Deep-learning convolutional neural network-based scatter correction for contrast enhanced digital breast tomosynthesis in both cranio-caudal and mediolateral-oblique views.

Purpose: Scatter radiation in contrast-enhanced digital breast tomosynthesis (CEDBT) reduces the image quality and iodinated lesion contrast. Monte Carlo simulation can provide accurate scatter estimation at the cost of computational burden. A model-based convolutional method trades off accuracy for processing speed. The purpose of this study is to develop a fast and robust deep-learning (DL) convolutional neural network (CNN)-based scatter correction method for CEDBT. Approach: Projection images and scatter maps of digital anthropomorphic breast phantoms were generated using Monte Carlo simulations. Experiments were conducted to validate the simulated scatter-to-primary ratio (SPR) at different locations of a breast phantom. Simulated projection images were used for CNN training and testing. Two separate U-Nets [low-energy (LE)-CNN and high-energy (HE)-CNN] were trained for LE and HE spectrum, respectively. CNN-based scatter correction was applied to a clinical case with a malignant iodinated mass to evaluate the influence on the lesion detection. Results: The average and standard deviation of mean absolute percentage error of LE-CNN and HE-CNN estimated scatter map are 2 % ± 0.4 % and 2.4 % ± 0.8 % , respectively. For clinical cases, the lesion signal difference to noise ratio average improvement was 190% after CNN-based scatter correction. To conduct scatter correction on clinical CEDBT images, the whole process of loading CNNs parameters and scatter correction for LE and HE images took < 4 s , with 9 GB GPU computational cost. The SPR variation across the breast agrees between experimental measurements and Monte Carlo simulations. Conclusions: We developed a CNN-based scatter correction method for CEDBT in both CC view and mediolateral-oblique view with high accuracy and fast speed.

A Competition, Benchmark, Code, and Data for Using Artificial Intelligence to Detect Lesions in Digital Breast Tomosynthesis.

Importance: An accurate and robust artificial intelligence (AI) algorithm for detecting cancer in digital breast tomosynthesis (DBT) could significantly improve detection accuracy and reduce health care costs worldwide. Objectives: To make training and evaluation data for the development of AI algorithms for DBT analysis available, to develop well-defined benchmarks, and to create publicly available code for existing methods. Design, Setting, and Participants: This diagnostic study is based on a multi-institutional international grand challenge in which research teams developed algorithms to detect lesions in DBT. A data set of 22 032 reconstructed DBT volumes was made available to research teams. Phase 1, in which teams were provided 700 scans from the training set, 120 from the validation set, and 180 from the test set, took place from December 2020 to January 2021, and phase 2, in which teams were given the full data set, took place from May to July 2021. Main Outcomes and Measures: The overall performance was evaluated by mean sensitivity for biopsied lesions using only DBT volumes with biopsied lesions; ties were broken by including all DBT volumes. Results: A total of 8 teams participated in the challenge. The team with the highest mean sensitivity for biopsied lesions was the NYU B-Team, with 0.957 (95% CI, 0.924-0.984), and the second-place team, ZeDuS, had a mean sensitivity of 0.926 (95% CI, 0.881-0.964). When the results were aggregated, the mean sensitivity for all submitted algorithms was 0.879; for only those who participated in phase 2, it was 0.926. Conclusions and Relevance: In this diagnostic study, an international competition produced algorithms with high sensitivity for using AI to detect lesions on DBT images. A standardized performance benchmark for the detection task using publicly available clinical imaging data was released, with detailed descriptions and analyses of submitted algorithms accompanied by a public release of their predictions and code for selected methods. These resources will serve as a foundation for future research on computer-assisted diagnosis methods for DBT, significantly lowering the barrier of entry for new researchers.

Wavelet Guided 3D Deep Model to improve Dental Microfracture Detection.

Epidemiological studies indicate that microfractures (cracks) are the third most common cause of tooth loss in industrialized countries. An undetected crack will continue to progress, often with significant pain, until the tooth is lost. Previous attempts to utilize cone beam computed tomography (CBCT) for detecting cracks in teeth had very limited success. We propose a model that detects cracked teeth in high resolution (hr) CBCT scans by combining signal enhancement with a deep CNNbased crack detection model. We perform experiments on a dataset of 45 ex-vivo human teeth with 31 cracked and 14 controls. We demonstrate that a model that combines classical wavelet-based features with a deep 3D CNN model can improve fractured tooth detection accuracy in both micro-Computed Tomography (ground truth) and hr-CBCT scans. The CNN model is trained to predict a probability map showing the most likely fractured regions. Based on this fracture probability map we detect the presence of fracture and are able to differentiate a fractured tooth from a control tooth. We compare these results to a 2D CNN-based approach and we show that our approach provides superior detection results. We also show that the proposed solution is able to outperform oral and maxillofacial radiologists in detecting fractures from the hr-CBCT scans. Early detection of cracks will lead to the design of more appropriate treatments and longer tooth retention.

Structure Correction for Robust Volume Segmentation in Presence of Tumors.

CNN based lung segmentation models in absence of diverse training dataset fail to segment lung volumes in presence of severe pathologies such as large masses, scars, and tumors. To rectify this problem, we propose a multi-stage algorithm for lung volume segmentation from CT scans. The algorithm uses a 3D CNN in the first stage to obtain a coarse segmentation of the left and right lungs. In the second stage, shape correction is performed on the segmentation mask using a 3D structure correction CNN. A novel data augmentation strategy is adopted to train a 3D CNN which helps in incorporating global shape prior. Finally, the shape corrected segmentation mask is up-sampled and refined using a parallel flood-fill operation. The proposed multi-stage algorithm is robust in the presence of large nodules/tumors and does not require labeled segmentation masks for entire pathological lung volume for training. Through extensive experiments conducted on publicly available datasets such as NSCLC, LUNA, and LOLA11 we demonstrate that the proposed approach improves the recall of large juxtapleural tumor voxels by at least 15% over state-of-the-art models without sacrificing segmentation accuracy in case of normal lungs. The proposed method also meets the requirement of CAD software by performing segmentation within 5 seconds which is significantly faster than present methods.

A Lightweight Multi-Section CNN for Lung Nodule Classification and Malignancy Estimation.

The size and shape of a nodule are the essential indicators of malignancy in lung cancer diagnosis. However, effectively capturing the nodule's structural information from CT scans in a computer-aided system is a challenging task. Unlike previous models that proposed computationally intensive deep ensemble models or three-dimensional CNN models, we propose a lightweight, multiple view sampling based multi-section CNN architecture. The model obtains a nodule's cross sections from multiple view angles and encodes the nodule's volumetric information into a compact representation by aggregating information from its different cross sections via a view pooling layer. The compact feature is subsequently used for the task of nodule classification. The method does not require the nodule's spatial annotation and works directly on the cross sections generated from volume enclosing the nodule. We evaluated the proposed method on lung image database consortium (LIDC) and image database resource initiative (IDRI) dataset. It achieved the state-of-the-art performance with a mean 93.18% classification accuracy. The architecture could also be used to select the representative cross sections determining the nodule's malignancy that facilitates in the interpretation of results. Because of being lightweight, the model could be ported to mobile devices, which brings the power of artificial intelligence (AI) driven application directly into the practitioner's hand.