologically evident, the correct diagnosis is more difficult to render. Differential diagnosis of SCSCC in such a scenario is broad and includes both reactive and benign spindle cell proliferations such as nodular fasciitis; fibromatosis; bizarre post-irradiation granulation tissue; leiomyoma; inflammatory myofibroblastic tumor; and malignant lesions such as fibrosarcoma, malignant fibrous histiocytoma, rhabdomyosarcoma, leiomyosarcoma, malignant peripheral nerve sheath tumor, melanoma, and myofibroblastic sarcoma [3, 4, 8, 9]. In cases where the squamous component is not evident at light microscopy, it may be elucidated by expression of epithelial markers by IHC, electron microscopy, or at the molecular level. Rate of immunopositivity with epithelial markers has been reported to vary from 48 to 75% in various studies [2, 4, 10, 11]. Epithelial marker expression in SCSCC diminishes as the degree of epithelial differentiation decreases and may be lost completely; hence, a negative epithelial IHC result does not rule out the diagnosis of this tumor [3]. Additionally, metaplastic elements and aberrant expression of mesenchymal marker (vimentin, calponin, SMA, and S100) are also reported in SCSCC, further confounding the differential diagnosis [3, 4, 9, 10]. Thompson et al. noted vimentin, SMA, S100, and desmin expression in 100%, 33%, 5%, and 2% of their SCSCC cases respectively [2]. We noted SMA positivity in 3 (27.2%) cases; however, staining was weak and patchy in all cases, and desmin was negative, thus ruling out leiomyosarcoma. S100 expression was seen in 5 (45.5%) cases, which was focal and weak in most cases, and only one case showed diffuse positivity. Use of more lineage-specific markers may provide clues for correct diagnosis in such cases [2, 12–14].

Though the morphologic diagnosis of SCSCC is challenging, thorough understanding of morphological and IHC features with a systematic approach is critical for accurate diagnosis to allow correct patient management. Dysplasia of surface epithelium and component of invasive SCC in a predominantly spindled tumor are important morphological clues. In the absence of obvious epithelial differentiation, a panel of IHC markers (CK and p40) is helpful in identifying epithelial differentiation in many cases. In our study, a combination of IHC with morphology yielded a diagnostic rate of 72.7% (8/11) versus 54.5% (6/11) on morphology alone. Judicious use of IHC also aids in ruling out morphological differential diagnoses such as sarcoma. Awareness regarding aberrant expression of certain mesenchymal markers and variable positivity to possible negativity for epithelial markers in SCSCC is vital to avoid misdiagnosis. It is also important to remember that primary head and neck sarcomas are very rare tumors (comprising only 1% of all head and neck malignancies), and so even if all IHC markers are negative in a mucosal-based spindle cell malignant neoplasm, SCSCC remains the first diagnostic consideration [3, 4, 8, 15].

Disease progression in SCSCC is characterized by recurrences and metastases, and clinical outcome remains poor. Comparison between prognoses of SCSCC versus conventional SCC remains contentious. Some studies report that SCSCC carries a poorer prognosis despite aggressive surgical intervention and mortality rate is higher than conventional SCC with tendency to recur and metastasize easily, especially in oral cavity and oropharyngeal tumors [16–18]. Survival rates for oral and oropharyngeal SCSCCs are lower than laryngeal and hypopharyngeal SCSCCs (27.5% to 36% vs 56.8%) [19–21]. Other authors have found that the clinical behavior of SCSCC parallels that of conventional SCC with the outcome being largely determined by tumor location and clinical stage [2–6]. Since most tumors are polypoidal and present with obstructive symptoms such as hoarseness, dysphagia, and dyspnea, they are detected early, and these cases tend to have better prognosis. Batsakis et al. reported certain features to be prognostically favorable, including (a) low stage, (b) polypoidal growth, (c) glottis site of origin, (d) relatively shallow depth of the sarcomatoid process, and (e) no prior radiation [22]. In contrast, poor prognostic factors include tumors diagnosed at higher stages, large tumors (> 3 cm) with a predominance of epithelial component, nonglottic tumors, fixed vocal cords, history of radiotherapy, metastasis to regional lymph nodes, and distant metastasis. Surgery is considered to be the mainstay in the management [13, 23, 24]. The effectiveness of radiotherapy was suggested by Ballo et al. in 1998; however, more recent studies did not confirm the impact of radiotherapy on survival [23, 24]. Radiotherapy may be an acceptable alternative in inoperable cases or in patients with nodal metastasis; however, the role of cytotoxic chemotherapy is unclear. Optimal treatment of this tumor is still unclear as it is a rare entity, and only retrospective reports have been published.

In conclusion, SCLCC is a rare tumor, morphological diagnosis of which is challenging with several histological mimics because of spindled morphology. Systematic morphological approach aided by appropriate IHC is helpful in clinching the accurate diagnosis in most cases. We found a basic IHC panel comprising CK, p40, p63, and vimentin to be most helpful in this regard. However, small sample size remains a limitation of our study. Finally, it is essential to reprise few key points:

1. Extensive tumor sampling coupled with diligent search will highlight the epithelial component in more than half of cases excluding the need for IHC.

2. IHC will highlight epithelial differentiation in a fair percentage of cases which show epithelial ulceration and/or lack of definite squamous component on light microscopy.

3. IHC for epithelial markers may be negative in SCSCC.

4. Aberrant mesenchymal marker expression in SCSCC is a known phenomenon and likely represents epithelial dedifferentiation and/or true divergent mesenchymal differentiation in the tumor.

5. Absence of epithelial markers with aberrant expression of mesenchymal markers can confound the picture further and may tempt the pathologist to make an erroneous diagnosis of sarcoma.

6. Evaluating the nature of mesenchymal marker positivity (weak and focal vs strong and diffuse) and use of more lineage-specific markers (e.g., SOX10 and neurofilament over S100 for neural differentiation, HMB45 over S100 for melanocytic differentiation, and desmin over SMA for smooth muscle differentiation) may provide clues for correct diagnosis. However, not many studies are available in this regard, and more such are required for validation.

7. Even in the absence of SCC component and negative epithelial markers, SCSCC remains the first differential diagnosis since both primary sarcomas of larynx/hypopharynx and benign mesenchymal laryngeal neoplasms are extremely rare and SCSCC is still more likely.

8. It may be wise to view all atypical spindle cell neoplasms of larynx as SCSCCs in order to ensure appropriate therapy, rather than diagnose these lesions by another term that would deliver inappropriate therapy, as suggested by other studies as well.

Author Contribution

Bakshi Neha: Design, Literature search, data acquisition, data analysis, manuscript preparation, manuscript editing. Dhawan Shashi: Concepts, definition of intellectual content, manuscript editing, and manuscript review. Rao Seema: Concepts, design, definition of intellectual content, manuscript editing, and manuscript review. Gaurantor: Dr. Seema Rao.

Data Availability

Not applicable.

Code Availability

Not applicable.

Declarations

Ethics Approval

Not sought (retrospective observational study).

Consent to Participate

Retrospective observational study.

Consent for Publication

Retrospective observational study.

Conflict of Interest

The authors declare no competing interests.