Abstract

Laser Raman spectroscopy (LRS) is a highly specific biomolecular technique which has been shown to have the ability to distinguish malignant and normal breast tissue. This paper discusses significant advancements in the use of LRS in surgical breast cancer diagnosis, with an emphasis on statistical and machine learning strategies employed for precise, transparent and real-time analysis of Raman spectra. When combined with a variety of “machine learning” techniques LRS has been increasingly employed in oncogenic diagnostics. This paper proposes that the majority of these algorithms fail to provide the two most critical pieces of information required by the practicing surgeon: a probability that the classification of a tissue is correct, and, more importantly, the expected error in that probability. Stochastic backpropagation artificial neural networks inherently provide both pieces of information for each and every tissue site examined by LRS. If the networks are trained using both human experts and an unsupervised classification algorithm as gold standards, rapid progress can be made understanding what additional contextual data is needed to improve network classification performance. Our patients expect us to not simply have an opinion about their tumor, but to know how certain we are that we are correct. Stochastic networks can provide that information.

Highlights

  • Breast cancer is the second most commonly diagnosed cancer among American women after skin cancer, with approximately 1 in 8 women (12.8%) receiving a diagnosis of invasive breast cancer within their lifetime [1]

  • The standard of care for early stage invasive breast cancer includes breast irradiation and breast conserving surgery (BCS), during which the surgeon attempts to excise all tumors with a negative margin, that is margins of the resected tissue have no evidence of tumor [2,3]

  • The biggest risk with BCS is the chance of local reoccurrence (LR) with 15% to 35% of patients who opt for BCS requiring a second surgery to obtain negative margins [4,5]

Read more

Summary

Introduction

Breast cancer is the second most commonly diagnosed cancer among American women after skin cancer, with approximately 1 in 8 women (12.8%) receiving a diagnosis of invasive breast cancer within their lifetime [1]. If the electrons are in an excited state, the photons can gain energy in what is termed an anti-Stokes shift This scattered light is collected by a spectrometer, creating a spectrum, which showcases a series of peaks, referred to as bands, corresponding to the characteristic vibrational frequencies of the scattering molecules [22]. When applied to cancer diagnostics, this unique fingerprint can be a direct indicator of the inherent biochemical differences between malignant and healthy tissue Such biochemical specificity obtained from the surgical field can be instrumental in helping a surgeon achieve negative margins during BCS. This paper discusses the most significant research advancements in the use of LRS as a real-time in vivo tool for surgical breast cancer detection, with an emphasis on the statistical and machine learning methods employed for spectral classification. We posit that (1) the majority of so called “machine learning” algorithms are, statistical tools; and, (2) that stochastic backpropagation algorithms are the only true learning algorithm inherently providing the two most critical pieces of information: the Bayesian probability of correct classification and an estimate of the certainty of each classification probability for each target

Statistical and Machine Learning Methods
Data Pre-Processing
Data Optimization and Dimension Reduction
Supervised Data Classification
Bayesian Probabilities of Correct Classification—Stochastic Neural Networks
Review of Major Research Advancements
Findings
51 BCS surfaces
Conclusions
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call