eduzhai > Applied Sciences > Engineering >

Weakly Supervised Multi-Organ Multi-Disease Classification of Body CT Scans

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: We designed a multi-organ, multi-label disease classification algorithm forcomputed tomography (CT) scans using case-level labels from radiology textreports. A rule-based algorithm extracted 19,255 disease labels from reports of13,667 body CT scans from 12,092 subjects. A 3D DenseVNet was trained tosegment 3 organ systems: lungs pleura, liver gallbladder, and kidneys. Frompatches guided by segmentations, a 3D convolutional neural network providedmulti-label disease classification for normality versus four common diseasesper organ. The process was tested on 2,158 CT volumes with 2,875 manuallyobtained labels. Manual validation of the rulebased labels confirmed 91 to 99 accuracy. Results were characterized using the receiver operatingcharacteristic area under the curve (AUC). Classification AUCs for lungs pleuralabels were as follows: atelectasis 0.77 (95 confidence intervals 0.74 to0.81), nodule 0.65 (0.61 to 0.69), emphysema 0.89 (0.86 to 0.92), effusion 0.97(0.96 to 0.98), and normal 0.89 (0.87 to 0.91). For liver gallbladder, AUCswere: stone 0.62 (0.56 to 0.67), lesion 0.73 (0.69 to 0.77), dilation 0.87(0.84 to 0.90), fatty 0.89 (0.86 to 0.92), and normal 0.82 (0.78 to 0.85). Forkidneys, AUCs were: stone 0.83 (0.79 to 0.87), atrophy 0.92 (0.89 to 0.94),lesion 0.68 (0.64 to 0.72), cyst 0.70 (0.66 to 0.73), and normal 0.79 (0.75 to0.83). In conclusion, by using automated extraction of disease labels fromradiology reports, we created a weakly supervised, multi-organ, multi-diseaseclassifier that can be easily adapted to efficiently leverage massive amountsof unannotated data associated with medical images.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×