1990, The British Journal of Radiology, 63, 202-205
Nasopharyngeal carcinoma: pattern of skeletal metastases By Jonathan S. T. Sham, MBBS, DMRT, FRCR, *Y. K. Cheung, MBBS, FRCR, *F. L Chan, MBBS, FRCR and D. Choy, MBBS, DMRT, FRCR Department of Radiotherapy and Oncology and *Department of Diagnostic Radiology, Queen Mary Hospital, Hong Kong (Received May 1989 and in revised form August 1989)
Abstract. The records of 153 nasopharyngeal carcinoma patients with skeletal metastases were reviewed. The skeletal system was the most common site of distant metastases. The patients who developed skeletal metastases were significantly younger than those without skeletal metastases, although there was no difference between sexes. The pattern of skeletal involvement conforms to the general pattern, the spine and pelvis being the common sites. The first region of involvement was lumbar spine (28.4%), then dorsal spine (27.7%), sacrum and pelvis (16.3%), femur (9.9%), rib and sternum (7.8%), humerus (5.0%), cervical spine (3.5%) and skull vault (1.4%). Radiologically, the lesions were lytic in 66.0%, mixed lytic and sclerotic in 12.8% and sclerotic in 21.2%. The time to development of symptomatic skeletal metastases of mixed or sclerotic nature was significantly longer than lytic lesions, and patients with mixed or sclerotic metastases also had better survival.
Compared with other head and neck tumours, nasopharyngeal carcinoma has a higher incidence of distant metastases (O'Brien et al, 1971; Probert et al, 1974; Merino et al, 1977). Based on clinical information without autopsy, 14.9-40.5% of distant metastases have been reported for nasopharyngeal carcinoma (Perez et al, 1969; Thompson et al, 1970; Moench et al, 1972; Hoppe et al, 1976; Bedwinek et al, 1980; Mesic et al, 1981), compared with 5.3-23.6% reported for other head and neck cancers (Castigliano et al, 1954; Arons et al, 1961; Hoye et al, 1962; Rubenfeld et al, 1962; Berger & Fletcher, 1971; Probert et al, 1974; Merino et al, 1977). While the other head and neck cancers involved the lungs as the most common site of distant metastases, the most common distant failure site for nasopharyngeal carcinoma was bone (Moench et al, 1972; Hoppe et al, 1976; Merino et al, 1977; Mesic et al, 1981; Sham et al, 1990). Huang (1980) reported axial skeleton and the long bones as the most common sites of involvement, but there has been no detailed description of the nature of bone metastases from nasopharyngeal carcinoma in the English literature. We have reviewed the records of 153 patients with nasopharyngeal carcinoma who developed skeletal metastases with emphasis on the pattern of involvement, radiological features and the impact of skeletal metastases on survival. Patients and methods
From January 1976 to December 1983, 841 patients with newly diagnosed nasopharyngeal carcinoma were referred to our department. Of these, 45 either refused treatment or died before the completion of radiotherapy For correspondence and reprints: Dr Jonathan S. T. Sham, Department of Radiotherapy and Oncology, Queen Mary Hospital, Pokfulam, Hong Kong. 202
from advanced disease; another 37 had distant metastases (Stage V disease) at presentation. These 82 patients were excluded from the present study. Of the remaining 759 Stage I to IV patients, 153 developed skeletal metastases after the completion of treatment, and they formed the basis of the present analysis. Details of patient characteristics, staging procedures and treatment of this group of 759 patients have been reported previously (Sham & Choy, 1990). After completion of radical radiotherapy all patients were followed up every 4-6 weeks for the first year, every 2 months for the second year, every 3 months for the third year and then half-yearly. At each visit the nasopharynx, the neck and the abdomen were carefully examined. Chest radiograph was performed yearly. When skeletal metastasis was suspected because of bone pain, plain radiographs of the appropriate region of skeleton were obtained. When the pain was persistent, or when the clinical suspicion was high but the radiographs were not conclusive, bone scintigraphy was performed. For analysis of the pattern of involvement, the skeletal system was divided into eight regions: cervical spine, lumbar spine, dorsal spine, sacrum and pelvis, femur, humerus, rib and sternum, and skull vault (excluding base of skull involvement by direct extension). A region was said to be involved if there was pain corresponding to radiological evidence of skeletal metastases. Metastases that brought about the destruction of bone were designated as osteolytic and metastases that brought about new bone formation were designated as osteoblastic. When both destruction and proliferation areas were present, the metastases were designated as mixed type. Patients with skeletal metastases shown by bone scintigraphy were excluded from the analysis on the pattern of involvement because the sensitivity of bone scintigraphy often reveals multiple The British Journal of Radiology, March 1990
Nasopharyngeal carcinoma: pattern of skeletal metastases
Months Figure 1. Time to manifestation of skeletal metastases: 1, patients with lytic bone metastases; 2, patients with mixed or sclerotic bone metastases; 3, patients with skeletal metastases diagnosed by bone scintigraphy (p = 0.04 for groups 1 and 2, p = 0.13 for groups 1 and 3, p = 0AS for groups 2 and 3).
involvement before some of the lesions become symptomatic. The time to development of skeletal metastasis, the actuarial risk of developing a symptomatic lesion in another skeletal region and the actuarial survival, starting from the date of diagnosing skeletal metastases, were computed by the Kaplan-Meier method and the difference between curves tested by the Mantel-Cox test (Kaplan & Meier, 1958; Mantel & Haenszel, 1959). Computations were performed by the 1L module of the BMDP Statistical Software. Results
Out of the 759 patients, 276 (36.4%) patients developed distant metastases in one or more organ-systems: 153 patients developed skeletal metastases, 102 developed lung metastases, and 72 developed liver metastases. Of the 153 patients with skeletal metastases, there were 119 males (77.8%) and 34 females (22.2%). Their
age ranged from 18 to 78 years, with median age of 45.9 years. These 153 patients had significantly more advanced stage than the 606 patients who had no diagnosed skeletal metastases (p< 0.0001, by Mann Whitney U-test). There was no significant difference in sex distribution between the 153 patients with skeletal metastases and the 606 patients without. The group of patients with skeletal metastases however was significantly younger than group without (median age of 45.9 years and 49.0 years respectively, p = 0.006, by Student's Mest). Pain was the main presenting symptom in all patients. The diagnosis of bone metastases was by radiographs in 141 patients and bone scintigraphy in the remaining 12 patients. These 12 patients had persistent bone pain but repeated radiographs were normal while bone scintigraphy showed multiple areas of increased uptake in the skeleton, with one or more areas corresponding to the symptomatic sites. For the 141 patients with radiographic evidence of bone metastases, the lesions were lytic in 93 (66.0%), mixed lytic and sclerotic in 18 (12.8%) and sclerotic in 30(21.2%). The time to development of symptomatic skeletal metastases from diagnosis of nasopharyngeal carcinoma ranged from 2 months to 113 months, the median time being 9.5 months. Eighty per cent of skeletal metastases were manifested by 24 months. The time to development of symptomatic skeletal metastases of mixed or sclerotic nature was significantly longer than lytic lesions (p = 0.04, Fig. 1). The first and second skeletal region with symptomatic metastases in the 141 patients with radiological evidence are shown in Table I. The first skeletal region in descending order of frequency was lumbar spine, 40 patients (28.4%); dorsal spine, 39 (27.7%); sacrum and pelvis, 23 (16.3%); femur, 14 (9.9%); rib and sternum, 11 (7.8%); humerus, 7 (5.0%); cervical spine, 5 (3.5%) and skull vault (excluding base of skull involvement by direct extension), 2 (1.4%). The actuarial risk of developing symptomatic metas-
Table I. First and second skeletal region with symptomatic metastases, and the total number of patients who developed symptomatic metastases in each skeletal region Region
Number of patients with this region involved as first site 5 (3.5%) (27.7%) (28.4%) (16.3%) (5.0%) (9.9%) (7.8%) (1.4%)
Cervical spine Dorsal spine Lumbar spine Sacrum-pelvis Humerus Femur Rib-sternum Skull vault
39 40 23 7 14 11 2
Total number of patients
141
Vol. 63, No. 747
Number of patients with this region involved as second site 0 18 29 17 3 7 4 2 80
(0.0%) (12.8%) (20.6%) (12.1%) (2.1%) (5.0%) (2.8%) (1.4%)
Total number of patients with this region involved in the course of disease 7 67 80 56 13 32 27 5
(5.0%) (47.5%) (56.7%) (39.7%) (9.2%) (22.7%) (19.1%) (3.5%)
141
203
J. S. T. Sham, Y. K. Cheung, F. L. Chan and D. Choy
0.9
o.a ,54 0.7
C? 0.6
J°* u •^5 0.3
\2
0.2 0.1 0
Months Figure 2. Actuarial risk of developing symptomatic lesion in another skeletal region (for patients with lytic, mixed or sclerotic lesions).
tases in another region is shown in Fig. 2. Patients who died before the development of symptomatic lesion in another skeletal region were considered censored. Twenty-four patients (17%) had simultaneous symptomatic lesions in more than one region at the first documentation of skeletal metastases. By 5.5 months, 50% of patients had developed symptomatic lesions in another skeletal region. There was no statistical difference in this aspect of metastases between patients with different radiological patterns of involvement (/> = 0.42). More than half the patients developed symptomatic metastatic lesions in two or more skeletal regions before death or the last follow-up (Table II). The total number of patients who developed symptomatic lesions in each skeletal region is shown in Table I. The actuarial survival starting from the date of diagnosis of skeletal metastases is shown in Fig. 3. Patients with mixed or sclerotic lesions had significantly better survival compared with those with lytic lesions and those with skeletal involvement documented by bone scintigraphy (p = 0.03 and 0.005 respectively). The median survival after documentation of skeletal metastases were 9 months, 5 months and 5 months respectively. Discussion This study confirmed the reported experience (Moench et al, 1972; Hoppe et al, 1976; Merino et al, Table II. Total number of skeletal regions with symptomatic metastases Number of skeletal regions
Number of patients
1 2 3 4 5 >5
61 (43.3%) 41 (29.1%) 19 (13.5%) 14 (9.9%) 4 (2.8%) 2 (1.4%)
204
10
20
30
40
BO
60
70
SO
BO
100
Months Figure 3. Actuarial survival after diagnosis of skeletal metastases: 1, patients with lytic bone metastases; 2, patients with mixed or sclerotic bone metastases; 3, patients with skeletal metastases diagnosed by bone scintigraphy (p = 0.03 for groups 1 and 2, p = 0.32 for groups 1 and 3, /> = 0.005 for groups 2 and 3).
1977; Mesic et al, 1981) that for nasopharyngeal carcinoma the skeleton is the most common site of distant involvement. The common sites of metastatic skeletal involvement by nasopharyngeal carcinoma conform to that described by Huang (1980) and the general pattern as described for other metastatic tumours, namely the spine, the pelvis, the ribs and the humerus and femur (Tofe et al, 1975; Krishnamurthy et al, 1977; Murray & Jacobson, 1977; Murray & Watt, 1987). This pattern of relative frequency was maintained in the second region of involvement, as well as the total number of patients with symptomatic lesions in each region. The axial skeleton and the proximal long bones, which contain red marrow, are expected to be preferentially affected as most skeletal metastases are haematogenous in origin, whether by retrograde venous flow or via the general arterial circulation after venous or lymphatic access (Tofe et al, 1975). The access of the nasopharynx to the prevertebral venous plexus (Batson, 1942) does not make the distribution of skeletal metastases different from other tumours with no access to the prevertebra! venous plexus, but perhaps accounts for the higher incidence of skeletal metastases compared with other head and neck tumours. While the majority of bony metastases are osteolytic in general, sclerotic metastases are well known for prostatic and breast cancers. Other primary sites that are well recognized as being associated with sclerotic metastases include lung (Napoli et al, 1973), pancreas (Joffe & Antonioli, 1978), carcinoid (Peavy et al, 1973), urinary bladder (Evison et al, 1981), melanoma (Fon et al, 1981), colon (Seife, 1973) and soft tissue sarcoma (Wong et al, 1982). Although they can be associated with any metastatic cancer, the high frequency of mixed and sclerotic metastases in the present series again The British Journal of Radiology, March 1990
Nasopharyngeal carcinoma: pattern of skeletal metastases makes nasopharyngeal carcinoma unusual among head and neck tumours. Another interesting finding in the present study is the significantly better prognosis of patients with mixed or sclerotic bony metastases compared with patients with lytic bony metastases. Although it has been a common belief that osteoblastic metastases tend to have a slower growth rate compared with osteolytic metastases, it has not been well substantiated (Greenfield, 1986). In their study of bone metastases from breast cancer, Barry et al (1981) showed that a lytic pattern correlated with progressive tumour growth, whereas blastic pattern correlated with clinical remission. The present study adds to the literature by showing that mixed or sclerotic bony metastases of nasopharyngeal carcinoma presented later compared with lytic lesions and also that patients with mixed or sclerotic metastases had significantly better survival after the diagnosis of skeletal involvement. Acknowledgments This study was supported by grants from the Hong Kong Anti-Cancer Society and the Asia Oceanean Clinical Oncology Association. References
JOFFE, N. & ANTONIOLI, D. A., 1978. Osteoblastic bone
metastases secondary to adenocarcinoma of the pancreas. Clinical Radiology, 29, 4 1 ^ 6 . KRISHNAMURTHY, G. T., TUBIS, M., HISS, J. & BLAHD, W. H.,
1977. Distribution pattern of metastatic bone disease: a need for total body skeletal image. Journal of the American Medical Association, 237, 2504-2506. MERINO, O. R., LINDBERG, R. D. & FLETCHER, G. H., 1977. An
analysis of distant metastases from squamous cell carcinoma of the upper respiratory and digestive tracts. Cancer, 40, 145-151. MESIC,
J.
B., FLETCHER,
G.
H.
& GOEPFERT,
H., 1981.
Megavoltage irradiation of epithelial tumors of the nasopharynx. International Journal of Radiation Oncology, Biology, Physics, 7, 447-453. MOENCH, H. C. & PHILLIPS, T. L., 1972. Carcinoma of the
nasopharynx, review of 146 patients with emphasis on radiation dose and time factor. American Journal of Surgery, 124, 515-518. MURRAY,
R.
O.
&
JACOBSON,
H.
G.,
1977.
Malignant
neoplasms. In The Radiology of Skeletal Disorder, Exercises in diagnosis, Vol. I, second edn. Ed. by R. O. Murray and H. G. Jacobson (Churchill Livingstone, London), pp.581-649. MURRAY, R. O. & WATT, I., 1987. Tumours and tumour-like
conditions of bone. In Textbook of Radiology and Medical Imaging, Vol. I, fourth edn. Ed. by D. Sutton (Churchill Livingstone, London), pp. 136-196.
ARONS, M. S. & SMITH, R. R., 1961. Distant metastasis and
NAPOLI, L. D., HANSEN, H. H., MUGGIA, F. M. & TWIGG, H.
local recurrence in head and neck cancer. Annual of Surgery, 154, 235-240.
L., 1973. The incidence of osseous involvement in lung cancer, with special reference to the development of osteoblastic changes. Radiology, 108, 17-21.
BARRY, W. F. JR, WELLS, S. A. J R , COX, C. E. & HAAGENSEN,
D. E. JR, 1981. Clinical and radiographic correlations in breast cancer patients with osseous metastases. Skeletal Radiology, 6, 27-32. BATSON, O. V., 1942. Veins of pharynx. Archives of Otolaryngology, 36, 212-219. BEDWINEK, J. M., PEREZ, C. A. & KEYS, D. J., 1980. Analysis of
failures after definitive irradiation of epidermoid carcinoma of the nasopharynx. Cancer, 45, 2725-2729. BERGER, D. S. & FLETCHER, G. H., 1971. Distant metastasis
following local control of squamous cell carcinoma of the nasopharynx, tonsillar fossa, and base of the tongue. Radiology, 100, 141-143.
O'BRIEN, P. H., CARLSON, R., STEUBNER, E. A. & STALEY,
C. T., 1971. Distant metastasis in epidermoid cell carcinoma of the head and neck. Cancer, 27, 304-307. PEAVY, P. W., ROGERS, J. V. JR, CLEMENTS, J. L. JR & BURNS, J.
B., 1973. Unusual osteoblastic metastases from carcinoid tumors. Radiology, 107, 327-330. PEREZ, C. A., ACKERMAN, L. V., MILL, W. B., OGURA, J. H. &
POWERS, W. E., 1969. Cancer of the nasopharynx, factors influenceing prognosis. Cancer, 24, 1-17. PROBERT, J. C , THOMPSON, R. W. & BAGSHAW, M. A., 1974.
Patterns of spread of distant metastasis in head and neck cancer. Cancer, 33, 127-133.
Distant
RUBENFELD, S., KAPLAN, G. & HOLDER, A. A., 1962. Distant
metastasis from carcinoma of the oral cavity. American Journal of Roentgenology, 72, 997-1006.
metastasis from head and neck cancer. American Journal of Roentgenology, 87, 441-448. SEIFE, B., 1973. Osseous metastases from carcinoma of the large bowel. American Journal of Roentgenology, 119, 414-418.
CASTIGLIANO,
S.
G.
&
ROMINGER,
C.
J.,
1954.
EVISON, G., PIZZY, N. & ROYLANCE, J., 1981. Bone formation
associated with osseous metastases from bladder carcinoma. Clinical Radiology, 32, 303-309. FON, G. T., WONG, W. S., GOLD, R. H. & KAISER, L. R., 1981.
SHAM,
J.
S.
T.,
CHOY,
D.
&
CHOI,
P.
H.
K.,
1990.
Nasopharyngeal carcinoma: the significance of neck node involvement in relation to the pattern of distant failure. British Journal of Radiology, 63, 108-113.
Skeletal metastases of melanoma: radiographic, scintigraphic, and clinical review. American Journal of SHAM, J. S. T. & CHOY, D., 1990. Prognostic factors of Roentgenology, 137, 103-108. nasopharyngeal carcinoma: a review of 759 patients. British GREENFIELD, G. B., 1986. Cardinal roentgen features. In Journal of Radiology, 63, 51-58. Radiology of Bone Disease, 4th edn by G. B. Greenfield (J. B. THOMPSON, R. W., DOGGET, R. L. S. & BAGSHAW, M. A. A., Lippincott, Philadelphia), pp. 393-552. 1970. Ten-year experience with linear accelerator irradiation HOPPE, R. T., GOFFINET, D. R. & BAGSHAW, M. A., 1976. of cancer of the nasopharynx. Radiology, 97, 149-155. Carcinoma of the nasopharynx, eighteen years' experience TOFE, A. J., FRANCIS, M. D. & HARVEY, W. J., 1975. with megavoltage radiation therapy. Cancer, 37, 2605-2612. Correlation of neoplasms with incidence and localization of HOYE, R. C , HERROLD, K. M., SMITH, R. R. & THOMAS, L. B., skeletal metastases: an analysis of 1355 diphosphonate bone 1962. A clinicopathological study of epidermoid carcinoma scans. Journal of Nuclear Medicine, 16, 986-989. of the head and neck. Cancer, 15, 741-749. HUANG, S. C , 1980. Nasopharyngeal cancer: a review of 1605 WONG, W. S., KAISER, L. R., GOLD, R. H. & FON, G. T., 1982. Radiographic features of osseous metastases of soft tissue patients treated radically with cobalt 60. International sarcoma. Radiology, 143, 1X-1A. Journal of Radiation Oncology, Biology, Physics, 6, 401-^407.
Vol. 63, No. 747
205
Nasopharyngeal carcinoma: pattern of skeletal metastases By Jonathan S. T. Sham, MBBS, DMRT, FRCR, *Y. K. Cheung, MBBS, FRCR, *F. L Chan, MBBS, FRCR and D. Choy, MBBS, DMRT, FRCR Department of Radiotherapy and Oncology and *Department of Diagnostic Radiology, Queen Mary Hospital, Hong Kong (Received May 1989 and in revised form August 1989)
Abstract. The records of 153 nasopharyngeal carcinoma patients with skeletal metastases were reviewed. The skeletal system was the most common site of distant metastases. The patients who developed skeletal metastases were significantly younger than those without skeletal metastases, although there was no difference between sexes. The pattern of skeletal involvement conforms to the general pattern, the spine and pelvis being the common sites. The first region of involvement was lumbar spine (28.4%), then dorsal spine (27.7%), sacrum and pelvis (16.3%), femur (9.9%), rib and sternum (7.8%), humerus (5.0%), cervical spine (3.5%) and skull vault (1.4%). Radiologically, the lesions were lytic in 66.0%, mixed lytic and sclerotic in 12.8% and sclerotic in 21.2%. The time to development of symptomatic skeletal metastases of mixed or sclerotic nature was significantly longer than lytic lesions, and patients with mixed or sclerotic metastases also had better survival.
Compared with other head and neck tumours, nasopharyngeal carcinoma has a higher incidence of distant metastases (O'Brien et al, 1971; Probert et al, 1974; Merino et al, 1977). Based on clinical information without autopsy, 14.9-40.5% of distant metastases have been reported for nasopharyngeal carcinoma (Perez et al, 1969; Thompson et al, 1970; Moench et al, 1972; Hoppe et al, 1976; Bedwinek et al, 1980; Mesic et al, 1981), compared with 5.3-23.6% reported for other head and neck cancers (Castigliano et al, 1954; Arons et al, 1961; Hoye et al, 1962; Rubenfeld et al, 1962; Berger & Fletcher, 1971; Probert et al, 1974; Merino et al, 1977). While the other head and neck cancers involved the lungs as the most common site of distant metastases, the most common distant failure site for nasopharyngeal carcinoma was bone (Moench et al, 1972; Hoppe et al, 1976; Merino et al, 1977; Mesic et al, 1981; Sham et al, 1990). Huang (1980) reported axial skeleton and the long bones as the most common sites of involvement, but there has been no detailed description of the nature of bone metastases from nasopharyngeal carcinoma in the English literature. We have reviewed the records of 153 patients with nasopharyngeal carcinoma who developed skeletal metastases with emphasis on the pattern of involvement, radiological features and the impact of skeletal metastases on survival. Patients and methods
From January 1976 to December 1983, 841 patients with newly diagnosed nasopharyngeal carcinoma were referred to our department. Of these, 45 either refused treatment or died before the completion of radiotherapy For correspondence and reprints: Dr Jonathan S. T. Sham, Department of Radiotherapy and Oncology, Queen Mary Hospital, Pokfulam, Hong Kong. 202
from advanced disease; another 37 had distant metastases (Stage V disease) at presentation. These 82 patients were excluded from the present study. Of the remaining 759 Stage I to IV patients, 153 developed skeletal metastases after the completion of treatment, and they formed the basis of the present analysis. Details of patient characteristics, staging procedures and treatment of this group of 759 patients have been reported previously (Sham & Choy, 1990). After completion of radical radiotherapy all patients were followed up every 4-6 weeks for the first year, every 2 months for the second year, every 3 months for the third year and then half-yearly. At each visit the nasopharynx, the neck and the abdomen were carefully examined. Chest radiograph was performed yearly. When skeletal metastasis was suspected because of bone pain, plain radiographs of the appropriate region of skeleton were obtained. When the pain was persistent, or when the clinical suspicion was high but the radiographs were not conclusive, bone scintigraphy was performed. For analysis of the pattern of involvement, the skeletal system was divided into eight regions: cervical spine, lumbar spine, dorsal spine, sacrum and pelvis, femur, humerus, rib and sternum, and skull vault (excluding base of skull involvement by direct extension). A region was said to be involved if there was pain corresponding to radiological evidence of skeletal metastases. Metastases that brought about the destruction of bone were designated as osteolytic and metastases that brought about new bone formation were designated as osteoblastic. When both destruction and proliferation areas were present, the metastases were designated as mixed type. Patients with skeletal metastases shown by bone scintigraphy were excluded from the analysis on the pattern of involvement because the sensitivity of bone scintigraphy often reveals multiple The British Journal of Radiology, March 1990
Nasopharyngeal carcinoma: pattern of skeletal metastases
Months Figure 1. Time to manifestation of skeletal metastases: 1, patients with lytic bone metastases; 2, patients with mixed or sclerotic bone metastases; 3, patients with skeletal metastases diagnosed by bone scintigraphy (p = 0.04 for groups 1 and 2, p = 0.13 for groups 1 and 3, p = 0AS for groups 2 and 3).
involvement before some of the lesions become symptomatic. The time to development of skeletal metastasis, the actuarial risk of developing a symptomatic lesion in another skeletal region and the actuarial survival, starting from the date of diagnosing skeletal metastases, were computed by the Kaplan-Meier method and the difference between curves tested by the Mantel-Cox test (Kaplan & Meier, 1958; Mantel & Haenszel, 1959). Computations were performed by the 1L module of the BMDP Statistical Software. Results
Out of the 759 patients, 276 (36.4%) patients developed distant metastases in one or more organ-systems: 153 patients developed skeletal metastases, 102 developed lung metastases, and 72 developed liver metastases. Of the 153 patients with skeletal metastases, there were 119 males (77.8%) and 34 females (22.2%). Their
age ranged from 18 to 78 years, with median age of 45.9 years. These 153 patients had significantly more advanced stage than the 606 patients who had no diagnosed skeletal metastases (p< 0.0001, by Mann Whitney U-test). There was no significant difference in sex distribution between the 153 patients with skeletal metastases and the 606 patients without. The group of patients with skeletal metastases however was significantly younger than group without (median age of 45.9 years and 49.0 years respectively, p = 0.006, by Student's Mest). Pain was the main presenting symptom in all patients. The diagnosis of bone metastases was by radiographs in 141 patients and bone scintigraphy in the remaining 12 patients. These 12 patients had persistent bone pain but repeated radiographs were normal while bone scintigraphy showed multiple areas of increased uptake in the skeleton, with one or more areas corresponding to the symptomatic sites. For the 141 patients with radiographic evidence of bone metastases, the lesions were lytic in 93 (66.0%), mixed lytic and sclerotic in 18 (12.8%) and sclerotic in 30(21.2%). The time to development of symptomatic skeletal metastases from diagnosis of nasopharyngeal carcinoma ranged from 2 months to 113 months, the median time being 9.5 months. Eighty per cent of skeletal metastases were manifested by 24 months. The time to development of symptomatic skeletal metastases of mixed or sclerotic nature was significantly longer than lytic lesions (p = 0.04, Fig. 1). The first and second skeletal region with symptomatic metastases in the 141 patients with radiological evidence are shown in Table I. The first skeletal region in descending order of frequency was lumbar spine, 40 patients (28.4%); dorsal spine, 39 (27.7%); sacrum and pelvis, 23 (16.3%); femur, 14 (9.9%); rib and sternum, 11 (7.8%); humerus, 7 (5.0%); cervical spine, 5 (3.5%) and skull vault (excluding base of skull involvement by direct extension), 2 (1.4%). The actuarial risk of developing symptomatic metas-
Table I. First and second skeletal region with symptomatic metastases, and the total number of patients who developed symptomatic metastases in each skeletal region Region
Number of patients with this region involved as first site 5 (3.5%) (27.7%) (28.4%) (16.3%) (5.0%) (9.9%) (7.8%) (1.4%)
Cervical spine Dorsal spine Lumbar spine Sacrum-pelvis Humerus Femur Rib-sternum Skull vault
39 40 23 7 14 11 2
Total number of patients
141
Vol. 63, No. 747
Number of patients with this region involved as second site 0 18 29 17 3 7 4 2 80
(0.0%) (12.8%) (20.6%) (12.1%) (2.1%) (5.0%) (2.8%) (1.4%)
Total number of patients with this region involved in the course of disease 7 67 80 56 13 32 27 5
(5.0%) (47.5%) (56.7%) (39.7%) (9.2%) (22.7%) (19.1%) (3.5%)
141
203
J. S. T. Sham, Y. K. Cheung, F. L. Chan and D. Choy
0.9
o.a ,54 0.7
C? 0.6
J°* u •^5 0.3
\2
0.2 0.1 0
Months Figure 2. Actuarial risk of developing symptomatic lesion in another skeletal region (for patients with lytic, mixed or sclerotic lesions).
tases in another region is shown in Fig. 2. Patients who died before the development of symptomatic lesion in another skeletal region were considered censored. Twenty-four patients (17%) had simultaneous symptomatic lesions in more than one region at the first documentation of skeletal metastases. By 5.5 months, 50% of patients had developed symptomatic lesions in another skeletal region. There was no statistical difference in this aspect of metastases between patients with different radiological patterns of involvement (/> = 0.42). More than half the patients developed symptomatic metastatic lesions in two or more skeletal regions before death or the last follow-up (Table II). The total number of patients who developed symptomatic lesions in each skeletal region is shown in Table I. The actuarial survival starting from the date of diagnosis of skeletal metastases is shown in Fig. 3. Patients with mixed or sclerotic lesions had significantly better survival compared with those with lytic lesions and those with skeletal involvement documented by bone scintigraphy (p = 0.03 and 0.005 respectively). The median survival after documentation of skeletal metastases were 9 months, 5 months and 5 months respectively. Discussion This study confirmed the reported experience (Moench et al, 1972; Hoppe et al, 1976; Merino et al, Table II. Total number of skeletal regions with symptomatic metastases Number of skeletal regions
Number of patients
1 2 3 4 5 >5
61 (43.3%) 41 (29.1%) 19 (13.5%) 14 (9.9%) 4 (2.8%) 2 (1.4%)
204
10
20
30
40
BO
60
70
SO
BO
100
Months Figure 3. Actuarial survival after diagnosis of skeletal metastases: 1, patients with lytic bone metastases; 2, patients with mixed or sclerotic bone metastases; 3, patients with skeletal metastases diagnosed by bone scintigraphy (p = 0.03 for groups 1 and 2, p = 0.32 for groups 1 and 3, /> = 0.005 for groups 2 and 3).
1977; Mesic et al, 1981) that for nasopharyngeal carcinoma the skeleton is the most common site of distant involvement. The common sites of metastatic skeletal involvement by nasopharyngeal carcinoma conform to that described by Huang (1980) and the general pattern as described for other metastatic tumours, namely the spine, the pelvis, the ribs and the humerus and femur (Tofe et al, 1975; Krishnamurthy et al, 1977; Murray & Jacobson, 1977; Murray & Watt, 1987). This pattern of relative frequency was maintained in the second region of involvement, as well as the total number of patients with symptomatic lesions in each region. The axial skeleton and the proximal long bones, which contain red marrow, are expected to be preferentially affected as most skeletal metastases are haematogenous in origin, whether by retrograde venous flow or via the general arterial circulation after venous or lymphatic access (Tofe et al, 1975). The access of the nasopharynx to the prevertebral venous plexus (Batson, 1942) does not make the distribution of skeletal metastases different from other tumours with no access to the prevertebra! venous plexus, but perhaps accounts for the higher incidence of skeletal metastases compared with other head and neck tumours. While the majority of bony metastases are osteolytic in general, sclerotic metastases are well known for prostatic and breast cancers. Other primary sites that are well recognized as being associated with sclerotic metastases include lung (Napoli et al, 1973), pancreas (Joffe & Antonioli, 1978), carcinoid (Peavy et al, 1973), urinary bladder (Evison et al, 1981), melanoma (Fon et al, 1981), colon (Seife, 1973) and soft tissue sarcoma (Wong et al, 1982). Although they can be associated with any metastatic cancer, the high frequency of mixed and sclerotic metastases in the present series again The British Journal of Radiology, March 1990
Nasopharyngeal carcinoma: pattern of skeletal metastases makes nasopharyngeal carcinoma unusual among head and neck tumours. Another interesting finding in the present study is the significantly better prognosis of patients with mixed or sclerotic bony metastases compared with patients with lytic bony metastases. Although it has been a common belief that osteoblastic metastases tend to have a slower growth rate compared with osteolytic metastases, it has not been well substantiated (Greenfield, 1986). In their study of bone metastases from breast cancer, Barry et al (1981) showed that a lytic pattern correlated with progressive tumour growth, whereas blastic pattern correlated with clinical remission. The present study adds to the literature by showing that mixed or sclerotic bony metastases of nasopharyngeal carcinoma presented later compared with lytic lesions and also that patients with mixed or sclerotic metastases had significantly better survival after the diagnosis of skeletal involvement. Acknowledgments This study was supported by grants from the Hong Kong Anti-Cancer Society and the Asia Oceanean Clinical Oncology Association. References
JOFFE, N. & ANTONIOLI, D. A., 1978. Osteoblastic bone
metastases secondary to adenocarcinoma of the pancreas. Clinical Radiology, 29, 4 1 ^ 6 . KRISHNAMURTHY, G. T., TUBIS, M., HISS, J. & BLAHD, W. H.,
1977. Distribution pattern of metastatic bone disease: a need for total body skeletal image. Journal of the American Medical Association, 237, 2504-2506. MERINO, O. R., LINDBERG, R. D. & FLETCHER, G. H., 1977. An
analysis of distant metastases from squamous cell carcinoma of the upper respiratory and digestive tracts. Cancer, 40, 145-151. MESIC,
J.
B., FLETCHER,
G.
H.
& GOEPFERT,
H., 1981.
Megavoltage irradiation of epithelial tumors of the nasopharynx. International Journal of Radiation Oncology, Biology, Physics, 7, 447-453. MOENCH, H. C. & PHILLIPS, T. L., 1972. Carcinoma of the
nasopharynx, review of 146 patients with emphasis on radiation dose and time factor. American Journal of Surgery, 124, 515-518. MURRAY,
R.
O.
&
JACOBSON,
H.
G.,
1977.
Malignant
neoplasms. In The Radiology of Skeletal Disorder, Exercises in diagnosis, Vol. I, second edn. Ed. by R. O. Murray and H. G. Jacobson (Churchill Livingstone, London), pp.581-649. MURRAY, R. O. & WATT, I., 1987. Tumours and tumour-like
conditions of bone. In Textbook of Radiology and Medical Imaging, Vol. I, fourth edn. Ed. by D. Sutton (Churchill Livingstone, London), pp. 136-196.
ARONS, M. S. & SMITH, R. R., 1961. Distant metastasis and
NAPOLI, L. D., HANSEN, H. H., MUGGIA, F. M. & TWIGG, H.
local recurrence in head and neck cancer. Annual of Surgery, 154, 235-240.
L., 1973. The incidence of osseous involvement in lung cancer, with special reference to the development of osteoblastic changes. Radiology, 108, 17-21.
BARRY, W. F. JR, WELLS, S. A. J R , COX, C. E. & HAAGENSEN,
D. E. JR, 1981. Clinical and radiographic correlations in breast cancer patients with osseous metastases. Skeletal Radiology, 6, 27-32. BATSON, O. V., 1942. Veins of pharynx. Archives of Otolaryngology, 36, 212-219. BEDWINEK, J. M., PEREZ, C. A. & KEYS, D. J., 1980. Analysis of
failures after definitive irradiation of epidermoid carcinoma of the nasopharynx. Cancer, 45, 2725-2729. BERGER, D. S. & FLETCHER, G. H., 1971. Distant metastasis
following local control of squamous cell carcinoma of the nasopharynx, tonsillar fossa, and base of the tongue. Radiology, 100, 141-143.
O'BRIEN, P. H., CARLSON, R., STEUBNER, E. A. & STALEY,
C. T., 1971. Distant metastasis in epidermoid cell carcinoma of the head and neck. Cancer, 27, 304-307. PEAVY, P. W., ROGERS, J. V. JR, CLEMENTS, J. L. JR & BURNS, J.
B., 1973. Unusual osteoblastic metastases from carcinoid tumors. Radiology, 107, 327-330. PEREZ, C. A., ACKERMAN, L. V., MILL, W. B., OGURA, J. H. &
POWERS, W. E., 1969. Cancer of the nasopharynx, factors influenceing prognosis. Cancer, 24, 1-17. PROBERT, J. C , THOMPSON, R. W. & BAGSHAW, M. A., 1974.
Patterns of spread of distant metastasis in head and neck cancer. Cancer, 33, 127-133.
Distant
RUBENFELD, S., KAPLAN, G. & HOLDER, A. A., 1962. Distant
metastasis from carcinoma of the oral cavity. American Journal of Roentgenology, 72, 997-1006.
metastasis from head and neck cancer. American Journal of Roentgenology, 87, 441-448. SEIFE, B., 1973. Osseous metastases from carcinoma of the large bowel. American Journal of Roentgenology, 119, 414-418.
CASTIGLIANO,
S.
G.
&
ROMINGER,
C.
J.,
1954.
EVISON, G., PIZZY, N. & ROYLANCE, J., 1981. Bone formation
associated with osseous metastases from bladder carcinoma. Clinical Radiology, 32, 303-309. FON, G. T., WONG, W. S., GOLD, R. H. & KAISER, L. R., 1981.
SHAM,
J.
S.
T.,
CHOY,
D.
&
CHOI,
P.
H.
K.,
1990.
Nasopharyngeal carcinoma: the significance of neck node involvement in relation to the pattern of distant failure. British Journal of Radiology, 63, 108-113.
Skeletal metastases of melanoma: radiographic, scintigraphic, and clinical review. American Journal of SHAM, J. S. T. & CHOY, D., 1990. Prognostic factors of Roentgenology, 137, 103-108. nasopharyngeal carcinoma: a review of 759 patients. British GREENFIELD, G. B., 1986. Cardinal roentgen features. In Journal of Radiology, 63, 51-58. Radiology of Bone Disease, 4th edn by G. B. Greenfield (J. B. THOMPSON, R. W., DOGGET, R. L. S. & BAGSHAW, M. A. A., Lippincott, Philadelphia), pp. 393-552. 1970. Ten-year experience with linear accelerator irradiation HOPPE, R. T., GOFFINET, D. R. & BAGSHAW, M. A., 1976. of cancer of the nasopharynx. Radiology, 97, 149-155. Carcinoma of the nasopharynx, eighteen years' experience TOFE, A. J., FRANCIS, M. D. & HARVEY, W. J., 1975. with megavoltage radiation therapy. Cancer, 37, 2605-2612. Correlation of neoplasms with incidence and localization of HOYE, R. C , HERROLD, K. M., SMITH, R. R. & THOMAS, L. B., skeletal metastases: an analysis of 1355 diphosphonate bone 1962. A clinicopathological study of epidermoid carcinoma scans. Journal of Nuclear Medicine, 16, 986-989. of the head and neck. Cancer, 15, 741-749. HUANG, S. C , 1980. Nasopharyngeal cancer: a review of 1605 WONG, W. S., KAISER, L. R., GOLD, R. H. & FON, G. T., 1982. Radiographic features of osseous metastases of soft tissue patients treated radically with cobalt 60. International sarcoma. Radiology, 143, 1X-1A. Journal of Radiation Oncology, Biology, Physics, 6, 401-^407.
Vol. 63, No. 747
205
