Neurophotonics and Biomedical Spectroscopy

Neurophotonics and Biomedical Spectroscopy

Editors

Robert R. Alfano

The City College of the City University of New York, New York, NY 10031, United States

Lingyan Shi

Columbia University in the City of New York, New York, NY 10027, United States

Table of Contents

Cover image

Title page

Copyright

Contributors

Preface

1. Optical Fiber-Probe Spectroscopy of Brain Tumors

1. Introduction

2. Current Diagnostic Modalities in Neuro-Oncology

3. Optical Methods

4. Methods for Spectral Analysis

5. Applications of Optical Spectroscopy in Neuro-Surgery

6. Conclusion

2. Noninvasive Optical Studies of the Brain: Contributions From Systemic Physiology

1. Background and Historical Overview of Noninvasive Optical Studies of the Brain

2. Systemic Contributions to Cerebral Optical Signals

3. Current Strategies to Minimize Systemic Contributions to Cerebral Optical Signals

4. Optical Techniques That Take Advantage of Systemic Effects on Cerebral Perfusion

5. Conclusions

3. Principle and Application of Fluorescence Lifetime Imaging for Neuroscience: Monitoring Biochemical Signaling in Single Synapses Using Fluorescence Lifetime Imaging

1. Introduction

2. FRET Biosensors for FLIM

3. Conclusions

4. Visible Resonance Raman Spectroscopy in Human Brain Tissues

1. Introduction

2. Molecular Histopathology of Human Primary Brain Tumors by VRR

3. Conclusion and Further Directions

Appendix I: Experimental Samples, Instruments, and Parameters

Appendix II: Theoretical Models for Brain Composition Changes Calculation and Disease Classification Glioma Grading by PCA, SVM

5. Label-Free Stimulated Raman Scattering Imaging of Neuronal Membrane Potential

1. Introduction

2. Materials and Methods

3. Results and Discussion

6. The Quest for Functional Biomarkers in the Prefrontal Cortex Using Functional Near-Infrared Spectroscopy (fNIRS)

1. Introduction

2. Functional Near-Infrared Spectroscopy

3. Conclusion

7. Optical Spectroscopy of Tryptophan Metabolites in Neurodegenerative Disease

1. Introduction

2. Abnormal α-Synuclein Aggregates

3. Conclusion

8. Optical-Resolution Photoacoustic Microscopy of Brain Vascular Imaging in Small Animal Tumor Model Using Nanosecond Solid-State Laser

1. Introduction

2. Background

3. In Vivo Vascular Imaging by Optical-Resolution Photoacoustic Microscopy

4. Functional Imaging by Optical-Resolution Photoacoustic Microscopy

5. OR-PAM Tumor Brain Imaging in Small Animal Model

6. Outlook and Conclusion

Appendix: Solid-State Nanosecond Laser

9. Quantum Processes in Neurophotonics and the Origin of the Brain's Spatiotemporal Hierarchy

1. Nonlinear Complexity and Subneural Dynamics

2. Quantum Brain Hypothesis

3. Optical Communication Channels in the Brain

4. Light Generated in the Brain

5. Light Signaling in the Brain

6. Scaling Back Up

7. Practical Applications

8. Concluding Remarks

10. Photoacoustic Microscopy of Cerebral Hemodynamic and Metabolic Responses to General Anesthetics

1. Introduction

2. Background and Motivation

3. Principle of Multiparametric Photoacoustic Microscopy

4. Instrument Design and Configuration

5. Experimental Procedures for Awake Brain Imaging

6. Cerebral Hemodynamic and Oxygen-Metabolic Responses to Isoflurane

7. Discussion and Perspectives

11. Basic Optical Scattering Parameter of the Brain and Prostate Tissues in the Spectral Range of 400–2400nm

1. Introduction

2. Scattering Parameters of Interest

3. Mie Theory and Calculation for Scattering Properties of Particles in Same Size as Nuclei of Prostate Tissues

4. Conclusion

Appendix—MATLAB Codes for Mie Scattering

12. Overview of Fluorescence Spectroscopy and Imaging for Early Cancer Detection

1. Fluorescence Spectroscopy

2. Fluorescence Imaging

3. Conclusion

13. Overview of Supercontinuum Sources for Multiphoton Microscopy and Optical Biopsy

1. Multiphoton Microscopy and Its Application in Optical Biopsy

2. SC-Source-Based MPM

3. SC MPM for Optical Biopsy and Neuro Imaging

4. Conclusion

14. Dynamic Vascular Optical Tomographic Imaging for Peripheral Artery Disease and Breast Cancer

1. Introduction

2. Peripheral Artery Disease

3. Breast Cancer Imaging

15. Stimulated Raman Scattering for Cell and Tissue Imaging

1. Introduction

2. Instrumentation

3. Imaging With SRS

4. Conclusions

16. Organic Molecules in Photonics, Cancer Phototherapy, and Neurophotonics

1. Introduction

2. Organic Dyes in Lasers for Cancer Phototherapy

3. Organic Dyes in Cancer Photodynamic Therapy

4. Organic Molecules in Light Sheet Microscopy

5. Organic Laser Optics for Optogenetics and Neurophotonics

6. Organic Molecules in Coherent Electrically Pumped Organic Semiconductors

17. Multimodal Optical Biopsy and Imaging of Skin Cancer

1. Introduction

2. Imaging Techniques for Skin Neoplasms Analysis

3. Optical Spectroscopy for Cancer

4. Multimodal Classification of Skin Tumors

5. In vivo Diagnostics of Malignant and Benign Neoplasms

6. Conclusion

7. Statement of Ethics

18. Light-Sheet Fluorescence Microscopy With Structured Light

1. Background and Motivation

2. Imaging With Sheets of Light

3. Gaussian, Bessel, and Airy Optical Modes

4. Structured Light and Wavefront Control in Light-Sheet Microscopy

5. Conclusions and Outlook

19. Tethered Capsule Endomicroscopy: A New Window Into the Upper Gastrointestinal Tract

1. Optical Coherence Tomography TCE

2. Spectrally Encoded Confocal Microscopy TCE

3. Conclusion

20. Plum Pudding Random Medium Model of Biological Tissue and Optical Biomedical Imaging in NIR and SWIR Spectral Windows

1. Introduction

2. Background Refractive Index Fluctuation

3. Plum Pudding Random Medium

4. Determination of Biological Tissue Structure From Scattering Spectroscopy

5. Optical Biomedical Imaging in NIR and SWIR Spectral Windows

6. Discussion and Conclusions

Appendix

21. Potential Roles for Spectroscopic Coherent Raman Imaging for Histopathology and Biomedicine

1. Introduction

2. Broadband Coherent Anti-Stokes Raman Scattering (CARS)

3. Uses for BCARS in Histopathology

Index

Copyright

Elsevier

Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

Copyright © 2019 Elsevier Ltd. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-0-323-48067-3

For information on all Elsevier publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Matthew Deans

Acquisition Editor: Simon Holt

Editorial Project Manager: Sabrina Webber

Production Project Manager: Surya Narayanan Jayachandran

Cover Designer: Victoria Pearson

Typeset by TNQ Technologies

Contributors

Robert R. Alfano

Institute for Ultrafast Spectroscopy and Lasers, Departments of Physics, The City College of the City University of New York, 160 Convent Avenue, New York, NY 10031, United States

Physics Program, Graduate School of the City University of New York, 365 5th Avenue, New York, NY 10016, United States

Mirella L. Altoé,     Department of Biomedical Engineering, Columbia University, 1210 Amsterdam Ave., MC8904, New York, NY 10027, United States

Suresh Anand

European Laboratory for Non-Linear Spectroscopy (LENS), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy

Center for Industrial Research and Innovation, Amrita Vishwa Vidyapeetham, Ettimadai, Coimbatore-641 112, Tamil Nadu, India

Afrouz Anderson,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Siamak Aram,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Dmitry N. Artemyev,     Samara National Research University, Department of Laser and Biotechnical Systems, Samara, Russia

Danielle R. Bajakian,     Department of Surgery - Division of Vascular Surgery and Endovascular Interventions, Columbia University Medical Center, Herbert Irving Pavilion, 161 Fort Washington Avenue, New York, NY 10032, United States

Renzhe Bi,     Singapore Bioimaging Consortium (SBIC), 11 Biopolis Way, Singapore City 138667, Singapore

Ivan A. Bratchenko,     Samara National Research University, Department of Laser and Biotechnical Systems, Samara, Russia

Charles H. Camp Jr. ,     National Institutes of Standards and Technology, 100 Bureau Drive, Gaithersburg, MD, United States

Rui Cao,     Department of Biomedical Engineering, University of Virginia, Charlottesville, VA 22908, United States

Jun Chen,     Department of Cardiology, Tianjin Medical University General Hospital, 154 Anshan Rd, Heping Qu, Tianjin, 300052, China

Ji-Xin Cheng,     Department of Biomedical Engineering, Department of Electrical and Computer Engineering, Photonics Center, Boston University, Boston, MA, United States

Fatima Chowdhry,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Riccardo Cicchi

European Laboratory for Non-Linear Spectroscopy (LENS), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy

National Institute of Optics-National Research Council (INO-CNR), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy

Marcus T. Cicerone

National Institutes of Standards and Technology, 100 Bureau Drive, Gaithersburg, MD, United States

Georgia Institute of Technology, Atlanta, GA, United States

Emma Condy,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Travis J.A. Craddock

Departments of Psychology & Neuroscience, Computer Science, and Clinical Immunology, Nova Southeastern University, Fort Lauderdale, FL, United States

Clinical Systems Biology Group, Institute for Neuro-Immune Medicine, Nova Southeastern University, Fort Lauderdale, FL, United States

Hadis Dashtestani,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Kishan Dholakia,     SUPA, School of Physics and Astronomy, University of St Andrews, North Haugh, St Andrews, KY16 9SS, United Kingdom

Jing Dong

Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, MA, United States

Department of Pathology, Massachusetts General Hospital, Boston, MA, United States

F.J. Duarte,     Interferometric Optics, Rochester, NY, United States

Sergio Fantini,     Department of Biomedical Engineering, Tufts University, 4 Colby Street, Medford, MA 02155, United States

Amir Gandjbakhche,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Stuart R. Hameroff,     Departments of Anesthesiology and Psychology, Center for Consciousness Studies, The University of Arizona Health Sciences Center, Tucson, AZ, United States

Dawn L. Hershman,     Departments of Medicine and Epidemiology, Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, 161 Fort Washington, New York, NY 10032, United States

Andreas H. Hielscher

Department of Biomedical Engineering, Columbia University, 1210 Amsterdam Ave., MC8904, New York, NY 10027, United States

Department of Electrical Engineering, Columbia University, 500 West 120th Street, Mudd 1310, New York, NY 10027, United States

Department of Radiology, Columbia University Medical Center, New York, NY 10032, United States

Song Hu,     Department of Biomedical Engineering, University of Virginia, Charlottesville, VA 22908, United States

Nicusor V. Iftimia,     Physical Sciences Inc., 20 New England Business Center Drive, Andover, MA 01810, United States

Jana M. Kainerstorfer,     Department of Biomedical Engineering, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA 15213, United States

Kevin Kalinsky,     Departments of Medicine and Epidemiology, Herbert Irving Comprehensive Cancer Center, Columbia University Medical Center, 161 Fort Washington, New York, NY 10032, United States

Nader Shahni Karamzadeh,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Riley Kermanian,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Hyun K. Kim,     Department of Radiology, Columbia University Medical Center, New York, NY 10032, United States

Sergey V. Kozlov,     Samara State Medical University, Department of Oncology, Samara, Russia

Philip Kurian

Department of Medicine, Howard University College of Medicine, Washington, DC, United States

Quantum Biology Laboratory, Howard University, Washington, DC, United States

Hyeon Jeong Lee,     Department of Biomedical Engineering, Department of Electrical and Computer Engineering, Photonics Center, Boston University, Boston, MA, United States

Martin Lee,     Cancer Research UK Edinburgh Centre, Institute of Genetics and Molecular Medicine, The University of Edinburgh, Edinburgh, United Kingdom

Jun Li,     Department of Anesthesiology, University of Virginia, Charlottesville, VA 22908, United States

Cheng-hui Liu,     Institute for Ultrafast Spectroscopy and Lasers, Department of Physics, The City College of the City University of New York, 160 Convent Avenue, New York, NY 10031, United States

Qinglei Ma,     Advanced Optowave Corporation (AOC), 105 Comac Street, Ronkonkoma, NY 11779, United States

Alessandro Marone,     Department of Biomedical Engineering, Columbia University, 1210 Amsterdam Ave., MC8904, New York, NY 10027, United States

Helga Miguel,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Haiding Mo,     Trumpf Photonics Inc., 2601 US -130, Cranbury, NJ 08512, United States

Alexander A. Moryatov,     Samara State Medical University, Department of Oncology, Samara, Russia

Oleg O. Myakinin,     Samara National Research University, Department of Laser and Biotechnical Systems, Samara, Russia

Jonathan Nylk,     SUPA, School of Physics and Astronomy, University of St Andrews, North Haugh, St Andrews, KY16 9SS, United Kingdom

Malini Olivo,     Singapore Bioimaging Consortium (SBIC), 11 Biopolis Way, Singapore City 138667, Singapore

Andrey E. Orlov,     Samara Regional Clinical Oncology Dispensary, Samara, Russia

Prabodh Kumar Pandey,     Department of Physics, Indian Institute of Technology, Kanpur, India, 208016

Francesco S. Pavone

European Laboratory for Non-Linear Spectroscopy (LENS), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy

National Institute of Optics-National Research Council (INO-CNR), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy

Asima Pradhan

Department of Physics, Indian Institute of Technology, Kanpur, India, 208016

Center for Lasers and Photonics, Indian Institute of Technology, Kanpur, India, 208016

Yang Pu

Institute for Ultrafast Spectroscopy and Lasers, The City College of the City University of New York, 160 Convent Avenue, New York, NY 10031, United States

MicroPhotoAcoustics (MPA), Inc., 105 Comac Street, Ronkonkoma, NY 11779, United States

Alexander Ruesch,     Department of Biomedical Engineering, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA 15213, United States

Lingyan Shi,     Department of Chemistry, Columbia University, New York, NY 10027, United States

Pankaj Singh,     Department of Physics, Indian Institute of Technology, Kanpur, India, 208016

Laura A. Sordillo

Institute for Ultrafast Spectroscopy and Lasers, Physics Department, The City College of New York, New York, NY, United States

The Grove School of Engineering, Electrical Engineering Department, The City College of New York, New York, NY, United States

Peter P. Sordillo

Institute for Ultrafast Spectroscopy and Lasers, Physics Department, The City College of New York, New York, NY, United States

Department of Medicine, Lenox Hill Hospital, New York, NY, United States

Gary Tearney

Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, MA, United States

Department of Pathology, Massachusetts General Hospital, Boston, MA, United States

Harvard Medical School, Boston, MA, United States

William J. Tipping,     Cancer Research UK Edinburgh Centre, Institute of Genetics and Molecular Medicine, The University of Edinburgh, Edinburgh, United Kingdom

Jack A. Tuszynski

Department of Experimental Oncology, Cross Cancer Institute, Edmonton, AB, Canada

Department of Physics, University of Alberta, Edmonton, AB, Canada

Department of Mechanical and Aerospace Engineering (DIMEAS), Polytechnic University of Turin, Turin, Italy

Wubao Wang

Institute for Ultrafast Spectroscopy and Lasers, The City College of the City University of New York, 160 Convent Avenue, New York, NY 10031, United States

Departments of Physics, The City College of the City University of New York, 160 Convent Avenue, New York, NY 10031, United States

Binlin Wu,     Physics Department and CSCU Center for Nanotechnology, Southern Connecticut State University, 501 Crescent Street, New Haven, CT 06515, United States

Min Xu,     Department of Physics & Astronomy, Hunter College of CUNY, 695 Park Ave, New York, NY 10065, United States

Ryohei Yasuda,     Max Planck Florida Institute for Neuroscience, One Max Planck Way Jupiter, FL 33458, United States

Xinguang Yu,     The Department of Neurosurgery, PLA General Hospital, 28 Fuxing Road, Beijing 100039, China

Valery P. Zakharov,     Samara National Research University, Department of Laser and Biotechnical Systems, Samara, Russia

Rachel Zaragoza,     National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD, United States

Youbo Zhao,     Physical Sciences Inc., 20 New England Business Center Drive, Andover, MA 01810, United States

Yan Zhou,     Air Force General Hospital, PLA, 30 Fuchenglu, Haidian District, Beijing 100142, China

Ke Zhu,     Institute of Physics, Chinese Academy of Sciences (CAS), Beijing 100190, China

Zhiyi Zuo,     Department of Anesthesiology, University of Virginia, Charlottesville, VA 22908, United States

Preface

Over the past three decades, a number of impressive advances in biophotonics have been made by using the salient properties of light toward better understanding of the fundamental interactions of light with tissues and cells and toward developing novel linear and nonlinear optical-based techniques to monitor, image, and diagnose diseases such as cancer, atherosclerosis, and Alzheimer's disease. Biophotonics is playing more and more important roles in imaging the brain and other organs and tissues for clinical, medical, and biomedical studies. Indeed light optical spectroscopy and imaging offer a new armamentarium for medicine and biomedical research. There is a need to review the advances and achievements of the technologies in neuroscience and other biomedical studies.

The primary goals of this book, Neurophotonics and Biomedical Spectroscopy, are to cover the major experimental, theoretical, and numerical advances in the biomedical photonics field by reviewing the fundamentals in biophotonics and going over emerging new novel optical spectroscopy methods and applications. Leading experts in the field discussed their works with photonic and optical spectroscopy to advance our knowledge in detection of cancer and understand the brain better using linear and nonlinear optical spectroscopy and imaging. Ultrafast laser beams, complex vector beams, and supercontinuum enable the applications of nonlinear optical processes for probing electronic and vibrational states of biomedical media, like cells and tissues, with higher spatial resolution than magnetic resonance imaging (MRI) or other methods.

The book consists of 21 chapters in two parts. The first part from Chapter 1 to 11 focuses on topics in neurophotonics. The second part consists of the rest 10 chapters, introducing studies of multiple spectroscopic technologies in various biomedical fields.

Chapter 1 by Anand, Cicchi, and Pavone overviews the fundamentals of spectroscopic techniques, different fiber optic probes used in the spectroscopic systems, and spectroscopic modalities for brain study. Chapter 2 by Fantini, Ruesch, and Kainerstorfer discusses the application of noninvasive near-infrared spectroscopy in the brain and their sensitivity to systemic physiological changes. Chapter 3 by Yasuda introduces the principle and application of fluorescence lifetime imaging, particularly on imaging protein activity in single dendritic spines. In Chapter 4, Alfano and coworkers first review the applications of nonresonance and resonance Raman spectroscopy on brain research, and then reveal their achievement of visible resonance Raman technique in diagnosing brain tumor, which opens a new window for molecular optics–based histopathology. Chapter 5 presents the recent study by Lee and Cheng of vibrational spectroscopic signature of neuronal membrane potentials using stimulated Raman scattering imaging. The next two chapters study biomarkers and metabolites in neurophotonics, where Chapter 6 by Gandjbakhche's group at the National Institutes of Health (NIH) of the United States introduces functional biomarkers that can be driven from functional near-infrared spectroscopy in imaging prefrontal cortex, and Chapter 7 by Sordillo et al. investigates tryptophan metabolites in neurodegeneration disease using optical spectroscopic techniques. Chapter 8 by Bi et al. focuses on the application of optical-resolution photoacoustic microscopy in monitoring the vascular disrupting agent therapeutic effect of glioma tumor. In Chapter 9 Craddock reviews the quantum process in cognition and consciousness and discusses the origin of brain's hierarchy. Chapter 10 by Cao et al. proposes the newly developed photoacoustic microscopy technique for simultaneous label-free imaging of blood perfusion, oxygenation, and flow in single microvessel of awake mouse brain. In Chapter 11 Pu et al. offer theoretical and experimental studies on optical and structural parameters, such as the scattering coefficient and the anisotropy factor, in brain and prostate tissues at the spectral range from 400 to 2400  nm.

In part 2, Chapter 12 by Pradhan's group reviews the fluorescence spectroscopy and imaging techniques, including spatially resolved fluorescence, polarized fluorescence, and synchronous fluorescence spectroscopy, on cancer diagnosis. Zhao and Iftimia in Chapter 13 offer the introduction of supercontinuum generation in photonic crystal fibers as the light source for multiphoton microscopy, second harmonic generation, and fluorescence and overview optical biopsy studies using supercontinuum. In Chapter 14 Hielscher and coworkers report the latest advances of dynamic optical tomographic imaging in diagnosing and monitoring peripheral artery disease and breast cancer. Chapter 15 by Lee and Tipping covers the basic principles and techniques of stimulated Raman scattering imaging and explores the recent developments in both research and clinical environments. Frank Duarte in Chapter 16 reviews organic molecules and their applications in photonics. Chapter 17 by Zakharov et al. presents and discusses multiple optical modalities for skin cancer detection. Chapter 18 by Nylk and Dholakia introduces the theories of an emerging imaging technique, light-sheet microscopy, and light and wavefront control in light-sheet microscopy. In Chapter 19 Dong and Tearney review a new technology, tethered capsule endomicroscopy, that is used to improve upper gastrointestinal tract diagnosis and research conducted with this technology. Chapter 20 by Xu and Alfano introduces a plum pudding random medium model of biological tissue and studies the breast and brain imaging in near-infrared and shortwave infrared regions. The last chapter, Chapter 21 by Cicerone and Camp at the National Institute of Standards and Technology (NIST) of the United States, presents the new application of coherent Raman imaging in computer-aided diagnosis of diseases.

As will be seen, many novel optical spectroscopy imaging techniques in neurophotonics and biomedical fields are covered in the book. Theories and principles of these optical spectroscopy and imaging techniques are also covered. But still, we could not cover all related articles of importance due to the scope and size of this book, and much remains for the future. The book provides researchers with up-to-date technologies and proposes them directions to adapt to their own or to explore new optical and spectroscopic technologies. This book is also an essential reference for graduate students and new investigators in fields including optics, photonics, neuroscience, neurophotonics, biophotonics, oncology, physics, biomedical engineering, etc.

Finally, the editors would like to thank all authors who devoted their precious time to contribute very interesting and knowledgeable chapters, all who helped in preparation of the book, authors and publishers for the permission of reproducing their figures in this book, and the editorial staff of the publisher.

Robert R. Alfano

Lingyan Shi,     New York, New York

August 2018

1

Optical Fiber-Probe Spectroscopy of Brain Tumors

Suresh Anand¹,², Riccardo Cicchi¹,³, and Francesco S. Pavone¹,³     ¹European Laboratory for Non-Linear Spectroscopy (LENS), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy     ²Center for Industrial Research and Innovation, Amrita Vishwa Vidyapeetham, Ettimadai, Coimbatore-641 112, Tamil Nadu, India     ³National Institute of Optics-National Research Council (INO-CNR), Via Nello Carrara, 1 - 50019, Sesto Fiorentino, Italy

Abstract

In the current clinical scenario, surgical resection of tumors remains the front-line treatment modality for the management of abnormalities in both brain and central nervous system. The primary goal of surgery is to remove as much tumor as possible sparing the normal tissue. Needless removal of normal brain tissue could lead to detrimental neurological deficits causing cognitive dysfunction and motor impairment. Surgical biopsy and other imaging modalities have the disadvantages of being invasive and time-consuming, and brain shift is a major problem when using preoperative magnetic resonance images. Alternately, in the recent years, optical spectroscopy has emerged as a diagnostic modality for the detection of abnormalities in the brain. The resulting spectral features from tissues can provide information related to the biochemistry, morphology, and physiological conditions of the tissue. This technique can probe tissues in vivo or minimally invasive, and real time. In this chapter, after having covered current diagnostic modalities used in pediatric neuro-surgery, a brief outline on the fundamentals of different spectroscopic techniques is explained in detail. In the following section, a brief description of fiber optic probes and different fiber probe configurations implemented in an optical spectroscopic system is presented. Additionally, details related to different methods for spectral analysis are also discussed. Next, we review the different optical point spectroscopic modalities for the demarcation of abnormalities in the pediatric brain. We finally conclude with a discussion on promises, hopes, and opportunities of this emerging technology in the neuro-oncological application.

Keywords

Brain tumor; Neuro-surgery; Optical spectroscopy

1. Introduction

Tumors in central nervous system (CNS) and brain are the frequent solid tumors in pediatric oncology, accounting for approximately 21% of pediatric cancers and also one of the leading cause of cancer-related mortality in children [1,2]. The overall 5-year survival rate is around 67% for all primary tumors in the CNS for subjects in the age from 0 through 19  years [3]. Some of the most common tumors associated with the childhood include medulloblastoma, astrocytoma, ependymoma, and gliomas [4]. Management of pediatric subjects with tumors in the CNS and brain is quite complex and is influenced by an extensive range of factors including the type and site of tumor in CNS and age of the pediatric subject. Surgical delineation of tumor margins to maximize tumor excision while sparring healthy tissue and minimize the postoperative neurological complications is a key issue for a successful treatment. If the tumors are centrally located, there is a rare possibility that these regions could be totally resected without any neurological deficits and even a biopsy could be fatal. The therapeutic options for pediatric brain tumors are multifaceted and could range from neurosurgical resection that removes affected part of the brain to chemotherapy and radiotherapy [5].

Neurosurgical intervention remains the first-line therapeutic option with the intent for gross total resection and improves the patient survival rates [6]. In addition, the extent of resection in turn directly associates with the length of survival [7]. Needless removal of normal brain tissue could lead to multiple sequelae including neurological deficits such as cognitive dysfunction, endocrine dysfunction, motor impairment, ototoxicity, and secondary malignancies [8–11]. To avoid these shortfalls and better demarcation of brain tumor margins, different types of techniques and imaging modalities such as magnetic resonance imaging, computed tomography, and ultrasound have been employed. Despite these advances, there are always technical limitations and practical disadvantages associated with the current technologies. Existing methodologies need to be complemented by improved intraoperative delineation margins that could aid in a more precise resection.

In light of this, optical point spectroscopic techniques have been gaining momentum for tissue diagnostics. The most widely used spectroscopic techniques include diffuse reflectance, fluorescence, and Raman spectroscopy. The application of optical spectroscopy in biomedical research has been the subject to manifold research and has included clinical studies for detecting tumors in wide-ranging organ sites [12–15], for monitoring response of tissues to different types of therapy [16–18], and for guided biopsies [19]. Additionally, this technique has been used to detect atherosclerosis [20–22], diabetes [23–25], noninvasively measure glucose in diabetic patients [26–28], and monitor the wound healing process in diabetic foot ulcers [29]. Optical spectroscopy finds its interest in minimally invasive tissue diagnostics, which may serve as an adjunct technique to tissue biopsy.

In this chapter, we present an overview of the current techniques based on optical spectroscopy for the demarcation of brain tumor margins in pediatric subjects. We first begin with a discussion on current diagnostic practices used for the detection of pediatric tumors and its technical and practical limitations. We then describe different spectroscopic techniques and their roles in disease diagnostics. Also, we provide information related to different fiber probe configurations and various methods for spectral analysis. In the later part, we provide an overview of different optical point spectroscopic methods and classification algorithms used in the delineation of tumor margins in the brain. Finally, we conclude with a discussion on the promises, hopes, and opportunities of this emerging technology in tissue diagnostics for neuro-oncological application.

2. Current Diagnostic Modalities in Neuro-Oncology

Significant progress has been made in the field of diagnostic pathology for the delineation of tumor margins during neuro-surgery. Proper detection of tumors of the brain in pediatric subjects requires an integrated comprehensive approach utilizing different diagnostic modalities. This could help in enabling the proper choice of treatment, prolonging event-free-survival and improving the curing rates. Many tools have been developed for the detection of tumors in the brain including stereotactic brain surgery, magnetic resonance imaging, and ultrasound. Histopathology is the most commonly used method and gold standard to determine the type and grade of brain tumors. Histology is based on the staining of hematoxylin and eosin (H & E), a simple dye combination capable of highlighting a range of intracellular and extracellular tissue structures. This method plays an important role in specific diagnosis and care for the subject under investigation. On the downside, biopsy is time-intensive and labor-intensive, subjective and invasive, since it requires tissue biopsy.

2.1. Stereotactic Brain Surgery

Stereotactic brain surgery is a minimally invasive diagnostic procedure coupled with a near-infrared or an electromagnetic tracking system in which a neurosurgeon uses preoperative three-dimensional (3D) images usually from advanced imaging modalities like computed tomography (CT) and magnetic resonance imaging (MRI) scans [30]. CT is hardly acceptable in the case of pediatric subjects as it concerns the dose of ionizing radiation [31]. These images are then imported into a computer system that providing a 3D image of the patient's brain. Neurosurgeon uses these images to determine the exact location of brain tumors in order to facilitate a pathway for removing as much tumor as possible leaving behind normal tissues undamaged. The problems associated with this preoperatively acquired imaging approach include errors in the image registration process due to intraoperative shift and deformations of brain or the brain shift [32]. A variety of factors including the drainage of cerebrospinal fluid (CSF), surgical resection of brain tumors, and the application of diuretics contribute to the brain shift during the course of neuro-surgery [33]. This reduces the diagnostic ability and effectiveness of stereotactic image-guided surgery using images acquired prior to neuro-surgery.

2.2. Intraoperative Ultrasound (iUS)

Ultrasound imaging is based on the principle of transmitting ultrasound pulses and recording the reflected echoes from tissue structures and components, which provides important information about tissue morphology. The penetration depth of ultrasound is dependent on the frequency used, and at about 7–15  MHz, the penetration depth is in the range through 2–7  cm [34]. For a phased array probe at 5  MHz, the radial and axial resolution is in the order of 0.5 and 1.0  mm [35]. Additionally, iUS is cheap, easy to use, and also used for identifying brain shift between intraoperative and preoperative images to improve diagnostic accuracy during neuro-surgery [31]. Severe practical pitfalls associated with ultrasound include time-consuming, intense training by a radiologist, and the fact that the probes used requires bigger craniotomies and this edges the concept of minimally invasive surgery [31,36]. Also, ultrasound does not provide any functional information [37].

2.3. Intraoperative MRI (iMRI)

The technical limitations associated with stereotactic brain surgery and iUS have been overcome by iMRI systems. The nuclear spin of the hydrogen atoms in various molecules is exploited in MRI for creating contrast, allowing the visualization of different anatomical structures. MRI systems are classified by their field strength, magnet configuration, and the manner by which the subject is brought inside the scanner [38]. iMRI systems can be classified into two classes: low- and high-field imaging systems. Low-field imaging systems employ a magnetic field strength in the range from 0.12 to 0.2  T [39]. The problems associated with low-field strength systems include increased scanning time, low resolution, decreased signal-to-noise ratio, and poor image quality [40]. While the high-field systems (magnetic field strength in the range 1.5–3.0  T [33]) come with a better image quality but with increased costs. Additionally, high-field MRI systems could perform additional competencies such as magnetic resonance spectroscopy (MRS), functional MRI (fMRI), diffusion weighted imaging, and chemical shift imaging [40]. Potential benefits of iMRI include real time, increasing the extent of tumor margins during surgical resection and excellent soft tissue resolution [41,42]. Though iMRI has its advantages, it has been limited by the following factors: being expensive, increased running costs, each time when the image is updated, the magnet has to be moved to a new position which may affect the regular surgical work flow [43].

2.4. Other High-Energy Imaging Modalities

In addition to the above, diagnostic methodologies based on functional imaging techniques such as positron emission tomography (PET) and single-photon emission computed tomography (SPECT) have been implemented in clinical neuro-oncology for the demarcation of brain tumor margins. PET involves injecting a radiotracer intravenously prior to the scanning procedure and diseased tissues present a differential absorption that serves as a contrast for tissue classification [44]. Some of the most commonly used radiotracers employed in PET include 2-¹⁸F-fluoro-deoxy-D-glucose (¹⁸F-FDG), Methyl-¹¹C-L-methionine (¹¹C-MET) and 3′-deoxy-3′-¹⁸F fluoro-L-thymidine (¹⁸F-FLT) [45]. Unlike MRI, which provides morphological and anatomical information, PET delivers insight into tissue biochemistry and physiological processes depending on the type of radiotracer [46]. On the downside, issues related to the lack of availability, especially in countries with low-resource-settings, and high costs are the major factors limiting the wide-spread application of PET [47]. SPECT is a version of PET that integrates a gamma camera to obtain 3-D images. This gamma camera rotates around the patient to acquire different images of the area under investigation at different angles in order to provide tomographic information. SPECT uses gamma-emitting radionuclides such as technetium-99m, iodine-123, and thallium-201 [48]. Some of the limitations of SPECT include costs, harmful due to ionization potential, intrinsic dependence on radio pharmaceuticals, and lower spatial resolution than PET [49].

3. Optical Methods

Each diagnostic technique has its own advantages and limitations. It has become increasingly natural for the neuro-surgeon and scientists to take full advantage of the diverse information provided by different diagnostic techniques to provide anatomical and functional information about tissue pathology. In light of this, different optical diagnostic methods on the basis of different types of light interaction with tissue have been developed for the detection of brain tumors. These optical diagnostic techniques based on spectroscopy can provide information related to pathological changes in tissue structures in a noninvasive or minimally invasive and objective way. In the following section, we explain different concepts of light interaction with tissue, which forms the basis for optical techniques in disease diagnostics.

3.1. Light–Tissue Interaction

Tissue is a complex, multilayered, and turbid medium with different constituents such as absorbers, scattering medium, and fluorophores inhomogeneously distributed in the tissue structure. Fig. 1.1 demonstrates different types of light interaction with tissue, which can be utilized to identify the inherent characteristics to distinguish normal and diseased tissue types. When a beam of light impinges on a tissue, it can interact with the tissue in a number of ways. For instance, light could be absorbed by specific tissue constituents, scattered due to the inhomogeneities in the refractive index, or emitted at a wavelength different from that of the excitation wavelength. The re-emitted spectral intensity and shape from the tissue after these interactions contain biochemical and morphological information relevant to the tissue status. Here, we explain three different types of light–tissue interaction that play an important role in classifying normal and diseased tissues: (1) light scattering, (2) absorption, and (3) fluorescence. The following subsections provide a detailed description of these light–tissue interactions.

3.2. Light Scattering

Light transmitted into the tissue will be deflected due to inhomogeneities in the refractive index between different tissue constituents. Light scattering can be separated into elastic scattering and inelastic scattering [50]. In elastic scattering, the light is deviated at the same wavelength during the scattering process. Elastically scattered light emerging at the tissue surface could either have undergone a single scattering event (typically arising from the tissue surface) or multiple scattering events (in which light penetrates deeper into the tissue structures, undergoes a sequence of multiple scattering before re-emerging from the tissue). Because of the increased penetration depth, multiply scattered light contains information related to deeper tissue structures such as a stromal-collagen matrix. The amount light scattered depends on the wavelength of the incident light and on the size and density of scatterers [51]. The major contributors of scattering in a biological tissue include collagen fibers, nucleus, lipid membranes, and mitochondria [52]. In the case of inelastic scattering (otherwise known as Raman scattering), the absorption of a photon causes a transfer of energy between the incident photon and the scattering molecule [53]. This in turn furthers an ensuing transition of molecule from one vibrational level to another, resulting in the emission of a photon at different wavelength from the incident wavelength used to excite the molecule. This phenomenon is commonly known as Raman scattering. Unlike fluorescence and reflectance techniques, Raman spectrum provides narrow band molecular fingerprints unique to each biomolecule which changes along with the disease progression. Raman spectroscopy can probe changes in tissue biomolecules such as proteins, lipids, and nucleic acids [14].

Figure 1.1  Schematic representation of different types of light interaction with a biological tissue: specular and diffuse reflectance, fluorescence, elastic and inelastic scattering processes.

3.3. Absorption

Absorption occurs when the energy of the incident is equal to the energy difference between the two allowed states. This phenomenon causes a reduction in intensity of the incident light as some of the photons are absorbed by the molecule. The tissue molecules that absorb light at specific wavelength are referred to as chromophores. The most important chromophores in a biological tissue include hemoglobin (in its oxygenated and deoxygenated forms), melanin, water, and beta carotene [15]. The process of absorption determines how deep light can penetrate into a specific tissue; also it strongly depends on the wavelength.

3.4. Fluorescence

Some molecules in the tissue absorb light at specific wavelengths, exciting an electronic transition, and then re-emit light at a different wavelength with the electron decaying from the excited to the ground state. This phenomenon is known as fluorescence [54]. The molecules that exhibit this phenomenon are referred to as fluorophores. The fluorophores contributing to fluorescence phenomenon native to the biological tissue include metabolic constituents (nicotine adenine dinucleotide [phosphate] [NAD(P)H], flavin adenine dinucleotide [FAD]), amino acids (tryptophan, tyrosine), connective tissue proteins (collagen, elastin), and porphyrins (protoporphyrin IX [PPIX]) [12,13,55]. In addition to native fluorophores, external fluorophores (exogenous fluorophores) primarily tested as photosensitizers for photodynamic therapy (PDT) have been administered orally/intravenously to the tissue. This creates a preferential absorption of exogenous fluorophores in normal and diseased tissues [56]. The concept of preferential absorption of exogenous contrast agents forms the basic principle behind fluorescence image-guided surgeries [57]. Some of the exogenous fluorophores used in tissue diagnostics include 5-aminolevulinic acid (ALA), hematoporphyrin derivative (HpD), and tetra(m-hydroxyphenyl)chlorin (mTHPC) [55]. However, issues related to drug safety and toxicity, low specificity and waiting period for some time after the application of exogenous fluorophore to reach the tissue site of interest inhibit the application of exogenous fluorophores in a clinical environment [12,58]. Fluorescence emitted from complex turbid media like tissues is affected by a number of factors such as: (1) concentration and the type of fluorophore located at different tissue depths [12,55]; (2) concentration of chromophores such as hemoglobin and scattering molecules in the tissue which alters the fluorescence intensity and spectral shape [59]; and (3) probe geometry [60].

3.5. Multimodal Optical Spectroscopy

Recently, the application of combining information from different spectroscopic techniques for disease diagnostics has been investigated extensively. The complementary information such as biochemical, morphological and molecular fingerprints derived from different spectroscopic techniques has been known to increase the diagnostic accuracy. Since different wavelengths are used in a multimodal approach, the spectral measurements about diverse biomolecules embedded at different depths in the tissue matrix could provide depth sensitive information related to tissue pathology. This approach has been successfully implemented in the tissue organ sites including cervix [61], Barrett's esophagus [62], skin [63], and bronchus [64]. These optical technologies, both in a single-mode or in a multimodal implementation, are very often implemented within fiber probes in order to have more flexibility, ease-of-use, and reduced translational complexity.

3.6. Fiber Optic Probes

Light is delivered and, after interaction with tissues, is collected back by means of fiber optic probes. The major considerations that go into the application of fiber-optics in spectroscopy include numerical aperture, diameter of the core and core material. Fig. 1.2 presents a basic schematic of a fiber-based optical spectroscopic system. Fiber probes used in tissue optical spectroscopy are classified into two different categories (1) single fiber, and (2) bifurcated fiber configuration. In single fiber configuration, the same fiber is used for light illumination and collection, while bifurcated fiber configuration consists of one or more illumination fibers surrounded by a concentric ring of multiple collection fibers. Single fiber configuration has better light collection efficiency; on the downside, it requires additional optics such as a beam splitter to isolate excitation and emission lights. In the case of Raman spectroscopy, filtered probes are used. The illumination fiber is equipped with a bandpass filter that allows only the transmission of excitation light to illuminate the tissues of interest while blocking the fluorescence light that intrinsically originates along the path of the excitation fiber.

Figure 1.2  Basic schematic representing different components in an optical fiber-based spectroscopy system.

In addition, a long pass or notch filter is placed in front of the collection fibers to prevent the elastically scattered light from reaching the detector and from exciting unwanted signal within the collection fibers.

3.7. Probe Geometry

The following parameters have a significant impact on the optical signals emitted from a tissue: (1) source to detector separation (SDS); (2) probe to target distance (PTD); (3) numerical aperture of the fibers; (4) angle of the fibers. In a study, Zhu and associates [60] performed a series of modeling studies using Monte-Carlo simulation to investigate the effects of two different probe geometries referred to as variable aperture and multidistance probe geometries. The study indicated that with variable aperture geometry, the probe is sensitive to epithelial layer while multidistance probe geometry provides information from the deeper tissue structures. In addition, with a numerical aperture of 0.37, the fluorescence collection increases for both the probe configurations involved in this study. Pfefer et al. [65] implemented single-fiber probe configuration and another probe with different SDS and measured fluorescence from a two-layered synthetic tissue model. Significant differences in the fluorescence emission spectra were observed between measurements from a single-fiber probe and probes with different SDS. For instance, an increase in the fiber diameter shifts the fluorescence emission peak resulting in the fluorescence originating from the bottom layers. Using the fiber configuration with varying SDS, the spectral sensitivity is toward the superficial layers when there is a significant overlap between the illumination and collection fibers. An increase in SDS allows collecting spectral information originating from deeper tissue structures. The experimental results from Pfefer et al. [65] provide a solid evidence for the simulation results by Zhu et al. [60]. Liu and coworkers [66] explored the application of angled probe configuration for depth-resolved measurements on tissue simulating synthetic tissue models. They demonstrated that the larger the angle of illumination, the greater the fluorescence sensing from the top layers. In a similar study, Schwarz et al. [67] revealed the application of a ball lens for depth-resolved detection in fluorescence and reflectance spectroscopy. Depth-resolved detection can be achieved by altering the SDS and the diameter of the ball lens. Papaioannou and associates [68] demonstrated the application of probe configuration, PTD separation, and scattering anisotropy in fiber-based spectroscopy measurements using a physical tissue model. Their results demonstrated that the probe design and PTD had a strong influence on the signal collection efficiency. Additionally, single-fiber configuration had a better signal collection efficiency than bifurcated fiber probes. Also, the sampling depth was dependent on PTD and the type of fiber probe used for spectroscopic measurements.

4. Methods for Spectral Analysis

In order to extract meaningful information to differentiate normal and diseased tissues, various diagnostic algorithms have been implemented including multivariate statistical methods [64,69,70], diffusion approximation [71], Monte-Carlo modeling [72–75], and empirical algorithms [76–79]. One of the most commonly used multivariate methods implemented in tissue optical spectroscopy includes principal component analysis (PCA) and cross-validation. PCA is a method that uses eigen-vector decomposition of the covariance matrix and reduces the dimensions of very large data sets into fewer principal component (PC) scores that contain maximum variance. First few PC's accounts for maximum variance in the data set. This in turn allows minimum number of variables for evaluating the discrimination accuracy. The predictive ability of the model can be evaluated using cross-validation. One of the robust methods to develop and implement tissue classification algorithms is to divide the data set in separate training set in order to fit the algorithm and validation set that is used to test the generalization ability. This classification algorithm for tissue differentiation in the validation set establishes an unbiased estimation of the algorithm. When the sample size is small, it may not be possible to use separate training and validation set. In such cases, another type of cross-validation approach, leave-one-out-cross-validation (LOOCV) is preferably used. In LOOCV, one spectral data is removed from the overall data set (N) and the model is recalculated from the remaining (N-1) spectral data. This process is repeated until all the spectra in the data set are validated. One of the problems associated with LOOCV is that N different spectral data needs to be validated that adds to computational complexity.

In addition to multivariate statistical models, empirical algorithms based on the ratio of two different intensities in spectroscopic data have also been used for tissue classification. In this regard, Subhash et al. [76] performed a pilot study to evaluate the application of diffuse reflectance ratios at 540 and 575  nm (R540/R575) for classifying normal and malignant lesions in the oral mucosa. Their study demonstrated a decrease in the R540/R575 ratio for malignant tissues in comparison with normal tissue. As an extension of this study, the same group implemented R545/R575 for determining different grades of oral malignancy [78]. The ratios were found lower for normal tissues and the value increases based on the grade of malignancy. This approach resulted in sensitivity of 100% and specificity of 86% for classifying dysplasia from hyperplasia and 97% sensitivity and 86% specificity for differentiating hyperplastic tissues from normal tissues. Chandra et al. [79] implemented an integrated spectroscopic approach involving fluorescence and reflectance spectroscopy for delineating pancreatic adenocarcinoma, chronic pancreatitis, and normal tissues. The algorithms implemented include diffuse reflectance ratios at R470/R650, fluorescence ratios at F400/F600 and area under each normalized fluorescence spectrum for tissue classification. They achieved a sensitivity of 85% and specificity of 89% for ex vivo tissue models compared with histopathology. In the case of multivariate statistical algorithm such as PCA, it takes into account the entire spectral range for calculating the PCs; while in ratiometric approach, only selected wavelengths are taken into account for tissue classification. There could be a possibility that the other wavelengths that would offer a better classification accuracy are missed.

In analytical models such as diffusion approximation and computational models based on Monte-Carlo model for light transport, tissue is often considered as a homogeneous semi-infinite turbid medium. Reflectance spectra are modeled as a function of absorption coefficient (μa), reduced scattering coefficient (μs’), and SDS between illumination and collection fibers. μa determines the concentrations of absorbers such as oxy-hemoglobin, deoxy-hemoglobin, and beta carotene [71]; μs’ gives information related to density and size of scatterers [80], which can be determined by Mie theory [81,82], power-law approximation [83,84], and Van de Hulst approximation [85,86]. Diffusion approximation is based on solving differential equations for photon propagation in a turbid medium. This approach breaks down when the SDS is less than the inverse of μs’. This method models the propagation of a single photon at a specific instant of time and evaluates the probability of absorption and scattering based on the input values of tissue optical properties. The restrictions applied to diffusion approximation are not valid for Monte-Carlo modeling.

Exact determination of intrinsic fluorescence is essential for estimating the actual fluorophore concentration in spectroscopy of turbid media. One of the difficulties in quantitative measurements of fluorophore and chromophore concentrations is that the optical spectrum is distorted by tissue optical properties such as absorption and scattering [59]. This makes it difficult to extract clinically useful information from spectroscopic measurements. Intrinsic fluorescence is defined as the spectrum obtained free from the related effects of absorption and scattering. Different algorithms have been previously reported to extract intrinsic fluorescence from turbid samples based on tissue optical properties [87,88]. In photon migration model, reflectance spectrum is measured in the same spectral range as that of fluorescence emission. This is then used to extract the intrinsic fluorescence free from the losses due to absorption and scattering [87]. Another study by Palmer and Ramanujam [88] used Monte-Carlo modeling-based approach to extract the intrinsic fluorescence. The advantage with this method is that it can be used for any probe configuration for measuring fluorescence spectra, and is applicable to highly absorbing medium and small SDS where diffusion approximation fails.

5. Applications of Optical Spectroscopy in Neuro-Surgery

In the recent years, few research groups have investigated the application of tissue optical spectroscopy for the delineation of brain tumor margin in clinical studies. Here, we describe the application of different optical spectroscopic techniques used in neuro-surgery based on different types of light interaction with tissues. We will further present a section related to the specific investigations related to the application of optical spectroscopy in demarcating tumor margins in pediatric subjects and also the application of multimodal spectroscopy in the area of brain tumor diagnostics.

5.1. Fluorescence and Diffuse Reflectance Spectroscopy

Croce et al. [89] performed an ex vivo clinical study to assess the application of inherent fluorescence originating from biological tissues to differentiate glioblastoma margins from normal tissues during neurosurgery. Spectroscopic measurements were made by using a mercury lamp as an excitation source coupled along with appropriate filters to provide a monochromatic excitation. The fiber optic probe consisted of four illumination fibers and 16 randomly arranged collection fibers around the illumination fibers to detect the fluorescence. The resulting fluorescence is then directed toward an optical multichannel analyzer to discern light at different wavelengths. The outcome of this investigation indicated the changes in the spectral intensity and line-shape between glioblastoma tissues in comparison with cortex and white matter.

Stummer et al. [90] performed an initial study to investigate the application of porphyrin-induced fluorescence for the delineation of malignant glioma during neurosurgery. 5-ALA was used as a fluorescence contrast agent and tumor tissues were differentiated based on their ability to accumulate PpIX fluorescence. In yet another preliminary investigation, the same group used porphyrin-induced fluorescence in nine subjects for the intraoperative detection of malignant glioma. This study resulted in 85% sensitivity and 100% specificity [91]. In a controlled clinical trial [92], 139 patients were administered with 5-ALA to undergo fluorescence-guided resection, while 131 patients underwent surgery with conventional white light. The results of this randomized controlled multicenter investigation revealed that patients underwent fluorescence-guided resection had a higher probability of complete tumor removal in comparison with those with microsurgery using conventional white light. A more detailed review of the clinical background for 5-ALA drug-induced fluorescence guided biopsies is provided in [93].

Stepp and coworkers [94] performed photodynamic diagnosis and therapy using multiple radial diffuser along with 5-ALA for the discrimination of normal cortex and malignant glioma. Their study demonstrated a 100-fold increase in PpIX fluorescence in tumor tissues. In addition, PDT seems to be an effective alternative for the treatment of recurrent and inoperable gliomas. In another prospective study, Eljamel and associates [95] investigated the application of 5-ALA and photofrin for treating glioblastoma multiforme. Here, 5-ALA acted as a fluorescence contrast agent while photofrin was used for photodynamic therapy (PDT). In this study, they used an approach of applying light for diagnosis and treatment. The investigation resulted in eliminating residual tumors, thereby increasing the quality of life and decreasing the tumor progression in agreement with the previous studies.

Kuroiwa [96] developed an intraoperative microscope system using fluorescein as a contrast agent along with appropriate filters for the detection of malignant glioma. Morofuji et al. [97] reported the application of 5-ALA in photodynamic diagnosis for the detection of meningioma with the invasion of cranial matter. During surgery, tumor presented an increased fluorescence intensity when compared to the surrounding dura mater. Recently, Valdés and coworkers [98] reported up to 90% classification for the delineation of meningioma and normal dura mater. Valdés et al. [99] studied the application of 5-ALA for fluorescence-guided resection and developed a biomechanical deformation that utilizes the ultrasound images to compensate for the brain shift. In a subsequent study [100], they examined the probability of PpIX fluorescence for the discrimination of a wide range of intracranial brain tumor pathologies including low- and high-grade glioma, meningioma, and metastatic brain tissue. ALA as a fluorescence contrast agent resulted in 84% sensitivity and 94% specificity. More recently, the same authors investigated the application of 5-ALA and the induced PpIX accumulation for the detection of low-grade gliomas [101]. The results indicate, though there is an accumulation of PpIX, the levels are very low to detect below the ability of current fluorescent techniques. In the study by Valdes and associates [102], fluorescence spectroscopy with 5-ALA as an exogenous contrast agent and reflectance spectroscopy were implemented in tandem for discerning normal and brain tumor tissues. Their group implemented a light transport model to quantify the localized concentration of PpIX, total hemoglobin concentration, oxygen saturation, and tissue optical parameters. Here, a discrimination accuracy of 94% sensitivity and 94% specificity was achieved. Hosseini et al. [103] developed a hand-held optical pointer based on 5-ALA-induced fluorescence for healthy and malignant brain tissue. The device incorporated optimum delivery of light to the tissue to avoid photo bleaching and also to prevent the interference of ambient light in the operating room with the spectral measurements. Subsequently, the same group coupled hand-held pointer coupled along with ultrasound navigational system to improve the delineation margins in brain tumor surgery [104].

Lin and associates [105] performed an ex vivo study employing diffuse reflectance and fluorescence spectroscopy for the demarcation of normal brain and tumor tissues. The emission peak at 460  nm was lower in tumor tissues when compared with the normal brain tissue. In the case of diffuse reflectance, the spectra from the gray matter were similar to that of brain tumor tissues at wavelengths greater than 600  nm. A simple algorithm taking into account the fluorescence intensity at 460  nm in terms of calibrated units yielded a sensitivity and specificity of 97% and 96% for classifying normal brain and primary brain tumor tissues. Another empirical algorithm incorporating a scatter plot of the ratio between fluorescence intensity at 460  nm and reflectance intensity at 625  nm was implemented, resulting in 94% sensitivity and 90% specificity for discriminating normal brain tissue and secondary brain tumors. Lin et al. [106] also presented an in vivo study to differentiate brain tumor tissues and infiltrating tumor margins from the normal brain. Here, they implemented a two-step empirical algorithm involving the fluorescence and diffuse reflectance intensities at 460 and 625  nm, achieving 100% sensitivity and 76% specificity for characterizing normal brain tissue and infiltrating tumor margins. Toms and associates [107] developed an intraoperative optical spectroscopy system with fluorescence at 337  nm and reflectance spectroscopy for the classification of solid brain tumors and infiltrating tumor margins. Different types of empirical algorithms developed previously have been implemented in this study [105,106]. They obtained 80% sensitivity and 89% specificity for classifying normal brain tissue and solid tumors, and a sensitivity of 94% and a specificity of 93% for distinguishing normal brain tissue and infiltrating tumor margins.

Nie et al. [108] developed a combined spectroscopic system involving fluorescence lifetime and reflectance spectroscopy for the intraoperative detection of brain tumor margins. Different spectrometers were used for collecting the spectral data: (1) acousto-optic tunable filter (AOTF) for collecting the time-resolved fluorescence data, and (2) grating-based spectrometer for discerning steady-state fluorescence and reflectance spectra. A dual-mode fiber optic probe was used to measure the fluorescence and reflectance data. In addition, for measuring reflectance data, three different sources to detector separations 0.23, 0.59 and 1.67  mm were used. This combined spectroscopic system was tested on tissue simulating phantoms and on surgically excised brain tumor specimens. Their preliminary results indicated a change in the fluorescence lifetime and differences in tissue optical properties between normal and pathological brain tissue specimens. Recently, Du Le et al. [109] implemented the same experimental setup of above for discriminating glioblastoma multiforme and low-grade glioma. They revealed changes in the tissue optical properties such as absorption and scattering coefficients, in addition to the ratio of fluorescence and reflectance intensities at 460  nm. A discriminative accuracy of about 100% sensitivity and 90% specificity was achieved.

Gebhart et al. [110] performed reflectance and transmission spectral measurements using an integrating sphere setup for estimating the tissue optical properties between white matter, grey matter, or glioma. Studies were performed in the wavelength range from 400  nm through 1300  nm and optical properties were estimated using inverse-adding and doubling method. They concluded that there were changes in the absorption parameters in the wavelengths below 600  nm dominated by hemoglobin while water absorption dominates at wavelengths greater than 900  nm. Furthermore, differences in the reduced scattering coefficient were due to the variations in scatterer size and density between the tissues involved in the study.

Gong and coworkers [111] performed an ex vivo analysis of brain tissues using light scattering spectroscopy. Three different algorithms based on spectral slopes, principal component analysis (PCA), and artificial neural networks (ANN) were used for tissue classification. ANN provided the best classification accuracy revealing 80% sensitivity and 93% specificity between normal and brain tumor tissues. In another study, Canpolat et al. [112] applied light scattering spectroscopy on 29 excised brain tumor specimens. Tissue classification was possible with a sensitivity of 93% and specificity of 100%.

5.2. Raman Spectroscopy

Leslie et al. [113] demonstrated the potential of Raman spectroscopy for characterizing various pediatric brain tumors ex vivo. They demonstrated a classification accuracy of 100% sensitivity and 96% specificity for differentiating high-grade ependymoma from low-grade ependymoma. Additionally, the authors obtained 91.5% sensitivity and 97.8% specificity for distinguishing normal brain tissue from low-grade glioma. Wills and coworkers [114] evaluated the application of frozen brain specimens to categorize pediatric brain tumors. The study established very good correspondence between freshly excised and frozen tissue specimens. Moreover, discrimination between normal brain tissue and tumor subtypes was possible with a better classification accuracy. Auner et al. [115] performed a study to explore the application of Raman spectroscopy for the detection of solid tumors in pediatric subjects including malignancies in the brain, kidney, and adrenalin gland. With respect to brain tumors, Raman spectroscopy was able to provide 95.1% accuracy for the training set and 88.9% for the validation set using discriminant function analysis. Koljenovic' et al. [116] analyzed the spectral differences between vital tumor and necrotic brain in glioblastoma tissue samples. Pseudo-colored Raman maps were generated from the Raman microspectroscopy and compared with conventional histopathological slides. The data were analyzed using PCA and K-means clustering, which resulted in 100% classification accuracy. Krafft et al. [117] performed Raman spectral mapping to characterize normal brain tissue, intracranial tumors, and meningioma. The changes in the tissue pathologies were analyzed by using cluster analysis. The study showed that normal brain tissues had an increased lipid content, while intracranial tumors were presented with an increased hemoglobin content and decreased protein-to-lipid ratios. Furthermore, an increased collagen content was found in meningioma. In a similar study, the same authors [118] discussed the changes in Raman biomarkers, such as collagen, nucleic acids, lipids and hemoglobin, for the discrimination of intracranial tumors. Daković et al. [119] evaluated the application of Raman spectroscopy coupled with tissue classification algorithms such as PCA and independent component analysis (ICA) to assess the variations in chemical composition between gray matter and white matter. The study provides insights into different structures of