Left, Left temporal bone of a 52-year-old man who had transitional cell carcinoma of the urinary bladder. Tumor (arrows) has invaded the petrous apex (PA), internal auditory canal (IAC), facial nerve (F), middle ear (ME), and mastoid (M) (hematoxylin-eosin, original magnification ×7). Right, Higher magnification of the same temporal bone showing tumor (T) in the mastoid (M) (hematoxylin-eosin, original magnification ×50).
Left temporal bone of a 53-year-old man who had an intracranial malignant meningioma. Tumor (T) had invaded most of the temporal bone (hematoxylin-eosin, original magnification ×4).
Left, Right temporal bone of a 61-year-old man who had undifferentiated carcinoma of the lungs. Tumor (arrows) involvement was limited to the mastoid (M) (hematoxylin-eosin, original magnification ×7). Right, Higher magnification of the same temporal bone showing solitary tumor (T) involvement of the mastoid within the temporal bone (hematoxylin-eosin, original magnification ×140).
Left temporal bone of a 37-year-old woman who had follicular carcinoma of the thyroid. Tumor (T) invaded the substance of the facial nerve (F), and the patient had left facial paralysis (hematoxylin-eosin, original magnification ×7).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Gloria-Cruz TI, Schachern PA, Paparella MM, Adams GL, Fulton SE. Metastases to Temporal Bones From Primary Nonsystemic Malignant Neoplasms. Arch Otolaryngol Head Neck Surg. 2000;126(2):209–214. doi:10.1001/archotol.126.2.209
To compare histopathological and clinical findings of metastasis to the temporal bone with previous reports and to determine the prevalence of these metastases in patients with nonsystemic cancer.
Autopsy records of 864 patients were screened to select those with primary nondisseminated malignant neoplasms. These were evaluated histopathologically for metastasis to and site of involvement within the temporal bone, and histological characteristics of the tumor. Clinical records and autopsy reports were reviewed for demographic data, clinical course, otologic and vestibular manifestations, site of primary and its histological features, extent of metastasis, and mode of spread.
Of 212 patients with primary nondisseminated malignant neoplasms, 47 had metastases to the temporal bone (76 temporal bones). Twenty different primary tumors had metastasized, most commonly breast cancer. Hearing loss was the most common otologic symptom (seen in 19 patients [40%]), while 17 (36%) had no otologic or vestibular symptoms. Temporal bone involvement was bilateral in 29 patients (62%). Most metastases to the temporal bone demonstrated hematogenous spread in 58 temporal bones (76.7%), and petrous apex was the most common site of metastases in 63 temporal bones (82.9%). Temporal bone metastases were not observed in cases where the primary tumor was adequately treated.
In the largest series to date, we found temporal bone metastases more frequently than previously reported. Absence of temporal bone involvement in cases in which the primary tumor was adequately treated stresses the need for early management of cancer. Metastatic disease must be considered as a cause of hearing loss in patients with a history of malignant neoplasm.
REPORTS OF metastases to the temporal bone seem to have undergone a gradual evolution since those of Proctor and Lindsay,1 Maddox,2 and Schuknecht et al.3 From the 1940s to the present, investigators have described the number of cases, case reports, histopathological findings in the temporal bone, associated signs and symptoms, site of the primary tumor and its histological features, mode of spread, and extent of metastases. Except for increasing numbers of case reports (occasionally unique), Berlinger and coworkers' description of 5 patterns of involvement,4 and the report by Saito et al5 on the pathological features of associated facial paralysis, there seems to be much similarity in findings and conclusions.
In this study, we compare our findings in metastatic tumors from the collection of temporal bones at the University of Minnesota, Minneapolis, with findings in previous reports. This is the largest series examined to date, to our knowledge. Because our review considered the entire collection and selected from among all 212 donors who had cancer, our review is also unique in approximating the rate of prevalence of tumors metastatic to the temporal bone from primary nondisseminated malignant neoplasms.
Autopsy reports of 864 patients from the temporal bone collection at the University of Minnesota were screened to select cases with primary nondisseminating malignant neoplasms (excluding those with systemic malignant diseases such as lymphoma, leukemia, and multiple myeloma). These so-called systemic malignant conditions proliferate initially into the bone marrow or lymph nodes before disseminating to the peripheral blood, spleen, and ultimately other tissues.
Primary nondisseminated malignant neoplasms were found in 212 patients. Of these 212 patients, 415 of their temporal bones were obtained at autopsy, after consent of next of kin was obtained; placed in fixative solution (10% buffered formalin); decalcified with 5% trichloroacetic acid; embedded in celloidin; and cut into sections at a thickness of 20 µm. Every 10th section had been stained routinely with hematoxylin-eosin.
We examined each temporal bone under light microscopy for the presence of tumor cells and noted the site of involvement. Corresponding clinical records and autopsy reports were reviewed to retrieve demographic data and note clinical course, otologic and vestibular manifestations, site of the primary malignant neoplasm and its histological features, the extent of metastasis, and the mode of spreading. The mode of spreading was determined by noting the other organ systems in which the tumor was found at autopsy and evaluating the most likely pathway to the ear, with Berlinger and coworkers'4 patterns of involvement used as a guide. These included (1) isolated metastases from a distant primary tumor, (2) direct extension from a regional primary tumor, (3) direct meningeal carcinomatosis, (4) leptomeningeal extension from an intracranial primary tumor, and (5) leukemic or lymphomatous infiltration. Hematogenous spread was determined by the presence of tumor cells in the bone marrow portions of the temporal bone.
Histopathological evaluation of the 415 temporal bones from 212 patients with primary nondisseminated malignant neoplasms showed that 76 temporal bones (18.3%) of 47 patients (22.2%) had cancer cells characteristic of the primary tumor. These cancer cells presumably developed the capacity for dissemination to other parts of the body, in these cases the temporal bone. The 47 patients with temporal bone metastases had had 20 different primary cancers (Table 1), the most common being breast cancer. The ages of the patients ranged from 2 to 87 years, with most (13 patients) being aged 51 to 60 years. Twenty-five were male, 22 female. From the clinical and autopsy reports, we noted otologic and vestibular manifestations (Table 2), most commonly hearing loss; 17 patients (36.2%) had neither otologic nor vestibular symptoms. We found 7 patients (14.9%) with facial palsy.
Most cases of metastasis to the temporal bone demonstrated hematogenous spread in 57 (75%) of temporal bones (Figure 1), but in no case was the temporal bone the sole site of hematogenous spread. Four cases (adenocarcinomas of the lung, stomach, and prostate and a malignant melanoma of the skin of the back) had both hematogenous and meningeal spread. Two cases (4.2%), an undifferentiated carcinoma of the lung and a rhabdomyosarcoma of the soleus, had meningeal spread alone. All tumors that involved the meninges with subsequent spread to the temporal bone involved the internal auditory canal. Two cases (4.2%) had intracranial spread from primary brain tumors; one was an astrocytoma, involving only the internal auditory canal, and the other was a malignant meningioma that invaded most of the temporal bone (Figure 2). Two cases (4.2%) had direct extension; both were squamous cell carcinomas of the oral cavity and the oropharynx that involved the eustachian tube and the carotid artery and extended into the middle ear and mastoid. One case had both direct extension and meningeal spread; it was a malignant melanoma of the skin over the parotid area that involved the internal auditory canal, petrous apex, facial nerve, carotid artery, otic capsule, vestibule, cochlea, and bony semicircular canals. No case metastasized from a malignant neoplasm in the kidney.
Metastases extended to sites other than the temporal bone (Table 3). The most common sites were lung and gastrointestinal tract. Four of our 40 cases involving the petrous apex had relatively limited metastasis elsewhere. They included an adenocarcinoma of the stomach that involved the marrow of vertebral bone, squamous cell carcinoma of the tonsils that spread to the cervical nodes, an infiltrating ductal carcinoma of the breast that spread to the marrow of vertebral bone and to the pleura, and a hepatocellular carcinoma involving only the adjacent duodenum.
Of the 212 patients with malignant neoplasms, 12 had occult (found at autopsy) primary tumors. These included 6 adenocarcinomas of the prostate, 2 adenocarcinomas of the stomach, 1 adenocarcinoma of the lung, 1 scirrhous carcinoma of the breast, 1 renal cell carcinoma of the kidney, and 1 astrocytoma of the brain. Of the 12 patients with occult primary tumors, only 1 had metastasis to the temporal bone; he had adenocarcinoma of the prostate with metastasis to the marrow of the long bones, found at autopsy, and involvement of the temporal bone limited to the petrous apex. This patient did not exhibit any otologic or vestibular manifestations.
Review of autopsy reports showed that 18 patients had no residual tumor. Primary tumors had been from the breast (1 infiltrating ductal, 1 adenocarcinoma), prostate (4 adenocarcinomas), cervix (1 squamous cell carcinoma), rectum (1 adenocarcinoma), larynx (1 squamous cell carcinoma), tongue (1 squamous cell carcinoma), colon (1 adenocarcinoma), kidney (1 renal cell carcinoma), and arm (1 liposarcoma). These had apparently been treated adequately by surgery, radiotherapy, and chemotherapy, alone or in combination. The interval between treatment and death ranged from 2 months to 18 years. None of these patients had metastasis to the temporal bone.
Table 4 shows areas within the temporal bone to which tumors metastasized, most commonly the petrous apex, which was the only site in 24 bones (31.6%). The mastoid was involved in 21 (27.6%) of temporal bones, by hematogenous spread in most cases, but in 1 case from direct extension of a squamous cell carcinoma of the oral cavity, and in another by intracranial extension of an astrocytoma of the brain. Only 1 mastoid involvement (1.3%, an undifferentiated carcinoma of the lungs) represented the sole metastatic site (Figure 3). The internal auditory canal was involved in 27.6% of temporal bones. It was the only site of metastasis within the temporal bone in 2 cases: an intracranial astrocytoma and an undifferentiated carcinoma of the lungs that involved the brain.
The facial nerve was involved in 15 temporal bones of 14 patients. Of these, 6 patients (42.8%) had facial paralysis; in these 6, the tumor had invaded the substance of the nerve in the internal auditory canal (Figure 4) in 3 patients, in the mastoid segment in 1 patient, and in both in 2 patients. In most cases, invasion of the external ear, middle ear, and carotid artery resulted from hematogenous spread, but there were 2 cases of direct extension from tumors of the oral cavity or the oropharynx. No patient in whom the membranous labyrinth was involved had any vestibular manifestations.
Jahn et al6 reported that, apart from an early attempt by Nylen,7 no study had determined the true incidence of metastases to the temporal bone. Previous reports, such as those of Jahn et al,6 Belal,8 and Nelson and Hinojosa,9 have noted the number of temporal bones with metastases in their series, rather than giving a percentage. Our temporal bone collection represents a patient population selected because they were under the care of an academic institution. Temporal bones are harvested from all available autopsies and are not selected because of specific diseases or conditions. Because we do not process the temporal bones of every patient who dies of cancer, the true prevalence of metastases to the temporal bone will remain undetermined. However, our large series (212 patients with cancer) showed an approximate prevalence of metastasis to the temporal bone of 22.2% of patients (18.3% of their 415 temporal bones).
As in previous reports, our patients ranged in age from 2 to 87 years, most (13 patients) being 51 to 60 years old. As early as the work of Maddox,2 the average age of patients with secondary lesions in the temporal bone has generally corresponded to the age at which most primary cancers have their onset, 55 years in Maddox's study. Schuknecht et al3 reported that 6 of their 10 patients were also in their 50s. Nelson and Hinojosa9 reported similar results: most of their patients were 51 to 70 years old.
We found a preponderance of males (male-female ratio, 25:22), a reversal of the more frequently reported ratio. Schuknecht et al,3 Adams et al,10 Belal,8 and Nelson and Hinojosa9 all reported finding more females than males, corresponding to the high occurrence of breast carcinoma. Jahn et al6 reported having more males than females. In our study with greater numbers of bones, the ratio of male to female patients approached unity.
We found 20 different origins of primary cancers (Table 1), most commonly breast cancer, as in other reports. In contrast to Maddox's findings, where the kidney was the most common source of metastasis to the temporal bone,2 we found no metastasis from the kidney to the temporal bone from our patient with a primary renal malignant neoplasm. We had only 1 renal malignant neoplasm, however, among our 212 patients with primary malignant neoplasms. Adenocarcinoma was the most common histopathological type of tumor found to metastasize to the temporal bone.
Review of our autopsy reports noted 18 patients with no residual tumor; none of these cases showed isolated metastasis to the temporal bone. On the other hand, all temporal bones in which we found metastasis had also been found at autopsy to have coexisting metastasis elsewhere. It therefore seems that if the primary tumor is adequately removed, metastasis to the temporal bone is not likely to occur.
Of our 212 patients with a malignant neoplasm, 12 had occult primary tumors, and only 1 tumor had metastasized to the temporal bone. In this case, metastasis was limited to the petrous apex, and the patient had no otologic or vestibular manifestations. As in all previous reports, hearing loss was the most common manifestation in the ear (Table 2), but we did not have detailed otologic histories for some of our patients. The actual incidence of otologic or vestibular symptoms may have been higher. Nonetheless, in corroboration of reports by Jahn et al6 and by Nelson and Hinojosa9 that, even in advanced cases, metastases to the temporal bone are often asymptomatic, we found that 17 patients (36.2%) had no otologic or vestibular signs or symptoms.
Previously, a high incidence of facial paralysis has been reported, 34% in Maddox's series.2 He thus emphasized otalgia, periauricular swelling, and palsy of the facial nerve as being suggestive of metastasis to the temporal bone. Schuknecht and coworkers3 also reported a high incidence, 50% with facial palsy. We found relatively less, 14.9% with a history of facial palsy.
Of our 15 temporal bones from 14 patients in which the facial nerve was involved, only 6 patients (42.8%) had facial paralysis. This is consistent with findings in previous reports that invasion of the facial canal does not necessarily produce facial paralysis. Saito and coworkers5 found that only 50% of patients with destruction of the facial canal exhibited facial paralysis, but 100% of patients who had tumors invading beyond the epineural sheath had complete paralysis. In our 6 patients who had a history of facial paralysis, the tumor had invaded the substance of the nerve.
The internal auditory canal was involved in 21 (27.6%) of our temporal bones. Tumor invasion of the osseous and membranous labyrinths, however, was infrequent. We can therefore corroborate others' findings that the inner ear seems resistant to invasion by tumors. One interesting finding was that none of our patients with involvement of the membranous labyrinth had any vestibular manifestations. In general, vestibular signs and symptoms were less frequent than otologic, reflecting the resistance of the otic capsule to destruction by tumors.
Most of our metastasis to the temporal bone (58 bones [76.7%]) demonstrated hematogenous spread, the most commonly reported of the 4 modes described by Berlinger and colleagues.4 Its basis was reported as early as 1947, when Proctor and Lindsay1 described the capacity of marrow to filter out tumorous cells from the circulating blood and give rise to metastatic growth. The sluggish blood flow in the sinusoidal capillaries of the marrow presumably favors the deposition of tumor cells in these areas of the temporal bone.
As in all reported series, we found the petrous apex to be the site most commonly affected within the temporal bone (Table 4). In 24 of our temporal bones (31.2%), the petrous apex was the only metastatic site in the temporal bone. In contrast, we found only 1 case (1.3%) of sole metastasis to the mastoid (from the lungs); it seems that, as a rule, tumorous cells that spread hematogenously must first pass the petrous apex before involving the mastoid. Maddox2 and Adams et al10 reported metastases to the pneumatized areas of the mastoid as unusual. We found 21 (27.6%) of our temporal bones to have mastoid involvement, most resulting from hematogenous spread, 1 from direct extension of a squamous cell carcinoma of the oral cavity, and another from intracranial extension of an astrocytoma of the brain.
Nelson and Hinojosa9 reported the presence of diffuse metastasis throughout the body in all cases when the petrous apex was involved hematogenously. We found 4 of 40 cases involving the petrous apex to have relatively limited metastasis, indicating the possibility of a tumorous predilection for the temporal bone that should be studied further in additional cases, but we never found the temporal bone to be the sole site of metastasis for hematogenously spread tumors.
In the largest series to date, our findings support the current belief that temporal bone metastasis is not as uncommon as previously reported. Metastatic disease must be considered as a cause of hearing loss in a patient with a history of malignant neoplasm. The absence of temporal bone involvement in cases where the primary tumor was adequately treated stresses the importance of early diagnosis and management of cancer. Aside from the battery of tests for hearing status, appropriate radiographic studies must be done. All of the patients in Maddox's2 series on whom films were obtained exhibited some radiological abnormality. By far the most common was a defect or osteolytic lesion of the squamous portion. Morton et al11 reported that tumor that manifests early with no bony erosion may be missed on computed tomography and magnetic resonance imaging, being mistaken for mastoiditis. However, some lesions that were missed on computed tomography were well visualized on magnetic resonance imaging. It is expected that, with improvements in imaging abilities, earlier diagnosis of these lesions will be possible.
Metastasis to the temporal bone almost always presents a complex clinical situation with many interacting factors that necessitate individualization of therapy. Foremost of these factors are the nature and behavior of the primary malignant neoplasm, the extent of metastasis in the temporal bone and other organs, the availability of effective therapy, and the patient's general well-being. We must take all of these into consideration in our quest to provide the best treatment for our patients.
Accepted for publication September 15, 1999.
This work was supported in part by grant P50DC03093 from the National Institutes of Health, Bethesda, Md; the International Hearing Foundation, Minneapolis, Minn; and Lions Clubs International, Minneapolis.
Reprints: Patricia A. Schachern, Lions Research Laboratories, Room 206, 2001 Sixth St SE, Minneapolis, MN 55455 (e-mail email@example.com).
Create a personal account or sign in to: