Deep Learning for Detection of Complete Anterior Cruciate Ligament Tear View Full Text


Ontology type: schema:ScholarlyArticle     


Article Info

DATE

2019-03-11

AUTHORS

Peter D. Chang, Tony T. Wong, Michael J. Rasiej

ABSTRACT

Deep learning for MRI detection of sports injuries poses unique challenges. To address these difficulties, this study examines the feasibility and incremental benefit of several customized network architectures in evaluation of complete anterior cruciate ligament (ACL) tears. Two hundred sixty patients, ages 18-40, were identified in a retrospective review of knee MRIs obtained from September 2013 to March 2016. Half of the cases demonstrated a complete ACL tear (624 slices), the other half a normal ACL (3520 slices). Two hundred cases were used for training and validation, and the remaining 60 cases as an independent test set. For each exam with an ACL tear, coronal proton density non-fat suppressed sequence was manually annotated to delineate: (1) a bounding-box around the cruciate ligaments; (2) slices containing the tear. Multiple convolutional neural network (CNN) architectures were implemented including variations in input field-of-view and dimensionality. For single-slice CNN architectures, validation accuracy of a dynamic patch-based sampling algorithm (0.765) outperformed both cropped slice (0.720) and full slice (0.680) strategies. Using the dynamic patch-based sampling algorithm as a baseline, a five-slice CNN input (0.915) outperformed both three-slice (0.865) and single-slice (0.765) inputs. The final highest performing five-slice dynamic patch-based sampling algorithm resulted in independent test set AUC, sensitivity, specificity, PPV, and NPV of 0.971, 0.967, 1.00, 0.938, and 1.00. A customized 3D deep learning architecture based on dynamic patch-based sampling demonstrates high performance in detection of complete ACL tears with over 96% test set accuracy. A cropped field-of-view and 3D inputs are critical for high algorithm performance. More... »

PAGES

1-7

References to SciGraph publications

  • 2015. U-Net: Convolutional Networks for Biomedical Image Segmentation in MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION – MICCAI 2015
  • Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/s10278-019-00193-4

    DOI

    http://dx.doi.org/10.1007/s10278-019-00193-4

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1112683236

    PUBMED

    https://www.ncbi.nlm.nih.gov/pubmed/30859341


    Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
    Incoming Citations Browse incoming citations for this publication using opencitations.net

    JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Artificial Intelligence and Image Processing", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Information and Computing Sciences", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "University of California, Irvine Medical Center", 
              "id": "https://www.grid.ac/institutes/grid.417319.9", 
              "name": [
                "Center for Artificial Intelligence in Diagnostic Medicine, University of California Irvine Medical Center, 101 The City Drive South, Building 55, Suite 201, 92868, Orange, CA, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Chang", 
            "givenName": "Peter D.", 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Columbia University", 
              "id": "https://www.grid.ac/institutes/grid.21729.3f", 
              "name": [
                "Department of Radiology, Columbia University Irving Medical Center, 622 West 168th St., MC 28, 10032, New York, NY, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Wong", 
            "givenName": "Tony T.", 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Columbia University", 
              "id": "https://www.grid.ac/institutes/grid.21729.3f", 
              "name": [
                "Department of Radiology, Columbia University Irving Medical Center, 622 West 168th St., MC 28, 10032, New York, NY, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Rasiej", 
            "givenName": "Michael J.", 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "https://doi.org/10.1136/bjsm.2004.010900", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1000039211"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/j.jsams.2008.07.005", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1008134660"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.3109/17453679009006504", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1008199874"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-24574-4_28", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1017774818", 
              "https://doi.org/10.1007/978-3-319-24574-4_28"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/j.cmpb.2016.12.006", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1039502687"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/j.knee.2008.11.004", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1046258888"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1302/0301-620x.92b3.22424", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1046902699"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1177/0363546508330136", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1063821339"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1177/0363546508330136", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1063821339"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1177/0363546516629944", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1063822946"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1177/0363546516629944", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1063822946"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1177/0363546516629944", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1063822946"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1148/radiology.167.3.3363138", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1079704860"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1302/0301-620x.68b3.3755441", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1079934530"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/j.jacr.2017.02.019", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1084083884"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1002/jor.23632", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1085993094"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/cvpr.2016.90", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093359587"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/iccv.2015.123", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093828312"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2019-03-11", 
        "datePublishedReg": "2019-03-11", 
        "description": "Deep learning for MRI detection of sports injuries poses unique challenges. To address these difficulties, this study examines the feasibility and incremental benefit of several customized network architectures in evaluation of complete anterior cruciate ligament (ACL) tears. Two hundred sixty patients, ages 18-40, were identified in a retrospective review of knee MRIs obtained from September 2013 to March 2016. Half of the cases demonstrated a complete ACL tear (624 slices), the other half a normal ACL (3520 slices). Two hundred cases were used for training and validation, and the remaining 60 cases as an independent test set. For each exam with an ACL tear, coronal proton density non-fat suppressed sequence was manually annotated to delineate: (1) a bounding-box around the cruciate ligaments; (2) slices containing the tear. Multiple convolutional neural network (CNN) architectures were implemented including variations in input field-of-view and dimensionality. For single-slice CNN architectures, validation accuracy of a dynamic patch-based sampling algorithm (0.765) outperformed both cropped slice (0.720) and full slice (0.680) strategies. Using the dynamic patch-based sampling algorithm as a baseline, a five-slice CNN input (0.915) outperformed both three-slice (0.865) and single-slice (0.765) inputs. The final highest performing five-slice dynamic patch-based sampling algorithm resulted in independent test set AUC, sensitivity, specificity, PPV, and NPV of 0.971, 0.967, 1.00, 0.938, and 1.00. A customized 3D deep learning architecture based on dynamic patch-based sampling demonstrates high performance in detection of complete ACL tears with over 96% test set accuracy. A cropped field-of-view and 3D inputs are critical for high algorithm performance.", 
        "genre": "research_article", 
        "id": "sg:pub.10.1007/s10278-019-00193-4", 
        "inLanguage": [
          "en"
        ], 
        "isAccessibleForFree": false, 
        "isPartOf": [
          {
            "id": "sg:journal.1100894", 
            "issn": [
              "0897-1889", 
              "1618-727X"
            ], 
            "name": "Journal of Digital Imaging", 
            "type": "Periodical"
          }
        ], 
        "name": "Deep Learning for Detection of Complete Anterior Cruciate Ligament Tear", 
        "pagination": "1-7", 
        "productId": [
          {
            "name": "readcube_id", 
            "type": "PropertyValue", 
            "value": [
              "a10eb25d6af1b9387c1d53ce4ea80fd91c780a210305d145c038b98e96242c24"
            ]
          }, 
          {
            "name": "pubmed_id", 
            "type": "PropertyValue", 
            "value": [
              "30859341"
            ]
          }, 
          {
            "name": "nlm_unique_id", 
            "type": "PropertyValue", 
            "value": [
              "9100529"
            ]
          }, 
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/s10278-019-00193-4"
            ]
          }, 
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1112683236"
            ]
          }
        ], 
        "sameAs": [
          "https://doi.org/10.1007/s10278-019-00193-4", 
          "https://app.dimensions.ai/details/publication/pub.1112683236"
        ], 
        "sdDataset": "articles", 
        "sdDatePublished": "2019-04-11T11:31", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-uberresearch-data-dimensions-target-20181106-alternative/cleanup/v134/2549eaecd7973599484d7c17b260dba0a4ecb94b/merge/v9/a6c9fde33151104705d4d7ff012ea9563521a3ce/jats-lookup/v90/0000000357_0000000357/records_99319_00000002.jsonl", 
        "type": "ScholarlyArticle", 
        "url": "https://link.springer.com/10.1007%2Fs10278-019-00193-4"
      }
    ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/s10278-019-00193-4'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/s10278-019-00193-4'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/s10278-019-00193-4'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/s10278-019-00193-4'


     

    This table displays all metadata directly associated to this object as RDF triples.

    123 TRIPLES      21 PREDICATES      41 URIs      18 LITERALS      7 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/s10278-019-00193-4 schema:about anzsrc-for:08
    2 anzsrc-for:0801
    3 schema:author N9020dd626dcc4280a99ccfb583382c24
    4 schema:citation sg:pub.10.1007/978-3-319-24574-4_28
    5 https://doi.org/10.1002/jor.23632
    6 https://doi.org/10.1016/j.cmpb.2016.12.006
    7 https://doi.org/10.1016/j.jacr.2017.02.019
    8 https://doi.org/10.1016/j.jsams.2008.07.005
    9 https://doi.org/10.1016/j.knee.2008.11.004
    10 https://doi.org/10.1109/cvpr.2016.90
    11 https://doi.org/10.1109/iccv.2015.123
    12 https://doi.org/10.1136/bjsm.2004.010900
    13 https://doi.org/10.1148/radiology.167.3.3363138
    14 https://doi.org/10.1177/0363546508330136
    15 https://doi.org/10.1177/0363546516629944
    16 https://doi.org/10.1302/0301-620x.68b3.3755441
    17 https://doi.org/10.1302/0301-620x.92b3.22424
    18 https://doi.org/10.3109/17453679009006504
    19 schema:datePublished 2019-03-11
    20 schema:datePublishedReg 2019-03-11
    21 schema:description Deep learning for MRI detection of sports injuries poses unique challenges. To address these difficulties, this study examines the feasibility and incremental benefit of several customized network architectures in evaluation of complete anterior cruciate ligament (ACL) tears. Two hundred sixty patients, ages 18-40, were identified in a retrospective review of knee MRIs obtained from September 2013 to March 2016. Half of the cases demonstrated a complete ACL tear (624 slices), the other half a normal ACL (3520 slices). Two hundred cases were used for training and validation, and the remaining 60 cases as an independent test set. For each exam with an ACL tear, coronal proton density non-fat suppressed sequence was manually annotated to delineate: (1) a bounding-box around the cruciate ligaments; (2) slices containing the tear. Multiple convolutional neural network (CNN) architectures were implemented including variations in input field-of-view and dimensionality. For single-slice CNN architectures, validation accuracy of a dynamic patch-based sampling algorithm (0.765) outperformed both cropped slice (0.720) and full slice (0.680) strategies. Using the dynamic patch-based sampling algorithm as a baseline, a five-slice CNN input (0.915) outperformed both three-slice (0.865) and single-slice (0.765) inputs. The final highest performing five-slice dynamic patch-based sampling algorithm resulted in independent test set AUC, sensitivity, specificity, PPV, and NPV of 0.971, 0.967, 1.00, 0.938, and 1.00. A customized 3D deep learning architecture based on dynamic patch-based sampling demonstrates high performance in detection of complete ACL tears with over 96% test set accuracy. A cropped field-of-view and 3D inputs are critical for high algorithm performance.
    22 schema:genre research_article
    23 schema:inLanguage en
    24 schema:isAccessibleForFree false
    25 schema:isPartOf sg:journal.1100894
    26 schema:name Deep Learning for Detection of Complete Anterior Cruciate Ligament Tear
    27 schema:pagination 1-7
    28 schema:productId N5e8527e0852d4cd3bb1c47b1ab3b883d
    29 N798775af7dd84a2582b5321c4440d835
    30 Nb3c21eb3ff4e484ea7bd6160bb37d144
    31 Nbf684018c29c45bdab16fa7c1df72c31
    32 Nd84dddc97be84499a4395e81f9225c4d
    33 schema:sameAs https://app.dimensions.ai/details/publication/pub.1112683236
    34 https://doi.org/10.1007/s10278-019-00193-4
    35 schema:sdDatePublished 2019-04-11T11:31
    36 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    37 schema:sdPublisher Nd3bfb580bd6d431e9d770f789c470361
    38 schema:url https://link.springer.com/10.1007%2Fs10278-019-00193-4
    39 sgo:license sg:explorer/license/
    40 sgo:sdDataset articles
    41 rdf:type schema:ScholarlyArticle
    42 N5e8527e0852d4cd3bb1c47b1ab3b883d schema:name doi
    43 schema:value 10.1007/s10278-019-00193-4
    44 rdf:type schema:PropertyValue
    45 N675ced31cf3442458f82186268aa211a schema:affiliation https://www.grid.ac/institutes/grid.21729.3f
    46 schema:familyName Wong
    47 schema:givenName Tony T.
    48 rdf:type schema:Person
    49 N73cb8391849d4c078f70051d4f92632f schema:affiliation https://www.grid.ac/institutes/grid.417319.9
    50 schema:familyName Chang
    51 schema:givenName Peter D.
    52 rdf:type schema:Person
    53 N7798fbb6744a426985c8dc329fad201c rdf:first N675ced31cf3442458f82186268aa211a
    54 rdf:rest N86e1ffd283074cda8dd36df77b2aca35
    55 N798775af7dd84a2582b5321c4440d835 schema:name readcube_id
    56 schema:value a10eb25d6af1b9387c1d53ce4ea80fd91c780a210305d145c038b98e96242c24
    57 rdf:type schema:PropertyValue
    58 N86e1ffd283074cda8dd36df77b2aca35 rdf:first N8c5c49b9406d46f395d9507fe4d2a739
    59 rdf:rest rdf:nil
    60 N8c5c49b9406d46f395d9507fe4d2a739 schema:affiliation https://www.grid.ac/institutes/grid.21729.3f
    61 schema:familyName Rasiej
    62 schema:givenName Michael J.
    63 rdf:type schema:Person
    64 N9020dd626dcc4280a99ccfb583382c24 rdf:first N73cb8391849d4c078f70051d4f92632f
    65 rdf:rest N7798fbb6744a426985c8dc329fad201c
    66 Nb3c21eb3ff4e484ea7bd6160bb37d144 schema:name pubmed_id
    67 schema:value 30859341
    68 rdf:type schema:PropertyValue
    69 Nbf684018c29c45bdab16fa7c1df72c31 schema:name dimensions_id
    70 schema:value pub.1112683236
    71 rdf:type schema:PropertyValue
    72 Nd3bfb580bd6d431e9d770f789c470361 schema:name Springer Nature - SN SciGraph project
    73 rdf:type schema:Organization
    74 Nd84dddc97be84499a4395e81f9225c4d schema:name nlm_unique_id
    75 schema:value 9100529
    76 rdf:type schema:PropertyValue
    77 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
    78 schema:name Information and Computing Sciences
    79 rdf:type schema:DefinedTerm
    80 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
    81 schema:name Artificial Intelligence and Image Processing
    82 rdf:type schema:DefinedTerm
    83 sg:journal.1100894 schema:issn 0897-1889
    84 1618-727X
    85 schema:name Journal of Digital Imaging
    86 rdf:type schema:Periodical
    87 sg:pub.10.1007/978-3-319-24574-4_28 schema:sameAs https://app.dimensions.ai/details/publication/pub.1017774818
    88 https://doi.org/10.1007/978-3-319-24574-4_28
    89 rdf:type schema:CreativeWork
    90 https://doi.org/10.1002/jor.23632 schema:sameAs https://app.dimensions.ai/details/publication/pub.1085993094
    91 rdf:type schema:CreativeWork
    92 https://doi.org/10.1016/j.cmpb.2016.12.006 schema:sameAs https://app.dimensions.ai/details/publication/pub.1039502687
    93 rdf:type schema:CreativeWork
    94 https://doi.org/10.1016/j.jacr.2017.02.019 schema:sameAs https://app.dimensions.ai/details/publication/pub.1084083884
    95 rdf:type schema:CreativeWork
    96 https://doi.org/10.1016/j.jsams.2008.07.005 schema:sameAs https://app.dimensions.ai/details/publication/pub.1008134660
    97 rdf:type schema:CreativeWork
    98 https://doi.org/10.1016/j.knee.2008.11.004 schema:sameAs https://app.dimensions.ai/details/publication/pub.1046258888
    99 rdf:type schema:CreativeWork
    100 https://doi.org/10.1109/cvpr.2016.90 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093359587
    101 rdf:type schema:CreativeWork
    102 https://doi.org/10.1109/iccv.2015.123 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093828312
    103 rdf:type schema:CreativeWork
    104 https://doi.org/10.1136/bjsm.2004.010900 schema:sameAs https://app.dimensions.ai/details/publication/pub.1000039211
    105 rdf:type schema:CreativeWork
    106 https://doi.org/10.1148/radiology.167.3.3363138 schema:sameAs https://app.dimensions.ai/details/publication/pub.1079704860
    107 rdf:type schema:CreativeWork
    108 https://doi.org/10.1177/0363546508330136 schema:sameAs https://app.dimensions.ai/details/publication/pub.1063821339
    109 rdf:type schema:CreativeWork
    110 https://doi.org/10.1177/0363546516629944 schema:sameAs https://app.dimensions.ai/details/publication/pub.1063822946
    111 rdf:type schema:CreativeWork
    112 https://doi.org/10.1302/0301-620x.68b3.3755441 schema:sameAs https://app.dimensions.ai/details/publication/pub.1079934530
    113 rdf:type schema:CreativeWork
    114 https://doi.org/10.1302/0301-620x.92b3.22424 schema:sameAs https://app.dimensions.ai/details/publication/pub.1046902699
    115 rdf:type schema:CreativeWork
    116 https://doi.org/10.3109/17453679009006504 schema:sameAs https://app.dimensions.ai/details/publication/pub.1008199874
    117 rdf:type schema:CreativeWork
    118 https://www.grid.ac/institutes/grid.21729.3f schema:alternateName Columbia University
    119 schema:name Department of Radiology, Columbia University Irving Medical Center, 622 West 168th St., MC 28, 10032, New York, NY, USA
    120 rdf:type schema:Organization
    121 https://www.grid.ac/institutes/grid.417319.9 schema:alternateName University of California, Irvine Medical Center
    122 schema:name Center for Artificial Intelligence in Diagnostic Medicine, University of California Irvine Medical Center, 101 The City Drive South, Building 55, Suite 201, 92868, Orange, CA, USA
    123 rdf:type schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...