| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
ABSTRACT |
Tuesday, May 8, 1:30 PM–3:30 PM
Abstracts 119-129
Moderator(s): Jessica Leung and Alfred Watson
1:30 PM
Keynote Address: Outcomes Research in Breast Imaging: A Retrospective
Ed Sickles, San Francisco, CA
119. Evaluation of Breast Cancer with a Computer-Aided Detection System and Computed Radiography for Mammography
1:40 PM
Brem R. F.1*; Hoffmeister J.2; Rapelyea J. A.1; Knapp J.2; Worrell S.2 1. George Washington University Medical Center, Washington, DC; 2. iCAD, Nashua, NH
Address correspondence to R. Brem (rbrem{at}mfa.gwu.edu)
Objective: To evaluate the performance of a computer-aided detection (CAD) system with computed radiography (CR) for mammography in the detection of breast cancer based on breast density, mammographic appearance, histopathology and lesion size.
Materials and Methods: Fifty-three cases with breast cancer from clinical trials conducted at 2 sites with CR mammography (Fuji CR for Mammography) were evaluated by a CAD system (iCAD Second Look®, version 7.2). The sensitivity of cancer detection by CAD and the corresponding mammographic density, appearance mammographically as masses or calcifications, cancer histopathologies and cancer sizes were assessed. Breast Imaging Reporting and Data System (BI-RADS®) 1 and 2 density cases were classified as nondense breasts; BI-RADS 3 and 4 density cases were classified as dense breasts. Normal cases (n = 36) were used to determine the system false-positive (FP) rate.
Results: Overall, 47 (89%) of 53 cancer cases were detected by the CAD system; 30/33 (91%) cancers in nondense breasts and 17/20 (85%) cancers in dense breasts were detected. CAD detected 92% (11/12) of cancers manifesting as calcifications and 88% (36/41) as masses. Twenty-two (54%) of 41 masses were spiculated and CAD correctly marked 91% (20/22). For invasive ductal carcinoma (n = 35) the detection rate was 89%; invasive lobular and other carcinomas (n = 5) 80%; DCIS (n = 13) 92%. CAD sensitivity for cancers 1–10 mm (n = 18) was 83%; 11–20 mm (n = 17) was 88%; 21–30 mm (n = 12) was 92%; and greater than 30 mm (n = 6) was 100%. The CAD system FP rate was 2.2 marks per 4-image case.
Conclusion: CAD had a high sensitivity of 89% with CR mammography that was maintained even in conditions that may lower the sensitivity of mammography - 85% sensitivity with dense breasts and 83% sensitivity for the smallest lesions 1 cm or less in size. CAD was also highly effective in detecting mammographic appearances that are clinically important - 92% of calcifications and 91% of spiculated masses. Therefore, the use of CAD with CR mammography will be effective in assisting radiologists with the detection of breast cancer.
120. Computer-Aided Detection with Full-Field Digital Mammography and Screen-Film Mammography
Melton A. R.1*; Worrell S. W.2; Knapp J.2; Hoffmeister J. W.2 1. Columbia University Medical Center, New York, NY; 2. iCAD, Inc., Nashua, NH
Address correspondence to A. Melton (am2454{at}columbia.edu)
Objective: To evaluate the performance of a computer-aided detection (CAD) system with full-field digital mammography (FFDM) compared with screen-film mammography (SFM).
Materials and Methods: Forty-five cases with breast cancer detected with FFDM (GE Healthcare Senographe® 2000D) and 899 detected with SFM were evaluated by a CAD system (iCAD Second Look®, version 7.2). The CAD system was developed with CAD algorithms that adapt to the characteristics of FFDM and SFM, so the system performance was expected to be consistent with both mammographic exams. The sensitivity of cancer detection by CAD was assessed with FFDM and SFM based on mammographic appearance as masses or calcifications, where architectural distortions and focal asymmetric densities were included with masses. Normal cases with FFDM (n = 38) and SFM (n = 147) were used to compare the system false-positive (FP) rate.
Results: The sensitivity of CAD with FFDM was 89%, with CAD detecting 40 of 45 cancers, and CAD sensitivity with SFM was 90%, where CAD detected 809 of 899 cancers. Therefore, the detection rate of CAD with FFDM and SFM was the same (p = 0.81). CAD with FFDM detected 100% (8/8) of cancers appearing mammographically as calcifications and 86% (32/37) as masses. CAD correctly marked 88% (29/33) of masses that were spiculated. The CAD system FP rate was 1.6 marks per 4-image case with FFDM and 2.0 with SFM.
Conclusion: Although the effectiveness of CAD with SFM has been extensively reported in the scientific literature, there is less available for CAD with FFDM. Since CAD was designed to be consistent with FFDM and SFM, system performance is expected to be similar. This study shows that CAD sensitivity is the same with FFDM and SFM and that FP rate is similar. Thus, the results indicate that CAD with FFDM will enhance a radiologist's effectiveness in detecting breast cancer consistent with the enhancement demonstrated for CAD with SFM.
121. Breast Specific Gamma Imaging as an Adjunct Imaging Modality for the Diagnosis of Breast Cancer
Brem R. F.*; Floerke A. C.; Rapelyea J. A.; Teal C.; Kelly T.; Mathur V. The George Washington University, Washington, DC
Address correspondence to R. Brem (rbrem{at}mfa.gwu.edu)
Objective: To determine the sensitivity and specificity of breast-specific gamma imaging (BSGI) for the detection of breast cancer.
Materials and Methods: Retrospective review of 146 patients undergoing BSGI and breast biopsy was performed. Patients underwent BSGI with intravenous injection of 30 mCi of 99mTc-sestamibi and were imaged in CC and MLO projections. Studies were classified as positive (focal increased radiotracer uptake) or negative (no focal increased radiotracer uptake or scattered heterogeneous physiologic uptake) and compared to biopsy results. The sensitivity, specificity, PPV, and NPV were determined.
Results: Among the 146 patients, 167 lesions underwent biopsy which demonstrated 83 malignant lesions; 16 DCIS and 67 invasive cancers. Of the 84 nonmalignant lesions, 82 were benign and 2 showed atypical histology (1 atypical lobular hyperplasia [ALH] and 1 lobular carcinoma in situ [LCIS]). BSGI detected cancer in 80 of 83 cancer cases (DCIS or invasive cancer) with a sensitivity of 96.4% (95% CI, 89%–99%) and correctly identified as negative for cancer 50 of 84 cases without cancer with a specificity of 59.5% (95% CI 48–70%). The PPV for 80 of 114 lesions was 70.2% (95% CI, 60%–78%) and the NPV for 50 of 53 lesions was 94.3% (95% CI, 84%-98%).The smallest invasive cancer detected was 1 mm and the smallest DCIS detected was 1 mm BSGI detected occult cancer not visualized with mammography or sonography in 6 patients.
Conclusion: Breast-specific gamma imaging has high sensitivity and moderate specificity in detecting breast cancers and detected 6 occult cancers not found with clinical examination or other imaging modalities.
122. Architectural Distortion Identified at Mammography: Correlation with Imaging Outcome and Pathology
Leung J. W.*; Sickles E. A.; Chen J.; Lvoff N.; Brenner R. J. University of California, San Francisco, San Francisco, CA
Address correspondence to J. Leung (Jessica.Leung{at}ucsfmedcftr.org)
Objective: To determine the frequency, imaging outcome, and pathologic diagnosis of architectural distortion identified at mammography.
Materials and Methods: This is a retrospective cohort study of data collected over 20 years. At time of interpretation, the principal mammographic finding of each abnormal mammographic exam was recorded by the radiologist in a database. Our study cohort consisted of all cases with the recorded finding of architectural distortion. We examined radiology records to obtain information on the use of ultrasound and MR imaging as adjunctive diagnostic tools and pathology records to determine histologic diagnosis.
Results: Architectural distortion was recorded for 368 of 192,063 consecutive screening (0.19%) and 121 of 31,590 consecutive diagnostic exams (0.38%). Four screening exams and 1 diagnostic exam were excluded for failure to comply with management recommendations. Biopsy was performed for 78 (21%) of the 364 screening exams. Fifty-four cancers were identified, resulting in a positive predictive value (PPV1) of 15% (54/364). Biopsy was performed for 110 (92%) of the 120 diagnostic exams. Sixty cancers were identified, resulting in a PPV1 of 55% (60/110). Nine cancers identified at diagnostic exam involved screening recalls, so 105 cancers (54 screen-detected lesions; 51 with clinical symptoms) were identified in total. Histologic details were available for 73 of these 105 cancers: invasive ductal (34), lobular (23), tubular (6), mixed ductal and lobular (5), ductal carcinoma in-situ (3), mixed tubular and lobular (1), angiosarcoma (1). The most common benign histologic diagnoses were radial scar (17), stromal fibrosis (12), sclerosing adenosis (7). Ultrasound was performed for 60 of the 105 cancers: solid mass (33), hypoechoic focus or distortion of tissue (22), no sonographic correlate (5). Of the 13 cancers that underwent MR imaging, all demonstrated MR abnormalities: non-mass enhancement (7), enhancing mass (6).
Conclusion: Architectural distortion is more likely to be malignant when identified at diagnostic mammography than screening (PPV1 of 55% versus 15%). However, even when detected at screening mammography, the likelihood of malignancy is sufficiently high to justify biopsy. The absence of a sonographic correlate did not exclude malignancy. MRI appears to be sensitive for cancer in cases of architectural distortion detected at mammography.
123. Utility of the 6-month Mammogram Following a Concordant Benign Breast Biopsy
Fowler A.*; Salkowski L.; Burnside B.; Sisney G. University of Wisconsin School of Medicine & Public Health, Madison, WI
Address correspondence to A. Fowler (afowler{at}wisc.edu)
Objective: In a comprehensive breast imaging facility, breast masses and microcalcifications are routinely biopsied and assessed for pathologic concordance with the original imaging findings. It has been traditionally practiced, based on limited evidence, that all patients with concordant benign biopsies have 6-month short-term surveillance to monitor for stability. This study aims to evaluate the utility of 6-month interval diagnostic mammography of benign breast lesions that demonstrate radiographic and pathologic concordance.
Materials and Methods: Our experimental design consists of an IRB-approved retrospective chart review. We included all cases that met the following criteria: (1) breast biopsies (stereotactic and ultrasound-guided) performed at the University of Wisconsin Department of Radiology Breast Center from November 2001 to December 2002 yielding benign findings that were concordant with findings by mammography and/or ultrasound, and (2) patients have a least a 12-month follow-up mammogram after biopsy.
Results: We found that 364 concordant benign biopsies in 334 patients were performed during the time frame of our study consisting of 159 masses and 205 microcalcifications. A total of 238 of these biopsies met our inclusion criteria having follow-up of at least one year (range 11–57 months, average 37.2 months) and, of these, 103 (43.3%) had a 6-month follow-up mammogram. Analysis of the 238 biopsies meeting our inclusion criteria revealed that no additional interventions were performed at the biopsy site at 6 months. At 12-months, only one patient had re-biopsy of the previous biopsy site for increasing microcalcifications again revealing benign results. No masses required additional biopsy. No cancers developed at the original benign biopsy site.
Conclusion: In our study, 6-month mammographic follow-up after a benign concordant biopsy for masses or microcalcifications did not change mammo-graphic or clinical management. No cancers developed at the site of initial biopsy. Therefore, based on our results, it may be reasonable to resume annual mammography after a benign concordant image guided biopsy. Prospective investigation would be helpful to validate our results because elimination of short-term mammographic surveillance following a concordant breast biopsy has the potential to streamline patient management protocols, decrease health care costs, and reduce patient anxiety without compromising quality.
124. High Mammographic Density Masks Prevalent Cancers and is Predictive of Future Breast Cancer Risk
Smith D. N.1*; Colditz G. A. 3; Tamimi R. M.2 1. Brigham and Women's Hospital, Boston, MA; 2. Channing Laboratory, Department of Medicine, Boston, MA; 3. Washington University, St. Louis, MO
Address correspondence to D. Smith (darrell_smith{at}hms.harvard.edu)
Objective: Increased mammographic density impedes the detection of breast cancer by mammography. This phenomenon interferes with the quantification of mammographic density as an independent risk factor for breast cancer. We used digitized analog mammograms from The Nurses' Health Study to determine risk of breast cancer due to increased mammographic density.
Materials and Methods: We carried out a prospective nested case-control study within The Nurses' Health Study with 857 breast cancer cases and 1444 matched controls. We examined the association of absolute and percent mammographic density with risk of breast cancer. We conducted a stratified analysis by time since last mammogram. Relative risk were adjusted for age, body mass index, parity/age at first birth, family history of breast cancer, previous benign breast disease, menopausal status, postmenopausal hormone use, alcohol consumption, age at menopause and menarche.
Results: Both percent mammographic density and absolute mammographic density were strong independent breast cancer risk factors. Women in the highest quintile of absolute and percent mammographic density were at a significant increased risk of breast cancer relative to those in the lowest (Percent density: OR = 4.1 95% CI 2.9–5.8, p for trend < 0.0001; absolute density: OR = 2.7, 95%CI 2.0–3.6; p for trend < 0.0001). The association between percent mammographic density and breast cancer risk stratified by time since mammogram was highest 1-2 year since mammography (RR = 5.5; 95% CI 1.9–16.1) comparing women in the highest quartile with those in the lowest. The relation between percent mammographic density and breast cancer risk persisted at more than 5 years since time of mammogram with (comparable RR = 2.7, 95% CI 1.6–4.8).
Conclusion: Although mammographic density is predictive of future breast cancer risk for 5+ years, these results also suggest that high mammographic density is masking prevalent cancers. Additional screening modalities need to be identified for women with high breast density.
125. Minimal Screening Mammography Recall Rates Using Standardized BI-RADS Criteria
Ghate S. V.*; Kim C.; Johnson K.; Baker J. A.; Soo M. C. Duke University Medical Center, Durham, NC
Address correspondence to S. Ghate (ghate001{at}mc.duke.edu)
Objective: To retrospectively determine minimal recall rates by following standardized BI-RADS lexicon and recall criteria for lesions seen on screening mammography.
Materials and Methods: As part of an ongoing study, between July 6, 2000 and October 26, 2000, 1546 screening mammograms were performed at our institution. Of these, 180/1546 (12%) women (mean age 55, range 34–91) with 223 lesions were recalled for additional imaging. Three independent radiologists, blinded to outcome, retrospectively classified lesions using standard BI-RADS lexicon. The lesion was considered a "strict" recall if 2 of 3 radiologists determined the lesion was classifiable using the BI-RADS lexicon and met strict criteria for recall (new, increasing or baseline-detected mass/calcification/asymmetry or distortion). Remaining lesions ("soft" recalls) were subcategorized into 1) lesions that 2 of 3 radiologists would recall but could not be strictly categorized according to BI-RADS lexicon; 2) lesions classifiable by BI-RADS but not recalled because typically benign or stable; 3) no definite lesion. Recall rates were determined for both "strict" and "soft" recall groups. Cancer detection rates and false negative rates for each group were calculated by pathology correlation or by determining at least two years stability. Agreement between readers was determined with kappa statistics.
Results: Two of 3 readers determined that 88 patients with 111 lesions (41 masses, 32 calcifications, 1 distortion, 37 asymmetries) met specific BI-RADS lexicon criteria resulting in a strict recall rate of 6%. Thirty five patients (2%) with 46 questionable lesions required recall although they could not strictly be categorized according to the BI-RADS lexicon. At least 2 of 3 readers determined that 4 patients (0.3%) with 6 stable or characteristically benign BI-RADS lexicon lesions did not require recall. In 53 patients (3%) with 60 questioned lesions, two of three readers could not detect a lesion for recall. Seven (0.45%) malignancies were detected in the strict BI-RADS lexicon recall group, and no cancers found in the others. Moderate agreement between readers was determined with an average kappa of 0.50.
Conclusion: Adhering to strict BI-RADS criteria reduced screening recalls by 50%, to the currently accepted rate of 5–10%, without change in the cancer detection rate.
126. Aids to Interpretation of Screening Mammography, Survey Results
Haygood T.; Whitman G.*; Moore C. A.; Atkinson E. N.; Nikolova R. G.; Dempsey P. M. D. Anderson Cancer Center, Houston, TX
Address correspondence to G. Whitman (gwhitman{at}di.mdacc.tmc.edu)
Objective: There is variability in the use of interpretive aids in screening mammography. Further studies are needed to assess the efficacy of various interpretive aids in film-screen and digital mammography.
Materials and Methods: Using e-mail, an 18-question survey was sent to 1703 members of the Society of Breast Imaging. The survey requested information about participants' use of computer assisted detection (CAD), magnification, and alteration of hanging protocols as aids to interpretation in screening mammography.
Results: Of 1703 surveys sent out, 204 were undeliverable, and one person replied that he was not a radiologist. Out of 1497 eligible respondents, 382 (26%) completed the survey. Of these 382, 187 (49%) interpret digital screening mammograms. Among those who read digital screening mammograms, 62% use computer-based magnification on 75% or more of their cases, 12% on 50–74% of cases, 11% on 25–49% of cases, 13% on 1–24% of cases, and 3% never use computer-based magnification. Values for use of optical magnification in digital mammography were 22% in >75%, 7% in 50–74%, 8% in 25–49%, 20% in 1–24%, and 43% never. 87% of respondents use optical magnification 75% of the time or more for film-screen mammograms. The most popular hanging methods for digital mammograms were one or two images per screen, with 75% and 72%, respectively, using those options as part of their routine screening display (participants were allowed to indicate more than one choice). 44% display images 4 on 1, either alone or in combination with other methods. For digital mammography, 68% of respondents use CAD 50–74% of the time, and for film-screen mammograms, 46% of respondents use CAD 50–74% of the time. For both modalities, this choice was more common than any other response.
Conclusion: There is a wide variation in the interpretive aids used by Society of Breast Imaging members in screening mammography. The majority use either computer-based or optical magnification most of the time. Relatively few never use it, but a sizable minority use it selectively for problem-solving. Though computer-based magnification is used more often, the majority of respondents (57%) use optical magnification at least occasionally in interpretation of digital mammograms. Since the majority of respondents use CAD 50–74% of the time, this suggests that the breast imagers use CAD selectively.
127. The Role of "Second Look" of Imaging Studies of Patients Presenting with Newly Diagnosed In Situ or Invasive Breast Cancer
Koelliker S.1*; Mainiero M. B.1; Barlow P.2; Dizon D.2; Gass J.2 1. Rhode Island Hospital/Brown Medical School, Providence, RI; 2. Women & Infants Hospital/Brown Medical School, Providence, RI
Address correspondence to S. Koelliker (skoelliker{at}lifespan.org)
Objective: To assess the role of the radiologist's "second look" of previously interpreted imaging studies for tumor board presentation in patients presenting with in situ or invasive breast cancer.
Materials and Methods: Tumor board notes of 416 consecutive patients presented at the institutional multidisciplinary breast tumor board, with newly diagnosed in situ or invasive breast cancer, were retrospectively reviewed. For each tumor board, the radiologist had carefully reviewed all related imaging studies and formally presented each patient. Patients with tumor board notes which recorded a radiologist recommendation for additional evaluation, biopsy, or more extensive surgery were further investigated with medical record review and/or film evaluation. Results were assessed as to change in management. The study period occurred immediately before breast MR was readily available and routinely adopted at our institution.
Results: 26 patients of 416 (6.3 %) were noted to need further investigation. In 5 patients, complete records could not be obtained or patients were lost to follow-up. In 21 patients, additional imaging and/or biopsy was performed as suggested. 14 underwent additional image-guided core needle biopsy, 5 received additional imaging only demonstrating no additional disease, 1 received ultrasound guided cyst aspiration with benign outcome, and 1 received surgery including an additional area in the lumpectomy which proved negative. Of the 14 (3.4%) patients undergoing additional core needle biopsy, 4 had benign histology, 2 had atypical ductal hyperplasia requiring excision which proved no further disease, 5 had additional ductal carcinoma in situ (DCIS), and 3 had additional invasive cancer. Overall, 1.9% (8 patients) proved to have significant additional disease (DCIS or invasive cancer) which required a change in management. 5 of these patients had multicentric ipsilateral disease leading to unilateral mastectomy, and 3 patients had contralateral disease necessitating contralateral lumpectomy.
Conclusion: "Second look" imaging review prior to tumor board presentation of patients presenting with newly diagnosed in situ or invasive breast cancer detected additional significant disease in 1.9% of patients.
128. Will Screening Mammography Save My Life?
Keen J. D.1*; Keen J. E. 1. John H. Stroger Hospital of Cook County, Chicago, IL
Address correspondence to J. Keen (jkeen{at}ccbhs.org)
Objective: Women often hear that "mammography saves lives" through a "relative mortality risk reduction" from screening. In the interest of promoting informed medical decision making, we wanted to calculate the absolute benefit of screening mammography, and to analyze the often advertised claim that "early detection saved my life."
Materials and Methods: We obtained the current cumulative probability of developing and dying from breast cancer over the next 15 years from ages 40 to 60 using national statistics. We adjusted this mortality risk upward to account for the positive effects of screening to get a current baseline absolute risk of breast cancer death for an average risk woman. We applied plausible relative mortality risk reductions (RRR) from repeated screening (10–25%) to get the absolute risk reduction (ARR) from repeated and baseline screening. The number needed to screen (NNS) to save one life is the inverse of the ARR.
Results: Assuming a RRR of 20%, the ARR of dying from breast cancer over the next 15 years for a 40-year-old woman is 1 life extended (or "saved") per 1000 women screened repeatedly. This increases to 1.8 lives extended at age 50, and 2.4 lives extended at age 60. For age 50, a 10% RRR extends 0.9 lives per 1000 screened repeatedly, while a 25% RRR extends 2.2 lives. Assuming 80% cumulative sensitivity, repeat screening extends the lives of 2% (10% RRR) to 5% (25% RRR) of women with screen detected breast cancers found over 15 years. For a baseline scan at age 50, 4.4% of women with cancer have their lives extended.
Conclusion: The absolute benefit of screening mammography increases with age and the RRR. Earlier treatment extends the lives of under 5% of women ages 40 to 60 over 15 years who have their breast cancer detected at screening. The total cost (including false positives and overdiagnosis) of repeated screening of 560 50-year-old women needs to be weighed against the absolute benefit of one life "saved." The best way to convey this information to women requires further research.
129. Breast Imaging in Adolescents
Joe A. I.1; Haugen S.2; Blane C. E.1*; Nees A.1; Helvie M. A.1 1. University of Michigan, Ann Arbor, MI. 2. Grand Rapids, MN
Address correspondence to C. Blane (cblane{at}umich.edu)
Objective: To report the Breast Imaging findings in adolescent patients under the age of 20 referred for evaluation of an abnormal clinical examination in this era of increased utilization of ultrasound (US).
Materials and Methods: IRB approval was obtained. Imaging and clinical records of 148 patients (age range 8–19 years) referred for breast imaging between 2000 and 2006 at a single institution, were retrospectively reviewed. The age, gender, clinical symptoms, imaging findings, biopsy results, pharmacologic contraception (C) history, and final outcome were recorded.
Results: 148 patients, 133 women and 16 men, fulfilled study criteria. Mean age was 15.7 years and median age 16 years. Reasons for referral included 112 (76 %) for mass, 21 (14%) pain, 20 (14%) nipple discharge, 8 (5%) thickening, 6 (4%) inflammation. Sonographic findings were 72/148 (49%) normal or gynecomastia, 49/148 (33%) solid masses, 22/148 (15%) cysts, and 6/148 (4%) other. 32/49 solid masses went on to biopsy. There were 2 malignant neoplasms (1 secretory cancer, 1 lymphoma) and the rest were benign, most commonly fibroadenoma 26/32 (81%). The US findings of the malignant neoplasms were suspicious. 22/148 (15%) of the patients had cysts identified on US. Data on PC was limited. There were more patients (10/27) on PC with the diagnosis of fibroadenoma than with cysts (2/22). In the men US demonstrated findings consistent with gynecomastia in 12/16 (75%) and US was normal in 4/16 (25%). Biopsy of 3 men revealed gynecomastia (2) and chronic inflammation (1).
Conclusion: While fibroadenoma remains the most common lesion, there were two (1.4%) malignancies. Ultrasound findings were suspicious in both cases. Compared to the earlier literature there is a marked increase in observed cysts in this population. This may in part be due to technological advances in instrumentation in US, increased utilization of US in this patient population or a true change in the distribution of cystic disease. Although we postulated this may be secondary to earlier and more prevalent use of PC we were unable to capture enough data to assess this influence.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
