Search

Objectives To evaluate the performance of radiomic features extracted from high-resolution computed tomography (HRCT) for the differentiation between cholesteatoma and middle ear inflammation (MEI), and to investigate the impact of post-reconstruction harmonization and data resampling. Methods One hundred patients were included in this retrospective dual-center study: 48 with histology-proven cholesteatoma (center A: 23; center B: 25) and 52 with MEI (A: 27; B: 25). Radiomic features (co-occurrence and run-length matrix, absolute gradient, autoregressive model, Haar wavelet transform) were extracted from manually defined 2D-ROIs. The ten best features for lesion differentiation were selected using probability of error and average correlation coefficients. A multi-layer perceptron feed-forward artificial neural network (MLP-ANN) was used for radiomics-based classification, with histopathology serving as the reference standard (70% of cases for training, 30% for validation). The analysis was performed five times each on (a) unmodified data and on data that were (b) resampled to the same matrix size, and (c) corrected for acquisition protocol differences using ComBat harmonization. Results Using unmodified data, the MLP-ANN classification yielded an overall median area under the receiver operating characteristic curve (AUC) of 0.78 (0.72–0.84). Using original data from center A and resampled data from center B, an overall median AUC of 0.88 (0.82–0.99) was yielded, while using ComBat harmonized data, an overall median AUC of 0.89 (0.79–0.92) was revealed. Conclusion Radiomic features extracted from HRCT differentiate between cholesteatoma and MEI. When using multi-centric data obtained with differences in CT acquisition parameters, data resampling and ComBat post-reconstruction harmonization clearly improve radiomics-based lesion classification. Key Points • Unenhanced high-resolution CT coupled with radiomics analysis may be useful for the differentiation between cholesteatoma and middle ear inflammation. • Pooling of data extracted from inhomogeneous CT datasets does not appear meaningful without further post-processing. • When using multi-centric CT data obtained with differences in acquisition parameters, post-reconstruction harmonization and data resampling clearly improve radiomics-based soft-tissue differentiation.

Existing quantitative imaging biomarkers (QIBs) are associated with known biological tissue characteristics and follow a well-understood path of technical, biological and clinical validation before incorporation into clinical trials. In radiomics, novel data-driven processes extract numerous visually imperceptible statistical features from the imaging data with no a priori assumptions on their correlation with biological processes. The selection of relevant features (radiomic signature) and incorporation into clinical trials therefore requires additional considerations to ensure meaningful imaging endpoints. Also, the number of radiomic features tested means that power calculations would result in sample sizes impossible to achieve within clinical trials. This article examines how the process of standardising and validating data-driven imaging biomarkers differs from those based on biological associations. Radiomic signatures are best developed initially on datasets that represent diversity of acquisition protocols as well as diversity of disease and of normal findings, rather than within clinical trials with standardised and optimised protocols as this would risk the selection of radiomic features being linked to the imaging process rather than the pathology. Normalisation through discretisation and feature harmonisation are essential pre-processing steps. Biological correlation may be performed after the technical and clinical validity of a radiomic signature is established, but is not mandatory. Feature selection may be part of discovery within a radiomics-specific trial or represent exploratory endpoints within an established trial; a previously validated radiomic signature may even be used as a primary/secondary endpoint, particularly if associations are demonstrated with specific biological processes and pathways being targeted within clinical trials. • Data-driven processes like radiomics risk false discoveries due to high-dimensionality of the dataset compared to sample size, making adequate diversity of the data, cross-validation and external validation essential to mitigate the risks of spurious associations and overfitting. • Use of radiomic signatures within clinical trials requires multistep standardisation of image acquisition, image analysis and data mining processes. • Biological correlation may be established after clinical validation but is not mandatory.

To determine, in patients with melanoma, the dependence of PET sensitivity on pulmonary metastasis size, and to determine patients who require further evaluation for definite staging.Of 183 melanoma patients who underwent (18)F-fluorodeoxyglucose PET/computed tomography (CT) for staging or follow-up between January 2008 and June 2011, 38 patients (18 women and 20 men; mean age 62.0 ± 14.7 years) with one or more pulmonary metastases visible on CT were included in the retrospective study. Each pulmonary metastasis was rated as positive or negative on PET, and lesion size (maximum transverse diameter) was assessed on CT. PET sensitivity was calculated according to the lesions' size, in 2-mm steps.A total of 181 pulmonary metastases were analysed. PET sensitivity was 7.9 % for lesions of 4-5 mm; 33.3 % for lesions of 6-7 mm; 56.8 % for lesions of 8-9 mm; 63.6 % for lesions of 10-11 mm; 100 % for lesions of 12-14 mm; and 100 % for lesions of at least 15 mm. The differences in sensitivity between the size groups were significant (P 0.001) CONCLUSIONS: With current state-of-the-art PET/CT technology, additional tests are necessary for definitive staging of melanoma patients who have one or more PET-negative lung nodules less than 12 mm in diameter on expiratory CT.• PET cannot rule out malignancy in pulmonary nodules less than 12 mm on expiratory CT. • Melanoma patients with PET-negative pulmonary nodules less than 12 mm require additional tests. • Knowledge of these factors can help interpretation of PET and PET/CT findings.

To compare fused gadoxetate-enhanced Ga-68-DOTANOC PET/MRI and Ga-68-DOTANOC PET/DWI (diffusion-weighted imaging) for the assessment of abdominal neuroendocrine tumours (NETs).Eighteen patients with suspected or histologically proven NETs of the abdomen were enrolled in this retrospective study. All patients underwent Ga-68-DOTANOC PET/CT for a primary search, staging, or restaging, and received an additional MRI, including dynamic gadoxetate-enhanced T1-weighted sequences and DWI (b-values 50, 300 and 600). Co-registered gadoxetate-enhanced PET/MRI and PET/DWI were separately analysed for NET lesions by a nuclear medicine physician and a radiologist in consensus. Sensitivity and specificity were calculated on a per-region, per-organ and per-patient basis.Eighty-seven out of 684 anatomical regions, and 23 out of 270 organs, were NET-positive in 14 out of 18 patients. Region-based sensitivities and specificities were 97.7 % and 99.7 % for gadoxetate-enhanced PET/MRI and 98.9 % and 99.7 % for PET/DWI. Organ-based sensitivities and specificities were 91.3 % and 99.6 % for gadoxetate-enhanced PET/MRI and 95.7 % and 99.6 % for PET/DWI. Finally, patient-based sensitivities and specificities were 100 % and 100 % for gadoxetate-enhanced PET/MRI and 100 % and 75 % for PET/DWI. Sensitivities and specificities of the two methods did not differ significantly.Gadoxetate-enhanced Ga-68-DOTANOC PET/MRI and Ga-68-DOTANOC PET/DWI are equally useful for the assessment of abdominal NETs.• Positron emission tomography and magnetic resonance imaging can both assess neuroendocrine tumours. • Fusion of PET/MR imaging provides helpful information. • Gadoxetate-enhanced Ga-68-DOTANOC PET/MRI and Ga-68-DOTANOC PET/DWI assess neuroendocrine tumours equally well. • PET/DWI is inherently simpler than gadoxetate-enhanced PET/MRI. • Only benign hepatic lesions pose a potential diagnostic dilemma for PET/DWI.

To determine the value of intravenous contrast medium in (68)Ga-DOTA-Phe(1)-Tyr(3)-octreotide - (68)Ga-DOTATOC - PET/CT for the detection of abdominal neuroendocrine tumours (NET).In fifty-five patients with known or suspected NETs of the abdomen PET/CT was performed on a 64-row multi-detector hybrid system. For PET, 150 MBq of (68)Ga-DOTATOC were injected intravenously. Full-dose unenhanced, and arterial- and venous-phase contrast-enhanced CT images were obtained. Unenhanced and contrast-enhanced PET/CT images were evaluated separately for the presence of NETs on a per-region basis, by two separate teams with different experience levels.On unenhanced PET/CT, sensitivity and specificity ranged from 89.3% (junior team) to 92% (senior team), and 99.1% (junior team) to 99.2% (senior team), respectively. On contrast-enhanced PET/CT, sensitivity and specificity ranged from 92.3% (junior team) to 98.5% (senior team), and 99.4% (junior team) to 99.5% (senior team), respectively. These increases in sensitivity and specificity, due to the use of contrast-enhanced images, were statistically significant (P0.05).Intravenous contrast medium only moderately, aleit significantly, improves the sensitivity of (68)Ga-DOTATOC PET/CT for the detection of abdominal NETs, and hardly affects specificity. Thus, while contrast enhancement is justified to achieve maximum sensitivity, unenhanced images may be sufficient for routine PET/CT in NET patients.Contrast media moderately improve the sensitivity of (68)Ga-DOTATOC PET/CT for neuroendocrine tumours. Contrast media hardly affect the specificity of (68)Ga-DOTATOC PET/CT for neuroendocrine tumours. Unenhanced PET/CT is sufficient for routine imaging of patients with neuroendocrine tumours.