Fine-needle aspiration (FNA) biopsy often is the first diagnostic procedure performed in patients with head and neck masses. When squamous cell carcinoma (SCC) is diagnosed, proper management and improved prognosis depends on identification of the primary tumor. Recent studies have indicated that human papillomavirus (HPV) infection is associated closely with oropharyngeal SCC and that these tumors have a distinct nonkeratinizing morphology. In this study, the authors explored the value of identifying HPV-related tumors in neck metastases to determine the origin of occult primary head and neck squamous cell carcinoma (HNSCC).
Thirty FNA biopsies of neck metastases from patients with HNSCC were recovered from the authors' files from 2004 to 2005. The primary sites included 13 oropharynx, 13 oral cavity, and 4 larynx/hypopharynx. All patients had corresponding tissue samples from the neck mass and the primary carcinoma. The FNA specimens and corresponding tissue samples were classified as either nonkeratinizing SCC (NKSCC) or keratinizing SCC (KSCC). In situ hybridization for HPV (HPV-ISH) was performed using ethanol-fixed, Papanicolaou-stained smears. A positive signal was defined as dark blue or black nuclear dots. Corresponding formalin-fixed, paraffin-embedded tissue sections also were processed for HPV-ISH.
Twenty of the 30 FNA specimens were KSCC, and 10 were NKSCC. Eight of the 10 NKSCCs originated in the oropharynx, and 2 had nonoropharyngeal origin. HPV was detected in 7 of 10 NKSCCs. Ten of 30 (33%) FNA biopsies were positive for HPV, and 9 of those biopsies were metastatic from the oropharynx. Nonkeratinzing morphology or HPV-positive ISH in FNA samples significantly predicted oropharyngeal origin (P < .0069 and P < .0004, respectively).
Patients who present with cervical lymph node metastasis from an upper aerodigestive tract (UADT) squamous cell carcinoma (SCC) usually have a clinically identifiable primary tumor.1, 2 However, in approximately 3% to 5% of patients, the origin of the tumor is not clinically evident.3 Even after extensive diagnostic workup, including various imaging modalities and panendoscopy with targeted biopsies, only 1 in 3 primary tumors is identified.4–8 When the primary tumor remains unknown, wide-field irradiation that includes the entire pharyngeal axis and larynx usually is applied, resulting in increased morbidity. Therefore, it is important to identify the primary tumor to provide targeted therapy.
A significant number of initially occult primary SCCs that are metastatic to cervical lymph nodes originate in the palatine tonsils.4 Many of these tumors are related to human papillomavirus (HPV). The HPV-related tumors are characterized by a nonkeratinizing, basaloid cell morphology.9–12 The objective of the current study was to investigate the possibility of determining the site of the primary tumors by identifying HPV-related carcinomas, both morphologically and by using in situ hybridization (ISH), in the metastatic cervical lymph nodes in ethanol-fixed, Papanicolaou-stained smears from fine-needle aspiration (FNA) biopsies.
MATERIALS AND METHODS
Patients and Specimens
Study approval was obtained from the Institutional Review Board at Washington University Medical Center. Thirty FNA biopsies from patients who had head and neck SCC (HNSCC) metastatic to cervical lymph nodes with known primary sites were identified from the Department of Pathology database from 2004 to 2005. Patients were included in the study if they underwent follow-up surgical excision and pathologic diagnosis of the metastatic lymph nodes as well as the primary tumors. The corresponding surgically excised lymph nodes were retrieved.
FNA biopsy specimens were reviewed and divided into 2 categories—nonkeratinizing SCC (NKSCC) or keratinizing SCC (KSCC)—according to morphologic criteria. The neoplastic cells of NKSCC are relatively monomorphic with low nuclear and cellular pleomorphism. The nuclear-to-cytoplasmic ratio is high, and they display cellular cohesion, forming cell clusters. The dense eosinophilic cytoplasm associated with keratinization is absent in NKSCC (Fig. 1). Conversely, the cells of KSCC show more nuclear and cellular pleomorphism, abundant cytoplasm, and cytoplasmic keratinization, and they display less cohesion (Fig. 2).
All 30 specimens were processed for HPV-ISH by using FNA ethanol-fixed, Papanicolaou-stained smears. INFORM HPV III Family 16 probe from Ventana (Tucson, Ariz) was used according to the manufacturer's instructions. The probe detects high-risk HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 66. The slides underwent digestion with the Ventana ISH Protease 3 (8 minutes) and were hybridized with the biotinylated HPV cocktail probe. Signals were amplified through the application of streptavidin to bind biotin and were observed by a chromagen reaction with nitroblue tretrazolium and 5-bromo-4-chloro-3-indolyl-phosphate. Counterstain was performed with Ventana Red Counterstain II (4 minutes). A positive signal was defined as the presence of dark blue or black dots in the cell nuclei.
Formalin-fixed and paraffin-embedded tissue sections were deparaffinized and heated in citrate buffer for antigen retrieval according to the procedure described for the FNA biopsies. Polymerase chain reaction-proven HPV16-positive and HPV16-negative FNA specimens from tonsilar squamous carcinoma served as positive and negative controls, respectively.
Statistical analysis was performed using InStat (GraphPad Software Inc., San Diego, Calif). All P values <.05, as determined by using 2-tailed Fisher exact tests or chi-square tests with Yates continuity correction, were considered statistically significant.
Site distribution and morphologic features of the FNA and tissue samples from all patients are shown in Table 1. Twenty samples (67%) were KSCC, and 10 samples (33%) were NKSCC. Of the 10 metastatic NKSCCs, 8 (80%) originated in the oropharynx, 1 (10%) metastasized from the oral cavity, and 1 (10%) was from the larynx/hypopharynx. Table 1 shows that the distribution of KSCCs included 5 of 13 oropharyngeal tumors, 12 of 13 oral tumors, and 3 of 4 laryngeal/hypopharyngeal tumors. The morphologic features of the FNA biopsies and tissue sections were correlated in 25 of 30 patients (83%) (Table 1).
Table 1. Profile of Morphology and Human Papillomavirus Detection in Fine-needle Aspiration Biopsies and Tissue Samples
HPV was detected in 10 of 30 FNA biopsy specimens (33%) (Table 1, Fig. 3). Nine of 10 HPV-positive tumors were oropharyngeal, whereas only 1 was of nonoropharyngeal origin. None of the 13 metastases from the oral cavity were positive for HPV. Surgical specimens were positive for HPV in 8 of 30 specimens (27%), and all 8 had metastasized from the oropharynx (Table 1). The HPV results in FNA biopsies and surgical samples were correlated in 28 of 30 patients (93%) (Table 1).
The presence of nonkeratinizing morphology in FNA biopsies significantly predicted HPV association. Seven of 10 NKSCCs (70%) were positive for HPV, and only 3 of 20 KSCCs (15%) of were positive for HPV (Tables 1, 2). The specimens with NKSCC morphology and HPV positivity were metastatic tumors from the oropharynx. Two specimens with KSCC morphology and HPV positivity also were from the oropharynx; however, 14 of 17 metastatic tumors (82%) of nonoropharyngeal origin had KSCC morphology and were HPV negative (Table 2). Identification of nonkeratinizing morphology or HPV by ISH in cervical lymph node metastases was highly predictive of oropharyngeal primary (P < .0069 and P < .0004, respectively (Table 3).
Table 2. Correlation of Human Papillomavirus Detection and Morphologic Type on Fine-needle Aspiration Biopsies
Table 3. Nonkeratinizing Squamous Cell Carcinoma Morphology and Human Papillomavirus Positivity on Fine-needle Aspiration Biopsies From Oropharyngeal and Nonoropharyngeal Metastases
No. of samples/total (%)
Oropharynx, n = 13
Nonoropharynx, n = 17
NKSCC indicates nonkeratinizing squamous cell carcinoma; HPV+, human papillomavirus positive.
NKSCC and HPV+
Despite extensive radiographic and clinical evaluation, including magnetic resonance imaging, computed tomography scans, positron emission tomography scans, and multiple endoscopic biopsies, in approximately 3% to 5% of patients who present with cervical lymph node metastases, the primary tumor remains occult.3 The lack of a clinically identifiable primary tumor usually leads to more aggressive therapy, which can result in higher morbidity.3 Because FNA biopsies often are the first diagnostic procedure performed in patients who present with a neck mass, it is important to assess the applicability of various diagnostic tests to FNA samples.13, 14 The discovery that high-risk HPV is associated with a distinct subset of HNSCC with a characteristic nonkeratinizing morphology and high prevalence in the oropharynx has led us to explore the possibility that, by identifying HPV-related SCC in FNA biopsies of cervical lymph nodes, it may be possible to determine the site of an occult primary.9, 10, 15, 16
Conventional HNSCC differs from HPV-related SCC epidemiologically, clinically, histologically, cytologically, and molecularly. HPV-related SCC is more common in patients aged <40 years and usually presents as a small or occult primary with advanced neck disease. Tobacco use and alcohol abuse are not prevalent risk factors, and sexual habits are important in the transmission of the virus.9, 10 Most HPV-related SCC, as discussed above, arises in the oropharynx, particularly in the tonsils and the base of the tongue. They are radiosensitive tumors and have a better prognosis than the conventional keratinizing HNSCC. The HPV-related SCC has a nonkeratinizing morphology, and it exhibits strong and diffuse reactivity to p16, low or negative p53 staining, and high Ki67 labeling scores.9, 10
In the current study, we were able to detect high-risk HPV by ISH using ethanol-fixed and Papanicolaou-stained smears from FNA biopsies. We observed a high correlation (93%) between the FNA samples and corresponding tissue sections (Table 1). In only 2 samples was concordance between the FNA smear and the surgical section lacking.
We also observed that HPV-positive tumors were associated significantly with nonkeratinizing cytomorphology and arose from the oropharynx (Tables 2, 3). In the FNA smears, the neoplastic cells formed clusters or sheets and had relatively uniform hyperchromatic nuclei, coarse chromatin pattern, and scant cytoplasm. NKSCC comprised 10 of 30 samples (33%), and 7 of those samples (70%) were HPV-positive and originated in the oropharynx. We also observed that 9 of 10 HPV-positive FNA biopsies (90%) arose from the oropharynx, whereas only 1 was nonoropharyngeal (10%), arising in the larynx/hypopharynx. All of our oral cavity tumors were negative for HPV.
Begum et al. reported HPV16-ISH results from 77 FNA biopsies using paraffin-embedded, formalin-fixed cell block sections.17 They detected HPV16 DNA integration in 10 of 19 oropharyngeal carcinomas and in 3 of 10 squamous cancers with unknown primary, whereas tumors of nonoropharyngeal origin were negative.17 Umudum et al. performed HPV-ISH using ethanol-fixed smears from 11 cervical lymph nodes with metastatic HNSCC and detected HPV DNA integration in 1 patient with a nasopharyngeal primary and in 1 patient with an oral cavity primary.18 Their series did not include any oropharyngeal tumors.18
Our findings support previous studies on the utility of identifying HPV-related carcinomas by ISH in FNA biopsies of metastatic cervical lymph nodes in the prediction of an occult oropharyngeal primary. In addition, we have demonstrated that using the characteristic cytologic features of HPV carcinomas, in itself, is a reliable method for the diagnosis HPV-related carcinomas and oropharyngeal primary tumors. Specimens in which the cytologic features are ambiguous ISH for HPV may be used to confirm the identity of the carcinoma.
The authors thank Mr. Kevin Selle for his excellent technical assistance.