| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Departments of Radiology (H.v.O., M.W.H., A.P.v.G.) and Pulmonology (A.C.R.), Leyenburg Hospital, Leyweg 245, 2545 CH, The Hague, the Netherlands; Departments of Radiology (K.B.) and Pulmonology (J.H.L.M.v.K., C.N.F.v.d.M.), St Anna Hospital, Geldrop, the Netherlands; and Department of Biostatistics and Epidemiology (B.E.H.), Erasmus University, Rotterdam, the Netherlands. From the 2002 RSNA scientific assembly. Received April 26, 2003; revision requested July 8; final revision received October 21; accepted November 20. Address correspondence to H.v.O. (e-mail: h.voverhagen@leyenburg-ziekenhuis.nl).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: One hundred seventeen consecutive patients (91 men and 26 women; mean age, 64.0 years) underwent palpation, US, and CT of supraclavicular regions and chest and upper abdominal CT. Fine-needle aspiration cytologic (FNAC) analysis was performed in patients with nodes with a short-axis diameter of 5 mm or greater; cytologic diagnosis was used as the standard of reference. Sensitivities of palpation, US, and CT were compared with McNemar testing. Relationship between size and palpability of nodes with metastasis was evaluated with logistic regression.
RESULTS: Supraclavicular metastases were diagnosed cytologically in 30 (26%) of 117 patients: eight (31%) of 26 patients with small cell lung cancer (SCLC) and 22 (24%) of 91 patients with non–small cell lung cancer (NSCLC). Sensitivities of US (1.00; 30 of 30 patients) and CT (0.83; 25 of 30 patients) for detection of metastases were significantly higher (P < .001 and P = .001, respectively) than that of palpation (0.33; 10 of 30 patients). Palpable nodes with metastasis (mean diameter, 25.2 mm) were significantly larger than nonpalpable nodes with metastasis (mean diameter, 13.7 mm) (P = .002). To have a 50% chance of being palpable, nodes with metastasis had to have a diameter of at least 22.3 mm. TNM stage was changed in three of 91 patients with NSCLC, and further invasive diagnostic procedures were prevented in 11 of such patients because it was proved that nonpalpable nodes had metastases.
CONCLUSION: Supraclavicular lung cancer metastases were cytologically proved in 26% of patients. Nodes with metastasis were only palpable when markedly enlarged. US tripled the sensitivity of palpation for detection of metastases. Results of US and US-guided FNAC analysis can change the work-up in patients with lung cancer.
© RSNA, 2004
Index terms: Lung neoplasms, metastases, 60.3213, 60.3214 • Lung neoplasms, staging, 60.3213, 60.3214 • Lymphatic system, biopsy, 997.12985 • Lymphatic system, CT, 997.12912, 997.12915 • Lymphatic system, US, 997.12981
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The prognosis of pulmonary carcinoma is primarily based on the histologic type of tumor and its stage at the time of diagnosis. Thus, accurate assessment of the anatomic extent of the primary tumor and of the presence or absence of lymphatic and hematogenous metastases is essential for selection of appropriate curative or palliative therapy. Despite improvement during the past decade in such imaging techniques as computed tomography (CT) and positron emission tomography (PET) with fluorine 18 (18F) fluorodeoxyglucose (FDG), cytologic or histologic proof of metastases is generally warranted in patients before their disease is considered inoperable. However, in patients with lung cancer, metastases are usually located in lymph nodes within the mediastinum, and these are difficult to reach with minimally invasive techniques (2). An exception is the moresuperficially located supraclavicular lymph nodes. These nodes are of substantial importance because the presence of supraclavicular metastases is associated with incurable disease (3).
Examination of the supraclavicular lymph nodes has traditionally been performed with palpation. In cases of nonpalpable nodes, supraclavicular metastases have been assumed to be absent. However, during the past 2 decades, results of studies of head and neck carcinoma (4), esophageal carcinoma (5), and melanoma (6) have shown that palpation is an unreliable method for the assessment of metastases in lymph nodes in the neck, supraclavicular regions, and groin. These investigators have suggested that noninvasive imaging techniques such as CT and ultrasonography (US) be used for improved detection of metastases in lymph nodes.
The aims of the present study were to compare US, CT, and palpation for diagnosing supraclavicular lung cancer metastases and to assess the effect of proved metastases on TNM stage and diagnostic work-up.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
For final inclusion in the study, histopathologic proof of primary pulmonary carcinoma—either for the primary intrathoracic tumor or for an extrapulmonary metastasis—had to be obtained.
Our study was approved by the medical ethics committees of both hospitals, and informed consent was obtained from all included patients.
Ninety-six of the 213 patients did not meet the inclusion criteria. Forty-four patients had either prior or coexisting malignancy elsewhere, 12 patients had a history of neck or thoracic surgery or radiation therapy, and 40 patients did not provide informed consent. Thus, 117 patients with cytologically or histopathologically proved primary pulmonary carcinoma were included in the study. Characteristics of these patients are shown in Table 1. There were 91 men (mean age, 65.4 years) and 26 women (mean age, 59.2 years). Ninety-one (78%) patients had NSCLC, and 26 (22%) had SCLC.
|
Palpation, CT, and US were performed by different examiners (H.v.O., K.B., J.H.L.M.v.K., C.N.F.v.d.M, A.C.R., A.P.v.G.), and the findings at each examination were documented independently, without knowledge of the findings at the other examinations.
Palpation
The supraclavicular region of each patient was palpated by one of the three referring experienced pulmonologists (J.H.L.M.v.K., C.N.F.v.d.M., and A.C.R., each of whom had >10 years of experience) for the presence of enlarged lymph nodes. The upper limit of the palpated area extended 4 cm above the clavicular heads. An enlarged node was considered present when an abnormally firm mass was palpable. The presence, number, and location of enlarged supraclavicular lymph nodes were documented.
CT Scanning and Image Interpretation
All patients underwent CT of the chest and upper abdomen from the pulmonary apices through the adrenal glands. One hundred milliliters of contrast material (iohexol, Omnipaque 300; Nycomed Amersham, Princeton, NJ) was administered intravenously to achieve better definition of the vascular structures in the neck and mediastinum. The supraclavicular regions were subsequently scanned from the cricoid cartilage to the pulmonary apices.
At St Anna Hospital, CT was performed with a Volume Zoom multisection CT scanner (Siemens, Erlangen, Germany) and the following parameters: a contrast material infusion rate of 2 mL/sec, a scanning delay of 50 seconds, collimation of 4.0 x 2.5 mm, a reconstruction increment of 7 mm with a pitch of 6 for the chest and upper abdomen, and a reconstruction increment of 5 mm with a pitch of 5 for the neck. At Leyenburg Hospital, a helical Somatom Plus CT scanner (Siemens) was used to obtain 10-mm-thick sections of the chest and 5-mm-thick contiguous sections of the neck with a contrast material infusion rate of 3 mL/sec, a scanning delay of 30 seconds, and a table speed of 10 mm/sec.
CT scans were interpreted by four radiologists (including A.P.v.G.) who each had more than 10 years of experience. Mediastinal and hilar lymph nodes were considered to be enlarged when they had a short-axis diameter of 10 mm or greater. The CT scans were also assessed for the size of the primary tumor; for the presence of pleural fluid; and for evidence of metastases in the adrenal glands, lungs, and skeleton. CT scans of the supraclavicular regions were assessed for adenopathy, which was defined as the presence of enlarged lymph nodes with a short-axis diameter of 5 mm or greater. All supraclavicular CT examinations were performed before US to prevent artifacts caused by cytologic sampling.
US and Image Interpretation
US of the supraclavicular regions (extending up to 4 cm above the clavicular heads) was performed by using a 5–10-MHz linear array transducer (Sonoline Elegra, Siemens; or ATL Ultramark 3000, Philips Medical Systems, Best, the Netherlands). US examinations were performed by one of four radiologists (including H.v.O. and K.B.), each of whom had more than 10 years of experience. Supraclavicular lymph nodes were considered to be enlarged when they had a short-axis diameter of 5 mm or greater.
The presence, number, site, and size of enlarged lymph nodes were documented, and US-guided fine-needle aspiration cytologic analysis was performed in all cases of lymph nodes with a short-axis diameter of 5 mm or greater. When multiple enlarged lymph nodes were present, the largest node was sampled.
US-guided Fine-Needle Aspiration Cytologic Analysis
US-guided fine-needle aspiration procedures were performed by one of the four radiologists who performed the US examinations, usually in conjunction with one of three pathologists (each of whom had >10 years of experience). A 23-gauge needle was used, local anesthesia was not provided, and the procedure typically was completed within 10 minutes. Fine-needle aspiration cytologic analysis was always initially performed in the largest supraclavicular node, and, if results in this node were positive (indicating metastasis), no further samples of these nodes were obtained. Cytologic smears were examined to guarantee that the specimens were diagnostically adequate. Smears were considered diagnostically adequate when they contained either tumor cells or lymphocytes. In cases of diagnostically inadequate specimens, cytologic sampling was repeated during the same session.
Change in Diagnostic Work-up
For each patient with cytologically proved supraclavicular metastases, the referring pulmonologist had to record, by using standardized questionnaires, whether the presence of these metastases had changed the diagnostic work-up and/or resulted in avoidance of more invasive diagnostic procedures. If diagnostic procedures had been avoided, the referring pulmonologist had to state which procedures (eg, mediastinoscopy, biopsy of the chest or abdomen) had specifically been prevented.
Statistical Analysis
The ages of the male and female patients were compared by using the Student t test. For each patient palpation, US or CT findings in the supraclavicular regions were scored as positive or negative. The examination results were regarded as positive when there was at least one enlarged node.
Results of palpation, CT, and US performed by the different examiners were collected and compared (by M.W.H.) with the results of cytologic examination. In those cases in which cytologic results were not available and there was discrepancy between findings at palpation, CT, and US, these examinations were repeated or their results were reassessed and effort was made to reach consensus regarding the presence or absence of supraclavicular metastasis.
The sensitivities of palpation, CT, and US were calculated and compared by using the McNemar test with exact significance values. Results of cytologic examination were used as the reference standard for the presence or absence of supraclavicular metastasis. Logistic regression was applied to relate the size (at US) and palpability of supraclavicular lymph nodes with metastasis. Cytologically proved supraclavicular metastases were related to the histologic findings in and the local extent of the primary tumor and to the presence of mediastinal lymph node involvement and/or distant metastasis at CT by using the 
2 test.
At both hospitals, clinical staging of the disease was performed at CT by using the TNM classification, with T indicating the size and site of the primary tumor, N indicating nodal involvement according to the site, and M indicating the presence or absence of distant metastasis. We assessed the number of patients with NSCLC in whom the TNM stage and the decision as to whether or not the cancer was operable were changed owing to the presence of these supraclavicular metastases. For these purposes, patients with NSCLC were separated into two groups: those with operable (stage I–IIIA) cancer and those with inoperable (stage IIIB or IV) cancer.
The same procedure was performed to determine how many patients with SCLC had had their disease classification changed from limited to extensive. These two groups were assumed to undergo different nonsurgical therapeutic regimens.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Enlarged supraclavicular lymph nodes were palpated in 15, detected at CT in 34, and detected at US in 38 of 117 patients. US-guided fine-needle aspiration cytologic analysis was performed in 38 patients. Thirty (79%) of 38 specimens were cytologically proved to contain metastasis, while the remaining eight specimens (21%) had lymphocytes but no malignant cells.
Supraclavicular metastases were proved in 30 (26%) of 117 patients: eight (31%) of 26 patients with SCLC and 22 (24%) of 91 patients with NSCLC (Table 2).
|
US (P < .001) and CT (P = .001) were significantly more sensitive than palpation for detecting supraclavicular metastases (Fig 1). Palpable nodes with metastasis (mean largest diameter, 25.2 mm; range, 10–50 mm) were significantly larger than nonpalpable nodes with metastasis (mean largest diameter, 13.7 mm; range, 10–27 mm) (P = .002). The relationship between the diameter and the palpability of supraclavicular lymph nodes with metastasis is shown in Figure 2, which illustrates the fact that these nodes had to have a diameter of 22.3 mm or greater to be palpated in 50% of cases.
|
|
|
In five patients who were initially suspected of having enlarged supraclavicular nodes at palpation, enlarged nodes or other abnormalities were not found at US and CT. Nodes without metastasis were found at CT and at US in nine and eight patients, respectively. In four of these patients, enlarged nodes were seen at CT only. US was repeated with the knowledge of the results of the CT examinations; however, the nodes were not seen and fine-needle aspiration cytologic analysis was therefore not performed. At reassessment, the CT results were considered to be false-positive. During clinical follow-up, supraclavicular metastases were not observed in these patients. Thus, the predictive value that an enlarged supraclavicular lymph node will contain metastasis was 0.67 (95% CI: 0.40, 0.94) at palpation, 0.74 (95% CI: 0.58, 0.89) at CT, and 0.79 (95% CI: 0.65, 0.93) at US.
The relationship between the size of the primary tumor at CT and the presence of supraclavicular metastases is shown in Table 3. The presence of supraclavicular metastases was independent of the extent of the primary tumor (P = .90).
|
|
|
In the group of patients with SCLC, disease in one patient with palpable contralateral supraclavicular metastases was reclassified from limited to extensive. Disease in another patient with nonpalpable ipsilateral supraclavicular metastases was reclassified only in terms of N status (from N0 to N3) and was still considered limited.
We were able to prevent more invasive diagnostic procedures from being performed in 18 of 22 patients with NSCLC and cytologically proved supraclavicular metastases (seven patients with palpable supraclavicular metastases and 11 patients with nonpalpable supraclavicular metastases) (Table 6). The remaining four patients with NSCLC and nonpalpable supraclavicular metastases had such extensive metastatic disease that their referring pulmonologists considered their cancers inoperable on the basis of their clinical condition and chest CT findings.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Therefore, improvement of the diagnostic work-up in patients with lung cancer so that the optimal therapeutic approach for the individual patient can be selected and unnecessary invasive procedures in patients with advanced disease can be avoided remains important.
The presence of metastases in supraclavicular lymph nodes in lung cancer is critical in NSCLC because it defines the N status as N3 and thus the stage as IIIB (inoperable).
In most staging protocols for patients with lung cancer, the presence of metastases in supraclavicular lymph nodes is evaluated with palpation alone, and a prevalence of 12% has been reported (9). However, at autopsy, these metastases have been found in 37.5% of cases (10). Because of this, the use of US (11–13) or extended cervical mediastinoscopy (14) has been advocated for better assessment of these lymph nodes in patients with lung cancer.
In other malignancies, such as head and neck cancer (4), esophageal cancer (5), and melanoma (6), US and US-guided fine-needle aspiration cytologic analysis have proved to be superior to palpation for the detection and diagnosis of metastasis in supraclavicular and cervical lymph nodes. Because of this, US and US-guided fine-needle aspiration cytologic analysis have become widely accepted diagnostic techniques in the work-up of these patients.
This encouraged us to use these techniques in the setting of lung carcinoma. In our study, supraclavicular metastases were cytologically proved in 30 (26%) of 117 patients with lung carcinoma and were underestimated with palpation alone. In addition, supraclavicular nodes with metastasis could be palpated only when they had reached a certain size.
Our finding that 26% of patients proved to have supraclavicular metastases is comparable to the results reported by Fultz and colleagues (15). The finding that 24% of our patients with NSCLC proved to have supraclavicular metastases is a slightly lower value than the 31% they reported (15). This difference may be explained by the fact that their patient group consisted of patients who were referred to a tertiary care center.
In our study, the sensitivity of US was triple that of palpation for the detection of metastases in supraclavicular lymph nodes (30 vs 10 patients). The sensitivities of US and CT did not differ significantly; this is consistent with results of assessment of supraclavicular metastases in patients with esophageal carcinoma (5). Nevertheless, US should be preferred to CT because of the possibility of combining US and US-guided fine-needle aspiration cytologic analysis in a contiguous procedure. In our experience, this combined procedure takes less than 10 minutes.
There was no correlation between the presence of supraclavicular metastases and advanced T stage at CT in our study. Metastases in supraclavicular lymph nodes do correlate with advanced lymph node metastases and hematogenous metastases at CT and therefore represent a sign of advanced disease. In our study, 28 of 30 patients with metastases in supraclavicular lymph nodes already had N2 or N3 lymph nodes at chest CT. Moreover, 40% of these 30 patients had radiologic signs of distant metastases.
It is therefore not surprising that clinical "upstaging" as a result of findings at US-guided fine-needle aspiration cytologic analysis occurred in only three patients with NSCLC and nonpalpable supraclavicular metastases. However, more invasive diagnostic procedures or unnecessary surgical procedures were prevented in these three patients. In addition, more invasive diagnostic procedures were prevented in eight other patients.
In the past few years, the use of whole-body 18F FDG PET in the preoperative staging of lung cancer has been strongly advocated in the literature (16–18). Because 18F FDG PET is costly and not yet commonly available, we suggest that US and US-guided fine-needle aspiration cytologic analysis should be incorporated in diagnostic protocols before 18F FDG PET.
There were limitations to our study. The reference standard we used for assessing metastasis in supraclavicular lymph nodes is limited because extended surgical lymph node dissection of the supraclavicular regions was not performed. Therefore, the true sensitivities of palpation, US, and CT may be lower than those we reported. However, in our opinion, comparison of the relative sensitivities of these techniques is justified.
Second, bias may have been introduced because fine-needle aspiration cytologic analysis is performed more easily with US guidance than with CT or palpation guidance. This implies that cytologic specimens of lymph nodes that were enlarged only at CT and not at US could not always be obtained.
In conclusion, in patients suspected of having lung cancer, a careful diagnostic approach must be used to prove the presence of a tumor, to determine the cell type, and to establish the stage while minimizing costs, discomfort, and risk to the patient.
US and US-guided fine-needle aspiration cytologic analysis of supraclavicular lymph nodes are simple procedures that helped to establish both cell type and the fact that a cancer was inoperable in 26% of patients with clinical or radiologic signs of lung cancer.
Therefore, we recommend that US and US-guided fine-needle aspiration cytologic analysis be used routinely in the diagnostic work-up of patients suspected of having lung cancer as early as possible and at least before any invasive or expensive procedure is undertaken.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, H.v.O., K.B.; study concepts, H.v.O., K.B.; study design, all authors; literature research, K.B., M.W.H., J.H.L.M.v.K., C.N.F.v.d.M., A.C.R.; clinical studies, H.v.O., K.B., M.W.H., J.H.L.M.v.K., C.N.F.v.d.M., A.C.R., A.P.v.G.; data acquisition, M.W.H., J.H.L.M.v.K., C.N.F.v.d.M., A.C.R., A.P.v.G.; data analysis/interpretation, H.v.O., K.B., M.W.H., B.E.H.; statistical analysis, B.E.H.; manuscript preparation, definition of intellectual content, revision/review, and final version approval, all authors; manuscript editing, H.v.O., K.B.
H.v.O. and K.B. contributed equally to this work.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Soudack, R. B. Shalom, O. Israel, Y. Ben-Arie, Z. Levy, and D. Gaitini Utility of Sonographically Guided Biopsy in Metabolically Suspected Recurrent Lymphoma J. Ultrasound Med., February 1, 2008; 27(2): 225 - 231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. M. Sung, K. S. Lee, B.-T. Kim, S. Kim, O J. Kwon, J. Y. Choi, and S.-O. Yang Nonpalpable Supraclavicular Lymph Nodes in Lung Cancer Patients: Preoperative Characterization with 18F-FDG PET/CT Am. J. Roentgenol., January 1, 2008; 190(1): 246 - 252. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. Choplin, D. J. Conces Jr, and S. D. Teague Invited Commentary RadioGraphics, May 1, 2005; 25(3): 786 - 788. [Full Text] [PDF]
|
|
|
|
 |
|
 M Kumaran, R E Benamore, R Vaidhyanath, S Muller, C J Richards, M D Peake, and J J Entwisle Ultrasound guided cytological aspiration of supraclavicular lymph nodes in patients with suspected lung cancer Thorax, March 1, 2005; 60(3): 229 - 233. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
mm).