Sage Journals: Discover world-class research

Abstract

Gastrointestinal cancer is a leading contributor to cancer-related morbidity and mortality worldwide. Early diagnosis currently plays a key role in the prognosis of patients with gastrointestinal cancer. Despite the advances in endoscopy over the last decades, missing lesions, undersampling and incorrect sampling in biopsies, as well as invasion still result in a poor diagnostic rate of early gastrointestinal cancers. Accordingly, there is a pressing need to develop non-invasive methods for the early detection of gastrointestinal cancers. Biomedical optical spectroscopy, including infrared spectroscopy, Raman spectroscopy, diffuse scattering spectroscopy and autofluorescence, is capable of providing structural and chemical information about biological specimens with the advantages of non-destruction, non-invasion and reagent-free and waste-free analysis and has thus been widely investigated for the diagnosis of oesophageal, gastric and colorectal cancers. This review will introduce the advances of biomedical optical spectroscopy techniques, highlight their applications for the early detection of gastrointestinal cancers and discuss their limitations.

Keywords

Gastrointestinal neoplasms early diagnosis optical spectroscopy autofluorescence endoscopy

Introduction

Gastrointestinal cancer (GIC) is a leading contributor to cancer-related morbidity and mortality worldwide.¹ Early diagnosis followed by surgical operations is critical to prolong the survival time of patients with GIC and save direct medical costs. Endoscopy with biopsy is currently the principal method to detect early GIC. However, it has 10%–15% rates of missed lesion depending on the skill and experience of endoscopists, preparation methods before examinations and endoscopic equipment.² The result of biopsy may be falsely negative due to undersampling, incorrect sampling adjacent to the lesion or pathologist error.³ In addition, occasionally unavoidable and unnecessary biopsies taken for precautionary reasons may be invasive to patients.⁴ Biomarkers that can improve patient satisfaction and enhance the predictive results for GICs have been intensively sought and validated in recent years. In spite of the great progress in this field, few biomarkers have become practical screening tools with high sensitivities and specificities.⁵

The emergence of optical spectroscopy (OS) offers great promise for wide use in medical fields, especially for early detection of GICs. The changes in the body leading to diseases such as cancer are thought to be due to structural alterations and biochemical transformations of proteins, nucleic acids, lipids and carbohydrates.^6,7 As a rapid photonic technology, OS is sensitive to characterize all of these changes and consequently generate a unique optical fingerprint (FP) via the absorption and scattering of photons after interactions with different chemical elements and compounds.^8,9 Thus, analysing optical spectra can reveal a superposition of all constituents in the measured sample and consequently provide a vast amount of potentially useful disease-related information.¹⁰ In the context of clinical diagnostics, the technique has the advantages of small sample sizes, ease of sample preparation and not destroying a sample’s biochemistry.¹¹ Furthermore, it has the ability to provide a real-time, non-invasive and in situ optical signature, thus allowing the analysis of complex biological substrates and biofluids.¹² Generally, OS is a cost-effective tool for diagnosis.^13–15

In this review, we provide an overview of the state-of-the-art OS techniques and their possible future applications in the early detection of GICs.

Techniques of OS for diagnosis

Infrared spectroscopy

Infrared (IR) spectroscopy is spectroscopy that considers the IR region of the electromagnetic (EM) spectrum. The IR portion of the EM spectrum is usually divided into three regions: the near infrared (NIR; 14,000–4000 cm⁻¹), middle infrared (MIR; 4000–400 cm⁻¹) and far infrared (FIR; 400–10 cm⁻¹), which are named for their relation to the visible spectrum.¹⁶ The MIR region provides the most important spectral information about the double-bond region (2000–1500 cm⁻¹ associated with C = C, C = N and C = O stretching) and the FP region (1500–600 cm⁻¹, which identifies bending and skeletal FP vibrations).¹⁷ Many biomolecular constituents are known to be absorbed at this region, such as Amide I at ~1650 cm⁻¹, Amide II at ~1550 cm⁻¹, Amide III at ~1260 cm⁻¹, lipids at ~1750 cm⁻¹, glycogen at ~1030 cm⁻¹ and carbohydrates at ~1155 cm⁻¹.¹⁸ The asymmetric and symmetric phosphate stretching vibrations of DNA are located at ~1225 and ~1080 cm⁻¹, respectively.¹⁹

Absorption in the NIR region is typically due to the H-containing functional groups of the constituent molecules such as C–H, N–H and O–H.²⁰ The absorption bands are typically broad and overlapping and primarily correspond to the overtone and combination of the fundamental molecular vibrations.²¹ Consequently, a quantitative analysis with NIR spectroscopy in complex clinical samples is complicated without distinct signatures of individual components. Therefore, a variety of chemometric methods have been used to analyse NIR spectra.²² Chen et al.²³ employed diverse statistical methods to extract NIR spectral features and construct effective diagnostic algorithms for classifying colorectal cancer (CRC) and healthy tissues, including partial least-square (PLS) discriminant analysis, K-nearest neighbour (KNN), classification and regression tree (CART) and random forest (RF) classifiers. The KNN, CART and RF models have the same performance (sensitivity 94.9%, specificity 100%) in the training set, but the CART and RF models exhibit lower misclassified ratios (4/94) than does KNN (7/93) in the test set. Fourier transform infrared (FTIR) spectroscopy is a promising technique that uses the mathematical Fourier transform process to extract quantitative information from the featureless IR spectra.²⁴ Recently, various new clinical applications of FTIR have been reported in the biomedical diagnosis of GICs. It has shown promising performance for identifying neoplasia in the stomach, with a sensitivity of 100% and a specificity of 83.3%.²⁵

Raman spectroscopy

Raman spectroscopy represents a unique optical technique that relies on the inelastic scattering of photons exchanging energy via molecular vibrations after irradiation with a light laser.^26,27 It is a powerful tool for providing details of the chemical composition and molecular structures in cells and tissues due to its major advantages of being label-free and non-destructive, not requiring sample preparation and having high spatial resolution, weak water scattering and no photobleaching, compared with the other optical diagnostic techniques.^28–30 However, the major obstacle in the development of clinical Raman applications is the weak Raman signal, which is attributed to only small fraction of photons (approximately 1 in 10⁶–10⁸) undergoing Raman scattering and strong fluorescence background in biological samples.³¹

Using NIR or ultraviolet (UV) wavelengths as excitation sources can minimize the spectral overlap between the Raman and fluorescence spectra, thereby reducing the fluorescence background.³² As a result of the sample degradation and photo damage at UV excitation, Raman spectroscopy with NIR laser light excitation is particularly attractive with an additional advantage of deeper penetration into the tissue.³³ Therefore, NIR Raman spectroscopy has been studied for the detection of malignant lesions in a number of organs, including the cervix,³⁴ skin,³⁵ breast,³⁶ and kidney.³⁷ High diagnostic accuracies have been achieved for discriminating early gastric cancer (GC; 93%) and CRC (93.2%) from normal tissues.^38,39

Surface-enhanced Raman spectroscopy (SERS) is also a promising solution that can greatly enhance the Raman signal level depending on the nanoscale metal or colloid surface. It is commonly thought that EM enhancement is the dominant contribution of SERS enhancement, which refers to the extent of 10⁴–10⁷.^40,41 An additional chemical contribution (referred to as chemical enhancement) can be observed with a minor extent (typically 10–10²) to the overall enhancement.⁴² Additionally, this technique has been expanded to novel nanoprobes (SERS tags) that fabricate antibodies or other targeting molecules onto metallic nanoparticles.⁴³ A chief advantage of SERS tags is the possibility for multiplexed measurements contributing to the narrow spectral features.⁴⁴ There is still very much work in progress for standard measures of different SERS tags within and between laboratories.⁴⁵ In addition, some sophisticated techniques, such as multi-excitation,⁴⁶ Wiener estimation,⁴⁷ and mathematical statistical methods,^48,49 have been employed to suppress or remove the fluorescence background.

Currently, Raman spectroscopy for the detection and evaluation of cancer and precancerous conditions is mostly focused on the FP region between 800 and 1800 cm⁻¹.^50,51 A comparison study by Short et al.⁵² suggested that the use of high-wavelength (HW) Raman spectroscopy (2050–3100 cm⁻¹) can improve the identification of neoplastic lesions compared with FP. This is probably because HW spectra can offer complimentary information related to both the CH₂/CH₃ stretching of lipids/proteins and the considerably reduced tissue fluorescence background.^53–56 Furthermore, the development of the integrated FP/HW Raman technology provides diagnostic advantages over either the FP or HW modality alone. The receiver operating characteristic (ROC) curve analysis showed that FP/HW Raman technology gave an area under the ROC curve (AUROC) of 0.92 (FP: 0.89, HW: 0.86) for gastric intestinal metaplasia classification.⁵⁷

Diffuse reflection spectroscopy

When a beam of incident light penetrates into an object and undergoes absorption and scattering, part of this radiation ultimately leaves the object in all directions and constitutes so-called diffusely reflected light.⁵⁸ Thus, changes in diffuse reflection spectroscopy (DRS) result primarily from a combination of optical absorption and scattering, which allows valuable information to be obtained on the biochemical composition and microarchitecture of the medium.^59,60 As is known, tumour growth depends on angiogenesis, which can lead to an increase in the haemoglobin concentration. Consequently, this alteration can increase the absorption coefficient and reduce the diffuse reflectance signal.⁶¹ Moreover, in the progression to cancer, dysplasia can be represented as an increase in epithelial thickness, which gives rise to a reduced quantity of light into the deeper tissues and results in a decrease in tissue scattering and, consequently, the intensity of diffuse reflectance.⁶² Currently, studies have incorporated DRS into current imaging modalities (e.g. magnetic resonance imaging, sonography) to improve the diagnostic performance.⁶³ The derivations of diffuse optics, such as diffuse optical tomography (DOT) and the time gated scanning diffuse optical imaging (DOI) method, also enhance the sensitivity to absorbing targets.⁶⁴

Autofluorescence

When fluorescent molecules absorb light energy in a specific wavelength, they are activated from the ground state to an excited singlet state and then return to their ground energy state by releasing energy that generates fluorescent light.⁶⁵ The signal amplitude and spectral shape depend strictly on the concentration, excitation wavelength, light scattering and absorption properties of the fluorescent molecules.^66,67 On the basis of the material source, fluorescent molecules can be either endogenous or exogenous fluorophores. Autofluorescence (AF) is a natural fluorescence of the organisms that differs from the ‘induced fluorescence’ that is obtained by the introduction of an exogenous fluorochrome.⁶⁸ Compared with induced fluorescence, AF has aroused widespread concern because of its convenience and biosecurity.

Neoplasia can cause oxidation of nicotinamide adenine dinucleotide (NADH) that yields decreased AF intensity.⁶⁹ The angiogenesis in tumour is able to lead the difference of absorption effect of haemoglobin and therefore influence the spectral characteristics of the measured fluorescence.⁷⁰ A number of biological fluorophores, such as flavin adenine dinucleotide (FAD), tryptophan, collagen, elastin and endogenous porphyrin, can give rise to fluorescence signals.^71–73 Thus, the fluorescence signals characterize changes in endogenous fluorophores related to the physiological state on the occurrence of a pathological process.⁷⁴ The logistic models constructed with 30 or 40 of the most important AF spectra selected by principal component analysis (PCA) could clearly differentiated between colorectal carcinoma and normal tissues with AUROCs of 0.88 and 0.91, respectively.⁷⁵ However, there are still some limitations to be addressed when applying AF to guide the clinical diagnosis of digestive tract tumours, such as the weak intensity of tumour AF,⁷⁶ unsatisfactory specificity with inflammation and hyperplasia.⁷⁷ To improve AF applications, a combination with DRS and narrow band image (NBI) is often used in tissue analysis.^78,79

In vivo OS based on endoscopy

Early detection followed by the adequate treatment of precancerous/premalignant lesions with the potential to progress to invasive cancer is a critical factor in the prognosis of patients with GIC. However, the performance of early detecting dysplasia and malignancy remains poor when using conventional white-light reflectance (WLR) imaging. Recent advancement of enhancement technologies has led to efficient evolutions in endoscopic images. Magnification endoscopy equipped with a system of movable lenses enables the size and visible details of mucosa to be magnified up to 150 times.⁸⁰ NBI via a narrow-bandwidth filter allows penetration at peak wavelengths of 415 (blue) and 540 (green) nm, which highlights the superficial mucosal vasculature.^81,82 However, these advances heavily rely on visual information of gross morphologic changes and not on biochemical or molecular analysis, resulting in poor ability for an early diagnosis.⁸³

Currently, endoscopic biopsy is the standard criterion for the clinical localization and diagnosis of GICs. However, a single biopsy from the worst-appearing area is associated with sampling errors and demands for doctors with abundant clinical experiences.^50,83 Furthermore, multiple-biopsy protocols can maximize disease ascertainment but at the same time bring a great deal of discomfort to patients.⁸⁴ Along with successful spectroscopic outcomes, the combination of spectroscopy and an endoscopic system achieved dramatic results in real time and objective diagnosis of malignant and premalignant gastrointestinal lesions with high specificity.

OS diagnosis based on tissues

Histopathology is the gold standard for cancer diagnosis, but it is invasive and prone to interobserver variability.^85,86 The kappa values between two pathologists ranged from 0.59 to 0.69 in tumour pathological diagnosis.^87–89 OS is able to permit non-invasive and objective analysis of biochemical and subcellular alterations of tissues, and the results are well correlated with histopathology. In the diagnosis of dysplasia in Barrett’s oesophagus (BE), a high level of agreement (kappa value 0.89) was observed between spectral data and consensus opinion of pathology.⁹⁰ Good correlation between DR spectroscopy and histopathology was also observed in the classification of intestinal ischemic-reperfusion injuries (r = 0.91).⁹¹ Moreover, OS offers the facility of analysis in vivo and targeting biopsy, which can assist in making a real-time decision during endoscopic operations. Hence, an enticing application of OS is to help distinguish malignant tissues from benign ones.

Oesophageal cancer

Oesophageal cancer is the sixth leading cause of cancer-related mortality and has 5-year survival rates of 15%–25%.⁹² With delayed symptoms, the disorder is frequently diagnosed at an advanced stage. To improve the outcomes, it is important to diagnose oesophageal cancer in the early stages. There is increasing evidence that OS is a powerful technique in the early diagnosis of cancer. Ishigaki et al.⁹³ performed an ex vivo Raman spectroscopy analysis in early-stage (stages 0 and I) oesophageal cancer tissues and obtained a discriminative sensitivity of 81.0% and a specificity of 94.0% between cancerous and normal samples.

BE is considered to be a precursor lesion to oesophageal adenocarcinoma.⁹⁴ It is thought to evolve through low-grade dysplasia (LGD), high-grade dysplasia (HGD) and adenocarcinoma.⁹⁵ Patients with BE carry a risk of adenocarcinoma 30–40 times greater than those without this condition.⁹⁶ In particular, the annual risk of malignant progression in BE patients with HGD is higher than nondysplastic BE (5%–20% vs 0.12%–0.5%).⁹⁷ Detection of dysplasia and early neoplasia in the setting of BE can potentially improve patient outcomes and prevent progression to invasive malignancy.⁹⁸ In an ex vivo study of 673 samples of BE (acquired from 62 patients), endoscopic Raman spectroscopy demonstrated a sensitivity of 86% and a specificity of 88% for distinguishing advanced neoplasia (HGD/adenocarcinoma) from LGD, nondysplastic BE and normal squamous oesophagus tissue.⁹⁹

BE tissues with various conditions show different optical physiological characteristics which are valuable for discriminant diagnosis. Bergholt et al.¹⁰⁰ utilized confocal Raman spectroscopy to detect BE in vivo and recorded a series of peaks that reflected expression levels of different substances (proteins, lipids and porphyrins) in columnar-lined epithelium (CLE) and nondysplastic BE and HGD BE, by which they obtained a sensitivity of 87% (67/77) and specificity of 84.7% (610/720) for detecting HGD in BE and AUROCs of 0.88, 0.84 and 0.90 for the discriminations of CLE versus nondysplastic and HGD BE, nondysplastic BE versus CLE and HGD BE and HGD BE versus CLE and nondysplastic BE, and the suspicious BE according to this technique was also confirmed by subsequent biopsy and histopathology.

GC

GC is the fifth most common cancer worldwide,¹⁰¹ with an estimated new cases of 951,000 (6.8% of the total) and 745,000 deaths (9.1% of the total) in 2012. The survival rate mainly depends on GC stages;¹⁰² therefore, detection in the early stage is vitally important in ensuring an excellent prognosis.

The carcinogenesis of GC is a multistep process that is largely initiated from infection with Helicobacter pylori (H. pylori), and it is progressively involved in the transformation from intestinal metaplasia (IM) to dysplasia before the onset of neoplastic transformation.¹⁰³ Accurate screening for early detection in high-risk pathological lesions is crucial to reducing mortality. Teh et al.¹⁰⁴ assessed the capability of NIR Raman spectroscopy for early diagnosis of H. pylori infection and IM lesions and found that the PCA–linear discriminant analysis (LDA) diagnostic algorithms were able to identify normal mucosa with a sensitivity of 91.7% and a specificity of 80.0%, H. pylori infection with a sensitivity of 80.0% and a specificity of 100% and IM with a sensitivity of 80.0% and a specificity of 92.7%. Huang et al.¹⁰⁵ acquired significant differences between normal and dysplastic gastric tissues within 0.5 s under the PCA-LDA algorithm and yielded a diagnostic sensitivity of 94.4% (17/18) and a specificity of 96.3% (52/54).

Some gastric ulcers are ulcerated gastric carcinomas; therefore, differentiating malignant from benign ulcers is quite important. Gastric ulcers have different optical physiological characteristics from gastric adenocarcinoma. Li et al.¹⁰⁶ measured the DR spectra of four types of stomach tissues (normal mucosa, chronic gastric ulcer, poorly differentiated and undifferentiated gastric adenocarcinomas) under wavelengths ranging from 400 to 650 nm and found that malignant tissues showed an increase in optical absorption at the HbO₂ band (DR spectral intensity: undifferentiated adenocarcinoma < differentiated adenocarcinomas < gastric ulcer < normal mucosa) with concomitant significant decrease in the DR ratio of R540/R575. Bergholt et al.¹⁰⁷ assessed the diagnostic potential of Raman endoscopy for in vivo discrimination of benign and malignant gastric ulcers. They used a partial least-square discriminant algorithm (PLS-DA) to analyse a total of 1102 Raman spectra from 71 patients, with a sensitivity of 90.8% and a specificity of 93.8% for identifying normal mucosa, 84.7% and 94.5% for identifying benign ulcers and 82.1% and 95.3% for identifying malignant ulcerous lesions.

CRC

In recent years, many published articles have reported the application of OS for the detection of CRC. Compared with normal tissues, malignant colorectal tissues in FTIR analysis showed increased bands at positions of 1460 cm⁻¹ (related to lipid), 3260 cm⁻¹ (related to nitrogen-containing compounds and water), 1640 and 1550 cm⁻¹ (amide groups related to protein) and 1080 cm⁻¹ (related to nucleic acid) and decreased peaks at positions of 2925, 1400 and 1740 cm⁻¹ (related to lipid) and 1160 cm⁻¹ (related to carbohydrate),¹⁰⁸ by which a diagnostic sensitivity of 96.6% was gained via Fisher’s discriminant analysis (FDA). With the integration of NIR AF and NIR DRS, normal and malignant colonic tissues were successfully classified with a high diagnostic accuracy of 95.8%.¹⁰⁹

CRCs are usually adenocarcinomas, and more than 70% of them occur from the transformation of adenomatous polyps.^110,111 Therefore, the early detection and removal of colorectal adenomas can effectively prevent CRC. Considerable efforts of colonoscopy in the field have been spent on optimizing the yield of polyp detection. Rodriguez-Diaz et al.¹¹² successfully applied elastic-scattering spectroscopy to the real-time classification of polyps as neoplastic or non-neoplastic lesions with a sensitivity of 91.5%, specificity of 92.2% and accuracy of 91.9%. In particular, a high-confidence rate of approximately 90% was associated with accuracies of 90.6% for polyps smaller than 1 cm and 90.1% for polyps equal to or smaller than 5 mm. Inelastic-scattering spectroscopy provides a diagnostic sensitivity of 90.9% and a specificity of 83.3% for differentiating adenoma from hyperplastic polyps.¹¹³ Another endoscopic method based on laser-induced AF spectroscopy has been considered as a complimentary technique for the real-time prediction of polyp histology.^114–116

OS diagnosis based on biofluids

Detection of biomarkers in biomedical fluids is the well-known and widely accepted laboratory test for tumour diagnosis. However, traditional tumour markers, such as carcinoembryonic antigen (CEA), CA19-9 and CA12-5, are not recommended for diagnosing oesophagus and stomach cancers due to their low sensitivities, and emerging biomarkers need more extensive validation.^117,118 OS can provide comprehensive information of a liquid sample, which is reasonable for diagnosis and coincided with the recent trend that combination of multiple bio-markers instead of single one for improving diagnostic performance.¹¹⁹

Blood

Peripheral blood samples are an ideal material for minimally invasive diagnosis, which can be performed expediently and repeatedly, and continuous sampling and testing are suitable for monitoring high-risk subjects.¹²⁰ A variety of studies have been conducted to evaluate the diagnostic performance of OS techniques in blood samples. Lualdi et al.¹²¹ found that a blood fluorescence emission peak around 615–635 nm significantly differed between 172 patients with CRC and 169 health donors, with an AUROC of 0.724. Multivariate analysis of the IR spectral characterization of peripheral blood mononuclear cells (PBMCs) and plasma principal components yielded a sensitivity of 81.5% and a specificity of 71.4% for distinguishing patients with CRC from healthy individuals, and when comparing CRC with the healthy population and patients with dysplastic adenomatous polyps, a sensitivity of 85% and a specificity of 60.8% were obtained.¹²²

Due to the advantages of weak water scattering, Raman spectroscopy analysis of blood has been applied for separating cancerous samples from normal ones on a finer scale. Li et al.¹²³ achieved a sensitivity of 94.4%, specificity of 100% and accuracy of 97.7% for the differentiation of oesophageal cancer (n = 36) and normal blood plasma (n = 50). Feng et al.¹²⁴ achieved a sensitivity of 100% and a specificity of 97% for non-invasive GC detection. Lin et al.¹²⁵ confirmed the capability of SERS for CRC detection, with a diagnostic sensitivity of 97.4% and a specificity of 100%.

The potential application of SERS for the early detection of cancer-related serum-constituents is of greatest concern. Ito et al.¹²⁶ compared the SERS peak heights of blood samples among patients with GC, CRC or benign diseases and found that the SERS peak heights of samples from patients with stage II–IV GC, stage II–IIIa CRC or stage IIIb–IV CRC were significantly higher than those from patients with benign diseases. A significant difference also existed between stage I and advanced GC. The peak heights of SERS spectra from patients with benign diseases were lower than those from patients with stage I GC, although no statistically significant differences (p = 0.068) were observed. These results indicated that SERS spectra intensified with progression of the disease. Feng et al.¹²⁷ explored the use of SERS-based blood plasma analysis for the early detection of precancerous lesions and found that the diagnostic sensitivity and specificity for differentiating polyp samples were 71.4% and 95.6% from CRC and 91.3% and 80% from normal subjects.

SERS detection is prone to interference from any constituent changes in the blood due to its high sensitivity.¹²⁸ To further improve this blood test, several reports demonstrated that SERS spectra of the purified serum proteins helped determine the magnitude of the SERS enhancement due to the elimination of other native plasma constituents and exogenous substances.^128,129 Serum albumin and globulin from CRC patients (n = 103) and healthy volunteers (n = 103) were analysed, and the diagnostic performance was evaluated by the algorithms combined with PCA or PLS and LDA.¹²⁹ The various models showed different diagnostic sensitivities and specificities: 100% and 100% for an albumin PCA model, 100% and 99% for a globulin PCA model, 95.6% and 91.3% for an albumin PLS model and 91.3% and 95.6% for a globulin PLS model. Similarly, using a purification procedure, SERS detection had a sensitivity and specificity of 100% for differentiating GC patients from healthy volunteers.¹²⁸

The levels of circulating nucleic acids are frequently associated with tumour development and have been proposed to be a liquid biopsy for the detection and monitoring of cancers.^130,131 Recently, SERS as a label-free tool for the ultrasensitive detection of RNA/DNA with the ability to detect a single base change has generated tremendous interest.^132–134 It has been shown that silver nanoparticles can be used as an excellent SERS-active substrate for detecting serum RNA in several investigations. The sensitivity and specificity of serum RNA SERS spectra were 100% and 94.1% for identifying GC (n = 31) from healthy volunteers (n = 34)¹³⁵ and 89.1% and 95.6% detecting CRC (n = 55) from normal controls (n = 45).¹³⁶

It has been confirmed that the efficient discrimination of circulating tumour cells (CTCs) is expected to provide valuable information for tumour diagnosis, prognosis, recurrence and metastasis.¹³⁷ An integrated approach based on magnetic trapping and SERS imaging has been reported for CTC detection and enumeration. Sun et al.¹³⁸ synthesized biotin-labelled aptamers specific to DLD-1 cells (a colorectal adenocarcinoma cell line) and subsequently conjugated them to magnetic beads via the biotin–streptavidin interaction to develop a capture system for CTCs, which identified CTCs readily and accurately with a SERS imaging technique. Shi et al.¹³⁹ used magnetic scattering nanoparticles (MNPs) decorated with folate (its receptor is overexpressed in many cancer cells) to target cancer cells and found that an increasing SERS signal was detected from the cells targeted by both SERS NPs and MNPs; this system was able to effectively trap cells within a large flow velocity range (0.2 to 12 cm s⁻¹).

Gastric juice

The feasibility of detecting GC in gastric juice has been investigated. Zhou et al.¹⁴⁰ demonstrated that the first peak of intrinsic fluorescence intensity (P1FI) in gastric juice was able to detect GC from benign gastric diseases with a sensitivity of 83.2%, specificity of 80.7% and accuracy of 82.0%, and it was also valuable for the diagnosis of early GC, with a sensitivity of 69.7%, specificity of 57.1% and accuracy of 63.2%.¹⁴¹ Furthermore, measuring the fluorescence intensity of gastric juice by mass spectra offers the possibility of detecting fluorescence biomarkers.¹⁴² Three aromatic amino acids in gastric juice could correctly predict gastric malignancies with AUROCs of 0.838, 0.856 and 0.816 for tyrosine, phenylalanine and tryptophan, respectively. Additional work is required to clarify the reasons for the changes in the fluorescence emission spectrum, and its diagnostic performance should be further validated in future studies.

Limitations and prospects

So far, OS techniques have not been conventionally used for clinical diagnosis. There are two obstacles to the clinical translation of the techniques: (1) Intrinsic limitations of OS. For example, conventional Raman spectroscopy suffers from weak signals and strong background AF. (2) Diagnostic validation of OS. For instance, it is difficult to compare the results from different research groups due to the variability of instruments, measurement protocols and procedures for data analysis.

Each spectroscopic technique suffers from certain drawbacks. The depth of penetration of IR radiation depends on the wavelength. MIR can only examine the penetration depth of ~10 µm through tissue, but NIR penetrates to great depths in the range of millimeter–centimeter.¹⁴³ In Raman spectroscopy, excitation of light can be UV, visible or NIR; therefore, different tissue types and samples require optimization of excitation wavelengths.⁷ The biggest drawback of SERS is the introduction of metal into the measurement system, the variation of which in size and structure may result in a large variability of the SERS spectrum.¹⁴⁴ The point monitoring of DR spectroscopy is time-consuming for screening a larger area.^145,146

Related studies involve different analytical techniques and methods, which makes it difficult to comprehensively assess the diagnostic value of a technology. For example, the diagnostic performance of AF spectroscopy is still controversial. Some studies have indicated that the detection rate of neoplasms was not significantly different between AF imaging and WL systems,^147,148 and some studies showed that AF endoscopy slightly elevated sensitivity for the detection of neoplasia, but its clinical value is limited probably due to over-diagnosis in regenerative hyperplasia, intestinal metaplasia, ulcerations or inflammation observations.^149,150 Therefore, there is a need for more investigations on reliability for large-scale applications.

The integrated Raman and NIR AF endoscopy method holds great promise for enhancing in vivo gastric tissue diagnosis and characterization,¹⁵¹ with an AUROC of 0.985 for diagnosing GC, which is superior to either the Raman or NIR AF alone. In addition, the combination of multiple excitation wavelengths can augment the available information. Liu et al.¹⁵² used 337, 375, 405 and 460 nm as excitation wavelengths to characterize the AF of fresh colon tissues and found that under 337 nm excitation, the AF spectrum provided more diagnostic information for recognizing neoplastic changes, and compared with single excitation, the combination of feature extractions of four excitation wavelengths by multispectral data analysis improved the specificity (91.4% vs 80.0%) and classification accuracy (90.3% vs 83.9%) while maintaining the same sensitivity (88.9%). Therefore, future studies may focus on an OS combination to overcome the limitations of any single one.

Technical development over the past decades has resulted in large and overlapping datasets that increase the difficulty of interpretation. Currently, to reduce the dimensionality of such datasets, many dimension reduction methods have been developed for this purpose. PCA is one of the most common and popular linear dimension reduction approaches that use an orthogonal transformation to convert a larger set of possibly correlated variables into a smaller set of uncorrelated variables with minimal loss of information.^153,154 However, many high-dimensional data in real-world applications may lie on a low-dimensional nonlinear manifold.¹⁵⁵ Recently, there have been some renewed interests on the algorithms for constructing nonlinear low-dimensional manifolds, such as ant colony optimization (ACO), KNN, CART, RF, supervised neighbourhood preserving embedding (SNPE) and support vector machine (SVM).^23,156–158

In summary, biomedical OS, including IR spectroscopy, Raman spectroscopy, diffuse scattering spectroscopy and AF, is capable of providing structural and chemical information about biological specimens with the advantages of non-destruction, non-invasion, reagent-free and waste-free analysis and has thus been widely investigated for the diagnosis of oesophageal, gastric and colorectal neoplasms. OS detected in situ via endoscopy is valuable for the early identification of premalignant and malignant lesions, and the OS of blood, particularly its components, is valuable in the diagnosis of GICs. OS techniques have their individual characteristics in cancer diagnosis (Table 1), and the multi-technical approach is a tendency to make more efficient diagnosis. During the investigation, innovative modalities for data analysis have been developed to reduce the dimensionality of spectroscopic datasets, and the combination of multiple spectral signals via mathematical models becomes a frequent method for the diagnostic application of spectral data because a single signal is usually not enough for highly sensitive and specific diagnoses. Incorporating different OS techniques with other biological techniques, such as nanotechnology, can effectively overcome their drawbacks alone. However, it is difficult to compare the results from different research groups due to the variability of instruments, measurement protocols and procedures for data analysis. It is expected that standardization in this field will occur, which is necessary for translating these approaches into routine clinical diagnoses.

Table 1.

Characteristics of various optical spectroscopy techniques in diagnosis.

	Raman spectroscopy	Near-infrared spectroscopy	Mid-infrared spectroscopy	Diffuse reflectance spectroscopy	Autofluorescence
Incident beam	Visible, near infrared or ultraviolet	Infrared	Infrared	Visible and near infrared	Laser, ultraviolet, visible and near infrared
Light type detected	Inelastic scattering of monochromatic light	Absorption of photons	Absorption of photons	Elastic scattering	Emission spectrum of endogenous fluorophores
Wave number range (cm⁻¹)	50–4000	4000–14,000	400–4000	NA	NA
Spatial resolution	~μm	~mm	~μm	~mm	NA
Sensitive molecules or molecular bonds	Polar (e.g. C–O, N–O and O–H) or vibrating chemical bonds (e.g. C=C, C≡C)	H-containing functional groups (e.g. C–H, N–H and O–H)	Double-bond region (e.g. C = C, C = N and C = O)	NA	Endogenous fluorophores (NADH, FAD, haemoglobin, tryptophan, porphyrin, etc.)
Label free and real time	Yes	Yes	Yes	Yes	Yes
Special sample preparation and tissue destruction	No	No	No	No	No
Detection of biological sample type	Liquid and solid	Solid; transmission methods for liquid samples	Solid and dried liquid samples	Liquid and solid	Liquid and solid
Detection in vivo in cooperation with endoscopy	Yes	Transmission in vivo analysis	Transmission in vivo analysis	Yes	Yes

NADH: nicotinamide adenine dinucleotide; FAD: flavin adenine dinucleotide; NA: not available.

Footnotes

Acknowledgements

Q.-S.W. prepared the article and T.W. and K.-H.Z. reviewed and provided beneficial comments and discussions.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by National Natural Science Foundation of China (No. 81560479) and Science and Technology Project of the Education Department of Jiangxi Province (No. KJLD13014).

References

Torre

Bray

Siegel

. Global cancer statistics, 2012. CA Cancer J Clin 2015; 65: 87–108.

Sugano

Screening of gastric cancer in Asia. Best Pract Res Clin Gastroenterol 2015; 29: 895–905.

Telford

Enns

RA.

Endoscopic missed rates of upper gastrointestinal cancers: parallels with colonoscopy. Am J Gastroenterol 2010; 105: 1298–1300.

Lau

O’Donoghue

. Quantitative spectroscopic imaging for non-invasive early cancer detection. Optics Express 2008; 16: 16227–16239.

Ahlquist

DA.

Molecular detection of colorectal neoplasia. Gastroenterology 2010; 138: 2127–2139.

Chen

Dai

Zhou

. Raman spectroscopy analysis of the biochemical characteristics of molecules associated with the malignant transformation of gastric mucosa. PLoS One 2014; 9: e93906.

Dukor

RK.

Vibrational spectroscopy in the detection of cancer. In: Chalmers

Griffiths

(eds) Handbook of vibrational spectroscopy. Chichester: John Wiley & Sons Ltd, 2002, pp. 3335–3361.

Evers

Hendriks

Lucassen

. Optical spectroscopy: current advances and future applications in cancer diagnostics and therapy. Future Oncol 2012; 8: 307–320.

Hendriks

Balthasar

Lucassen

. Nerve detection with optical spectroscopy for regional anesthesia procedures. J Transl Med 2015; 13: 380.

10.

Old

Fullwood

Scott

. Vibrational spectroscopy for cancer diagnostics. Anal Methods 2014; 6: 3901–3917.

11.

Lasch

Kneipp

Biomedical vibrational spectroscopy. Hoboken, NJ: John Wiley & Sons, 2008.

12.

Green

Cobb

Brennan

. Optical diagnostic techniques for use in lesions of the head and neck: review of the latest developments. Br J Oral Maxillofac Surg 2014; 52: 675–680.

13.

Shim

Song

Marcon

. In vivo near-infrared Raman spectroscopy: demonstration of feasibility during clinical gastrointestinal endoscopy. Photochem Photobiol 2000; 72: 146–150.

14.

Rodriguez-Diaz

Bigio

Singh

SK.

Integrated optical tools for minimally invasive diagnosis and treatment at gastrointestinal endoscopy. Robot Computer Integr Manuf 2011; 27: 249–256.

15.

Lewis

Ghosal

. Evaluation of FTIR spectroscopy as a diagnostic tool for lung cancer using sputum. BMC Cancer 2010; 10: 640.

16.

Cozzolino

Cynkar

Shah

. Can spectroscopy geographically classify Sauvignon Blanc wines from Australia and New Zealand? Food Chem 2011; 126: 673–678.

17.

Martin

Kelly

Llabjani

. Distinguishing cell types or populations based on the computational analysis of their infrared spectra. Nat Protoc 2010; 5: 1748–1760.

18.

Theophilou

Paraskevaidi

Lima

. Extracting biomarkers of commitment to cancer development: potential role of vibrational spectroscopy in systems biology. Exp Rev Mol Diagn 2015; 15: 693–713.

19.

Mitchell

Gajjar

Theophilou

. Vibrational spectroscopy of biofluids for disease screening or diagnosis: translation from the laboratory to a clinical setting. J Biophotonics 2014; 7: 153–165.

20.

Kilic

Caiazzo

Fabris

. Near-infrared spectroscopy-intravascular ultrasound: scientific basis and clinical applications. Eur Heart J Cardiovasc Imaging 2015; 16: 1299–1306.

21.

Chen

Lin

. Near-infrared spectroscopy as a diagnostic tool for distinguishing between normal and malignant colorectal tissues. Biomed Res Int 2015; 2015: 472197.

22.

Tan

Qin

An ensemble method based on a self-organizing map for near-infrared spectral calibration of complex beverage samples. Anal Bioanal Chem 2008; 392: 515–521.

23.

Chen

Lin

. Diagnosis of colorectal cancer by near-infrared optical fiber spectroscopy and random forest. Spectrochim Acta A Mol Biomol Spectrosc 2015; 135: 185–191.

24.

Garhwal

Fourier transform infrared (FT-IR) spectroscopy in an-overview. Res J Med Plant 2011; 5: 127–135.

25.

Wang

Ling

. Detection of gastric cancer with Fourier transform infrared spectroscopy and support vector machine classification. Biomed Res Int 2013; 2013: 942427.

26.

Butler

Ashton

Bird

. Using Raman spectroscopy to characterize biological materials. Nat Protoc 2016; 11: 664–687.

27.

Kourkoumelis

Balatsoukas

Moulia

. Advances in the in vivo Raman spectroscopy of malignant skin tumors using portable instrumentation. Int J Mol Sci 2015; 16: 14554–14570.

28.

Guze

Short

Zeng

. Comparison of molecular images as defined by Raman spectra between normal mucosa and squamous cell carcinoma in the oral cavity. J Raman Spectros 2011; 42: 1232–1239.

29.

Cals

FLJ

Schut

TCB

Koljenovi

. Method development Raman spectroscopy-based histopathology of oral mucosa. J Raman Spectros 2013; 44: 963–972.

30.

Kong

Kendall

Stone

. Raman spectroscopy for medical diagnostics – from in-vitro biofluid assays to in-vivo cancer detection. Adv Drug Deliv Rev 2015; 89: 121–134.

31.

Ellis

Goodacre

Metabolic fingerprinting in disease diagnosis: biomedical applications of infrared and Raman spectroscopy. Analyst 2006; 131: 875–885.

32.

DaCosta

Wilson

Marcon

NE.

Fluorescence and spectral imaging. ScientificWorldJournal 2007; 7: 2046–2071.

33.

Huang

McWilliams

Lui

. Near-infrared Raman spectroscopy for optical diagnosis of lung cancer. Int J Cancer 2003; 107: 1047–1052.

34.

Duraipandian

Zheng

. Near-infrared-excited confocal Raman spectroscopy advances in vivo diagnosis of cervical precancer. J Biomed Opt 2013; 18: 067007.

35.

Silveira

Pacheco

Bodanese

. Discrimination of non-melanoma skin lesions from non-tumor human skin tissues in vivo using Raman spectroscopy and multivariate statistics. Lasers Surg Med 2015; 47: 6–16.

36.

Liu

Zhou

Sun

. Resonance Raman and Raman spectroscopy for breast cancer detection. Technol Cancer Res Treat 2013; 12: 371–382.

37.

Wills

Kast

Stewart

. Diagnosis of Wilms’ tumor using near-infrared Raman spectroscopy. J Pediatr Surg 2009; 44: 1152–1158; discussion 1158.

38.

Kawabata

Kikuchi

Okazaki

. Near-infrared multichannel Raman spectroscopy with a 1064 nm excitation wavelength for ex vivo diagnosis of gastric cancer. J Surg Res 2011; 169: e137–e143.

39.

Chen

Zhang

. Identification and characterization of colorectal cancer using Raman spectroscopy and feature selection techniques. Optics Express 2014; 22: 25895–25908.

40.

Negri

Dluhy

RA.

Ag nanorod based surface-enhanced Raman spectroscopy applied to bioanalytical sensing. J Biophotonics 2013; 6: 20–35.

41.

Ling

Xiao

. Surface enhanced Raman spectroscopy on a flat graphene surface. Proc Natl Acad Sci USA 2012; 109: 9281–9286.

42.

Stiles

Dieringer

Shah

. Surface-enhanced Raman spectroscopy. Annu Rev Anal Chem 2008; 1: 601–626.

43.

Wang

Yan

Chen

SERS tags: novel optical nanoprobes for bioanalysis. Chem Rev 2013; 113: 1391–1428.

44.

Nolan

Duggan

Liu

. Single cell analysis using surface enhanced Raman scattering (SERS) tags. Methods 2012; 57: 272–279.

45.

Nolan

Duggan

Condello

Optimization of SERS tag intensity, binding footprint, and emittance. Bioconjug Chem 2014; 25: 1233–1242.

46.

McCain

Willett

Brady

DJ.

Multi-excitation Raman spectroscopy technique for fluorescence rejection. Optics Express 2008; 16: 10975–10991.

47.

Chen

Ong

Lin

. Optimization of advanced Wiener estimation methods for Raman reconstruction from narrow-band measurements in the presence of fluorescence background. Biomed Opt Express 2015; 6: 2633–2648.

48.

Lopez

Ruisanchez

Callao

MP.

Figures of merit of a SERS method for Sudan I determination at traces levels. Spectrochim Acta A Mol Biomol Spectrosc 2013; 111: 237–241.

49.

Zhao

Lui

McLean

. Automated autofluorescence background subtraction algorithm for biomedical Raman spectroscopy. Appl Spectrosc 2007; 61: 1225–1232.

50.

Sun

Huang

. Preliminary study of differentiating smears from cancerous and non-cancerous nasopharyngeal tissue using confocal Raman spectroscopy. J Cancer Res Clin Oncol 2016; 142: 823–831.

51.

Duraipandian

Zheng

. Simultaneous fingerprint and high-wavenumber confocal Raman spectroscopy enhances early detection of cervical precancer in vivo. Anal Chem 2012; 84: 5913–5919.

52.

Short

Tai

Owen

. Using high frequency Raman spectra for colonic neoplasia detection. Optics Express 2013; 21: 5025–5034.

53.

Santos

Wolthuis

Koljenovic

. Fiber-optic probes for in vivo Raman spectroscopy in the high-wavenumber region. Anal Chem 2005; 77: 6747–6752.

54.

Koljenovic

Bakker Schut

Wolthuis

. Tissue characterization using high wave number Raman spectroscopy. J Biomed Opt 2005; 10: 031116.

55.

Zheng

Low

. High wavenumber Raman spectroscopy for in vivo detection of cervical dysplasia. Anal Chem 2009; 81: 8908–8915.

56.

Lin

Cheng

Huang

Optical diagnosis of laryngeal cancer using high wavenumber Raman spectroscopy. Biosens Bioelectron 2012; 35: 213–217.

57.

Lin

Wang

Zheng

. Rapid fiber-optic Raman spectroscopy for real-time in vivo detection of gastric intestinal metaplasia during clinical gastroscopy. Cancer Prev Res 2016; 9: 476–483.

58.

Torrent

Barrón

Diffuse reflectance spectroscopy of iron oxides. In: AT

Hubbard

(ed.) Encyclopedia of surface and colloid science, Vol. 1. New York: Marcel Dekker, 2002, pp. 1438–1446.

59.

Bish

Rajaram

Nichols

. Development of a noncontact diffuse optical spectroscopy probe for measuring tissue optical properties. J Biomed Opt 2011; 16: 120505.

60.

Garcia-Uribe

Zou

Duvic

. In vivo diagnosis of melanoma and nonmelanoma skin cancer using oblique incidence diffuse reflectance spectrometry. Cancer Res 2012; 72: 2738–2745.

61.

Pimenta

Castanheira

Minas

Optical microsystem for analysis of diffuse reflectance and fluorescence signals applied to early gastrointestinal cancer detection. Sensors 2015; 15: 3138–3153.

62.

Pimenta

Carmo

Correia

. Characterization of silicon photodiodes for diffuse reflectance signal extraction. In: Jorge

Belinha

Martins

. (eds) 2015 IEEE 4th Portuguese meeting on bioengineering (ENBENG). New York: IEEE, 2015, pp. 1–4.

63.

La Yun

Kim

Jang

. Does adding diffuse optical tomography to sonography improve differentiation between benign and malignant breast lesions? Observer performance study. J Ultrasound Med 2015; 34: 749–757.

64.

Chen

Florian

Rajesh

. Recovering the superficial microvascular pattern via diffuse reflection imaging: phantom validation. Biomed Eng Online 2015; 14: 87.

65.

Behlke

Huang

Bogh

. Fluorescence and fluorescence applications. Coralville, IA: Integrated DNA Technologies, 2005.

66.

Venugopal

Nazeer

Balan

. Autofluorescence spectroscopy augmented by multivariate analysis as a potential noninvasive tool for early diagnosis of oral cavity disorders. Photomed Laser Surg 2013; 31: 605–612.

67.

Holz

Boerwinkel

Meijer

. Optimized endoscopic autofluorescence spectroscopy for the identification of premalignant lesions in Barrett’s oesophagus. Eur J Gastroenterol Hepatol 2013; 25: 1442–1449.

68.

Palumbo

Pratesi

Lasers and current optical techniques in biology. London: Royal Society of Chemistry, 2004.

69.

Angelova

Borisova

Zhelyazkova

. Fluorescence spectroscopy of gastrointestinal tumors: in vitro studies and in vivo clinical applications. In: Spigulis

Kuzmina

(eds) 1st international conference. Bellingham, WA: SPIE (International Society for Optics and Photonics), 2013, pp. 903209–903211.

70.

Tajiri

Autofluorescence endoscopy for the gastrointestinal tract. Proc Jpn Acad Ser B Phys Biol Sci 2007; 83: 248–255.

71.

Monici

Cell and tissue autofluorescence research and diagnostic applications. Biotechnol Annu Rev 2005; 11: 227–256.

72.

Yuvaraj

Udayakumar

Jayanth

. Fluorescence spectroscopic characterization of salivary metabolites of oral cancer patients. J Photochem Photobiol B 2014; 130: 153–160.

73.

Croce

Bottiroli

Autofluorescence spectroscopy and imaging: a tool for biomedical research and diagnosis. Eur J Histochem 2014; 58: 2461.

74.

Croce

De Simone

Freitas

. Human liver autofluorescence: an intrinsic tissue parameter discriminating normal and diseased conditions. Lasers Surg Med 2010; 42: 371–378.

75.

Duchac

Zavadil

Vranova

. Peroperative optical autofluorescence biopsy – verification of its diagnostic potential. Lasers Med Sci 2011; 26: 325–333.

76.

Ernst

Herth

FJ.

Principles and practice of interventional pulmonology. New York: Springer, 2012.

77.

Nakanishi

Ohsaki

Kurihara

. Color auto-fluorescence from cancer lesions: improved detection of central type lung cancer. Lung Cancer 2007; 58: 214–219.

78.

Zaric

Perin

Becker

. Combination of narrow band imaging (NBI) and autofluorescence imaging (AFI) videobronchoscopy in endoscopic assessment of lung cancer extension. Med Oncol 2012; 29: 1638–1642.

79.

Nguyen

Bashirzadeh

Hodge

. High specificity of combined narrow band imaging and autofluorescence mucosal assessment of patients with head and neck cancer. Head Neck 2013; 35: 619–625.

80.

Wasielica-Berger

Baniukiewicz

Wroblewski

. Magnification endoscopy and chromoendoscopy in evaluation of specialized intestinal metaplasia in Barrett’s esophagus. Dig Dis Sci 2011; 56: 1987–1995.

81.

Inoue

Kaga

Ikeda

. Magnification endoscopy in esophageal squamous cell carcinoma: a review of the intrapapillary capillary loop classification. Ann Gastroenterol 2015; 28: 41–48.

82.

Wanders

East

Uitentuis

. Diagnostic performance of narrowed spectrum endoscopy, autofluorescence imaging, and confocal laser endomicroscopy for optical diagnosis of colonic polyps: a meta-analysis. Lancet Oncol 2013; 14: 1337–1347.

83.

Wang

Lin

Zheng

. Comparative study of the endoscope-based bevelled and volume fiber-optic Raman probes for optical diagnosis of gastric dysplasia in vivo at endoscopy. Anal Bioanal Chem 2015; 407: 8303–8310.

84.

Stoler

Vichnin

Ferenczy

. The accuracy of colposcopic biopsy: analyses from the placebo arm of the Gardasil clinical trials. Int J Cancer 2011; 128: 1354–1362.

85.

Sreedhar

Varma

Nguyen

. High-definition Fourier transform infrared (FT-IR) spectroscopic imaging of human tissue sections towards improving pathology. J Vis Exp 2015; 95: 52332.

86.

Kleiner

Brunt

Van Natta

. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005; 41: 1313–1321.

87.

Fenger

Bak

Kronborg

. Observer reproducibility in grading dysplasia in colorectal adenomas: comparison between two different grading systems. J Clin Pathol 1990; 43: 320–324.

88.

McCluggage

Walsh

Thornton

. Inter- and intra-observer variation in the histopathological reporting of cervical squamous intraepithelial lesions using a modified Bethesda grading system. BJOG 1998; 105: 206–210.

89.

Landis

Koch

GG.

The measurement of observer agreement for categorical data. Biometrics 1977; 33: 159–174.

90.

Kendall

Stone

Shepherd

. Raman spectroscopy, a potential tool for the objective identification and classification of neoplasia in Barrett’s oesophagus. J Pathol 2003; 200: 602–609.

91.

Karakas

Sircan-Kucuksayan

Elpek

. Investigating viability of intestine using spectroscopy: a pilot study. J Surg Res 2014; 191: 91–98.

92.

Pennathur

Gibson

Jobe

. Oesophageal carcinoma. Lancet 2013; 381: 400–412.

93.

Ishigaki

Maeda

Taketani

. Diagnosis of early-stage esophageal cancer by Raman spectroscopy and chemometric techniques. Analyst 2016; 141: 1027–1033.

94.

Shaheen

Weinberg

Denberg

. Upper endoscopy for gastroesophageal reflux disease: best practice advice from the clinical guidelines committee of the American College of Physicians. Ann Intern Med 2012; 157: 808–816.

95.

Kendall

Day

Hutchings

. Evaluation of Raman probe for oesophageal cancer diagnostics. Analyst 2010; 135: 3038–3041.

96.

Wani

Falk

Hall

. Patients with nondysplastic Barrett’s esophagus have low risks for developing dysplasia or esophageal adenocarcinoma. Clin Gastroenterol Hepatol 2011; 9: 220–227; quiz e26.

97.

Hvid-Jensen

Pedersen

Drewes

. Incidence of adenocarcinoma among patients with Barrett’s esophagus. N Engl J Med 2011; 365: 1375–1383.

98.

Almond

Barr

Advanced endoscopic imaging in Barrett’s oesophagus. Int J Surg 2012; 10: 236–241.

99.

Almond

Hutchings

Lloyd

. Endoscopic Raman spectroscopy enables objective diagnosis of dysplasia in Barrett’s esophagus. Gastrointest Endosc 2014; 79: 37–45.

100.

Bergholt

Zheng

. Fiberoptic confocal Raman spectroscopy for real-time in vivo diagnosis of dysplasia in Barrett’s esophagus. Gastroenterology 2014; 146: 27–32.

101.

Ferlay

Soerjomataram

Dikshit

. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer 2015; 136: E359–E386.

102.

Huang

Zheng

. Consideration of tumor size improves the accuracy of TNM predictions in patients with gastric cancer after curative gastrectomy. Surg Oncol 2013; 22: 167–171.

103.

Correa

Human gastric carcinogenesis: a multistep and multifactorial process – First American Cancer Society Award Lecture on Cancer Epidemiology and Prevention. Cancer Res 1992; 52: 6735–6740.

104.

Teh

Zheng

. Near-infrared Raman spectroscopy for optical diagnosis in the stomach: identification of Helicobacter pylori infection and intestinal metaplasia. Int J Cancer 2010; 126: 1920–1927.

105.

Huang

Bergholt

Zheng

. In vivo early diagnosis of gastric dysplasia using narrow-band image-guided Raman endoscopy. J Biomed Opt 2010; 15: 037017.

106.

Wei

Guo

. Oxygenated hemoglobin diffuse reflectance ratio for in vitro detection of human gastric pre-cancer. Laser Phys 2010; 20: 1667–1672.

107.

Bergholt

Zheng

Lin

. Raman endoscopy for in vivo differentiation between benign and malignant ulcers in the stomach. Analyst 2010; 135: 3162–3168.

108.

Dong

Sun

Chao

. Evaluation of FTIR spectroscopy as diagnostic tool for colorectal cancer using spectral analysis. Spectrochim Acta A Mol Biomol Spectrosc 2014; 122: 288–294.

109.

Shao

Zheng

Huang

Polarized near-infrared autofluorescence imaging combined with near-infrared diffuse reflectance imaging for improving colonic cancer detection. Optics Express 2010; 18: 24293–24300.

110.

Lee

Hong

Kim

. Korean guidelines for colorectal cancer screening and polyp detection. Clin Endosc 2012; 45: 25–43.

111.

Dorundi

Bannerjea

Colorectal cancer: early diagnosis and predisposing causes. Surgery 2006; 24: 131–136.

112.

Rodriguez-Diaz

Huang

Cerda

. Endoscopic histological assessment of colonic polyps by using elastic scattering spectroscopy. Gastrointest Endosc 2015; 81: 539–547.

113.

Bergholt

Lin

Wang

. Simultaneous fingerprint and high-wave number fiber-optic Raman spectroscopy enhances real-time in vivo diagnosis of adenomatous polyps during colonoscopy. J Biophotonics 2016; 9: 333–342.

114.

Kuiper

Alderlieste

Tytgat

. Automatic optical diagnosis of small colorectal lesions by laser-induced autofluorescence. Endoscopy 2015; 47: 56–62.

115.

Rath

Tontini

Vieth

. In vivo real-time assessment of colorectal polyp histology using an optical biopsy forceps system based on laser-induced fluorescence spectroscopy. Endoscopy 2016; 48: 557–562.

116.

Lim

Bajbouj

von Delius

. Fluorescein-enhanced autofluorescence imaging for accurate differentiation of neoplastic from non-neoplastic colorectal polyps: a feasibility study. Endoscopy 2011; 43: 419–424.

117.

Sturgeon

Duffy

Hofmann

. National Academy of Clinical Biochemistry Laboratory Medicine Practice Guidelines for use of tumor markers in liver, bladder, cervical, and gastric cancers. Clin Chem 2010; 56: e1–e48.

118.

Lam

SC.

Discovery of diagnostic serum biomarkers of gastric cancer using proteomics. Proteomics Clin Appl 2008; 2: 219–228.

119.

Pierrakos

Vincent

J-L.

Sepsis biomarkers: a review. Crit Care 2010; 14: R15.

120.

Zeng

. Characterization and noninvasive diagnosis of bladder cancer with serum surface enhanced Raman spectroscopy and genetic algorithms. Sci Rep 2015; 5: 9582.

121.

Lualdi

Colombo

Leo

. Natural fluorescence spectroscopy of human blood plasma in the diagnosis of colorectal cancer: feasibility study and preliminary results. Tumori 2007; 93: 567–571.

122.

Barlev

Zelig

Bar

. A novel method for screening colorectal cancer by infrared spectroscopy of peripheral blood mononuclear cells and plasma. J Gastroenterol 2016; 51: 214–221.

123.

Feng

Huang

. Label-free detection of blood plasma using silver nanoparticle based surface-enhanced Raman spectroscopy for esophageal cancer screening. J Biomed Nanotechnol 2014; 10: 478–484.

124.

Feng

Chen

Lin

. Gastric cancer detection based on blood plasma surface-enhanced Raman spectroscopy excited by polarized laser light. Biosens Bioelectron 2011; 26: 3167–3174.

125.

Lin

Feng

Pan

. Colorectal cancer detection by gold nanoparticle based surface-enhanced Raman spectroscopy of blood serum and statistical analysis. Optics Express 2011; 19: 13565–13577.

126.

Ito

Inoue

Hasegawa

. Use of surface-enhanced Raman scattering for detection of cancer-related serum-constituents in gastrointestinal cancer patients. Nanomedicine 2014; 10: 599–608.

127.

Feng

Wang

Tai

. Label-free surface-enhanced Raman spectroscopy for detection of colorectal cancer and precursor lesions using blood plasma. Biomed Opt Express 2015; 6: 3494–3502.

128.

Lin

Chen

Feng

. A novel blood plasma analysis technique combining membrane electrophoresis with silver nanoparticle-based SERS spectroscopy for potential applications in noninvasive cancer detection. Nanomedicine 2011; 7: 655–663.

129.

Wang

Lin

. Label-free detection of serum proteins using surface-enhanced Raman spectroscopy for colorectal cancer screening. J Biomed Opt 2014; 19: 087003.

130.

Chan

YM.

Circulating tumour-derived nucleic acids in cancer patients: potential applications as tumour markers. Br J Cancer 2007; 96: 681–685.

131.

Schwarzenbach

Circulating nucleic acids as biomarkers in breast cancer. Breast Cancer Res 2013; 15: 211.

132.

Zheng

Shi

. A new enzyme-free quadratic SERS signal amplification approach for circulating microRNA detection in human serum. Chem Commun 2015; 51: 16271–16274.

133.

Driskell

Primera-Pedrozo

Dluhy

. Quantitative surface-enhanced Raman spectroscopy based analysis of microRNA mixtures. Appl Spectrosc 2009; 63: 1107–1114.

134.

Barhoumi

Zhang

Tam

. Surface-enhanced Raman spectroscopy of DNA. J Am Chem Soc 2008; 130: 5523–5529.

135.

Chen

Zheng

. Discrimination of gastric cancer from normal by serum RNA based on surface-enhanced Raman spectroscopy (SERS) and multivariate analysis. Med Phys 2012; 39: 5664–5668.

136.

Chen

Feng

. Label-free serum ribonucleic acid analysis for colorectal cancer detection by surface-enhanced Raman spectroscopy and multivariate analysis. J Biomed Opt 2012; 17: 067003.

137.

Wan

Zhang

KH.

Noninvasive detection of gastric cancer. Tumour Biol 2016; 37: 11633–11643.

138.

Sun

Zhang

Gao

. A rapid and simple method for efficient capture and accurate discrimination of circulating tumor cells using aptamer conjugated magnetic beads and surface-enhanced Raman scattering imaging. Anal Bioanal Chem 2015; 407: 8883–8892.

139.

Shi

Paproski

Moore

. Detection of circulating tumor cells using targeted surface-enhanced Raman scattering nanoparticles and magnetic enrichment. J Biomed Opt 2014; 19: 056014.

140.

Zhou

Lin

. The intrinsic fluorescence spectrum of dilute gastric juice as a novel diagnostic tool for gastric cancer. J Dig Dis 2011; 12: 279–285.

141.

Deng

Zhou

Lin

. A novel approach for the detection of early gastric cancer: fluorescence spectroscopy of gastric juice. J Dig Dis 2013; 14: 299–304.

142.

Deng

Lin

Zhou

. Three aromatic amino acids in gastric juice as potential biomarkers for gastric malignancies. Anal Chim Acta 2011; 694: 100–107.

143.

Padalkar

Pleshko

Wavelength-dependent penetration depth of near infrared radiation into cartilage. Analyst 2015; 140: 2093–2100.

144.

Kast

Tucker

Killian

. Emerging technology: applications of Raman spectroscopy for prostate cancer. Cancer Metastasis Rev 2014; 33: 673–693.

145.

Stephen

Jayanthi

Unni

. Diagnostic accuracy of diffuse reflectance imaging for early detection of pre-malignant and malignant changes in the oral cavity: a feasibility study. BMC Cancer 2013; 13: 278.

146.

Shaikh

Prabitha

Dora

. A comparative evaluation of diffuse reflectance and Raman spectroscopy in the detection of cervical cancer. J Biophotonics 2017; 10: 242–252.

147.

Odagaki

Sakamoto

Sekiguchi

. What is the accuracy of autofluorescence imaging in identifying non-polypoid colorectal neoplastic lesions when reviewed by trainees? A pilot study. Dig Endosc 2013; 25: 428–433.

148.

Kuiper

van den Broek

Naber

. Endoscopic trimodal imaging detects colonic neoplasia as well as standard video endoscopy. Gastroenterology 2011; 140: 1887–1894.

149.

Kaise

Advanced endoscopic imaging for early gastric cancer. Best Pract Res Clin Gastroenterol 2015; 29: 575–587.

150.

Uedo

Iishi

Tatsuta

. A novel videoendoscopy system by using autofluorescence and reflectance imaging for diagnosis of esophagogastric cancers. Gastrointest Endosc 2005; 62: 521–528.

151.

Bergholt

Zheng

Lin

. Combining near-infrared-excited autofluorescence and Raman spectroscopy improves in vivo diagnosis of gastric cancer. Biosens Bioelectron 2011; 26: 4104–4110.

152.

Liu

Nie

Lin

. Pattern recognition of multiple excitation autofluorescence spectra for colon tissue classification. Photodiagnosis Photodyn Ther 2013; 10: 111–119.

153.

Sainani

KL.

Introduction to principal components analysis. PM R 2014; 6: 275–278.

154.

Jolliffe

Cadima

Principal component analysis: a review and recent developments. Philos Trans A Math Phys Eng Sci 2016; 374: 20150202.

155.

Zhang

Z-Y

Zha

H-Y.

Principal manifolds and nonlinear dimensionality reduction via tangent space alignment. J Shanghai Univ 2004; 8: 406–424.

156.

Bergholt

Zheng

Lin

. In vivo diagnosis of gastric cancer using Raman endoscopy and ant colony optimization techniques. Int J Cancer 2011; 128: 2673–2680.

157.

Hou

Cai

. Terahertz spectroscopic investigation of human gastric normal and tumor tissues. Phys Med Biol 2014; 59: 5423–5440.

158.

Lee

Kim

Lee

. Improving the classification accuracy for IR spectroscopic diagnosis of stomach and colon malignancy using non-linear spectral feature extraction methods. Analyst 2013; 138: 4076–4082.

Biomedical optical spectroscopy for the early diagnosis of gastrointestinal neoplasms

Abstract

Keywords

Introduction

Techniques of OS for diagnosis

Infrared spectroscopy

Raman spectroscopy

Diffuse reflection spectroscopy

Autofluorescence

In vivo OS based on endoscopy

OS diagnosis based on tissues

Oesophageal cancer

GC

CRC

OS diagnosis based on biofluids

Blood

Gastric juice

Limitations and prospects

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

References