Two-Dimensional Mammography Imaging Techniques for Screening Women with Silicone Breast Implants: A Pilot Phantom Study.

This study aimed to evaluate the impact of three two-dimensional (2D) mammographic acquisition techniques on image quality and radiation dose in the presence of silicone breast implants (BIs). Then, we propose and validate a new International Atomic Energy Agency (IAEA) phantom to reproduce these techniques. Images were acquired on a single Hologic Selenia Dimensions® unit. The mammography of the left breast of a single clinical case was included. Three methods of image acquisition were identified. They were based on misused, recommended, and reference settings. In the clinical case, image criteria scoring and the signal-to-noise ratio on breast tissue (SNRBT) were determined for two 2D projections and compared between the three techniques. The phantom study first compared the reference and misused settings by varying the AEC sensor position and, second, the recommended settings with a reduced current-time product (mAs) setting that was 13% lower. The signal-difference-to-noise ratio (SDNR) and detectability indexes at 0.1 mm (d' 0.1 mm) and 0.25 mm (d' 0.25 mm) were automatically quantified using ATIA software. Average glandular dose (AGD) values were collected for each acquisition. A statistical analysis was performed using Kruskal-Wallis and corrected Dunn tests (p < 0.05). The SNRBT was 2.6 times lower and the AGD was -18% lower with the reference settings compared to the recommended settings. The SNRBT values increased by +98% with the misused compared to the recommended settings. The AGD increased by +79% with the misused settings versus the recommended settings. The median values of the reference settings were 5.8 (IQR 5.7-5.9), 1.2 (IQR 0.0), 7.0 (IQR 6.8-7.2) and 1.2 (IQR 0.0) mGy and were significantly lower than those of the misused settings (p < 0.03): 7.9 (IQR 6.1-9.7), 1.6 (IQR 1.3-1.9), 9.2 (IQR 7.5-10.9) and 2.2 (IQR 1.4-3.0) mGy for the SDNR, d' 0.1 mm, d' 0.25 mm and the AGD, respectively. A comparison of the recommended and reduced settings showed a reduction of -6.1 ± 0.6% (p = 0.83), -7.7 ± 0.0% (p = 0.18), -6.4 ± 0.6% (p = 0.19) and -13.3 ± 1.1% (p = 0.53) for the SDNR, d' 0.1 mm, d' 0.25 mm and the AGD, respectively. This study showed that the IAEA phantom could be used to reproduce the three techniques for acquiring 2D mammography images in the presence of breast implants for raising awareness and for educational purposes. It could also be used to evaluate and optimize the manufacturer's recommended settings.

Best imaging signs identified by radiomics could outperform the model: application to differentiating lung carcinoid tumors from atypical hamartomas.

OBJECTIVES: Lung carcinoids and atypical hamartomas may be difficult to differentiate but require different treatment. The aim was to differentiate these tumors using contrast-enhanced CT semantic and radiomics criteria. METHODS: Between November 2009 and June 2020, consecutives patient operated for hamartomas or carcinoids with contrast-enhanced chest-CT were retrospectively reviewed. Semantic criteria were recorded and radiomics features were extracted from 3D segmentations using Pyradiomics. Reproducible and non-redundant radiomics features were used to training a random forest algorithm with cross-validation. A validation-set from another institution was used to evaluate of the radiomics signature, the 3D 'median' attenuation feature (3D-median) alone and the mean value from 2D-ROIs. RESULTS: Seventy-three patients (median 58 years [43‒70]) were analyzed (16 hamartomas; 57 carcinoids). The radiomics signature predicted hamartomas vs carcinoids on the external dataset (22 hamartomas; 32 carcinoids) with an AUC = 0.76. The 3D-median was the most important in the model. Density thresholds < 10 HU to predict hamartoma and > 60 HU to predict carcinoids were chosen for their high specificity > 0.90. On the external dataset, sensitivity and specificity of the 3D-median and 2D-ROIs were, respectively, 0.23, 1.00 and 0.13, 1.00 < 10 HU; 0.63, 0.95 and 0.69, 0.91 > 60 HU. The 3D-median was more reproducible than 2D-ROIs (ICC = 0.97 95% CI [0.95‒0.99]; bias: 3 ± 7 HU limits of agreement (LoA) [- 10‒16] vs. ICC = 0.90 95% CI [0.85‒0.94]; bias: - 0.7 ± 21 HU LoA [- 4‒40], respectively). CONCLUSIONS: A radiomics signature can distinguish hamartomas from carcinoids with an AUC = 0.76. Median density < 10 HU and > 60 HU on 3D or 2D-ROIs may be useful in clinical practice to diagnose these tumors with confidence, but 3D is more reproducible. CRITICAL RELEVANCE STATEMENT: Radiomic features help to identify the most discriminating imaging signs using random forest. 'Median' attenuation value (Hounsfield units), extracted from 3D-segmentations on contrast-enhanced chest-CTs, could distinguish carcinoids from atypical hamartomas (AUC = 0.85), was reproducible (ICC = 0.97), and generalized to an external dataset. KEY POINTS: • 3D-'Median' was the best feature to differentiate carcinoids from atypical hamartomas (AUC = 0.85). • 3D-'Median' feature is reproducible (ICC = 0.97) and was generalized to an external dataset. • Radiomics signature from 3D-segmentations differentiated carcinoids from atypical hamartomas with an AUC = 0.76. • 2D-ROI value reached similar performance to 3D-'median' but was less reproducible (ICC = 0.90).

Comparative performances of machine learning algorithms in radiomics and impacting factors.

There are no current recommendations on which machine learning (ML) algorithms should be used in radiomics. The objective was to compare performances of ML algorithms in radiomics when applied to different clinical questions to determine whether some strategies could give the best and most stable performances regardless of datasets. This study compares the performances of nine feature selection algorithms combined with fourteen binary classification algorithms on ten datasets. These datasets included radiomics features and clinical diagnosis for binary clinical classifications including COVID-19 pneumonia or sarcopenia on CT, head and neck, orbital or uterine lesions on MRI. For each dataset, a train-test split was created. Each of the 126 (9 × 14) combinations of feature selection algorithms and classification algorithms was trained and tuned using a ten-fold cross validation, then AUC was computed. This procedure was repeated three times per dataset. Best overall performances were obtained with JMI and JMIM as feature selection algorithms and random forest and linear regression models as classification algorithms. The choice of the classification algorithm was the factor explaining most of the performance variation (10% of total variance). The choice of the feature selection algorithm explained only 2% of variation, while the train-test split explained 9%.