Quantitative Imaging and Imaging Biomarkers: The Search for Generalizability in Radiology

MRI Radiomics in Prostate Cancer: A Reliability Study

BACKGROUND

Radiomics can provide quantitative features from medical imaging that can be correlated to clinical endpoints. The challenges relevant to robustness of radiomics features have been analyzed by many researchers, as it seems to be influenced by acquisition and reconstruction protocols, as well as by the segmentation of the region of interest (ROI). Prostate cancer (PCa) represents a difficult playground for this technique, due to discrepancies in the identification of the cancer lesion and the heterogeneity of the acquisition protocols. The aim of this study was to investigate the reliability of radiomics in PCa magnetic resonance imaging (MRI).

METHODS

A homogeneous cohort of patients with a PSA rise that underwent multiparametric MRI imaging of the prostate before biopsy was tested in this study. All the patients were acquired with the same MRI scanner, with a standardized protocol. The identification and the contouring of the region of interest (ROI) of an MRI suspicious cancer lesion were done by two radiologists with great experience in prostate cancer (>10 years). After the segmentation, the texture features were extracted with LIFEx. Texture features were then tested with intraclass coefficient correlation (ICC) analysis to analyze the reliability of the segmentation.

RESULTS

Forty-four consecutive patients were included in the present analysis. In 26 patients (59.1%), the prostate biopsy confirmed the presence of prostate cancer, which was scored as Gleason 6 in 6 patients (13.6%), Gleason 3 + 4 in 8 patients (18.2%), and Gleason 4 + 3 in 12 patients (27.3%). The reliability analysis conversely showed poor reliability in the majority of the MRI acquisition (61% in T2, 89% in DWI50, 44% in DWI400, and 83% in DWI1,500), with ADC acquisition only showing better reliability (poor reliability in only 33% of the texture features).

CONCLUSIONS

The low ratio of reliability in a monoinstitutional homogeneous cohort represents a significant alarm bell for the application of MRI radiomics in the field of prostate cancer. More work is needed in a clinical setting to further study the potential of MRI radiomics in prostate cancer.

Delta-radiomics increases multicentre reproducibility: a phantom study

Texture analysis (TA) can provide quantitative features from medical imaging that can be correlated to clinical endpoints. The challenges relevant to robustness of radiomics features have been analyzed by many researchers, as it seems to be influenced by acquisition and reconstruction protocols. Delta-texture analysis (D-TA), conversely, consist in the analysis of TA feature variations at different acquisition times, usually before and after a therapy. Aim of this study was to investigate the influence of different CT scanners and acquisition parameters in the robustness of TA and D-TA. We scanned a commercial phantom (CIRS model 467, Gammex, Middleton, WI, USA), that is used for the calibration of electron density, two times by varying the disposition of plugs, using three different scanners. After the segmentation, we extracted TA features with LifeX and calculated TA features and D-TA features, defined as the variation of each TA parameters extracted from the same position by varying the plugs with the formula (Y-X)/X. The robustness of TA and D-TA features were then tested with intraclass coefficient correlation (ICC) analysis. The reliability of TA parameters across different scans, with different acquisition parameters and ROI positions has shown poor reliability in 12/37 and moderate reliability in the remaining 25/37, with no parameters showing good reliability. The reliability of D-TA, conversely, showed poor reliability in 10/37 parameters, moderate reliability in 10/37 parameters, and good reliability in 17/37 parameters. The comparison between TA and D-TA ICCs showed a significant difference for the whole group of parameters (p:0.004) and for the subclasses of GLCM parameters (p:0.033), whereas for the other subclasses of matrices (GLRLM, NGLDM, GLZLM, Histogram), the difference was not significant. D-TA features seem to be more robust than TA features. These findings reinforce the potentiality for using D-TA features for early assessment of treatment response and for developing tailored therapies. More work is needed in a clinical setting to confirm the results of the present study.

3-Dimensional Quantification of Composite Pleural Plaque Volume in Patients Exposed to Asbestos Using High-resolution Computed Tomography: A Validation Study

RATIONALE

As pleural plaque has been reported as a risk factor in the occurrence of lung cancer and mesothelioma, a reproducible and precise method of measurement of pleural plaque volume (PPV) is needed to further describe these relationships. The aim of the study was to assess the reproducibility of a 3-dimensional computed tomography (3D-CT) volumetric analysis of PPV in patients with occupational exposure to asbestos.

MATERIAL AND METHODS

A total of 28 patients were retrospectively randomly selected from the multicenter APEXS (Asbestos Post Exposure Survey) study, which was held between 2003 and 2005. All patients underwent a 3D-CT scan. Two readers specialized in chest radiology completed the 3D semiautomated quantification of lung volume using dedicated software. They also had to categorize the visual extent of pleural plaque in terms of thickness and circumference. Reproducibility of the continuous PPV variable was assessed using the intraclass correlation coefficient (ICC) and Bland-Altman analysis. Reproducibility of categorical variables was assessed using the κ test.

RESULTS

Intraobserver reproducibility of PPV was almost perfect (ICC=0.98 [95% interval: 0.97-0.99]), and interobserver reproducibility was very good (ICC=0.93 [0.88-0.97]). At Bland-Altman analysis, the mean differences were 0.1 (limit of agreement: -11.0 to 11.2) and 3.7 cc (-17.8 to 25.2), respectively. Visual analysis of both plaque in terms of thickness and circumference were fair to moderate, with κ values ranging from 0.30 to 0.60.

CONCLUSIONS

3D semiautomatic quantification of PPV is feasible and reproducible using CT in patients with occupational exposure to asbestos. PPV measurement may be useful to correlate with other asbestos-related disease outcomes and prognosis.