Metastatic carcinomatous cells were observed after HES staining in 33 (2.5%) of the 1328 lymph nodes. The frequencies of nodal metastases were 2.0% (six of 301) in the oral cavity, 0.9% (two of 226) in the oropharynx, 3.3% (25 of 755) in the hypopharynx and 0% (zero of 46) in the larynx (). Of the 178 cervical lymph node levels, 29 (16.3%) investigated were considered to be metastatic. The details of these results are presented in .
Results of histopathologic examination and CK19 mRNA real-time RT–PCR in the various cervical lymph node levels expressed as the number of positive levels/total number of levels investigated
Real-time RT–PCR assay
No amplification product was observed when no template was added. The amplified products obtained from RNA extracted from CAPAN-1 cell line for CK19 and HLAIC show the expected bands (75 and 63
bp) after agarose gel electrophoresis (data not shown). CK19 mRNA was not detected in either PBMN samples from healthy subjects or in RNA extracted from the 17 lymph nodes collected from patients without cancer (data not shown). Human leukocyte class IC expressed in histopathologically negative lymph nodes (Ct
=24.78±1.05) close to the levels observed in histopathologically positive lymph nodes (Ct
=24.35±1.25) has been demonstrated to be a particularly accurate internal control (P
CK19 mRNA was detected in 236 lymph nodes from 64 lymph node levels; 203 of these lymph nodes were negative after histopathology examination. Detailed results of CK19 mRNA detection by real-time RT–PCR according to tumour site are shown in . Real-time RT–PCR gave a positive signal in 36.0% of all lymph node levels, while histopathology detected carcinomatous cells in 16.3% of lymph node levels. On the basis of these results, molecular analysis using real-time RT–PCR for CK19 mRNA detection gave significantly different results from routine histopathologic examination (P<0.0001).
Individual results per patient are shown in for patients who underwent unilateral neck dissection and in for patients who underwent bilateral neck dissections. Patients were staged as pN according to the usual histopathologic criteria (Hermanek et al, 1998
) and mN according to CK19 mRNA real-time RT–PCR results. The nodal status of 13 (45.2%) of 31 patients (eight with unilateral neck dissection, five with bilateral neck dissections) would therefore have been changed if the molecular results were taken into account. Moreover, in nine of these 13 patients (five with unilateral neck dissection, four with bilateral neck dissections), nodal status would have been upstaged from pN0 to mN1, mN2b or mN2c, respectively (Hermanek et al, 1998
Results of histopathologic and molecular node status in patients with unilateral neck dissection
Results of histopathologic and molecular node status in patients with bilateral neck dissections
In the 16 patients with lymph nodes metastases, analysis of the pattern of carcinomatous spread within lymph node levels showed that nodes from the IIa and IIb levels were free of metastatic cells on pathologic examination in five (16%) of them: CK19 mRNA was detected, while HES was negative in IIa and IIb lymph node levels in two of those patients.
In the 64 lymph node levels showing CK19 mRNA expression, the number of CK19 mRNA copies detected by real-time RT–PCR was evaluated by the RQ calculation as shown in . The expression values of all histopathologically positive lymph node levels ranged from 4.85 × 10−4 to 4.15 × 10−1, values that were significantly higher than those for histologically negative lymph node levels, which ranged from 1.13 × 10−5 to 1.42 × 10−3 (P<0.0001).
Figure 2 RQ (relative quantification of CKmRNA copies) according to the equation: of CK19 mRNA expression in 64 lymph node levels (29 histopathologically positive; 35 histopathologically negative). Immunohistochemistry
In the 203 lymph nodes that were histopathologically negative but CK19 mRNA positive on real-time RT–PCR, no carcinomatous cell was immunodetected by IHC, while in all the five HES-positive lymph nodes, carcinomatous cells were immunolabelled ().
Immunohistochemistry on lymph nodes containing carcinomatous cells detected after HES staining (right) after incubation with an anticytokeratin monoclonal antibody (left) (final magnification × 400).
Carcinomatous cell diffusion in cervical lymph nodes is a major determinant of therapy and prognosis for patients with HNSCC (Shores et al, 2004
) as cure rates for patients with pathologically metastatic lymph nodes drop to one-half of those of patients without nodal involvement (Shah, 1990
). As histopathologic analysis of neck dissection specimens is usually performed on several 3–4
mm sections from each lymph node, and as micrometastases represent tumour deposits measuring less than 2
mm in diameter, they can be easily missed on routine light microscopy (Hermanek et al, 1999
; Genden et al, 2003
). For example, it has been reported that 21.9% of patients with cancer of the oral cavity have micrometastases with an average diameter of 1.36
mm (Hamakawa et al, 2000b
Various molecular markers, as mRNA expressed in tumour cells, particularly CK mRNAs, have therefore been investigated in order to detect the presence of occult tumour cells in lymph nodes (Shores et al, 2004
). Previous studies using such markers in HNSCC have suggested a low specificity (Kawamata et al, 1999
; Hamakawa et al, 2000a
), particularly because of false-positive results associated with benign epithelial rests in lymph nodes found in 1.6% of cervical nodes, but only in lymph nodes from level I (Gnepp, 2000
; Hamakawa et al, 2000b
). The assay used in the present study, that is, RT–PCR detection of CK19 mRNA, has been developed in our laboratory for the detection of occult carcinomatous cells in mediastinal lymph nodes in non-small-cell lung carcinoma (Saintigny et al, 2005
) and by other teams in various carcinomas (Van Trappen et al, 2001
; Inokuchi et al, 2003
; Favia et al, 2004
). The primers and probe for CK19 mRNA detection were designed to avoid the detection of the two known pseudogenes (Rund et al, 1999
; Van Trappen et al, 2001
; Saintigny et al, 2005
) and we used real-time RT–PCR with the Taqman chemistry. All blood cell samples from healthy donors and lymph nodes specimens collected from patients without cancer were negative, while CK19 mRNA was detected in all samples from the 29 lymph node levels shown to be metastatic on histopathology. We therefore consider that CK19 mRNA is a reliable marker for carcinomatous cell detection in HNCC as in other carcinomas (Van Trappen et al, 2001
; Saintigny et al, 2005
In this study, CK19 mRNA was detected in 35 of the 178 lymph node levels that were free of tumour cells on histopathology. Moreover, quantitative evaluation by the RQ of the amount of CK19 mRNA clearly showed a significant difference between histopathologically positive and negative lymph nodes. In a previous study, using dilution of CAPAN-1 cells in blood cells, we showed that the threshold of this method can be considered to detect one tumour cell within 104
PBMNS (Saintigny et al, 2005
These data suggest that real-time RT–PCR could potentially increase the sensitivity of routine pathological examination (Becker et al, 2004
; Ferris et al, 2005
), and could therefore be used routinely to examine all nodes removed during neck dissection that are negative on standard histopathologic examination.
Clinically, although the prognostic significance of micrometastases has not been established in HNSCC, a previous study suggested that the presence of occult micrometastases increases the risk of regional recurrence (Rhee et al, 2002
). According to Ferlito et al (2000)
, selective neck dissection in HNSCC provides important information for prognostic purposes and therapeutic decisions. In the present study, occult micrometastases were detected in 13 out of 31 patients, and if these findings were taken into account, they would have changed the patients' nodal status and nine pN0 patients would have been mN-upstaged.
According to the work of Rouvière (1938)
, the lymphatic drainage of the head and neck region follows a relatively constant and sequential route. Level II is the first node level in the majority of HNSCC, but other levels are also likely to be the first node level, especially in tumours of the oral cavity (level I) or hypopharynx (level III) (Schauer et al, 2004
). As a variable lymphatic drainage in HNSCC can be observed, tumour cells may be missed in unsuspected groups of lymph nodes (Byers et al, 1997
). In our study, this variability of lymphatic drainage was observed in five out of 31 patients. In two of these patients, CK19 mRNA was detected in IIa and IIb lymph node levels that were negative on histopathology. In support of this idea, Mozzillo et al (2001)
recently noted a 18% incidence of sentinel nodes outside the expected nodal drainage pattern, and Byers et al (1997)
found that 15.8% of 277 patients with squamous cell carcinoma of the oral tongue had skip metastases.
Serial sectioning and IHC have been extensively developed for the detection of occult metastasis in HNCC and in a wide range of solid tumours (Ambrosch and Brinck, 1996
). This method has the advantage of preserving cellular and tissue morphological features, but is expensive and time-consuming, and therefore cannot be used for most routine applications (Becker et al, 2004
). However, it has been demonstrated that it does not allow a significantly higher detection rate than histopathologic examination of HE-stained sections (Woolgar, 1999
; Hamakawa et al, 2000b
). In our study, immunodetection of carcinomatous cells of the 203 nodes that were HES negative and real-time RT–PCR positive was negative. However, the clinical relevance of this detection, by real-time RT–PCR only, with respect to prognosis and treatment needs to be investigated by further clinical studies.
In conclusion, these results indicate that real-time RT–PCR assay for the detection of CK19 mRNA is a sensitive and reliable method for the detection of carcinomatous cells in lymph nodes in patients with HNSCC. This type of method could be used to reassess lymph node status according to occult lymphatic spread in patients with HNSCC.