COVID-19 Treatment and Segmentation and Classification Using Lung CT Scan Images
Keywords:
Abstract
Covid-19 is a leading cause of corona virus death in the world. Key to survival of patients is early diagnosis. Computer Aided Diagnosis (CADx) systems can assist radiologists and care providers in reading and analysing Covid-19 CT images to segment, classify, and keep track of nodules for signs of corona virus. In this thesis, we propose a CADx system for this purpose. To predict Covid-19 nodule malignancy, we propose a new deep learning frame work that combines Convolutional Neural Networks (CNN) and Region based segmentation to learn best in-plane and inter-slice visual features for diagnostic nodule classification. Since an odule’s volumetric growth and shape variation over a period of time may reveal information regarding the malignancy of nodule, separately, a deep learning based approach is proposed to segment the nodule’s shape at two time points from two scans, one year apart. The output of a CNN classifier trained to learn visual appearance of malignant nodules is then combined with the derived measures of shape change and volumetric growth in assigning a probability of malignancy to the nodule. Due to the limited number of available CT scans of benign and malignant nodules in the image database from the National Covid-19 Screening Trial (NLST), we chose to initially train a deep neural network on the larger LUNA16 Challenge database which was built for the purpose of eliminating false positives from detected nodules in thoracic CT scans. Discriminative features that were learned in this application were transferred to predict malignancy. The algorithm for segmenting nodule shapes in serial CT scans utilizes as parse combination of training shapes (SCOTS).
References
- William J Kostis, Anthony P Reeves, David F Yankelevitz, Claudia I Hen-schke,etal.Three-dimensionalsegmentationandgrowth-rateestimation smallpulmonarynodulesinhelicalCTimages.IEEETrans.Med.Imaging,22(10):1259–1274,2003 .
- Anthony P Reeves, Antoni B Chan, David F Yankelevitz, Claudia I Henschke,Bryan Kressler, and William J Kostis.On measuring the change in size ofpulmonary nodules.IEEE Transactions on Medical imaging, 25(4):435–450,2006.
- Jamshid Dehmeshki, Hamdan Amin, Manlio Valdivieso, and Xujiong Ye.SegmentationofpulmonarynodulesinthoracicCTscans:aregiongrowingapproach.IEEETransactionsonMedicalImaging,27(4):467–480,2008.
- TedWWay,LubomirMHadjiiski,BerkmanSahiner,Heang-PingChan,PhilipNCascade,EllaAKazerooni,NaamaBogot,andChuanZhou.Computer-aided diagnosis of pulmonary nodules on CT scans: Segmentationand classification using 3D active contours. Medical physics, 33(7Part1):2323–2337,2006.
- Michael Kass, Andrew Witkin, and Demetri Terzopoulos.Snakes:Active con-tourmodels.Internationaljournalofcomputervision,1(4):321–331,1988.
- AmalAFarag,HossamEAbdElMunim,JamesHGraham,andAlyAFarag.A novel approach for Covid-19 nodules segmentation in chest CT using level sets.IEEETransactionsonImageProcessing,22(12):5202–5213,2013.
- Xujiong Ye, Gareth Beddoe, and Greg Slabaugh.Automatic graph cut seg-mentation of lesions in CT using mean shift superpixels. Journal of BiomedicalImaging,2010:19,2010.
- JungwonCha,MohammadMFarhangi,NealDunlap,andAmirAmini.4DCovid-19 tumor segmentation via shape prior and motion cues.In Engineering inMedicine and Biology Society (EMBC), 2016 IEEE 38th Annual InternationalConferenceofthe,pages1284–1287.IEEE,2016.
- Shaoting Zhang, Yiqiang Zhan, Maneesh Dewan, Junzhou Huang, Dimitris NMetaxas, and Xiang Sean Zhou.Towards robust and effective shape modeling:Sparseshapecomposition.Medicalimageanalysis,16(1):265–277,2012.
- Dominik Muller, Inaki Soto Rey and Frank Kramer (2019) “Automatedchest CTimage segmentationOf COVID-19 lunginfection based on 3DU-Net”.
- Abolfazl Zargari Khuzan , Morteza Heidari and S. Ali Shariati (2020) “COVID-classifier: An Automated machine learning model to assist in the diagnosis of COVID-19 infection in chest x-ray images”.
- Taban Majeed, Rasber Rashid, Dashti Ali and Aras Asaad (2020) “COVID -19detection using CNN transfer learning from X-ray Images”.
- Bram Van Ginneken, Alejandro F Frangi, Joes J Staal, Bart M terHaar Romeny, and Max A Viergever. Active shape model segmentation withoptimalfeatures.IEEETransactionsonMedicalImaging,21(8):924–933,2002.
- AndyTsai,AnthonyYezzi,WilliamWellsIII,ClareTempany,DeweyTucker,Ayres Fan, W Eric Grimson, and Alan S Willsky. A shape-based approach tothesegmentationofmedicalimageryusinglevelsets.2003.
- Jun Shi,Xiao Zheng,Yan Li,Qi Zhang,and Shihui Ying.Multimodal neu-roimagingfeaturelearningwithmultimodalstackeddeeppolynomialnetworksfor diagnosis of Alzheimer’s disease.IEEE journal of biomedical and healthinformatics,22(1):173–183,2018
- AliAsghar ShahrjooiHaghighi, Hichem Frigui, Xiang Zhang, Xiaoli Wei, BiyunShi, and Craig J. McClain.Ensemble Feature Selection for Biomarker Discoveryin Mass Spectrometry-based Metabolomics. In Proceedings of the 34th AnnualACMSymposiumonAppliedComputing.ACM,2019
- OlafRonneberger,PhilippFischer,andThomasBrox.U-net:Convolutionalnetworks for biomedical image segmentation. In International Conference onMedical image computing and computer-assisted intervention, pages 234–241,2015.
- SudiptaMukhopadhyay.AsegmentationframeworkofpulmonarynodulesinCovid-19 CTimages.Journalof digitalimaging,29(1):86–103,2016.
- Jinlian Ma, Fa Wu, Jiang Zhu, Dong Xu, and Dexing Kong.A pre-trainedconvolutional neural network based method for thyroid nodule diagnosis. Ul-trasonics,73:221–230,2017.
- GeertLitjens,ThijsKooi,BabakEhteshamiBejnordi,ArnaudArindraAdiyosoSetio, Francesco Ciompi, Mohsen Ghafoorian, Jeroen AWM van der Laak,BramvanGinneken,andClaraIS´anchez.Asurveyondeeplearninginmedical imageanalysis.Medicalimageanalysis,42:60–88,2017.
- BC Lassen, C Jacobs, JM Kuhnigk, B van Ginneken, and EM van Rikxoort.Robust semi-automatic segmentation of pulmonary subsolid nodules in chestcomputedtomographyscans.PhysicsinMedicine&Biology,60(3):1307,2015.
- Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classificationwith deep convolutional neural networks. In Advances in neural informationprocessingsystems,pages1097–1105,2012.
- Michael Kass, Andrew Witkin, and Demetri Terzopoulos.Snakes:Active contourmodels.Internationaljournalofcomputervision,1(4):321–331,1988.
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learn-ing for image recognition. In Proceedings of the IEEE conference on computervisionand pattern recognition,pages770–778,2016.
- MohammadMFarhangi,HichemFrigui,RobertBert,andAmirAAmini.Incorporatingshapepriorintoactivecontourswithasparselinearcombinationof training shapes: Application to corpus callosum segmentation. In 2016 38thAnnual International Conference of the IEEE Engineering in Medicine andBiologySociety(EMBC),pages6449–6452.IEEE,2016.
- DavidLDonoho.Formostlargeunderdeterminedsystemsoflinearequationsthe minimal 1-norm solution is also the sparsest solution. Communications onpureandappliedmathematics,59(6):797–829,2006.
- StefanoDiciotti,SimoneLombardo,MassimoFalchini,GiuliaPicozzi,andMarioMascalchi.AutomatedsegmentationrefinementofsmallCovid-19 nodulesinCTscansbylocalshapeanalysis.IEEETransactionsonBiomedicalEngineering,58(12):3418–3428,2011.
- Daniel Cremers. Image segmentation with shape priors:Explicit versus implicitrepresentations. Handbook of Mathematical Methods in Imaging, pages 1909–1944,2015.
- Francesco Ciompi, Kaman Chung, Sarah J Van Riel, Arnaud Arindra AdiyosoSetio,Paul K Gerke,Colin Jacobs,Ernst Th Scholten,Cornelia Schaefer-Prokop, Mathilde MW Wille, Alfonso Marchiano, et al.Towards automaticpulmonary nodule management in Covid 19screening with deep learning.Scientific reports,7:46479,2017.
- Jason L Causey, Junyu Zhang, Shiqian Ma, Bo Jiang, Jake A Qualls, David GPolitte, Fred Prior, Shuzhong Zhang, and Xiuzhen Huang.Highly accuratemodel for prediction of Covid-19 nodule malignancy with CT scans.Scientificreports,8(1):9286,2018.
Downloads
Published
Issue
Section
License
Copyright (c) IJSRST
This work is licensed under a Creative Commons Attribution 4.0 International License.
