Automatic quantification of lobular inflammation and hepatocyte ballooning in nonalcoholic fatty liver disease liver biopsies

Scott Vanderbeck, Joseph Bockhorst, David Kleiner, Richard Komorowski, Naga Chalasani, Samer Gawrieh

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Automatic quantification of cardinal histologic features of nonalcoholic fatty liver disease (NAFLD) may reduce human variability and allow continuous rather than semiquantitative assessment of injury. We recently developed an automated classifier that can detect and quantify macrosteatosis with greater than or equal to 95% precision and recall (sensitivity). Here, we report our early results on the classifier's performance in detecting lobular inflammation and hepatocellular ballooning. Automatic quantification of lobular inflammation and ballooning was performed on digital images of hematoxylin and eosin-stained slides of liver biopsy samples from 59 individuals with normal liver histology and varying severity of NAFLD. Two expert hepatopathologists scored liver biopsies according the nonalcoholic steatohepatitis clinical research network scoring system and provided annotations of lobular inflammation and hepatocyte ballooning on the digital images. The classifier had precision and recall of 70% and 49% for lobular inflammation, and 91% and 54% for hepatocyte ballooning. In addition, the classifier had an area under the curve of 95% for lobular inflammation and 98% for hepatocyte ballooning. The Spearman rank correlation coefficient for comparison with pathologist grades was 45.2% for lobular inflammation and 46% for hepatocyte ballooning. Our novel observations demonstrate that automatic quantification of cardinal NAFLD histologic lesions is feasible and offer promise for further development of automatic quantification as a potential aid to pathologists evaluating NAFLD biopsies in clinical practice and clinical trials.

Original languageEnglish
Pages (from-to)767-775
Number of pages9
JournalHuman Pathology
Volume46
Issue number5
DOIs
StatePublished - May 1 2015

Fingerprint

Hepatocytes
Inflammation
Biopsy
Liver
Hematoxylin
Eosine Yellowish-(YS)
Nonparametric Statistics
Area Under Curve
Non-alcoholic Fatty Liver Disease
Histology
Clinical Trials
Wounds and Injuries
Research
Pathologists

Keywords

  • Digital image analysis
  • Fatty liver
  • Hepatocyte ballooning
  • Lobular inflammation
  • Machine learning
  • NAFLD activity score

ASJC Scopus subject areas

  • Medicine(all)

Cite this

Automatic quantification of lobular inflammation and hepatocyte ballooning in nonalcoholic fatty liver disease liver biopsies. / Vanderbeck, Scott; Bockhorst, Joseph; Kleiner, David; Komorowski, Richard; Chalasani, Naga; Gawrieh, Samer.

In: Human Pathology, Vol. 46, No. 5, 01.05.2015, p. 767-775.

Research output: Contribution to journalArticle

Vanderbeck, Scott ; Bockhorst, Joseph ; Kleiner, David ; Komorowski, Richard ; Chalasani, Naga ; Gawrieh, Samer. / Automatic quantification of lobular inflammation and hepatocyte ballooning in nonalcoholic fatty liver disease liver biopsies. In: Human Pathology. 2015 ; Vol. 46, No. 5. pp. 767-775.
@article{9b2d31ac88dc47c8ba8bd24a9ea73f60,
title = "Automatic quantification of lobular inflammation and hepatocyte ballooning in nonalcoholic fatty liver disease liver biopsies",
abstract = "Automatic quantification of cardinal histologic features of nonalcoholic fatty liver disease (NAFLD) may reduce human variability and allow continuous rather than semiquantitative assessment of injury. We recently developed an automated classifier that can detect and quantify macrosteatosis with greater than or equal to 95{\%} precision and recall (sensitivity). Here, we report our early results on the classifier's performance in detecting lobular inflammation and hepatocellular ballooning. Automatic quantification of lobular inflammation and ballooning was performed on digital images of hematoxylin and eosin-stained slides of liver biopsy samples from 59 individuals with normal liver histology and varying severity of NAFLD. Two expert hepatopathologists scored liver biopsies according the nonalcoholic steatohepatitis clinical research network scoring system and provided annotations of lobular inflammation and hepatocyte ballooning on the digital images. The classifier had precision and recall of 70{\%} and 49{\%} for lobular inflammation, and 91{\%} and 54{\%} for hepatocyte ballooning. In addition, the classifier had an area under the curve of 95{\%} for lobular inflammation and 98{\%} for hepatocyte ballooning. The Spearman rank correlation coefficient for comparison with pathologist grades was 45.2{\%} for lobular inflammation and 46{\%} for hepatocyte ballooning. Our novel observations demonstrate that automatic quantification of cardinal NAFLD histologic lesions is feasible and offer promise for further development of automatic quantification as a potential aid to pathologists evaluating NAFLD biopsies in clinical practice and clinical trials.",
keywords = "Digital image analysis, Fatty liver, Hepatocyte ballooning, Lobular inflammation, Machine learning, NAFLD activity score",
author = "Scott Vanderbeck and Joseph Bockhorst and David Kleiner and Richard Komorowski and Naga Chalasani and Samer Gawrieh",
year = "2015",
month = "5",
day = "1",
doi = "10.1016/j.humpath.2015.01.019",
language = "English",
volume = "46",
pages = "767--775",
journal = "Human Pathology",
issn = "0046-8177",
publisher = "W.B. Saunders Ltd",
number = "5",

}

TY - JOUR

T1 - Automatic quantification of lobular inflammation and hepatocyte ballooning in nonalcoholic fatty liver disease liver biopsies

AU - Vanderbeck, Scott

AU - Bockhorst, Joseph

AU - Kleiner, David

AU - Komorowski, Richard

AU - Chalasani, Naga

AU - Gawrieh, Samer

PY - 2015/5/1

Y1 - 2015/5/1

N2 - Automatic quantification of cardinal histologic features of nonalcoholic fatty liver disease (NAFLD) may reduce human variability and allow continuous rather than semiquantitative assessment of injury. We recently developed an automated classifier that can detect and quantify macrosteatosis with greater than or equal to 95% precision and recall (sensitivity). Here, we report our early results on the classifier's performance in detecting lobular inflammation and hepatocellular ballooning. Automatic quantification of lobular inflammation and ballooning was performed on digital images of hematoxylin and eosin-stained slides of liver biopsy samples from 59 individuals with normal liver histology and varying severity of NAFLD. Two expert hepatopathologists scored liver biopsies according the nonalcoholic steatohepatitis clinical research network scoring system and provided annotations of lobular inflammation and hepatocyte ballooning on the digital images. The classifier had precision and recall of 70% and 49% for lobular inflammation, and 91% and 54% for hepatocyte ballooning. In addition, the classifier had an area under the curve of 95% for lobular inflammation and 98% for hepatocyte ballooning. The Spearman rank correlation coefficient for comparison with pathologist grades was 45.2% for lobular inflammation and 46% for hepatocyte ballooning. Our novel observations demonstrate that automatic quantification of cardinal NAFLD histologic lesions is feasible and offer promise for further development of automatic quantification as a potential aid to pathologists evaluating NAFLD biopsies in clinical practice and clinical trials.

AB - Automatic quantification of cardinal histologic features of nonalcoholic fatty liver disease (NAFLD) may reduce human variability and allow continuous rather than semiquantitative assessment of injury. We recently developed an automated classifier that can detect and quantify macrosteatosis with greater than or equal to 95% precision and recall (sensitivity). Here, we report our early results on the classifier's performance in detecting lobular inflammation and hepatocellular ballooning. Automatic quantification of lobular inflammation and ballooning was performed on digital images of hematoxylin and eosin-stained slides of liver biopsy samples from 59 individuals with normal liver histology and varying severity of NAFLD. Two expert hepatopathologists scored liver biopsies according the nonalcoholic steatohepatitis clinical research network scoring system and provided annotations of lobular inflammation and hepatocyte ballooning on the digital images. The classifier had precision and recall of 70% and 49% for lobular inflammation, and 91% and 54% for hepatocyte ballooning. In addition, the classifier had an area under the curve of 95% for lobular inflammation and 98% for hepatocyte ballooning. The Spearman rank correlation coefficient for comparison with pathologist grades was 45.2% for lobular inflammation and 46% for hepatocyte ballooning. Our novel observations demonstrate that automatic quantification of cardinal NAFLD histologic lesions is feasible and offer promise for further development of automatic quantification as a potential aid to pathologists evaluating NAFLD biopsies in clinical practice and clinical trials.

KW - Digital image analysis

KW - Fatty liver

KW - Hepatocyte ballooning

KW - Lobular inflammation

KW - Machine learning

KW - NAFLD activity score

UR - http://www.scopus.com/inward/record.url?scp=84930984362&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84930984362&partnerID=8YFLogxK

U2 - 10.1016/j.humpath.2015.01.019

DO - 10.1016/j.humpath.2015.01.019

M3 - Article

VL - 46

SP - 767

EP - 775

JO - Human Pathology

JF - Human Pathology

SN - 0046-8177

IS - 5

ER -