- Email: [email protected]
Diffusion-Weighted Imaging Characteristics of Primary Central Nervous System Germinoma with Histopathologic Correlation
Diffusion-Weighted Imaging Characteristics of Primary Central Nervous System Germinoma with Histopathologic Correlation: A Retrospective Study1 Annette C. Douglas-Akinwande, MD, Jun Ying, PhD, Zahirabbas Momin, MD, Amr Mourad, MD, Eyas M. Hattab, MD
Rationale and Objectives. The aim of this study was to quantify, using diffusion-weighted magnetic resonance imaging, the microscopic rate of water diffusion in pure germinoma and to determine whether or not the apparent diffusion coefficient (ADC) values correlated with the different histologic components. Materials and Methods. A retrospective analysis of echoplanar diffusion-weighted magnetic resonance images was conducted on 10 patients with 11 germinoma lesions. All images were obtained using 1.5-T magnets with a b value of 1000 s/mm2. Regions of interest were drawn separately within the solid and the cystic or necrotic components of each germinoma, as well as within the normal gray and white matter of the respective cases, to calculate ADCs. The diffusion characteristics of the germinomas were assessed using mean and normalized ADC values. Histologic samples from all cases were blindly reviewed and then correlated with the ADC values. Results. Data are expressed as mean standard error. Evaluation of the solid components revealed that 36% of germinomas (4 of 11) had predominantly restricted diffusion (ADCsolid, 694.71 106 74.54 106 s/mm2; ADC ratio, 0.84 0.07) compared to normal brain. The majority (55% [6 of 11]) had normal diffusion (ADCsolid, 947.64 106 54.38 106 s/mm2; ADC ratio, 1.14 0.10). Only 9% (1 of 11) had increased diffusion (ADCsolid, 1172.30 106 48.52 106 s/mm2; ADC ratio, 1.67 0.16). The cystic and necrotic components had a mean ADC ratio of 2.55 0.25. There was no significant correlation between the histologic components and the ADC values of germinomas. Conclusions. The vast majority of germinomas demonstrated predominantly restricted (36%) or normal (55%) diffusion. The histologic components were not correlated with the ADC values. Key Words. Germinoma; germ cell tumor; diffusion; central nervous system; magnetic resonance (MR); mean diffusivity. ª AUR, 2009
Acad Radiol 2009; 16:1356–1365 1
From the Department of Radiology (A.C.D.-A., J.Y., A.M.), Indiana University School of Medicine, 550 North University Boulevard, University Hospital 0279, Indianapolis, IN 46202; Department of Pathology and Laboratory Medicine (E.M.H.), Indiana University School of Medicine, CPL 4000, Indianapolis, IN and the Department of Radiology, Emory University, Atlanta, GA (Z.M.). Received April 3, 2009; accepted May 29, 2009. This paper was presented in part at the 90th Scientific Assembly and Annual Meeting of the Radiological Society of North America, Chicago, IL, November 28 to December 3, 2004. Address correspondence to: A.C.D.-A. e-mail: [email protected]
ª AUR, 2009 doi:10.1016/j.acra.2009.05.004
1356
Primary intracranial germinomas are rare tumors of children and young adults (1). They make up the bulk of central nervous system (CNS) germ cell tumors and account for approximately 0.06% to 2.24% of all intracranial neoplasms (1,2). Like other germ cell tumors, germinomas are situated in the midline, with the majority of lesions occurring in the pineal and suprasellar regions. Histologically, they are characterized by variable proportions of cellular sheets or lobules of uniform germinoma cells with large round nuclei, prominent nucleoli, and clear cytoplasm interspersed by septal bands of connective tissue rich in capillaries, lymphocytes, and at times granulomas (1,3). Cystic degeneration
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
and necrosis may be features of germinoma. Necrosis can be in the form of single cells, small microscopic foci, or in the most extreme cases, geographic zones. Magnetic resonance (MR) imaging plays a vital role in the diagnosis of germ cell tumors (1,4–9). Diffusion-weighted (DW) imaging is a fairly recently developed MR sequence that has been used for the evaluation of neurologic diseases (10–12). With fast echoplanar technology, DW imaging has become widely clinically available and is part of the routine brain imaging protocol at our institution. It is highly resistant to motion, and imaging times are <30 seconds. DW imaging provides image contrast that is dependent on the molecular (Brownian) motion of water (10–12). The apparent diffusion coefficient (ADC) is a rotationally invariant measurement of the amount of total diffusion within a given tissue (11,13), and quantification of the degree of water motion requires ADC maps (10,12). The rate of molecular water diffusion as depicted on DW imaging may be altered in various diseases (10,12). Although DW imaging has had its major impact in the evaluation of hyperacute infarctions (12,14), it is assuming an increasingly important role in the evaluation of other intracranial pathologies, such as infections (15–19) and neoplasms (15,20). DW imaging has been proposed as a mechanism for producing contrast in the demarcation of different brain neoplasms (21–24). Recent studies have demonstrated hyperintense DW imaging signal in certain primary CNS neoplasms, such as lymphoma, pineoblastoma, medulloblastoma, high-grade gliomas (17,19,20,22), and two germinoma cases (25). However, the single study that investigated germinoma included only two cases and did not report ADC signal or value. The purpose of this study was to quantify the diffusivity of germinoma as reflected by ADC values and to determine if the ADC values correlated with the various histologic components. We hypothesized that the ADC values would correlate with the different histologic components.
MATERIALS AND METHODS Patients The archival pathology files and tumor registry of this medical center were searched over a 6-year period for biopsyproven germinomas. This study was approved by the institutional review board, and the requirement for informed consent was waived. Inclusion criteria were (1) histopathologic confirmation of the diagnosis of pure germinoma, (2) the availability of DW images in electronic format prior to any treatment, (3) the ability to generate ADC maps, and (4) the ability to calculate ADC values. Exclusion criteria were (1) lesions < 1 cm3 situated along the hypothalamic-hypophyseal axis and (2) the presence of susceptibility artifacts on the DW sequences.
A search of the radiology database yielded 15 patients in whom DW imaging was performed. Five patients were excluded. In three patients, the images were not in electronic format, and ADC values could not be retrospectively generated; in one patient, susceptibility artifacts from a shunt catheter were present; and in another, a hypothalamic-hypophyseal axis lesion was <1 cm3. There were therefore 10 patients in whom trace DW imaging, ADC maps, and ADC values were available for analysis. Of these, two had synchronous lesions situated in the pineal region and along the hypothalamic-hypophyseal axis, resulting in a total of 12 lesions. However, one of the synchronous lesions situated along the hypothalamic-hypophyseal axis measured <1 cm3 and was excluded. Consequently, we studied 11 lesions for which DW images, ADC maps, and ADC values were available. Three lesions (27%) were located in the pineal region, and eight (73%) resided along the hypothalamic-hypophyseal axis. The study population of 10 patients comprised an equal number of male and female patients (mean age, 11 years; range, 5-22 years). Imaging Protocol All studies were performed using 1.5-T GE Genesis Signa magnets (GE Healthcare, Milwaukee, WI) equipped with a head coil. A typical examination consisted of T1-weighted (repetition time [TR], 450 ms; echo time [TE], 14 ms), T2-weighted (TR, 3000 ms; TE, 105 ms), and single-shot echoplanar DW images (TR, 7000 ms; TE, 118 ms). The DW images of the whole brain were obtained in three orthogonal directions, with b values of 0 and 1000 s/mm2 (field of view, 220 220 mm; slice thickness, 5 mm; interslice gap, 0.6 mm; matrix size, 256 256 [n = 5] or 128 128 [n = 6]; number of signals acquired, 1). T1-weighted (TR, 450 ms; TE, 14 ms) and/ or magnetization transfer (TR, 500 ms; TE, 9 ms; flip angle, 90 ) images were obtained after the infusion of gadopentate dimeglumine (Magnevist; Berlex Laboratories, Wayne NJ). In addition to these routine sequences, high-resolution precontrast and postcontrast T1-weighted (TR, 550 ms; TE, 20 ms) sagittal and coronal images (field of view, 160 160 mm; slice thickness, 3 mm; interslice gap, 0.3 mm; number of signals acquired, 1) were obtained for lesions located along the hypothalamic-hypophyseal axis. Images were in electronic format on a picture archiving and communication system (GE Healthcare) and were transmitted to an Advantage workstation (GE Healthcare), on which ADC maps were generated. Quantitative Analysis of the ADC Maps for Each Germinoma The trace DW imaging and ADC maps were loaded into ImageJ, a medical image-processing software package (National Institutes of Health, Bethesda, MD). All images were interpolated to a 768 768 matrix to ensure the visualization of finer details of the lesions. Because the DW images were
1357
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Figure 1. Placement of regions of interest (ROIs) in the solid and cystic and necrotic portions of a germinoma and in normal gray and white matter. (a) Transverse apparent diffusion coefficient (ADC) maps in the pineal region demonstrating ROI placements in the solid component of a germinoma located in the pineal region. (b) Transverse ADC map from a different patient with cystic and necrotic suprasellar germinoma demonstrating the placement of ROIs in the cystic and necrotic components and (c) in normal gray and white matter.
acquired with two different matrices (128 128 and 256 256), two circular region-of-interest (ROI) sizes were used (11.83 and 2.95 mm2, respectively, corresponding to 2 2 pixels on each original image). By so doing, the size of each ROI (2 2 pixels) was maintained uniform in context with the original uninterpolated images. The total number of ROIs for each lesion varied directly with the size and shape of the germinoma. All ROI placements were performed by a single neuroradiologist. On imaging, cystic and necrotic foci were defined as regions in which there were no enhancement and the signal intensities were similar to cerebrospinal fluid (CSF) on all imaging sequences. First, ROIs were drawn only within the solid components of the tumor. This was achieved by correlating the isointense or hypointense T2 signal and the enhancing portions of germinomas with the DW imaging and ADC maps while avoiding the cystic and necrotic components, peritumoral edema, and adjacent CSF (Fig 1a). Similarly, ROIs were placed only in the cystic and necrotic components of each lesion while avoiding the solid portions and the adjacent CSF (Fig 1b). The ADC values of normal gray and white matter were used as reference. These values were measured by the random placement of 10 standardized ROIs in normal gray and white matter while avoiding the CSF and the germinoma on each image with a germinoma (Fig 1c). The mean ADC values for the solid, cystic and necrotic components, and uninvolved brain for each lesion were used for statistical analysis. Analysis of Histopathology In one patient in whom synchronous lesions were found in the pineal and suprasellar regions, a biopsy specimen was obtained from the pineal region only. Therefore, there were a total of 10 biopsy specimens for 11 radiographic lesions. All histologic samples were reviewed by one neuropathologist, who was blinded to the diffusion patterns and ADC values. In each case, the presence of necrosis, cystic degeneration, and
1358
granuloma was noted. Necrosis was defined as the presence of more than just single-cell necrosis, while cystic degeneration was defined as the presence of small or large cystic spaces within the confines of viable tumor. However, because cystic degeneration and necrosis have identical MR imaging characteristics, they were combined for statistical analysis. Biopsy samples were additionally scored for connective tissue content, extent of lymphocytic infiltrate, and cellularity. Statistical Analysis For each germinoma, the diffusion pattern was determined by comparing the mean ADC values of the solid components of germinomas (ADCsolid) and uninvolved brain (ADCbrain) using a two-sample t test. A P value < .05 was considered to indicate statistical significance. Restricted diffusion was defined as an ADCsolid value significantly less than that of ADCbrain. Normal diffusion was defined as no significant difference between ADCsolid and ADCbrain, and increased diffusion was defined as an ADCsolid value significantly higher than that of ADCbrain. Tukey’s tests were used to perform multiple comparisons of means of the ADC values under circumstances in which fixed factors (the tissue type and diffusion pattern) in the mixed model were significant. Chi-square tests were used to investigate the associations between diffusion patterns and histologic findings. We compared the likelihood of the presence of a restricted diffusion pattern between patients with cystic degeneration or necrosis and patients without these histologic features using Fisher’s exact test. Data are expressed as mean standard error. RESULTS Solid Components of Germinoma Thirty-six percent of germinomas (4 of 11) had ADCsolid values significantly lower than those of ADCbrain, indicating
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
Figure 2. Magnetic resonance images from an 11-year-old male patient with a pinealregion germinoma show restricted diffusion. (a) Transverse T2-weighted image shows an isointense mass in the pineal region. (b) Transverse diffusion-weighted imaging sequence shows hyperintense signal. (c) Transverse apparent diffusion coefficient (ADC) map shows hypointense signal. The ADC value of this lesion was relatively reduced compared to uninvolved brain. The mean ADCsolid value of this germinoma (656.59 106 26 106 s/mm2) was significantly lower than ADCbrain (834 106 16.37 106 s/mm2) and was consistent with restricted diffusion. (d) Contrast-enhanced transverse T1-weighted image shows enhancement of the mass.
restricted diffusion (Fig 2). Most (55% [6 of 11]) had ADCsolid values not significantly different from ADCbrain, indicating normal diffusion (Fig 3). One germinoma (9%) showed increased diffusion, with an ADCsolid value significantly higher than the ADCbrain value but significantly lower than the ADCcystic value (Fig 4). These findings are summarized in Table 1.
Cystic and Necrotic Components of Germinomas The ADC values of the cystic and necrotic components in all germinoma cases were similar (P > .05), regardless of the diffusion patterns observed in the solid components (Table 2). The mean ADC value of the cystic and necrotic components of germinomas (2124.78 106 210.99 106 s/mm2) was significantly larger than those of the solid
1359
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Figure 3. Magnetic resonance images from a 9-year-old female patient with an intrasellar and suprasellar germinoma show normal diffusion. Transverse T2-weighted image (a), diffusion-weighted (DW) imaging (b), apparent diffusion coefficient (ADC) maps (c), and contrast-enhanced coronal T1-weighted images (d) from a germinoma in the sellar and suprasellar region shows a germinoma with normal diffusion. The solid components have isointense signal on T2-weighted imaging, DW imaging, and ADC maps. The mean ADC value of this germinoma (843.03 106 26 106 s/mm2) is not significantly different from normal brain ADCbrain (844.00 106 16.37 106 s/mm2).
components as well as uninvolved brain. The mean ADCcystic/ADCbrain ratio was 2.55 0.25. Normal Gray and White Matter (Uninvolved Brain) ADC values from normal gray and white matter were indistinguishable for all the germinoma cases, regardless of the diffusion pattern exhibited in any given case (Table 2), and ranged from 809.72 106 to 870.99 106 s/mm2 (mean, 15.37 106 29.27 106 s/mm2; Table 1). The mean ADCbrain value was 833.82 106 5.32 106 s/mm2.
1360
Histologic Analysis of Germinoma The diagnosis of germinoma (Fig 5) was confirmed in all 10 patients. Table 3 summarizes the relationship between the ADC values and the histologic findings. No correlation was detected between the presence of cystic degeneration or necrosis, cellularity, or connective tissue content and the mean ADCsolid value of germinomas. Statistically, there may be a weak correlation between germinomas with restricted diffusion and the presence of lymphocytic infiltrate and granuloma.
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
Figure 4. Magnetic resonance images from a 7-year-old female patient with a suprasellar germinoma show relatively increased diffusion in the solid component. (a) Axial T2-weighted image shows an isointense mass in the suprasellar cistern. (b) The axial diffusion-weighted imaging sequence shows hypointense signal, and the axial apparent diffusion coefficient (ADC) map (c) shows mixed isointense and hyperintense signal. The ADC value of this lesion was relatively increased compared to uninvolved brain, despite effort to avoid the cystic and necrotic components. The mean ADCsolid value of this germinoma (1172.3 106 48.52 106 s/mm2) was significantly higher than the ADCbrain value (845.66 106 10.72 106 s/mm2) and was consistent with increased diffusion.
DISCUSSION The imaging characteristics of primary intracranial germinoma on conventional MR have been extensively described (5–9,26–30). Unfortunately, these imaging findings are nonspecific and may be shared by a variety of benign
and malignant conditions. The distinction among these lesions is important, because the management and prognosis often differ. Therefore, the search for additional imaging modalities and testing of new ones is constantly needed. DW imaging is a technique that has been used to evaluate gliomas and lymphomas (20,21,23–25). It is thought to
1361
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Table 1 Summary of ADC Values ( 106 s/mm2) by Individual Cases
Case
Germinoma Solid Component
Uninvolved Brain
Germinoma Cystic/ Necrotic Component
P*
Diffusion Patterny
Solid/Uninvolved Brain
Cystic/Necrotic/ Uninvolved Brain
1 2 3 4 5 6 7 8 9 10 11
747.35 26.00 667.66 30.56 656.89 14.97 714.34 8.24 984.96 67.63 944.69 68.90 849.43 31.92 843.03 13.69 794.16 15.95 850.90 20.57 1172.30 48.52
809.72 15.37 820.18 9.52 833.71 18.28 834.14 16.25 819.85 9.83 849.57 22.15 870.99 29.27 844.00 13.33 823.11 12.07 821.14 13.34 845.66 10.72
920.35 0.66 1296.88 156.60 1478.06 46.03 NA 2833.77 169.80 2545.19 264.61 NA 2330.39 138.98 2540.03 32.63 2726.29 215.22 2452.06 210.18
.0473 .0009 <.0001 <.0001 .0705 .0934 .6518 .9599 .1724 .2289 .0008
Restricted Restricted Restricted Restricted Normal Normal Normal Normal Normal Normal Increased
0.92 0.04 0.81 0.04 0.79 0.02 0.86 0.02 1.20 0.08 1.11 0.09 0.98 0.05 1.00 0.02 0.96 0.02 1.04 0.03 1.39 0.06
1.14 0.02 1.58 0.19 1.77 0.07 NA 3.46 0.21 3.00 0.32 NA 2.76 0.17 3.09 0.06 3.32 0.27 2.90 0.25
ADC, apparent diffusion coefficient. Data are expressed as mean standard error. Standard errors in columns 7 and 8 were calculated from those in columns 2 to 4 using Tylor’s expansion method. * P values are from two-sample t tests to compare means of ADC values between columns 2 and 3; the pooled method was used if equality of variance held, and the Satterthwaite method was used otherwise. y Quantitative results are based on comparing means in columns 2 and 3 and P values in column 5. Table 2 Comparison of Apparent Diffusion Coefficient Values ( 106 s/mm2) Quantitative Diffusion Pattern* Tissue Type Solid germinoma Uninvolved brain Cystic/necrotic germinoma Solid/uninvolved brain Cystic/necrotic/uninvolved brain
Restricted 694.71 74.54 aA 828.62 8.56 dB 2574.16 125.42 eC 0.84 0.09 3.11 0.15
Normal
Increased
947.64 54.38 bD 833.74 6.46 dD 1875.84 247.16 eE 1.14 0.07 2.25 0.30
1408.44 131.64 cF 845.66 16.68 dF 2452.06 210.18 eF 1.67 0.16 2.90 0.26
Data are expressed as mean standard error. Standard errors in rows 5 and 6 were estimated from means and standard errors in rows 2 to 4 using Tylor’s expansion method. * Means followed by different lowercase letters in a row indicate a statistical difference by Tukey’s test (P < .05). Means followed by different uppercase letters in a column in rows 2 to 4 indicate a statistical difference by Tukey’s test (P < .05).
enable the demarcation of histopathologic components of gliomas that may not be clearly distinguishable on conventional MR (21,23,24). The ADC, a rotationally invariant measurement of the amount of total diffusion within a given tissue (11,13), is required to quantify the degree of water motion (diffusivity) (11,12). Additionally, it has been inferred from prior studies (20,21,23–25) that ADC values provide an indication of the relative intracellular and extracellular volumes within the tissue being studied. On the basis of the radiologic appearance, higher ADC values have been attributable to cystic foci, whereas values approaching that of normal gray and white matter were attributed to dense, highly cellular tumor tissue (21,23,24). However, the diffusivity of CNS germinomas as reflected by ADC values or maps has not been previously reported. We set
1362
out to quantify the rate of water diffusion of primary CNS germinomas and to determine if there was correlation with the histologic components.
Solid Components of Germinoma The ADC values of the solid components (range, 656.89 106 to 1172.30 106 s/mm2; standard error, 8.24 106 to 68.90 106 s/mm2) of all 11 germinoma lesions in this study were lower than the reported ADC values of enhancing high-grade gliomas (range, 900 106 to 1400 106 s/mm2) using similar techniques (20,23,24). In addition, the normalized ADC values as depicted by ADCsolid/ADCbrain ratios (Table 1) for the group of germinomas with restricted and normal diffusion were lower than
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
Germinomas with Predominantly Normal Diffusion The majority (55%) of germinomas in this series exhibited ADC values that were not significantly different from those of normal gray and white matter. However, the mean ADCsolid value for these cases (947.6410654.38106 s/mm2) was lower than the mean ADC values (1100 106 and 1600 106 s/mm2) reported for enhancing high-grade gliomas (24). The normalized ADC values as represented by the mean ADCsolid/ADCbrain ratio (1.14) was much lower than the previously reported (20) ADC ratio for high-grade astrocytoma (1.68) and very similar to the ADC ratio for CNS lymphoma (1.15). As such, our results are expected for a hypercellular neoplasm such as germinoma. Figure 5. Classic germinoma showing lobules of densely packed tumor cells interspersed by thin bands of fibrous tissue sprinkled with lymphocytes. (Hematoxylin and eosin staining; magnification, 200.)
the reported (20) ADC ratios of lymphoma, a very cellular neoplasm. Because the leading hypothesis for the rate of water diffusion in CNS neoplasms is related to cellularity (20,23,24,31), the histologically characteristic hypercellularity of all germinomas might account for the reduced diffusivity that was observed in this cohort.
Germinomas with Predominantly Restricted Diffusion The rate of water diffusion in the solid components of germinomas was significantly lower than that of uninvolved brain (P < .05) in approximately 36% of cases, and the mean ADCsolid value for these cases (694.9110674.54 106 s/mm2) was lower than the mean ADC values (1100 106 and 1600 106 s/mm2, respectively) reported for enhancing and nonenhancing high-grade gliomas (24). Moreover, the mean ADC value for this group of germinomas was lower than that for lymphoma (870 106 s/mm2) (20). Additionally, the normalized ADC values, as represented by the mean ADCsolid/ADCbrain ratio (0.84) for germinomas with restricted diffusion, was yet lower than that reported (1.15) for CNS lymphoma (20). The low ADC values found in this subset of germinomas are similar to the ADC values reported for restricted diffusion seen in acute infarctions (32) and the cystic portions of abscesses (15). This may reflect the precision with which the ADC values were obtained. Alternatively, it may indicate that correlation of the ADC values with the underlying pathology and interpretation may be more complicated than previously hypothesized. Future analysis of diffusivity with more advanced techniques such as diffusion tensor imaging in a prospective manner might help in understanding the pathophysiology of germinomas.
Germinomas with Predominantly Increased Diffusion The solid component of a single germinoma (9%) demonstrated increased rates of water diffusion (Fig 4). The mean ADCsolid value (1172.3 106 48 106 s/mm2) and the normalized ADC values (ADCsolid/ADCbrain, 1.39) were similar to prior reports for high-grade gliomas (20,21,23,24,33,34) but higher than that for lymphoma (20). This germinoma was radiologically heterogeneous, had extensive cystic and necrotic foci (Fig 4), and histologically demonstrated extensive necrosis. From a biologic perspective, this is consistent with the presence of microscopic foci of necrosis within the solid more cellular germinomas, resulting in a relatively large extracellular volume. Another possible reason for this outlier is that the moderate standard error of the mean ADCsolid value for this lesion may reflect overlap between the solid and cystic/necrotic foci in the ROI placement in this germinoma, despite efforts to avoid them. Similar issues with volume averaging and ROI placements have been previously reported and are known to possibly affect ADC measurements in routine clinical imaging(14).
Cystic and Necrotic Components of Germinoma The ADC values of the cystic and necrotic components of germinomas in this series are in agreement with those of other intracranial neoplasms (21,23,24). Likewise, in this study, the ADC values were highest in cystic and necrotic foci and were significantly higher than the ADC values of any of the solid components. The ADC values of the cystic and necrotic components of germinomas may be used to distinguish these lesions from intracranial abscess, because the reported ADC values for abscess (15,19,35) are significantly lower than those for the cystic components of tumors. The identification of cystic and necrotic foci in germinomas may have prognostic implications, as suggested by Moon et al (29), who demonstrated that the response of germinoma to radiation
1363
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Table 3 Comparison of Diffusion Pattern and Histopathologic Results Quantitative Diffusion Pattern Component of Germinoma Cystic/necrosis Connective tissue Granuloma Lymphocytic infiltrate Cellularity
Classification
Restricted
Normal
Increased
P
3 1 0 3 0 3 0 3 0 0 2 1
3 3 4 2 4 2 1 5 0 0 3 3
0 1 0 1 1 0 1 0 0 0 1 0
.55
Absent Present Low High Absent Present Mild Intense <25% 25%–50% 50%–75% >75%
therapy correlated negatively with the presence of a cystic region. Analysis of Normal Gray and White Matter The ADC values for normal gray and white matter in this series were consistent with previous reports in adult patients (24,36) and were therefore suitable reference standards. The ADCbrain values in this series also suggest that the ADCbrain values of children aged $5 years are similar to those of adults. Study Limitations It may be argued that one potential limitation of this study is that some of the lesions were located along the hypothalamic-hypophyseal axis, and DW echoplanar imaging is known to be sensitive to nonuniformities in the magnetic field. A major source of such nonuniformity is the differences in magnetic susceptibility at air-tissue and bone-tissue interfaces (24,37), such as the skull base and particularly in the sellar region. However, such susceptibility artifacts are less problematic with single-shot echoplanar techniques (12,37), which were used in this study. In addition, inclusion criteria for this study required large lesions that extended above the suprasellar cistern. Moreover, in this study, there was no difference in the DW characteristics of germinomas that were located along the hypothalamic-hypophyseal axis and the pineal region. Because of the relatively inaccessible locations of germinomas, neoplasms that are sensitive to adjuvant therapy, no attempts are usually made for complete surgical excision. Instead, the goal is to obtain an adequate biopsy specimen of the solid lesion for histopathologic diagnosis. This approach usually yields small samples that are just sufficient for diagnosis.
1364
.11 .10 .09 .61
CONCLUSIONS Germinomas had heterogeneous diffusivity; the solid components of 36%, 55%, and 9% of germinomas predominantly exhibited restricted, normal, and increased diffusivity, respectively. There was no correlation between the histologic components and the ADC values of germinomas. The ADC values may be used to target biopsy sites and to noninvasively monitor response to therapy, as suggested previously (38).
REFERENCES 1. Kleihues P, Cavenee WK. International Agency for Research on Cancer. Pathology and genetics of tumours of the nervous system. Lyon, France: IARC Press, 2000. 2. Horowitz MB, Hall WA. Central nervous system germinomas. A review. Arch Neurol 1991; 48:652–657. 3. Salzman KL, Rojiani AM, Buatti J, et al. Primary intracranial germ cell tumors: clinicopathologic review of 32 cases. Pediatr Pathol Lab Med 1997; 17:713–727. 4. Chang T, Teng MM, Guo WY, Sheng WC. CT of pineal tumors and intracranial germ-cell tumors. AJNR Am J Neuroradiol 1989; 10:1039–1044. 5. Higano S, Takahashi S, Ishii K, Matsumoto K, Ikeda H, Sakamoto K. Germinoma originating in the basal ganglia and thalamus: MR and CT evaluation. AJNR Am J Neuroradiol 1994; 15:1435–1441. 6. Kim DI, Yoon PH, Ryu YH, Jeon P, Hwang GJ. MRI of germinomas arising from the basal ganglia and thalamus. Neuroradiology 1998; 40:507–511. 7. Konno S, Oka H, Utsuki S, et al. Germinoma with a granulomatous reaction. Problems of differential diagnosis. Clin Neuropathol 2002; 21:248–251. 8. Koren A. Bifocal primary intracranial germinoma in a child. Case report. Radiol Oncol 2001; 35:179–183. 9. Ballesteros MD, Duran A, Arrazola J, et al. Primary intrasellar germinoma with synchronous pineal tumor. Neuroradiology 1997; 39:860–862. 10. Schaefer PW, Grant PE, Gonzalez RG. Diffusion-weighted MR imaging of the brain. Radiology 2000; 217:331–345. 11. Le Bihan D. Molecular diffusion nuclear magnetic resonance imaging. Magn Reson Q 1991; 7:1–30. 12. Castillo M, Mukherji SK. Diffusion-weighted imaging in the evaluation of intracranial lesions. Semin Ultrasound CT MR 2000; 21:405–416.
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
13. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, LavalJeantet M. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988; 168:497–505. 14. DeLano MC, Cooper TG, Siebert JE, Potchen MJ, Kuppusamy K. High-bvalue diffusion-weighted MR imaging of adult brain: image contrast and apparent diffusion coefficient map features. AJNR Am J Neuroradiol 2000; 21:1830–1836. 15. Cartes-Zumelzu FW, Stavrou I, Castillo M, Eisenhuber E, Knosp E, Thurnher MM. Diffusion-weighted imaging in the assessment of brain abscesses therapy. AJNR Am J Neuroradiol 2004; 25:1310–1317. 16. Guzman R, Barth A, Lovblad KO, et al. Use of diffusion-weighted magnetic resonance imaging in differentiating purulent brain processes from cystic brain tumors. J Neurosurg 2002; 97:1101–1107. 17. Hartmann M, Jansen O, Heiland S, Sommer C, Munkel K, Sartor K. Restricted diffusion within ring enhancement is not pathognomonic for brain abscess. AJNR Am J Neuroradiol 2001; 22:1738–1742. 18. Desprechins B, Stadnik T, Koerts G, Shabana W, Breucq C, Osteaux M. Use of diffusion-weighted MR imaging in differential diagnosis between intracerebral necrotic tumors and cerebral abscesses. AJNR Am J Neuroradiol 1999; 20:1252–1257. 19. Tung GA, Evangelista P, Rogg JM, Duncan JA III. Diffusion-weighted MR imaging of rim-enhancing brain masses: is markedly decreased water diffusion specific for brain abscess? AJR Am J Roentgenol 2001; 177:709–712. 20. Guo AC, Cummings TJ, Dash RC, Provenzale JM. Lymphomas and highgrade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002; 224:177–183. 21. Brunberg JA, Chenevert TL, McKeever PE, et al. In vivo MR determination of water diffusion coefficients and diffusion anisotropy: correlation with structural alteration in gliomas of the cerebral hemispheres. AJNR Am J Neuroradiol 1995; 16:361–371. 22. Eis M, Els T, Hoehn-Berlage M. High resolution quantitative relaxation and diffusion MRI of three different experimental brain tumors in rat. Magn Reson Med 1995; 34:835–844. 23. Krabbe K, Gideon P, Wagn P, Hansen U, Thomsen C, Madsen F. MR diffusion imaging of human intracranial tumours. Neuroradiology 1997; 39:483–489. 24. Tien RD, Felsberg GJ, Friedman H, Brown M, MacFall J. MR imaging of high-grade cerebral gliomas: value of diffusion-weighted echo-planar pulse sequences. AJR Am J Roentgenol 1994; 162:671–677.
25. Okamoto K, Ito J, Ishikawa K, Sakai K, Tokiguchi S. Diffusion-weighted echo-planar MR imaging in differential diagnosis of brain tumors and tumor-like conditions. Eur Radiol 2000; 10:1342–1350. 26. Edwards MS, Hudgins RJ, Wilson CB, Levin VA, Wara WM. Pineal region tumors in children. J Neurosurg 1988; 68:689–697. 27. Liang L, Korogi Y, Sugahara T, et al. MRI of intracranial germ-cell tumours. Neuroradiology 2002; 44:382–388. 28. Matsumura N, Kurimoto M, Endo S, Fukuda O, Takaku A. Intracranial germinoma associated with Down’s syndrome. Report of 2 cases. Pediatr Neurosurg 1998; 29:199–202. 29. Moon WK, Chang KH, Han MH, Kim IO. Intracranial germinomas: correlation of imaging findings with tumor response to radiation therapy. AJR Am J Roentgenol 1999; 172:713–716. 30. Osborn AG. Diagnostic imaging. Brain. Salt Lake City, UT: Amirsys, 2004. 31. Gupta RK, Sinha U, Cloughesy TF, Alger JR. Inverse correlation between choline magnetic resonance spectroscopy signal intensity and the apparent diffusion coefficient in human glioma. Magn Reson Med 1999; 41:2–7. 32. Oppenheim C, Grandin C, Samson Y, et al. Is there an apparent diffusion coefficient threshold in predicting tissue viability in hyperacute stroke? Stroke 2001; 32:2486–2491. 33. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999; 9:53–60. 34. Castillo M, Smith JK, Kwock L, Wilber K. Apparent diffusion coefficients in the evaluation of high-grade cerebral gliomas. AJNR Am J Neuroradiol 2001; 22:60–64. 35. Camacho DL, Smith JK, Castillo M. Differentiation of toxoplasmosis and lymphoma in AIDS patients by using apparent diffusion coefficients. AJNR Am J Neuroradiol 2003; 24:633–637. 36. Helenius J, Soinne L, Perkio J, et al. Diffusion-weighted MR imaging in normal human brains in various age groups. AJNR Am J Neuroradiol 2002; 23:194–199. 37. Maier SE, Bogner P, Bajzik G, et al. Normal brain and brain tumor: multicomponent apparent diffusion coefficient line scan imaging. Radiology 2001; 219:842–849. 38. Chenevert TL, McKeever PE, Ross BD. Monitoring early response of experimental brain tumors to therapy using diffusion magnetic resonance imaging. Clin Cancer Res 1997; 3:1457–1466.
1365
Rationale and Objectives. The aim of this study was to quantify, using diffusion-weighted magnetic resonance imaging, the microscopic rate of water diffusion in pure germinoma and to determine whether or not the apparent diffusion coefficient (ADC) values correlated with the different histologic components. Materials and Methods. A retrospective analysis of echoplanar diffusion-weighted magnetic resonance images was conducted on 10 patients with 11 germinoma lesions. All images were obtained using 1.5-T magnets with a b value of 1000 s/mm2. Regions of interest were drawn separately within the solid and the cystic or necrotic components of each germinoma, as well as within the normal gray and white matter of the respective cases, to calculate ADCs. The diffusion characteristics of the germinomas were assessed using mean and normalized ADC values. Histologic samples from all cases were blindly reviewed and then correlated with the ADC values. Results. Data are expressed as mean standard error. Evaluation of the solid components revealed that 36% of germinomas (4 of 11) had predominantly restricted diffusion (ADCsolid, 694.71 106 74.54 106 s/mm2; ADC ratio, 0.84 0.07) compared to normal brain. The majority (55% [6 of 11]) had normal diffusion (ADCsolid, 947.64 106 54.38 106 s/mm2; ADC ratio, 1.14 0.10). Only 9% (1 of 11) had increased diffusion (ADCsolid, 1172.30 106 48.52 106 s/mm2; ADC ratio, 1.67 0.16). The cystic and necrotic components had a mean ADC ratio of 2.55 0.25. There was no significant correlation between the histologic components and the ADC values of germinomas. Conclusions. The vast majority of germinomas demonstrated predominantly restricted (36%) or normal (55%) diffusion. The histologic components were not correlated with the ADC values. Key Words. Germinoma; germ cell tumor; diffusion; central nervous system; magnetic resonance (MR); mean diffusivity. ª AUR, 2009
Acad Radiol 2009; 16:1356–1365 1
From the Department of Radiology (A.C.D.-A., J.Y., A.M.), Indiana University School of Medicine, 550 North University Boulevard, University Hospital 0279, Indianapolis, IN 46202; Department of Pathology and Laboratory Medicine (E.M.H.), Indiana University School of Medicine, CPL 4000, Indianapolis, IN and the Department of Radiology, Emory University, Atlanta, GA (Z.M.). Received April 3, 2009; accepted May 29, 2009. This paper was presented in part at the 90th Scientific Assembly and Annual Meeting of the Radiological Society of North America, Chicago, IL, November 28 to December 3, 2004. Address correspondence to: A.C.D.-A. e-mail: [email protected]
ª AUR, 2009 doi:10.1016/j.acra.2009.05.004
1356
Primary intracranial germinomas are rare tumors of children and young adults (1). They make up the bulk of central nervous system (CNS) germ cell tumors and account for approximately 0.06% to 2.24% of all intracranial neoplasms (1,2). Like other germ cell tumors, germinomas are situated in the midline, with the majority of lesions occurring in the pineal and suprasellar regions. Histologically, they are characterized by variable proportions of cellular sheets or lobules of uniform germinoma cells with large round nuclei, prominent nucleoli, and clear cytoplasm interspersed by septal bands of connective tissue rich in capillaries, lymphocytes, and at times granulomas (1,3). Cystic degeneration
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
and necrosis may be features of germinoma. Necrosis can be in the form of single cells, small microscopic foci, or in the most extreme cases, geographic zones. Magnetic resonance (MR) imaging plays a vital role in the diagnosis of germ cell tumors (1,4–9). Diffusion-weighted (DW) imaging is a fairly recently developed MR sequence that has been used for the evaluation of neurologic diseases (10–12). With fast echoplanar technology, DW imaging has become widely clinically available and is part of the routine brain imaging protocol at our institution. It is highly resistant to motion, and imaging times are <30 seconds. DW imaging provides image contrast that is dependent on the molecular (Brownian) motion of water (10–12). The apparent diffusion coefficient (ADC) is a rotationally invariant measurement of the amount of total diffusion within a given tissue (11,13), and quantification of the degree of water motion requires ADC maps (10,12). The rate of molecular water diffusion as depicted on DW imaging may be altered in various diseases (10,12). Although DW imaging has had its major impact in the evaluation of hyperacute infarctions (12,14), it is assuming an increasingly important role in the evaluation of other intracranial pathologies, such as infections (15–19) and neoplasms (15,20). DW imaging has been proposed as a mechanism for producing contrast in the demarcation of different brain neoplasms (21–24). Recent studies have demonstrated hyperintense DW imaging signal in certain primary CNS neoplasms, such as lymphoma, pineoblastoma, medulloblastoma, high-grade gliomas (17,19,20,22), and two germinoma cases (25). However, the single study that investigated germinoma included only two cases and did not report ADC signal or value. The purpose of this study was to quantify the diffusivity of germinoma as reflected by ADC values and to determine if the ADC values correlated with the various histologic components. We hypothesized that the ADC values would correlate with the different histologic components.
MATERIALS AND METHODS Patients The archival pathology files and tumor registry of this medical center were searched over a 6-year period for biopsyproven germinomas. This study was approved by the institutional review board, and the requirement for informed consent was waived. Inclusion criteria were (1) histopathologic confirmation of the diagnosis of pure germinoma, (2) the availability of DW images in electronic format prior to any treatment, (3) the ability to generate ADC maps, and (4) the ability to calculate ADC values. Exclusion criteria were (1) lesions < 1 cm3 situated along the hypothalamic-hypophyseal axis and (2) the presence of susceptibility artifacts on the DW sequences.
A search of the radiology database yielded 15 patients in whom DW imaging was performed. Five patients were excluded. In three patients, the images were not in electronic format, and ADC values could not be retrospectively generated; in one patient, susceptibility artifacts from a shunt catheter were present; and in another, a hypothalamic-hypophyseal axis lesion was <1 cm3. There were therefore 10 patients in whom trace DW imaging, ADC maps, and ADC values were available for analysis. Of these, two had synchronous lesions situated in the pineal region and along the hypothalamic-hypophyseal axis, resulting in a total of 12 lesions. However, one of the synchronous lesions situated along the hypothalamic-hypophyseal axis measured <1 cm3 and was excluded. Consequently, we studied 11 lesions for which DW images, ADC maps, and ADC values were available. Three lesions (27%) were located in the pineal region, and eight (73%) resided along the hypothalamic-hypophyseal axis. The study population of 10 patients comprised an equal number of male and female patients (mean age, 11 years; range, 5-22 years). Imaging Protocol All studies were performed using 1.5-T GE Genesis Signa magnets (GE Healthcare, Milwaukee, WI) equipped with a head coil. A typical examination consisted of T1-weighted (repetition time [TR], 450 ms; echo time [TE], 14 ms), T2-weighted (TR, 3000 ms; TE, 105 ms), and single-shot echoplanar DW images (TR, 7000 ms; TE, 118 ms). The DW images of the whole brain were obtained in three orthogonal directions, with b values of 0 and 1000 s/mm2 (field of view, 220 220 mm; slice thickness, 5 mm; interslice gap, 0.6 mm; matrix size, 256 256 [n = 5] or 128 128 [n = 6]; number of signals acquired, 1). T1-weighted (TR, 450 ms; TE, 14 ms) and/ or magnetization transfer (TR, 500 ms; TE, 9 ms; flip angle, 90 ) images were obtained after the infusion of gadopentate dimeglumine (Magnevist; Berlex Laboratories, Wayne NJ). In addition to these routine sequences, high-resolution precontrast and postcontrast T1-weighted (TR, 550 ms; TE, 20 ms) sagittal and coronal images (field of view, 160 160 mm; slice thickness, 3 mm; interslice gap, 0.3 mm; number of signals acquired, 1) were obtained for lesions located along the hypothalamic-hypophyseal axis. Images were in electronic format on a picture archiving and communication system (GE Healthcare) and were transmitted to an Advantage workstation (GE Healthcare), on which ADC maps were generated. Quantitative Analysis of the ADC Maps for Each Germinoma The trace DW imaging and ADC maps were loaded into ImageJ, a medical image-processing software package (National Institutes of Health, Bethesda, MD). All images were interpolated to a 768 768 matrix to ensure the visualization of finer details of the lesions. Because the DW images were
1357
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Figure 1. Placement of regions of interest (ROIs) in the solid and cystic and necrotic portions of a germinoma and in normal gray and white matter. (a) Transverse apparent diffusion coefficient (ADC) maps in the pineal region demonstrating ROI placements in the solid component of a germinoma located in the pineal region. (b) Transverse ADC map from a different patient with cystic and necrotic suprasellar germinoma demonstrating the placement of ROIs in the cystic and necrotic components and (c) in normal gray and white matter.
acquired with two different matrices (128 128 and 256 256), two circular region-of-interest (ROI) sizes were used (11.83 and 2.95 mm2, respectively, corresponding to 2 2 pixels on each original image). By so doing, the size of each ROI (2 2 pixels) was maintained uniform in context with the original uninterpolated images. The total number of ROIs for each lesion varied directly with the size and shape of the germinoma. All ROI placements were performed by a single neuroradiologist. On imaging, cystic and necrotic foci were defined as regions in which there were no enhancement and the signal intensities were similar to cerebrospinal fluid (CSF) on all imaging sequences. First, ROIs were drawn only within the solid components of the tumor. This was achieved by correlating the isointense or hypointense T2 signal and the enhancing portions of germinomas with the DW imaging and ADC maps while avoiding the cystic and necrotic components, peritumoral edema, and adjacent CSF (Fig 1a). Similarly, ROIs were placed only in the cystic and necrotic components of each lesion while avoiding the solid portions and the adjacent CSF (Fig 1b). The ADC values of normal gray and white matter were used as reference. These values were measured by the random placement of 10 standardized ROIs in normal gray and white matter while avoiding the CSF and the germinoma on each image with a germinoma (Fig 1c). The mean ADC values for the solid, cystic and necrotic components, and uninvolved brain for each lesion were used for statistical analysis. Analysis of Histopathology In one patient in whom synchronous lesions were found in the pineal and suprasellar regions, a biopsy specimen was obtained from the pineal region only. Therefore, there were a total of 10 biopsy specimens for 11 radiographic lesions. All histologic samples were reviewed by one neuropathologist, who was blinded to the diffusion patterns and ADC values. In each case, the presence of necrosis, cystic degeneration, and
1358
granuloma was noted. Necrosis was defined as the presence of more than just single-cell necrosis, while cystic degeneration was defined as the presence of small or large cystic spaces within the confines of viable tumor. However, because cystic degeneration and necrosis have identical MR imaging characteristics, they were combined for statistical analysis. Biopsy samples were additionally scored for connective tissue content, extent of lymphocytic infiltrate, and cellularity. Statistical Analysis For each germinoma, the diffusion pattern was determined by comparing the mean ADC values of the solid components of germinomas (ADCsolid) and uninvolved brain (ADCbrain) using a two-sample t test. A P value < .05 was considered to indicate statistical significance. Restricted diffusion was defined as an ADCsolid value significantly less than that of ADCbrain. Normal diffusion was defined as no significant difference between ADCsolid and ADCbrain, and increased diffusion was defined as an ADCsolid value significantly higher than that of ADCbrain. Tukey’s tests were used to perform multiple comparisons of means of the ADC values under circumstances in which fixed factors (the tissue type and diffusion pattern) in the mixed model were significant. Chi-square tests were used to investigate the associations between diffusion patterns and histologic findings. We compared the likelihood of the presence of a restricted diffusion pattern between patients with cystic degeneration or necrosis and patients without these histologic features using Fisher’s exact test. Data are expressed as mean standard error. RESULTS Solid Components of Germinoma Thirty-six percent of germinomas (4 of 11) had ADCsolid values significantly lower than those of ADCbrain, indicating
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
Figure 2. Magnetic resonance images from an 11-year-old male patient with a pinealregion germinoma show restricted diffusion. (a) Transverse T2-weighted image shows an isointense mass in the pineal region. (b) Transverse diffusion-weighted imaging sequence shows hyperintense signal. (c) Transverse apparent diffusion coefficient (ADC) map shows hypointense signal. The ADC value of this lesion was relatively reduced compared to uninvolved brain. The mean ADCsolid value of this germinoma (656.59 106 26 106 s/mm2) was significantly lower than ADCbrain (834 106 16.37 106 s/mm2) and was consistent with restricted diffusion. (d) Contrast-enhanced transverse T1-weighted image shows enhancement of the mass.
restricted diffusion (Fig 2). Most (55% [6 of 11]) had ADCsolid values not significantly different from ADCbrain, indicating normal diffusion (Fig 3). One germinoma (9%) showed increased diffusion, with an ADCsolid value significantly higher than the ADCbrain value but significantly lower than the ADCcystic value (Fig 4). These findings are summarized in Table 1.
Cystic and Necrotic Components of Germinomas The ADC values of the cystic and necrotic components in all germinoma cases were similar (P > .05), regardless of the diffusion patterns observed in the solid components (Table 2). The mean ADC value of the cystic and necrotic components of germinomas (2124.78 106 210.99 106 s/mm2) was significantly larger than those of the solid
1359
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Figure 3. Magnetic resonance images from a 9-year-old female patient with an intrasellar and suprasellar germinoma show normal diffusion. Transverse T2-weighted image (a), diffusion-weighted (DW) imaging (b), apparent diffusion coefficient (ADC) maps (c), and contrast-enhanced coronal T1-weighted images (d) from a germinoma in the sellar and suprasellar region shows a germinoma with normal diffusion. The solid components have isointense signal on T2-weighted imaging, DW imaging, and ADC maps. The mean ADC value of this germinoma (843.03 106 26 106 s/mm2) is not significantly different from normal brain ADCbrain (844.00 106 16.37 106 s/mm2).
components as well as uninvolved brain. The mean ADCcystic/ADCbrain ratio was 2.55 0.25. Normal Gray and White Matter (Uninvolved Brain) ADC values from normal gray and white matter were indistinguishable for all the germinoma cases, regardless of the diffusion pattern exhibited in any given case (Table 2), and ranged from 809.72 106 to 870.99 106 s/mm2 (mean, 15.37 106 29.27 106 s/mm2; Table 1). The mean ADCbrain value was 833.82 106 5.32 106 s/mm2.
1360
Histologic Analysis of Germinoma The diagnosis of germinoma (Fig 5) was confirmed in all 10 patients. Table 3 summarizes the relationship between the ADC values and the histologic findings. No correlation was detected between the presence of cystic degeneration or necrosis, cellularity, or connective tissue content and the mean ADCsolid value of germinomas. Statistically, there may be a weak correlation between germinomas with restricted diffusion and the presence of lymphocytic infiltrate and granuloma.
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
Figure 4. Magnetic resonance images from a 7-year-old female patient with a suprasellar germinoma show relatively increased diffusion in the solid component. (a) Axial T2-weighted image shows an isointense mass in the suprasellar cistern. (b) The axial diffusion-weighted imaging sequence shows hypointense signal, and the axial apparent diffusion coefficient (ADC) map (c) shows mixed isointense and hyperintense signal. The ADC value of this lesion was relatively increased compared to uninvolved brain, despite effort to avoid the cystic and necrotic components. The mean ADCsolid value of this germinoma (1172.3 106 48.52 106 s/mm2) was significantly higher than the ADCbrain value (845.66 106 10.72 106 s/mm2) and was consistent with increased diffusion.
DISCUSSION The imaging characteristics of primary intracranial germinoma on conventional MR have been extensively described (5–9,26–30). Unfortunately, these imaging findings are nonspecific and may be shared by a variety of benign
and malignant conditions. The distinction among these lesions is important, because the management and prognosis often differ. Therefore, the search for additional imaging modalities and testing of new ones is constantly needed. DW imaging is a technique that has been used to evaluate gliomas and lymphomas (20,21,23–25). It is thought to
1361
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Table 1 Summary of ADC Values ( 106 s/mm2) by Individual Cases
Case
Germinoma Solid Component
Uninvolved Brain
Germinoma Cystic/ Necrotic Component
P*
Diffusion Patterny
Solid/Uninvolved Brain
Cystic/Necrotic/ Uninvolved Brain
1 2 3 4 5 6 7 8 9 10 11
747.35 26.00 667.66 30.56 656.89 14.97 714.34 8.24 984.96 67.63 944.69 68.90 849.43 31.92 843.03 13.69 794.16 15.95 850.90 20.57 1172.30 48.52
809.72 15.37 820.18 9.52 833.71 18.28 834.14 16.25 819.85 9.83 849.57 22.15 870.99 29.27 844.00 13.33 823.11 12.07 821.14 13.34 845.66 10.72
920.35 0.66 1296.88 156.60 1478.06 46.03 NA 2833.77 169.80 2545.19 264.61 NA 2330.39 138.98 2540.03 32.63 2726.29 215.22 2452.06 210.18
.0473 .0009 <.0001 <.0001 .0705 .0934 .6518 .9599 .1724 .2289 .0008
Restricted Restricted Restricted Restricted Normal Normal Normal Normal Normal Normal Increased
0.92 0.04 0.81 0.04 0.79 0.02 0.86 0.02 1.20 0.08 1.11 0.09 0.98 0.05 1.00 0.02 0.96 0.02 1.04 0.03 1.39 0.06
1.14 0.02 1.58 0.19 1.77 0.07 NA 3.46 0.21 3.00 0.32 NA 2.76 0.17 3.09 0.06 3.32 0.27 2.90 0.25
ADC, apparent diffusion coefficient. Data are expressed as mean standard error. Standard errors in columns 7 and 8 were calculated from those in columns 2 to 4 using Tylor’s expansion method. * P values are from two-sample t tests to compare means of ADC values between columns 2 and 3; the pooled method was used if equality of variance held, and the Satterthwaite method was used otherwise. y Quantitative results are based on comparing means in columns 2 and 3 and P values in column 5. Table 2 Comparison of Apparent Diffusion Coefficient Values ( 106 s/mm2) Quantitative Diffusion Pattern* Tissue Type Solid germinoma Uninvolved brain Cystic/necrotic germinoma Solid/uninvolved brain Cystic/necrotic/uninvolved brain
Restricted 694.71 74.54 aA 828.62 8.56 dB 2574.16 125.42 eC 0.84 0.09 3.11 0.15
Normal
Increased
947.64 54.38 bD 833.74 6.46 dD 1875.84 247.16 eE 1.14 0.07 2.25 0.30
1408.44 131.64 cF 845.66 16.68 dF 2452.06 210.18 eF 1.67 0.16 2.90 0.26
Data are expressed as mean standard error. Standard errors in rows 5 and 6 were estimated from means and standard errors in rows 2 to 4 using Tylor’s expansion method. * Means followed by different lowercase letters in a row indicate a statistical difference by Tukey’s test (P < .05). Means followed by different uppercase letters in a column in rows 2 to 4 indicate a statistical difference by Tukey’s test (P < .05).
enable the demarcation of histopathologic components of gliomas that may not be clearly distinguishable on conventional MR (21,23,24). The ADC, a rotationally invariant measurement of the amount of total diffusion within a given tissue (11,13), is required to quantify the degree of water motion (diffusivity) (11,12). Additionally, it has been inferred from prior studies (20,21,23–25) that ADC values provide an indication of the relative intracellular and extracellular volumes within the tissue being studied. On the basis of the radiologic appearance, higher ADC values have been attributable to cystic foci, whereas values approaching that of normal gray and white matter were attributed to dense, highly cellular tumor tissue (21,23,24). However, the diffusivity of CNS germinomas as reflected by ADC values or maps has not been previously reported. We set
1362
out to quantify the rate of water diffusion of primary CNS germinomas and to determine if there was correlation with the histologic components.
Solid Components of Germinoma The ADC values of the solid components (range, 656.89 106 to 1172.30 106 s/mm2; standard error, 8.24 106 to 68.90 106 s/mm2) of all 11 germinoma lesions in this study were lower than the reported ADC values of enhancing high-grade gliomas (range, 900 106 to 1400 106 s/mm2) using similar techniques (20,23,24). In addition, the normalized ADC values as depicted by ADCsolid/ADCbrain ratios (Table 1) for the group of germinomas with restricted and normal diffusion were lower than
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
Germinomas with Predominantly Normal Diffusion The majority (55%) of germinomas in this series exhibited ADC values that were not significantly different from those of normal gray and white matter. However, the mean ADCsolid value for these cases (947.6410654.38106 s/mm2) was lower than the mean ADC values (1100 106 and 1600 106 s/mm2) reported for enhancing high-grade gliomas (24). The normalized ADC values as represented by the mean ADCsolid/ADCbrain ratio (1.14) was much lower than the previously reported (20) ADC ratio for high-grade astrocytoma (1.68) and very similar to the ADC ratio for CNS lymphoma (1.15). As such, our results are expected for a hypercellular neoplasm such as germinoma. Figure 5. Classic germinoma showing lobules of densely packed tumor cells interspersed by thin bands of fibrous tissue sprinkled with lymphocytes. (Hematoxylin and eosin staining; magnification, 200.)
the reported (20) ADC ratios of lymphoma, a very cellular neoplasm. Because the leading hypothesis for the rate of water diffusion in CNS neoplasms is related to cellularity (20,23,24,31), the histologically characteristic hypercellularity of all germinomas might account for the reduced diffusivity that was observed in this cohort.
Germinomas with Predominantly Restricted Diffusion The rate of water diffusion in the solid components of germinomas was significantly lower than that of uninvolved brain (P < .05) in approximately 36% of cases, and the mean ADCsolid value for these cases (694.9110674.54 106 s/mm2) was lower than the mean ADC values (1100 106 and 1600 106 s/mm2, respectively) reported for enhancing and nonenhancing high-grade gliomas (24). Moreover, the mean ADC value for this group of germinomas was lower than that for lymphoma (870 106 s/mm2) (20). Additionally, the normalized ADC values, as represented by the mean ADCsolid/ADCbrain ratio (0.84) for germinomas with restricted diffusion, was yet lower than that reported (1.15) for CNS lymphoma (20). The low ADC values found in this subset of germinomas are similar to the ADC values reported for restricted diffusion seen in acute infarctions (32) and the cystic portions of abscesses (15). This may reflect the precision with which the ADC values were obtained. Alternatively, it may indicate that correlation of the ADC values with the underlying pathology and interpretation may be more complicated than previously hypothesized. Future analysis of diffusivity with more advanced techniques such as diffusion tensor imaging in a prospective manner might help in understanding the pathophysiology of germinomas.
Germinomas with Predominantly Increased Diffusion The solid component of a single germinoma (9%) demonstrated increased rates of water diffusion (Fig 4). The mean ADCsolid value (1172.3 106 48 106 s/mm2) and the normalized ADC values (ADCsolid/ADCbrain, 1.39) were similar to prior reports for high-grade gliomas (20,21,23,24,33,34) but higher than that for lymphoma (20). This germinoma was radiologically heterogeneous, had extensive cystic and necrotic foci (Fig 4), and histologically demonstrated extensive necrosis. From a biologic perspective, this is consistent with the presence of microscopic foci of necrosis within the solid more cellular germinomas, resulting in a relatively large extracellular volume. Another possible reason for this outlier is that the moderate standard error of the mean ADCsolid value for this lesion may reflect overlap between the solid and cystic/necrotic foci in the ROI placement in this germinoma, despite efforts to avoid them. Similar issues with volume averaging and ROI placements have been previously reported and are known to possibly affect ADC measurements in routine clinical imaging(14).
Cystic and Necrotic Components of Germinoma The ADC values of the cystic and necrotic components of germinomas in this series are in agreement with those of other intracranial neoplasms (21,23,24). Likewise, in this study, the ADC values were highest in cystic and necrotic foci and were significantly higher than the ADC values of any of the solid components. The ADC values of the cystic and necrotic components of germinomas may be used to distinguish these lesions from intracranial abscess, because the reported ADC values for abscess (15,19,35) are significantly lower than those for the cystic components of tumors. The identification of cystic and necrotic foci in germinomas may have prognostic implications, as suggested by Moon et al (29), who demonstrated that the response of germinoma to radiation
1363
DOUGLAS-AKINWANDE ET AL
Academic Radiology, Vol 16, No 11, November 2009
Table 3 Comparison of Diffusion Pattern and Histopathologic Results Quantitative Diffusion Pattern Component of Germinoma Cystic/necrosis Connective tissue Granuloma Lymphocytic infiltrate Cellularity
Classification
Restricted
Normal
Increased
P
3 1 0 3 0 3 0 3 0 0 2 1
3 3 4 2 4 2 1 5 0 0 3 3
0 1 0 1 1 0 1 0 0 0 1 0
.55
Absent Present Low High Absent Present Mild Intense <25% 25%–50% 50%–75% >75%
therapy correlated negatively with the presence of a cystic region. Analysis of Normal Gray and White Matter The ADC values for normal gray and white matter in this series were consistent with previous reports in adult patients (24,36) and were therefore suitable reference standards. The ADCbrain values in this series also suggest that the ADCbrain values of children aged $5 years are similar to those of adults. Study Limitations It may be argued that one potential limitation of this study is that some of the lesions were located along the hypothalamic-hypophyseal axis, and DW echoplanar imaging is known to be sensitive to nonuniformities in the magnetic field. A major source of such nonuniformity is the differences in magnetic susceptibility at air-tissue and bone-tissue interfaces (24,37), such as the skull base and particularly in the sellar region. However, such susceptibility artifacts are less problematic with single-shot echoplanar techniques (12,37), which were used in this study. In addition, inclusion criteria for this study required large lesions that extended above the suprasellar cistern. Moreover, in this study, there was no difference in the DW characteristics of germinomas that were located along the hypothalamic-hypophyseal axis and the pineal region. Because of the relatively inaccessible locations of germinomas, neoplasms that are sensitive to adjuvant therapy, no attempts are usually made for complete surgical excision. Instead, the goal is to obtain an adequate biopsy specimen of the solid lesion for histopathologic diagnosis. This approach usually yields small samples that are just sufficient for diagnosis.
1364
.11 .10 .09 .61
CONCLUSIONS Germinomas had heterogeneous diffusivity; the solid components of 36%, 55%, and 9% of germinomas predominantly exhibited restricted, normal, and increased diffusivity, respectively. There was no correlation between the histologic components and the ADC values of germinomas. The ADC values may be used to target biopsy sites and to noninvasively monitor response to therapy, as suggested previously (38).
REFERENCES 1. Kleihues P, Cavenee WK. International Agency for Research on Cancer. Pathology and genetics of tumours of the nervous system. Lyon, France: IARC Press, 2000. 2. Horowitz MB, Hall WA. Central nervous system germinomas. A review. Arch Neurol 1991; 48:652–657. 3. Salzman KL, Rojiani AM, Buatti J, et al. Primary intracranial germ cell tumors: clinicopathologic review of 32 cases. Pediatr Pathol Lab Med 1997; 17:713–727. 4. Chang T, Teng MM, Guo WY, Sheng WC. CT of pineal tumors and intracranial germ-cell tumors. AJNR Am J Neuroradiol 1989; 10:1039–1044. 5. Higano S, Takahashi S, Ishii K, Matsumoto K, Ikeda H, Sakamoto K. Germinoma originating in the basal ganglia and thalamus: MR and CT evaluation. AJNR Am J Neuroradiol 1994; 15:1435–1441. 6. Kim DI, Yoon PH, Ryu YH, Jeon P, Hwang GJ. MRI of germinomas arising from the basal ganglia and thalamus. Neuroradiology 1998; 40:507–511. 7. Konno S, Oka H, Utsuki S, et al. Germinoma with a granulomatous reaction. Problems of differential diagnosis. Clin Neuropathol 2002; 21:248–251. 8. Koren A. Bifocal primary intracranial germinoma in a child. Case report. Radiol Oncol 2001; 35:179–183. 9. Ballesteros MD, Duran A, Arrazola J, et al. Primary intrasellar germinoma with synchronous pineal tumor. Neuroradiology 1997; 39:860–862. 10. Schaefer PW, Grant PE, Gonzalez RG. Diffusion-weighted MR imaging of the brain. Radiology 2000; 217:331–345. 11. Le Bihan D. Molecular diffusion nuclear magnetic resonance imaging. Magn Reson Q 1991; 7:1–30. 12. Castillo M, Mukherji SK. Diffusion-weighted imaging in the evaluation of intracranial lesions. Semin Ultrasound CT MR 2000; 21:405–416.
Academic Radiology, Vol 16, No 11, November 2009 DIFFUSIVITY OF GERMINOMA AND HISTOLOPATHOLOGIC CORRELATION
13. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, LavalJeantet M. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988; 168:497–505. 14. DeLano MC, Cooper TG, Siebert JE, Potchen MJ, Kuppusamy K. High-bvalue diffusion-weighted MR imaging of adult brain: image contrast and apparent diffusion coefficient map features. AJNR Am J Neuroradiol 2000; 21:1830–1836. 15. Cartes-Zumelzu FW, Stavrou I, Castillo M, Eisenhuber E, Knosp E, Thurnher MM. Diffusion-weighted imaging in the assessment of brain abscesses therapy. AJNR Am J Neuroradiol 2004; 25:1310–1317. 16. Guzman R, Barth A, Lovblad KO, et al. Use of diffusion-weighted magnetic resonance imaging in differentiating purulent brain processes from cystic brain tumors. J Neurosurg 2002; 97:1101–1107. 17. Hartmann M, Jansen O, Heiland S, Sommer C, Munkel K, Sartor K. Restricted diffusion within ring enhancement is not pathognomonic for brain abscess. AJNR Am J Neuroradiol 2001; 22:1738–1742. 18. Desprechins B, Stadnik T, Koerts G, Shabana W, Breucq C, Osteaux M. Use of diffusion-weighted MR imaging in differential diagnosis between intracerebral necrotic tumors and cerebral abscesses. AJNR Am J Neuroradiol 1999; 20:1252–1257. 19. Tung GA, Evangelista P, Rogg JM, Duncan JA III. Diffusion-weighted MR imaging of rim-enhancing brain masses: is markedly decreased water diffusion specific for brain abscess? AJR Am J Roentgenol 2001; 177:709–712. 20. Guo AC, Cummings TJ, Dash RC, Provenzale JM. Lymphomas and highgrade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002; 224:177–183. 21. Brunberg JA, Chenevert TL, McKeever PE, et al. In vivo MR determination of water diffusion coefficients and diffusion anisotropy: correlation with structural alteration in gliomas of the cerebral hemispheres. AJNR Am J Neuroradiol 1995; 16:361–371. 22. Eis M, Els T, Hoehn-Berlage M. High resolution quantitative relaxation and diffusion MRI of three different experimental brain tumors in rat. Magn Reson Med 1995; 34:835–844. 23. Krabbe K, Gideon P, Wagn P, Hansen U, Thomsen C, Madsen F. MR diffusion imaging of human intracranial tumours. Neuroradiology 1997; 39:483–489. 24. Tien RD, Felsberg GJ, Friedman H, Brown M, MacFall J. MR imaging of high-grade cerebral gliomas: value of diffusion-weighted echo-planar pulse sequences. AJR Am J Roentgenol 1994; 162:671–677.
25. Okamoto K, Ito J, Ishikawa K, Sakai K, Tokiguchi S. Diffusion-weighted echo-planar MR imaging in differential diagnosis of brain tumors and tumor-like conditions. Eur Radiol 2000; 10:1342–1350. 26. Edwards MS, Hudgins RJ, Wilson CB, Levin VA, Wara WM. Pineal region tumors in children. J Neurosurg 1988; 68:689–697. 27. Liang L, Korogi Y, Sugahara T, et al. MRI of intracranial germ-cell tumours. Neuroradiology 2002; 44:382–388. 28. Matsumura N, Kurimoto M, Endo S, Fukuda O, Takaku A. Intracranial germinoma associated with Down’s syndrome. Report of 2 cases. Pediatr Neurosurg 1998; 29:199–202. 29. Moon WK, Chang KH, Han MH, Kim IO. Intracranial germinomas: correlation of imaging findings with tumor response to radiation therapy. AJR Am J Roentgenol 1999; 172:713–716. 30. Osborn AG. Diagnostic imaging. Brain. Salt Lake City, UT: Amirsys, 2004. 31. Gupta RK, Sinha U, Cloughesy TF, Alger JR. Inverse correlation between choline magnetic resonance spectroscopy signal intensity and the apparent diffusion coefficient in human glioma. Magn Reson Med 1999; 41:2–7. 32. Oppenheim C, Grandin C, Samson Y, et al. Is there an apparent diffusion coefficient threshold in predicting tissue viability in hyperacute stroke? Stroke 2001; 32:2486–2491. 33. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999; 9:53–60. 34. Castillo M, Smith JK, Kwock L, Wilber K. Apparent diffusion coefficients in the evaluation of high-grade cerebral gliomas. AJNR Am J Neuroradiol 2001; 22:60–64. 35. Camacho DL, Smith JK, Castillo M. Differentiation of toxoplasmosis and lymphoma in AIDS patients by using apparent diffusion coefficients. AJNR Am J Neuroradiol 2003; 24:633–637. 36. Helenius J, Soinne L, Perkio J, et al. Diffusion-weighted MR imaging in normal human brains in various age groups. AJNR Am J Neuroradiol 2002; 23:194–199. 37. Maier SE, Bogner P, Bajzik G, et al. Normal brain and brain tumor: multicomponent apparent diffusion coefficient line scan imaging. Radiology 2001; 219:842–849. 38. Chenevert TL, McKeever PE, Ross BD. Monitoring early response of experimental brain tumors to therapy using diffusion magnetic resonance imaging. Clin Cancer Res 1997; 3:1457–1466.
1365