MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel
  1. Home
  2. Meeting Program
  3. All
  4. Evaluation of the Stability of Radiomics Features Using 4D-CT and Across Radiomics Platforms for Lung and Liver Tumors

Evaluation of the Stability of Radiomics Features Using 4D-CT and Across Radiomics Platforms for Lung and Liver Tumors

X Wang1*, C Ma1, H Wang2, Y Zhang1, N Yue1, K Nie1, (1) Rutgers-Cancer Institute of New Jersey, Rutgers-Robert Wood Johnson Medical School, New Brunswick, NJ, (2) Shanghai Chest Hospital, Shanghai Jiaotong University, Shanghai, China

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: To evaluate the stability of radiomics features using 4D-CT as an alternative to test-retest CT-scans, and the agreement/consistency across radiomics platforms.


Methods: 4D-CT images with 10 breathing phases were acquired for 10 patients with lung tumors and 10 patients with liver tumors. For each patient, two contours (GTV and GTV-1mm) were delineated on each breathing phase. GTV-1mm was generated by subtracting a 1mm inner margin from GTV to investigate the impact of contouring accuracy. Sixteen radiomics features were extracted using two open-source radiomics platforms. The intraclass correlation coefficient (ICC) of each feature was calculated from all phases to evaluate test-retest stability. The concordance correlation coefficient (CCC), the ICC(A,1) with an absolute agreement definition, the ICC(C,1) with a consistency definition, as well as Pearson correlation coefficient (PCC) were calculated to evaluate the agreement/consistency of the two radiomics platforms. Features with ICC/CCC above 75% were deemed stable features.


Results: Regardless of difference in contours, all features were identified as stable in test-retest for both radiomics platforms, with ICC ranging from 0.785 to 1. The two platforms only concorded on one feature (correlation in Gray Level Co-occurrence Matrix) for GTV in liver tumors (ICC(A,1)=0.88, CCC=0.879). The two platforms showed consistency on two features: short run emphasis (in Gray Level Run Length Matrix) for GTV-1mm in lung tumors (ICC(C,1)=0.813, PCC=0.813), and entropy (in first-order) in liver tumors for GTV (ICC(C,1)=0.999, PCC=0.999), and for GTV-1mm (ICC(C,1)=0.993, PCC=0.994).


Conclusion: The results show that the stability of radiomics features can be evaluated using 4D-CT as an alternative to test-retest CT-scans. The two radiomics platforms show some consistency, but low agreement. The stability evaluation depends on the contouring accuracy and disease site. Overall, we are able to identify stable radiomics features in test-retest and across platforms, which may have potential to be used for radiomics-guided clinical studies.

Keywords

CT, Feature Extraction, Statistical Analysis

Taxonomy

IM/TH- Image Analysis Skills (broad expertise across imaging modalities): Feature extraction, texture analysis, radiomics

Contact Email

Additional Links