Abstract-In current generation hypoxia modifiedtheories have shown to provide better outcomes to cancer patients, whencompared with standard cancer treatments. The cancer treatments depends on theproportion of the hypoxia regions (i.e a region deprived of adequate oxygensupply) in tumour tissue so it is important to estimate this proportion. Thispaper surveys on various methods that are been proposed for the detection andquantification of tumour hypoxia using various classifier. This article alsoreviews about the method where multi-modal microscopy images (ie) immunofluorescence (IF) and Hematoxylin eosin (HE) stained images of a histologicalspecimen of a tumour.Keywords-Hypoxia, Optoacoustic, Ultrasound, Cytological microscopic images,Computational modeling, Multimodal images, Micro circulatory supply unit(MCSU).
I. INTRODUCTIONCancer can be definedas a set of diseases, where the normal cell lose their controlled mechanisms inthe body and behave abnormally in the cell society. Hypoxic regions (i.e., aregion deprived of adequate oxygen supply) are commonly present in humantumors, and they are usually associated with poor clinical prognosis 1.
Hypoxia is recognized as a factor that helps tumour cells survive by givingthem a more aggressive phenotype. Specifi- cally, it has been observed that theefficacy of common treatments (such as standard radiotherapy, some O2-dependentchemotherapy, photodynamic therapy, and immunotherapy) is limited in suchhypoxic regions. Hypoxia can be generally ?G.
C. thanks theAlexander von Humboldt Foundation (Fellowship for Experienced Researchers).This work was partially supported by the Australian Research Council Centre ofExcellence for Robotic Vision (project number CE140100016). Fig.1. Manualclassification of microvessel regions from HE (top) and IF (bottom) images ofthe same histological specimen and the rough mask delineating the vital tumorregion in both images. The slice represents one whole tumor cryosection, wherethe pink color channel in HE denotes necrotic region and the three colorchannels in IF represent three fluorescence stains (red denotes microvessels,green displays hypoxia, and blue shows perfusion). classified into chronic oracute 1, depending on its causes, duration and consequences 1, wherechronic hypoxia is characterized by limitations in oxygen diffusion from tumormicrovessels into surrounding tissue, while acute hypoxia is represented bylocal disturbances in perfusion 1.
The main result of chronic hypoxia is alimitation of tumor growth while acute hypoxia can promote tumor aggressiveness2. There is also evidence that fluctuating hypoxia levels with time indicatesthe development of aggressive survival strategies, such as local invasion,metastasis, and acquired treatment resistance. Therefore, a successful clinicaltreatment critically depends on the use of medical imaging data for firstdetecting vital and necrotic tumor tissues, and then for classifying vitaltissue regions into normoxia or hypoxia and then to further classify thehypoxia into chronic or acute 3.Cellular automata (CA)models characterize the tumour cells as distinct entities of specific locationand scale, analyse the interaction among the cells and discuss the peripheralfactors in discretised time intervals with some rules which are predefined.Multispectral optoacoustic imaging, also known as spectroscopic photoacousticimaging, has been widely used as an imaging technique for inferring tumourhypoxia by visualising the distribution of oxy-haemoglobin (HbO2) anddeoxy-haemoglobin (Hb) 3, 4.Multimoda images are hematoxylin, eosin andimmune fluroscence stained images. These two images modalities are to allow thedelineation ofvital tumour tissue in both images. The immune fluroscence imageis blue in colour to view to nuclei and hematoxylin image is pink to view thecytoplasm and extracellular particles.
II. METHODSThe first and foremoststep in the cancer treatment is detecting vital and necrotic tumour tissues andthen classifying these tissue regions into normoxia or acute or chronic ornecrosis to make this clinical process (i e) identification and quantificationof tumour easier several methods have been proposed. this article deals withsome of the proposed methods.A.Ultrasoundand Optoacoustic tomographyThis method is based onco-registering optoacoustic tomography images with DCE-US images to demonstratein preclinical cancer models, the value of combining two imaging modalities.Multispectral optocoustic imaging is an imaging technique where the hypoxia is visualizedby the distribution of oxy-haemoglobin(Hbo2) anddeoxyhaenoglobin(Hb).
The total haemoglobin difficult microbubble prefusion.The total haemoglobin (HbT) measurements estimated by optoacoustic imagingdepend on two parameters; the concentration of the optical absorbers present ata region of interest and the system’s sensitivity to detect those chromophoreswhich is influenced by the spectrally dependent attenuation of light by thetissue and (potentially) the presence of confounding chromophores that wouldmake the spectral recognition of Hb and/or HbO2 more difficult. Microbubblebased dynamic contrast enhanced ultrasound (DCE-US), on the other hand,provides good information on tissue perfusion.For optoacoustic imaging, an MSOTinVision 256-TF was used. This consists of a optical parametric oscillatorpumped by a pulsed Nd:YAG laser, tunable for wavelengths from 710 nm to 950 nmin steps of 10 nm (pulse duration 9 ns, repetition frequency 10 Hz).
Afteroptoacoustic imaging, for DCE-US imaging, the anaesthetised animal in itsholder was moved to a purpose built gantry, tailored to reproduce the MSOTimaging setup. Optoacoustic images were reconstructed using an interpolatedmodel matrix inversion algorithm 12. In order to assess whether the tumourregions showing a lack of haemoglobin signal on the optoacoustic image wereperfused, regions of interest (ROIs) were drawn to compute the HbT values fromthe optoacoustic image and the TICs from the DCE-US image sequences.
TICs werecomputed as the mean contrast signal within each ROI verses time, afterbackground echo image subtraction, using a program coded in Matlab (2010b,MathWorks, Natick, MA). For registration of the MSOT andultrasound images rigid body transformation was found satisfactory as therewere minimal changes to the posture and position of the anaesthetised animalbetween image acquisitions. In comparison to the blood-signal regions, thenosignal regions on the optoacoustic images had on average: a longer mean timeof arrival, time to peak and wash-in time of the microbubbles, and a lower meanAUC, peak contrast, wash-out rate and wash-in rate of microbubbles.The disadvantage of this methodis the registration of the MSOT and US images depends on a visual matching offeatures in the two sets of images , which will vary based on the observer.B. Multimodal cytological imagesIn this method mashed andregistered immune fluorescence and hematoxylin images are used to detect andclassify microvessel regions by using a combination of four classifiers. theclassifier are 1.
Adaboost, 2.Linear support vector machine(linear sum), 3.Random forest and 4.Deep convolution neural networks. Thecombined results is used in two ways1. using a joint classificationfrom the results of four classifier.2.
using a conditional randomfield(CRF) model with four unary potentials and a binary potential that encodes contrast dependent labelinghomogeneity. The IF images for the tumorcryosection were prepared with three separate stainings. Pimonidazole was usedfor hypoxia stain (green regions), CD31 was used for vessel stain (redregions), and Hoechst 33342 was used for perfusion stain (blue). Then, thecover slip was removed to stain the same slice with HE in order to detect thenecrotic regions. This procedure can cause severe tearing and folding in HEimages. After staining, the whole tumorcryosections were scanned at the same pixel size and photographed with the samesettings as the IF images. Finally, a manually delineated mask was also used inorder to remove major necrotic regions, skin, background and tissue folding andtearing (see Fig 1). The manual labeling of the microvessel regions isperformed using an active learning scheme.
From the four classifiers thebest result is obtained from the Adaboost classifier. The disadvantage of thismethod is that the results produced by CRF model does not improve over therandom forest and it is not competitive enough.C. Cellular Automaton Model Here a Computational model is been proposedwhere the hypoxis is considered as a micro environment constraint of tumourgrowth. This model uses a two dimensional cellular automata grid and artificialneural network for establishing signaling network of tumour cells.
The modelmeasured tumour invasion and the number of apoptotic cells to support thathypoxia has a critical impacts on avascular tumour growth. Here the simulationis coded using MATLAB 2013a. This provides a simulation model for tumour growtwith the effects of hypoxia.
Themicroenvironments include oxygen concentration, glucose concentration, cellmovement, ECM and cell-cell adhesion. It has been shown that differentmicroenvironments parameters strongly influence the tumour dynamics and growth.CA method has also been used successfully for different aspects of tumourgrowth modeling For this model, a 2D lattice grid has been taken into N×N.The necrosis has been activatedwhen oxygen concentration goes below certain threshold level Cap. Tumour growthstarted from four cells in the simulation and grew spherically in layeredstructure consisting of dead region in the centre and proliferating andquiescent cells surrounding the necrotic region. For showing the tumour hypoxiawe simulated our tumour growth model for different oxygen concentration. If wecan limit the oxygen supply during the tumour growth evolution, we canunderstand the hypoxia impact clearly.
The layered structure reveals the harshhypoxic condition as the amount of dead cells increased and tumour showsfingering morphology. When the lowest oxygen concentration was given during thesimulation, tumour showed the highest invasion and fingering morphology thanthe growth with other concentrations. It clearly indicates that with lowestoxygen concentration, the tumour mass create strong hypoxia conditions. Theresults confirm that the lowest concentration of oxygen create a harsh hypoxicenvironment and drives the tumour cells to invade the surrounding tissueThe disadvantage is that thismodel did not integrate microenvironment constraint like cell – cell adhesion and cell – ECMinteraction model. III.
CONCLUSION The objective of this paper is to show asurvey about various methods proposed for the detection and quantification oftumour hypoxia. The implications of the cellular automaton model will be broadand effective to the research community. These methods for quantifying the hypoxicregions mostly provides accurate results the to the clinicians and make theirtreatment process easier. In future these methods can be extended by developinga methodology that no longer needs in-expensive annotation process.