Remarks on sentinel node biopsy in head and neck cancer

Sir, The status of the regional lymphatics is one of the most important prognostic parameters in patients with head and neck cancer, and the presence or absence, level, and size of metastatic neck disease are crucial for treatment and survival. Due to the limited sensitivity and specificity of the usual diagnostic tools like ultrasound, CT, MRI, and PET (Stuckensen et al, 2000), a pathohistological staging of the neck was generally adopted to remove and detect occult metastases, which could not be detected by these imaging techniques. A large number of elective neck dissections (ND) where the pathohistological examination of the surgical neck specimen did not reveal any positive nodes was accepted. This surgical procedure was associated with risks and morbidity of the patients concerned. In the last years, beginning with a case report by Alex and Krag (1996), sentinel node biopsy (SNB) in head and neck cancer became a very interesting field of clinical investigation. Other investigators followed, and the recent article of Hoft et al (2004) mentioned some of them. Because the method of SNB was interlaced with the N0 neck, that is, a neck without clinically detectable nodal metastasis, the procedure generally adopted by investigators like Hoft et al (2004) was to carry out an elective ND parallely to SNB. The pathohistological results were compared, with the sentinel nodes carrying metastasis being the false-negative results of the imaging techniques. Experience still is limited, follow-up too short. We felt, however, that Hoft et al (2004) did not discuss some relevant topics in depth. The first problem is the definition of the N0 neck, in other words the method of examination on which the clinical diagnosis was based. Hoft et al (2004) based staging of the neck on ultrasound examination and had a rate of 24% positive sentinel nodes in 50 patients. Contributors to the first multicentre study on sentinel node biopsy in head and neck cancer (Ross et al, 2004) based their staging on either clinical palpation or radiological imaging techniques like CT and had an upstaging rate of 34% in 134 patients. Kovacs et al (2001, 2004a, c) had an upstaging rate of 515% following neck staging based on PET in 15 and 38 patients. Thus, SNB reflected the accuracy of the clinical and radiological staging methods, and the ideal diagnostic prerequisite for SNB is not yet found. The second problem is the dimension of the primary. Hoft et al (2004) stated that ‘if a complete peritumoral injection of the tracer is not possible, the patient is not eligible for the sentinel node method’. We agree that peritumoral accessibility is more important than T classification, but large T3 and T4 primaries pose problems due to destroyed lymphatic drainage. Local intraarterial induction chemotherapy, however, did not seem to alter lymphatic drainage (Kovacs et al, 2004b) and might be a modality of treatment that can be added prior to SNB reducing tumour size. The third and main problem is the omission of an elective ND, which would potentially achieve a benefit for the patient concerning risk, morbidity, and life quality. This would depend on the reliability of SNB. In the study of Hoft et al (2004), ‘no patient with tumor-free sentinel nodes was found to have a metastasis in a nonsentinel lymph node’. False-negative results have been very rare in all previous studies, too. This would encourage the omission of elective ND in favour of SNB. However, Hoft et al (2004) falsely stated that ‘So far, only Ross et al (2002) have reported on a study of a true biopsy of the sentinel lymph node without elective neck dissection’. Kovacs et al (2001) reported on true biopsy of the sentinel node without elective ND, and all consecutive patients have been treated that way (Kovacs et al, 2001, 2004a –c). Diagnostics using PET in combination with SNB considerably reduced the number of elective ND, and the inconspicuous follow-up time of 80 patients to date surpassing a median of 2 years makes it not likely that this will be paired with hazard. Some contributors of the mentioned multicentre study also adopted this procedure, and there is hope that SNB without elective ND will be the staging procedure of the future in a large number of head and neck cancer patients.


Sir,
The status of the regional lymphatics is one of the most important prognostic parameters in patients with head and neck cancer, and the presence or absence, level, and size of metastatic neck disease are crucial for treatment and survival. Due to the limited sensitivity and specificity of the usual diagnostic tools like ultrasound, CT, MRI, and PET (Stuckensen et al, 2000), a pathohistological staging of the neck was generally adopted to remove and detect occult metastases, which could not be detected by these imaging techniques. A large number of elective neck dissections (ND) where the pathohistological examination of the surgical neck specimen did not reveal any positive nodes was accepted. This surgical procedure was associated with risks and morbidity of the patients concerned.
In the last years, beginning with a case report by Alex and Krag (1996), sentinel node biopsy (SNB) in head and neck cancer became a very interesting field of clinical investigation. Other investigators followed, and the recent article of Höft et al (2004) mentioned some of them. Because the method of SNB was interlaced with the N0 neck, that is, a neck without clinically detectable nodal metastasis, the procedure generally adopted by investigators like Höft et al (2004) was to carry out an elective ND parallely to SNB. The pathohistological results were compared, with the sentinel nodes carrying metastasis being the false-negative results of the imaging techniques. Experience still is limited, follow-up too short. We felt, however, that Höft et al (2004) did not discuss some relevant topics in depth.
The first problem is the definition of the N0 neck, in other words the method of examination on which the clinical diagnosis was based. Höft et al (2004) based staging of the neck on ultrasound examination and had a rate of 24% positive sentinel nodes in 50 patients. Contributors to the first multicentre study on sentinel node biopsy in head and neck cancer (Ross et al, 2004) based their staging on either clinical palpation or radiological imaging techniques like CT and had an upstaging rate of 34% in 134 patients. Kovács et al (2001Kovács et al ( , 2004a had an upstaging rate of 5-15% following neck staging based on PET in 15 and 38 patients. Thus, SNB reflected the accuracy of the clinical and radiological staging methods, and the ideal diagnostic prerequisite for SNB is not yet found. The second problem is the dimension of the primary. Höft et al (2004) stated that 'if a complete peritumoral injection of the tracer is not possible, the patient is not eligible for the sentinel node method'. We agree that peritumoral accessibility is more important than T classification, but large T3 and T4 primaries pose problems due to destroyed lymphatic drainage. Local intraarterial induction chemotherapy, however, did not seem to alter lymphatic drainage (Kovács et al, 2004b) and might be a modality of treatment that can be added prior to SNB reducing tumour size.
The third and main problem is the omission of an elective ND, which would potentially achieve a benefit for the patient concerning risk, morbidity, and life quality. This would depend on the reliability of SNB. In the study of Höft et al (2004), 'no patient with tumor-free sentinel nodes was found to have a metastasis in a nonsentinel lymph node'. False-negative results have been very rare in all previous studies, too. This would encourage the omission of elective ND in favour of SNB. However, Höft et al (2004) falsely stated that 'So far, only Ross et al (2002) have reported on a study of a true biopsy of the sentinel lymph node without elective neck dissection'. Kovács et al (2001) reported on true biopsy of the sentinel node without elective ND, and all consecutive patients have been treated that way (Kovács et al, 2001(Kovács et al, , 2004a. Diagnostics using PET in combination with SNB considerably reduced the number of elective ND, and the inconspicuous follow-up time of 80 patients to date surpassing a median of 2 years makes it not likely that this will be paired with hazard. Some contributors of the mentioned multicentre study also adopted this procedure, and there is hope that SNB without elective ND will be the staging procedure of the future in a large number of head and neck cancer patients.

Sir,
Thank you for giving us the opportunity to respond to the commentary letter by Dr Kovacs.
In the third paragraph, Dr Kovacs is referring to staging methods. . Dr Kovacs' conclusion that 'sentinel lymph node biopsy reflected the accuracy of the clinical and radiological staging methods, and the ideal diagnostic prerequisite for SNB is not yet found' is quite similar to ours, namely the 'better the staging methods are in detecting small metastases the less occult metastases will be overlooked and the more valuable will be the impact of an additional sentinel lymph node procedure'. However, the rate of positive sentinel lymph nodes is not only influenced by the staging procedure but also by the pathohistologic work-up. The more the intensive sentinel lymph nodes are examined, the more the occult metastases will be detected, raising the percentage of patients with occult disease (Höft et al, 2002;Höft et al, 2004). In our study, we applied ultrasound examinations for staging of the neck as a radiological method and also performed fine-needle aspiration cytologies. The accuracy of US-guided aspiration cytology has been shown to be significantly better than that of CT or MRI (van den Brekel et al, 1991).
The fourth paragraph concerns the obstacle of performing a sentinel node biopsy on large tumours. In our group of patients, this is due to the fact that it was difficult to perform an endoscopic peritumoral injection on large tumours located in the pharynx and the larynx (Höft et al, 2004). Dr Kovacs adds that large tumours pose problems due to destruction of the lymphatic drainage, but he fails to provide data or literature on which his opinion is based. As we did not perform histologic studies of the lymphatics, we cannot add new information whether large tumours are more aggressive in destroying lymphatic vessels on their rim than are small tumours. However, the technique of the peritumoral injection is to inject the tracer adjacent to, and not into, the tumour. Thus, it does not really matter if a tumour destroys lymphatic vessels within its borders, as there will be intact lymphatics surrounding the tumour to take up the tracer even in large carcinomas.
In his last sentence of the fourth paragraph, Dr Kovacs suggests that an intra-arterial induction chemotherapy might be a modality in reducing tumour size prior to sentinel node biopsy. It is an interesting suggestion we would have liked to discuss in our article. However, Dr Kovacs' results were still in press when his letter of comment was written. Thus, the information he refers to was not accessible to us. Yet, although Dr Kovacs states that intraarterial chemotherapy did not seem to alter lymphatic drainage (Kovacs et al, 2004b), there is ample evidence that preoperative chemotherapy affects lymph nodes. Cohen et al (2000) describe lymph nodes showing areas of fibrosis, fat necrosis, histiocytic accumulation and granulation formation after neoadjuvant chemotherapy. Furthermore, metastatic foci can be completely obliterated by chemotherapy. Accordingly, Nason et al (2000) concluded in their study that preoperative chemotherapy is associated with an unacceptable high false negative rate for sentinel lymph node detection. If these changes are evident in chemotherapy, we would expect to find them in patients with intra-arterial chemotherapy, too. We are looking forward to the publication of Dr Kovacs' results and histologic findings of the lymphatics.
In the fifth paragraph, Dr Kovacs refers to the omission of an elective neck dissection potentially achieving a benefit for the *Correspondence: Dr S Hoft, E-mail: hoft@hno.uni-kiel.de