Screening for Distant Metastases in Patients With Head and Neck Cancer†
Presented at the International Symposium on Metastases in Head and Neck Cancer, Kiel, Germany, January 15–18, 1998.
Objectives The detection of distant metastases at initial evaluation may alter the selection of therapy in patients with head and neck squamous cell carcinoma (HNSCC). In this study the value of screening for distant metastases is evaluated.
Study Design Retrospective analysis.
Methods The results of screening for distant metastases were retrospectively analyzed in 101 consecutive HNSCC patients with high-risk factors who were scheduled for major surgery. All patients had computed tomography (CT) scan of the thorax, bone scintigraphy, examination of the liver by ultrasound and/or CT scan, and blood tests.
Results Distant metastases were found in 17% of the patients. Patients with four or more clinical lymph node metastases or low jugular lymph node metastases had the highest incidence of distant metastases (33%). CT scan of the thorax detected in 12 patients, lung metastases; in 4, mediastinal lymph node metastases; and in 2, primary lung tumors. Bone scintigraphy detected in four patients bone metastases; in all four patients lung or mediastinal lymph node metastases were also found. Ultrasound and/or CT scan of the liver revealed one patient with metastases. Blood tests did not show any significant difference between patients with or without bone or liver metastases.
Conclusions Screening in patients with three or more lymph node metastases, bilateral lymph node metastases, lymph nodes of 6 cm or larger, low jugular lymph node metastases, locoregional tumor recurrence, and second primary tumors revealed distant metastases in 10% or more. CT scan of the thorax is currently the single most important diagnostic technique for screening of distant metastases.
Approximately one third of the patients with head and neck squamous cell carcinoma (HNSCC) present with early-stage disease and two thirds present with advanced disease. HNSCCs grow locally invasive and have a proclivity to metastasize to regional lymph nodes rather than to spread hematogenously. Distant metastases usually occur late during the course of the disease. As locoregional results of treatment of HNSCC have improved over the last decades as a result of advances in surgery and radiotherapy, more patients are exposed to the risk of developing distant metastases. The overall incidence of clinically detected distant metastases in HNSCC is 4% to 25%. 1–10 Only a limited number of studies have reported on the incidence of distant metastases at presentation. Dennington et al. 3 and Black et al. 4 found distant metastases at presentation in 7% and 12% of patients with advanced-stage disease, respectively.
The lungs, the skeletal system, and the liver are the most frequent sites of distant metastases. 3,4,10–12 Distant spread to other sites is less frequent. The detection of distant metastases at the time of initial evaluation influences the selection of treatment modality. Patients with distant metastases are generally not considered curable and often receive only palliative treatment. No effective systemic treatment for disseminated HNSCC is currently available.
Given the low incidence of distant metastases at presentation, it is mandatory to establish certain guidelines for determining which patients should undergo evaluation for distant metastases. There is a clear correlation between higher nodal stage, number of lymph node metastases, and involvement of the lower neck lymph nodes on the one side and development of distant metastases on the other side. 3,6–8,10 Patients with three or more lymph node metastases have a risk of almost 50% of developing distant metastases. 8 Furthermore, the incidence of distant metastases is significantly higher in patients with residual or recurrent tumors above the clavicles. 2,12 On the other hand, in patients scheduled for major “quality of life impact” surgery (e.g., total glossectomy) and salvage surgery for recurrent tumors, it is important to exclude distant metastases. This might also be warranted in patients with second primary tumors in previous treated areas of the head and neck, in whom locoregional cure is still possible.
Because the lungs, bone, and liver are the most frequent sites of distant metastases, patients at risk for distant metastases and patients in whom exclusion of distant metastases is important often undergo laboratory investigations, computerized tomography (CT) of the lungs, bone scintigraphy, and ultrasound or CT of the liver. The purpose of this study is to determine the value of screening for distant metastases and of the different diagnostic techniques in the management of high-risk head and neck cancer patients who are scheduled to undergo curative major surgery.
MATERIALS AND METHODS
One hundred one consecutive patients (29 women and 72 men) with high-risk factors for distant metastases who were scheduled for major surgery underwent screening for distant metastases between March 1993 and October 1996. The mean age was 57 years and ranged from 31 to 83 years. All patients had clinically advanced-stage (n = 68) or recurrent (n = 33) HNSCC. At the time of screening these patients had 106 tumors; two patients had double primary tumors, and one patient had four simultaneous primary tumors. Primary tumor sites included the oral cavity (n = 34), oropharynx (n = 32), hypopharynx (n = 14), larynx (n = 17), cervical esophagus (n = 4), nose and paranasal sinus (n = 4), and lymph node metastases of unknown primary tumor (n = 1). Clinical (palpation, CT, magnetic resonance imaging [MRI], and/or ultrasound-guided fine-needle aspiration cytology) indications for screening for distant metastases were three or more lymph node metastases (n = 19; 9 of these patients had four or more lymph node metastases), bilateral lymph node metastases (n = 30), lymph nodes of 6 cm or larger (n = 10), low jugular lymph node metastases (n = 12), locoregional tumor recurrence (n = 33), second primary tumors (n = 13) and major “quality of life impact” surgery (i.e., total glossectomy [n = 11]). Some patients had more than one reason for screening for distant metastases.
In all 101 patients a CT scan of the thorax was performed to supplement routine plain chest x-ray films for exclusion of lung metastases and mediastinal lymph node metastases. Spiral CT scans were obtained with a fourth-generation Siemens Somaton Plus (Siemens AG, Erlangen, Germany) after intravenous administration of contrast medium (Ultravist, Schering AG, Berlin, Germany). Contiguous axial scanning planes were used at 10-mm slice thickness without interslice gap. Radiological criteria were for lung metastases, multiple, smooth and mostly peripherally located lesions; for bronchiogenic carcinoma, solitary, speculated, and mostly centrally located lesions; and for mediastinal lymph node metastases, a diameter of more than 10 mm. In peripherally located solitary lesions thoracoscopy was performed for histopathological confirmation. Every patient also underwent bone scintigraphy according the procedure guidelines for bone scintigraphy of the Society of Nuclear Medicine. If not obvious on bone scintigraphy, the diagnosis of bone metastases was made when osteolytic disease on plain x-ray films coincided with the scintigraphic abnormality. 13 When no explanation (e.g., osteolysis, osteoarthritis, degenerative changes) for a scintigraphic abnormality was found on plain x-ray films, a CT or an MRI scan was made. For evaluation of possible liver metastases, all patients underwent ultrasound and/or CT scan of the abdomen. Five of the 92 patients who were examined with ultrasound underwent, in addition, CT scan of the abdomen. Nine other patients underwent only a CT scan of the abdomen.
Laboratory tests for detection of bone or liver metastases included serum calcium, phosphate, alkaline phosphatase, γ-glutamyl transpeptidase, aspartate aminotransferase, alanine aminotransferase, and bilirubin studies. Serum calcium was corrected for low albumin when necessary.
The incidence of distant metastases was 17%. With the exception of patients scheduled for total glossectomy, the incidence of distant metastases varied from 10% to 33%. Patients with four or more clinical lymph node metastases or low jugular lymph node metastases had the highest incidence of distant metastases. The incidence of distant metastases for the different indications is shown in Table I.
Table Table 1.. Results of Screening for Distant Metastases for the Different Indications in 101 Patients With Advanced Head and Neck Cancer.
Computed tomography of the thorax detected lung metastases in 12 patients. In 10 patients multiple smooth peripheral lesion in the lung parenchyma were seen. Three patients, who had a solitary peripheral lesion on CT scan of the thorax, underwent a thoracoscopy with excision of the lesion. In two patients histopathological examination revealed a squamous cell carcinoma. Mediastinal lymph node metastases were detected in four patients. In two patients a primary bronchogenic carcinoma was detected. In one patient a lobectomy was performed. Histopathological examination revealed a bronchiogenic carcinoma. Because a diagnosis could not be made on the basis of the CT scan, thoracoscopy with biopsy of the lesion was performed in two patients. This showed malignancy in one of them. In eight patients CT scan was advised by the radiologist to evaluate findings on chest x-ray. Five of these CT scans showed lung or mediastinal lymph node metastases. Thus chest x-ray did not show 13 of the 18 malignant lesions in the thorax that were found by CT scan.
Bone scintigraphy was abnormal in 48 patients. In two patients bone scintigraphy showed obvious bone metastases. In addition, 87 plain x-ray films of several anatomical regions (the spine, shoulder, arm, hip, leg, ribs, and skull), one CT scan (of the shoulder), and 7 MRI scans (of the spine, skull, and leg) were made, which revealed bone metastases in two other patients.
Ultrasound examination of the liver showed metastases in one patient without any sign of an abdominal tumor as cause of the liver metastases. Because of abnormal ultrasound, five additional CT scans of the abdomen were performed, which did not result in any finding of metastases. These CT scans showed hemangioma, adenoma, and cysts.
The results of the laboratory biochemical blood tests are summarized in Table II. There were no significant differences in biochemical tests between patients with or without bone or liver metastases.
Table Table 2.. Results of Different Blood Tests for Screening for Bone and Liver Metastases.
*Number of patients with elevated levels per number of patients with available values.
†Calcium concentration corrected for low albumin; 5 patients had low calcium.
AP = alkaline phosphatase; GGT = γ-glutamyl transpeptidase; AST = aspartate aminotransferase; ALT = alanine aminotransferase; NS = not significant.
In 17 of the 101 patients screened, distant metastases were found (17%); in total, 16 metastases in the thorax (12 lung and 4 mediastinal lymph node metastases), 4 bone metastases, and 1 liver metastases were found. All four patients with bone metastases also had lung or mediastinal lymph node metastases. The patient with liver metastases had no distant metastases at other sites. In two patients a synchronous primary bronchogenic carcinoma was detected. Patients with distant metastases detected by screening did not undergo surgery but received palliative radiotherapy or chemotherapy for better local control for a better quality of life.
The 17% incidence of distant metastases at presentation found in this study in a selected group of patients with HNSCC is higher than in other studies. 3,4 Selection criteria for screening for distant metastases may have been different in these studies. All indications (risk factors) for screening for distant metastases in this study, except major “quality of life impact” surgery (e.g., total glossectomy) had an incidence of distant metastases of 10% or more. Therefore screening for distant metastases is valuable in HNSCC patients with three or more lymph node metastases, bilateral lymph node metastases, lymph nodes of 6 cm or larger, low jugular lymph node metastases, locoregional tumor recurrence, and second primary tumors.
Computed tomography scan of the thorax detected in 18 patients malignant lesions in the thorax (12 lung metastases, 4 mediastinal lymph node metastases, and 2 primary bronchogenic carcinomas). Thirteen of these 18 malignant lesions in the thorax were not seen on chest x-ray. In this study the sensitivity for the detection of metastases in the thorax by chest x-ray was only 28%, if CT scan of the thorax was used as gold standard. Troell and Terris 9 and Houghton et al. 14 found a sensitivity of only 50% and 21%, respectively, for the detection of lung metastases or second primary tumors in head and neck cancer patients. This low sensitivity may be due in part to the fact that the majority of lung metastases are peripherally situated, where chest x-ray films tend to be less reliable. 4
Besides lung metastases, CT of the thorax can also detect mediastinal lymph node metastases and primary bronchogenic carcinomas. The latter are not infrequent in this population. 15 Houghton et al. 14 and Houghton et al. 16 found a primary bronchogenic carcinoma or lung metastases on CT of the thorax in 9% of 111 and in 15% of 81 patients with HNSCC. Mediastinal lymph node metastases are found in 6 of 546 (3%), 2 of 26 (8%), and 14 of 145 (10%) of the patients with clinically detected distant metastases. 2,8,9 At autopsy even 23% to 34% of the patients with distant metastases had mediastinal lymph node metastases. 7,11 Moreover, because bone metastases are most likely to be found in vertebrae and ribs, CT scan of the thorax may detect metastases at these sites as well. Findings on 1 of 13 CT scans suspected for lung metastases and 1 of 3 suspected for bronchiogenic carcinoma were false-positive as shown by histopathological examination of biopsy specimens taken during thoracoscopy.
Screening for bone metastases by bone scintigraphy revealed four patients with bone metastases. In initial staging using bone scintigraphy, bone metastases were found in 1% to 4% of head and neck cancer patients 17–21
In the present study, 87 additional plain x-ray films in 46 of 101 patients were made, which resulted in findings of four bone metastases in one patient. Eight additional CT and MRI scans in eight patients resulted in one bone metastasis. The two other patients with bone metastases did not require additional x-ray films for confirmation. In 93 patients with advanced-stage head and neck cancer, Ampil et al. 13 found 15 bone scintigrams with increased uptake. Additional plain x-ray films showed no bone metastases. Bone scintigraphy is still the most sensitive technique for detection of bone metastases, but a positive bone scintigram is nonspecific. 20 Plain x-ray films (sometimes in combination with CT and MRI) and open biopsy may be helpful in determining the kind of lesion. 4,21 The high incidence of inconclusive evidence of bone metastases necessitates additional examinations and, consequently, delay, costs, and confusion for patients and physicians. Moreover, there is a low probability of a positive test result and there is a potential risk carried by diagnostic nuclear medicine. Therefore, the performance of bone scintigraphy for metastatic workup in head and neck cancer patients without clinical symptoms seems to be generally unnecessary. 13,17–21
In only one patient were liver metastases found by screening. Liver metastases are often present for some time before causing symptoms or resulting in abnormal liver function tests. Ultrasound, CT, and MRI scan of the liver are reliable techniques for the detection of liver metastases. CT scan with intravenous contrast agents is currently the recommended procedure. 22 Because of the low frequency of liver metastases found by ultrasound or CT scan, screening for liver metastases seems not to be warranted.
Laboratory biochemical tests have been used as a screening method for bone and liver metastases. No difference was found in liver function tests (alkaline phosphatase, γ-glutamyl transpeptidase, aspartate aminotransferase, alanine aminotransferase, and bilirubin studies) and calcium and alkaline phosphatase values in patients with liver or bone metastases. Degardin et al. 23 found that bone metastases were the cause of hypercalcemia in only 43 of 173 head and neck cancer patients (25%). Troell and Terris 9 found that alkaline phosphatase has a sensitivity of only 20% and a specificity of 98% for the detection of bone metastases. In head and neck cancer patients the reliability of liver function tests as screening tests to identify liver metastases is reduced by the high incidence of alcohol abuse. 1,5 Troell and Terris 9 found that biochemical tests for evaluation of liver metastases are not very sensitive and are extremely nonspecific. Korver et al. 24 found that abnormal liver function test results are commonly encountered, but are of little value in identifying patients with liver metastases during initial assessment. Therefore in screening for distant metastases laboratory biochemical tests are not useful.
The high incidence of distant metastases found by screening in this study may also reflect better diagnostic techniques. CT of the thorax is currently the most valuable technique for screening of distant metastases in comparison to other currently available diagnostic techniques. New diagnostic techniques for detection of distant metastases (i.e., thallium-201 single photon emission computed tomography [SPECT], fluorine-18-fluorodeoxyglucose pos-itron emission tomography [FDG PET], and radioimmunoscintigraphy [RIS]) have been developed but need further evaluation for clinical applicability. Recently at the University Hospital “Vrije Universiteit” in Amsterdam a study has been initiated in which the diagnostic value of whole-body PET will be evaluated in screening for distant metastases.
The lungs are the most frequent initial site of distant metastases. Moreover, lung metastases occur often in combination with distant metastases at other sites. In this study, almost all (94%) of the patients with distant metastases had metastases in the thorax (12 lung metastases and 4 mediastinal lymph node metastases), and all bone metastases occurred in combination with lung metastases. When only a CT scan of the thorax was made for screening for distant metastases, the presence of distant metastases was missed in only one patient (1%). Also, liver metastases occur rarely in the absence of other distant metastases, particularly lung metastases. 9 Leemans et al. 8 reported on 26 of 281 patients with distant metastases. Twenty-one of the 22 lung metastases were the first manifestation of distant metastases. Nishijima et al. 12 found that the lungs were not involved at autopsy in only 15% of the patients with distant metastases. In 24%, only the lungs were involved, and in 61%, the lungs were involved in combination with other sites. Because of the high incidence of lung metastases and the frequent combination of distant metastases at other sites with lung metastases, examination of the thorax as single investigation seems to be sufficient in screening for distant metastases. In our institution in patients without high-risk factors, the preoperative chest x-ray film is the sole screening method for distant metastases.
Regarding cost-effectiveness, CT scan of the thorax is the best diagnostic technique. In this study the detection of one patient with lung or mediastinal lymph node metastases cost $1,333; with bone metastases, $9,720; and with liver metastases, $8,858.
Recently, studies have reported on screening of distant metastases using CT scan of the thorax, in which it was concluded that the role of CT scan of the thorax was limited in head and neck cancer patients. However, in these studies, patients were not selected based on high-risk factors and indications for screening of distant metastases were not analyzed. 25,26
Computed tomography scan of the thorax is the single most important diagnostic technique that is currently available for screening of distant metastases in head and neck cancer patients. Patients with three or more clinical lymph node metastases or low jugular lymph node metastases have the highest incidence of distant metastases. Laboratory investigations, bone scintigraphy, and ultrasound or CT scan of the liver are of little value in screening for distant metastases. Therefore our current policy in screening patients with HNSCC who have three or more lymph node metastases, bilateral lymph node metastases, lymph nodes of 6 cm or larger, low jugular lymph node metastases, locoregional tumor recurrence, or second primary tumors is CT scan of the thorax only. The value of FDG PET is currently under investigation.