Computed tomography in astrocytomas. A statistical analysis of the parameters of malignancy and the positive contrast-enhanced CT scan

Radiomics MRI Phenotyping with Machine Learning to Predict the Grade of Lower-Grade Gliomas: A Study Focused on Nonenhancing Tumors

Objective

To assess whether radiomics features derived from multiparametric MRI can predict the tumor grade of lower-grade gliomas (LGGs; World Health Organization grade II and grade III) and the nonenhancing LGG subgroup.

Materials and Methods

Two-hundred four patients with LGGs from our institutional cohort were allocated to training (n = 136) and test (n = 68) sets. Postcontrast T1-weighted images, T2-weighted images, and fluid-attenuated inversion recovery images were analyzed to extract 250 radiomics features. Various machine learning classifiers were trained using the radiomics features to predict the glioma grade. The trained classifiers were internally validated on the institutional test set and externally validated on a separate cohort (n = 99) from The Cancer Genome Atlas (TCGA). Classifier performance was assessed by determining the area under the curve (AUC) from receiver operating characteristic curve analysis. An identical process was performed in the nonenhancing LGG subgroup (institutional training set, n = 73; institutional test set, n = 37; and TCGA cohort, n = 37) to predict the glioma grade.

Results

The performance of the best classifier was good in the internal validation set (AUC, 0.85) and fair in the external validation set (AUC, 0.72) to predict the LGG grade. For the nonenhancing LGG subgroup, the performance of the best classifier was good in the internal validation set (AUC, 0.82), but poor in the external validation set (AUC, 0.68).

Conclusion

Radiomics feature-based classifiers may be useful to predict LGG grades. However, radiomics classifiers may have a limited value when applied to the nonenhancing LGG subgroup in a TCGA cohort.

Evidence and context of use for contrast enhancement as a surrogate of disease burden and treatment response in malignant glioma

The use of contrast enhancement within the brain on CT or MRI has been the gold standard for diagnosis and therapeutic response assessment in malignant gliomas for decades. The use of contrast enhancing tumor size, however, remains controversial as a tool for accurately diagnosing and assessing treatment efficacy in malignant gliomas, particularly in the current, quickly evolving therapeutic landscape. The current article consolidates overwhelming evidence from hundreds of studies in the field of neuro-oncology, providing the necessary evidence base and specific contexts of use for consideration of contrast enhancing tumor size as an appropriate surrogate biomarker for disease burden and as a tool for measuring treatment response in malignant glioma, including glioblastoma.

History and evolution of brain tumor imaging: insights through radiology

This review recounts the history of brain tumor diagnosis from antiquity to the present and, indirectly, the history of neuroradiology. Imaging of the brain has from the beginning held an enormous interest because of the inherent difficulty of this endeavor due to the presence of the skull. Because of this, most techniques when newly developed have always been used in neuroradiology and, although some have proved to be inappropriate for this purpose, many were easily incorporated into the specialty. The first major advance in modern neuroimaging was contrast agent-enhanced computed tomography, which permitted accurate anatomic localization of brain tumors and, by virtue of contrast enhancement, malignant ones. The most important advances in neuroimaging occurred with the development of magnetic resonance imaging and diffusion-weighted sequences that allowed an indirect estimation of tumor cellularity; this was further refined by the development of perfusion and permeability mapping. From its beginnings with indirect and purely anatomic imaging techniques, neuroradiology now uses a combination of anatomic and physiologic techniques that will play a critical role in biologic tumor imaging and radiologic genomics.

Survival prediction model of children with diffuse intrinsic pontine glioma based on clinical and radiological criteria

BACKGROUND

Although diffuse intrinsic pontine glioma (DIPG) carries the worst prognosis of all pediatric brain tumors, studies on prognostic factors in DIPG are sparse. To control for confounding variables in DIPG studies, which generally include relatively small patient numbers, a survival prediction tool is needed.

METHODS

A multicenter retrospective cohort study was performed in the Netherlands, the UK, and Germany with central review of clinical data and MRI scans of children with DIPG. Cox proportional hazards with backward regression was used to select prognostic variables (P < .05) to predict the accumulated 12-month risk of death. These predictors were transformed into a practical risk score. The model's performance was validated by bootstrapping techniques.

RESULTS

A total of 316 patients were included. The median overall survival was 10 months. Multivariate Cox analysis yielded 5 prognostic variables of which the coefficients were included in the risk score. Age ≤3 years, longer symptom duration at diagnosis, and use of oral and intravenous chemotherapy were favorable predictors, while ring enhancement on MRI at diagnosis was an unfavorable predictor. With increasing risk score categories, overall survival decreased significantly. The model can distinguish between patients with very short, average, and increased overall survival (medians of 7.0, 9.7, and 13.7 mo, respectively). The area under the receiver operating characteristic curve was 0.68.

CONCLUSIONS

We developed a DIPG survival prediction tool that can be used to predict the outcome of patients and for stratification in trials. Validation of the model is needed in a prospective cohort.

Radiation dose response for supratentorial low-grade glioma--institutional experience and literature review

PURPOSE

To examine radiation dose response for low-grade glioma (LGG) based on our institutional experience and to review the literature on this topic.

METHODS AND MATERIALS

Sixty-seven patients with supratentorial low-grade nonpilocytic astrocytomas (n=36) or oligodendrogliomas (n=31) were treated with postoperative radiation therapy (RT). Twenty-seven patients (group A) received 5520 cGy; 24 patients (group B) received 5940 cGy; and 16 patients (group C) received 6375 cGy. The corresponding median follow-up was 60, 35 and 91 months, respectively.

RESULTS

The disease-specific survival (DSS) at 5 and 10 years were 90.2% and 56.2%, 67.6% and 47.3%, and 62.5% and 50% for groups A, B and C, respectively (P=0.40). Only a greater extent of surgical resection and absence of contrast enhancement predicted DSS on multivariate analyses. Patients receiving higher doses of RT had higher complication rates.

CONCLUSION

Our data confirmed the lack of radiation dose response for supratentorial LGG as demonstrated in the previous randomized trials. The radiation dose should not exceed 5520 cGy because dose escalation did not result in an improvement of DSS and it also increased the complication rates. Future research should focus on the eradication of radioresistant clones either by the improvement of surgical resection or the use of cytotoxic agents that can target on the radioresistant tumor cells.

Awake craniotomy for aggressive resection of primary gliomas located in eloquent brain

OBJECTIVE

To determine with intraoperative neurologic and language examinations the maximal tumor resection achievable with acceptable postoperative neurologic dysfunction in patients undergoing awake stereotactic glial tumor resection in eloquent regions of the brain.

PATIENTS AND METHODS

Between October 1995 and December 2000, 65 patients underwent frameless stereotactic resection of glial tumors located in functioning tissue. During the resection, continuous examinations by a neurologist and speech pathologist were performed. The goal of surgery was to resect the maximum neurologically permissible tumor volume defined on preoperative T2 imaging. Tumor resection was stopped at the onset of neurologic dysfunction. Novel segmentation software was used to measure tumor cytoreduction based on pre- and postoperative magnetic resonance imaging. All patients underwent 3-month postoperative neurologic examinations to determine functional outcomes.

RESULTS

The cortical and subcortical white matter tracts at risk for injury were the left frontal operculum in 15 patients, the central lobule in 38, the insula in 11, and the left angular gyrus in 1. Thirty-four (52%) had a greater than 90% reduction in T2 signal postoperatively. In 26 patients thought to have low-grade tumors based on preoperative imaging, 12 proved to have grade 3 gliomas. Forty-eight patients (74%) developed intraoperative deficits; 34 (71%) recovered to a modified Rankin grade of 0 or 1 at 3 months postoperatively, 11 (23%) achieved a modified Rankin grade of 2, and 3 patients (6%) achieved a modified Rankin grade of 3 or 4 at 3-month follow-up. There was no operative mortality; 17 patients (26%) died from tumor progression during the follow-up period.

CONCLUSIONS

Combining frameless computer-guided stereotaxis with cortical stimulation and repetitive neurologic and language assessments facilitates tumor resection in functioning brain regions. Resecting tumor until the onset of neurologic deficits allows for a good functional recovery. Imaging software can objectively and accurately measure preoperative and postoperative tumor volumes.

Cerebellar astrocytomas in childhood

According to CT appearance and surgical observations, cerebellar astrocytomas can be separated into three types. On CT scan, cystic astrocytomas have a typical mural nodule; with contrast injection only the nodule becomes hyperdense; the wall of the cyst is not modified. In these cases, only the mural nodule is removed since the wall does not contain tumor cells. In contrast, false cystic astrocytomas present an irregular wall, diffusely enhanced and thick. Then the wall is invaded by tumor cells, it must be totally removed. Solid astrocytomas may invade the peduncle, the IV ventricle, and the subarachnoid spaces. Removal is sometimes questionably total. As recurrences are not frequently observed in these cases, radiotherapy is not always recommended. Rather, radiotherapy is only used in cases of undoubted partial removal or after partial removal of a recurrence. Of 63 cases, early postoperative mortality was 4.7% and late recurrence 6.3%.

Quantitative 201Tl SPET imaging in the follow-up of treatment for brain tumour: a sensitive tool for the early identification of response to chemotherapy?

The aim of this study was to establish if repeated quantitative 201Tl SPET scanning during follow-up of astrocytoma therapy can provide information that is relevant for clinical management. Sixteen consecutive patients, with histopathologically verified highly malignant astrocytoma, were followed during PCV chemotherapy. Imaging with 201Tl SPET and CT was performed repeatedly over 8-16 weeks until treatment discontinuation, with a maximum follow-up of 74 weeks. Tumour uptake volume (TUV), a measure of metabolically active tumour tissue, was calculated from the SPET images. The reliability of early identification of treatment failure, defined as > 25% tumour volume increase, following one course (week 8) and three courses (week 24) of chemotherapy, was calculated for the two imaging methods. 201Tl SPET positive patients (> 25% tumour volume increase) were compared with 201Tl SPET negative patients in terms of time to treatment discontinuation (TTD) and survival time (ST). The patients were followed with a total of 59 SPET examinations, and treatment was continued for a median 27 weeks (range 16-78 weeks). The comparative reliability of SPET and CT showed the highest sensitivity and accuracy for SPET in the early identification of astrocytoma treatment failure at the week 24 assessment. Patients with positive 201Tl SPET after three courses of chemotherapy had a significantly reduced TTD (P = 0.040) but not significantly reduced ST. Of the ten patients who received concomitant radiation and chemotherapy, five had a small (0-10 ml) TUV at the week 24 assessment. Patients with a TUV > 10 ml at this assessment had a shorter TTD (P = 0.016) and a reduced ST (P = 0.024) compared to patients with a TUV < 10 ml. In conclusion, the assessment of progressive disease by quantitative 201Tl SPET appears to provide information on treatment response, earlier and with a higher reliability than CT. Repeated 201Tl SPET scanning during follow-up of astrocytoma treatment is an alternative tool for the early identification of treatment failure.

The role of surgery in low grade gliomas

Several prognostic factors appear to be more important than treatment factors in determining the outcome of patients with low grade glioma. The role of surgery in establishing a histological diagnosis or in relieving neurological deficit is well established. The role of surgery in prolonging survival of patients with low grade glioma however, remains controversial, largely because studies in the literature on the management of this disease have been predominantly descriptive and observational, which precludes the setting of standards for the management of these patients. Several viable management options are discussed.

CT number distribution and its association with local control and as a marker of lung tumor response to radiation

An early noninvasive indicator of tumor response to therapy and the ability to predict clinical outcome may potentially enhance disease management. Currently, however, tumor response to therapy is often delayed, potentially compromising disease management. We examined the computed tomography (CT) number or Hounsfield unit distribution to follow lung tumor response to radiation treatment. To help interpret the results, we examined whether the CT number distribution follows a simple two-component model. The CT number distribution was derived from a CT-simulator for 11 patients with lung cancer before and after the initial radiation treatment (1-1.5 months, average 3,407 cGy). Clinical outcomes were followed in 8 patients who received 5,580-6,660 cGy. All patients were scanned serially, using identical radiation imaging parameters (voltage, current, scan time, and slice thickness) in a CT-simulator. The lung tumors were digitally contoured, and software windows were applied to avoid inclusion of lung tissue in the analysis. Histograms and statistical analysis of the CT numbers for the tumor were generated. Radiation-induced CT number or Hounsfield unit (HU) shifts exceeding a threshold (13 HU) in lung tumors were associated with (P=0.04) local control (> or = 10 months). Initial lung tumor size (below 100 cm3) was less well-associated with local control (P=0.26). The change in standard deviation of the CT numbers (derived from the more careful contouring and using software windows) induced by radiation treatment correlated with the change in average CT number (R2=0.71). The change in standard deviation did not correlate with a change in tumor volume (R2=0.02). Radiation treatments reduced the average CT number (P < 0.001). In summary, radiation reduces the CT number and this reduction may be associated with local control at 10 months. A two-component model is consistent with lung tumor number distribution and its response to radiation.

Prognostic significance of CT contrast enhancement within histological subgroups of intracranial glioma

We report the prognostic significance of tumor CT contrast enhancement within histological subgroups in 831 consecutive adult glioma patients of high-grade (n = 516) and low-grade (n = 315) histology. In the present report, a negative prognostic factor is associated with shortened survival.

METHODS

Survival analysis including Kaplan-Meier plots, log-rank tests, Cox analysis, and Aalen's linear model as implemented in SPSS and S-PLUS.

RESULTS

Sensitivity and specificity of contrast enhancement as a test for high-grade glioma was 0.87 and 0.79, respectively. Enhancement was a strong negative prognostic factor comparable to high-grade histology in the total patient population. Enhancement was also a negative prognostic factor within the subgroups adult high-grade (Grade 3-4), anaplastic (Grade 3), and low-grade (Grade 1-2) gliomas (p < 0.001). The prognostic implications of initial enhancement declined in high-grade patients surviving beyond 36 months. Tumor contrast enhancement or calcifications (in parentheses) were present in 96% (3.6%) of glioblastomas, in 87% (7.4%) of high-grade gliomas, in 56.5% of anaplastic gliomas, and in 21% (16.2%) of low-grade gliomas. Calcification was a positive prognostic factor within the high-grade group of patients (p < 0.0001).

CONCLUSION

Enhancement was a major prognostic factor comparable to high-grade histology in this glioma patient population. Enhancement was a negative prognostic factor within each of the adult subgroups high-grade, anaplastic (grade 3), and low-grade gliomas. Enhancement was strongly associated with but not pathognomonic for high-grade histology.

The significance of lack of MR contrast enhancement of supratentorial brain tumors in adults: histopathological evaluation of a series

BACKGROUND

To correlate magnetic resonance imaging (MRI) findings of non-enhancement of supratentorial brain neoplasms in adults with histopathologic findings.

METHODS

Forty adult patients whose preoperative MRI studies demonstrated a non-enhancing supratentorial brain neoplasm were identified retrospectively. Biopsy material for all patients was then reviewed by a board-certified neuropathologist.

RESULTS

Histopathologic examination identified 24 (60%) low-grade gliomas: 4 (10%) low-grade astrocytomas, 10 (25%) low-grade gliomas (not further classified), 8 (20%) low-grade oligodendrogliomas, and 2 (5%) low-grade mixed oligoastrocytomas. However, 16 (40%) nonenhancing lesions were classified as anaplastic gliomas: 12 (30%) anaplastic astrocytomas, 1 (2.5%) anaplastic mixed oligoastrocytoma, 1 (2.5%) anaplastic oligodendroglioma, and 2 (5%) anaplastic infiltrating gliomas of indeterminate subtype.

CONCLUSION

Non-enhancement of supratentorial brain neoplasms in adults does not equate with low-grade malignancy. This fact should be taken into account when biopsy and treatment are being planned in patients with nonenhancing brain tumors. More aggressive and/or surgical therapy might be indicated for such lesions, particularly those in the nondominant hemisphere or nonmotor areas.

Evaluation of malignancy in ring enhancing brain lesions on CT by thallium-201 SPECT

OBJECTIVE

To investigate patients with cystic enhancing lesions on CT and to determine whether thallium-201 (201Tl) SPECT adds to further preoperative information in differential diagnosis between gliomas and abscesses.

METHODS

Twenty one patients with cystic ring enhancing CT findings were studied and uptake indices were compared with CT enhancement volumes, histopathology, and survival times.

RESULTS

Fourteen high grade gliomas, three low grade gliomas, and four abscesses were found. Uptake was higher in the highly malignant glioma group (median thallium index (TI)=2.1), than in the low grade glioma group (median TI=1.4) or among the abscesses (median TI=1.6). Overlapping indices were found between high and low malignant cystic gliomas as well as between either one of the glioma groups and the infectious lesions, and there were no significant differences between groups. There was a level at the value 2, where TI > or = 2 correlated with tumour diagnosis. One low grade tumour had an extremely high index and a very high enhancement volume. Indices correlated significantly with CT enhancement volumes (P=0.005). There was no significant correlation between Tl indices and patient survival times among the high grade gliomas. One patient with a highly malignant tumour but low Tl uptake < 2, had a survival > five years.

CONCLUSIONS

It is concluded that high 201Tl uptake in enhancing cystic lesions is an indicator of highly malignant glioma. However, the differentiation between the high malignant gliomas and abscesses or low malignant gliomas by 201TL SPECT is only partial with an overlap between these groups.

Neural networks analysis of astrocytic gliomas from MRI appearances

A three-layered backpropagation neural network was developed to differentiate malignant from benign brain tumors in a group of patients with astrocytic gliomas. The MRI findings of 43 patients were reviewed before biopsy by three neuroradiologists independently. This provided a database made up of 129 patients' records each of which comprised 13 parameters derived from pre- and post-contrast MR images. The network's generalizing ability was then tested to predict the outcome of biopsy in 36 new cases and its performance compared to that of radiologist using ROC analysis. The output of the network with and without radiologists' impression yielded a better diagnostic performance with relative ROC areas of 0.94 and 0.91, respectively; compared to 0.84 obtained by radiologist. These results demonstrate that the neural network can effectively differentiate malignant from benign brain tumors.

Uptake of iodine-123-alpha-methyl tyrosine by gliomas and non-neoplastic brain lesions

Using single-photon emission tomography (SPET), the radiopharmaceutical l-3-iodine-123-alpha-methyl tyrosine (IMT) has been applied to the imaging of amino acid transport into brain tumours. It was the aim of this study to investigate whether IMT SPET is capable of differentiating between high-grade gliomas, low-grade gliomas and non-neoplastic brain lesions. To this end, IMT uptake was determined in 53 patients using the triple-headed SPET camera MULTISPECT 3. Twenty-eight of these subjects suffered from high-grade gliomas (WHO grade III or IV), 12 from low-grade gliomas (WHO grade II), and 13 from non-neoplastic brain lesions, including lesions after effective therapy of a glioma (five cases), infarctions (four cases), inflammatory lesions (three cases) and traumatic haematoma (one case). IMT uptake was significantly higher in high-grade gliomas than in low-grade gliomas and non-neoplastic lesions. IMT uptake by low-grade gliomas was not significantly different from that by non-neoplastic lesions. Diagnostic sensitivity and specificity were 71% and 83% for differentiating high-grade from low-grade gliomas, 82% and 100% for distinguishing high-grade gliomas from non-neoplastic lesions, and 50% and 100% for discriminating low-grade gliomas from non-neoplastic lesions. Analogously to positron emission tomography with radioactively labelled amino acids and fluorine-18 deoxyglucose, IMT SPET may aid in differentiating high-grade gliomas from histologically benign brain tumours and non-neoplastic brain lesions; it is of only limited value in differentiating between non-neoplastic lesions and histologically benign brain tumours.

Neurological assessment of high grade astrocytomas following high dose radiotherapy as sole treatment

Between January 1985 and June 1991, 19 patients, in whom the site and extension of the tumour prevented surgical excision, were treated with high dose radiotherapy as sole treatment for high grade astrocytomas. Quality of life, according to functional capacity, was measured prospectively before treatment and 4 weeks later. High dose radiation improved the functional capacity in only four of the patients (21%). The mean duration of improvement was 12 weeks and median survival 22 weeks (range 4-80). We conclude that high dose radiotherapy is not very useful and is probably not justified in this group of patients.

Endothelial differentiation in intracerebral and subcutaneous experimental gliomas

Blood-brain barrier (BBB) properties of endothelial cells have on impact on brain tumor behavior, diagnosis, and response to therapy. Biochemical BBB properties are expressed by endothelial cells within intracerebral (IC) gliomas but little is known regarding the expression of BBB-associated proteins within gliomas established subcutaneously (SC), a site that is frequently used in experimental glioma models. We compared the expression of two BBB proteins, glucose transporter type-1 (Glut1) and endothelial barrier antigen (EBA), in IC and SC rat 9L and F98 gliomas. The percentage of microvessels with immunohistochemically-detectable Glut1 and EBA in IC 9L tumors (31-98%) contrasted with that found in SC 9L tumors (0-3.9%) (P < 0.0001). Likewise, the percentage of immunohistochemically-positive vessels in IC F98 tumors (35-66%) differed markedly from that in SC F98 tumors (0%) (P < 0.0001). These differences were not explained by effects of tumor location on vessel density or tumor histology. These findings demonstrate that the peritumoral environment influences endothelial differentiation within glial tumors and suggest that glioma cells maintain but do not induce the expression of barrier properties in vessels that infiltrate tumor from surrounding tissue.

Vascular differentiation and glucose transporter expression in rat gliomas: effects of steroids

The GLUT1 isoform of the glucose transporter is normally expressed at high levels in differentiated brain vessels that also express a permeability barrier. In contrast, malignant brain neoplasms have relatively undifferentiated vessels that are highly permeable, proliferate to high vascular densities, and often lose GLUT1 expression. Using the rat intracerebral 9L glioma model, we investigated whether dexamethasone-induced changes in permeability are associated with the appearance of other differentiated vascular properties. The percentage of vessels expressing immunohistochemically detectable GLUT1 (74.2 +/- 6.1%) and the tumor vessel density as assessed by laminin immunostaining (282 +/- 37 vessels/mm2) did not vary with control tumor size. Dexamethasone treatment resulted in an 83% reduction of vascular permeability to intravenous Evans blue, an increased percentage of vessels expressing GLUT1 (106.4 +/- 10.5%), lower vascular density (102 +/- 64 vessels/mm2), and smaller tumor size (control cross-sectional area, 17.0 +/- 3.4 mm2; treated, 4.6 +/- 1.0 mm2). Essentially all vessels became GLUT1-positive after dexamethasone treatment. Increased GLUT1 expression by glioma vessels in association with the appearance of other signs of differentiation (low vascular density, slow tumor growth) suggests that immunostaining for GLUT1 may identify neoplasms that are biologically less aggressive.

A Comparison of CT Contrast Enhancement and BUDR Labeling Indices in Moderately and Highly Anaplastic Astrocytomas of the Cerebral Hemispheres

Contrast enhancement on computerized tomography (CT) scans has been used in directing therapy for presumed intracranial gliomas. However, for moderately anaplastic astrocytomas (MOAAS) and highly anaplastic astrocytomas (HAAS), it provides no information about proliferative potential. The bromodeoxyuridine (BUDR) labeling index (LI), however, indicates proliferative potential, correlating with histologic malignancy and survival. An LI < 1% is a favorable indicator; LI > 5% suggests more aggressiveness. To determine the correlation, if any, between BUDR LI and contrast enhancement, CT scans of 71 patients with cerebral hemisphere tumors labeled with BUDR were retrospectively reviewed. Among 36 MOAAS, the BUDR LI was < 1% in 77% of enhanced tumors and 61% of unenhanced tumors. Among 35 HAAS, it was < 5% in 56% of enhanced tumors and 90% of unenhanced tumors. Therefore, contrast enhancement on CT scans does not always correctly predict proliferative potential in these tumors, and biopsy and labeling studies are recommended before therapy.

Molecular anatomy of the blood-brain barrier as defined by immunocytochemistry

This review outlines the recent developments and improvements of our knowledge concerning the molecular composition of the BBB as revealed by immunocytochemistry. Data have been accumulated which show that the BBB exhibits a specific collection of structural and metabolic properties which are also found in tight transporting epithelia. This conclusion is substantiated by (i) the implementation of antibodies which recognize proteins of non-BBB origin, to show that these biochemical markers and the functions that they represent are localized in the BBB endothelium; and (ii) the characterization of target molecules to which polyclonal or monoclonal antibodies which have been generated to epitopes of the BBB endothelium or brain homogenates. According to these data the protein assemblies comprising the phenotypical appearance of the BBB can therefore be defined by the particular selection as well as topological expression of common epithelial antigens, rather than the expression of BBB-unique molecular species. In this respect the immunocytochemical data corroborate the physiological assumption that the BBB possesses the character of a specific polarized epithelium. Attention is also given to the description of developmental expression of BBB-related immunomarkers. By collecting the data from different sources we introduce a classification of the BBB marker proteins according to their developmental appearance. Three groups of proteins are classified with respect to their sequential expression around the time of BBB closure: Phase E (early) markers which appear before BBB closure, phase I (intermediate) markers which are expressed at the time of BBB tightening, and phase L (late) markers which are detectable after the closure of the BBB. Such a scheme may to be useful in better defining the maturation process of BBB, which apparently is not a momentary event in brain development, but rather consists of a temporally sequenced process of hierarchically structured gene expression which finally define the molecular properties of the BBB. This process continues even after parturition, especially with regard to the achievement of immunological properties of the mature BBB. By examining the developmental spatio-temporal expression of different BBB markers we conclude that the mechanisms governing the pattern of BBB maturation are not limited to the interactions occurring between glial and endothelial cells. We therefore suggest a heuristic model in a triangular interrelationship that includes differentiation effects of neurons on glia and of glia cells on the BBB endothelium.(ABSTRACT TRUNCATED AT 400 WORDS)

Early prognosis of supratentorial grade 2 astrocytomas in adult patients after resection or stereotactic biopsy. An analysis of 50 cases operated on between 1984 and 1988

50 adult supratentorial low-grade astrocytomas operated upon between 1984 and 1988 were analysed retrospectively with respect to postoperative condition and progression-free survival. Pilocytic lesions were excluded. In 32 instances the tumour was macroscopically completely removed and partially in 4. In 14 cases a stereotactic biopsy was performed only. 10 patients received postoperative radiotherapy with 55 to 65 Gy. 1 patient died perioperatively from pulmonary embolism. 39 patients could resume their previous activities after discharge from the hospital, 10 were significantly disabled by neurological deficit, reduced neuropsychological performance or medically intractable epilepsy. Postoperatively, most patients required continuous anti-epileptic medication, 10 recurrences or tumour progressions of incompletely removed or merely biopsied lesions were observed within the mean follow-up period of 22 months. All recurrences after gross total removal, that were reoperated, had progressed to a malignant glioma. Of the prognostic tumour characteristics analysed, a histologically well-delineated tumour demarcation was most clearly associated with a favourable prognosis. Concerning treatment modalities, gross total resection was associated with a favourable prognosis. Radiotherapy was associated with an unfavourable outcome but this is probably due to selection of otherwise unfavourable cases.

Adult supratentorial tumors

MR imaging has represented a significant advance in the diagnosis and management of adult supratentorial neoplasms, and indeed MR has become the modality of choice for the initial diagnosis and follow-up in these cases. It is clear that the overall sensitivity has increased significantly, due to both the use of MR itself and to the implementation of IV contrast agents. It appears that some inroads have been made with regard to specificity, but significant overlap still exists in the imaging appearance of many of these entities. A thorough knowledge of neuroanatomy, pathophysiology, and neuropathology aid markedly in the imaging diagnosis of these processes.

Malignancy criteria in computed tomography of primary supratentorial tumors in infancy and childhood

The CT findings of 11 malignant primary supratentorial tumors in children are compared with those of 38 benign tumors. The malignant tumors were more often laterally placed and surrounded by edema. Irregular tumor shape and high or "mixed" density were more frequent in the malignant group where larger tumor volume was found. The contrast enhancement pattern was also different. Thus, ring enhancement with central lucency was more often seen in the malignant tumors. Seven of the malignant tumors had both irregular shape and edema, while this combination was not present in any of the 38 benign tumors.

Correlation of neuropathologic findings, computerized tomographic and high-resolution ultrasound scans of canine avian sarcoma virus-induced brain tumors

The location, size, geometry and neuropathological findings of anaplastic astrocytomas (AA), gliosarcomas and sarcomas induced by the avian sarcoma virus (ASV) in dogs were compared with images generated using computerized tomography (CT) and real time high-resolution ultrasound (HRUS). Seven AA showed a wide range of findings on CT. Pre-contrast CT scans showed that the tumors could be hyper, hypo, or isodense. Three of seven AA had no contrast enhancement; two of these tumors were also isodense which resulted in a false-negative CT exam. Partial enhancement was seen in one tumor. This resulted in a sensitivity of detection of 72%. Real time HRUS was able to define tumor location, size and geometry of the AA missed or incompletely imaged by CT. All tumors were hyperechoic. Inhomogeneity of the echo pattern was due to hemorrhage, cyst formation, and necrosis within the tumors. Such secondary tumor characteristics were more accurately defined by HRUS compared to CT. Vasogenic edema in the brain surrounding tumors was of low density on CT and hypoechoic or indistinguishable from normal brain on US. Similar findings were seen in six gliosarcomas, two of which were not visualized by either pre- or post-contrast enhanced CT scans (sensitivity of 66%). Sarcomas differed in that they were either hyper or isodense; none were hypodense. The area of increased density matched the tumor geometry and correlated with dense cellularity and reticulin deposition. All 13 sarcomas showed contrast enhancement (100% sensitivity), but in two tumors, contrast enhanced CT underestimated the size of the tumor. Because of the large size and multiplicity of the sarcomas, HRUS imaging was not able to resolve the entire tumor volume because of limited imaging access. Intravenously injected horseradish peroxidase (HRP) crossed the tumor blood-brain barrier (BBB) only in those tumors in which contrast enhancement was seen. These studies suggest that intraoperative HRUS imaging may be useful in detecting and delineating human AA incompletely visualized by CT.

Primary glioma: diagnosis with magnetic resonance imaging

Seventeen patients with surgically documented primary glial-origin brain tumors were evaluated by magnetic resonance imaging and high-resolution computed tomography. The exclusion of CT ring-enhancing lesions directed the focus of this study toward lower grade tumors that were more difficult to diagnose. The computed tomography abnormalities were often subtle and included areas of low attenuation, mass effect, and focal enhancement. Spin-echo sequences with both heavy T1 and T2 weighting were utilized. Prolonged T1 and T2 values were observed in all tumors. The T2-weighted spin-echo 1000/120 sequence was the most sensitive in tumor detection and was positive in all cases. Magnetic resonance imaging was superior to computed tomography in tumor detection, tumor localization, assessment of tumor extent, and determination of associated changes, ie, brain stem encroachment. All the magnetic resonance sequences used showed an increase in severity of imaging changes with increasing tumor grade. The T2-weighted sequence showed progressive margin irregularity, whereas the T1-weighted (inversion recovery) sequence showed increasing severity of internal tissue changes. The superior resolution of these changes by magnetic resonance imaging may have implications for better assessment of tumor grade in the future than is currently possible with computed tomography.

Supratentorial ependymomas. Neuroimaging and clinicopathological correlation

In 10 patients with supratentorial ependymomas, the tumors exhibited hyperdensity on computerized tomography (CT) scanning prior to contrast infusion and, with one exception, all tumors were mixed lesions with the low densities suggesting cystic or necrotic portions. Eighty percent of the tumors contained small calcifications. Characteristically, the tumors were well demarcated and demonstrated moderate to marked enhancement after the intravenous administration of contrast material. Angiograms obtained in some patients showed mild hypervascular tumor staining and absence of large feeding arteries. The degree of contrast enhancement, angiographic vascularity, and tumor stain was compared to the pathological anaplasia of the tumors. No correlation was observed. Of four patients who were still alive during a follow-up period of 4 years or longer, three had recurrences with inoperable tumors; the remaining patient is without recurrence after craniospinal radiation. This same patient belonged to a group of five patients with a diagnosis of high-grade ependymoma, four of whom had recurrence. Follow-up CT accurately recorded the clinical course of each patient. Annual routine follow-up examinations are proposed for patients with low-grade ependymomas, and for those with high-grade ependymomas follow-up CT should be performed every 6 months. The characteristic appearance and behavior of these tumors include several distinctive features on angiographic and CT images.

Prediction of survival in glioma patients by means of positron emission tomography

The aim of this study was to determine whether positron emission tomography (PET) with fluorine-18 (18F)-2-deoxyglucose (FDG) can be used as a prognostic test in patients with high-grade cerebral gliomas, regardless of the treatment given. Forty-five patients with astrocytoma Grade III or IV were included in this analysis. The mean survival time of patients with tumors exhibiting high glucose utilization as determined by PET-FDG was 5 months, whereas patients with gliomas showing lower glucose utilization had a mean survival period of 19 months. It is postulated that PET-FDG scans reflect the biological behavior of high-grade astrocytomas and may be used to predict the survival time of patients harboring such neoplasms.

Computed tomographic and histopathological studies of pontine glioma

Correlative study of computed tomography (CT) and pathologic findings was performed in eight cases of pontine glioma. All patients except one had chemotherapy and radiotherapy, with no surgical intervention. On initial CT scans, all patients had hypodense lesions in part or the whole of the pons, and there was evidence of mass effect. Four of eight cases showed contrast enhancement. After chemotherapy and radiotherapy, swelling of the pons and the width of the hypodense areas decreased. The hypodense areas sometimes became isodense in correlation with clinical amelioration. After several months of remission in responding cases, ring-enhanced lesions reappeared at the primary site, together with recurrent neurological signs. Pathological study postmortem was focused on the histological counterparts of the CT findings of central and perifocal hypodense areas and contrast enhancement. In six of seven treated cases, the central hypodense area surrounded by ring enhancement was shown to be coagulation necrosis. Higher cellularity and hypervascularity with glomeruluslike structures of small vessels were generally observed in enhanced areas. The areas diffusely infiltrated by tumor cells, but not enhanced in CT scans, had few abnormal vessels. Tumor cells were seen not only in hypodense areas around the enhanced portion but also in areas far beyond the enhanced portion. Exophytic expansion of tumor was observed in two cases on postmortem examination. One of these was detected by CT scans before death.

Microvascular permeability in induced astrocytomas and peritumor neuropil of rat brain. A high-voltage electron microscope-protein tracer study

Brain tumors, benign and malignant, are characteristically more permeable to various types of tracer molecules than the neuropil in which they are embedded. Impermeability of brain neuropil capillaries is imparted by the blood-brain barrier, the anatomic basis of which is the network of interendothelial zonulae occludentes that seal capillary endothelial cells. To explore both the vascular elements of brain neoplasms and the route of tracer extravasation from them, as well as the possible effects of brain tumors on the permeability of peritumoral neuropil capillaries, brain tumors were induced in newborn Wistar rats by intracerebral (i.c.) injection of C-6 astrocytoma cells. The protein tracer horseradish peroxidase (HRP) was injected systemically into both normal and tumor-bearing rats to mark the pathway along which it flowed into the tumor parenchyma tissue spaces, and to signal any concomitant tracer loss from the tumor extracellular compartment or peritumoral brain capillaries, into the neuropil extracellular milieu. Electron-microscopic examination of thin plastic sections of tumor and peritumoral neuropil revealed massive extravasation of tracer into the tumor tissue spaces, but none was seen outside of the capillaries in the surrounding brain neuropil. Zonulae occludentes of both tumor capillary endothelium and brain capillary endothelium were devoid of tracer and judged tight (sealed). Tracer was seen in pinocytotic vesicles in the highly attenuated endothelium of tumor capillaries and also in cytoplasmic vesicles within the tumor cells. The peritumoral and contralateral neuropil capillary endothelium exhibited reaction product-filled pinocytotic vesicles and vesiculo-tubular conduits. Often, one end of a HRP-filled vesiculo-tubular channel appeared continuous with either the luminal or abluminal plasmalemma. High-voltage electron microscopy of these conduits often showed them to be continuous with both luminal and abluminal surfaces of the endothelium, thus forming a continuum across the capillary wall. In addition, these transendothelial channels, clearly constituted as chains of fused vesicles, were often seen in close proximity to, or fused with, dense bodies in the endothelial cytoplasm. In spite of the presence of HRP-filled structures in the peritumoral neuropil capillary endothelium of tumor-bearing rats, no evidence of tracer extravasation from these vessels was apparent. These results suggest that although peritumoral and contralateral neuropil capillaries possess the machinery for extravasation of tracer, likely as a response to the presence of the neoplasm, tracer is not lost but, instead, is degraded by endothelial enzymes.(ABSTRACT TRUNCATED AT 400 WORDS)

Transient neurologic disturbances, brain tumors, and normal computed tomography scans

During a 4-year period, four patients presented with transient disturbances in neurologic function that were diagnosed as seizures in two and transient ischemic attacks in the other two. Computed tomography (CT scan), both with and without contrast, was normal in all four patients. Isotopic brain scans (3 patients), cerebral angiograms (4 patients), and lumbar punctures (4 patients) were normal. Electroencephalograms (EEG) were normal in two patients and abnormal in two patients (consisting of focal slowing). Within 4.5 months, all patients developed symptoms and signs of a brain tumor, and in all four, CT scan now revealed a large mass lesion which at surgery was shown to be a malignant astrocytoma. These four patients constituted 4% of the total number of patients with malignant astrocytomas that were seen at the NYU Medical Center during this same time period. It is stressed that the CT scan may be normal early in the course of patients with brain tumors, particularly if they present with a transient disturbance in neurologic function. The first evidence of the tumor in such patients may be a slow-wave abnormality on the EEG. Patients who are suspected of having a brain tumor should, if the initial CT scan is normal, have the scan repeated later.

Regional blood-to-tissue transport in ethylnitrosourea-induced brain tumors

Regional blood-to-tissue transport, expressed as a unidirectional transfer rate constant (K), was measured in experimental brain tumors using alpha-aminoisobutyric acid (AIB) labeled with carbon 14 and quantitative autoradiography. A total of sixteen oligodendrogliomas, four mixed gliomas, three astrocytomas, two diffuse gliomatosis, one anaplastic astrocytoma, one ependymoma and four malignant schwannomas were studied in 9 rats. The mean Ks for all glioma classifications were similar, averaging 2.6 +/- 0.4 (standard error of the mean) ml . kg-1 . min-1, and were only slightly higher than those for nontumorous parietal cortex (2.1 ml . kg-1 . min-1), corpus callosum (0.9 ml . kg-1 . min-1), and a comparable region of brain tissue in the contralateral hemisphere (1.3 ml . kg-1 . min-1). Values of K varied minimally in the intracerebral gliomas and were marginally correlated with tumor cell morphology in only two tumors. In some (but not all) of the larger gliomas, increased vascularity, with or without endothelial proliferation, was associated with a 3- to 15-fold increase in K. Regional K values in malignant schwannomas were highly variable (4 to 207 ml . kg-1 . min-1) and generally were not correlated with specific histological features of the tumor, except in some regions with increased vascularity. Estimates of the average fractional extraction of AIB by the intracranial gliomas and malignant schwannomas were 0.01 and 0.2, respectively; average fractional extractions for nontumorous brain were approximately 0.003.

Computerized tomographic and pathologic studies of the untreated, quiescent, and recurrent glioblastoma multiforme

Pathological findings in 20 cases of glioblastoma multiforme were correlated with clinical histories and computerized tomographic (CT) scans. This was done to define the neoplasm in three stages: before treatment, during remission, and during recurrence. The untreated lesions were markedly cellular neoplasms composed predominantly of small anaplastic cells. The radiographic central region of low density was necrosis, the enhancing rim was a cellular zone of viable neoplasm, and the perilesional low-density area was edema with infiltrating tumor. In these 20 cases, all of the identifiable neoplasms lay within the zone of peritumoral edema or contrast enhancement, although small anaplastic cells may have been present in more distant regions. The lesions in remission were remarkable for their minimal mass effect, discrete nature, extensive necrosis, and content of large bizarre glia. The large cells were confined to the original tumor bed and were consistent with neoplastic cells inactivated and immobilized by radio- and chemotherapy. These lesions were accurately localized by CT scanning. The recurrent lesions were heterogeneous, but most were formed of widely disseminated small anaplastic cells. The highly cellular regions of such lesions could be localized by CT scanning, but CT could not detect less cellular foci in the cerebrum, cerebellum, or brain stem. In one patient, the contrast-enhancing lesions of "recurrence," were foci of radionecrosis, underscoring the difficulty in distinguishing this entity from recurrent neoplasm.

Image processing and quantitative assessment in computer tomography for non-surgical treatment of brain tumours

Using the EMI computer tomographic (CT) system, EMI-1010, a series of new programmes were developed for the digital analysis of the CT images in order to make a more objective and quantitative assessment possible of two non-surgical methods of treatment of brain tumours such as irradiation and chemotherapy. Amongst the various therapeutic effects demonstrable from the Ct data, a reduction of the mass effect was found to lower the average CT number, with a dilatation of the cisterns and ventricles. In contrast, an improvement in the amount of perifocal oedema increased the average CT number of the region, however the changes in CT number of the tumour itself may be variable. The separate evaluation of these factors, therefore, gives more information about the results of the treatment than a simple analysis of the histogram of the region. Circumscribed tumours are fairly well evaluated with our programme for the statistical analysis of the volume and the CT weight of tumours and the degree of contrast enhancement using histograms and subtraction scans. For the digital analysis of the ventricular system, the subarachnoid space, perifocal oedema, and irregularly shaped infiltrating tumours, our programmes for the character-image print-out and edge correction for the partial-volume effect of skull and air are much more useful than the CRT display for data extraction and geographic-pattern recognition.

Angiography and computed tomography in the diagnosis of primary lymphoma of the brain

Three patients with primary lymphomas of the brain are presented. The angiographic and computed tomographic (CT) appearances resembled the few cases reported previously. In all three cases angiography showed avascular tumours with little vessel displacement or involvement whereas the CT scans showed large strongly and homogeneously enhancing tumour masses. Neither method alone gives pathognomonic appearances, but taken together angiography and CT should suggest the diagnosis of this rare tumour.

Computed tomography of benign supratentorial astrocytomas of infancy and childhood

The CT findings of 15 benign supratentorial astrocytomas in children less than 15 years of age are compared with the CT findings of 19 supratentorial tumors of other histological types in the same age group. Astrocytomas were more often hypodense, lacked calcification and showed greater contrast enhancement than other tumors. Seven of the 15 astrocytomas were hypodense, without calcification and showed contrast enhancement of more than 10 Hounsfield units, whereas this coexistence was not present in any of the 19 tumors of the other histological types.

Contrast enhancement with time in gliomas. Stereotactic computer tomography following contrast medium infusion

Dynamic enhancement of astrocytomas has been investigated with CT following infusion of contrast medium. The areas of measurement corresponded with obtained stereotactic biopsies. The resulting types of enhancement are discussed in terms of a 3-compartment model. Great differences exist in the degree of enhancement of astrocytomas of grade I to II and those of grade III to IV. In the latter tumours a dynamic response appears to occur, distinctly different from that in infarctions.