Comparison of Gadomer-17 and gadopentetate dimeglumine for differentiation of benign from malignant breast tumors with MR imaging.

RATIONALE AND OBJECTIVES: This study compared gadopentetate dimeglumine (molecular weight, 0.5 kD), a standard contrast medium, and Gadomer-17 (apparent molecular weight, approximately 35 kD), a new, clinically applicable, large-molecular contrast medium, with respect to their microvascular characterizations of experimentally induced breast tumors at magnetic resonance (MR) imaging. MATERIALS AND METHODS: A spectrum of breast tumors, benign through highly malignant, was induced in Sprague-Dawley rats (n = 30) by intraperitoneal administration of N-ethyl-N-nitrosourea (ENU), a potent carcinogen. All animals underwent three-dimensional spoiled gradient-recalled MR imaging, with precontrast imaging and dynamic postcontrast imaging after injection of gadopentetate dimeglumine (0.1 mmol/kg) and Gadomer-17 (0.03 mmol/kg), administered in a random order at a 24-hour interval. Several microvascular parameters were compared: the endothelial transfer coefficient (K(PS)), a measure of microvascular permeability; the fractional plasma volume (fPV), and the plasma equivalent volume. Each MR imaging parameter was correlated with histopathologic findings. RESULTS: With Gadomer-17, the mean values for K(PS) and fPV were significantly greater in carcinomas than in fibroadenomas (P < .004 and .04, respectively). With gadopentetate dimeglumine, the mean values for fPV and PEV were significantly greater in carcinomas (P <. 004 and .02, respectively). Because of the high variability within both fibroadenoma and carcinoma groups, however, there were no significant correlations between K(PS), fPV, or PEV and histopathologic tumor grade as indicated by the Scarff-Bloom-Richardson score, for either agent. CONCLUSION: Although the K(PS) and fPV estimates obtained from dynamic MR imaging data with Gadomer-17 enhancement offer some potential for characterizing breast tumors, none of the quantitative microvascular parameters derived with either agent were significantly correlated with histopathologic tumor grade.

Histopathological study of experimental poststreptococcal pneumonia in mice. Group A, type 50, streptococcal infection of murine nares controls with Staphylococcus aureus and E. coli.

Microscopic methods (light and electron microscopy, histochemistry, immunohistochemistry) have been used to assess previously unknown pulmonary inflammatory responses of specific pathogen-free (SPF) mice secondary to infection via the nares by group A, type 50, streptococci suspended in saline ("strep group mice"). As controls for the strep group mice, the animals were either injected with saline alone via nares (no lesions were seen), or with Staphylococcus aureus in saline ("staph group mice") or with E. coli ("E. coli group mice"). The three different bacterial species caused clearly different histological changes in the lung. In the strep group mice, the microscopic findings were consistent with the diagnosis of lymphocytic interstitial pneumonia of bronchiolovascular bundles, secondary to exaggerated pulmonary recirculation of lymphocytes, concomitant with vasoconstrictive angiopathy of encased pulmonary artery branches and nodular inflammatory cell aggregates in lung parenchyma. These aggregates either consisted predominantly of lymphocytes, or of mixed cells (neutrophils, lymphocytes, macrophages) or of activated macrophages only. In 18 of 22 inflamed lungs of strep group mice, no bacteria could be cultured from lung tissue. In staph group mice the microscopic findings are consistent with the diagnosis of lymphocytic interstitial pneumonia of bronchiolovascular bundles, secondary to exaggerated pulmonary recirculation of lymphocytes only. In 12 of 17 inflamed lungs of staph group mice, no bacteria could be cultured from lung tissue. In E. coli group mice the microscopic findings were consistent with the diagnosis of distal terminal bronchiolitis and early pleural-based pneumonitis, in which lymphocytes and neutrophils mingled with macrophages. In 10 of 11 inflamed lungs of E. coli group mice, no bacteria could be cultured from lung tissue. The morphologic approaches described here may have potential for unravelling the complex inflammatory processes underlying different forms of interstitial and parenchymal pneumonia.

Correlation of dynamic contrast-enhanced MR imaging with histologic tumor grade: comparison of macromolecular and small-molecular contrast media.

OBJECTIVE: The endothelial integrity of microvessels is disrupted in malignant tumors. Quantitative assays of tumor microvascular characteristics based on dynamic MR imaging were correlated with histopathologic grade in mammary soft-tissue tumors. MATERIALS AND METHODS: A spectrum of tumors, benign through highly malignant, was induced in 33 female rats by administration of N-ethyl-N-nitrosourea, a potent carcinogen. Dynamic contrast-enhanced MR imaging was performed using a small-molecular contrast medium (gadopentetate, molecular weight = 0.5 kDa) and a macromolecular contrast medium (albumin-(Gd-DTPA)30, molecular weight = 92 kDa) at an interval of 1-2 days. Permeability surface area product (PS), as estimated by the corresponding endothelial transfer coefficient (K(PS)), and fractional plasma volume (fPV) were calculated for each tumor and each contrast agent using a two-compartment bidirectional kinetic model. MR imaging microvascular characteristics were correlated with histopathologic tumor grade. RESULTS: Tumor permeability to macromolecular contrast medium, characterized by K(PS), showed a highly positive correlation with tumor grade (r2 = .76, p < 10(-10)). K(PS) values were zero for all benign and some low-grade carcinomas, greater than zero in all other carcinomas, and increased in magnitude with higher tumor grade. A considerably smaller but significantly positive correlation was found between fPV and tumor grade using macromolecular contrast medium (r2 = .25, p < .003). No correlation between K(PS) or fPV values and tumor grade was found using gadopentetate (r2 = .01, p > .95 and r2 = .03, p > .15, respectively). CONCLUSION: Quantitative tumor microvascular permeability assays generated with macromolecular MR imaging contrast medium correlate closely with histologic tumor grade. No significant correlation is found using small-molecular gadopentetate.

Histologic confirmation of microvascular hyperpermeability to macromolecular MR contrast medium in reperfused myocardial infarction.

A macromolecular MR contrast medium (MMCM) designed to permit histochemical staining and specific tissue localization, albumin-(biotin)10-(Gd-DTPA)25 (Bio-Alb-Gd), was used in a rat model of reperfused myocardial infarction to confirm the presence and distribution of microvascular hyperpermeability. T1-weighted spin-echo images were acquired before and after administration of Bio-Alb-Gd. An avidin-biotin-complex (ABC) stain, specific for the biotinylated MR contrast medium, was used to define the MMCM distribution and to detect any regional change in microvascular permeability related to infarction. Immediately after Bio-Alb-Gd administration, the infarcted region was enhanced, with greatest signal intensity noted at the rim and less at the center. There was a gradual increase in signal intensity of the initially hypointense central region. The steady increase in signal intensity of the central region suggested convection transport of MMCM through the interstitial space and its influx into cellular compartment after leakage from the vascular compartment. Histologic findings confirmed regional microvascular hyperpermeability corresponding to the site of infarction and a predominant rim distribution of the MMCM. Bio-Alb-Gd was identified at high microscopic power in the intravascular, interstitial, and intracellular spaces at the periphery of reperfused infarcted myocardium. Bio-Alb-Gd can be used as an MR contrast medium in reperfused infarcted myocardium to confirm the existence and to localize altered microvascular permeability to macromolecules. Bio-Alb-Gd contrast technique removes all the ambiguity between the distribution of the MR or other imaging contrast agent and the distribution of the substrate for histochemical staining.

Disseminated neocortical and subcortical encephalopathy (DNSE) with widespread activation of brain macrophages: a new dementia disorder? Autopsy reports of two postmenopausal women from families with mitochondrial DNA mutations.

In this study we present 2 postmenopausal women who showed clinical symptoms that resembled those of a rather well-defined group of vascular dementia disorders, termed subcortical dementia (Binswanger disease, CADASIL). Patient 1 exhibited mitochondrial DNA (mtDNA) variants in the ND5 gene at position 13,708 and the Cytb gene at position 15,257. These DNA variants have been described in a number of neurologic disorders, but their pathogenetic potential is unclear. Patient 2 showed the same DNA alterations and an additional mtDNA variant at position 15,812 in the Cytb gene. The principal neurohistologic features of the 2 atrophic brains presented here include: subtotal selective neuronal cell loss in the cortex and, to a lesser degree, in the basal ganglia (claustrum, putamen, globus pallidus), sparing palaeocortex and periarchaeocortex, and a very characteristic and diagnostic feature was detachment of astrocytic processes from capillary walls resulting in pericapillary space formation. These pericapillary spaces were partially filled with macrophages. The spaces were not associated with total breakdown of the blood vessel walls as demonstrated by the absence of erythrocytes, lymphocytes, or polymorphonuclear leukocytes outside the vascular bed of the brain; progressive subcortical encephalopathy, as it is seen in subcortical dementia (Binswanger), but lacking arterial lipohyalinosis. The cerebral grey and white matter revealed cuffing of arteries and arterioles by adventitial macrophages. The neocortical and subcortical changes were accompanied by myriads of activated macrophages filled with lipids. The pathology of our 2 cases differs from that of other neurodegenerative disorders and we suggest the term of "disseminated neocortical and subcortical encephalopathy (DNSE) with widespread activation of brain macrophages".

Correlation of dynamic contrast-enhanced magnetic resonance imaging with histologic tumor grade: comparison of macromolecular and small-molecular contrast media.

BACKGROUND: The endothelial integrity of microvessels is disrupted in malignant tumors. Quantitative assays of tumor microvascular characteristics based on dynamic magnetic resonance imaging (MRI) were correlated with histopathologic grade in mammary soft tissue tumors. MATERIALS AND METHODS: A spectrum of tumors, benign through highly malignant, was induced in 33 female rats by administration of N -ethyl-N -nitrosourea (ENU), a potent carcinogen. Dynamic contrast-enhanced MRI was performed using a small-molecular contrast medium [gadopentetate, MW = 0.5 kDa] and a macromolecular contrast medium [albumin-(Gd-DTPA)30, MW = 92 kDa] at an interval of 1-2 days. Permeability surface area product (PS), as estimated by the corresponding endothelial transfer coefficient (KPS), and fractional plasma volume (fPV) were calculated for each tumor and each contrast agent using a two-compartment bi-directional kinetic model. MRI microvascular characteristics were correlated with histopathologic tumor grade. RESULTS: Tumor permeability to macromolecular contrast medium, characterized by KPS, showed a highly positive correlation with tumor grade (r 2 = 0.76, P < 10(-10)). KPS values were zero for all benign and some low-grade carcinomas, greater than zero in all other carcinomas, and increased in magnitude with higher tumor grade. A considerably smaller but significantly positive correlation was found between fPV and tumor grade using macromolecular contrast medium (r 2 = 0.25, P < 0.003). No correlation between KPS or fPV values and tumor grade was found using gadopentetate (r 2 = 0.01, P > 0.95 and r2 = 0.03, P > 0.15, respectively). CONCLUSION: Quantitative tumor microvascular permeability assays generated with macromolecular MRI contrast medium correlate closely with histologic tumor grade. No significant correlation is found using small-molecular gadopentetate.

Osteogenic sarcoma: noninvasive in vivo assessment of tumor necrosis with diffusion-weighted MR imaging.

PURPOSE: To evaluate diffusion-weighted magnetic resonance (MR) imaging for detecting tumor necrosis in an animal model of osteogenic sarcoma. MATERIALS AND METHODS: Twelve rats with osteogenic sarcoma underwent T1-weighted unenhanced and gadolinium-enhanced spin-echo and diffusion-weighted spin-echo MR imaging. Histologic correlation was performed. Signal intensities, T2 relaxation times, normalized apparent diffusion coefficients, and relative signal intensity increases were calculated. RESULTS: On diffusion-weighted images, necrotic tumor showed low signal intensity (mean normalized apparent diffusion coefficient, 0.46 +/- 0.20 [1 standard deviation]), indicating rapid diffusion of water molecules as a result of loss of membrane integrity, while viable tumor showed high signal intensity (mean normalized apparent diffusion coefficient, 0.16 +/- 0.05; P < .0001). Differences in the T2 relaxation times and relative signal intensity increases between viable and necrotic tumor were not statistically significant. CONCLUSION: Normalized apparent diffusion coefficients are more accurate in differentiating between viable and necrotic tumor than are T2 relaxation times or relative signal intensity increases on contrast-enhanced images. Signal intensity overlap between viable and necrotic tumor on gadolinium-enhanced images may be caused by the small molecular size of the agent, which permeates the interstitial space freely, thereby also enhancing necrosis. Diffusion-weighted MR imaging depicts differences in diffusion and, ultimately, in membrane integrity between viable and necrotic tumor and may be used to monitor tumor viability during treatment.

Definition of liver tumors in the presence of diffuse liver disease: comparison of findings at MR imaging with positive and negative contrast agents.

PURPOSE: The potential to define liver tumors at magnetic resonance (MR) imaging was compared with a positive and a negative contrast agent (gadoxetic acid disodium, or gadolinium EOB-DTPA [a hepatocyte-directed agent], and ferumoxides, or superpara-magnetic iron oxide particles [a Kupffer cell-directed agent], respectively) in normal rats and in rats with induced acute hepatitis, fatty liver, or cirrhosis. MATERIALS AND METHODS: Rats with implanted liver adenocarcinomas were divided into four groups: no diffuse liver disease ("normal" [n = 6]) and diffuse liver diseases (induced acute hepatitis [n = 6], fatty liver [n = 6], or cirrhosis [n = 6]). Rats first received gadoxetic acid disodium (50 mumol/kg) and then, 45 minutes later, ferumoxides (10 mumol/kg). Liver signal intensity enhancement and tumor-to-liver contrast-to-noise ratio (C/N) were measured in each group. RESULTS: Mean liver signal intensity enhancement values with gadoxetic acid disodium and ferumoxides were excellent in the normal liver model (176% and -62%, respectively; P < .01) but were significantly reduced in the acute hepatitis model (82% and -36%, respectively). In the fatty livers compared with the normal livers, enhancement with gadoxetic acid disodium was reduced (57%) but with ferumoxides was excellent (-55%). In the cirrhotic livers compared with the normal livers, enhancement with gadoxetic acid disodium (174%) was virtually the same but was impaired with ferumoxides (-43%). CONCLUSION: Hepatic enhancement and tumor-to-liver C/N with either positive or negative liver-enhancing agents can be impaired by the presence of underlying liver disease. Prior knowledge of the type of diffuse liver disease may influence the choice of contrast agent for tumor detection.

[Macromolecular contrast media for MR mammography. A new approach to characterizing breast tumors]. / Makromolekulare Kontrastmittel für die MR-Mammographie. Ein neuer Ansatz für die Charakterisierung von Mammatumoren.

The value of macromolecular contrast agents (MMCM) for the characterization of benign and malignant breast tumors will be demonstrated in this review. Animal studies suggest a high potential of MMCM to increase the specificity of MR-mammography. The concept of tumor differentiation is based on the pathological hyperpermeability of microvessels in malignant tumors. MMCM show a leak into the interstitium of carcinomas, whereas they are confined to the intravascular space in benign tumors. Capabilities and limitations of the MMCM-prototype. Albumin-Gd-DTPA, for breast tumor characterization will be summarized and compared to the standard low molecular weight contrast agent Gd-DTPA. Initial experience with new MMCM, such as Dendrimers, Gd-DTPA-Polylysine and MS-325 will be outlined. The potential of "blood-pool"-iron oxides, such as AMI-227 for the evaluation of tumor microvascular permeabilities will be discussed.

Monitoring neovascularity as an indicator to response to chemotherapy in osteogenic and Ewing sarcoma using magnetic resonance angiography.

Histologic studies on resected specimen have shown that tumor neovascularity is related to prognosis and response to therapy in a variety of human neoplasms. In nine patients with osteogenic or Ewing sarcoma, we evaluated the use of magnetic resonance angiography (MRA) to assess neovascularity non-invasively in vivo and to monitor response to chemotherapy. Seven patients with osteosarcoma and two patients with Ewing sarcoma were studied before and after chemotherapy by MRA (2-D time-of-flight gradient-echo sequence, TR = 50 msec, TE = 9.5 msec, theta = 50 degrees, acquisition time 13 min). MR angiograms were assessed for chemotherapy-induced changes in neovascularity. MRA showed both feeder vessels and neovascularity. Six patients responded to chemotherapy ( > or = 90% histologic tumor necrosis). MRA demonstrated marked reduction in neovascularity in all responders. Three patients did not respond to chemotherapy ( < 90% histologic tumor necrosis). MRA demonstrated persistent or increased neovascularity in the non-responders. MRA provides a unique opportunity to study tumoral neovascularity noninvasively in vivo and helps to assess response to chemotherapy in patients with osteogenic or Ewing sarcoma. These general principles may be applicable to other human tumors.

Early detection of acute tubular injury with diffusion-weighted magnetic resonance imaging in a rat model of myohemoglobinuric acute renal failure.

We evaluated the feasibility of magnetic resonance imaging (MRI) for early detection of tubular injury by monitoring changes in the apparent diffusion coefficient (ADC) of renal water in a rat model of myohemoglobinuric glycerol-induced acute renal failure (ARF). Diffusion-weighted MRI was performed concurrently with measurements of serum creatinine and blood urea nitrogen (BUN), evaluation of renal perfusion with dynamic contrast-enhanced MRI, and renal morphological examination. ADC values in the cortex and outer medulla significantly declined within minutes after the glycerol administration (70-75% of control at 4 min and 50-60% of control at 15 min). Contrast-enhanced MRI demonstrated renal hypoperfusion at 20 min after the onset of injury. Light microscopy showed normal glomeruli and edematous tubular epithelial cells at 10 and 30 min, with more severe swelling and protein casts at 30 min. No changes in serum creatinine or BUN levels were detected. We hypothesize that decrease in renal ADC may be attributed to renal ischemia and to subsequent intracellular accumulation of diffusion-restricted water. Similar imaging evaluation in other experimental models of ARF, and in patients, will define the diagnostic value of renal ADC changes in early detection of acute tubular injury.

Musculoskeletal neoplasm: perineoplastic edema versus tumor on dynamic postcontrast MR images with spatial mapping of instantaneous enhancement rates.

PURPOSE: To evaluate the utility of fast, contrast-enhanced, sequential magnetic resonance (MR) imaging in differentiating between extraosseous tumor and perineoplastic edema. MATERIALS AND METHODS: Fourteen patients underwent sequential MR imaging (3.5 seconds per image) after bolus administration of gadopentetate dimeglumine. Initial rates of enhancement (initial slope) were calculated on a pixel-by-pixel basis and displayed as a "slope image"' in which pixel intensity reflected the slope value. Close correlation with wedge biopsy specimens was performed. RESULTS: Mean initial slope values were viable extraosseous tumor, 9.33 (standard deviation, 2.23); infiltrated muscle, 9.07 (2.31); edematous muscle without tumor infiltration, 5.48 (1.27); normal muscle, 1.11 (0.65). Differences in initial slope between all neoplastic and nonneoplastic tissues were statistically significant. Within individual patients, initial slope of edematous muscle was always 20% or more lower than that of neoplastic tissue. Slope images highlighted areas of viable extraosseous tumor and infiltrated muscle against edematous and normal tissues. CONCLUSION: Computer-generated slope images derived from sequential postcontrast MR images allow differentiation between tumor and nonneoplastic edema and may thereby guide the surgeon in planning limb-sparing procedures.

[CT and MRI in elastofibroma. A rare, benign soft tissue tumor]. / CT und MRT bei Elastofibrom. Ein seltener, gutartiger Weichteiltumor.

Elastofibroma is a rare, benign soft-tissue tumor that occurs along the chest wall in the periscapular region. We identified five patients with a surgical diagnosis of elastofibroma who had had CT and/or MR scans. CT and MR imaging were compared to histologic findings in two patients. CT was available in three patients and demonstrated periscapular soft-tissue tumor with an attenuation similar to that of adjacent muscle. In one patient, the tumors appeared homogeneous. In two patients, the tumor showed strands of low density similar to that of adjacent subcutaneous fat on CT. On MR imaging, all five tumors showed predominantly low-signal intensity on both T1 and T2-weighted images. The tumors contained focal areas with high-signal intensity on T1-weighted scans that showed intermediate to high-signal intensity with T2-weighting. The areas of low-signal intensity on both T1 and T2-weighted images corresponded histologically to fibroelastic tissue. The areas of high signal intensity on T1-weighting and intermediate to high-signal intensity on T2-weighting reflected focal areas of fat within the tumor. Although CT and MR imaging findings are not specific, the presence of a soft-tissue tumor along the chest wall in a periscapular location that shows predominantly low intensity on all sequences reflecting fibrous tissue with interspersed focal areas of fatty tissue is highly suggestive of elastofibroma.

Primary bone tumors: value of MR angiography for preoperative planning and monitoring response to chemotherapy.

OBJECTIVE: The purposes of our study were to investigate the use of MR angiography with two- (2D) and three-dimensional (3D) displays in evaluating vascular morphology of musculoskeletal neoplasms for preoperative planning of limb-salvage surgery and to assess the use of MR angiography for monitoring changes in neovascularity and evaluating response to chemotherapy. SUBJECTS AND METHODS: We used MR angiography (2D time-of-flight) to study 13 patients with primary bone tumors (nine osteogenic sarcomas, two Ewing's sarcomas, and two primary lymphomas of bone) at the time of initial presentation. Eight patients (all of whom had osteogenic sarcoma) also underwent MR angiography following chemotherapy before limb-salvage surgery. Two-dimensional maximum intensity projections were obtained. Three-dimensional reconstructions of vascular structures were created from the angiographic source images and were displayed simultaneously with 3D reconstructions of tumor and normal bone generated from conventional MR images. RESULTS: Two-dimensional maximum intensity projections were useful for evaluating small vessel neovascularity; 3D displays demonstrated spatial relationships of tumor, feeder vessels, and normal vascular structures. Tumor encroachment onto or encasement of normal vascular structures was shown in four patients on 2D maximum intensity projections and on 3D displays. The eight patients with osteogenic sarcoma who had follow-up imaging showed marked neovascularity prior to chemotherapy. Five patients responded to chemotherapy (> or = 90% tumor necrosis at histology); MR angiography showed marked reduction in tumor neovascularity in these patients. Three patients did not respond to chemotherapy; MR angiography showed unchanged neovascularity in one and increased neovascularity in two of these patients. CONCLUSION: MR angiography provides good visualization of peripheral vascular branches and tumor neovascularity in patients with primary bone tumors. MR angiography demonstrates encroachment onto and encasement of major vessels by the tumor mass and appears to be useful for assessing response to chemotherapy in osteogenic sarcoma and possibly other primary bone tumors by detecting treatment-induced changes in tumor neovascularity.

Cellular proliferation in prostatic adenocarcinoma as assessed by bromodeoxyuridine uptake and Ki-67 and PCNA expression.

Prostatic carcinomas vary in their biological potential, even when stratified by grade and stage. Measurement of cellular proliferation by various methods has been shown to correlate with outcome for several human cancers, including prostatic carcinoma. Uptake of bromodeoxyuridine (BrdUrd), a thymidine analogue, has been accepted as a measure of cellular proliferative rate. However, the technique is somewhat complex, requiring incubation with fresh tissue. We compared cellular proliferation as measured by BrdUrd uptake with two more simple immunohistochemical methods in 44 prostatic adenocarcinoma specimens and correlated the results with standard clinical parameters. The tissue was obtained via needle biopsy, channel transurethral resection, and radical prostatectomy. Specimens were incubated in vitro with BrdUrd and then fixed and paraffin embedded. Sections were immunohistochemically stained with antibodies to BrdUrd, proliferating cell nuclear antigen (PCNA), and Ki-67. At least 1,000 cells were scored, and a labeling index (LI) was calculated (number of positive cells/total number of cells). The mean LI determined by all three indices was low (BrdUrd = 3.0, PCNA = 7.0, Ki-67 = 3.4), consistent with the knowledge that prostatic tumors grow slowly. In 36 patients who had not been treated at the time of analysis, the LI as determined by all three methods correlated well with clinical stage and pathological grade. Furthermore, the LIs discriminated between those with tumor confined to the prostate and those with extension to the seminal vesicles, lymph nodes, or bone (P = 0.003, 0.004, 0.008 for BrdUrd, PCNA, and Ki-67, respectively). The LIs for PCNA and Ki-67 correlated well with that for BrdUrd (r = 0.84; r = 0.85), while the LIs for Ki-67 and PCNA correlated slightly less well with each other (r = 0.78). PCNA and Ki-67 expression appear to reflect essentially the same biological process as BrdUrd uptake. Either can substitute for BrdUrd as a measure of cellular proliferation, and Ki-67 seems to offer the fewest technical problems.

Radiation-induced telangiectasia in the brain simulates cryptic vascular malformations at MR imaging.

PURPOSE: To analyze the magnetic resonance (MR) imaging, clinical, and pathologic features of radiation-induced telangiectasia of the brain. MATERIALS AND METHODS: The clinical and radiation therapy records were reviewed of 20 patients who developed focal hypointense lesions on T2-weighted MR images obtained after radiation therapy of the central nervous system. Pathologic material was reviewed in six patients. RESULTS: Eleven patients had solitary lesions, and nine had multiple foci on MR images. Fourteen of the 20 patients were less than 20 years old. The appearance ranged from small hypointense foci to larger regions of acute hemorrhage. Hematomas occurred at the site of a previously identified focus of T2 shortening in five patients. Pathologic findings included ectatic thin-walled vessels surrounded by hemosiderin and gliosis, with minimal evidence of necrosis. CONCLUSION: Radiation-induced telangiectasia in the brain results in varying amounts of hemorrhage and, occasionally, parenchymal hematomas, and may appear similar to cryptic vascular malformations on T2-weighted MR images.