Virtual Library

Start Your Search

Hong-Di Li



Author of

  • +

    P1.16 - Treatment of Early Stage/Localized Disease (Not CME Accredited Session) (ID 948)

    • Event: WCLC 2018
    • Type: Poster Viewing in the Exhibit Hall
    • Track:
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/24/2018, 16:45 - 18:00, Exhibit Hall
    • +

      P1.16-40 - Evaluating the Tumor Heterogeneity in Lung Cancer by Constructing Tumor Heterogeneity Index (THI) from Magnetic Resonance Imaging (ID 13134)

      16:45 - 18:00  |  Author(s): Hong-Di Li

      • Abstract

      Background

      To improve the evaluation of primary lung cancer heterogeneity using clinical routine magnetic resonance imaging (MRI), we proposed a method based on basic measurements from T1- and T2-weighted MRI.

      a9ded1e5ce5d75814730bb4caaf49419 Method

      As a novel technique of magnetic resonance imaging analysis, we investigated a total of 203 patients with biopsy-proven primary lung cancer and with different T stages. All patients previously received positron emission tomography/computed tomography (PET/CT) scan. Gross lesions were manually contoured on T1-weighted, T1-enhanced, T2-weighted and T2 fat suppression (T2fs) images. The ratios of standard deviation (SD) / mean tumor value from each sequence were calculated. Correlation analyses were performed between T stages and the ratios. P value <0.05 was defined as statistical significant. Then a linear regression was performed to determine the weight of each related ratio. A model was built to calculate Tumor Heterogeneous Index (THI). One hundred and one patients were analyzed as the training set and another 102 as validating set.

      4c3880bb027f159e801041b1021e88e8 Result

      There were 56 patients diagnosed with T1 disease, 60 with T2 disease, 51 with T3 disease and 36 with T4 disease. Pair matching was performed between training set and validating set. As a result of the correlation analyses, SD/mean ratio showed significantly correlations with T stages in T1-enhanced (p=0.003), T2-weighted (p<0.0001) and T2fs sequences (p=0.002). Based on a linear regression model, THI was established for assessing the heterogeneity of lung tumor, consisting the three ratio measurements. Correlation analysis demonstrated that Higher THI was significantly related to more advanced T stages (p<0.0001).

      fig.1.jpg

      8eea62084ca7e541d918e823422bd82e Conclusion

      The proposed SD/mean ratio measurements and the calculation of THI according to clinical routine MR images could be clinical biomarkers that correlated with T stages, and were capable of evaluating heterogeneity of lung cancers.

      6f8b794f3246b0c1e1780bb4d4d5dc53