Understanding metric-related pitfalls in image analysis validation

Validation metrics are key for tracking scientific progress and bridging the current chasm between artificial intelligence research and its translation into practice. However, increasing evidence shows that, particularly in image analysis, metrics are often chosen inadequately. Although taking into account the individual strengths, weaknesses and limitations of validation metrics is a critical prerequisite to making educated choices, the relevant knowledge is currently scattered and poorly accessible to individual researchers. Based on a multistage Delphi process conducted by a multidisciplinary expert consortium as well as extensive community feedback, the present work provides a reliable and comprehensive common point of access to information on pitfalls related to validation metrics in image analysis. Although focused on biomedical image analysis, the addressed pitfalls generalize across application domains and are categorized according to a newly created, domain-agnostic taxonomy. The work serves to enhance global comprehension of a key topic in image analysis validation.

Transformer-based Deep Neural Network for Breast Cancer Classification on Digital Breast Tomosynthesis Images

Purpose

To develop an efficient deep neural network model that incorporates context from neighboring image sections to detect breast cancer on digital breast tomosynthesis (DBT) images.

Materials and Methods

The authors adopted a transformer architecture that analyzes neighboring sections of the DBT stack. The proposed method was compared with two baselines: an architecture based on three-dimensional (3D) convolutions and a two-dimensional model that analyzes each section individually. The models were trained with 5174 four-view DBT studies, validated with 1000 four-view DBT studies, and tested on 655 four-view DBT studies, which were retrospectively collected from nine institutions in the United States through an external entity. Methods were compared using area under the receiver operating characteristic curve (AUC), sensitivity at a fixed specificity, and specificity at a fixed sensitivity.

Results

On the test set of 655 DBT studies, both 3D models showed higher classification performance than did the per-section baseline model. The proposed transformer-based model showed a significant increase in AUC (0.88 vs 0.91, P = .002), sensitivity (81.0% vs 87.7%, P = .006), and specificity (80.5% vs 86.4%, P < .001) at clinically relevant operating points when compared with the single-DBT-section baseline. The transformer-based model used only 25% of the number of floating-point operations per second used by the 3D convolution model while demonstrating similar classification performance.

Conclusion

A transformer-based deep neural network using data from neighboring sections improved breast cancer classification performance compared with a per-section baseline model and was more efficient than a model using 3D convolutions.Keywords: Breast, Tomosynthesis, Diagnosis, Supervised Learning, Convolutional Neural Network (CNN), Digital Breast Tomosynthesis, Breast Cancer, Deep Neural Networks, Transformers Supplemental material is available for this article. © RSNA, 2023.

Fully automated detection of breast cancer in screening MRI using convolutional neural networks

Current computer-aided detection (CADe) systems for contrast-enhanced breast MRI rely on both spatial information obtained from the early-phase and temporal information obtained from the late-phase of the contrast enhancement. However, late-phase information might not be available in a screening setting, such as in abbreviated MRI protocols, where acquisition is limited to early-phase scans. We used deep learning to develop a CADe system that exploits the spatial information obtained from the early-phase scans. This system uses three-dimensional (3-D) morphological information in the candidate locations and the symmetry information arising from the enhancement differences of the two breasts. We compared the proposed system to a previously developed system, which uses the full dynamic breast MRI protocol. For training and testing, we used 385 MRI scans, containing 161 malignant lesions. Performance was measured by averaging the sensitivity values between 1/8-eight false positives. In our experiments, the proposed system obtained a significantly ([Formula: see text]) higher average sensitivity ([Formula: see text]) compared with that of the previous CADe system ([Formula: see text]). In conclusion, we developed a CADe system that is able to exploit the spatial information obtained from the early-phase scans and can be used in screening programs where abbreviated MRI protocols are used.

Classifying symmetrical differences and temporal change for the detection of malignant masses in mammography using deep neural networks

We investigate the addition of symmetry and temporal context information to a deep convolutional neural network (CNN) with the purpose of detecting malignant soft tissue lesions in mammography. We employ a simple linear mapping that takes the location of a mass candidate and maps it to either the contralateral or prior mammogram, and regions of interest (ROIs) are extracted around each location. Two different architectures are subsequently explored: (1) a fusion model employing two datastreams where both ROIs are fed to the network during training and testing and (2) a stagewise approach where a single ROI CNN is trained on the primary image and subsequently used as a feature extractor for both primary and contralateral or prior ROIs. A "shallow" gradient boosted tree classifier is then trained on the concatenation of these features and used to classify the joint representation. The baseline yielded an AUC of 0.87 with confidence interval [0.853, 0.893]. For the analysis of symmetrical differences, the first architecture where both primary and contralateral patches are presented during training obtained an AUC of 0.895 with confidence interval [0.877, 0.913], and the second architecture where a new classifier is retrained on the concatenation an AUC of 0.88 with confidence interval [0.859, 0.9]. We found a significant difference between the first architecture and the baseline at high specificity with [Formula: see text]. When using the same architectures to analyze temporal change, we yielded an AUC of 0.884 with confidence interval [0.865, 0.902] for the first architecture and an AUC of 0.879 with confidence interval [0.858, 0.898] in the second setting. Although improvements for temporal analysis were consistent, they were not found to be significant. The results show our proposed method is promising and we suspect performance can greatly be improved when more temporal data become available.