Clinical biomarkers for cancer recognition and prevention: A novel approach with optical measurements

Abstract

Cancer is the most important cause of death worldwide, and early cancer detection is the most fundamental factor for efficacy of treatment, prognosis, and increasing survival rate. Over the years great effort has been devoted to discovering and testing new biomarkers that can improve its diagnosis, especially at an early stage. Here we report the potential usefulness of new, easily applicable, non-invasive and relatively low-cost clinical biomarkers, based on abnormalities of oral mucosa spectral reflectance and fractal geometry of the vascular networks in several different tissues, for identification of hereditary non-polyposis colorectal cancer carriers as well for detection of other tumors, even at an early stage. In the near future the methodology/technology of these procedures should be improved, thus making possible their applicability worldwide as screening tools for early recognition and prevention of cancer.

Keywords

Clinical biomarkers cancer oral mucosal network abnormalities fractal analysis oral mucosal spectral reflectance

1. Introduction

Cancer is a class of diseases characterized by the uncontrolled growth and spread of abnormal cells. The basic cellular mechanism that permits cancer development consists in alterations of multiple genes and of their expression that confer a survival advantage and proliferative potential to somatic or germinal cells. Such alterations primarily affect three main classes of genes: (proto) oncogenes, tumor suppressor, and DNA repair genes. These alterations (mutated genes) collectively contribute to the development of cancer genotypes and phenotypes that show high resilience to the death mechanism(s) intrinsic in cells (e.g., apoptosis), and miss-regulated cell proliferation [94, 170].

As is well-known, the impact of this disease is remarkable. It has been estimated that 14.1 million new cancer cases and 8.2 million cancer-related deaths occurred in 2012 worldwide [68], with a huge increase foreseen in the next few years. Projections based on the GLOBOCAN 2012 forecast an increase to 19.3 million new cancer cases per year by 2025, due to growth and aging of the global population. In the USA alone 1.5 million of people were diagnosed with cancer in 2010, causing in the same year around 570,000 cancer deaths [1]. Colorectal cancer (CRC) is the third most common form of cancer (after Lung and Breast) and the fourth leading cause of death among cancers worldwide, after lung, stomach, and liver cancers [34]. CRC gives rise to almost 1.4 million new cases each year [34, 69]. Advances in care and treatment have increased the survival rate for some cancers, but the most important factor for prognosis is early detection, reflected in a 30% reduction in colorectal cancer mortality in individuals enrolled in regular screening programs [105, 165].

The importance of detecting cancer at an early stage is also suggested by the circumstance that 90% of patients diagnosed with stage I CRC survive 5 years, while patients diagnosed with stage II or III cancer have a 5-year survival expectation from 60% up to 70%, which falls to 10% for stage IV [69, 85]. Similar findings have been reported for other cancers, although the exact benefits of screening are still questioned.

In this perspective, a key point of discussion is that despite the widespread availability of multiple options for CRC screening, participation in screening programs remains suboptimal: Over one-third of adults in the US population have not been screened, far below the goal of 80% set by the Centers For Disease Control [97]. In fact, the currently available tests have limitations (e.g., fecal sampling, bowel preparation, procedural risks) that present significant hurdles to participation. On the other hand, blood-based screening, which is usually well accepted, is not well-established for CRC and seems too expensive in comparison with procedures such as fecal occult blood test, although some interesting clues on blood-based assay for CRC screening are emerging (e.g., Septin9 marker) [42].

For all these reasons further efforts are needed to discover and test new biomarkers that can improve the diagnosis, prognosis, and treatment of different types of cancer [20, 40, 95, 134, 139]. Biomarker utilization can be fruitful for early cancer detection and for the deployment of prevention campaigns in an attempt to increase the average survival expectations.

2. Biomarkers for cancer detection

According to the National Cancer Institute (NCI), part of the National Institutes of Health’s (NIH), a biomarker is “a biological molecule found in blood, other body fluids, or tissues that is a sign of a normal or abnormal process, or of a condition or disease. A biomarker may be used to see how well the body responds to a treatment for a disease or condition. Also called molecular marker and signature molecule.” The World Health Organization (WHO) defines a biomarker as “any substance, structure or process that can be measured in the body or its products and influences or predicts the incidence of outcome or disease” [167]. Others accept a more general biomarker definition, which can be any measurable phenotypic parameter that characterizes an organism’s state of health or disease, or a response to a particular therapeutic intervention [83, 99, 125, 142].

Cancer biomarker investigation dates back to 1847, with the casual discovery in the urine of a multiple myeloma patient [18, 100] of a protein that was subsequently recognized as a tumor-produced free antibody immunoglobulin light chain (aka monoclonal globulin protein). The “Bence-Jones protein” was present in the urine in large quantities and could be revealed by simple heat denaturation. Recently, the Bence-Jones protein was also found in serum [153], and in 1998 a practical immunodiagnostic test was approved by the Food and Drug Administration (FDA) [145]. Nowadays, hormones produced by tumors are also utilized for their detection facility [37]. For instance, Adrenocorticotropic hormone (ACTH), and chorionic gonadotropin (hCG) are elevated in particular cancer types, but they do not achieve the high tumor specificity featured by the Bence-Jones proteins [11, 145].

A systematic approach to finding new protein biomarkers of various cancer types is functional genomics or proteomics, as discussed in [156]. Proteomics is the study of the complete protein complement, or the proteome of the cell. In contrast to the genome, the proteome is dynamic and is in constant flux due to a combination of factors. These include differential splicing of the respective mRNAs, post-translational modifications, and temporal and functional regulation of gene expression [63, 155]. Proteomic technologies allow for identification of the protein changes caused by the disease process with fair to modest accuracy. At the protein level, distinct changes occur during the transformation of a healthy cell into a neoplastic cell, including altered expression, differential protein modification, changes in specific activity, and aberrant localization, all of which may affect cellular function. Identifying and understanding these changes is the underlying theme in cancer proteomics [63, 155].

An important contribution to cancer detection and comprehension has emerged from genetic studies, habitually termed genomics. In fact, it is usually believed that underlying mechanisms at the cellular level are associated with the cancer and promote its development. From a biological standpoint there is a close link between proteomic and genomic cancer biomarkers, since proteins are produced in cells by synthesis driven by the DNA. If an abnormal level of a protein efficiently marks the cancer disease, this protein is presumably produced by some (mutant) cancer cells, being thus associated with one or more gene modifications. Initially, the field of genomics was motivated by the rush to create the first complete map of the human genome. The sequencing of the human genome, along with the development of high throughput technologies, moved the focus to gene expression profiling and the quest to identify genetic signatures of diseases [74]. Many different gene expression profiling technologies are currently in use, including cDNA microarrays, oligonucleotide arrays, and serial analysis of genes [87]. Although molecular mechanisms of tumor genesis are unclear at present, carcinogenic agents could contribute to fields of genomic instability localized to specific areas of the breast. Understanding the functional importance of genomic instability in breast cancer is important for improvement of diagnostic and treatment strategies [61, 143] gave a self-contained account of the role of genomics in the study of breast cancer and its specific biomarkers.

Cancer is also driven by epigenetic changes such as DNA methylation and altered patterns of histone modifications [15, 20, 86, 92, 101, 129]. These processes lead to alterations in the chromatin condensation status there by regulating the activation and deactivation of certain sets of specific genes. Epigenetic embraces all stable changes of phenotypic traits that are not coded in the DNA sequence [16, 66, 140]. Interestingly, epigenetic features affecting gene expression and chromatin organization very often constitute heritable characteristics. Epigenetic inheritance is a critical mechanism that allows a remarkably stable propagation of gene activity states over many cell generations [148]. Recently, the field of epigenetic has received remarkable attention, owing to the awareness that epigenetic inheritance is essential for critical cellular processes such as gene transcription, differentiation, and protection against viral genomes. In a broader sense, epigenetics can be considered as an interface between genotype and phenotype. Epigenetic mechanisms play a critical role in normal physiological responses to environmental stimuli allowing the establishment of appropriate gene expression patterns.

Recent investigations conjecture that exosomes could play an active role in cancer growth and that they could serve as reliable cancer biomarkers [93, 157, 139]. Exosomes are one of the various types of cell-derived vesicles, having a mean diameter between 30 nm and 100 nm, and transporting molecular contents of the cells from which they originate. This last circumstance is the main motivation of the conjectured ability of these entities to map cancer activity and presence. Exosome investigations can be considered as something midway in between genomics and proteomics, and the principal obstacle hindering the utilization of these structures as cancer marker is the complexity of their detection and the related molecular profiling [91].

In recent years, genomics and proteomics have generated many cancer biomarker candidates [135]. Unfortunately, only a few of these markers exhibits superior prognostic or diagnostic ability, and even fewer have been validated and approved. Several factors might play a role in the slow progression of identifying cancer biomarkers. On the one hand, available proteomics technology has limited power to detect low-abundance cancer biomarkers by distinguishing them in the mess of highly abundant plasma proteins, and many of the best markers may hence be missed until recognition technology improves. On the other hand, the ability to verify and validate existing candidate markers (through rigorous testing in large sample sets from many diseases) is limited, and in the past 20 years, fewer than 12 biomarkers have been approved by the US FDA for monitoring response, surveillance or recurrence of cancer [7, 118]. Thus it is even possible that optimal biomarkers have already been proposed but not still validated [81].

A number of books and review papers discuss promising cancer markers, but they often focus on proven or well-developed markers, or specific disease states. The case of an overproduced tumor-specific protein that reliably marks the cancer and can be easily detected is the exception rather than the rule of biomarker research. Hence, in the nearly 160 years since Bence-Jones’ discovery, about ten proteins have been approved by FDA as cancer diagnostic tests, and most of them show poor sensitivity and specificity for cancer.

Figure 1.

The process of carcinogenesis, showing opportunities of identifying biomarkers.

There seems to be a growing consensus that panels of markers may be able to supply the specificity and sensitivity that individual markers lack [46, 69]. A panel combining four known biomarkers (leptin, prolactin, osteopontin, insulin-like growth factor II), none of which used alone could distinguish patients from the controls, achieved a sensitivity and specificity of 95% for the diagnosis of ovarian cancer [127]. In this case a combination of known proteins in a novel panel provided a significant advance. Shen et al. in 2005 [151] identified 299 proteins in tissue culture and then used ELISA to test 13 of the most interesting in serum. They reported that CD98, fascin, the secreted chain of the polymeric immunoglobulin receptor and 14-3-3 eta provide greater sensitivity when used together as a panel than any of the markers used alone.

García-Bilbao et al. have achieved a tentative biomarker consisting in a panel of seven genes capable of correctly classifying cancerous and non-cancerous colon samples [69].

Campbell et al. [32] investigated the ability of eight independent parameters to predict breast cancer morbidity. These parameters were age, presence of lump symptoms, lumpiness, pain, and nipple discharge, signs of cancer, HRT use, and family history. These eight observables gave a 30% improvement in prediction accuracy of breast cancer. As an instance, patients with a breast lump were significantly more likely to have breast cancer than patients without a lump. According to Pedersen and Wandall [134] circulating antibodies produced by the patient’s own immune system after exposure to cancer proteins are emerging as potential biomarkers available for early detection of the disease. An advantage of using autoantibodies as biomarkers is their production in large quantities irrespective of the presence of a relatively small amount of corresponding antigen. Autoantibodies are also expected to have persistent concentrations and long half-lives due to limited proteolysis and clearance.

The discussed background of cancer biomarkers is well-represented by the scheme shown in Fig. 1, which is an extension of a similar picture pointed out by Bhatt et al. [20]. The most important aspects are the strict logical links between genetic mutations, altered gene expressions, modified protein status, and metabolites. At the first level we find genetic biomarkers associated to genetic mutations, while at level two there are epigenetic markers dependent on changed gene transcription. Both these phenomena can give rise to altered protein status (level three), which is followed by altered metabolites in level four. Antibodies and protein biomarkers are located at these two levels.

Nonetheless, additional opportunities of cancer detection have been often neglected in previous studies. These supplementary chances of cancer recognition are related to the possible modifications of tissue structure and its physical properties that could be due to ineffective gene expression, as well as alterations of protein status and metabolites. An important possible relationship between tissues’ physical properties and cancer occurrence can be found at the genotype level. Today various types of hereditary cancers are known, such as hereditary nonpolyposis colorectal cancer (HNPCC) or Lynch Syndrome. The specific gene combinations that are responsible for cancer occurrence in such subjects will also have an effect, although as yet unknown, on the mesenchyme formation and the physical properties of tissues. Therefore, one can expect the presence of some odd or unexpected characteristics of tissues at the phenotypic level, such as the alteration of some physical parameters. The observation or measurement of such changed physical characteristics of tissues might be employed as a marker of cancer or an oncogenesis-related factor. While great effort is being devoted to the emerging landscape of oncogenic signatures across different human cancers [43], the synthesis of such genetic and epigenetic panorama might be summarized in a few physical parameters observable in many tissues of a potential patient with a simple and inexpensive procedure. This reasoning could be weakened by the existence of tumors with no genetic or epigenetic cause, thus unable to affect the structure and physical properties of tissues. And this seems to be the case with most of those malignant neoplasms, which do not have any hereditary component and so are named “sporadic cancers” [107]. However, recent investigations have shown that most sporadic cancers have a deficiency in DNA repair due to mutation in a DNA repair gene, or a deficiency in DNA repair due to epigenetic alterations that reduce or silence DNA repair gene expression [4, 79, 126, 136, 151].

The occurrence of one or both of these deficiencies (genetic and epigenetic) might be in some unknown way be related to specific gene combinations of the affected individuals gene combinations that could influence the physical properties of many biological tissues.

To sum up, cancer clinical biomarkers can be found as changed physical properties of specific tissues as a potential effect of modifications at genetic or epigenetic levels. This new approach to cancer detection has been analyzed in various works [21, 22, 23, 24, 25, 26, 52, 53]. Usually, the physical parameters that are monitored for the purpose of tumor detection are derived from optical measurements carried out over the skin or the mucosa of the patients, at selected anatomical sites. These optical measurements fall into one of the following two main classes: Spectroscopic analysis, and imaging observations, although many experiments are being performed which include the application of hyperspectral imaging devices [9, 10, 113, 115, 119, 120, 121]. In this last case the instrument is able to gather a number of monochromatic images of adjacent spectral bands recording into a single dataset both spatial and spectral features of the observed tissue. Sometimes these two applications significantly overlap each other. Spectroscopic analysis also includes color measurement, a diagnostic tool that is also widely adopted to monitor patients by pulse-oximetry, to evaluate the state of newborns (e.g., the APGAR score) as well as the occurrence of inflammatory states and erythema [59]. There is no precise limit differentiating spectroscopic from colorimetric measurements, since color is strictly dependent on the lower resolution reflectance or transmittance spectrum of the sample.

Various researches [3, 77, 119, 121] have pointed out the ability of Diffuse Reflectance Spectroscopy, often termed as DRS, to measure in-vivo many metabolic and physiologic parameters relevant to define the patient healthiness and possibly to detect cancer structures. The typical DRS instrument is made up of a light source that illuminates the studied anatomical region and a spectrophotometer that measures the radiation retro-diffused by the skin or the mucosa. It is worth noting that the observed radiation is partially originated by the skin or mucosal surface due to the steep change of refraction index passing from the air into the tissue. The other and most important contribution is originated under the tissue surface by back scattering and reflection from underlying fat and bones. Often, but not necessarily, the electromagnetic radiation (i.e. the light) is guided from the source to the observed site, and from this back to the spectrophotometer by a miniaturized optical fiber bundle, thus allowing an easy development of small probes that are employed for noninvasive in-vivo measurements. The DRS instrument collects reflectance spectra of the observed tissue that usually cover the visible and near IR spectral regions [121].

Obviously, the DRS can perform accurate assessment of the total hemoglobin content and hemoglobin saturation. The ability of this method to detect cancer lesions is claimed to be connected to the reduced hemoglobin saturation recorded in malignant tissues [88, 89], and possibly to their modified vascularity mapped by the total hemoglobin content [158, 159].

Figure 2.

Transmission spectra of normal and malignant ovarian tissues observed in the IR spectral region (adapted from Mehrotra in 2010) [121].

High resolution spectral measurements are usually performed to recognize the spectral signature of selected molecules or substances that are essential to defining the state of a patient. This is the case of infrared (IR) reflectometry whose utilization for assessing a patient’s glycemia has been attempted repeatedly in recent years [119].

Another instance of the detection power of material spectral signatures is represented by the application of infrared spectroscopy from 2 $\mu$ m up to 25 $\mu$ m of electromagnetic wavelength. Spectral transmittance of biological samples in this spectral range seems to be able to differentiate cancerous tissues from normal ones, as reported in a number of research studies that have examined several types of cancers [3, 44, 121, 144, 152]. Usually, the spectral transmittance of malignant tissues is compared to those measured over normal tissues of the same anatomic district, revealing impressive differences. A clear example of such spectroscopic differences is shown in Fig. 2.

From a physical standpoint the absorbance (transmittance) of tissues at IR wavelengths is mainly related to the vibrational and rotational absorption spectra of molecules embedded in the observed samples (e.g., N-H and C=O stretches). Therefore, the IR transmission of the samples indirectly maps the biological composition of the observed tissues in terms of proteins, peptides, nucleic acids, and so forth. And this map would evidently be able to carefully highlight the presence of an abnormal (e.g., malignant) tissue. Usually, transmittance measurements are performed utilizing a Fourier Transform Spectrometer (FTS), and are referenced to as Fourier Transform Infrared Spectroscopy or FTIR [12, 67, 121, 144, 150, 152, 171]. An FTS instrument does not include any dispersive element such as gratings and prisms, and it is believed to provide the user with measurements of higher Signal-to-Noise Ratio (SNR) than traditional dispersive spectroscopy [12, 39, 67, 171, 138]. However, recent studies in this field have demonstrated that this expectation of a higher accuracy is not realistic and that FTIR measurements would convey a lower information content than dispersive ones [12]. Unfortunately, the above examples of FTIR application to cancer recognition necessarily implies an invasive methodology, reducing its appeal as a diagnostic tool for cancer prevention.

Since the 1990s several attempts have been made to extract physiological information regarding a specific disease from fluorescence spectroscopic data measured in vivo in the skin, head and neck, genito-urinary tract, gastro-intestinal tract, breast, and brain [39, 171]. Fluorescence is a physical phenomenon in which a material that absorbs electromagnetic radiation of short wavelengths (e.g., ultraviolet (UV) radiation at 330 nm of wavelength) emits in response a distribution of radiation of longer wavelengths (e.g., in the red part of the visible spectrum). The exact values of the excitation and re-emission wavelengths depends on the chemical composition of the considered material and its state and conditions (e.g., its temperature). Fluorescence measurements can be made at high temporal resolution, a case that is termed time-resolved fluorescence, or with long integration time that produces steady-state fluorescence data. This technique is believed to be a powerful diagnostic tool of high specificity and sensitivity also for recognizing neoplastic growth in many epithelial tissues [9, 138]. Let us note that fluorescence instruments usually employ a short wavelength light source in order to excite the observed sample. Such a light source is frequently constituted by an UV to NIR laser, hence the measurement technique is called Laser-Induced Fluorescence (LIF). Despite these expectations and results, the best accomplishments involving LIF techniques are related to cancer imaging and management in the infrared spectral range, a domain known as NIRF (Near-infrared fluorescence). One of the best applications relying on this technology is the real-time imaging of even single cancer cells during surgeries, a methodology that could considerably improves surgery results and prognosis [108, 124]. From a biological standpoint important progress has been made with the development of fluorescent nanoprobes, resulting from the integration of nanotechnology with molecular biology. The availability of suitable nanoprobes can improve detection sensitivity and specificity in molecular imaging [124]. “Functional nanomaterial-based molecular probes could target tumors either through the Enhanced Permeability and Retention (EPR) effect of the tumor microvasculature or by specific binding with tumor-associated biomarkers, such as tumor cell receptors, tumor extracellular matrix and enzymes.” [84].

Often spectroscopic measurements (of reflectance, transmittance, and fluorescence) concern a single point or a small anatomical region and do not have imaging ability. An additional characteristic of spectroscopic observations is that they aim at deducing the surface spectral reflectance, or the volume spectral transmittance or fluorescence of a tissue sample. Colorimetric observations can be made even with digital still cameras [102, 103] or similar equipment, and customarily incorporates full imaging capability.

The utilization of digital images of the skin or other tissues can also be aimed at investigating tissue structure and morphology, a case in which the apparent color is a less important feature. This kind of approach can be assisted by original geometrical concepts, such as fractal analysis, and can produce important results in identifying new biomarkers useful for cancer detection [21, 22, 23, 24, 25, 26, 53].

3. Morphological approaches

Morphological approaches for the characterization of cancer cell and tissues appearance can be based on quantitative image analysis. These promising techniques could be used as prognostic biomarkers to detect cancer or to follow-up anti-cancer therapies. In effect, quantitative morphometric analysis of shapes of cells and organelles or tissue textures are opening up new areas of interest. Traditional imaging methods can be performed at gross level, by optical or electron microscopy, or directly in the human subject by micro-endoscopy [115, 128, 162], computed tomography or nuclear magnetic resonance. Recently, new technical approaches are gaining visibility such as the Diffuse Optical Tomography (DOT) [17, 168], the Optical Coherence Tomography (OCT) [5, 75, 90, 132, 160, 158, 159], and the multispectral DOT [101]. DOT and OCT measurements collect three dimensional (3d) data volumes that can be organized as a stratification of images at increasing tissue depths. The OCT technique employs a fiber-optic interferometer (e.g. a Mach Zender device) to select the geometrical depth of the tissue layer to be imaged. Instead DOT utilizes the natural propagation of electromagnetic radiation throughout the biological sample in order to retrieve structural 3d information of the observed sample. DOT instrument can also include a pulsed radiation source to improve translucent tissue visibility. Both DOT and OCT can be implemented by means of cheap miniaturized instruments and seems provide clinicians with improved cancer diagnostic ability [30, 48, 70, 115, 159, 168]. Moreover, by virtue of their small size these instruments can be easily adopted for in vivo noninvasive examinations [72]. Multispectral DOT [78, 168] is realized by using light sources of controlled spectrum (color) in order to retrieve multispectral information from the observed tissue. This technique adds spectral information to the diagnostic ability of the basic imaging technique, and retrieve a 4d multispectral image data volume.

Several works [6, 36, 45, 72, 115, 154] utilize the confocal microscopy to recognize neoplastic tissues into the oral cavity. This technique is able to reconstruct 3d panchromatic images of the observed tissue with a spatial resolution of a few microns. Adopting miniaturized optical fiber instruments, the confocal microscopy can even be adopted for in vivo analysis as in [6, 47, 115, 120, 137], and has been adopted to investigate non-malignant tissue disorders like psoriasis [73].

We should note that microscopy may be also performed at the molecular level using immunohistochemistry and immunofluorescence techniques, also providing new potential biomarkers for cancer diagnosis [2, 41, 60, 122, 133].

3.1 Cells

Apoptosis disturbance, presence of micronuclei, and shape changes of the nucleus, are some of the most interesting approaches.

3.1.1 Apoptosis

Cancer cells harbor several features that distinguish them from normal cells, one of which is the ability to evade programmed cell death or apoptosis. It is characterized morphologically by chromatin condensation and nucleus destruction, it may be identified step by step by immunohistochemistry [60, 62, 161]. In normal tissues, apoptosis plays a pivotal role in the maintenance of tissue homeostasis and the development of the immune system. Disturbance of this process in tumor cells results in the impaired removal of mutated cells and contributes to tumor progression. In addition, evasion of apoptosis enables malignant cells to escape from tumor immune surveillance and to acquire resistance to cancer therapy. In retrospective studies, the status of the apoptotic pathway in a tumor was shown to be of prognostic value, for example in colorectal cancer, using immunohistochemistry [29], although due to the complex biology of the pathway, a combination of several markers in order to establish an “apoptotic tumor profile” is needed to have a better predictive biomarker [19, 60].

3.1.2 Micronuclei

During cancer cell division, defects in the nuclear lamina, filaments that provide support and stability to the cell’s nucleus cause the nuclear envelope surrounding micronuclei (small, erratic extra nuclei containing whole chromosomes or their fragments) to catastrophically collapse, leading to the loss of basic nuclear functions such as replication, transcription, and DNA damage recognition and repair, and causing the acquisition of one of the hallmarks of cancer such as the development of genomic instability [80]. Immunohistochemistry and immunofluorescent microscopy have recently revealed that in solid tumors more than 60 percent of those micronuclei undergo this dysfunction [82]. The damaged DNA can then enter the next generation of daughter cells and undergo chromothripsis, a rearrangement of genomic information in one chromosome, which leads to massive DNA damage and the formation of tumors. Surely, in pulmonary adenocarcinoma, bone cancer, melanoma and other forms of lung cancer, micronuclei disruption plays an important role in the development of the neoplasia, giving us a new potent quantitative biomarker at the microscopic level.

Interestingly, the micronucleus assay, applied in different human cells, is one of the best validated methods for biomonitoring human populations to evaluate chromosomal damage as well as exposure to carcinogenic agents [27]. Furthermore, there is now solid evidence from molecular epidemiological studies evaluating large cohorts of disease-free subjects that an increased MN frequency in peripheral lymphocytes was associated with an increased risk of cancer at the population level, supporting the notion that this biomarker could be used as an early predictor [28, 29]. In addition, the simplicity and rapidity of the MN assay make it a valuable and useful screening tool for identifying underlying repair deficiencies caused by genetic defects in order to define the development of a predictive multivariate biomarker [55].

3.1.3 Nucleus shape change

Alterations in nuclear morphology are a hallmark of cancer cells. Nuclei of cancer cells divide uncontrollably and contain extra chromosomes. Compared with nuclei in non-cancerous cells, nuclei in cancer cells are larger, closer together and more irregular in morphology. The optical density also appears modified. The changes in cellular and nuclear morphology correlate with clinic-pathological parameters and with the level of expression of oncogenes, for example in melanoma in light microscopy, measuring shape, size and optical density of the nuclei [123], or in Mycosis Fungoides, making use of fractal analysis of the nucleus of lymphocytes present in the skin observed in Transmission Electron Microscopy, to distinguish early cancer status by dermatitis with 100% specificity, as shown in [24]. Moreover, the detection and classification of early cancer through nuclear morphometry may also be performed, for example in breast cancer, making use of the multispectral diffuse optical tomography based on light scattering [78], a novel technique for differentiating benign from malignant lesions.

3.2 Tissues

Because histopathological interpretation by the pathologist to distinguish cancer from non-malignant status is subjective and often lacking in sensitivity, many attempts have been made to establish objective methods to perform differential diagnosis by molecular biomarkers identified by immunohistochemistry or by morphometry (classical Euclidean morphometry parameters or, better, reaching higher sensitivity and specificity, by fractal indexes). For example, in oral cancer several molecular biomarkers have been identified by making use of immunohistochemistry or immunofluorescent techniques, for detection of epithelial dysplasia with an increased risk of developing cancer [2] or by fractal analysis of the epithelial-connective tissue interface in the basal cell carcinoma of the skin where a 100% specificity resulted in distinguishing low- and high-risk tumors in order to evaluate their ability to develop metastasis [25], as well in order to distinguish among the different classes of myelodysplastic syndromes, analyzing standard bone marrow Hematoxylin-Eosin slides [23], as shown also by us.

Regarding early detection of cancer or the follow-up of anti-cancer therapies, the morphometry of the microvessel network appears is also able to provide important biomarkers. Healthy vasculature exhibits a hierarchical branching structure in which, on average, vessel radius and length change systematically with branching order. In contrast, tumor vasculature exhibits less hierarchy and more variability in its branching patterns. Regarding changes in vasculature geometric complexity, fractal analysis is emerging as a reliable model for quantifying and describing morphological aspects of microvessels, with potential implications in future clinical and surgical applications, as well as providing indications for anti-angiogenic therapy in pituitary adenoma, glioblastoma multiforme and other malignant brain tumors [57, 58]. To identify a genetic status with high risk of developing neoplasia, fractal analysis of microvessel patterns also appeared successful. An example of this approach in our hands is the analysis of the microvessel network complexity from images of the oral vestibular mucosa, which reveals a network phase transition in HNPCC [53], as well in other genetic diseases [21, 26, 103], as shown by us. Potent mathematical models based on fractal analysis of tumor vasculature are growing in order to perform diagnosis and predict response to cancer therapies [147, 163].

3.3 Fractal geometry analysis

Theoretical fractal can be described mathematically as an iterative application of a mathematical equation, or a repeated geometric transformation of a structure. Theoretical fractals have that iterative algorithm repeated at infinite level, so producing a property known as self-similarity. A self-similar structure has the same geometrical properties when observed at different levels or scales.

To understand geometrical fractals, let us consider the example below (Fig. 3).

In the above figure we start our description with a simple equilateral triangle (Initiator) with side length, $r_{0}$ . The iterative algorithm that generates the set of von Koch curves (prefractal curve) consists of recursive reduction of the straight-line segments (or scales) by 1/3 exchanging repeatedly the middle third of each side of the initiator with two sides of a smaller equilateral triangle whose side is one-third the length of the previous side. The result after the first iteration (the stage of construction, $z=$ 1) is shown in Fig. 3b, after the second construction (the stage of construction, $z=$ 2), in Fig. 3c. For the Koch set, the length of a segment at the $z$ th stage of construction ( $r_{z}$ ) and the number of segments at the same stage ( $N_{z}$ ) are respectively [8]:

$\displaystyle r_{z}=\frac{r_{0}}{3^{z}}$ $\displaystyle N_{z}=3\times 4^{z}$

In other terms, the 3 generating elements in Fig. 1a can be subdivided into 12 generating elements to form Fig. 1b, and so on.

Natural objects also have the same self-similarity property even if, due to the discrete organization of matter (molecules, atoms, elementary particles), it is present only at certain levels of magnification or scales. Mandelbrot showed that a suitable way to define the fractal property of a natural object, such as a coast line or the texture of a biological tissue, is to extend the “self-similarity” concept utilized for defining with the abstract (theoretical) fractals. A natural self-similar structure has the same appearance and the same geometric properties when examined at various levels of magnification at least two log scales). In this sense this structure is said to be self-similar.

Unlike a smooth line, a fractal curve, which has a dimension between 1 and 2, is irregular or wrinkly. In effect, Mandelbrot’s concept of fractal geometry [164] has improved our ability to characterize natural structures and shapes, bringing essential contributions to the comprehension of complex systems and chaos.

We must consider, also, that the term “fractal” is associated with a curve, a surface or any other geometrical domain that subtends the mathematical property of this entity of having fractional dimensions. In effect, this concept characterizes the fractal geometry and distinguishes it from traditional Euclidean geometry, in which any object must have integer dimension. More irregular or wrinkly is the border of a natural object and higher is its fractal dimension, which may be described as the space-filling property of an object.

Figure 3.

The Koch (mathematical) fractal set. (a) Initiator, a simple equilateral triangle with side length $r_{0}$ ; (b) The first stage of construction ( $z=$ 1), below the drawing of the corresponding generating element; (c) The second stage of construction ( $z=$ 2), below the drawing of the corresponding generating element.

Figure 4.

An histologic image of the bone marrow in a healthy subject (top) and its skeletonized contours (bottom). Hematoxylin-Eosin, X50.

Natural structures having fractal appearance have been found in a variety of natural and biological entities: River deltas, trees, or, in the human body, cardiopulmonary structures, the ramifying tracheo-bronchial tree, the His-Purkinje network [71] the cardiac muscle bundles [14], as well as the placenta’s arterial. The meaning of these fractal structures in the human body is profound. The self-similar tracheo-bronchial tree provides an enormous surface area for exchange of gases at the vascular-alveolar interface, coupling pulmonary and cardiac functions. For vasculature, fractal branching provides a rich network for distribution of nutrients and oxygen, as well for collecting metabolic waste products [71]. A variety of other organ systems contain fractal structures that have particular functions, e.g., the fractal organization of connective tissue in the aortic leaflets relates to the efficient distribution of mechanical forces. It is worth noting that Mandelbrot also developed the mathematical branch known today as “fractal analysis”, which extends calculus to the concept of fractal geometrical dimension.

The first application of fractal geometry has been applied to histopathology, cytopathology, and other biomedical domains [109, 146]. Today, we know that is possible to make differential diagnosis between the early stage of tumors and flogosis [34] or among the different stages of Basal Cell Cancer [25], as well as to investigate the subtle alterations of the nuclear patterns in human breast tumors [111] or evaluate brain tumors [56]. Fractal analysis showed a high ability for automatic classification of cancer cells in urinary smears [65], as well as to identify prostate cancer cells [169]. Fractal analysis has been shown able to provide quantitative information on the link among molecular, cellular and histomorphological tissue changes in the development of tumors [110, 112].

In effect, to day, basic researchers and clinicians are aware of the remarkable importance of the fractal approaches in the morphological study of cells and tissues. The fractal dimension analysis may be seen as a useful objective tool in describing tissue complexity providing information for understanding pathological changes. The fractal dimension is a real-valued index, which scores the space-filling properties of an object [117]. The most commonly used methods for measuring the fractal dimension in the field of life sciences and biomedicine [21, 22, 23, 24, 25, 26] are the mass dimension, the fractal dimension of the minimum path and the box-counting dimension (perhaps the easiest to implement [64, 96].

Figure 5.

Log-log plot of the outlines of the bone marrow in acute leukemia. Outlines of the contours present in the histologic images of the bone marrow in an acute leukemia in a X40 microscopy image (left picture), and its log-log plot (right picture). The slope of the plot is the fractal dimension (D $=$ 1.92) of the pattern, which is linked to its space-filling property [23].

Figure 6.

The slope of the log log plots (fractal dimension, D or space-filling properties of an image) is lower in an healthy tissue (left, prostate, $D=$ 1.6) than in prostatic cancer (right, $D=$ 1.89).

Figure 7.

Fractal analysis of the angiogenetic pattern in stage IA pulmonary adenocarcinoma, histologic images. Fractal dimension of the microvascular network, its geometric complexity, is higher when the prognosis is worse for each growth pattern [114]. Group 1: 10 patients who developed distant metastasis within the first 36 months after surgery; Group 2: 10 patients free from disease after a follow up $>$ 96 months. Sgp $=$ Solid growth pattern; Agp $=$ Acinar growth pattern; BAgp $=$ Broncoalveolar growth pattern.

3.3.1 Box-counting technique

Performing the box-counting technique, target contours are extracted by means of a specific image- processing algorithm (Fig. 3), then this skeletonized image is superimposed over a grid (lattice) constituted by $L$ boxes covering the entire image. The box-counting method consists in counting the number of boxes $N_{b}$ overlapped to a single-pixel contour (black). By increasing the number of boxes held into the grid, it is possible to estimate the number of black boxes $N_{b}(L)$ as a function of the grid resolution $L$ , a quantity that is strictly linked to the fractal dimension of the target: A self-similar contour has a higher number $N_{b}$ of black boxes when magnified (if observed with a finer grid). The linearity of the log-log plot of $N_{b}(L)$ versus $1/L$ reveals the self-similarity of the pattern (the contour has fractal shape), and its slope is an unbiased estimation of its fractal dimension of the distribution [116]. Higher is the space-filling properties of the object and higher will result its fractal dimension.

Figure 4 shows a biologic image, the bone marrow in a healthy subject and its skeletonized contour. Figure 5 shows the Log-log plot of the contours revealed in a bone marrow image, at histological level, in an acute leukemia subject. Figure 6 shows the log log plot in a healthy tissue and in a malignant tissue (prostatic cancer).

3.3.2 Microvascular complexity in cancer

Performing fractal analysis of the angiogenetic pattern in stage IA pulmonary adenocarcinoma surgically resected link fractal dimension of the microvascular pattern to prognosis (Fig. 7). This tumor shows a lower fractal dimension in patients free from disease at 96 months than in patients with metastasis within 36 months; the fractal dimension of the microvascular network (its geometric complexity) is higher when the prognosis is worse [114].

The fractal dimension appeared able to accurately describe the altered microvascular complexity of the oral mucosa revealed in cancer-related diseases [53, 114] and in several genetic diseases [21, 26]. Figure 8 shows the outcome from a pilot study of a cancer-related genetic disease (HNPCC or Lynch syndrome); the oral mucosa has a higher geometric complexity of the microvascular network in comparison to healthy subjects of the same ages and sex [53]. This greater geometric complexity of the microvascular network in the cancer-related genetic disease could be linked to changes in the extracellular matrix (ECM); it is indicated by high values of the microvessel network geometric complexity observed by us in other pathological conditions where changes in the ECM have been described [50, 52, 53] and supported by the modifications of the oral mucosa light reflectance spectra observed in the HNPCC.

Figure 8.

Fractal analysis of the microvessel pattern. Oral mucosa is photographed (top, left), and a small portion of it is enlarged (top right, healthy subject). Vessels are extracted: microvessel pattern in a healthy subject (bottom, left), and in a HNPCC patient (bottom, right). Geometric complexity is increased in the HNPCC patients: fractal analysis permits exact quantification of the modified pattern in the cancer-related condition (table below, D calculated for low and high scale, left and right respectively [21, 53]).

Figure 9.

Comparison of oral mucosa light reflectance spectra between hereditary non-polyposis colorectal cancer (HNPCC) carriers and controls, both including and excluding blood vessels. Subtraction of blood vessels results in an increase in the reflectance values between 480 and 600 nm, without showing any difference between HNPCC and controls (after [35]).

Figure 10.

ROC curve. Y axis $=$ sensitivity; X axis $=$ 1 $-$ specificity, AUC $=$ Area Under curve, that is equal to 1 when sensitivity and specificity are equal to 1, or 100%, both. Fractal analysis of healthy tissue (prostate, $n=$ 31) vs. prostatic cancer ( $n=$ 15, Gleason index $=$ 7), Hematoxylin-Eosin stained samples. The discriminant power of fractal analysis is very high.

4. Oral mucosal color change and cancer detection

Early cancer diagnosis can be made using screening tests [31, 76]; therefore, identification of novel and reliable biomarkers for life-threatening cancers is a priority in clinical medicine [8, 95, 130]. In this context, color may be a useful parameter in clinical medicine [49]; in fact, the radiation emitted or reflected by an object is known to provide relevant information about its physical state and chemical composition [13, 38]. Some years ago we identified color changes in oral mucosa in HNPCC carriers [54]. In this regard, Extracellular Matrix (ECM), a major component of the cellular microenvironment, is an essential physical factor in establishing the resulting color of a tissue. Statistically significant lower oral mucosal spectral reflectance of HNPCC carriers has been observed in the 590700 nm wavelength range ( $p\leqslant$ 0.0004) [50, 51]. A reflectance cutoff value $\leqslant$ 47.9% at 700 nm of wavelength discriminated between HNPCC carriers and control subjects, with 100% sensitivity and specificity. This result is very interesting, considering the objective limitations of clinical and molecular criteria in identifying HNPCC families: Even today we are unable to identify gene mutations in more than onethird of cases. As we are not able to identify subjects at high risk in such families, extensive clinical-instrumental surveillance is recommended for all family members [54]. Our observation was not confirmed by the work performed by [35], whose main result is depicted in Fig. 9. As pointed out by [51], the apparently different results could be ascribed to the experimental methodology adopted, as well as the instruments utilized for a correct theoretical interpretation of the collected data. In particular Carrara et al. [35] stated that oxy-hemoglobin and deoxyhemoglobin have a negligible effect on the reflectance spectrum of biological tissues, an assumption that is only partially true [51] In addition they do not consider the possible role of the scattering effects of the hemoglobin, and of additional chromophores (e.g., carotenoids, fat, bilirubin). Finally, they used an image segmentation algorithm, which (excluding the darker image regions portraying vessels) could produce a bias in the spectral data.

Nonetheless it should be noted that although the difference is smaller than the one reported by [50], also in the paper by [35], the reflectance exhibited by HNPCC carriers is slightly lower with respect to controls, and such difference is systematic and increases with increasing the wavelength. In our opinion such difference is intriguing, and could reflect a very different optical behavior.

A more recent study [102] found a similar connection between various types of cancer and the color of the gingival and vestibular oral mucosa [141]. Once again these observations point out the relevance of the modalities of signal source acquisition for the subsequent interpretation. Further data are needed before considering a widespread, routine utilization of this phenotypic marker.

5. Discussion

The ability of these new approaches towards morphology or physical properties in order to perform differential diagnosis is huge.

For what concerns imaging, let us show the sensitivity equal to 100% and specificity equal to 96% in a comparison between a prostatic cancer (Gleason index $=$ 7) and a healthy prostate after fractal analysis of Hematoxylin-Eosin slides (Fig. 10). We have here particularly focused on oral mucosa, where fractal analysis is a powerful tool to identify genetic diseases analyzing the microvessel distribution, as shown by our pilot studies [21, 52, 53]. For the same anatomical district several authors have recently presented other new imaging techniques able to identify diseases with high specificity, like as optic confocal reflectance microscopy [115], high-resolution microendoscopy [113, 128, 162], and optical coherence tomography [132, 158, 159, 160].

We would underline, in particular, that the data discussed here indicate the diagnostic ability of these new clinical physical and structural biomarkers in oral mucosa for identifying HNPCC carriers.

To date, the mechanisms underlying the observed data are unknown. We advanced the hypothesis that the above mentioned findings may be the result of a subclinical extracellular matrix (ECM) abnormality. It is well-known that ECM is constituted by different composite structures; it is a substance produced by cells and secreted into the intercellular environment, which contains various amounts of collagen, and chromophores actually transported in the blood. ECM plays an important role in modulating the behavior of the cells, interacting with and influencing their behavior. A number of families of ECM proteins are known, including collagens, proteoglycans, and laminins. As the chemical composition of ECM plays a pivotal role in cellular and tissue development, regeneration, and differentiation, ECM may play a crucial role in several human pathophysiological processes, including cancer [50, 53, 102].

In particular, ECM is a natural barrier that can prevent cancer development, which is a complex process that relies on interaction and communication between a number of cellular compartments [98, 106]. In fact ECM is a major component of the cellular and tumor microenvironment and a significant contributor to cancer progression. In this regard it is well-known that tumors are characterized by ECM remodeling and stiffening and ECM proteins are essential in the regulation of tumor angiogenesis [33, 131]. Moreover, cancer-produced ECM proteins can be key elements in cancer growth and metastasis. As a consequence, these observations suggest that ECM degradation, facilitating tumor development, may derive from the tumor itself or the tumor microenvironment [104].

The new potential clinical biomarkers derived from spectroscopic and colorimetric measurements or mucosa imaging are particularly interesting because they imply non-invasive, easily applicable and cheap methodologies. Similar conclusions have also been pointed out by other works [88, 89, 141]. In the near future, it would be beneficial to use these procedures as screening tools for cancer detection. However, the technology embedded in these applications needs to be improved in order to facilitate their widespread application for cancer detection and screening campaigns. However, let us note that these possible clinical biomarkers have already shown their importance for early detection of cancer, thus increasing their relevance, prognostic utility and potential.

Footnotes

Acknowledgments

We wish to dedicate this work to the memory of two friends who actively collaborated in the development of the methodology and its application, i.e., Prof. Ricardo Laurini (Tromsø, Norway) an unforgotten friend who inspired our research on the subject and Mr. Roberto Casini (Prato, Florence) who supported with his farsighted vision our preliminary applications of the concepts behind this review.

Abbreviations

References

American Cancer Society, Cancer Facts and Figures 2010, Atlanta, GA, American Cancer Society, 2010.

Abdul Majeed

A.A.

and Farah

C.S.

, Can immunohistochemistry serve as an alternative to subjective histopathological diagnosis of oral epithelial dysplasia? Biomarkers Cancer 5 (2013), 49–60.

Ali

and Dixit

, Infrared spectroscopic analysis of skin tumor of mice treated with several medicinal plants, Asian Pac J Trop Dis 3(5) (2013), 405–408.

Amatu

Sartore-Bianchi

Moutinho

Belotti

Bencardino

Chirico

Cassingena

Rusconi

Esposito

Nichelatti

Esteller

and Siena

, Promoter CpG Island hypermethylation of the DNA repair enzyme MGMT predicts clinical response to dacarbazine in a phase II study for metastatic colorectal cancer, Clin Cancer Res 19(8) (2013), 2265–2272.

Shen

T.T.

and Wang

, High speed spectral domain optical coherence tomography for retinal imaging at 500,000 A-lines per second, Biomed Opt Express 2(10) (2011), 2770–2783.

Qin

and Wang

R.K.

, Ultrahigh sensitive optical microangiography for in vivo imaging of microcirculations within human skin tissue beds, Opt Express 18(8) (2010), 8220–8228.

Anderson

N.L.

and Anderson

N.G.

, The human plasma proteome: History, character, and diagnostic prospects, Mol Cell Proteomics 1 (2002), 845–867.

Andersen

J.D.

et al., Identification of candidate biomarkers in ovarian cancer serum by depletion of highly abundant proteins and differential in-gel electrophoresis, Electrophoresis 31 (2010), 599–661.

Andersson-Engels

Canti

Cubeddu

Eker

Klinteberg

Pifferi

Svanberg

Taroni

Valentini

and Wang

, Preliminary evaluation of two fluorescence imaging methods for the detection and the delineation of basal cell carcinomas of the skin, Lasers Surg Med 26 (2000), 76–82.

10.

Bailey

M.J.

Verma

Fradkin

Lam

MacAulay

Poh

Markey

M.K.

and Sokolov

K.V.

, Detection of precancerous lesions in the oral cavity using oblique polarized reflectance spectroscopy: A clinical feasibility study, J Biomed Opt 22(6) (2017), 65002.

11.

Bali

Singh Padmavathi

M.P.

Khorate

and Ahmed

, Malignant fibrous histiocytoma – an unusual transformation from benign to malignant, Journal of Cancer Science and Therapy 2 (2010), 53–57.

12.

Barducci

, Information-theoretic approach to Fourier transform spectrometry, Journal of the Optical Society of America B 28(4) (2011), 637–648.

13.

Barducci

et al., Algorithm for the retrieval of columnar water vapor from hyperspectral remotely sensed data, Appl Opta 43 (2004), 5552–5563.

14.

Bassingthwaighte

J.B.

Liebovitch

L.S.

and West

B.J.

, Fractal Physiology, New York, Oxford University Press, 1994.

15.

Bayli

S.B.

and Ohm

J.E.

, Epigenetic gene silencing in cancer – a mechanism for early oncogenic pathway addiction? Nature Rev Cancer 6(107) (2006), 17.

16.

Baylin

S.B.

, DNA methylation and gene silencing in cancer, Nat Clin Pract Oncol 2(Suppl 1) (2005), S4–S11.

17.

Bélanger

Abran

Intes

Casanova

and Lesage

, Real-time diffuse optical tomography based on structured illumination, Journal of biomedical optics 15 (2010), 016006.

18.

Bence-Jones

, Papers on chemical pathology lecture III, Lancet (1847), 269–272.

19.

Bendardaf

et al., Apoptotic index and bcl-2 expression as prognostic factors in colorectal carcinoma, Oncology 64(4) (2003), 435–442.

20.

Bhatt

Mathur

Farooque

Verma

and Dwarakanath

B.S.

, Cancer biomarkers – current perspectives, Indian J Med Res 132 (2010), 129–149.

21.

Bianciardi

De Felice

Cattaneo

Parrini

Monaco

and Latini

, Fractal analysis of vascular network pattern in human diseases, Fractals in Biology and Medicine Birkhauser, Vol. IV, Locarno, Switzerland, 2005.

22.

Bianciardi

Leoncini

Lazzi

Lalinga

and Luzi

, Fractal analysis of the bone marrow in refractory anemia and acute leukemia, Rivista Di Biologia 2 (2000), 337–340.

23.

Bianciardi

and Luzi

, Fractal analysis of the bone marrow in myelodysplastic syndromes, Current Bioinformatics 9 (2014), 408–413.

24.

Bianciardi

Miracco

De Santi

M.M.

and Luzi

, Differential diagnosis between mycosis fungoides and chronic dermatitis by fractal analysis, J Dermatl Sci 33 (2003), 184–186.

25.

Bianciardi

Miracco

Lazzi

and Luzi

, Fractal analysis of epithelial-connective tissue interface in basal cell carcinoma of the skin, Current Bionformatics 8 (2013), 257–261.

26.

Bianciardi

Traversi

Cattaneo

De Felice

Monaco

Tosi

Parrini

and Latini

, Phase transition of microvascular network architecture in human pathologies, Theor Biol Forum 1 (2012), 37–45.

27.

Bolognesi

and Fenech

, Micronucleus assay in human cells: Lymphocytes and buccal cells, Methods Mol Boll 1044 (2013), 191–207.

28.

Bonassi

et al., An increased micronucleus frequency in peripheral blood lymphocytes predicts the risk of cancer in humans, Carcinogenesis 28 (2007), 625–631.

29.

Bonassi

El-Zein

Bolognesi

and Fenech

, Micronuclei frequency in peripheral blood lymphocytes and cancer risk: Evidence from human studies, Mutagenesis 26 (2011), 93–100.

30.

Boudoux

Yun

S.H.

W.Y.

White

W.M.

Iftimia

N.V.

Shishkov

et al., Rapid wavelength-swept spectrally encoded confocal microscopy, Opt Exp 13(20) (2005), 8214–822.

31.

Boursi

et al., First report of screening an asymptomatic population for cancer: The yield of an integrated cancer prevention center, Isr Med Assoc J 12 (2010), 21–25.

32.

Campbell

Durning

Cheema

and Naisby

, A simple tool for rapid access to a symptomatic breast clinic, Eur J Surg Oncol 30 (2004), 248–251.

33.

Campbell

N.E.

et al., Extracellular matrix proteins and tumor angiogenesis, J Oncol (2010), 586905. http://www.ncbi.nlm.nih.gov/pubmed?term=%22Petrik%20J%22%5BAuthor%5D.

34.

Cancer Research

UK.

,World cancer factsheet, published under the joint sponsorship of the International Agency for Research on Cancer (World Health Organization) and the Cancer Research UK URL http://publications.cancerresearchuk. org/downloads/Product/CS_REPORT_WORLD.pdf January 2014.

35.

Carrara

et al., Hereditary non-polyposis colorectal cancer carriers and abnormal light reflectance of oral mucosa, Gut 57(2) (2008), 279–279.

36.

Cavanagh

H.D.

Jester

J.V.

Essepian

Shields

and Lemp

M.A.

, Confocal microscopy of the living eye, Clao J 16(1) (1990), 65–73.

37.

Chan

D.W.

and Sell

, Tumor markers, in: Tietz Textbook of Clinical Chemistry Burtis

C.A.

and Ashwood

E.R.

, eds, WB Saunders Company, Philadelphia, 1999, pp. 722–724.

38.

Chandrasekhar

, Radiative Transfer, Dover Publications Inc., New York, USA, 1960.

39.

Chang

S.K.

Follen

Malpica

Utzinger

Staerkel

Cox

Atkinson

E.N.

MacAulay

and Richards-Kortum

, Optimal excitation wavelengths for discrimination of cervical neoplasia, IEEE Transactions on Biomedical Engineering 49(10) (2002), 1102–1110.

40.

Cherneva

Georgiev

and Petrova

, Early detection of lung cancer: A call for novel biomarkers-review, Biotechnol Biotechnol Equip 21(2) (2007), 145–150.

41.

Choo

J.R.

and Nielsen

T.O.

, Biomarkers for basal-like breast cancer, Cancers 2 (2010), 1040–1065.

42.

Church

T.R.

Wandell

Lofton-Day

Mongin

S.J.

Burger

Payne

S.R.

Castaños-Vélez

Blumenstein

B.A.

Rösch

Osborn

Snover

Day

R.W.

and Ransohoff

D.F.

, Prospective evaluation of methylated SEPT9 in plasma for detection of asymptomatic colorectal cancer, Gut 63(2) (2014), 317–325.

43.

Ciriello

Miller

M.L.

Aksoy

B.A.

Senbabaoglu

Schultz

and Sander

, Emerging landscape of oncogenic signatures across human cancers, Nature Genetics 45(10) (2013), 1127–1135.

44.

Colagar

A.H.

Chaichi

M.J.

and Khadjvand

, Fourier transform infrared microspectroscopy as a diagnostic tool for distinguishing between normal and malignant human gastric tissue, Journal of Biosciences 36(4) (2011), 669–677.

45.

Collier

Shen

de Pradier

Sung

K.B.

and Richards-Kortum

, Near real time confocal microscopy of amelanotic tissue: Dynamics of aceto-whitening enable nuclear segmentation, Opt Exp 6(2) (2000), 40–48.

46.

Conrads

T.P.

Zhou

Petricoin III

E.F.

Liotta

and Veenstra

T.D.

, Cancer diagnosis using proteomic patterns, Expert Rev Mol Diagn 3 (2003), 411–420.

47.

Corcuff

Gonnord

Pie’rard

G.E.

and Leveque

J.L.

, In vivo confocal microscopy of human skin: A new design for cosmetology and dermatology, Scanning 18(5) (1996), 351–355.

48.

Dabbs

and Glass

, Fiber-optic confocal microscope: FOCON, Appl Opt 31(16) (1992), 3030–3035.

49.

De Felice

et al., Predictive value of skin color for illness severity in the high-risk newborn, Pediatr Res 51 (2002), 100–105.

50.

De Felice

et al., Abnormal oral mucosal light reflectance: A new clinical marker of high risk for colorectal cancer, Gut 55 (2006), 1436–1439.

51.

De Felice

et al., Authors’ response to the letter titled “hereditary non-polyposis colorectal cancer carriers and abnormal light reflectance of oral mucosa”, Gut 57(2) (2008), 279–280.

52.

De Felice

Barducci

Latini

Bianciardi

and Parrini

, Node degree distribution in complex microvascular networks: A potential new diagnostic tool for extracellular matrix-related diseases, Fractals 14(4) (2006), 251–258.

53.

De Felice

Latini

Bianciardi

Parrini

Fadda

G.M.

Marini

Laurini

R.N.

and Kopotic

R.J.

, Abnormal vascular network complexity: A new phenotypic marker in hereditary non-polyposis colorectal cancer syndrome, Gut D52(12) (2003), 1764–1767.

54.

De Vos tot Nederveen Cappel

W.H.

et al., Colorectal surveillance in Lynch syndrome families, Fam Cancer 12(2) (2013), 261–265.

55.

Dhillon

V.S.

et al., Genetic polymorphisms of genes involved in DNA repair and metabolism influence micronucleus frequencies in human peripheral blood lymphocytes, Mutagenesis 26(1) (2011), 33–42.

56.

Di Leva

, Angioarchitectural morphometrics of brain tumors: Are they any potential histopathological biomarkers? Microvasc Res 80(3) (2010), 522–533.

57.

Di Ieva

, Fractal analysis of microvascular networks in malignant brain tumors, Clinical Neuropathology 31(5) (2012), 342–351.

58.

Di Ieva

et al., Microvascular morphometrics of the hypophysis and pituitary tumors: From bench to operating theatre, Microvasc Res 89 (2013), 7–14.

59.

Dolotov

L.E.

Sinichkin

Tuchin

V.V.

Utz

S.R.

Altshuler

G.B.

and Yaroslavsky

I.V.

, Design and evaluation of a novel portable erythema-melanin-meter, Lasers Surg Med 34 (2004), 127–135.

60.

Eliane

C.M.

Zeestraten

Benard

M.S.

and Reimers

, The prognostic value of the apoptosis pathway in colorectal cancer: A review of the literature on biomarkers identified by immunohistochemistry, Biomarkers in Cancer 5 (2013), 13–29.

61.

Ellsworth

D.L.

Ellsworth

R.E.

Liebman

M.N.

Hooke

J.A.

and Shriver

C.D.

, Genomic instability in histologically normal breast tissues: Implications for carcinogenesis, The Lancet Oncology 5(12) (2004), 753–758.

62.

Evan

G.I.

and Vousden

K.H.

, Proliferation, cell cycle and apoptosis in cancer, Nature 411(6835) (2001), 342–348.

63.

Van Eyk

J.E.

, Proteomics: Unraveling the complexity of heart disease and striving to change cardiology, Curr Opin Mol Ther 3 (2001), 546–553.

64.

Falconer

, Fractal Geometry: Mathematical Foundations and Application, Chichester, John Wiley, 1990.

65.

Farahnaz

et al., Fractal study on nuclear boundary of cancer cells in urinary smears, Iran J Pathol 6(2) (2011), 63–67.

66.

Feinberg

A.P.

Ohlsson

and Henikoff

, The epigenetic progenitor origin of human cancer, Nat Rev Genet 7 (2006), 21–33.

67.

Fellgett

P.B.

, The nature and origin of multiplex Fourier spectrometry, Notes Rec R Soc 60(1) (2006), 91–93.

68.

Ferlay

Soerjomataram

Ervik

Dikshit

Eser

Mathers

Rebelo

Parkin

D.M.

Forman

and Bray

, Cancer incidence and mortality worldwide: Sources, methods and major patterns in GLOBOCAN 2012, International Journal of Cancer 136 (2015), E359–E386, doi:10.1002/ ijc.29210, PMID:25220842.

69.

García-Bilbao

Armañanzas

Ispizua

Calvo

Alonso-Varona

Inza

Larrañaga

López-Vivanco

Suárez-Merino

and Betanzos

, Identification of a biomarker panel for colorectal cancer diagnosis, BMC Cancer 12(43) (2012), 1–13.

70.

Gmitro

A.F.

and Aziz

, Confocal microscopy through a fiber-optic imaging bundle, Opt Lett 18(8) (1993), 565–567.

71.

Goldberger

A.L.

Rigney

D.R.

and West

B.J.

, Chaos and fractals in human physiology, Sci Am 262 (1990), 42–49.

72.

Gonzalez

Rajadhyaksha

and Anderson

R.R.

, Non-invasive (real time) imaging of histologic margin of a proliferative skin lesion in vivo , J Invest Dermatol 111(3) (1998) 538–539.

73.

Gonzalez

Rajadhyaksha

Rubinstein

and Anderson

R.R.

, Characterization of psoriasis in vivo by reflectance confocal microscopy, J Med 30(5-6) (1999), 337–356.

74.

Grogan

R.H.

Mitmaker

E.J.

and Clark

O.H.

, The evolution of biomarkers in thyroid cancer – from mass screening to a personalized biosignature, Cancers 2(2) (2010), 885–912.

75.

Grulkowski

et al., High-precision, high-accuracy ultralong-range swept-source optical coherence tomography using vertical cavity surface emitting laser light source, Opt Lett 38(5) (2013), 673–675.

76.

Hade

E.M.

et al., Intraclass correlation estimates for cancer screening outcomes: Estimates and applications in the design of group-randomized cancer screening studies, J Natl Cancer Inst Monogr 2010 (2010), 97–103.

77.

Hafez

Hamadah

and Bachir

, Mapping of healthy oral mucosal tissue using diffuse reflectance spectroscopy: Ratiometric-based total hemoglobin comparative study, Lasers Med Sci 30(8) (2015), 2135–2141.

78.

Hajihashemi

M.R.

Grobmyer

S.R.

Al-Quran

S.Z.

and Jiang

, Noninvasive evaluation of nuclear morphometry in breast lesions using multispectral diffuse optical tomography, PLoS One 7(9) 2012.

79.

Halford

Rowan

Sawyer

Talbot

and Tomlinson

, O6-methylguanine methyltransferase in colorectal cancers: Detection of mutations, loss of expression, and weak association with G:C>A:T transitions, Gut 54(6) (2005), 797–802.

80.

Hanahan

and Weinberg

R.A.

, Hallmarks of cancer: The next generation, Cell 144(5) (2011), 646–674.

81.

Hassanein

Clay Callison

Callaway-Lane

et al., The state of molecular biomarkers for the early detection of lung cancer, Cancer Prev Res 5 (2012), 992–1006.

82.

Hatch

E.M.

Fischer

A.H.

Deerinck

T.J

and Hetzer

M.W.

, Catastrophic nuclear envelope collapse in cancer cell micronuclei, Cell 154(1) (2013), 47–60.

83.

Hawkridge

A.M.

and Muddiman

D.C.

, Mass spectrometry-based biomarker discovery: Toward a global proteome index of individuality, Annu Rev Anal Chem 2 (2009), 265–277.

84.

Gao

Gambhir

S.S.

and Cheng

, Near-infrared fluorescent nanoprobes for cancer molecular imaging: Status and challenges, Trends Mol Med 16(12) (2010), 574–583.

85.

Henley

S.J.

King

J.B.

German

R.R.

Richardson

L.C.

and Plescia

, Surveillance of screening – detected cancers (colon and rectum, breast, and cervix), MMWR Morb Mortal Wkly Rep 59(SS-9) (2010), 1–26.

86.

Herceg

and Hainaut

, Genetic and epigenetic alterations as biomarkers for cancer detection, diagnosis and prognosis, Mol Oncol 1 (2007), 26–41.

87.

Hoheisel

J.D.

, Microarray technology beyond transcript profiling and genotype analysis, Nat Rev Genet 7(3) (2006), 200–210.

88.

, et al., Assessment of the sensitivity and specificity of tissue-specific-based and anatomical-based optical biomarkers for rapid detection of human head and neck squamous cell carcinoma, Oral Oncology 50(9) (2014), 848–856.

89.

Vishwanath

Erkanli

Mulvey

, et al., Rapid determination of oxygen saturation and vascularity for cancer detection, PLoS One 8(12) (2013), e82977, doi:101371/journal.pone.0082977.

90.

Huang

et al., Optical coherence tomography, Science 254(5035) (1991), 1178–1181.

91.

Shao

Park

Y.I.

Peterson

V.M.

Castro

C.M.

Weissleder

and Lee

, Label-free detection and molecular profiling of exosomes with a nano-plasmonic sensor, Nat Biotechnol 32 (2014), 490–495.

92.

Jaenisch

and Bird

, Epigenetic regulation of gene expression: How the genome integrates intrinsic and environmental signals, Nature Genet 33 (2003), 245–254.

93.

Jørgensen

Bæk

Pedersen

Søndergaard

E.K.L.

Kristensen

S.R.

and Varming

, Extracellular vesicle (EV) array: Microarray capturing of exosomes and other extracellular vesicles for multiplexed phenotyping, J Extracell Vesicles 2 (2013), 20920–20929.

94.

Karley

Gupta

and Tiwari

, Biomarker for cancer: A great promise for future, World J Oncol 2(4) (2011), 151–157.

95.

Kas

, On the technicalities of discovering and applying protein biomarkers for cancer prevention, European Journal of Cancer Prevention 13 (2004), 437–446.

96.

King

R.D.

George

A.T.

Jeon

Hynan

L.S.

Youn

T.S.

Kennedy

D.N.

and Dickerson

, The Alzheimer’s disease neuroimaging initiative, “characterization of atrophic changes in the cerebral cortex using fractal dimensional analysis”, Brain Imaging and Behavior 3(2), (2009), 154–166.

97.

Klabunde

C.N.

Joseph

D.A.

King

J.B.

White

and Plescia

, Vital signs: Colorectal cancer screening test use – United States, 2012, MMWR 62(44) (2013), 881–888.

98.

Kostourou

and Papalazarou

, Non-collagenous ECM proteins in blood vessel morphogenesis and cancer, Biochim Biophys Acta 14 (2014), S0304-416500079-8.

99.

Kulasingamand

and Diamandis

E.P.

, Strategies for discovering novel cancer biomarkers through utilization of emerging technologies, Nat Clin Pract Oncol 5 (2008), 588–599.

100.

Kyle

R.A.

, Multiple myeloma: How did it begin? Mayo Clinical Proceedings 69 (1994), 680–3.

101.

Lambert

M.-P.

and Herceg

, Epigenetics and cancer, 2nd IARC meeting, Lyon, France, 6 and 7 December 2007, Mol Oncol 2 (2008), 33–40.

102.

Latini

et al., Oral mucosal color changes as a clinical biomarker for cancer detection, 21 (2012), 360–366.

103.

Latini

Bianciardi

Parrini

Laurini

R.N.

and De Felice

, Abnormal oral vascular network pattern geometry: A new clinical sign of Down syndrome, J Pediatr 148 (2006), 132–137.

104.

Lauden

et al., TGF-β-induced (TGFBI) protein in melanoma: A signature of high metastatic potential, J Invest Dermatol 134(6) (2014), 1675–85.

105.

Levin

Lieberman

D.A.

McFarland

et al., Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology, Gastroenterology 134(5) (2008), 1570–1595.

106.

Gao

Liu

and Pan

, ADAM-12 as a diagnostic marker for the proliferation, migration and invasion in patients with small cell lung cancer, PLoS One 9(1) (2014), e85936.

107.

Lichtenstein

Holm

N.V.

Verkasalo

P.K.

Iliadou

Kaprio

Koskenvuo

Pukkala

Skytthe

and Hemminki

, Environmental and heritable factors in the causation of cancer — analyses of cohorts of twins from sweden, denmark, and finland, N Engl J Med 343(2) (2000), 78–85.

108.

Liu

Njuguna

Matthews

Akers

W.J.

Sudlow

G.P.

Mondal

Tang

Gruev

and Achilefu

, Near-infrared fluorescence goggle system with complementary metal-oxide-semiconductor imaging sensor and see-through display, J Biomed Opt 18(10) (2013), 101303-1–101303-7.

109.

Losa

G.A.

and Nonnemacher

T.F.

, Self-similarity and fractal irregularity in pathologic tissues, Mod Pathol 9 (1996), 174–182.

110.

Losa

G.A.

, Fractals and their contribution to biology and medicine, Medicographia 34(3) (2012), 364–374.

111.

Losa

G.A.

and Castel

, Nuclear patterns of human breast cancer cells during apoptosis: Characterization by fractal dimension and (GLCM) co-occurrence matrix statistics, Cell Tissue Res 322 (2005), 257–267.

112.

Losa

G.A.

et al., Contribution of connective and epithelial tissue components to the morphologic organization of canine trichoiblastoma, Connect Tissue Res 50 (2009), 28–29.

113.

and Fei

, Medical hyperspectral imaging: A review, Journal of Biomedical Optics 19(1) (2014) 010901-1–010901-23.

114.

Luzzi

et al., Prognostic relevance of the angiogenetic pattern in stage IA pulmonary adenocarcinoma surgically resected, in: Third EACTS/ESTS Joint Meeting, Leipzig, Germany, 2004.

115.

Maitland

K.C.

Gillenwater

A.M.

Williams

M.D.

El-Naggar

A.K.

Descour

M.R.

and Richards-Kortum

R.R.

, In vivo imaging of oral neoplasia using a miniaturized fiber optic confocal reflectance microscope, Oral Oncology 44(11) (2008), 1059–1066, https://doiorg/10.1016/j.oraloncology.2008.02.002.

116.

Mandelbrot

B.B.

, How long is the coast of Britain? Statistical self-similarity and fractional dimension, Science 155 (1967), 636–640.

117.

Mandelbrot

B.B.

, Fractals: Form, Chance and Dimension, San Francisco, Freeman, 1977.

118.

Manne

Srivastava

R.-G.

and Srivastava

, Recent advances in biomarkers for cancer diagnosis and treatment, Drug Discovery Today 10(14) (2005), 965–976.

119.

Marbach

Koschinsky

T.H.

Gries

F.A.

and Heise

H.M.

, Non-invasive blood glucose assay by near-infrared diffuse reflectance spectroscopy of the human inner lip, Appl Spectrosc 47(7) (1993), 875–881.

120.

Masters

B.R.

and Thaer

A.A.

, In vivo human corneal confocal microscopy of identical fields of subepithelial nerve plexus, basal epithelial, and wing cells at different times, Microsc Res Tech 29(5) (1994), 350–356.

121.

Mehrotra

Tyagi

Jangir

D.K.

Dawar

and Gupta

, Analysis of ovarian tumor pathology by Fourier Transform Infrared Spectroscopy, J Ovarian Res 3(27) (2010), 1–6.

122.

Meulenberg

E.P.

, Immunochemical methods for ochratoxin a detection: A review, Toxins 4 (2012), 244–266.

123.

Mijovic

et al., Correlation of nuclear morphometry of primary melanoma of the skin with clinicopathological parameters and expression of tumor suppressor proteins (p53 and p16(INK4a)) and bcl-2 oncoprotein, Journal BUON 18(2) (2013), 471–476.

124.

Minchin

R.F.

and Martin

D.J.

, Nanoparticles for molecular imaging – an overview, Endocrinology 151 (2010), 474–481.

125.

Mishra

and Verma

, Cancer biomarkers: Are we ready for the prime time? Cancers 2 (2010), 190–208.

126.

Mokarram

Zamani

Kavousipour

Naghibalhossaini

Irajie

Moradi Sarabi

and Hosseini

S.V.

, Different patterns of DNA methylation of the two distinct O6-methylguanine-DNA methyltransferase (O6-MGMT) promoter regions in colorectal cancer, Mol Biol Rep 40(5) (2012), 3851–3857.

127.

Mor

Visintin

Lai

et al., Serum protein markers for early detection of ovarian cancer, Proc Natl Acad Sci USA 102 (2005), 7677–7682.

128.

Muldoon

T.J.

Anandasabapathy

Maru

and Richards-Kortum

, High-resolution imaging in barrett’s esophagus: A novel, low-cost endoscopic microscope, Gastrointest Endosc 68 (2008), 737–744.

129.

Mulero-Navarro

and Esteller

, Epigenetic biomarkers for human cancer: The time is now, Crit Rev Oncol Hemat 68 (2008), 1–11.

130.

Mydlarz

W.K.

Hennessey

P.T.

and Califfato

J.A.

, Advances and perspectives in the molecular diagnosis of head and neck cancer, Expert Opin Med Diagn 4 (2010), 53–65.

131.

Naba

et al., Extracellular matrix signatures of human mammary carcinoma identify novel metastasis promoters, e-Life (2014), e01308.

132.

Nankivil

et al., Coherence revival multiplexed, buffered swept source optical coherence tomography: 400 kHz imaging with a 100 kHz source, Opt Lett 39(13) (2014), 3740–3743.

133.

Odell

I.D.

and Cook

, Immunofluorescence Techniques, J Invest Dermatol 133 (2013), e41–4.

134.

Pedersen

J.W.

and Wandall

H.H.

, Autoantibodies as biomarkers in cancer, Labmedicine 42(10) (2011), 5623–628.

135.

Pritzker

V.P.

, Cancer biomarkers: Easier said than done, Clin Chem DC US 48 (2002), 1147–1150.

136.

Psofaki

Kalogera

Tzambouras

Stephanou

Tsianos

Seferiadis

and Kolios

, Promoter methylation status of hMLH1, MGMT, andCDKN2A/p16in colorectal adenomas, World J Gastroenterol 16(28) (2010), 3553–3560.

137.

Rajadhyaksha

Grossman

Esterowitz

Webb

R.H.

and Anderson

, In vivo confocal scanning laser microscopy of human skin: Melanin provides strong contrast, J Invest Dermatol 104(6) (2008), 946–952.

138.

Ramanujiam

, Fluorescence spectroscopy of neoplastic and non-neoplastic tissues, Neoplasia 2(1-2) (2000), 89–117.

139.

Roccaro

A.M.

Sacco

et al., BM mesenchymal stromal cell – derived exosomes facilitate multiple myeloma progression, J Clin Invest 123(4) (2013), 1542–1555.

140.

Rountree

M.R.

Bachman

K.E.

Herman

J.G.

and Baylin

S.B.

, DNA methylation, chromatin inheritance, and cancer, Oncogene 20 (2001), 3156–3165.

141.

Roy

H.K.

and Lynch

H.T.

, Diagnosing Lynch syndrome: Is the answer in the mouth? Gut 52 (2003), 1665–1667.

142.

Rusling

J.F.

Kumar

C.V.

Gutkinde

J.S.

and Patel

, Measurement of biomarker proteins for point-of-care early detection and monitoring of cancer, Analyst 135 (2010), 2496–2511.

143.

Safaei

Rezaei-Tavirani

Sobhi

and Esmaeil Akbari

, Breast cancer biomarker discovery: Proteomics and genomics approaches, Iran J Cancer Prev 6 (2013), 1–13.

144.

Sahu

R.K.

and Mordechai

, Fourier transform infrared spectroscopy in cancer detection, Future Ocnol 1(5) (2005), 635–647.

145.

Sai

Y.R.K.M.

Dattatreya

Anand

S.Y.

and Mahalakshmi

, Biomarkers and their role in premonition, interpretation and treatment of cancer, Journal of Cancer Science and Therapy S17 (2011).

146.

Sanders

H.W.

Smith

A.G.

and Crocker

, The potential values of fractal dimension measurement in histopathology, J Pathol 168(1) (1992), 86–94.

147.

Savage

V.M.

Herman

A.B.

West

G.B.

and Leu

, Using fractal geometry and universal growth curves as diagnostics for comparing tumor vasculature and metabolic rate with healthy tissue and for predicting response to drug therapies, Discrete Cont Dyn-B 18(4) (2013), 1077–1108.

148.

Sawan

Vaissiere

Murr

and Herceg

, Epigenetic drivers and genetic passengers on the road to cancer, Mutat Res 642(1-2) (2008), 1–13.

149.

Sawyers

C.L.

, The cancer biomarker problem, Nature 452(7187) (2008), 548–552.

150.

Sellar

R.G.

and Boreman

G.D.

, Comparison of relative signal-to-noise ratios of different classes of imaging spectrometer, Appl Opt 44(9) (2005), 1614–1624.

151.

Shen

Kondo

Rosner

G.L.

Xiao

et al., MGMT promoter methylation and field defect in sporadic colorectal cancer, J Natl Cancer Inst 97(18) (2005), 1330–1338.

152.

Simonova

and Karamancheva

, Application of fourier transform infrared spectroscopy for tumor diagnosis, Biotechnol Biotechnol Equip 27(6) (2013), 4200–4207.

153.

Singh

R.K.

Sudhakar

and Lokeshwar

B.L.

, Role of chemokines and chemokine receptors in prostate cancer development and progression, J Cancer Sci Ther 2 (2010), 89–94.

154.

Smithpeter

C.L.

Dunn

A.K.

Drezek

Collier

and Richards-Kortum

, Near real time confocal microscopy of cultured amelanotic cells: Sources of signal, contrast agents, and limits of contras, J Biomed Opt 3(4) (1998), 429–436.

155.

Srinivas

P.R.

Kramer

B.S.

and Srivastava

, Trends in biomarker research for cancer detection, Lancet Oncol 2 (2001), 698–704.

156.

Srinivas

P.R.

Verma

Zhao

and Srivastava

, Proteomics for cancer biomarker discovery, Clin Chem 48(8) (2002), 1160–1169.

157.

Suetsugu

Honma

Saji

Moriwaki

Ochiya

and Hoffman

R.M.

, Imaging exosome transfer from breast cancer cells to stroma at metastatic sites in orthotopic nude-mouse models, Adv Drug Delivery Rev 65(3) (2013), 383.

158.

Tsai

M.T.

et al., Effective indicators for diagnosis of oral cancer using optical coherence tomography, Opt Express 16(20) (2008), 15847–15862.

159.

Tsai

M.T.

et al., Noninvasive structural and microvascular anatomy of oral mucosae using handheld optical coherence tomography, Biomed Opt Express 8(11) (2017), 5001–5012.

160.

Tsai

M.T.

and Chan

M.C.

, Simultaneous 0.8, 1.0, and 1.3 μm multispectral and common-path broadband source for optical coherence tomography, Opt Lett 39(4) (2014), 865–868.

161.

Vaux

D.L.

and Korsmeyer

S.J.

, Cell death in development, Cell 96(2) (1999), 245–254.

162.

Vila

P.M.

et al., Discrimination of benign and neoplastic mucosa with a high-resolution microendoscope (HRME) in head and neck cancer, Ann Surg Oncol (2012), Epub DOI:101245/s10434-012-2351-1.

163.

Wang

Whiteaker

J.R.

and Paulovich

A.G.

, The evolving role of mass spectrometry in cancer biomarker discovery, Cancer Biol Ther 8(12) (2009), 1083–1094.

164.

Weibel

E.R.

, Fractal geometry: A design principle for living organisms, Am J Physiol 261 (1991), 361–69.

165.

Whitlock

E.P.

Lin

J.S.

Liles

et al., Screening for colorectal cancer: A targeted, updated systematic review for the US Preventive Services Task Force, Ann Intern Med 149 (2008), 638–658.

166.

Wilson

, Confocal Microscopy, San Diego, Academic Press, 1990.

167.

World Health Organization International Programme on Chemical Safety, Biomarkers and Risk Assessment: Concepts and Principles published under the joint sponsorship of the United Nations environment Programme, the International Labour Organisation, and the World Health Organization, http://www.who.int/iris/handle/10665/39037#sthash.B3afJqJq.dpuf, Geneva, World Health Organization, 1993.

168.

Yang

Zhang

Yang

and Jiang

, A fast multispectral diffuse optical tomography system for in vivo three-dimensional imaging of seizure dynamics, Appl Opt 51(16) (2012), 3461–3469.

169.

et al., Detection of prostate cancer on histopathology using color fractals and Probabilistic Pairwise Markov model, in: Conf Proc IEEE Eng Med Biol Soc, 2011, pp. 3427–3430.

170.

Zhang

Dan

D.L.

Weijun

Yanhua

Yanan

and Rong

, Detection of Cancer Biomarkers with Nanotechnology, Am J Biochem Biotechnol 9(1) (2013), 71–89

171.

Zuluaga

A.F.

Utzinger

Durkin

Fuchs

Gillenwater

Jacob

Kemp

Fan

and Richards-Kortum

, Fluorescence excitation emission matrices of human tissue: A system for in vivo measurement and method of data analysis, Appl Spectrosc 53(3) (1999), 302–311.