Research article Special Issues

FCM-DNN: diagnosing coronary artery disease by deep accuracy fuzzy C-means clustering model


  • Cardiovascular disease is one of the most challenging diseases in middle-aged and older people, which causes high mortality. Coronary artery disease (CAD) is known as a common cardiovascular disease. A standard clinical tool for diagnosing CAD is angiography. The main challenges are dangerous side effects and high angiography costs. Today, the development of artificial intelligence-based methods is a valuable achievement for diagnosing disease. Hence, in this paper, artificial intelligence methods such as neural network (NN), deep neural network (DNN), and fuzzy C-means clustering combined with deep neural network (FCM-DNN) are developed for diagnosing CAD on a cardiac magnetic resonance imaging (CMRI) dataset. The original dataset is used in two different approaches. First, the labeled dataset is applied to the NN and DNN to create the NN and DNN models. Second, the labels are removed, and the unlabeled dataset is clustered via the FCM method, and then, the clustered dataset is fed to the DNN to create the FCM-DNN model. By utilizing the second clustering and modeling, the training process is improved, and consequently, the accuracy is increased. As a result, the proposed FCM-DNN model achieves the best performance with a 99.91% accuracy specifying 10 clusters, i.e., 5 clusters for healthy subjects and 5 clusters for sick subjects, through the 10-fold cross-validation technique compared to the NN and DNN models reaching the accuracies of 92.18% and 99.63%, respectively. To the best of our knowledge, no study has been conducted for CAD diagnosis on the CMRI dataset using artificial intelligence methods. The results confirm that the proposed FCM-DNN model can be helpful for scientific and research centers.

    Citation: Javad Hassannataj Joloudari, Hamid Saadatfar, Mohammad GhasemiGol, Roohallah Alizadehsani, Zahra Alizadeh Sani, Fereshteh Hasanzadeh, Edris Hassannataj, Danial Sharifrazi, Zulkefli Mansor. FCM-DNN: diagnosing coronary artery disease by deep accuracy fuzzy C-means clustering model[J]. Mathematical Biosciences and Engineering, 2022, 19(4): 3609-3635. doi: 10.3934/mbe.2022167

    Related Papers:

    [1] Zhe Su, Yiying Tong, Guo-Wei Wei . Persistent de Rham-Hodge Laplacians in Eulerian representation for manifold topological learning. AIMS Mathematics, 2024, 9(10): 27438-27470. doi: 10.3934/math.20241333
    [2] Ruzhi Song, Fengling Li, Jie Wu, Fengchun Lei, Guo-Wei Wei . Multi-scale Jones polynomial and persistent Jones polynomial for knot data analysis. AIMS Mathematics, 2025, 10(1): 1463-1487. doi: 10.3934/math.2025068
    [3] Huan Xu, Tao Yu, Fawaz E. Alsaadi, Madini Obad Alassafi, Guidong Yu, Jinde Cao . Some spectral sufficient conditions for a graph being pancyclic. AIMS Mathematics, 2020, 5(6): 5389-5401. doi: 10.3934/math.2020346
    [4] Li Shen, Jian Liu, Guo-Wei Wei . Evolutionary Khovanov homology. AIMS Mathematics, 2024, 9(9): 26139-26165. doi: 10.3934/math.20241277
    [5] Wei Xue, Pengcheng Wan, Qiao Li, Ping Zhong, Gaohang Yu, Tao Tao . An online conjugate gradient algorithm for large-scale data analysis in machine learning. AIMS Mathematics, 2021, 6(2): 1515-1537. doi: 10.3934/math.2021092
    [6] Shama Liaqat, Zeeshan Saleem Mufti, Yilun Shang . Newly defined fuzzy Misbalance Prodeg Index with application in multi-criteria decision-making. AIMS Mathematics, 2024, 9(8): 20193-20220. doi: 10.3934/math.2024984
    [7] Jinfang Li, Chunjiang Wang, Li Zhang, Jian Zhang . Multi-solitons in the model of an inhomogeneous optical fiber. AIMS Mathematics, 2024, 9(12): 35645-35654. doi: 10.3934/math.20241691
    [8] Qingbing Liu, Aimin Xu, Shuhua Yin, Zhe Tu . A note on the preconditioned tensor splitting iterative method for solving strong M-tensor systems. AIMS Mathematics, 2022, 7(4): 7177-7186. doi: 10.3934/math.2022400
    [9] Muhammad Riaz, Khadija Akmal, Yahya Almalki, S. A. Alblowi . Cubic m-polar fuzzy topology with multi-criteria group decision-making. AIMS Mathematics, 2022, 7(7): 13019-13052. doi: 10.3934/math.2022721
    [10] Changlong Yu, Jing Li, Jufang Wang . Existence and uniqueness criteria for nonlinear quantum difference equations with p-Laplacian. AIMS Mathematics, 2022, 7(6): 10439-10453. doi: 10.3934/math.2022582
  • Cardiovascular disease is one of the most challenging diseases in middle-aged and older people, which causes high mortality. Coronary artery disease (CAD) is known as a common cardiovascular disease. A standard clinical tool for diagnosing CAD is angiography. The main challenges are dangerous side effects and high angiography costs. Today, the development of artificial intelligence-based methods is a valuable achievement for diagnosing disease. Hence, in this paper, artificial intelligence methods such as neural network (NN), deep neural network (DNN), and fuzzy C-means clustering combined with deep neural network (FCM-DNN) are developed for diagnosing CAD on a cardiac magnetic resonance imaging (CMRI) dataset. The original dataset is used in two different approaches. First, the labeled dataset is applied to the NN and DNN to create the NN and DNN models. Second, the labels are removed, and the unlabeled dataset is clustered via the FCM method, and then, the clustered dataset is fed to the DNN to create the FCM-DNN model. By utilizing the second clustering and modeling, the training process is improved, and consequently, the accuracy is increased. As a result, the proposed FCM-DNN model achieves the best performance with a 99.91% accuracy specifying 10 clusters, i.e., 5 clusters for healthy subjects and 5 clusters for sick subjects, through the 10-fold cross-validation technique compared to the NN and DNN models reaching the accuracies of 92.18% and 99.63%, respectively. To the best of our knowledge, no study has been conducted for CAD diagnosis on the CMRI dataset using artificial intelligence methods. The results confirm that the proposed FCM-DNN model can be helpful for scientific and research centers.



    Human milk stored in human milk banks (HMBs) is the gold standard for feeding preterm neonates, especially those born before 32 weeks of age. To reduce the infectious risk linked to a possible contamination by harmful bacteria through human milk of these preterm neonates, pasteurization (mainly standard Holder pasteurization at 62.5 °C for 30 min) is carried out in HMBs to kill bacteria present in raw human milk. However, B. cereus endospores may not completely be eliminated by pasteurization and can therefore give rise to new vegetative forms of B. cereus in the milk post-pasteurization [1]. Additionally, several reports show that pasteurized human milk can be contaminated with Bacillus cereus [2][4]. This is a cause for concern as preterm neonates have previously been reported to be susceptible to severe cases of B. cereus infections [5],[6] and human breast milk was pointed out as a possible source of contamination [5],[7]. The most usual symptoms of B. cereus infections are diarrheal and emetic [8]. Diarrheal symptoms are caused by a variety of toxins such as pore-forming toxins hemolysin BL (Hbl) and non-hemolytic enterotoxin (Nhe) which both comprise three subunits, as well as cytotoxin K (cytK) [8],[9]. Additionally, enterotoxins entFM and bceT, encoded respectively by entFM and bceT, have also been proposed to play a part in diarrheic symptoms witnessed in B. cereus food poisoning [10],[11]. However, the enterotoxic properties of bceT were later questioned [12]. Meanwhile, the emetic form of B. cereus infection is driven by the action of cereulide toxin, a cyclic dodecadepsipeptide resistant to heat, proteolysis and acidic pH [9]. The latter toxin is encoded by genes found on a megaplasmid and its prevalence in B. cereus strains is relatively low, as compared to other toxins [9],[13]. A classification of B. cereus strains in seven different toxin profiles: A (nhe+, hbl+, cytK+), B (nhe+, cytK+, ces+), C (nhe+, hbl+), D (nhe+, cytK+), E (nhe+, ces+), F (nhe+) and G (cytK+) has been proposed [14].

    In Amiens-Picardie Human Milk Bank (APHMB), contamination rates with B. cereus of pasteurized milk have been shown to be the leading cause of milk rejection post-microbiological quality control [3]. Therefore, we endeavored to assess whether this contamination could (i) be due to B. cereus strains presenting a worrisome toxin expression profile and (ii) arise from the hospital environment, especially APHMB rooms, linen and materials, or simply from donating mothers. The toxin gene patterns from various B. cereus strains have been studied and seem to be influenced by their geographical and/or foodstuff origins [9]. It was hypothesized that toxin gene patterns could also help us in discerning B. cereus strains originating from human milk and the environment. Hence, the aims of this work were first to evaluate the pathogenic potential of strains isolated from human milk and then to compare toxin expression profiles in order to determine if the B. cereus contamination of pasteurized human milk in APHMB could arise from environmental strains or not. In a further attempt to differentiate between strains isolated from breast milk and those from the hospital environment, the results obtained were compared to those gathered through analysis of Fourier-transform infra-red (FTIR) spectra, a promising new method for the discrimination of bacterial isolates of the same species [15].

    A total of 63 B. cereus strains coming from human milk (HM) donations made to the Centre Hospitalier Universitaire APHMB were collected during a previously described study [3]. The study design was validated by the ad hoc Research Commission at the Centre Hospitalier Universitaire Amiens-Picardie (institutional review board) following the French Regulation [16]. In accordance with the European General Data Protection Regulation [17], the opinion of the Research Commission on the registration of the databank built for this study by the national commission in charge of data protection (Commission Nationale de l'Informatique et des Libertés) was also sought. The registration was deemed unnecessary (decision date: 24 April 2019).

    Twenty-seven environmental strains were retrieved from samples routinely collected through the hospital environment surveillance program. Those included samples coming from APHMB rooms (13 isolates) as well as other wards (4 isolates), linen (7 isolates) and endoscopes (3 isolates).

    All strains were identified using matrix-assisted laser desorption ionisation-time of flight mass spectrometry (MALDI Biotyper 2.2; Bruker Daltonik GmbH, Bremen, Germany). They were kept at –20 °C on cryobeads (Mast Diagnostic, Amiens, France) until use.

    Additionally, for the detection of toxin genes, several collection strains were included in the analysis as controls: B. cereus DSM 31, B. cereus DSM 4312 and B. cereus DSM 4313, (Deutsche Sammlung für Mikroorganismen und Zellkulturen, Braunschweig, Germany).

    DNA extraction was performed using DNeasy Tissue extraction kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions for Gram positive bacteria. Following the extraction procedure, DNA contents in extracts were measured using the Nanodrop apparatus (Thermo Fisher Scientific, Illkirch, France). Amplification of 16S rDNA (positive control), ces, bceT, cytK, nhe(ABC), and hbl(ACD) genes was performed with a Verity thermal cycler (Applied Biosystems, France) using primers previously described (Table 1). A 25 µL reaction volume consisting of 12.5 µL DreamTaq PCR Master Mix (Thermo Fisher Scientific), 1µL of forward and 1 µL of reverse primer (final concentration 0.2 to 0.5 µM), 4.5 µL molecular biology grade water and 5 µL of template DNA was submitted to amplification. The amplification typically consisted of 1 cycle at 94 °C for 5 min followed by 35 cycles including 1 min at 94 °C, 1 min at the mentioned annealing temperature and 2 min at 72 °C. A final elongation cycle at 72 °C for 5 min completed the amplification run. For ces amplification, the PCR protocol was composed of a denaturation step at 95 °C for 15 min followed by five cycles of 1 min at 95 °C, 75 s at 53 °C, and 50 s at 72 °C and then by 25 cycles of 1 min at 95 °C, 75 s at 58 °C, and 50s at 72 °C. A final elongation step consisting of 72 °C for 5 min ended the amplification procedure.

    Table 1.  Primer sequences and annealing conditions used in this study.
    Target gene Sequence (5′-3′) Annealing temperature (°C) Amplicon size (bp) Reference
    16S rDNA Fa: ACTCCTACGGGAGGCAG
    Ra: ATTACCGCGGCTGCTGGCA
    55 196 [18]
    bceT F: CGTATCGGTCGTTCACTCGG
    R: GTTGATTTTCCGTAGCCTGGG
    55 661 [19]
    ces F: GGTGACACATTATCATATAAGGTG
    R: GTAAGCGAACCTGTCTGTAACAACA
    53/58 1271 [20]
    cytK F: CGACGTCACAAGTTGTAACA
    R: CGTGTGTAAATACCCCAGTT
    54 565 [21]
    entFM F: GTTCGTTCAGGTGCTGGTAC
    R: AGCTGGGCCTGTACGTACTT
    54 486 [21]
    hblA F: AAGCAATGGAATACAATGGG
    R: AGAATCTAAATCATGCCACTGC
    56 1154 [19]
    hblC F: GATAC(T,C)AATGTGGCAACTGC
    R: TTGAGACTGCTCG(T,C)TAGTTG
    58 740 [19]
    hblD F: ACCGGTAACACTATTCATGC
    R: GAGTCCATATGCTTAGATGC
    58 829 [19]
    nheA F: TAAGGAGGGGCAAACAGAAG
    R: TGAATGCGAAGAGCTGCTTC
    54 759 [21]
    nheB F: CAAGCTCCAGTTCATGCGG
    R: GATCCCATTGTGTACCATTG
    54 935 [21]
    nheC F: ACATCCTTTTGCAGCAGAAC
    R: CCACCAGCAATGACCATATC
    54 618 [21]

    a: F=Forward primer/R=Reverse primer

    Amplification products were visualized by electrophoresis on a 1.5% agarose gel containing GelRed© nucleic acid stain (Merck Millipore, Molsheim, France) and run for 1h at 100V followed by UV transillumination on the IBright 1500© system (Thermo Fisher Scientific). Their size was estimated using Generuler DNA ladder (Thermo Fisher Scientific).

     | Show Table
    DownLoad: CSV

    All B. cereus strains were handled similarly to minimize differences in the carbohydrate, lipid and protein structures of the cell wall that might arise from variations in culture conditions. Frozen strains were subcultured twice on plate count agar (Biomérieux, Marcy l'étoile, France) for 18 to 24 hours at 36 ± 1 °C. Well individualized colonies were then processed using IR Biotyper R© kit (Bruker Daltonik GmbH) according to the manufacturer's instruction. Three replicates of each strain were submitted to FTIR analysis using the IR-Biotyper© system (Bruker Daltonik GmbH) in transmission mode in the spectral range of 4,000–500 cm–1 (mid-IR). For each run, quality control was performed with the InfraRed Test Standards (IRTS 1 and 2) in the IR Biotyper R© kit. Resulting spectra were analyzed using OPUS software V8.2.28 (Bruker Daltonik GmbH). Principal Component Analysis (PCA) was applied for multivariate analyses, using the IR Biotyper R© V3.1 software functionality.

    Comparison of the prevalences of toxin genes and toxin profiles between environmental and HM B. cereus strains was performed using Fisher exact test. A p-value below 0.05 was considered as significant.

    PCR efficiency was first checked for all DNA extracts with the amplification of a control gene (16S rDNA). All extracts gave positive results, validating the absence of inhibitory substances that might impair the amplification process. The prevalence of toxin genes and toxin profiles [16] are summarized in Table 2 and Figure 1, respectively. Even though no statistically significant differences in toxin gene prevalences were found between HM and environmental B. cereus isolates, prevalences of nheA, nheB, and ces in HM isolates failed to be qualified as higher than that of environment strains by a narrow margin (p = 0.054; p = 0.063 and p = 0.12, respectively; Fisher exact test). Similarly, no differences in toxin profiles could be highlighted by the statistical analysis comparing HM and environmental isolates.

    Table 2.  Prevalences of toxin genes in Bacillus cereus strains included in this study.
    Toxin genes Prevalences
    Overall (90)b Human Milk (63) Environment (27)
    bceT 38 (34) 38 (24) 37 (10)
    ces 27 (24) 32 (20) 15 (4)
    cytk 73 (66) 76 (48) 67 (18)
    entFM 92 (83) 92 (58) 93 (25)
    hblA 30 (27) 33 (21) 22 (6)
    hblC 29 (26) 32 (20) 22 (6)
    hblD 59 (53) 62 (39) 52 (14)
    nheA 86 (77) 90 (57) 74 (20)
    nheB 93 (84) 97 (61) 85 (23)
    nheC 90 (81) 92 (58) 85 (23)

    a: results expressed as percentage (number of positive isolates)

    b: number of isolates per category

     | Show Table
    DownLoad: CSV
    Figure 1.  Toxin profiles of Bacillus cereus strains included in this study. Toxin profiles are defined as follows: A = nhe+, hbl+, cytK+; B = nhe+, cytK+, ces+; C = nhe+, hbl+; D = nhe+, cytK+; E = nhe+, ces+; F = nhe+; G = cytK+ and Other = combination of genes different from those described for A to G profiles.

    A scatter-plot of PCA results from the 90 analyzed strains is displayed in Figure 2. A resulting calculated cut-off distance between isolates of 0.161 was set by the software to discriminate clusters of B. cereus isolates. A total of 33 clusters were thus obtained with 18 clusters containing a single isolate, 8 clusters containing either 2 or 3 isolates (4 clusters each). The largest clusters held 5 (1 cluster), 7 (3 clusters), 8 (2 clusters) and 11 (1 cluster) isolates, respectively.

    B. cereus is a facultative anaerobic, Gram-positive, spore-forming opportunistic pathogen. It is a common cause of food poisoning, the most usual symptoms of which are diarrheal and emetic [8]. One of its prominent virulence features is its ability to produce various toxins. Surprisingly, while plenty of literature can be found on the prevalence and distribution of toxin genes in B. cereus strains isolated from cow milk, dairy products and powdered milks, very few works report on these parameters in HM isolates. In the only study found available on the subject, forty-nine isolates (including 3 strains recovered post-pasteurization) were tested for the occurrence of toxin genes [22].

    Figure 2.  2D plot of the first two components arising from the primary component analysis comparing the 90 B. cereus strains included in this study (3 FITR spectra acquired per strain). Color coding for B. cereus isolates as follows: Raw Human Milk; Pasteurized Human Milk; Endoscope; Linen; Human Milk Bank Room 1; Human Milk Bank Room 2; Human Milk Bank Room 3; Other hospital wards. Shape coding of the 7 B. cereus main (5 isolates and more) clusters as follows: Lower ○ Cluster 1 (5 isolates); Upper ○ Cluster 2 (8 isolates); + Cluster 3 (7 isolates); Upper▼Cluster 4 (7 isolates); Lower ▼Cluster 5 (7 isolates); ▲Cluster 6 (8 isolates); ● Cluster 7 (11 isolates).

    Similar to the high prevalence witnessed in our work, all three nhe genes were nearly systematically detected in all but one isolates by these authors. High prevalences ranging from 78 to 100% were also reported for nhe in B. cereus strains isolated from other raw/pasteurized milk sources and fit with the values found in this work [23][27]. As for our hbl and cytK prevalences, they also fell within the ranges previously described in milk isolated (2 to 90% and 46 to 73%, respectively) [23],[25][27]. Regarding entFM and bceT presence in milk isolates, prevalences were less widely investigated in the literature. While our high entFM prevalence was in accordance with the ones already described [24],[25], the 24% prevalence witnessed for bceT was lower than those previously published [25],[27]. Finally, ces prevalence was null or very low in most previous works reporting on milk isolates [22][24],[26] when ces prevalence was higher (27%) in the isolates investigated here. Only two studies reported high prevalences for ces (10 and 16%, respectively) [25],[27] but none as high as the one found in this work, which is worrisome and supports the current French policy of discarding pasteurized milk displaying a bacterial contamination of 2 Colony Forming Units per milliliter or above [28]. This discrepancy in observed ces prevalences might be due to the fact that, as very few B. cereus were identified pre-pasteurization in this study, most of the isolates came from pasteurized milk. Indeed, Radmehr et al. already noted that isolates harvested post-pasteurization displayed more virulence genes than raw milk ones [26]. It could also be due to the spread of a B. cereus clone carrying ces in APHMB donations and environment.

    As for using toxin gene profiles to discriminate between our isolates, no statistically significant difference in toxin prevalences could be highlighted between environmental and HM isolates investigated in this work. However, ces could once more be qualified as concerning as its prevalence was close to being statistically higher in HM isolates than the prevalence in environmental ones. Nevertheless, according to the classification previously described by Ehling-Schulz et al., no discriminating pattern between environmental and HM isolates could be identified [14]. However, it must be underscored that this classification did not take into account strains carrying nhe, hbl, cytK and ces, nor did it include entFM. As a consequence, 13 of our isolates could not be assigned to a given toxin gene profile. We therefore sought another possibility to discriminate between our isolates, such as FTIR spectroscopy.

    FTIR spectroscopy has previously been described as being able to discriminate clones between strains of a given species in a time and cost-efficient manner both for Gram-negative and Gram-negative bacteria [15],[29][30]. It has also been reported as a helpful tool in discriminating B. cereus isolates [26]. The golden standard for discriminating bacterial isolates within a species is the whole genome sequencing (WGS) technique [15]. This method has allowed to drastically improve the discriminatory power over other molecular-based typing methods such as 16S rDNA sequencing, Pulse Field Gel Electrophoresis (PFGE) or Multilocus Sequence Typing (MLST) [15]. However, these molecular-based typing methods are universally recognized as time-consuming, costly and labor-intensive [31]. Several studies have described the interest of spectroscopic/spectrometric-based methods such as Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and FTIR in typing bacterial isolates. A major advantage of these techniques is that, once the apparatus is present in a facility, they are easily integrated in the laboratory daily workflow. While MALDI-TOF MS is now routinely used for species identification in medical laboratories, its discrimination power when it comes to clonality assessment has been questioned [31],[32]. The choice of using FITR was therefore made for this work.

    Applied to our panel of isolates, FTIR showed that around 42% (38 isolates out of 90) were classified in clusters consisting of a single or 2 to 3 isolates. Hence, no common origin on which a corrective action might be implemented to mitigate the spread of B. cereus in pasteurized HM could be identified for those isolates. Nevertheless, more than half of our isolates were grouped in seven clusters of 5 or more isolates (Figure 2). Two of those (Clusters 4 & 6) were only constituted by HM isolates while in the five remaining ones, at least one isolate arose from the hospital environment. Interestingly, in clusters 2 (8 isolates) and 7 (11 isolates), isolates from environmental samples belonging to APHMB rooms were found to aggregate with HM ones. When a closer look was taken at the time sequence in which clusters 2 & 7 isolates were recovered, we found in cluster 7 that the only B. cereus strain isolated in raw HM was the first one to be recovered along with its post-pasteurization counterpart and in cluster 2, it was a pasteurized HM isolate. According to FTIR spectra analysis, similar isolates were thereafter harvested from all 3 APHBM rooms in cluster 7 and 2 out of the three APHBM rooms in cluster 2, pointing out that an environmental contamination by the HM donation of one mother could have spread to the environment and other HM donations. It is noteworthy that the first post-pasteurization isolate in cluster 7 was positive for ces but only one environmental isolate in the same cluster carried ces. Also of interest is the fact that this link between an environmental isolate and some HM ones found in cluster 7 was also highlighted in a previous analysis using rep-PCR [3]. However, this latter method is less suited to a routine determination of isolates' proximity as it is much more time-consuming and costly as well as less discriminating than FTIR spectroscopy. In our case, a closer monitoring of the environment and systematic real-time use of FTIR spectroscopy on B. cereus isolates could have helped in limiting the spread of this clone in APHBM as well as the one identified in cluster 2 for which a similar pattern. It might therefore be interesting to implement such a monitoring routinely. Comparing the results obtained through this technique with those of WGS would also help strengthening the daily use of FITR if a good correlation for clonality assessment is found.

    This study is one of the first papers reporting on toxin gene prevalences in HM B. cereus isolates and highlights that about one fourth of those are expressing cereulide toxin, which could be worrisome from a public health point of view and call for preventative measures. It is also one of the first reports on B. cereus clustering through FTIR Spectroscopy, which proved efficient, time and cost-effective. The use of this technique pointed out that some HM B. cereus isolates clustered with APHMB environmental isolates. This is an indication that mitigation measures such as a thorough cleaning procedure following several occurrences of B. cereus isolation in a short time span and/or B. cereus detection in routine environmental samples could be implemented to reduce the isolation of B. cereus in HM donations and the rejection of thusly contaminated donations.



    [1] F. Jiang, Y. Jiang, H. Zhi, Y. Dong, H. Li, S. Ma, et al., Artificial intelligence in healthcare: past, present and future, Stroke Vasc. Neurol., 2 (2017). https://doi.org/10.1136/svn-2017-000101 doi: 10.1136/svn-2017-000101
    [2] M. D. McCradden, E. A. Stephenson, J. A. Anderson, Clinical research underlies ethical integration of healthcare artificial intelligence, Nat. Med., 26 (2020), 1325-1326. https://doi.org/10.1038/s41591-020-1035-9 doi: 10.1038/s41591-020-1035-9
    [3] K. H. Yu, A. L. Beam, I. S. Kohane, Artificial intelligence in healthcare, Nat. Biomed. Eng., 2 (2018), 719-731. https://doi.org/10.1038/s41551-018-0305-z doi: 10.1038/s41551-018-0305-z
    [4] O. Asan, A. E. Bayrak, A. Choudhury, Artificial intelligence and human trust in healthcare: focus on clinicians, J. Med. Internet Res., 22 (2020), e15154. https://doi.org/10.2196/15154 doi: 10.2196/15154
    [5] D. Shen, G. Wu, H. I. Suk, Deep learning in medical image analysis, Annu. Rev. Biomed. Eng., 19 (2017), 221-248. https://doi.org/10.1146/annurev-bioeng-071516-044442 doi: 10.1146/annurev-bioeng-071516-044442
    [6] G. Litjens, T. Kooi, B. E. Bejnordi, A. A. A. Setio, F. Ciompi, M. Ghafoorian, et al., A survey on deep learning in medical image analysis, Med. Image Anal., 42 (2017), 60-88. https://doi.org/10.1016/j.media.2017.07.005 doi: 10.1016/j.media.2017.07.005
    [7] M. I. Razzak, S. Naz, A. Zaib, Deep learning for medical image processing: Overview, challenges and the future, Classif. BioApps, (2018), 323-350.
    [8] J. H. Thrall, D. Fessell, P. V. Pandharipande, Rethinking the approach to artificial intelligence for medical image analysis: the case for precision diagnosis, J. Am. Coll. Radiol., 18 (2021), 174-179. https://doi.org/10.1016/j.jacr.2020.07.010 doi: 10.1016/j.jacr.2020.07.010
    [9] Y. Zhang, Z. Wang, J. Zhang, C. Wang, Y. Wang, H. Chen, et al., Deep learning model for classifying endometrial lesions, J. Transl. Med., 19 (2021), 1-13. https://doi.org/10.1186/s12967-020-02660-x doi: 10.1186/s12967-020-02660-x
    [10] C. Zheng, L. Chen, J. Jian, J. Li, Z. Gao, Efficacy evaluation of interventional therapy for primary liver cancer using magnetic resonance imaging and CT scanning under deep learning and treatment of vasovagal reflex, J. Supercomput., 77 (2021), 7535-7548. https://doi.org/10.1007/s11227-020-03539-w doi: 10.1007/s11227-020-03539-w
    [11] G. A. Roth, C. Johnson, A. Abajobir, F. Abd-Allah, S. F. Abera, G. Abyu, et al., Global, regional, and national burden of cardiovascular diseases for 10 causes, 1990 to 2015, J. Am. Coll. Cardiol., 70 (2017), 1-25.
    [12] K. H. Miao, J. H. Miao, Coronary heart disease diagnosis using deep neural networks, Int. J. Adv. Comput. Sci. Appl., 9 (2018), 1-8. https://doi.org/10.14569/IJACSA.2018.091001 doi: 10.14569/IJACSA.2018.091001
    [13] A. Gupta, H. S. Arora, R. Kumar, B. Raman, DMHZ: a decision support system based on machine computational design for heart disease diagnosis using z-alizadeh sani dataset, in 2021 International Conference on Information Networking (ICOIN), (2021), 818-823. https://doi.org/10.1109/ICOIN50884.2021.9333884
    [14] A. D. Villa, E. Sammut, A. Nair, R. Rajani, R. Bonamini, A. Chiribiri, Coronary artery anomalies overview: the normal and the abnormal, World J. Radiol., 8 (2016), 537. https://doi.org/10.4329/wjr.v8.i6.537 doi: 10.4329/wjr.v8.i6.537
    [15] R. Alizadehsani, M. Abdar, M. Roshanzamir, A. Khosravi, P. M. Kebria, F. Khozeimeh, et al., Machine learning-based coronary artery disease diagnosis: A comprehensive review, Comput. Biol. Med., 111 (2019), 103346. https://doi.org/10.1016/j.compbiomed.2019.103346 doi: 10.1016/j.compbiomed.2019.103346
    [16] T. M. Williamson, C. Moran, A. McLennan, S. Seidel, P. P. Ma, M. L. Koerner, T. S. Campbell, Promoting adherence to physical activity among individuals with cardiovascular disease using behavioral counseling: A theory and research-based primer for health care professionals, Prog. Cardiovasc. Dis., (2020). https://doi.org/10.1016/j.pcad.2020.12.007
    [17] J. H. Joloudari, E. H. Joloudari, H. Saadatfar, M. Ghasemigol, S. M. Razavi, A. Mosavi, et al., Coronary artery disease diagnosis; ranking the significant features using a random trees model, Int. J. Environ. Res. Public Health, 17 (2020), 731. https://doi.org/10.3390/ijerph17030731 doi: 10.3390/ijerph17030731
    [18] M. V. Dyke, S. Greer, E. Odom, L. Schieb, A. Vaughan, M. Kramer, et al., Heart disease death rates among blacks and whites aged ≥ 35 years-United States, 1968–2015, MMWR Surveillance Summaries, 67 (2018), 1. https://doi.org/10.15585/mmwr.ss6705a1
    [19] D. Mozaffarian, E. J. Benjamin, A. S. Go, D. K. Arnett, M. J. Blaha, M. Cushman, et al., Heart disease and stroke statistics—2015 update: a report from the American heart association, Circulation, 131 (2015), e29-e322.
    [20] E. J. Benjamin, S. S. Virani, C. W. Callaway, A. M. Chamberlain, A. R. Chang, S. Cheng, et al., Heart disease and stroke statistics—2018 update: a report from the American heart association, Circulation, 137 (2018), e67-e492. https://doi.org/10.1161/CIR.0000000000000573 doi: 10.1161/CIR.0000000000000573
    [21] H. Larochelle, Y. Bengio, J. Louradour, P. Lamblin, Exploring strategies for training deep neural networks, J. Mach. Learn. Res., 10 (2009).
    [22] R. O. Bonow, D. L. Mann, D. P. Zipes, P. Libby, Braunwald's Heart Disease E-Book: A Textbook of Cardiovascular Medicine, Elsevier Health Sciences, 2011.
    [23] E.G. Nabel, E. Braunwald, A tale of coronary artery disease and myocardial infarction, New Engl. J. Med., 366 (2012), 54-63. https://doi.org/10.1056/NEJMra1112570 doi: 10.1056/NEJMra1112570
    [24] İ. Babaoglu, O. Findik, E. Ülker, A comparison of feature selection models utilizing binary particle swarm optimization and genetic algorithm in determining coronary artery disease using support vector machine, Expert Syst. Appl., 37 (2010), 3177-3183. https://doi.org/10.1016/j.eswa.2009.09.064 doi: 10.1016/j.eswa.2009.09.064
    [25] M. Kumar, R. B. Pachori, U. R. Acharya, Characterization of coronary artery disease using flexible analytic wavelet transform applied on ECG signals, Biomed. Signal Proces. Control, 31 (2017), 301-308. https://doi.org/10.1016/j.bspc.2016.08.018 doi: 10.1016/j.bspc.2016.08.018
    [26] R. Alizadehsani, J. Habibi, M. J. Hosseini, R. Boghrati, A. Ghandeharioun, B. Bahadorian, et al., Diagnosis of coronary artery disease using data mining techniques based on symptoms and ecg features, Eur. J. Scientific Res., 82 (2012), 542-553.
    [27] R. Alizadehsani, J. Habibi, M. J. Hosseini, H. Mashayekhi, R. Boghrati, A. Ghandeharioun, et al., A data mining approach for diagnosis of coronary artery disease, Comput. Meth. Prog. Bio., 111(2013), 52-61. https://doi.org/10.1016/j.cmpb.2013.03.004 doi: 10.1016/j.cmpb.2013.03.004
    [28] R. Alizadehsani, J. Habibi, Z. A. Sani, H. Mashayekhi, R. Boghrati, A. Ghandeharioun, et al., Diagnosing coronary artery disease via data mining algorithms by considering laboratory and echocardiography features, Res. Cardiov. Med., 2 (2013), 133. https://doi.org/10.5812/cardiovascmed.10888 doi: 10.5812/cardiovascmed.10888
    [29] R. Alizadehsani, M. H. Zangooei, M. J. Hosseini, J. Habibi, A. Khosravi, M. Roshanzamir, et al., Coronary artery disease detection using computational intelligence methods, Knowledge-Based Syst., 109 (2016), 187-197. https://doi.org/10.1016/j.knosys.2016.07.004 doi: 10.1016/j.knosys.2016.07.004
    [30] A. D. Dolatabadi, S. E. Z. Khadem, B. M. Asl, Automated diagnosis of coronary artery disease (CAD) patients using optimized SVM, Comput. Meth. Prog. Bio., 138 (2017), 117-126. https://doi.org/10.1016/j.cmpb.2016.10.011 doi: 10.1016/j.cmpb.2016.10.011
    [31] Z. Arabasadi, R. Alizadehsani, M. Roshanzamir, H. Moosaei, A. A. Yarifard, et al., Computer aided decision making for heart disease detection using hybrid neural network-Genetic algorithm, Comput. Meth. Prog. Bio., 141 (2017), 19-26. https://doi.org/10.1016/j.cmpb.2017.01.004 doi: 10.1016/j.cmpb.2017.01.004
    [32] R. Alizadehsani, M. J. Hosseini, A. Khosravi, F. Khozeimeh, M. Roshanzamir, N. Sarrafzadegan, et al., Non-invasive detection of coronary artery disease in high-risk patients based on the stenosis prediction of separate coronary arteries, Comput. Meth. Prog. Bio., 162 (2018), 119-127. https://doi.org/10.1016/j.cmpb.2018.05.009 doi: 10.1016/j.cmpb.2018.05.009
    [33] M. Abdar, W. Książek, U. R. Acharya, R. S. Tan, V. Makarenkov, P. Pławiak, A new machine learning technique for an accurate diagnosis of coronary artery disease, Comput. Meth. Prog. Bio., 179 (2019), 104992. https://doi.org/10.1016/j.cmpb.2019.104992 doi: 10.1016/j.cmpb.2019.104992
    [34] C. Blake, UCI Repository of Machine Learning Databases, 1998. Available from: http://www.ics.uci.edu/~mlearn/MLRepository.html.
    [35] R. W. Hamersvelt, M. Zreik, M. Voskuil, M. A. Viergever, I. Išgum, T. Leiner, et al., Deep learning analysis of left ventricular myocardium in CT angiographic intermediate-degree coronary stenosis improves the diagnostic accuracy for identification of functionally significant stenosis, Eur. Rad., 29 (2019), 2350-2359. https://doi.org/10.1007/s00330-018-5822-3 doi: 10.1007/s00330-018-5822-3
    [36] U. R. Acharya, H. Fujita, O. S. Lih, M. Adam, J. H. Tan, C. K. Chua, Automated detection of coronary artery disease using different durations of ECG segments with convolutional neural network, Knowledge-Based Syst., 132 (2017), 62-71. https://doi.org/10.1016/j.knosys.2017.06.003 doi: 10.1016/j.knosys.2017.06.003
    [37] A. L. Goldberger, L. A. Amaral, L. Glass, J. M. Hausdorff, P. C. Ivanov, R. G. Mark, et al., PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals, Circulation, 101 (2000), e215-e220. https://doi.org/10.1161/01.CIR.101.23.e215 doi: 10.1161/01.CIR.101.23.e215
    [38] J. H. Tan, Y. Hagiwara, W. Pang, I. Lim, S. L. Oh, M. Adam, et al., Application of stacked convolutional and long short-term memory network for accurate identification of CAD ECG signals, Comput. Biol. Med., 94 (2018), 19-26. https://doi.org/10.1016/j.compbiomed.2017.12.023 doi: 10.1016/j.compbiomed.2017.12.023
    [39] U. R. Acharya, H. Fujita, M. Adam, O. S. Lih, V. K. Sudarshan, T. J. Hong, et al., Automated characterization and classification of coronary artery disease and myocardial infarction by decomposition of ECG signals: A comparative study, Inform. Sci., 377 (2017), 17-29. https://doi.org/10.1016/j.ins.2016.10.013 doi: 10.1016/j.ins.2016.10.013
    [40] U. R. Acharya, H. Fujita, S. L. Oh, Y. Hagiwara, J. H. Tan, M. Adam, et al., Deep convolutional neural network for the automated diagnosis of congestive heart failure using ECG signals, Appl. Intell., 49 (2019), 16-27. https://doi.org/10.1007/s10489-018-1179-1 doi: 10.1007/s10489-018-1179-1
    [41] M. M. Ghiasi, S. Zendehboudi, A. A. Mohsenipour, Decision tree-based diagnosis of coronary artery disease: CART model, Comput. Meth. Prog. Bio., 192 (2020), 105400. https://doi.org/10.1016/j.cmpb.2020.105400 doi: 10.1016/j.cmpb.2020.105400
    [42] L. Verma, S. Srivastava, P. Negi, A hybrid data mining model to predict coronary artery disease cases using non-invasive clinical data, J. Med. Syst., 40 (2016), 178. https://doi.org/10.1007/s10916-016-0536-z doi: 10.1007/s10916-016-0536-z
    [43] N. M. Idris, Y. K. Chiam, K. D. Varathan, W. A. W. Ahmad, K. H. Chee, Y. M. Liew, Feature selection and risk prediction for patients with coronary artery disease using data mining, Med. Biol. Eng. Comput., 58 (2020), 3123-3140. https://doi.org/10.1007/s11517-020-02268-9 doi: 10.1007/s11517-020-02268-9
    [44] D, . Velusamy, K. Ramasamy, Ensemble of heterogeneous classifiers for diagnosis and prediction of coronary artery disease with reduced feature subset, Comput. Meth. Prog. Bio., 198 (2020), 105770. https://doi.org/10.1016/j.cmpb.2020.105770 doi: 10.1016/j.cmpb.2020.105770
    [45] I. Goodfellow, D. Warde-Farley, M. Mirza, A. Courville, Y. Bengio, Maxout networks, in International Conference On Machine Learning, PMLR, (2013), 1319-1327.
    [46] J. H. Joloudari, M. Haderbadi, A. Mashmool, M. GhasemiGol, S. S. Band, A. Mosavi, Early detection of the advanced persistent threat attack using performance analysis of deep learning, IEEE Access, 8 (2020), 186125-186137. https://doi.org/10.1109/ACCESS.2020.3029202 doi: 10.1109/ACCESS.2020.3029202
    [47] Y. Ito, Approximation of functions on a compact set by finite sums of a sigmoid function without scaling, Neural Networks, 4 (1991), 817-826. https://doi.org/10.1016/0893-6080(91)90060-I doi: 10.1016/0893-6080(91)90060-I
    [48] N. Hassan, N. Akamatsu, A new approach for contrast enhancement using sigmoid function, Inter Arab J. Inf. Techn., 1 (2004).
    [49] X. Li, X. Zhang, W. Huang, Q. Wang, Truncation cross entropy loss for remote sensing image captioning, IEEE Transactions Geosci. Remote Sens., 59 (2020), 5246-5257. https://doi.org/10.1109/TGRS.2020.3010106 doi: 10.1109/TGRS.2020.3010106
    [50] C. Otto, D. Wang, A. K. Jain, Clustering millions of faces by identity, IEEE Transactions Pattern Anal. Mach. Intell., 40 (2017), 289-303. https://doi.org/10.1109/TPAMI.2017.2679100 doi: 10.1109/TPAMI.2017.2679100
    [51] E. H. Ruspini, A new approach to clustering, Inform. Control, 15 (1969), 22-32. https://doi.org/10.1016/S0019-9958(69)90591-9 doi: 10.1016/S0019-9958(69)90591-9
    [52] R. O. Duda, P. E. Hart, Hart PE Pattern Classification And Scene Analysis, New York: Wiley, 1973.
    [53] R. Veloso, F. Portela, M. F. Santos, A. Silva, F. Rua, A. Abelha, et al., A clustering approach for predicting readmissions in intensive medicine, Procedia Technol., 16 (2014), 1307-1316. https://doi.org/10.1016/j.protcy.2014.10.147 doi: 10.1016/j.protcy.2014.10.147
    [54] H. S. Park, C. H. Jun, A simple and fast algorithm for K-medoids clustering, Expert Syst. Appl., 36 (2009), 3336-3341. https://doi.org/10.1016/j.eswa.2008.01.039 doi: 10.1016/j.eswa.2008.01.039
    [55] R. O. Duda, P. E. Hart, Pattern Classification And Scene Analysis, Wiley New York, 1973.
    [56] J. C. Dunn, Well-separated clusters and optimal fuzzy partitions, J. Cybern., 4 (1974), 95-104. https://doi.org/10.1080/01969727408546059 doi: 10.1080/01969727408546059
    [57] J. C. Bezdek, Objective function clustering, in Pattern Recognition With Fuzzy Objective Function Algorithms, Springer, (1981), 43-93. https://doi.org/10.1007/978-1-4757-0450-1_3
    [58] M. S. Yang, A survey of fuzzy clustering, Math. Comput. Model., 18 (1993), 1-16. https://doi.org/10.1016/0895-7177(93)90202-A doi: 10.1016/0895-7177(93)90202-A
    [59] G. Govaert, M. Nadif, Clustering with block mixture models, Pattern Recogn., 36 (2003), 463-473. https://doi.org/10.1016/S0031-3203(02)00074-2 doi: 10.1016/S0031-3203(02)00074-2
    [60] S. Bandyopadhyay, U. Maulik, A. Mukhopadhyay, Multiobjective genetic clustering for pixel classification in remote sensing imagery, IEEE Trans. Geosci. Remote Sens., 45 (2007), 1506-1511. https://doi.org/10.1109/TGRS.2007.892604 doi: 10.1109/TGRS.2007.892604
    [61] R. Xu, D. Wunsch, Survey of clustering algorithms, IEEE Trans. Neural Networks, 16 (2005), 645-678. https://doi.org/10.1109/TNN.2005.845141 doi: 10.1109/TNN.2005.845141
    [62] J. H. Joloudari, H. Saadatfar, A. Dehzangi, S. Shamshirband, Computer-aided decision-making for predicting liver disease using PSO-based optimized SVM with feature selection, Inform. Medicine Unlocked, 17 (2019), 100255. https://doi.org/10.1016/j.imu.2019.100255 doi: 10.1016/j.imu.2019.100255
    [63] M. Abdar, M. Zomorodi-Moghadam, R. Das, I. H. Ting, Performance analysis of classification algorithms on early detection of liver disease, Expert Syst. Appl., 67 (2017), 239-251. https://doi.org/10.1016/j.eswa.2016.08.065 doi: 10.1016/j.eswa.2016.08.065
    [64] C. H. Weng, C. K. Huang, R. P. Han, Disease prediction with different types of neural network classifiers, Telemat. Inform., 33 (2016), 277-292. https://doi.org/10.1016/j.tele.2015.08.006 doi: 10.1016/j.tele.2015.08.006
    [65] M. Diwakar, A. Tripathi, K. Joshi, M. Memoria, P. Singh, Latest trends on heart disease prediction using machine learning and image fusion, Materials Today: Proceed., 37 (2021), 3213-3218. https://doi.org/10.1016/j.matpr.2020.09.078 doi: 10.1016/j.matpr.2020.09.078
    [66] J. H. Moon, W. C. Cha, M. J. Chung, K. S. Lee, B. H. Cho, J. H. Choi, Automatic stenosis recognition from coronary angiography using convolutional neural networks, Comput. Meth. Prog. Bio., 198 (2021), 105819. https://doi.org/10.1016/j.cmpb.2020.105819 doi: 10.1016/j.cmpb.2020.105819
  • Reader Comments
  • © 2022 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(3808) PDF downloads(133) Cited by(8)

Figures and Tables

Figures(13)  /  Tables(8)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog