Virtual Library

Start Your Search

H. Kim



Author of

  • +

    P1.19 - Poster Session 1 - Imaging (ID 179)

    • Event: WCLC 2013
    • Type: Poster Session
    • Track: Imaging, Staging & Screening
    • Presentations: 1
    • +

      P1.19-015 - Clinical application of thin section chest CT with computer-aided detection system for the detection of small metastatic nodules in pulmonary metastasectomy. (ID 3428)

      09:30 - 09:30  |  Author(s): H. Kim

      • Abstract

      Background
      It is important to identify possible metastatic nodule prior to metastasectomy, in order to perform complete resection. Thin section chest CT (TSCT) can detect small metastatic nodules more accurately than conventional chest CT. Conventional CT is known to miss metastatic nodules that are manually palpable by up to 40%. Reported sensitivity of helical CT cannot exceed 90%, also. We adopted TSCT and applied computer-aided detection (CAD) system for the search of small metastatic nodules to improve the detection power of TSCT.

      Methods
      From March 2009 to February 2013, 333 patients were referred to thoracic surgeon for pulmonary metastasectomy. TSCT with CAD was performed in every candidate for pulmonary metastasectomy. Every nodule detected by TSCT and CAD was annotated by radiologists and the whole map and annotation numbers of nodules were reported before operation. Intraoperatively, bi-manual palpation in open metastasectomy and finger palpation in video assisted thoracoscopic surgery (VATS) were used to detect the nodules. Only operations with complete available data were included in the study and 251 operations from 239 patients were analyzed.

      Results
      A total of 1021 nodules were identified by TSCT with CAD, while 115 nodules (11.3%) were additionally detected by CAD. Among the annotated nodules, 742 nodules were palpable during operation. Including intraoperatively detected 49 nodules, a total of 824 nodules were resected. Of 527 pathologically proven malignant nodules, 518 (62.9%) nodules were metastatic lesions. Among metastatic lesions, 496 were detected by TSCT (95.8%), 9 by CAD (1.7%) and 13 by manual palpation (2.5%). Also, 11 metastatic nodules were no palpated during operation but resected according to the annotation on CT. The overall sensitivity of thin section chest CT with CAD was 97.5%. Though overall specificity of TSCT with CAD was very low as 12.1% because that of radiologist’s reading was 14.2%, specificity only for CAD was 45.6%.

      Conclusion
      TSCT could detect small subcentimetre nodules and CAD enabled detecting additional small nodules. With help of TSCT with CAD, even the metastatic nodules, which were not palpable during operation, could successfully be resected. Furthermore, TSCT with CAD could detect tiny nodules with rather high specificity compare to radiologist’s reading only. Although the number of true metastatic nodules was small and several metastatic nodules were undetected, the TSCT with CAD system increased the detection sensitivity and would be helpful in complete metastasectomy.