| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Abdominal Imaging |
1 Department of Radiology, Division of Abdominal Imaging and Intervention,
Massachusetts General Hospital, 55 Fruit St., White 270, Boston, MA
02114.
2 Present address: Faculty of Health Sciences, University of Dublin, Trinity
College, Dublin 2, Ireland.
Received December 19, 2003; accepted after revision March 16, 2004.
Address correspondence to P. F. Hahn
(phahn{at}partners.org).
OBJECTIVE. Our aim was to determine portable abdominal CT image quality and clinical content in a consecutive series of scans obtained in ICUs at a single center.
MATERIALS AND METHODS. All helical portable abdominal CT scans obtained between June 1999 and December 2000 were reviewed by two observers. Image quality was assessed for feature detection and overall quality compared with the patients' contemporaneous stationary CT scans. Hospital records were used to determine patient demographics, scanning indications, and clinical impact and to verify portable CT findings when possible.
RESULTS. One hundred twenty-two helical portable CT scans (47 contrast-enhanced) and 41 contemporaneous stationary CT scans in 107 patients were included. IV contrast material improved portable CT scan quality, but quality scores for portable CT scans were consistently lower than those for stationary CT scans, both with and without contrast material. Thirty-three conditions suspected before scanning were supported by findings on portable CT scans, which detected evidence of infection in 18 and hemorrhage in 16 cases and motivated seven laparotomies and six percutaneous drainage procedures. Thirty-three portable CT scans (27%) contributed to a change in patient treatment. Results of surgery or autopsy confirmed portable CT findings in 12 of 17 cases.
CONCLUSION. Although image quality is inferior to conventional stationary CT, portable abdominal CT provides important diagnostic information without requiring patient transport outside the ICU. Radiologists should avoid overconfident interpretation of portable CT scans.
Technical developments have added helical CT to the abdominal imaging techniques that can be performed at the bedside. CT scanners used worldwide, capable of portable operation, have been deployed in many large hospitals [1]. However, little information has been published about the performance of such scanners or how they are being used. The purpose of this study was to evaluate a consecutive series of portable helical abdominal CT scans at a single institution to assess image quality and clinical impact.
Materials and Methods
This study was approved by the Subcommittee on Human Studies, the hospital institutional review board. Because it was a retrospective study, patient informed consent was not required.
Portable CT Hardware
In all cases, portable CT scans were obtained on an Anatom 2000 (Analogic),
marketed as the Tomoscan M (Philips Medical Systems)
(Fig. 1). Components of the
Anatom 2000 system include a lightweight gantry with built-in computer and
power supplies, patient table, and operating console with display
[1]. The gantry, driven by a DC
motor, can translate horizontally for a total distance of approximately 356
mm, so the patient remains stationary during scanning
[2]. For scanning of the
abdomen, patients must be moved from their beds, usually onto the patient
table placed at the bedside. A slide board resting between the foot of the
patient's bed and a small table serves as the scanning table when bedside
space is limited.
|
For 18 months before the study, 68 portable CT examinations were performed using an axial CT protocol that did not permit use of IV contrast enhancement. Subsequently, a software upgrade enabled portable abdominal CT examinations to be performed with a helical protocol. In this mode, two volumes 285–355 mm long were scanned in 5-mm contiguous sections, each acquired during 70 sec at 120–130 kV, 30–40 mA, and a pitch of 1.0–1.5:1. Study indication and patient condition determined whether IV contrast material was administered during helical portable CT.
Case Material
Helical portable CT scans obtained after the software upgrade in June 1999
through December 2000 were identified using the radiology information system
(RIS) (IDXrad, IDX Systems). For RIS purposes, the portable CT scanner was
assigned a specific unique examination location, ensuring a complete and
consecutive listing of all portable CT scans. Those examinations with
abdominal examination codes were included, except that 19 portable abdominal
CT scans during the study period were excluded because they were obtained with
the older axial protocol. All CT scans (portable and stationary nonportable)
obtained at our institution during the study period were archived digitally
and without lossy compression on the PACS (Agfa). During evaluation of the
portable CT scans of each patient, the investigators determined from PACS
whether a stationary (nonportable) CT scan had been obtained within 14 days of
any portable CT scans. If so, the temporally closest stationary scan was
included for comparison, regardless of whether it matched the portable CT scan
in contrast agent usage. More than one stationary CT scan was included for
comparison if the patient had more than one portable CT scan. All portable and
stationary abdominal CT scans evaluated included those of the pelvis and the
abdomen.
Assessment of Image Quality
All helical portable abdominal CT scans obtained during the study period
were retrieved from PACS and displayed on a diagnostic-quality work-station.
In addition, any conventional abdominal CT scan obtained within 2 weeks of the
portable scan was also retrieved for the purpose of comparison with the
portable CT scan.
Each abdominal CT scan retrieved was assessed in consensus by two radiologists with 6 and 17 years' experience in CT interpretation. Although both portable and conventional nonportable studies were included, no attempt was made to blind the observers as to which type they were reviewing. Indeed, the acquisition of separate volumes to cover the entire abdomen and pelvis readily identified the portable CT scans.
First, the presence of IV and oral contrast material was determined by inspection of the images. The observers then assessed the quality of the scans. The quality assessment proceeded by determining the visibility of nine selected anatomic structures at different levels in the abdominal CT scan. Visualization of each of the selected anatomic structures was assessed and graded as 1 point (seen) or 0 points (not seen). The anatomic structures in the liver chosen as indicators of image quality included the main portal vein, both right and left portal veins, and secondary portal vein branches. Three points were awarded if all three sets of structures were visible; otherwise, the point score was reduced. An additional point was awarded for visualization of the lobular pattern of the pancreas (Fig. 2). Renal visualization was assessed for identification of the renal collecting system at the hilum (1 point) and the intrarenal collecting system (1 point). The quality of small-bowel visualization at the level of the iliac crest was determined from the definition of bowel against mesenteric fat (1 point), the bowel lumen against the bowel wall (1 point), and the valvular pattern (1 point).
|
The CT scan was then assessed for major technical problems such as streak artifact, motion artifact, skip areas of coverage, or other major defects limiting the diagnostic value of the study. The scan received an additional point if no major technical problems existed. An optimal scan would receive 3 points for visualization of the portal vein, 1 point for visualization of the lobular pattern of the pancreas, 2 points for visualization of the collecting system of the kidney, 3 points for small-bowel visualization at the iliac crest, and 1 point for absence of technical problems. Thus, the maximal score that could be awarded was 10 points.
The overall scan quality was then assessed following a modification of the qualitative scoring system of Stockberger et al. [3]. Scans were graded from 1 to 5 (5, exceptional quality; 4, standard quality; 3, less than standard quality; 2, diagnosis in doubt because of poor scan quality; 1, inadequate quality for diagnosis). The same scoring system was used to assess both the portable scans and the contemporaneous stationary scans.
Assessment of Clinical Impact
All portable CT scans were reviewed on the PACS by the two radiologists,
first without knowledge of the original interpretation and then in comparison
with the radiology report recorded in the RIS. Any discrepancies in
interpretation between the two reviews were recorded. Contemporaneous
stationary CT scans, whether obtained before or after the same patient's
portable CT scans, were also evaluated by the observers.
Using the database of patients generated from the RIS, we retrospectively reviewed the computerized hospital medical record for patient demographics, including age, sex, principal reason for hospital admission, date of admission, date of performance of portable CT, and indication for portable CT. Using this information, we calculated the interval between hospital admission and date of portable abdominal CT. In addition, patients were categorized into seven groups on the basis of the principal medical problem that led to hospital admission: neurologic disease, cardiovascular disease, multiorgan failure, congenital disease, postoperative complications, trauma, and respiratory failure.
Portable abdominal CT scans were evaluated for positive radiologic findings, particularly those suggesting infection, hemorrhage, or important unsuspected diagnoses. In addition, portable CT findings were correlated with the indication for CT, and a decision was made as to whether the portable CT findings confirmed the clinical indication that precipitated the test.
The hospital record was then evaluated for outcome in the aftermath of portable CT. The reports of all abdominal imaging examinations performed during the subsequent hospital course were evaluated. The reports of any examinations performed within 2 weeks of the portable CT were compared with findings on portable CT. In addition, the question of whether the portable CT led to a change in patient treatment was addressed. Finally, the findings on portable CT were correlated with information from other events in the 7-day period after portable CT was performed, including abdominal surgery and autopsy.
Statistical Analysis
Characteristics of portable and nonportable CT scans were compared using
the chi-square test with continuity correction. Means of summed scores for
feature visualization were compared using the two-tailed unpaired Student's
t test. Paired scores for portable CT scans and stationary CT scans
in the same patient were compared using the Wilcoxon's signed rank test. The
null hypothesis was rejected if the difference would have arisen by chance in
less than 5% of trials.
Results
Patient Characteristics
From June 22, 1999, to December 31, 2000, 122 helical portable abdominal CT
studies were performed on 107 patients (64 males, 43 females; age range,
1–86 years; mean age, 55 years). Except for two patients less than 3
years old, all were 19 years old or older. Eleven patients underwent two
helical portable CT examinations; two patients had three. In addition, four
patients who underwent helical portable CT examinations each had one axial
portable CT (of 19 axial portable CT examinations during this period) not
included in the analysis. Portable CT examinations were performed at a mean of
16 days (1–134 days) after hospital admission. All portable CT
examinations were performed in medical, cardiac, pediatric, neurologic, or
surgical ICUs. The patients' principal medical problems were as follows:
neurologic disease (n = 36), cardiovascular disease (n =
20), multiorgan failure (n = 34), congenital disease (n =
2), postoperative complications (n = 5), trauma (n = 4), and
respiratory failure (n = 6).
Assessment of Portable CT Quality
IV contrast agent was not administered for 31 portable CT examinations in
which the indication for the study was exclusion of hemorrhage. This decision
followed departmental protocol for abdominal–pelvic CT performed for
exclusion of this condition. Contrast material was withheld in the remaining
44 patients because of renal insufficiency or allergy to iodinated contrast
material.
One hundred twenty-two helical portable CT scans averaged 6.19 ± 1.64 (mean ± SD) of a possible 10 points for feature visualization and 1.49 ± 0.55 of a possible 5 points on the modified qualitative scoring system of Stockberger et al. [3]. Contrast administration improved portable CT scans for feature visualization, which averaged 6.83 ± 1.49 for 47 contrast-enhanced portable CT scans and 5.79 ± 1.60 for 75 scans without contrast enhancement (p < 0.001) and on the qualitative scale 1.64 ± 0.57 and 1.40 ± 0.52 (p < 0.02), respectively. When scores for portable CT scans were compared with 41 nearly contemporaneous stationary CT scans in this cohort, the stationary CT scans were rated better both for feature visualization (7.22 ± 1.44, p < 0.001) and on the qualitative scale (2.02 ± 0.72, p < 0.0001). Inferior ratings of portable CT versus stationary CT persisted when only the 75 unenhanced portable CT scans were compared with 26 unenhanced stationary CT scans (5.79 ± 1.60 vs 6.92 ± 1.55, p < 0.01 and 1.40 ± 0.52 vs 1.92 ± 0.74, p < 0.001) and for comparison of 47 contrast-enhanced portable CT scans with 15 contrast-enhanced stationary CT scans (6.83 ± 1.49 vs 7.73 ± 1.10, p < 0.05 and 1.64 ± 0.57 vs 2.20 ± 0.68, p < 0.01). Chi-square analysis showed no difference in contrast agent use between portable CT and stationary CT (p = 0.82).
Table 1 shows, for each of nine features scored, the number of portable and stationary CT scans on which that feature could be identified. An additional point was awarded to scans with no major problems. Stationary CT outscored portable CT in all 10 features. In four of the 10 (left and right portal vein, branch portal vein, intrarenal collection system, and absence of serious technical problems), the differences were statistically significant. Thirty-nine portable CT scans (32%) had a total of 41 major problems including motion (n = 23) or streak (n = 4) artifact, gaps in coverage (n = 13), and equilibrium phase contrast enhancement (n = 1). In comparison, only three (7%) of the same patients' 41 stationary CT scans had major problems (streak artifact [n = 1], motion [n = 1], and inadequate field of view [n = 1], p < 0.01).
|
In 24 cases, portable and stationary CT examinations of the same patient were performed within 2 weeks of each other, and both were performed either with (10 cases) or without (14 cases) IV contrast material. Two patients were represented twice among these 24 cases. In these 24 cases, both summed scores for feature visualization and qualitative scores for image quality could be compared between pairs of portable and stationary CT scans. The mean 10-point summed assessment scores for these patients were 6.2 ± 1.68 for portable CT and 7.2 ± 1.53 for stationary CT. The mean 5-point qualitative assessment scores were 1.40 ± 0.51 and 2.10 ± 0.78, respectively. For both 10- and 5-point scales, the scores rated quality of stationary CT better than that of portable CT in the same patients. For both scales, the differences between the two cohorts were significant by the Wilcoxon's signed rank test (p < 0.02 and 0.01, respectively).
Portable CT Indications and Findings
In 116 cases, the indication for portable CT fell into a single indication
category, whereas in six cases, the indications fell into two categories, for
a total of 128 indications among 122 scans. The two most commonly stated
indications for portable CT were suspected infection (n = 79) and
hemorrhage (n = 31). Eighteen portable CT studies included
indications other than these. For 33 (26%) of 128 indications, portable CT
confirmed the suspected diagnosis, whereas in 95 cases (74%), the imaging
findings did not support the suspected clinical diagnosis prompting the CT
scan.
Air-space or pleural disease was present in virtually all patients, reducing the value of clinical signs of infection. When abdominal infection was the indication for portable CT (79 scans), potential sources were found on 15 scans (19%). Categories of infection included infectious colitis (n = 4) (Fig. 3), focal intraabdominal collection (n = 10), acalculous cholecystitis (n = 2), and bowel perforation (n = 1) (Figs. 4A, 4B, and 4C). In 64 other portable CT examinations (81%) performed for a suspected intraabdominal infection, no source of intraabdominal infection was identified. In one CT scan obtained to exclude infection, the findings were reported as negative for infection, but the result of laparotomy showed necrotic bowel. Retrospective review showed evidence of bowel wall thickening.
|
|
|
|
Among 43 portable CT examinations performed for indications that did not include infection, three scans (7%) had findings suggestive of infection. Specifically, these cases all showed colonic thickening suspicious for infectious colitis.
On 31 portable CT scans obtained because of clinical concern for hemorrhage, hemorrhage was found in 11 cases (35%) (Figs. 5 and 6). In one patient, a portable CT scan obtained to exclude hemorrhage showed a rectus sheath hematoma that was not identified prospectively. Portable CT also revealed hemorrhage in five (5%) of 91 scans obtained for indications that did not include suspected hemorrhage. Similarly, among this group, a psoas hematoma was not reported on an initial CT but was detected on retrospective review. Thus, hemorrhage was a finding in 13% (16/122) of the portable CT scans in the series. As mentioned previously, in two cases, retrospective evaluation of portable CT scans showed hemorrhage that was not detected at the initial interpretation.
|
|
Among 18 scans obtained for an indication other than infection or hemorrhage, findings of portable CT scans supported the clinically suspected diagnosis in six cases (33%). These included one case each of advanced metastatic disease and fetal hydrops and four cases of acute pancreatitis (Figs. 7A and 7B).
|
|
Other than three scans with findings unexpectedly raising concern for infection and five unexpectedly showing hemorrhage, eight portable CT scans had eight unexpected findings. Portable abdominal CT revealed two pneumothoraces. There was one large obstructing ventral hernia and an additional case of small-bowel obstruction. One patient had shock liver, one additional patient had acute tubular necrosis, and portable CT detected unsuspected pneumatosis coli in two patients.
Impact of Portable CT on Patient Treatment
Findings on portable CT contributed to a change in patient treatment in 33
cases (27%). Thirteen patients with bleeding had correction of their
coagulopathy within 48 hr of the portable CT. In six patients, findings of
infectious colitis resulted in a change of antibiotic regimen (Figs.
8A and
8B). The finding of advancing
metastatic disease influenced the decision to withdraw support of one patient.
Portable CT led to laparotomy for seven patients.
|
|
In six patients, portable CT precipitated percutaneous catheter drainage. In two patients, intraabdominal collections depicted on portable CT were drained under portable sonographic guidance. Two other patients traveled to the radiology department for abscess drainage, either under stationary CT (1 patient) (Figs. 4A, 4B, and 4C) or under combined fluoroscopy and sonography. Two patients with distended gallbladders visualized on their portable CT scans underwent portable sonographically guided cholecystostomy for suspected acalculous cholecystitis.
Surgical and Autopsy Correlation
Within 7 days after their portable CT, 17 patients underwent either autopsy
or abdominal surgery. Ten patients underwent autopsy within 7 days of portable
CT (mean, 3 days; range, 1–7 days). Seven patients had surgery at a mean
duration of 1.2 days (0–5 days) after portable CT.
In seven (70%) of the 10 patients presented for postmortem examination, there was concordance between the portable CT and the autopsy findings. Diagnoses not visualized on CT included hepatic necrosis (n = 1) and pancreatitis (n = 1). In a third case, intraperitoneal hemorrhage was suspected on portable CT performed 5 days before the patient's death, but there was no support for this finding at autopsy.
The findings on portable abdominal CT were confirmed in five (71%) of seven patients who had a laparotomy. In two cases, colitis was not detected on preoperative portable CT, one each from ischemia and infection–necrosis. One patient underwent a laparotomy that proved to have negative findings despite a portable CT scan showing pneumatosis coli and portal venous gas (Figs. 9A and 9B).
|
|
On review of all portable CT scans, three significant findings were not seen at the time of the initial evaluation: infectious colitis with necrosis found at laparotomy as described previously (n = 1), psoas hematoma (n = 1), and rectus sheath hematoma (n = 1). In a fourth patient in whom necrotic bowel was diagnosed at laparotomy, preoperative portable CT findings were remarkably minimal, even when reviewed retrospectively.
Discussion
Since the introduction of portable CT to our hospital in 1998, use of this technology has steadily increased. Cranial CT remains the principal application for portable scanning, which accounts for the disproportionate number of neurologic patients whose abdomens also were imaged portably. Unlike cranial CT, which can be accomplished with the patient still lying in his or her ICU bed, portable acquisition of an abdominal scan requires moving the patient. Nevertheless, all the portable abdominal CT scans were acquired at the bedside, without the need of patient transport away from the ICU.
On the basis of the stated indication or indications for performing abdominal CT, portable CT confirmed one or more clinical suspicions in more than one quarter of the cases. The most common indication for portable CT was suspected intraabdominal infection, which was supported by the portable CT findings in 15 (19%) of 79 cases and detected in an additional three cases when infection was not the stated indication for the examination. This result is similar to that for stationary CT in the study of Barkhausen et al. [4], who detected an abdominal source in 16% of septic patients.
Understandably, our proportion of positive portable CT findings for infection is not as large as the 58% reported in patients after trauma [5], which represented (including post-operative patients) only 7% (9/122) of the portable CT scans in our series. It was reassuring to find that in five (71%) of seven patients in whom portable CT findings precipitated laparotomy, the suspected abnormality was confirmed. The two false-negative cases in which laparotomy was performed despite normal portable CT findings revealed ischemia and in fection–necrosis. However, these two may have represented cases in which the ischemia was predominantly on the mucosal side and therefore may have been better detected at endoscopy. One patient underwent a laparotomy that proved to have negative results despite a portable CT scan showing pneumatosis coli and portal vein gas (Figs. 9A and 9B). This case reemphasizes that pneumatosis coli and portal vein gas can occasionally be benign findings and do not always imply bowel ischemia.
Many patients who underwent portable CT subsequently left the ICU for special procedures or surgery or for other imaging procedures performed on nonportable equipment. In some cases, the portable CT forced an intervention that could not be performed at the bedside, and in others, the patients' conditions improved to the point at which they could be safely moved. Nevertheless, undoubtedly in some patients, abdominal CT was performed portably more for convenience than for true medical necessity. McCunn et al. [6] surveyed ICU physicians who had access to a portable CT scanner. Extracorporeal life support and cardiovascular, respiratory, or neurologic instability were the principal indications for performing CT portably. However, when portable CT was unavailable, these physicians willingly transported their patients to a stationary CT scanner to obtain diagnostic information.
ICU staff reluctance to transport unstable patients to the radiology department is understandable. Risks associated with transport of ICU patients to hospital locations outside the ICU have been documented in several studies [7–9], although serious injuries and deleterious effects on outcome are rare. In a comparison of neurologic ICU patients transported for stationary head CT scans with similar patients who underwent portable scans, Gunnarsson et al. [10] found that the 20–25% incidence of complications and deterioration associated with transport fell to 0–4% when the examination was performed portably. Those researchers also noted a beneficial effect on staff workload when transport could be avoided.
Besides portability, several other technical factors distinguish the portable CT instrument used in our study from conventional stationary CT scanners. The lower tube current (maximum, 50 mA) imposes a prolonged scanning time (2 sec) and results in a reduced milliampere-second compared with conventional CT. Even with high-efficiency detectors [11], both the longer scanning time and lower milliampere-second have their greatest effect in scanning the abdomen, in which the combined effects of motion and patient density are most severe. Motion sensitivity was readily apparent on many of the portable CT scans in our study and caused major artifacts in 23 (19%) of the 122 scans. In many cases, monitoring equipment or support devices such as intraaortic balloons showed characteristic artifacts (Fig. 10). Reduced milliampere-second correlates with increased image noise, the major determinant obscuring small differences in contrast [12, 13]. Among the effects of reduced contrast resolution might be diminished capacity to detect hemorrhage as abnormal hyperdensity, particularly in patients whose anemia already limits this effect. Hemorrhage was the second most common indication for portable CT in our series. Two of the three cases with missed findings visible in retrospect had hematomas, and in the third case, hemorrhagic ascites provided a clue to the diagnosis of colonic infarction subsequently found at laparotomy.
|
Battery charge and tube heating place limits not applicable to stationary units on the number of sections that can be performed on portable CT. Particularly when the chest and abdomen are to be scanned together, these factors restrict portable multiphase scanning. As a result, in a study performed for hemorrhage, the operator's choice was between unenhanced portable CT optimized for detection of blood and the benefits in scanning quality associated with IV contrast administration [14].
During this study, most stationary scans were obtained on either single-detector helical CT scanners or 4-MDCT scanners. Despite this, analysis of image quality showed that 10% of stationary scans had a major defect (Table 1). On initial analysis, this finding seems disappointing but can be explained, in most cases, by excessive respiratory motion artifacts in this severely ill cohort of patients and by streak artifact from monitoring equipment and support devices. The continued technologic refinements currently being seen with MDCT should improve the image quality of scans acquired in this patient cohort. Although spatial resolution of MDCT is equal to that of single-detector CT scanners, substantial improvement in temporal resolution with MDCT, because of markedly reduced acquisition times, reduces artifacts caused by voluntary or involuntary patient motion [15].
The value of CT for noninvasive diagnosis of acute abdominal conditions has been well established [16, 17]. However, even in the same cohort of severely ill patients, stationary nonportable abdominal CT offers a level of quality that the portable CT could not provide. Physicians caring for ICU patients must avoid thinking of portable abdominal CT as a conventional abdominal CT performed portably. Portable CT should be regarded as a different examination with special limitations that do not apply to stationary CT. Whenever possible, patients who need the valuable diagnostic information that CT can provide should be moved to a stationary CT installation for their study in preference to a portable CT scanner.
The unique imaging opportunity afforded by portable abdominal CT is itself the major limitation of this study. Although many of the positive findings on portable CT were subsequently confirmed by the intervention they required, most negative studies have no correlation. We do not know the accuracy of portable CT, especially the negative predictive value. Portable CT examinations must be interpreted with special caution. Uncertainties inherent in each scan should be part of the report communicated to the physicians caring for the patient.
Nevertheless, portable abdominal CT offers a window on the critically ill population that would not otherwise be available. Alternative portable imaging techniques do not offer reliable detection or exclusion of many conditions that CT can image very well, including bowel obstruction, interloop abscess, and many forms of hemorrhage [16]. Radiologists who interpret CT scans in major medical centers will probably see continued use of portable CT. They must learn to recognize abnormalities on CT scans of less than the usual quality and to overcome the effects of artifacts. Most important, radiologists should avoid overstating the information provided by any given portable CT scan. A balanced assessment of each study will prevent overconfident reliance on portable CT for decisions in patient treatment and encourage the use of portable CT for abdominal imaging only when transport for stationary CT would expose the most fragile patients to high risk of decompensation.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
