Detection of Lymph Node Metastases in the Neck: Radiologic Criteria

Regional metastasis is one of the most important factors in the prognosis and treatment of patients with head and neck squamous cell cancer. In addition, because lymphatic metastasis is a frequent event that impacts prognosis, a decision to treat the lymph nodes in the neck has to be made in almost all patients, even if metastases are not apparent clinically. It is therefore important to assess as reliably as possible whether a patient has regional lymph node metastases.

It is well known that palpation is an inaccurate technique to stage cancer in the neck. In a recent decision-analysis study, a risk of occult neck metastases (in a palpatory-negative neck) above 20% was found to be indicative for elective neck treatment, either radiation therapy or surgery. This risk of occult metastasis, which can occur in both sides of the neck, is determined by characteristics of the primary tumor such as size, site, and several biological criteria (1). Because of the increased risk of nodal metastases, even in clinically negative necks, most patients with tumors staged as T2 or larger undergo some form of elective neck treatment. The disadvantage of this policy is that the majority of patients do not harbor metastases and, therefore, will be subjected to the morbidity of unnecessary treatment. By detecting some otherwise clinically occult adenopathy, modern imaging techniques may have increased sensitivity for detecting positive nodes, and consequently, may decrease the risk of occult metastasis to below 20%. If this can be accomplished, the clinician may refrain from a neck dissection or radiation, and adapt a wait-and-see policy with careful follow-up to detect a neck metastasis as early as possible (2).

Imaging techniques like CT, MR, and sonography are more accurate than palpation. Most clinicians have maintained, however, that the accuracy of these techniques is not high enough to justify a change of policy. Indeed, in 25% of pathologically verified tumor-positive neck dissections, only micrometastases smaller than 3 mm, which are undetectable by most techniques, are present (3). Lymph nodes 2–3 mm in size can be seen as nodules on CT and MR images, and may even be better seen with high-resolution scanners. Nonetheless, differentiation between benign and malignant metastatic disease still remains a problem. Recently, other techniques such as radioimmunoscintigraphy (4) and positron emission tomography (5) have been explored, but these expensive techniques still have to prove their value in clinical practice.

Sonographic criteria, such as nodal size and configuration of the lesion, and Doppler sonographic criteria have been studied extensively for their value in differentiating between benign and malignant lymphatic disease in the neck. The minimal axial diameter appears to be the most accurate size criterion, compared to the maximal axial diameter and the longitudinal diameter (6, 7). Regarding the aspect of lymph nodes on sonograms, the echogenic hilus appears to be a reliable parameter (7). The configuration (shape) of the node might be important, but some authors doubt its value (8). Sonography-guided fine-needle aspiration cytology (FNAC) has been shown to be very accurate in the evaluation of regional metastatic disease. It combines the high sensitivity of sonography with the excellent specificity of FNAC. The reported sensitivity of sonography-guided FNAC in the N0 neck ranges from 48% to 73% (6, 9, 10), whereas the reported specificity is 100% (11). In the United States, this technique has received less acceptance because it is labor-intensive and operator-dependent. False-negative results may be the result of sampling the wrong node or the wrong part of the correct node. Furthermore, the cytopathologist may overlook small nests or single tumor cells.

The potential value of Doppler sonographic criteria (avascular pattern, scattered pattern, peripheral vascularity) as an adjunct to differentiate between benign and metastatic lymph nodes has been the topic of various reports. Because gray-scale criteria are not very accurate, there is a great need for additional criteria for small lymph nodes. In this issue of the AJNR, Yonetsu et al (page 163) report on the Doppler sonographic findings in 338 lymph nodes from 73 patients with head and neck cancer in an effort to improve the accuracy of conventional sonography. The authors were able to define a new, more accurate combination of size and Doppler criteria. They report that the combined use of short-axis diameter and Doppler blood flow pattern (the absence or presence of “normal” hilar flow) increased the diagnostic accuracy compared to the use of short axis diameter alone. These combined criteria yield a very high sensitivity (> 89%) and specificity (> 94%). The authors performed a histologically verified study in which they analyzed criteria for metastatic lymphadenopathy. Although the authors performed a large study and were able to improve the results of conventional sonography significantly, the article raises some methodological questions. To allow comparison of results of different authors, the reporting of the results should be as uniform as possible. For neck imaging, the sensitivity and specificity per neck side should be reported, ideally for the clinically N0 neck separately, as this is the most clinically relevant. The size criteria used should be defined, as well as the histopathologic techniques used to assess the specimen. If the neck is categorized into different levels, the definition of the levels as proposed by the American Academy of Otolaryngology should be used. Yonetsu et al compared sonographic-histopathologic findings per node, and correlated the nodes on the sonogram and in the surgical specimen on the basis of relation to surrounding structures and size of the nodes. This method may cause false sonographic-histopathologic correlations, however, especially in the case of small nodes. Because only a limited number of nodes per neck side were correlated, there might have been small metastatic nodes in the specimen that were not seen on sonograms and thus not included in the study. The authors introduced a new classification of neck levels without indicating how it differed from the internationally accepted classification. Furthermore, the authors defined their short-axis diameter as “the greatest diameter on the maximum axial cross-sectional area of a node.” Actually, this appears to be the maximum axial diameter. The paramount question is what value can be assigned to these findings for making treatment decisions in patients with head and neck cancer?

First, the predictive value of power Doppler parameters, such as hilar blood flow, remains controversial. In the March 2000 issue of the AJNR, this same group compared in a multivariate study the gray-scale and power Doppler parameters (7). In contrast to their current article, they reported that Doppler features did not add significant predictive value to gray-scale criteria in differentiating metastatic from reactive nodes.

Second, clinicians are especially interested in the accuracy of modern imaging techniques for staging of the N0 neck. The high sensitivity and specificity reported in this article only have a limited clinical value as they were calculated per lymph node, and probably a majority of the metastatic lymph nodes were palpable. As a consequence, the sensitivity would have been lower if the study had been limited to an N0-neck population.

Third, clinicians are more interested in the status of the entire neck than in the presence or absence of metastatic disease in lymph nodes separately. Thus, after having defined the optimal criteria in single nodes, it would have been very interesting to see if these new criteria were advantageous for the entire neck as well.

Yonetsu et al studied the potential of a combination of gray-scale and duplex sonography criteria to discriminate between benign and malignant disease, and found a possibly valuable new combination of criteria. They should be given credit for this observation and report. The clinical value and accuracy of this new combination of criteria will have to be assessed in a population of patients with cancer and stage N0 (clinically nonpalpable) neck disease. If the duplex criteria are accurate in very small lymph nodes, there might be an important impact on clinical decision making. On the other hand, a combination of gray-scale and duplex sonography criteria may also be helpful in selecting nodes for sonography-guided FNAC. The use of contrast-enhanced Doppler sonography may also further increase the reliability and accuracy of duplex criteria in small lymph nodes.