In a follow-up study of 179 patients with glioblastoma verified histologically at Tohoku University Hospital during the 30 years from 1947 to 1977, we investigated the relationships between the postoperative survival rate and the operation era, sex of the patients, age of admission, duration of symptoms, location of the tumor, histological malignancy, method of resection, resection times, external decompression, radiotherapy, chemotherapy, and combination of treatments. Furthermore, we listed a clinical summary in 14 cases of long-term survivals, who survived postoperatively for 5 years or more.
Of the 179 patients, 6 could not be followed up, 10 died preoperatively and 27 died within the first postoperative month. The postoperative survival rates were 55%, 24%, 12% and 4%, respectively at 1, 3, 5 and 7 years. A significantly higher survival rate was revealed in younger cases (especially teenagers), in cases of Grade I or II on the first operation, and those with preoperative symptoms persisting for more than 2 years. No correlations were observed between the survival rate and the operation era, sex of the patients, and location of the tumor. The effect of radiotherapy and chemotherapy were doubtful. Although external decompression and/or shunt operations were irrelevant to the survival rate, the result was better when the tumor was removed as extensively as possible.

View full abstract

Adequate follow-ups were obtained in 173 of 179 cases (96.6%) of histologically verified glioblastoma patients experienced at Tohoku University Hospital during the 30 years from 1947 to 1977 Excluding 10 pre and the 27 postoperative deaths within one month, the survival rates of the remaining 136 postoperative cases after one, 3 and 5 years were 55%, 24% and 12% respectively.
Of the 14 cases who survived more than 5 years, 6 cases (4/6 males, 12 52 years old) are still alive and were the subject of this immunological study. The results of the study revealed that; 1) cytotoxicity of lymphocytes against KS-1 cells (established cell line derived from human glioblastoma) was augmented both preoperatively and within one month postoperatively in glioblastoma cases without radiation and chemotherapy. In long-term survivals of glioblastoma, cytotoxicity of lymphocytes was depressed significantly (p<0.01) to the level of healthy persons. In cytotoxicity-augumented glioblastoma cases, the administration of the anti-cancer drug made the cytotoxicity of lymphocytes depressed; 2) value of T-cells and count of peripheral lymphocytes were not significantly different among the preoperative patients, postoperative cases and the long-term survivals of glioblastoma; other preoperative brain tumor cases; and healthy persons; 3) tuberculin reaction was negative in 2 and positive in 4 of 6 longterm survivals; and 4) titer of serum immunoglobulin in long-term survivals were within the normal range.
Judging from the above results, we concluded that immunity in long-term survivals is similar to healthy individuals.

View full abstract

In the past several years a number of immunological studies of human brain tumors suggesting evidence of the existence of certain brain tumor-specific antigens have been extensively performed. However, there is other evidence that suppressive factors not only exist in the sera of patients with malignant brain tumors but also bind to lymphoid cells and that they can specifically modify the cell-mediated immune mechanisms to the tumors of the central nervous system. Therefore, humoral immunity seems to be playing an important role in the cell-mediated immune responses against brain tumors especially in relation to both antibody dependent cell-mediated cytotoxicity (ADCC) and blocking factor(s). In the present study, 23 human brain tumors (9 glioblastomas, 5 astrocytomas, 1 ependymoma, 1 medulloblastoma, 5 ineningiomas and 2 metastatic cancers) obtained at craniotomy have been examined to determine the presence or absence of immunoglobulin G (IgG) and immunoglobulin M (IgM) in the tumor tissue by the direct immunoperoxidase method using enzyme-labeled antigen binding Fab' fragments.
Of 23 human brain tumors examined, 3 glioblastomas, I ependymoma, 1 medulloblastoma, 5 meningiomas and 1 metastatic cancer displayed unequivocally positive immunocytochemical reaction for IgG in the tumor tissue. The intensity of staining differed among the tumor cells and IgG was mainly localized on the surface of the tumor cells by immunoelectron microscopy. On the other hand, IgM was not observed in any tumor tissues except for the positive staining limited within the capillary lumen of the tumors. In the meningiomas, the positive staining for IgG was more prominent around the capillaries, psammoma bodies and in the whorl formation, suggesting possible contribution of humoral immune reaction in the mechanisms of whorl or psammoma body formation. Although it is not clear as to whether or not IgG distributed on the surface of tumor cells acts to suppress or accelerate cellular immunity and complement-dependent cytolysis of tumor cells in vivo, our present immunocytochemical findings suggest tumor-specific antibodies among IgG “coating” on the cells of certain human brain tumors.

View full abstract

Experimental model of epidural spinal cord compression was produced in rabbits by injecting VX₂ tumor cell suspension anterior to the Th-13 vertebral bodies. The tumor grows into the epidural space directly through the intervertebral foramina or by destroying the vertebral bodies to compress the spinal cord and produce paraplegia in 3 to 4 weeks.
Using this animal model, edema of the spinal cord compressed by epidural neoplasms was studied. The water content of compressed spinal cord (67.2±1.2%) was significantly increased, compared with that of normal spinal cord (65.7±0.7%). The ratio of sodium and potassium content had a tendency to increase in the compressed spinal cord. The uptake of ^99mTc pertechnetate in the compressed spinal cord was significantly higher than that in the adjacent cord at the early stage of symptoms. The uptake was increased in proportion to the degree of neurologic symptoms. There was leakage of horseradish peroxidase into the gray matter of the compressed spinal cord. In areas of gray matter extravasation, the indicator appeared in the neurons.
These results indicate that breakdown of the blood-brain barrier and vasogenic edema develop in the spinal cord compressed by epidural neoplasms, and progress as the degree of compression increases.

View full abstract

A consecutive series of 112 cases of hypertensive intracerebral hematoma from June 1976 to April 1978 were studied with repeated scanning by CT of EMI model 1010 and ACTA model 0100. Ninety-four cases were diagnosed as basal ganglionic bleeding and 9 were subcortical bleeding in this series. Fifty-four of basal ganglionic bleeding were operated and 50 were not operated.
These 103 cases of supratentorial hematomas were analyzed by CT based on density of hematoma and perifocal low density area surrounding hematoma chronologically. It was suggested that indication and timing for surgery of hypertensive intracerebral hemorrhage should be decided on hematoma factors (namely, hematoma size, midline shift and ventricular rupture) and chronological change of perifocal area surrounding hematoma on CT.
1. A progressive and concentric diminution in density of hematoma always occurred from the periphery. The high density in the center of hematoma did not change for a certain period until the area of the density decreased to 1-1.5 cm in size. Then, the high density area of 1-1.5 cm in size started to decrease its density and reached isodensity with the surrounding brain parenchyma during 18-20 days. As could be predicted, the larger the hematoma the longer the resolution time, but the general pattern of resolution as described above can be identified.
2. The rim of lucency surrounding hematoma probably represented secondary changes of edema, degeneration and necrosis of the adjacent parenchyma. So, the perilucent area varied according to the age of intracerebral hematoma. The area surrounding hematoma was small within 6 hours from onset and then increased greatly until the 4th day. The peak of perilucent area ranged between the 4th and the 15th day. The area was noted to decrease after the 15th day and to disappear within a month after.
3. Mortality of patients with basal ganglionic hematoma correlated with hematoma size, midline shift and ventricular rupture. Thirteen of our 33 non-operative cases with hematoma of less than 2.5 cm in diameter and midline shift of less than 5 mm were all alive. Fourteen cases fell between 2.5 and 4.5 cm in diameter and within 10 mm in shift. In 3 of these 14 cases, rupture of the intracerebral hematoma into the ventricular system was observed. One of them died and the others survived. The other 6 cases showed hematoma of more than 4.5 cm in diameter and/or shift of more than 10 mm and rupture of the hematoma into the ventricular system. All of these cases died.
4. Positive contrast enhancement in non-operative cases was observed first on the 14th day from the onset and frequently had a ring formed configuration contrast enhancement after the 20th day. Ratio of positive contrast enhancement changed according to hematoma size. Large hematoma was accompanied by contrast enhancement of longer duration. Four types of configurations of contrast enhancement; namely, ring, ring with inside high density, incomplete ring, and spot form, were observed in these cases.

View full abstract

Carotid endarterectomy was performed on 36 patients with stenosis of extra-cranial carotid artery due to atherosclerosis. This study was undertaken to determine criteria for this procedure based on clinical symptoms, angiographical findings versus post-operative pathological findings and post-operative clinical courses. This series includes 7 cases of RIND'(s) and 29 cases of completed stroke. These cases were divided into four categories according to residual lumen of stenotic lesion demonstrated by arteriograms of 31 cases. Grade 1 : Less than 2 mm in diameter; Grade 2: Smaller in diameter than carotid artery; Grade 3: Greater in diameter than carotid artery, smaller than common carotid artery; and Grade 4: Greater in diameter than common carotid artery. Pathological findings were divided into four categories: ulcer, mural thrombus, hemorrhage and atherosclerosis. Angiographical findings were compared with pathological findings. Three cases were classified as Grade 1 and 17 cases were classified as Grade 2. Among these, 4 cases showed ulcer, 3 cases mural thrombus, 2 cases hemorrhage, and 8 cases atherosclerosis. Seven cases were classified as Grade 3. Among these 4 cases showed ulcer and 3 cases mural thrombus. Four cases were classified as Grade 4. Among these one case showed Mural thrombus and one case atherosclerosis. Among cases which had more than four irregularities in the arteriograms, 4 cases had mural thrombus, 5 cases ulcer and 5 cases atherosclerosis. The number of irregularities were related with severeness of abnormal findings in pathology. Among cases in which contrast medium remained more than 5 seconds after vanishing of laminal flow of common carotid artery, 6 cases had ulcer, 5 cases mural thrombus and 3 cases atherosclerosis. Of all operated cases 25 patients had revealed no worsening of clinical findings during a follow-up period from 5 months to 3 years and 2 months.
From our data described above, the authors propose the following surgical indications of carotid endarterectomy: Grade 1 cases definitely should be operated. Grade 2 cases may be good indication. Grade 3 and Grade 4 cases are relatively suitable for operation, if they show “irregularities” and “long resting time of contrast medium.”

View full abstract

Localized thalamic infarction was obtained in extremely high frequency utilizing our cerebral occlusion model in dogs. The ischemic cerebral tissue of 20 dogs subjected to occlusions ranging from 15 minutes to 24 hours and immediately sacrificed thereafter, were examined by electron microscope.
In the group subjected to 15-minute occlusion, no changes were observed in the nerve cells. In the 30-minute group, microvacuoles were found in the cytoplasm of some of the nerve cells in the ischemic tissue (microvacuolation). These microvacuoles were swollen mitochondria and rough endoplasmic reticulurn. In the one to 3 hours group, numerous nerve cells showed shrinkage with perineuronal vacuolation. The degree of shrinkage, however, varied from severe in some to slight in others. Mitochondria, endoplasmic reticulurn and ribosomal rosettes were altered in the slightly shrunken neurons. In the severely shrunken neurons the nucleus was also altered. In the 6 to 12 hours group, the number of slightly shrunken neurons decreased and the number of swollen and disrupted ones increased. Severely shrunken ones further deteriorated. In the 24 hours group, severely shrunken neurons were fragmentated and showed necrosis.

View full abstract

The present experiment was undertaken to clarify the correlation between pathological findings of primary brainstem damage and alteration of the far field acoustic response of the rhesus monkey by linear acceleration head impact using a HYGE sled and slider impactor system.
Brainstem damage was produced by the impact upon the frontal or occipital part of the head, at an linear acceleration in parallel to orbito-meatal line, 1, 000 1, 500 G and duration 3 ?? 5 msec.
In 4 out of 12 monkeys, the electron microscopic cellular changes and disturbance of microcirculation at the lower brainstem were observed and they died within one hour after the impact. The far field acoustic response disappears just after the impact without any relation to cortical EEG.
In the 6 monkeys that survived following the impact, the far field acoustic response did not disappear even just after the impact. Normal microcirculation and slight electron microscopic cellular changes were observed at the brainstem.
Monkeys that died of cervical cord injury or intrathoracic bleeding by the impact, had slight electron microscopic changes with normal microcirculation at the brainstem as in survival group. The far field acoustic responses did not disappear just after the impact but disappeared when EEG became flat by death.
According to these results, it is concluded that: 1) The primary brainstem damage consists of electron microscopic changes and disturbances of microcirculation is induced by pure linear acceleration head impact without any cranial fracture, intracranial hematoma or cerebral contusion; and 2) This primary brainstem damage is fatal in monkeys whenever all waves of the far field acoustic responses disappear.

View full abstract