Advanced

Automated image analysis of cyclin D1 protein expression in invasive lobular breast carcinoma provides independent prognostic information

Tobin, Nicholas P.; Lundgren, Katja L.; Conway, Catherine; Anagnostaki, Lola; Costello, Sean and Landberg, Göran LU (2012) In Human Pathology 43(11). p.2053-2061
Abstract
The emergence of automated image analysis algorithms has aided the enumeration, quantification, and immunohistochemical analyses of tumor cells in both whole section and tissue microarray samples. To date, the focus of such algorithms in the breast cancer setting has been on traditional markers in the common invasive ductal carcinoma subtype. Here, we aimed to optimize and validate an automated analysis of the cell cycle regulator cyclin D1 in a large collection of invasive lobular carcinoma and relate its expression to clinicopathologic data. The image analysis algorithm was trained to optimally match manual scoring of cyclin D1 protein expression in a subset of invasive lobular carcinoma tissue microarray cores. The algorithm was capable... (More)
The emergence of automated image analysis algorithms has aided the enumeration, quantification, and immunohistochemical analyses of tumor cells in both whole section and tissue microarray samples. To date, the focus of such algorithms in the breast cancer setting has been on traditional markers in the common invasive ductal carcinoma subtype. Here, we aimed to optimize and validate an automated analysis of the cell cycle regulator cyclin D1 in a large collection of invasive lobular carcinoma and relate its expression to clinicopathologic data. The image analysis algorithm was trained to optimally match manual scoring of cyclin D1 protein expression in a subset of invasive lobular carcinoma tissue microarray cores. The algorithm was capable of distinguishing cyclin D1 positive cells and illustrated high correlation with traditional manual scoring (kappa = 0.63). It was then applied to our entire cohort of 483 patients, with subsequent statistical comparisons to clinical data. We found no correlation between cyclin D1 expression and tumor size, grade, and lymph node status. However, overexpression of the protein was associated with reduced recunrrence-free survival (P = .029), as was positive nodal status (P < .001) in patients with invasive lobular carcinoma. Finally, high cyclin D1 expression was associated with increased hazard ratio in multivariate analysis (hazard ratio, 1.75; 95% confidence interval, 1.05-2.89). In conclusion, we describe an image analysis algorithm capable of reliably analyzing cyclin D1 staining in invasive lobular carcinoma and have linked overexpression of the protein to increased recurrence risk. Our findings support the use of cyclin D1 as a clinically informative biomarker for invasive lobular breast cancer. (C) 2012 Elsevier Inc. All rights reserved. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Automated image analysis, Cyclin D1, Invasive lobular breast carcinoma, Recurrence-free survival
in
Human Pathology
volume
43
issue
11
pages
2053 - 2061
publisher
Elsevier
external identifiers
  • wos:000310654500032
  • scopus:84867880533
ISSN
1532-8392
DOI
10.1016/j.humpath.2012.02.015
language
English
LU publication?
yes
id
2e3d9542-fcb0-4dcc-bfbd-2cde087e5b6f (old id 3259384)
date added to LUP
2013-01-07 09:38:53
date last changed
2017-07-23 03:18:54
@article{2e3d9542-fcb0-4dcc-bfbd-2cde087e5b6f,
  abstract     = {The emergence of automated image analysis algorithms has aided the enumeration, quantification, and immunohistochemical analyses of tumor cells in both whole section and tissue microarray samples. To date, the focus of such algorithms in the breast cancer setting has been on traditional markers in the common invasive ductal carcinoma subtype. Here, we aimed to optimize and validate an automated analysis of the cell cycle regulator cyclin D1 in a large collection of invasive lobular carcinoma and relate its expression to clinicopathologic data. The image analysis algorithm was trained to optimally match manual scoring of cyclin D1 protein expression in a subset of invasive lobular carcinoma tissue microarray cores. The algorithm was capable of distinguishing cyclin D1 positive cells and illustrated high correlation with traditional manual scoring (kappa = 0.63). It was then applied to our entire cohort of 483 patients, with subsequent statistical comparisons to clinical data. We found no correlation between cyclin D1 expression and tumor size, grade, and lymph node status. However, overexpression of the protein was associated with reduced recunrrence-free survival (P = .029), as was positive nodal status (P &lt; .001) in patients with invasive lobular carcinoma. Finally, high cyclin D1 expression was associated with increased hazard ratio in multivariate analysis (hazard ratio, 1.75; 95% confidence interval, 1.05-2.89). In conclusion, we describe an image analysis algorithm capable of reliably analyzing cyclin D1 staining in invasive lobular carcinoma and have linked overexpression of the protein to increased recurrence risk. Our findings support the use of cyclin D1 as a clinically informative biomarker for invasive lobular breast cancer. (C) 2012 Elsevier Inc. All rights reserved.},
  author       = {Tobin, Nicholas P. and Lundgren, Katja L. and Conway, Catherine and Anagnostaki, Lola and Costello, Sean and Landberg, Göran},
  issn         = {1532-8392},
  keyword      = {Automated image analysis,Cyclin D1,Invasive lobular breast carcinoma,Recurrence-free survival},
  language     = {eng},
  number       = {11},
  pages        = {2053--2061},
  publisher    = {Elsevier},
  series       = {Human Pathology},
  title        = {Automated image analysis of cyclin D1 protein expression in invasive lobular breast carcinoma provides independent prognostic information},
  url          = {http://dx.doi.org/10.1016/j.humpath.2012.02.015},
  volume       = {43},
  year         = {2012},
}