I F A brain tumor recurs after a primary

Transcription

1 The Value of Postoperative Brain Scan in Patients with Supratentorial Intracranial Tumors* ROBERT H. WILKIN$, M.D., FELIX J. PIRCHER, 5I.D., AND GUY L. ODOM, M.D. Division of Neurosurgery and the Department of Radiology, Duke University Medical Center, Durham, North Carolina I F A brain tumor recurs after a primary resection, as most gliomas do, the recurrence probably has started before the patient leaves the operating room. During the ensuing months and years, the recurrent neoplasm grows and eventually produces seizures, headaches, or other symptoms and signs that indicate its presence. However, some of the same symptoms and signs can be produced by postoperative gliosis, fibrosis, hydrocephalus, and vascular insufficiency, and the diagnosis of significant tumor recurrence may be difficult to establish by history and physical examination in the early stages. 1' Because of these processes and the mechanical distortions produced by the operation, electroencephalography, echoencephalography, and plain roentgenography may also be of little value. Until recently, the clinician evaluating the initial evidences of recurrence has been faced with the choice of waiting for further manifestations or of subjecting his patient to the expense and inconvenience of hospitalization and to the discomfort and risks of angiography, pneumoencephalography, or ventriculography. Further delay is perhaps of little importance in patients with malignant gliomas, but in patients with benign tumors it may result in significant neurological damage that could have been prevented. This situation has been improved greatly by the development of two new diagnostic procedures. The first employs tantalum powder, which is sprinkled along the zone of resection at the conclusion of the primary tumor removal. The subsequent displacement of this tantalum powder by the enlarging neoplasm then can be demonstrated roentgenographically. 6,9 The second pro- Received for publication November ~5, * Supported in part by U.S. Public Health Training Grant 1-T01-NB055~8-01. Presented in part at the ~7th Annual Meeting of the American Academy of Neurological Surgery, San Francisco, California, October 17, cedure is isotope encephalography, which heretofore has been used mainly for preoperative localization of intraeranial neoplasms, although it also has been used in evaluating patients whose tumors have been treated by radiotherapy. 7'14-16 In 1953, Ashkenazy proposed that isotope eneephalography be used to identify the recurrences of brain tumors after primary resections,' and this idea has been mentioned by others since. However, relatively few studies of postoperative brain scans have been published, and in none has there been a detailed analysis of the patterns of postoperative isotope retention. 2,4,5,1~ At the beginning of our own experience with postoperativebrain scanswe worried that superficial retention of isotope in the scalp and skull might m~tsk underlying tumor recurrences and that further confusion might arise from isotope retention in deeper areas of fibrosis and gliosis. The present study was undertaken to analyze these and the other problems of interpretation mentioned below. Material and Method Seventy-three patients had 83 brain scans at various times after surgery. Of these, ~0 patients also had preoperative scans. Commercial rectilinear scanners, with a S- or 5-inch scintillation crystal and a coarse S-inch focusing collimator, were used for all examinations. For the postoperative scans, various agents were given intravenously or orally; two scans were performed at ~4 and 48 hours after the injection of 11~1 labeled human serum albumin, nine at 1 to 3 hours after the injection of Hg 2~ labeled neohydrin, 1~ at 1 to 3 hours after the injection of Hg 19~ labeled neohydrin, and 60 at 30 to 90 minutes after the injection or ingestion of Tc 99m as perteehnetate. From two to four views (anteroposterior, posteroanterior, left lateral, and right lateral) were obtained at each examination. Roentgenograms of the skull

2 11~ Robert H. Wilkins, Felix J. Pireher and Guy L. Odom made either in the scanning position immediately before the brain scan or at other times in the Department of Radiology were used for rough estimates of the sizes and locations of craniectomy and craniotomy defects. However, because of magnification, these fihns could not be superimposed exactly on the brain scans for accurate analysis. The retention in all postoperative scans was classified as superficial (within or adjacent to the calvarium) or deep, after comparison of the sagittal and coronal views. In each case, further classification was achieved by relating the areas of superficial isotope retention to the operative defect in the skull. In addition, the intensity of abnormal retention in each view was arbitrarily estimated as l-k- to 3+. These data were then correlated with similar data from the preoperative scans and with the following clinical parameters: location of original lesion, operative approach, pathological diagnosis, extent of resection, type of closure, radiotherapy, postoperative infection, chronological relation of scan to operation, and clinical or pathological evidence of residual or recurrent tumor at the time of brain scan. The sharpest definition of the retention patterns was obtained with technetium 99m. Superficial Retention The various patterns of superficial postoperative isotope retention were best demonstrated in the patients with non-neoplastic lesions. The four patients with chronic subdural hematoma or subdural empyema had postoperative patterns of superficial retention similar to their preoperative patterns, which consisted of diffuse retention on the lateral scan and crescent-shaped superficial retention on the anteroposterior and posteroanterior scans?,s The following case is illustrative. Case 1. E. C., a 60-year-old woman under treatment with warfarin sodium for a pulmonary embolus, fell and later gradually developed confusion and lethargy. On December 15, 1965, a brain scan (Tc 99m) showed superficial retention compatible with a left subdural hematoma, and this was confirmed by carotid arteriography and left temporal craniectomy. Hematoma membranes were found at operation. Tantalum dust was placed on the cortical surface and tantalum clips on the dura mater) 2 Plain roentgenograms on January ~1, 1966, showed the tantalum dust closer to the dural clips than previously. A second brain scan (TO 9m) on January ~6, 1966, again showed isotope retention in a crescent-shaped superficial pattern. When last examined on July 5, 1966, 7 months postoperatively, the patient was doing well and had no evidence of recurrent subdural hematoma. Plain roentgenograms showed the tantalum dust immediately adjacent to the dural clips, and a third brain scan (Tc 99m) still showed residual superficial retention, though less marked than previously (Fig. 1). A second pattern of superficial retention was observed in 16 scans. In these scans the abnormal isotope retention occurred in a fairly uniform distribution within the area of the craniectomy or eraniotomy defect. The following case demonstrates this type of uniform superficial retention. Case 2. J. W., a 45-year-old man, sustained a blow to the left temporoparietal area on August 15, 1964, resulting in a depressed skull fracture, dysi)hasia, and seizures. A brain scan (Hg 197) was performed on August :FIG. 1. Case 1. Anteroposterior brain scan with Tc 99m, 7 months after the evacuation of a chronic subdural hematoma from the surface of the left hemisphere. Crescent-shaped abnormal superficial retention is still present on the left, despite absence of a residual hematoma by other criteria.

3 Brain Scan in Supratentorial Intracranial Tumors 113 FIG. 2. Case 2. Left lateral roentgenogram of the skull made during a pneumoeneephalogram, showing the size and location of the left parietal craniectomy that had been performed for evacuation of an extradural and intracerebral abscess. ~4, 1964, showing increased superficial retention of isotope in the area of the injury. Three days later, after a carotid arteriogram and pneumoencephalogram, a left parietal craniectomy was performed, with evacuation of an extradural and intracerebral abscess. The patient improved, and on September 8, 1964, a second brain scan(hg I9~) showed abnormal superficial retention only in the area of the parietal craniectomy (Figs. and 3). The most common pattern of superficial isotope retention was present in 40 brain scans. In these scans the abnormal retention was also limited to the craniectomy or eraniotomy defect, but it tended to occur along the margins of the defect or flap rather than uniformly throughout the area concerned. This type of marginal superficial retention is demonstrated by the following ease. Case 3. L. W., a ~9-year-old woman, was injured in an automobile accident on June ~1, 1964, sustaining a blowout fracture of the left orbital roof. The same day a left frontal craniotomy was performed, with removal of orbital bone fragments and insertion of a dural graft. The frontal bone flap was replaced at the end of the procedure. The patient developed seizures, and a brain scan (Te 99m) was performed on April ~8, 1965, showing typical marginal superficial retention (Figs. 4 and 5). No unusual superficial retention was FIG. 3. Case 2. Left lateral brain scan with Hg 197, 12 days after the evacuation of the extradural and intracerebral abscess. Abnormal retention is rather uniformly distributed over the area of the craniectomy defect. This retention was shown to be superficial on an anteroposterior view. noted in the seven brain scans performed from 7 to 79 months after Torkildscn shunt operations. The type and the intensity of superficial postoperative retention were recorded for the 53 brain scans performed on the 18 patients with biopsies or partial resections of non-neoplastic lesions or benign neoplasms and the ~t8 patients with total resections of neoplastic or non-neoplastic lesions and no clinical evidences of recurrence. Of the 53 scans, 41 showed abnormal superficial retchtion, and 4 demonstrated this between 1~ and ~0 years postoperatively. Uniform or marginal superficial retention p~tterns were present in varying intensities in ahnost all of the postoperative time periods, but there was a distinct tendency for the abnormal superficial retention to become less pronounced with time. When a similar analysis of the 36 scans of patients in whom the bone flaps were replaced at surgery was compared with an analysis of tlle 17 scans of patients in who,ll the bone was discarded, no striking differences were detected. Five patients had postoperative craniotomy infections requiring secondary removal of the bone flap. These patients were scanned between 13 days and 14 months after removal of the infected bone, but no unusual intensity or patterns of retention were encountered. Likewise, radiotherapy appeared to have no effect on superficial isotope retention.

4 114 Robert H. Wilkins, Felix J. Pireher and Guy L. Odom FIG. 4. Case 3. Left lateral roentgenogram of the skull, showing the size and location of the left frontal craniotomy that had been performed for repair of a blowout fracture of the left orbital roof. Deep Retention Deep retention was demonstrated in 16 of 18 patients with clinical evidence of postoperative tumor recurrence; 12 of these had confirmation of recurrence by a second operation or by autopsy. The other two patients had recurrent astrocytomas, which ordinarily do not retain isotopes well. Because of the overlying vascular tissues, such as the temporalis muscles, recurrences in the temporal lobe were more difficult to identify by isotope encephalography than were recurrent neoplasms in the frontal, parietal, or occipital lobes, as shown by the following case. Case 4. H. W., a 45-year-old man, developed seizures, confusion, headaches, and left facial weakness after a blow to the head on March 4, Roentgenograms of the skull showed a shift of the pineal to the left, and on March ~2, 1965, a right temporal craniectomy was performed. The suspected subdural hematonm was not found. A right carotid arteriogram then showed evidence of a large mass in the right temporal lobe. The craniectomy was enlarged, and resection of an ependymoblastoma was performed. All gross tunmr was removed. Radiotherapy was given from April 6 to May 18, 1965 (5980r tumor dose). By December, 1965, the patient had developed seizures and lethargy. There were evidences of tumor recurrence by electroencephalogram and pneumoencephalogram. A brain scan (To 99m) on December 17, 1965, was interpreted as showing only superficial isotope retention, but in retro- FIG. 5. Case 3. Left lateral brain scan with Tc 99m, 10 months after the repair of the blowout fracture of the left orbital roof. Abnormal retention is mainly confined to the margins of the craniotomy defect. spect it demonstrated deep retention in the right temporal lobe (Fig. 6). The patient died on March ~1, 1966, and at autopsy a large recurrence was confirmed. Five patients with clinical evidences of tumor recurrence had two or more postoperative brain scans. In all five cases the area of deep retention enlarged with time, in keeping with the clinical progression of the recurrent neoplasm. The following case illustrates this development. Case 5. H. W., a 33-year-old woman, developed seizures, and on September s 1957, an astrocytoma of the left temporal lobe was resected. All gross tumor was re moved but the bone flap was not replaced because of cerebral edema. Between October 2 and November 9, 1957, the patient was given radiotherapy (4700r tumor dose). The patient did well for the next 7 years, but by April, 1965, she had developed a very mild right hemiparesis. On April 9, 1965, a brain scan (Tc 99m) showed a faint area of deep isotope retention in the temporo-occipital region. The patient then gradually developed dysphasia and headache. A second brain scan (Tc 99m) on May 12, 1966, showed a much larger and more intense area of abnormal retention (Figs. 7 and 8). Four days later the recurrence was verified by a second craniotomy. Microscopically, the recurrent astrocytoma appeared very anaplastic. Two patients with previous removals of cerebral metastases developed clinical signs

5 Brain Scan in Supratentorial Intracranial Tumors 115 of recurrent lesions, but their brain scans, when compared to preoperative scans, demonstrated that they had developed metastases in other areas of the brain. Three other patients had brain scans before and after radiotherapy, with no intervening tumor resection, and demonstrated reduction in the size and intensity of their previous areas of deep isotope retention. Only three of 34 patients with no clinical evidence of postoperative tumor recurrence had deep isotope retention. Two of the three had oligodendrogliomas resected 5 years and 9 years prior to brain scan, and may yet develop recurrence. The third case, which is illustrated below, showed deep retention initially but a later scan was negative. Case 6. T. F., a 5-year-old girl, developed headaches, vomiting, and papilledema. A Tc 99m brain scan on March 9, 1965, showed a prominent area of abnormal retention in the region of the third ventricle. A ventriculogram 2 days later showed a mass between the third ventricle and the right lateral ventricle. A Torkildsen shunt was performed, and on March 16, 1965, an astrocytoma was resected through a right frontal craniotomy. The bone flap was replaced at the end of the procedure. From April 26 to May 28, 1965, FIG. 6. Case 4. Anteroposterior brain scan with Te 99m, 9 months after the resection of the ependymoblastoma from the right temporal lobe. The abnormal retention in the area of the right temporal lobe is masked to a great degree by isotope retention in the adjacent extracranial tissues. the patient received radiotherapy (4960r tumor dose). By August, 1965, the patient was asymi)tomatic and neurologically norreal, but a brain scan (Tc 99m) on August 12, 1965, still showed abnormal deep isotope retention. No further therapy was given, and in June, 1966, the patient was still asymptomatic and neurologically normal. Interestingly, a second postoperative brain scan (Tc 99m) on June ~1, 1966, was normal (Figs. 9 and 10). Of the 14 patients with non-neoplastic lesions, only one had deep isotope retention postoperatively and this probably was related to his missile wound rather than his operation. Case 7. W. S., a 12-year-old boy, was struck in the left frontal area on June ~6, 1963, by an object of unknown type thrown from a power lawn mower. On June 28, 1963, a left frontal craniectomy was performed, with debridement of a compound depressed skull fracture and evacuation of an intracerebral hematoma. Because of persistent headaches, a brain scan (Hg 2~ was obtained on July 20, This showed deep retention in the left frontal lobe. Three years later the patient was normal neurologically, and a brain scan (Tc 99m) showed no abnormal retention. Among the 73 patients who had cerebral operations, only six had postoperative brain scans within 2 months after surgery. In two of these cases (including Case 2), superficial brain abscesses were evacuated. Preoperative and postoperative (12 and 23 days) brain scans showed no abnormal deep isotope retention. A third patient (Case 7) did have deep retention 22 days after evacuation of a traumatic intracerebral hematoma, but none 3 years later. Two patients whose cerebral neoplasms (spongioblastoma and metastatic carcinoma) were only partially resected had deep isotope retention 7 and 46 days postoperatively. Neither one had had a preoperative scan. The former patient is still alive, but has had persistent deep retention on brain scans at 6 and 11 months postoperatively. Persistent metastatic carcinoma in the other patient was verified at autopsy. The sixth patient had a metastatic carcinoma that was identified by abnormal deep

6 116 Robert It. Wilkins, Felix J. Pircher and Guy L. Odom Fro. 7. Case 5. Left lateral brain scan with Tc 99m, 8 years after resection of an astrocytoma of the left temporal lobe. A faint area of isotope retention is present in the left temporal and occipital lobes. isotope retention on a preoperative brain scan. This metastasis was totally resected, and a second scan done ~5 days postoperatively showed no deep retention. Of the ~0 patients with preoperative brain scans, 13 had total resections of intraeranial neoplasms. In all 13 cases abnormal deep isotope retention was present preoperatively. Eleven of these patients had no clinical evidences of tumor recurrence at the time of their second brain scans from ~5 days to 30 months postoperatively, and 10 had no deep isotope retention. The one exception was Case 6 discussed above. The other two patients did have clinical evidences of recurrent tumor, and both had deep retention in their postoperative scans. In an additional case, abnormal deep isotope retention was present preoperatively, and persistent deep retention was noted in a second scan performed ~0 months after the partial resection of a mixed glioma. Discussion Certain difficulties in the interpretation of the brain scans became apparent early in our study. The occasional masking of retention in the temporal lobe by retention in the vascular structures anterior, inferior, and lateral to it has been mentioned above (Case 4). In addition, superficial isotope retention, when viewed in only one plane, can be confused with areas of deep retention; coronal as F~G. 8. Case 5. Left lateral brain scan with Te 99m, 9 years after the resection of the left temporal astrocytoma. There is a zone of abnormal retention in the left temporo-occipital area that is more marked than that demonstrated a year earlier (Fig. 7). well as sagittal views are required to separate them. Though not encountered in our 73 patients, the scan pattern of a superficially recurrent tumor, such as a recurrent meningioma over the convexity of a cerebral hemisphere, might be difficult to differentiate from the usual postoperative pattern of superficial isotope retention. After a period of time, however, the former should be larger and more pronounced on a second brain scan, whereas the latter sh0'u]d be smaller and less noticeable. As with postoperative roentgenograms of the skull, the interpretation of a postoperative brain scan is easier if it can be compared with previous scans from the same patient. For this reason, there may be value in obtaining a brain scan ~ or 3 months after the resection of a brain tumor as a baseline for later scans. As might be expected, the occurrence and intensity of superficial isotope retention were greatest in the initial postoperative period, during the time of the maximum injury to the scalp and skull and the greatest rate of wound healing. However, it was surprising to find superficial isotope retention as long as ~0 years postoperatively. This is probably related to permanently altered vascularity in the healed areas. Also surprising was the lack of significant

7 Brain Scan in Supratentorial Intracranial Tumors 117 isotope retention in postoperative areas of cerebral gliosis and fibrosis, even as early as ~5 days after the total resection of a metastatic carcinoma. With the few exceptions mentioned above, deep retention in our patients meant tumor recurrence. This deep isotope retention was not masked by the superficial retention. Fro. 9. Case 6. Preoperative posteroanterior brain scan with Tc 99m, demonstrating the deep nature of the abnormal retention in the astrocytoma. :FIo. 10. Case 6. Anteroposterior brain scan with Tc 9~, 15 months after resection and 13 months after irradiation of the paraventricular astrocytoma. The previously abnormal area of deep isotope retention (Fig. 9) is no longer present. Summary Postoperative recurrences of supratentorial intracranial neoplasms may be difficult to diagnose in the early stages. The present study demonstrates the value of postoperative brain scans in the identification of these recurrences. Seventy-three patients had 83 postoperative brain scans, which were analyzed and compared with the patients' roentgenograms and preoperative brain scans. Superficial retention in the craniotomy or cranicctomy defects was common, especially in the initial postoperative period, but deep retention along the edges of cerebral resections was not seen. Deep retention was found usually to indicate tumor recurrence, and this was not masked by superficial retention. In the patients studied, isotope cncephalography was both accurate and useful in the identification of recurrent supratentorial neoplasms. References 1. ASHKENAZY, M. The detection of intracranial lesions, especially recurrent tumors, by the use of radioactive isotopes and tile scintillation counter. J. Am. pharm. Ass., 1953, 42:1~5-F26. ~. VAN ECK, J. H.M. Clinical value of isotope encephalography. J. Neurol. Neurosurg. Psychial., 1966, 29: GERMAN, W. J., FLANmAN, S., and DAVEY, L. M. Remarks on subdural hematoma and aphasia. Clin. Neurosurg., 1966, 12: GREEN, J., McDoNALD, J., YANG, G., and RUBIN, P. The usefulness of the brain scan in post-treatment brain tumor patients. Presented at 51st annual meeting of the Radiological Society of North America, Chicago, KvI~AI~, V., and BOU~EK, J. Die Bedeutung der Gammaenzephalographie fiir die Diagnostik van Hirngeschwulsbiezidiven. Zentbl. Neurochir., 196~, 23:1~2-~. 6. LINDGREN, M., LOFGREN, Y. 0., and LUNDBERG, N. Tantalum powder as an indicator of the brain turnour region for postoperative radiotherapy and the diagnosis of recurrence. Acta radiol., Stockh., 1957, 48:17-~5. 7. McArEE, J. G., and TAXDAL, D.R. Comparison of radioisotope scanning with cerebral angiography and air studies in brain tumor localization. Radiology, 1961, 77:~07-~. 8. MEYERSON, S. B., WALTERS, G. S., and REYNOLDS, D.H. Brain scans and subdural hematomas. J. Fla. reed. Ass., 1966, 53:~91-~95... ODOM, G. L., WOODHALL, B., and GOREE, J. Unpublished observations. 10. PLANIOL, T., and PERTVmET, B. Diagnostic des r6cidives de tumeurs intracraniennes par la gammaene6phalographie (Etude de 200 eas). Neuro- CMrurg~e, 196~2, 8:14-21.

8 118 Robert H. Wilkins, Felix J. Pircher and Guy L. Odom 11. RAY, B. S. Surgery of recurrent intracranial tumors. Clin. Neurosurg., 1964, 10: ~. VIETII, R. G., TINDALL, G. W., and ODOM, G. L. The use of tantalum dust as an adjunct in the postoperative management of subdural hematomas. J. Neurosurg., 1966, 24: VAN VLIET, P. D., TAUXE, W. N., SVIEN, H. J., and JENKINS, D.A. The effect of craniotomy on the brain scan. J. Neurosurg., 1965, 23: VOUTILAINEN, i., and PIItKANEN, T. Gammaen- cephalography as a method of checking on irradiation of brain tumours. Annls. Med. intern. Fenn., 1965, 54: WENDE, S. Das radiologisch ausgelsste Hirn(idem und seine Verhtitung. Fortschr. Geb. R6ntgStrahl., 1963, 98: ZIMMEH.MAN, C., and BLUESTEIN, S.G. Diagnosing benign and malignant intracranial disease with mercury 2~ neohydrin photoscanning. J. Mr. Sinai Hosp., 1965, 32: