HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) CME Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Porter, G. J. R. Articles by Chakrabarti, J. Search for Related Content PubMed PubMed Citation Articles by Porter, G. J. R. Articles by Chakrabarti, J. Social Bookmarking                      What's this?

Influence of Mammographic Parenchymal Pattern in Screening-Detected and Interval Invasive Breast Cancers on Pathologic Features, Mammographic Features, and Patient Survival

¹ All authors: Nottingham Breast Institute, City Hospital, Hucknall Rd., Nottingham, England, NG5 1PB.

Received November 4, 2005; accepted after revision August 2, 2006.

 
CME

This article is available for CME credit. See www.arrs.org for more information.

Address correspondence to G. J. R. Porter (gareth.porter{at}phnt.swest.nhs.uk).

FOR YOUR INFORMATION

This article is available for CME credit. See www.arrs.org for more information.

Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
OBJECTIVE. The aim of our study was to assess the effect of mammographic parenchymal pattern on patient survival, mammographic features, and pathologic features of breast cancer in a screened population.

MATERIALS AND METHODS. We classified the parenchymal pattern (according to BI-RADS) of 759 screened women who presented with a screening-detected (n = 455) or interval (n = 304) invasive breast cancer. Pathologic details (tumor size, histologic grade, lymph node stage, vascular invasion, and histologic type) and mammographic appearances were recorded. Breast cancer-specific survival was ascertained, with a median follow-up of 9.0 years.

RESULTS. An excess of interval cancers was seen in women with dense breasts (p < 0.0001). Screening-detected (but not interval) tumors were significantly smaller in fatty breasts (p = 0.014). Tumor grade, lymph node stage, vascular invasion, and histologic type did not vary significantly with mammographic parenchymal pattern in screening-detected or interval cancers. Screening-detected cancers in fatty breasts were more likely to appear as indistinct (p = 0.003) or spiculated (p = 0.002) masses in contrast to cancers in dense breasts, which more commonly appeared as architectural distortions (p < 0.0001). No significant breast cancer-specific survival difference was seen by mammographic parenchymal pattern for screening-detected cancers (p = 0.75), interval cancers (p = 0.82), or both groups combined (p =0.12).

CONCLUSION. The prognosis of screened women presenting with breast cancer is unrelated to dense mammographic parenchymal pattern despite an excess of interval cancers and larger screening-detected tumors in this group. These data support the mammographic screening of women with dense parenchymal patterns.

Keywords: breast cancer • mammography • screening

Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Mammography is a well-accepted screening technique for the early detection of breast cancer and has been shown to reduce the breast cancer mortality rate in women who are 40-69 years old [1-5]. Recent studies have shown a 17% reduction in mortality rate (95% CI, 3-29%) in screened women 55-64 years old and a 27% reduction in mortality rate (13-39%) in screened women 60-69 years old [1-3].

The mammographic appearance of the female breast varies from completely fatty to extremely dense, depending on the proportion of fibroglandular tissue present. Because of the early work published by Wolfe [6, 7] in the 1970s and Ciatto and Zappa [8], we know that the mammographic parenchymal pattern or the mammographic density of the breast is related to the risk of developing breast cancer, with a two- to sixfold increase in risk in women with a dense parenchymal pattern compared with women with a fatty parenchymal pattern.

A dense parenchymal pattern is more commonly seen in young women, and breast density naturally diminishes with age. The use of hormone replacement therapy slows breast involution and increases mammographic density; this is particularly seen in women on continuous combined hormone replacement therapy [9]. Nulliparous women and women with later first pregnancies (age > 20 years) have denser breasts than multiparous women [10], and women with mutations of the BRCA1 gene have also been shown to have denser breasts than age-matched controls [11].

Mammographic parenchymal pattern has a major impact on breast screening because a dense pattern often makes identification of cancers difficult, resulting in lowered sensitivity and a higher rate of interval cancers [12]. An interval cancer is a breast cancer diagnosed in the interval between scheduled screening episodes in a woman screened and assessed to have normal screening results. It would seem plausible that the mammographic parenchymal pattern might influence the mammographic features of screening-detected breast cancer. Several groups have addressed the issue of whether parenchymal pattern influences the well-established prognostic factors of breast cancer such as nodal stage, histologic grade, size, and vascular invasion status in patients with screening-detected or interval breast cancer [13-15]. One of those studies showed that interval cancers occurring in mammographically dense breasts were more likely to be estrogen-receptor negative, histologically high-grade, and larger than interval tumors in fattier breasts [13]. Another study showed no correlation between breast density and tumor size, nodal stage, histologic grade, or estrogen-receptor status in screening-detected cancers [15].

To our knowledge, only two studies including both interval and screening-detected cancers have addressed the issue of prognostic factors and their association with mammographic parenchymal patterns. One group of authors concluded that not only does mammographic parenchymal pattern serve as an indicator of risk of breast cancer but also a dense parenchymal pattern is associated with nodal positivity and high-grade disease [16, 17]. Another study noted that increased breast density in screening-detected cancers was positively related to tumor size, nodal positivity, and vascular invasion. That study also found that increased breast density in interval cancers was associated with a low histologic grade [14]. To our knowledge, no study has considered whether breast density has any influence on the actual survival of screened patients who have screening-detected or interval breast cancer.

The aim of our study was to assess the effect of mammographic parenchymal pattern on the mammographic and pathologic features of screening-detected and interval breast cancers and subsequent survival in a screened population. We also aimed to assess any survival difference by parenchymal pattern in these groups.

Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
At our institution, a database of the pathologic and the mammographic features of all screening-detected and interval breast cancers has been prospectively maintained since the initiation of screening in 1987. During this period, women 50-64 years old have been invited for mammographic screening every 3 years, which is the standard screening interval in the British National Health Service Breast Screening Programme. Two-view mammography was performed at the first and the subsequent screenings. Women older than 64 years were not routinely invited but could attend if they wished. The mammograms were interpreted by one of several radiologists without the assistance of a computer-aided detection system.

Data regarding tumor type, grade, maximum invasive diameter, lymph node status, and the presence or absence of vascular invasion were recorded. Pathologic data were incomplete in a number of cases (Tables 1, 2 and 3). These patients were excluded from the chi-square analysis of association of that particular pathologic feature with a specific mammographic parenchymal pattern.

Tumor grade was determined using the Nottingham method described by Elston and Ellis [18]. In the analysis of lymph nodes attained at either axillary sampling (aiming to remove at least four nodes) or axillary dissection: stage 1 was node-negative (with a minimum of four nodes retrieved), stage 2 was three or fewer positive nodes, and stage 3 was four or more positive nodes. If in a node-negative patient fewer than four nodes were retrieved, that patient was excluded from nodal stage analysis. During the time period of the study, a few women (< 5) had sentinel node biopsy. If those nodes were disease-free, those patients were included in the lymph node-negative group and were not excluded from lymph node stage analysis.

Tumor histologic type was defined according to previously described protocols [19]. In our category "tumors with lobular features," we included pure lobular cancers and tumors with mixed morphology but showing lobular features.

The study group consisted of 813 patients with primary operable invasive breast cancers either detected from the prevalent (initial) and incident (subsequent) screening rounds between January 1987 and December 1997, or detected as interval cancers in patients whose last screening was during this time period. During this time, 148,080 screenings were performed. Retrieving mammograms for 54 (6.6%) patients was not possible, and those patients were therefore excluded. The remaining study population consisted of 759 patients comprising 252 prevalent, 203 incident, and 304 interval invasive cancers. Noninvasive in situ cancers were not included in this study. Ductal carcinoma in situ is almost exclusively detected mammographically by calcifications [20]. The mammographic conspicuity of calcifications is much less affected by the mammographic parenchymal pattern than other mammographic features. Therefore, these noninvasive cases were excluded.

A retrospective analysis of the mammographic parenchymal pattern was performed by one of three radiologists. The study radiologists had 1, 12, and 15 years' experience in interpreting at least 5,000 screening mammograms annually. No formal assessment of inter- or intraobserver variability was performed. The mammographic parenchymal pattern of the contralateral breast at the time of diagnosis with invasive breast cancer in screening-detected cancers, or at the time of the last screening in interval cancers, was assessed using the BI-RADS classification [21]. This system consists of parenchymal pattern 1 indicating a fatty breast; pattern 2, a fatty breast with scattered fibroglandular densities; pattern 3, a heterogeneously dense breast; and pattern 4, extremely dense parenchyma. The BI-RADS classification has proven to have good reproducibility and good inter- and intraobserver variability [22]. Women presenting with locally advanced interval breast cancer that was unsuitable for primary surgical management did not routinely undergo diagnostic mammography.

Deceased women who were known to have metastatic breast cancer were presumed to have died from breast cancer. The cause of death of other women was ascertained by reviewing patient case notes or linking with the local or national cancer registry. Patient survival was assessed by checking with the patient's general practitioner for all women who had not recently been hospitalized.

Statistical analysis was performed as follows: Differences in survival were assessed using Kaplan-Meier survival curves, with the log rank test for difference, with a median follow-up of 9.0 years. Patients who died from non-breast cancer causes were excluded from further survival analysis from the time of death. The significance of any correlations was assessed using the chi-square and chi-square for trend tests. A p value of < 0.05 was considered significant. Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS), version 12.0.1, software.

Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Seven hundred fifty-nine women were included in the study. The interval and screening-detected cancer mammographic parenchymal patterns and mean group ages are listed in Tables 4 and 5. The interval breast cancer population had a significantly more radiodense parenchymal pattern than the screening-detected group (p < 0.0001). A significant trend was seen for younger women to have a more dense mammographic parenchymal pattern in both the screening-detected and interval breast cancer groups (p = 0.004 and < 0.0001, respectively).

Table 1 shows the pathologic features of screening-detected breast cancer by parenchymal pattern. This shows that women with denser mammograms have slightly larger tumors at detection (p = 0.01) but show no difference with regard to lymph node stage, the presence of vascular invasion, histologic grade, or tumor type.

Table 2 shows the relationship between parenchymal pattern and the pathologic features in interval breast cancers. The tumor size, grade, stage, and presence of vascular invasion do not differ with mammographic parenchymal density.

Table 3 details the pathologic features of both groups combined, again showing tumors occurring in dense breasts to be slightly larger (p = 0.01).

Table 6 shows the relationship between the presence of a mammographic feature and the parenchymal pattern of screening-detected cancers. A cancer may exhibit more than one mammographic feature. Table 7 lists the mammographic features by parenchymal pattern in interval cancers. Table 8 shows the mammographic features of both groups of cancers combined. These tables indicate that cancers detected in dense breasts are significantly more likely to manifest as an architectural distortion, whereas tumors detected in fatty breasts are more frequently identified as spiculated or indistinct masses.

Figure 1 shows that no significant breast cancer-specific survival difference by mammographic parenchymal pattern is seen for screening-detected cancers (p = 0.75) as assessed by Kaplan-Meier survival curves. Figure 2 shows survival curves for interval cancers by parenchymal pattern, which also show no significant survival difference (p = 0.82). Figure 3 shows both groups combined, again showing no difference in breast cancer-specific survival (p = 0.12). A nonsignificant trend was seen for women with denser breasts to have a better survival than women with radiographically fatty breasts.

View larger version (15K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1 —Survival by mammographic parenchymal pattern for screening-detected cancers.

View larger version (13K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2 —Survival by mammographic parenchymal pattern for interval cancers.

View larger version (14K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3 —Survival by mammographic parenchymal pattern for screening-detected and interval cancers combined.

Top Abstract Introduction Materials and Methods Results Discussion References

The reduced sensitivity of mammography in women with dense breasts has raised the suggestion that screening mammography may be of less benefit in that group. Our study does not support this view because women with dense parenchymal patterns had breast cancers with essentially similar pathologic prognostic features to those in women with fattier parenchymal patterns. The only significant difference was that women with a dense parenchymal pattern had screening-detected tumors that were slightly larger at the time of detection. This size difference was too small to cause a difference in either lymph node stage or lymphovascular invasion.

In addition, size has been shown to be relatively less important as a prognostic feature than histologic grade or nodal status in the established prognostic index, the Nottingham Prognostic Index [20]. Therefore, it is to be expected and we have found that survival of women with screening-detected breast cancer is similar to those with interval cancer, regardless of which mammographic pattern is seen. It is encouraging that screening-detected breast cancers in women with dense breast have an equally good prognosis as cancers detected in fatty breasts.

Our study also showed an excess of interval cancers in women with dense breasts. Our study did not show any difference in prognostic features by parenchymal pattern in this group. It is therefore not surprising that mammographic density was not associated with survival in the interval cancer group.

The mammographic features that enable detection of cancers at mammographic screening or at interval presentation are different in patients with different mammographic parenchymal patterns despite their pathologic features being similar. The higher proportions of spiculated masses found in fatty breasts and architectural distortions found in dense breasts are likely because in a dense breast a spiculated mass will appear as an architectural distortion with the central mass obscured by parenchymal density (Fig. 4), whereas in a fatty breast the central mass will not be hidden and the spiculated mass will be observed (Fig. 5). Indistinct masses are easy to see in fatty breasts but are obscured in dense breast tissue because they lack the more specific features of malignancy of either calcifications or spiculations. Calcifications and indistinct masses are common features of high-grade invasive cancer [27, 28]. In our study, we have shown that indistinct masses are less commonly seen in dense breasts. Therefore, it might seem logical to expect fewer screening-detected high-grade tumors in dense breasts than in fatty breasts in our study because of masking of ill-defined masses by dense breast tissue. However, we found no difference in the histologic grade of tumors detected in fatty and dense breasts.

View larger version (111K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4 —Carcinoma in heterogeneously dense breast (BI-RADS parenchymal pattern 3) of 55-year-old woman manifests as area of architectural distortion (arrow).

View larger version (110K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5 —Carcinoma in fatty breast (BI-RADS parenchymal pattern 1) in 59-year-old woman is shown as spiculated mass (arrow).

Two studies by Sala et al. [16, 17] found that cancers arising in dense breasts are of higher histologic grade than those arising in fatty breasts. Those studies, like ours, included interval cancers and screening-detected cancers. However, there are important differences in methodology between the studies by Sala et al. and ours that may explain the difference in findings. These studies by Sala et al. used the Wolfe mammographic parenchymal pattern classification system and also included in situ and invasive breast cancer. Ductal carcinoma in situ is almost exclusively detected mammographically by calcifications [29]. The mammographic detection of calcifications is much less likely to be affected by the background parenchymal pattern than are other mammographic features of malignancy.

A recent study by Roubidoux et al. [13] considered the prognostic factors of interval cancers occurring within 17 months of screening and found that such tumors arising in women with dense breasts were more likely to be high-grade, estrogen-receptor negative, and large when compared with such tumors arising in fatty breasts. That study also had major methodologic differences from our study. The age range of screened women varied from 34 to 86 years in the study by Roubidoux et al. compared with 50-64 years in our study. Their screening involved physical examination in addition to mammographic screening. Their screening interval was shorter, commonly being 1 year.

A previous study has shown that approximately 50% of histologic grade 3 invasive cancers detected at screening have extensive associated ductal carcinoma in situ [30]. These high-grade tumors can be detected at screening, even in dense breasts, because of mammographic calcifications. The remainder of grade 3 cancers with little or no associated ductal carcinoma in situ tend to appear as indistinct masses. In women with dense breasts, these indistinct masses may be obscured by dense tissue, and these cancers are more conspicuous in a fatty background at screening mammography.

In conclusion, we have shown that the prognosis of screening-detected or interval invasive breast cancer is unrelated to the mammographic parenchymal pattern. Our study supports the mammographic screening of women with mammographically dense breasts. The pathologic features of screening-detected and interval breast cancers are similar regardless of the parenchymal pattern. However, the mammographic features of screening-detected cancers vary according to the mammographic parenchymal pattern.

Acknowledgments
 
We thank L. J. Hamilton and A. R. M. Wilson for their contributions to this study.

References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Nystrom L, Andersson I, Bjurstam N, Frisell J, Nordenskjold B, Rutqvist L. Long-term effects of mammographic screening: updated overview of the Swedish randomised trials. Lancet2002; 359:909 -919[CrossRef][Medline]
2. Bjurstam N, Bjorneld L, Duffy S, et al. The Gothenburg Breast Screening Trial: first results on mortality, incidence and mode of detection for women ages 39-49 years at randomization. Cancer1997; 80:2091 -2099[CrossRef][Medline]
3. Andersson I, Janzon L. Reduced breast cancer mortality in women aged under 50: updated results from the Malmo Mammographic Screening Program. J Natl Cancer Inst Monogr 1997;22 : 63-67[Abstract/Free Full Text]
4. Tabar L, Yen M, Vitak B, Chen H, Smith R, Duffy S. Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening. Lancet2003; 361:1405 -1410[CrossRef][Medline]
5. Otto S, Fracheboud J, Looman C, et al. Initiation of population-based mammography screening in Dutch municipalities and effect on breast-cancer mortality: a systematic review. Lancet2003; 361:1411 -1417[CrossRef][Medline]
6. Wolfe J. Breast parenchymal patterns and their changes with age. Radiology 1976;121 : 545-552[Abstract]
7. Wolfe J. Breast patterns as an index of risk for developing breast cancer. AJR 1976;126 : 1130-1139[Abstract]
8. Ciatto S, Zappa M. A prospective study of the value of the mammographic patterns as indicators of breast cancer risk in a screening experience. Eur J Radiol 1993;17 : 122-125[CrossRef][Medline]
9. Laya M, Gallagher J, Schreimann J, Larson F, Watson P, Weinstein L. Effect of post-menopausal hormone replacement therapy on mammographic density and parenchymal pattern. Radiology 1995;53 : 433-437
10. Vachon C, Kuni C, Anderson K, Anderson V, Sellers T. Association of mammographically defined percent breast density with epidemiologic risk factors for breast cancer. Cancer Causes Control2000; 11:653 -662[CrossRef][Medline]
11. Huo Z, Giger M, Olepade O, et al. Computerized analysis of digitized mammograms of BRCA1 and BRCA2 gene mutation carriers. Radiology 2002;225 : 519-526[Abstract/Free Full Text]
12. Sala E, Warren R, McCann J, Duffy S, Day N, Luben R. Mammographic parenchymal patterns and mode of detection: implications for the breast screening programme. J Med Screen 1998;5 : 207-212[Abstract/Free Full Text]
13. Roubidoux M, Bailey J, Wray L, Helvie M. Invasive cancers detected after breast cancer screening yielded a negative result: relationship of mammographic density to tumor prognostic factors. Radiology. 2004;230 : 42-48[Abstract/Free Full Text]
14. Aiello E, Buist D, White E, Porter P. Association between mammographic breast density and breast cancer tumor characteristics. Cancer Epidemiol Biomarkers Prev 2005;14 : 662-668[Abstract/Free Full Text]
15. Morishita M, Ohtsuru A, Hayashi T, et al. Clinical significance of categorisation of mammographic density for breast cancer prognosis. Int J Oncol 2005;26 : 1307-1312[Medline]
16. Sala E, Solomon L, Warren R, et al. Size, node status and grade of breast tumours: association with mammographic parenchymal patterns. Eur Radiol 2000;10 : 157-161[CrossRef][Medline]
17. Sala E, Warren R, McCann J, Duffy S, Luben R, Day N. Mammographic parenchymal patterns and breast cancer natural history: a case-control study. Acta Oncol 2001;40 : 461-465[Medline]
18. Elston C, Ellis I. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology1991; 19:403 -410[Medline]
19. Ellis I, Galea M, Broughton N, Locker A, Blamey R, Elston C. Pathological prognostic factors in breast cancer. II. Histological type: relationship with survival in a large study with long-term follow-up. Histopathology 1992;20 : 479-489[Medline]
20. Blamey R. The design and clinical use of the Nottingham Prognostic Index in breast cancer. Breast 1996;5 : 156-157
21. American College of Radiology. Breast Imaging Reporting and Data System (BI-RADS), 4th ed. Reston, VA: American College of Radiology, 2003
22. Berg WA, Campassi C, Langenberg P, Sexton MJ. Breast Imaging Reporting and Data System: inter- and intraobserver variability in feature analysis and final assessment. AJR 2000;174 : 1769-1777[Abstract/Free Full Text]
23. Carlile T, Thompson DJ, Kopecky KJ, et al. Reproducibility and consistency in classification of breast parenchymal patterns. AJR 1983; 140:1 -7[Abstract/Free Full Text]
24. Harvey J, Bovbjerg V. Quantitative assessment of mammographic breast density: relationship with breast cancer risk. Radiology 2004;230 : 29-41[Abstract/Free Full Text]
25. Rosenberg R, Hunt W, Williamson M, et al. Effects of age, breast density, ethnicity, and estrogen replacement therapy on screening mammographic sensitivity and cancer stage at diagnosis: review of 183,134 screening mammograms in Albuquerque, New Mexico. Radiology1998; 209:511 -518[Abstract/Free Full Text]
26. Persson I, Thurfjell E, Holmberg L. Effect of estrogen and estrogen-progestin replacement regimens on mammographic parenchymal density. J Clin Oncol 1997;15 : 3201-3207[Abstract]
27. Lamb P, Perry N, Vivvicombe S, Wells C. Correlation between ultrasound characteristics, mammographic features and histological grade in patients with invasive ductal carcinoma of the breast. Clin Radiol 2000; 55:40 -44[CrossRef][Medline]
28. De Nunzio M, Evans A, Pinder S, et al. Correlations between the mammographic features of prevalent round screen-detected invasive breast cancer and pathological prognostic factors. Breast1997; 6:146 -149[CrossRef]
29. Evans AJ, Pinder S, Ellis O, et al. Screening-detected and symptomatic ductal carcinoma in situ: mammographic features with pathologic correlation. Radiology 1994;191 : 237-240[Abstract/Free Full Text]
30. Evans AJ, Pinder SE, Snead DR, Wilson AR, Ellis IO, Elston CW. The detection of ductal carcinoma in situ at mammographic screening enables the diagnosis of small, grade 3 invasive tumours. Br J Cancer 1997; 75:542 -544[Medline]
31. Elston C, Ellis I. Systemic pathology, vol. 13. The breast, 3rd ed. Edinburgh, Scotland: Churchill Livingstone, 1998:313 -315
32. NHSBSP guidelines for pathology reporting in breast disease (a joint document incorporating the 3rd edition of the NHS Breast Screening Programme's guidelines for pathology reporting in breast cancer screening and the 2nd edition of The Royal College of Pathologists minimum dataset for breast cancer histopathology). NHSBSP publication no. 58. London, England: National Health Service Breast Screening Programme,2005

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				K. Ghosh, K. R. Brandt, T. A. Sellers, C. Reynolds, C. G. Scott, S. D. Maloney, M. J. Carston, V. S. Pankratz, and C. M. Vachon Association of Mammographic Density with the Pathology of Subsequent Breast Cancer among Postmenopausal Women Cancer Epidemiol. Biomarkers Prev., April 1, 2008; 17(4): 872 - 879. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) CME Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Porter, G. J. R. Articles by Chakrabarti, J. Search for Related Content PubMed PubMed Citation Articles by Porter, G. J. R. Articles by Chakrabarti, J. Social Bookmarking                      What's this?