Accurate histopathology from low signal-to-noise ratio spectroscopic imaging data

Rohith K. Reddy, Rohit Bhargava

Research output: Contribution to journalArticle

Abstract

Fourier Transform Infrared (FT-IR) spectroscopic imaging is emerging as an automated alternative to human examination in studying development and disease in tissue. The technology's speed and accuracy, however, are limited by the trade-off with signal-to-noise ratio (SNR). Signal processing approaches to reduce noise have been suggested but often involve manual decisions, compromising the automation benefits of using spectroscopic imaging for tissue analysis. In this manuscript, we describe an approach that utilizes the spatial information in the data set to select parameters for noise reduction without human input. Specifically, we expand on the Minimum Noise Fraction (MNF) approach in which data are forward transformed, eigenimages that correspond mostly to signal selected and used in inverse transformation. Our unsupervised eigenimage selection method consists of matching spatial features in eigenimages with a low-noise gold standard derived from the data. An order of magnitude reduction in noise is demonstrated using this approach. We apply the approach to automating breast tissue histology, in which accuracy in classification of tissue into different cell types is shown to strongly depend on the SNR of data. A high classification accuracy was recovered with acquired data that was ∼10-fold lower SNR. The results imply that a reduction of almost two orders of magnitude in acquisition time is routinely possible for automated tissue classifications by using post-acquisition noise reduction.

Original languageEnglish (US)
Pages (from-to)2818-2825
Number of pages8
JournalAnalyst
Volume135
Issue number11
DOIs
StatePublished - Nov 1 2010

Fingerprint

histopathology
Signal-To-Noise Ratio
signal-to-noise ratio
Noise
Signal to noise ratio
Tissue
Imaging techniques
Noise abatement
Histology
histology
signal processing
automation
Automation
trade-off
Fourier transform
Fourier Analysis
Fourier transforms
Signal processing
gold
tissue

ASJC Scopus subject areas

  • Analytical Chemistry
  • Biochemistry
  • Environmental Chemistry
  • Spectroscopy
  • Electrochemistry

Cite this

Accurate histopathology from low signal-to-noise ratio spectroscopic imaging data. / Reddy, Rohith K.; Bhargava, Rohit.

In: Analyst, Vol. 135, No. 11, 01.11.2010, p. 2818-2825.

Research output: Contribution to journalArticle

@article{194f8ba0b72f4485a2533a58dba62015,
title = "Accurate histopathology from low signal-to-noise ratio spectroscopic imaging data",
abstract = "Fourier Transform Infrared (FT-IR) spectroscopic imaging is emerging as an automated alternative to human examination in studying development and disease in tissue. The technology's speed and accuracy, however, are limited by the trade-off with signal-to-noise ratio (SNR). Signal processing approaches to reduce noise have been suggested but often involve manual decisions, compromising the automation benefits of using spectroscopic imaging for tissue analysis. In this manuscript, we describe an approach that utilizes the spatial information in the data set to select parameters for noise reduction without human input. Specifically, we expand on the Minimum Noise Fraction (MNF) approach in which data are forward transformed, eigenimages that correspond mostly to signal selected and used in inverse transformation. Our unsupervised eigenimage selection method consists of matching spatial features in eigenimages with a low-noise gold standard derived from the data. An order of magnitude reduction in noise is demonstrated using this approach. We apply the approach to automating breast tissue histology, in which accuracy in classification of tissue into different cell types is shown to strongly depend on the SNR of data. A high classification accuracy was recovered with acquired data that was ∼10-fold lower SNR. The results imply that a reduction of almost two orders of magnitude in acquisition time is routinely possible for automated tissue classifications by using post-acquisition noise reduction.",
author = "Reddy, {Rohith K.} and Rohit Bhargava",
year = "2010",
month = "11",
day = "1",
doi = "10.1039/c0an00350f",
language = "English (US)",
volume = "135",
pages = "2818--2825",
journal = "Analyst",
issn = "0003-2654",
publisher = "Royal Society of Chemistry",
number = "11",

}

TY - JOUR

T1 - Accurate histopathology from low signal-to-noise ratio spectroscopic imaging data

AU - Reddy, Rohith K.

AU - Bhargava, Rohit

PY - 2010/11/1

Y1 - 2010/11/1

N2 - Fourier Transform Infrared (FT-IR) spectroscopic imaging is emerging as an automated alternative to human examination in studying development and disease in tissue. The technology's speed and accuracy, however, are limited by the trade-off with signal-to-noise ratio (SNR). Signal processing approaches to reduce noise have been suggested but often involve manual decisions, compromising the automation benefits of using spectroscopic imaging for tissue analysis. In this manuscript, we describe an approach that utilizes the spatial information in the data set to select parameters for noise reduction without human input. Specifically, we expand on the Minimum Noise Fraction (MNF) approach in which data are forward transformed, eigenimages that correspond mostly to signal selected and used in inverse transformation. Our unsupervised eigenimage selection method consists of matching spatial features in eigenimages with a low-noise gold standard derived from the data. An order of magnitude reduction in noise is demonstrated using this approach. We apply the approach to automating breast tissue histology, in which accuracy in classification of tissue into different cell types is shown to strongly depend on the SNR of data. A high classification accuracy was recovered with acquired data that was ∼10-fold lower SNR. The results imply that a reduction of almost two orders of magnitude in acquisition time is routinely possible for automated tissue classifications by using post-acquisition noise reduction.

AB - Fourier Transform Infrared (FT-IR) spectroscopic imaging is emerging as an automated alternative to human examination in studying development and disease in tissue. The technology's speed and accuracy, however, are limited by the trade-off with signal-to-noise ratio (SNR). Signal processing approaches to reduce noise have been suggested but often involve manual decisions, compromising the automation benefits of using spectroscopic imaging for tissue analysis. In this manuscript, we describe an approach that utilizes the spatial information in the data set to select parameters for noise reduction without human input. Specifically, we expand on the Minimum Noise Fraction (MNF) approach in which data are forward transformed, eigenimages that correspond mostly to signal selected and used in inverse transformation. Our unsupervised eigenimage selection method consists of matching spatial features in eigenimages with a low-noise gold standard derived from the data. An order of magnitude reduction in noise is demonstrated using this approach. We apply the approach to automating breast tissue histology, in which accuracy in classification of tissue into different cell types is shown to strongly depend on the SNR of data. A high classification accuracy was recovered with acquired data that was ∼10-fold lower SNR. The results imply that a reduction of almost two orders of magnitude in acquisition time is routinely possible for automated tissue classifications by using post-acquisition noise reduction.

UR - http://www.scopus.com/inward/record.url?scp=77958030904&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77958030904&partnerID=8YFLogxK

U2 - 10.1039/c0an00350f

DO - 10.1039/c0an00350f

M3 - Article

C2 - 20830324

AN - SCOPUS:77958030904

VL - 135

SP - 2818

EP - 2825

JO - Analyst

JF - Analyst

SN - 0003-2654

IS - 11

ER -