| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Johannes Gutenberg-University, Langenbeckstrasse 1,
Mainz 55131, Germany.
2 Department of Cardiothoracic and Vascular Surgery, Johannes
Gutenberg-University, Mainz 55131, Germany.
3 Department of Internal Medicine and Cardiology, Johannes Gutenberg-University,
Mainz 55131, Germany.
4 Department of Medical Biometry, Epidemiology and Informatics, Johannes
Gutenberg-University, 55131 Mainz, Germany.
Received June 30, 2003;
accepted after revision November 12, 2003.
Supported in part by Schering, Berlin, Germany.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty-three patients (41 postoperative Stanford type A, 12 type B dissections) were scanned at 1.5 T during a 3-year period. The study reviewed ECG-gated breath-hold black blood sequences and 3D contrast-enhanced MR angiography of the thoracic aorta supplemented by segmented cine and phase-contrast imaging as well as abdominal contrast-enhanced MR angiography. A retrospective separate analysis of black blood acquisitions and contrast-enhanced MR angiograms from a total of 72 examinations was performed by two radiologists to evaluate detection of intimal flaps and assess image quality.
RESULTS. Sensitivity and specificity of black blood sequences compared with those of contrast-enhanced MR angiography in detecting intimal flaps were 87% and 94% for the thoracic aorta, and 54% and 97% for the supraaortic branches, respectively. Contrast-enhanced MR angiography was subjectively rated as superior to black blood techniques for visualizing intimal flaps and yielded better overall image quality (p < 0.001). Aortic valve competence can be assessed on segmented cine techniques. Phase-contrast sequences enabled the quantification of regurgitant flow across the aortic valve and the analysis of flow patterns in the true and false channels.
CONCLUSION. Contrast-enhanced MR angiography is superior to black blood MRI in detecting the presence or absence of intimal flaps and is particularly useful in assessing supraaortic branch vessel involvement. Cine and phase-contrast techniques should be included in the imaging follow-up to diagnose possible complications of chronic aortic dissections.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI is an established technique for the assessment of aortic diseases in general and aortic dissection in particular [8–12]. The implementation of high-performance gradient systems enables the acquisition of most MRI sequences in a single breath-hold. Contrast-enhanced 3D MR angiography represents a substantial advancement for the morphologic assessment of thoracic aortic abnormalities [13–18]. The purpose of this retrospective study was to evaluate the impact and contribution of different ECG-gated breath-hold MRI techniques in a comprehensive protocol for the follow-up of chronic aortic dissections. We compared the advantages and disadvantages of these methods for assessing morphology, function, and flow, and we examined whether a single MRI sequence could be used to address all required aspects.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A total of 72 MRI examinations of the thoracic aorta were performed in 41 patients with Stanford type A dissection after surgical repair of the ascending aorta and 12 patients with chronic Stanford type B dissection, including one intramural hematoma. Thirty-six patients underwent a single examination; 15 were scanned twice, and two were scanned three times. Stanford type B dissections were managed conservatively in six patients (nine examinations). Five patients were examined in diagnostic work-up leading to replacement of the descending aorta (five examinations). Two MRI examinations were performed postoperatively in a patient with Stanford type B dissection.
MRI Protocol
MRI was performed on a 1.5-T whole-body MRI scanner (Magnetom Vision,
Siemens Medical Solutions) with 25-mT/m gradients enabling a 0.6-msec rise
time. A 4-MDCT body phased array coil was used for signal detection. All
sequences were obtained with breath-hold technique during deep inspiration,
and all but one were ECG-gated. Table
1 lists the parameters of the sequences performed. Two different
turbo spin-echo sequences were used for morphologic assessment of the thoracic
aorta by means of black blood MRI. A transverse T2-weighted HASTE sequence was
applied to acquire 7 slices per breath-hold. Additional sections were obtained
in the oblique sagittal plane if necessary. In approximately half of the
examinations, one or more complementary single-slice T1-weighted turbo
spin-echo sequences were added, mainly in parasagittal orientation, to
supplement the HASTE imaging.
|
A 3D spoiled fast low-angle shot (FLASH) parasagittal sequence analogous to the left anterior oblique projection was used for contrast-enhanced MR angiography of the thoracic aorta. Data acquisition was enhanced by power injection (Spectris, Medrad) of a single dose of 20 mL of gadopentetate dimeglumine (Magnevist, Schering) followed by a 20-mL saline flush, both at a flow rate of 2 mL/sec. The test bolus method was used to determine the individual time delay between the start of contrast administration and the initiation of data acquisition [18–20]. The time delay was optimized to cover the entire thoracic aorta as well as both lumina with sufficient intraluminal contrast, because the test bolus examination provided information on the time of peak gadolinium concentration not only in the ascending and descending aorta but also in the true and false channels of the dissection.
An MR angiogram of the abdominal aorta was added in 44 examinations (36 patients) to determine the course of the intimal flap or to evaluate suspected abnormalities of the abdominal aorta. Two sequential coronal fat-saturated contrast-enhanced MR angiograms without ECG-gating were acquired after repositioning the phased array coil and de nouveau application of 20 mL of gadopentetate dimeglumine. Ten seconds were empirically added to the thoracic time delay to time the contrast-enhanced MR angiography of the abdominal aorta. The total volume of contrast agent used for the test bolus and the MR angiographic display of the thoracic and abdominal aorta did not exceed 42 mL (median dose of gadopentetate dimeglumine, 0.22 mmol/kg of body weight; interquartile range, 0.14–0.25 mmol/kg of body weight).
Segmented cine FLASH sequences with echo-view sharing and a temporal resolution of 50 msec and segmented phase-contrast gradient-recalled echo through-plane flow measurements with a temporal resolution of 110 msec and velocity encoding ranging from ± 75 to ± 250 cm/sec were both performed during suspended respiration. The median time in the MRI scanner from the initiation of the localizer to the end of contrast-enhanced MR angiography was 31 min (interquartile range, 22–41 min).
Qualitative Image Analysis
All 72 MR images were retrieved from the hospital PACS (picture archiving
and communication system) or optical discs and evaluated on a satellite MRI
scanner console (Siemens Medical Solutions, with Numaris VB33D software). A
separate retrospective analysis of the black blood images and the
contrast-enhanced MR angiograms was performed by two radiologists in consensus
who were unaware of the patient's clinical history or type of dissection.
Double-oblique multiplanar reformations (MPR) of 3D contrast-enhanced MR
angiography data sets were interactively created by either radiologist as an
adjunct to the source images. Other imaging techniques that could have served
as a standard of reference were not available in all cases because the
retrospective nature of this study precluded having a standardized imaging
protocol. Contrast-enhanced MR angiography therefore served as an internal
standard of reference for the depiction of vascular morphology.
The presence and extent of dissection as well as supraaortic branch vessel involvement were assessed on black blood sequences and contrast-enhanced MR angiography. The evaluation of these features on black blood techniques was mainly based on HASTE sequences. T1-weighted turbo spin-echo acquisitions could not be compared with HASTE or contrast-enhanced MR angiography because of the limited number of examinations and slices obtained with this sequence type, so they were grouped with HASTE images in the black blood imaging category. The involvement of abdominal aortic branch vessels and iliac arteries, the relative size of the two lumina, the presence of intimal entry or reentry tears, and the anatomy of the supraaortic branches were evaluated on contrast-enhanced MR angiography. Aneurysmal dilatation, aortic valvular regurgitation, and other complications of chronic aortic dissection that were discovered on MRI follow-up were noted also.
Sensitivity and specificity of black blood sequences compared with contrast-enhanced MR angiography in assessing the presence of intimal flaps in the thoracic aorta and supraaortic branches were calculated on the basis of the consensus reviews. The parameter "supraaortic branch vessel involvement" was coded if the supraaortic branches were covered by at least two different transverse slices on black blood sequences and if they were depicted in diagnostic image quality (51 examinations).
Quantitative Image Analysis
The image quality of black blood imaging and contrast-enhanced MR
angiography was graded for the visualization of the intimal flap in the
thoracic aorta and supraaortic branches on a 3-point scale (1, good; 2,
moderate; 3, poor). Only those 47 cases with a concordant positive finding of
a dissection membrane in the thoracic aorta were included, so that the
depiction of the intimal flap could be compared between both imaging
techniques. The same 51 cases were analyzed with respect to image quality of
supraaortic branch vessels for sensitivity and specificity. Overall image
quality was rated on a 5-point scale (1, excellent; 2, good; 3, moderate; 4,
poor; 5, nondiagnostic).
The presence or absence of the jet phenomenon on cine MRI was used to
evaluate aortic valve competency. If a diastolic jet phenomenon indicating
aortic valve insufficiency was detected on cine MRI, phase-contrast
gradient-recalled echo sequences were performed to assess the degree of aortic
regurgitation. The regurgitant fraction (RF) of incompetent aortic valves was
quantified by means of velocity-encoded flow measurements that were orientated
orthogonally to the proximal ascending aorta. The severity of aortic
regurgitation was graded on a 3-point scale (mild [RF < 15%], moderate [RF
> 15% but < 30%], severe [RF > 30% but 
50%]).
Flow patterns in the true and false lumen of dissections involving the descending aorta were assessed on phase-contrast gradient-recalled echo sequences planned perpendicularly to its course in 32 examinations. Mean velocity, peak mean velocity, mean flow, mean area, forward and reverse volume, and net forward volume were evaluated for both channels using Argus V2.3 software (Siemens Medical Solutions).
Statistical Analysis
For the description of continuous parameters, both means and standard
deviations (mean ± SD) and ranges or medians with the interquartile
ranges Q1–Q2 are presented beside the number of
cases, n. Sensitivity and specificity in detecting intimal flaps were
estimated as measures of validity for the black blood techniques compared with
contrast-enhanced MR angiography. The respective 95% confidence intervals (CI)
are given in parentheses. The level of agreement between imaging methods in
detecting the presence of intimal flaps was evaluated by means of Cohen's
kappa coefficient with 95% CIs and the McNemar test for observation bias.
The intraindividual comparison of parallel quantitative assessments, like the grading of black blood sequences and contrast-enhanced MR angiography, was performed using the sign test for paired samples. Image quality ratings and their agreement were also evaluated by means of Cohen's kappa coefficient and the McNemar test after they were recoded dichotomously by combining codes 1 and 2, and combining codes 3, 4, and 5. The interindividual quantitative comparison of patient subgroups was based on two-sample Wilcoxon's tests, for example, for differences in flow parameters of the false lumen in relation to the presence of a parietal thrombosis.
No adjustment for multiple comparisons was made because of the exploratory character of our analyses, so a p value of 0.05 or less was considered as an indication of local statistical significance. All statistical analyses were performed using SPSS 11.0 software (SPSS) for Windows (Microsoft).
Correlative Imaging
We retrospectively collected correlative routine imaging studies that had
been requested by vascular surgeons or cardiologists and performed within 6
weeks before or after our contrast-enhanced MR angiograms. Our study was not
planned prospectively, so no standardized trial imaging protocol was applied.
The combinations of imaging techniques varied a great deal and did not allow
detailed statistical analysis. Nevertheless, we compared the presence and
extent of intimal flaps on the available correlative imaging studies with our
respective contrast-enhanced MR angiography findings to detect possible
systematic errors of the internal standard of reference.
Supplementary imaging included transthoracic and transesophageal echocardiography in 42 and 33 cases, respectively. CT angiography performed on helical scanners was available in 38 cases. Six patients underwent digital subtraction angiography or cardiac catheterization combined with aortography. Surgery was performed in 42 cases. Findings of four MRI examinations were not supported by any further correlative imaging or surgical proof.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
The sensitivity and specificity of T2-weighted HASTE and T1-weighted turbo
spin-echo sequences in detecting intimal flaps were 87% and 94% for the
thoracic aorta and 54% and 97% for the supraaortic branches, respectively
(Tables 2 and
3). The presence of thoracic
aortic dissection was missed on black blood sequences in five cases of small
flap remnants and on two examinations of a type B dissection (Figs.
1A,
1B,
1C and
2A,
2B). A false-positive
interpretation with respect to the presence of a dissection membrane occurred
both in the thoracic aorta and supraaortic branches. The extension of the
intimal flap into supraaortic branches was not detected on six of 13
examinations by black blood sequences. The aortic arch and supraaortic
branches were not covered by axial T2-weighted HASTE sequences on two and 16
examinations, respectively. The supraaortic vessels could not be adequately
assessed on black blood sequences in five cases. The agreement of these
imaging techniques in detecting intimal flaps is substantial for the thoracic
aorta (
= 0.73), but only moderate for supraaortic branches ( =
0.59; Tables 2 and
3). However, correlation was
excellent regarding the extent of dissection in all 47 cases with a concordant
positive finding.
|
|
|
|
Contrast-enhanced MR angiography was superior to black blood imaging for visualizing the intimal flap in the thoracic aorta as well as supraaortic branch vessels and yielded better overall image quality (p < 0.001, sign test) (Table 4). Cohen's kappa coefficients of approximately zero indicated discrepant categorical ratings. Accordingly, the McNemar test also revealed a significant trend for better image quality in contrast-enhanced MR angiograms (p < 0.001) (Table 4).
|
Because no phase dispersion jet was observed on contrast-enhanced MR angiography, aortic valve insufficiency could only be detected on cine MRI. Evaluation of incompetent aortic valves on phase-contrast imaging revealed 16 mild, 15 moderate, and eight severe aortic regurgitations. The analysis of flow patterns in the true and false channel showed significantly lower mean velocity and peak of mean velocities and a significantly higher mean cross-sectional area and proportion of retrograde flow volume in the false lumen compared to the true one (p = 0.008, sign test). Mean flow and antegrade and net forward volume did not differ between channels. A pronounced decrease in mean velocity, peak of mean velocities, mean flow, and antegrade and net forward volume was noted in false channels that had marked parietal thrombosis compared with those without apposition thrombus (p = 0.040, Wilcoxon's test).
Complications During Follow-Up
Thoracic dissecting aneurysms increased in diameter in eight patients
leading to elective surgery in two of them. Two anastomotic aneurysms were
discovered, and one required resection. An entry persisted after replacement
of the descending aorta in a patient with Stanford type B dissection. Three
new aortic valve insufficiencies were discovered (including one prosthesis).
Progression of existing aortic regurgitation was noted in 12 cases. The
severity of aortic regurgitation led to replacement of the aortic valve or
rerepair of the ascending aorta by means of a composite graft in four patients
(Fig. 3A,
3B,
3C,
3D).
|
|
|
|
Three patients presented with hemorrhagic pleural effusion, two with atelectasis. Mediastinal hematoma, simple pleural effusion, renal artery stenosis, and compression of left lower lobe bronchi were observed in one case each. These extraaortic abnormalities were only visible on black blood sequences, with the exception that the renal artery stenosis was depicted on abdominal contrast-enhanced MR angiography.
Correlative Imaging
Several discrepant findings were observed. One aneurysm at the cannulation
site of the aorta was missed on transesophageal echocardiography, probably
because of the blind spot. A Stanford type B dissection with retrograde
propagation of blood in a markedly dilated false channel next to the origin of
the left subclavian artery was misinterpreted as retrograde involvement of the
aortic arch by transesophageal echocardiography in a patient with a greatly
elongated aortic arch. The number of dissected supraaortal branches as
depicted by transesophageal echocardiography differed from the
contrast-enhanced MR angiography assessment in one case. No other
discrepancies were found.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Black blood techniques are inadequate for the diagnosis of supraaortic
branch vessel involvement, as is shown by their moderate ( = 0.59)
agreement and low (54%) sensitivity in this study and in others (67%)
[7,
12,
21]. Although sensitivities
and specificities of 98% have been reported for gated spin-echo pulse
sequences in the detection of acute aortic dissection, our data reveal a
slightly lower sensitivity of 87% for chronic dissections (95% CI,
0.78–0.96) [14,
21,
22]. This might be caused by
sample size differences, difficulties in depicting short residual intimal
flaps in the postoperative setting, and the problem of differentiating
artifacts from parietal thrombosis or intimal flaps
[12,
14]. Artificial increases in
intraluminal signal intensity can be caused by stagnant, retrograde, or
turbulent blood flow or by poor ECG-gating
[3,
12,
23,
24]. These factors account for
two false-positive findings observed on T2-weighted HASTE images. Black blood
sequences are not suited for assessing aortic dissection if no supplementary
techniques are used because their total percentage of inaccurate diagnoses in
our study exceeds 10% (8/72 or 11% for the thoracic aorta and 7/51 or 14% for
the supraaortic branches).
However, black blood sequences provide an anatomic overview and are an essential part of the imaging protocol for evaluating aortic wall abnormalities such as intramural hematoma, aortitis, penetrating aortic ulcer, effusion or hemorrhage in the pericardium, mediastinum, or pleural space. These features may be missed on contrast-enhanced MR angiography [14, 24]. Rapid black blood MRI using half-Fourier acquisition with relaxation enhancement is superior to ECG-triggered turbo spin-echo sequences for the evaluation of thoracic aortic diseases [24].
Fast imaging using steady-state free precession is a new bright blood approach for evaluating diseases of the thoracic aorta [25]. This technique is well suited for evaluating aortic dissection and can also depict extraaortic manifestations, although its value in the detection of intramural hematomas and branch vessel involvement has not yet been established [25].
The usefulness of contrast-enhanced MR angiography for imaging of thoracic aortic diseases in general has been shown in several studies [7, 13–18]. Our study specifically addresses chronic aortic dissection and is based on the largest series—to our knowledge—so far reported of cases of aortic dissection examined on breath-hold contrast-enhanced MR angiography. Krinsky et al. [14] reported a sensitivity of 96% and specificity of 100% for the detection of acute and chronic aortic dissection by non–breath-hold contrast-enhanced MR angiography. Contrast-enhanced MR angiography is independent of slow-flow phenomena and can therefore differentiate this kind of artifact in black blood images from an intimal flap or a parietal thrombosis [7, 13, 14] because intraluminal contrast relies on the shortening of the T1 of blood in response to the administration of gadolinium chelates. Contrast-enhanced MR angiography is superior to black blood techniques for detecting intimal flaps and assessing supraaortic branch involvement and gives the best overall image quality, which can partly be attributed to its higher spatial resolution and the good delineation of a hypointense intimal flap outlined on either side by gadolinium-enhanced bright blood in a double-barreled aorta [7, 14].
MPRs of 3D contrast-enhanced MR angiography data sets of both the thoracic and abdominal aorta are extremely helpful in evaluating the extent of the dissection, in delineating the course of the intraluminal membrane, in depicting short flap remnants and intimal tears, and in assessing false lumen patency and degree of parietal thrombosis. MPRs are also valuable for determining in which lumen the supraaortic and visceral branches originate or whether the dissection extends into some of them [26]. The external aortic diameter can be measured perpendicularly to the course of the aorta, which is recommended for reporting aneurysm size [14]. Contrast-enhanced breath-hold 3D MR angiography of the thoracic and abdominal aorta can provide all the information the referring clinicians need on the morphology of aortic dissection and can be regarded as the standard MRI sequence for depicting intimal flaps and assessing branch vessel involvement. However, contrast-enhanced MR angiography by itself lacks intrinsic properties needed to identify intramural hematomas, periprosthetic thickening, and aortic regurgitation.
The evaluation of flow patterns in a patent false lumen provides valuable information for the prediction of aneurysm formation because the amount of flow volume in the false channel correlates positively with aneurysm growth rate [27]. Thrombosed channels exhibit significantly diminished mean velocity and flow. This fact supports the observation that the formation of thrombus in the false lumen is a good prognostic sign [6]. A patent false lumen, which is found in as many as 90% of patients with aortic dissection, has a propensity for aneurysm formation and is associated with a more unfavorable prognosis than is the case with an obliterated false channel [4, 6]. Communication between the lumina via tears in the aortic arch or the proximal descending aorta keeps the false lumen patent distal to the site of entry [2, 3]. The velocity-encoded flow measurements can also be used to differentiate the true channel from the false lumen on the basis of the flow profile pattern for cases in which morphologic indications are ambiguous [27, 28]. Two sequential contrast-enhanced MR angiography acquisitions of the thoracic and abdominal aorta and time-resolved MR angiography can also provide information on false channel perfusion by qualitatively highlighting the passage of a contrast agent in the true and false lumen [29]. Both phase-contrast and MR angiographic techniques may be helpful in evaluating visceral ischemia caused by impaired perfusion of the true lumen or by branch vessel obstruction. A predominantly concave configuration of the dissection flap toward the false lumen during the cardiac cycle is an indirect sign for malperfusion of the true channel and the organs supplied by its branches [30].
The most common causes of death in acute and chronic aortic dissection are rupture of the aorta caused by aneurysm formation and the development of severe aortic regurgitation or prosthetic valve malfunction, respectively [1–3]. The progression of aortic regurgitation and aneurysm formation were also the two major causes for surgical reintervention in our study population. Objective documentation of aortic diameter on MRI and the high reproducibility in serial examinations allow the detection of even small changes during follow-up that may lead to altered patient treatment including preventive surgical measures [3, 5, 7, 8, 14, 21, 31]. The need to evaluate aortic valve competence is underscored by the high incidence of aortic valve insufficiency in our study group. Cine MRI therefore has to be considered not only as an option [16] but rather as essential for the follow-up of chronic aortic dissections. Because of the unique property of velocity-encoded MRI to measure flow and its high interstudy reproducibility, phase-contrast sequences are suited for quantifying and monitoring aortic regurgitation if any is present [32]. MRI can also detect postoperative complications such as perigraft thickening or thrombus, periprosthetic flow, and leakage or dehiscence of the anastomosis [3, 7–9, 31, 33]. Although a combination of different MRI sequences is needed to assess the features of morphology, function, and flow, it is just this combined approach that makes MRI so well suited to address all aspects of chronic aortic dissection and its possible complications in a single examination. MRI is therefore the imaging technique of choice for the follow-up of chronic aortic dissection by interdisciplinary teams at tertiary referral centers [3, 7–9]. This conclusion is also supported by the European Society of Cardiology Task Force on Aortic Dissection [1].
The following breath-hold MRI protocol is suggested for the evaluation of the aorta in patients with chronic aortic dissection and can be acquired in 15–25 min on cardiovascular MRI scanners:
We acknowledge several limitations in our study, including its retrospective design, the restrictions in sample size when analyzing qualitative readings, and the subjective criteria used for the consensus grading of black blood and contrast-enhanced MR angiography sequences. Although most patients underwent an examination with another imaging technique for correlation within approximately 6 weeks, a true evaluation of diagnostic accuracy was not possible retrospectively. Therefore, to compare the diagnostic impact of black blood techniques and contrast-enhanced MR angiography, we chose the latter as an internal standard of reference.
A comparison of the specific contributions and drawbacks of various diagnostic imaging techniques such as cardiovascular MRI, MDCT angiography, and multiplanar transesophageal echocardiography for the follow-up of chronic aortic dissections was outside the scope of our study. Those goals can only be achieved by a prospective trial using state-of-the-art equipment.
In conclusion, contrast-enhanced MR angiography yields better image quality than black blood techniques and is superior for determining the presence or absence of aortic intimal flaps and particularly for assessing supraaortic branch vessel involvement. Performance of only black blood sequences in patients with chronic aortic dissection would have led to false-positive and false-negative interpretations. Black blood imaging is nevertheless an essential part of the protocol because intramural hematomas and extraaortic complications can be missed if only contrast-enhanced MR angiography is used. Parietal thromboses can be depicted with higher diagnostic confidence on contrast-enhanced MR angiography than on the black blood approach. Incompetent aortic valves can be identified by cine imaging. Phase-contrast flow measurements allow the quantification of aortic regurgitation and the evaluation of false channel perfusion. A thorough assessment of chronic aortic dissection is achieved by combining all these MRI techniques. Such a comprehensive approach reduces the need for further imaging studies in the follow-up of this disease entity.
Acknowledgments
This study contains results from a doctoral thesis by B. K. Haag, which is
currently in preparation. The authors thank M. Schuez for assistance in
preparing the manuscript and A. Keuchel for photographic work.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
