Utilizing Digital Pathology Informatics Algorithms for Diffuse Large B-cell Lymphoma Subtyping Público

Abstract

 

Introduction:

 

Through gene expression profiling (GEP), two biologically distinct molecular subgroups of diffuse large B-cell lymphoma (DLBCL) have been identified – germinal center B-cell (GCB) and activated B-cell – with disparate prognostic implications. Because GEP is not easily implemented in clinical practice, immunohistochemical (IHC) algorithms have been developed and validated as surrogates for DLBCL subtype. The most commonly employed algorithm uses CD10, BCL6 and MUM1 to distinguish GCB and non-GCB subtypes, but has been shown to misclassify up to 20% of cases, partially due to the significant variability that exists in IHC interpretation. Advancements in digital pathology technology have led to the development of image analysis algorithms that allow for the extraction of quantitative descriptions of pathology features such as level of IHC staining. We explore the feasibility of developing precise, objective image analysis-based scoring algorithm for DLBCL digital pathology whole slide images that can effectively distinguish between GCB and non-GCB subtype.

 

 

 

Methods:

 

Tissue slides of the immunostains CD10, BCL6, and MUM1 for 40 DLBCL patients were digitized at 40x objective resolution. We developed and trained an image analysis algorithm on the IHC whole slide images that returns the percentage of positive regions over the DLBCL tissue area. Receiver operating characteristics (ROC) curves were calculated to assess ability of image analysis-based measurements of region positivity to predict pathologist classifications as positive or negative for CD10, BCL6, and MUM1. Using thresholds for percent positive regions established by the ROC curves, the patients were classified into GCB/non-GCB by the Hans algorithm using sequential application of the identified thresholds, and concordance with pathologist classification was calculated.

 

 

 

Results:

 

Area under the ROC curve (AUC) for predicting pathologist classifications of positivity from image analysis measurements for CD10, BCL6, and MUM1 were 0.92, 1.0, and 0.95 respectively. Thresholds from the ROC curves for CD10, BCL6, and MUM1 were 13%, 15%, and 20% respectively. Using these thresholds, classification by image analysis algorithm was concordant with pathologist classification in 82.5% (κ = 0.65) of cases.

 

 

 

Conclusion:

 

The image analysis algorithm could provide an effective support tool for pathologists, improving the IHC classification of DLBCL subtype.

Table of Contents

 

INTRODUCTION...............................................................................1.. 1

BACKGROUND................................................................................2.. 2

METHODS.........................................................................................7. 7

RESULTS..........................................................................................11. 11

DISCUSSION....................................................................................13.. 13

REFERENCES..................................................................................21. 21

TABLES/FIGURES..........................................................................33. 33

Figure 1. Hans Classification System...........................................33 33

Figure 2. Image Analysis Algorithm............................................34 34

Figure 3.  Heatmap of positive region density (right) for MUM1 immunohistochemical Whole Slide Image....................................35 35

Figure 4. Receiver operator curve comparing percent positive regions from image analysis algorithm to pathologist classification for CD10...................................36 36

Figure 5. Receiver operator curve comparing percent positive regions from image analysis algorithm to pathologist classification for BCL6......................................37 37

Figure 6. Receiver operator curve comparing percent positive regions from image analysis algorithm to pathologist classification for MUM1.....................................38 38

Figure 7. Sequential application of the Hans classification system using image analysis algorithm output to predict subtype.......................................................39 39

 

Table 1. Clinical Characteristics of DLBCL patients by Hans classification subtype...40 40

40

 

 

 

 

 

 

 

 

 

 

About this Master's Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Laney Graduate School
Department	Clinical Research
Degree	M.S.
Submission	Master's Thesis
Language	English
Research Field	Biology, Bioinformatics Biology, Histology Health Sciences, Oncology
Palabra Clave	diffuse large B-cell lymphoma digital pathology immunohistochemistry image analysis
Committee Chair / Thesis Advisor	Flowers, Christopher, Emory University
Committee Members	Qayed, Muna, Emory University Klein, Mitchel, Emory University

Última modificación

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	Utilizing Digital Pathology Informatics Algorithms for Diffuse Large B-cell Lymphoma Subtyping ()	2018-04-06 13:38:16 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions