An Artificial Intelligence System for Screening and Recommending the Treatment Modalities for Retinopathy of Prematurity.

PURPOSE: The purpose of this study was to develop an artificial intelligence (AI) system for the identification of disease status and recommending treatment modalities for retinopathy of prematurity (ROP). METHODS: This retrospective cohort study included a total of 24,495 RetCam images from 1075 eyes of 651 preterm infants who received RetCam examination at the Shenzhen Eye Hospital in Shenzhen, China, from January 2003 to August 2021. Three tasks included ROP identification, severe ROP identification, and treatment modalities identification (retinal laser photocoagulation or intravitreal injections). The AI system was developed to identify the 3 tasks, especially the treatment modalities of ROP. The performance between the AI system and ophthalmologists was compared using extra 200 RetCam images. RESULTS: The AI system exhibited favorable performance in the 3 tasks, including ROP identification [area under the receiver operating characteristic curve (AUC), 0.9531], severe ROP identification (AUC, 0.9132), and treatment modalities identification with laser photocoagulation or intravitreal injections (AUC, 0.9360). The AI system achieved an accuracy of 0.8627, a sensitivity of 0.7059, and a specificity of 0.9412 for identifying the treatment modalities of ROP. External validation results confirmed the good performance of the AI system with an accuracy of 92.0% in all 3 tasks, which was better than 4 experienced ophthalmologists who scored 56%, 65%, 71%, and 76%, respectively. CONCLUSIONS: The described AI system achieved promising outcomes in the automated identification of ROP severity and treatment modalities. Using such algorithmic approaches as accessory tools in the clinic may improve ROP screening in the future.

Deep Learning-assisted Diagnosis of Breast Lesions on US Images: A Multivendor, Multicenter Study.

Purpose: To evaluate the diagnostic performance of a deep learning (DL) model for breast US across four hospitals and assess its value to readers with different levels of experience. Materials and Methods: In this retrospective study, a dual attention-based convolutional neural network was built and validated to discriminate malignant tumors from benign tumors by using B-mode and color Doppler US images (n = 45 909, March 2011-August 2018), acquired with 42 types of US machines, of 9895 pathologic analysis-confirmed breast lesions in 8797 patients (27 men and 8770 women; mean age, 47 years ± 12 [SD]). With and without assistance from the DL model, three novice readers with less than 5 years of US experience and two experienced readers with 8 and 18 years of US experience, respectively, interpreted 1024 randomly selected lesions. Differences in the areas under the receiver operating characteristic curves (AUCs) were tested using the DeLong test. Results: The DL model using both B-mode and color Doppler US images demonstrated expert-level performance at the lesion level, with an AUC of 0.94 (95% CI: 0.92, 0.95) for the internal set. In external datasets, the AUCs were 0.92 (95% CI: 0.90, 0.94) for hospital 1, 0.91 (95% CI: 0.89, 0.94) for hospital 2, and 0.96 (95% CI: 0.94, 0.98) for hospital 3. DL assistance led to improved AUCs (P < .001) for one experienced and three novice radiologists and improved interobserver agreement. The average false-positive rate was reduced by 7.6% (P = .08). Conclusion: The DL model may help radiologists, especially novice readers, improve accuracy and interobserver agreement of breast tumor diagnosis using US.Keywords: Ultrasound, Breast, Diagnosis, Breast Cancer, Deep Learning, Ultrasonography Supplemental material is available for this article. © RSNA, 2023.

Cancer immunotherapy response prediction from multi-modal clinical and image data using semi-supervised deep learning.

BACKGROUND AND PURPOSE: Immunotherapy is a standard treatment for many tumor types. However, only a small proportion of patients derive clinical benefit and reliable predictive biomarkers of immunotherapy response are lacking. Although deep learning has made substantial progress in improving cancer detection and diagnosis, there is limited success on the prediction of treatment response. Here, we aim to predict immunotherapy response of gastric cancer patients using routinely available clinical and image data. MATERIALS AND METHODS: We present a multi-modal deep learning radiomics approach to predict immunotherapy response using both clinical data and computed tomography images. The model was trained using 168 advanced gastric cancer patients treated with immunotherapy. To overcome limitations of small training data, we leverage an additional dataset of 2,029 patients who did not receive immunotherapy in a semi-supervised framework to learn intrinsic imaging phenotypes of the disease. We evaluated model performance in two independent cohorts of 81 patients treated with immunotherapy. RESULTS: The deep learning model achieved area under receiver operating characteristics curve (AUC) of 0.791 (95% CI 0.633-0.950) and 0.812 (95% CI 0.669-0.956) for predicting immunotherapy response in the internal and external validation cohorts. When combined with PD-L1 expression, the integrative model further improved the AUC by 4-7% in absolute terms. CONCLUSION: The deep learning model achieved promising performance for predicting immunotherapy response from routine clinical and image data. The proposed multi-modal approach is general and can incorporate other relevant information to further improve prediction of immunotherapy response.

Comparative validation of machine learning algorithms for surgical workflow and skill analysis with the HeiChole benchmark.

PURPOSE: Surgical workflow and skill analysis are key technologies for the next generation of cognitive surgical assistance systems. These systems could increase the safety of the operation through context-sensitive warnings and semi-autonomous robotic assistance or improve training of surgeons via data-driven feedback. In surgical workflow analysis up to 91% average precision has been reported for phase recognition on an open data single-center video dataset. In this work we investigated the generalizability of phase recognition algorithms in a multicenter setting including more difficult recognition tasks such as surgical action and surgical skill. METHODS: To achieve this goal, a dataset with 33 laparoscopic cholecystectomy videos from three surgical centers with a total operation time of 22 h was created. Labels included framewise annotation of seven surgical phases with 250 phase transitions, 5514 occurences of four surgical actions, 6980 occurences of 21 surgical instruments from seven instrument categories and 495 skill classifications in five skill dimensions. The dataset was used in the 2019 international Endoscopic Vision challenge, sub-challenge for surgical workflow and skill analysis. Here, 12 research teams trained and submitted their machine learning algorithms for recognition of phase, action, instrument and/or skill assessment. RESULTS: F1-scores were achieved for phase recognition between 23.9% and 67.7% (n = 9 teams), for instrument presence detection between 38.5% and 63.8% (n = 8 teams), but for action recognition only between 21.8% and 23.3% (n = 5 teams). The average absolute error for skill assessment was 0.78 (n = 1 team). CONCLUSION: Surgical workflow and skill analysis are promising technologies to support the surgical team, but there is still room for improvement, as shown by our comparison of machine learning algorithms. This novel HeiChole benchmark can be used for comparable evaluation and validation of future work. In future studies, it is of utmost importance to create more open, high-quality datasets in order to allow the development of artificial intelligence and cognitive robotics in surgery.

Deep semi-supervised multiple instance learning with self-correction for DME classification from OCT images.

Supervised deep learning has achieved prominent success in various diabetic macular edema (DME) recognition tasks from optical coherence tomography (OCT) volumetric images. A common problematic issue that frequently occurs in this field is the shortage of labeled data due to the expensive fine-grained annotations, which increases substantial difficulty in accurate analysis by supervised learning. The morphological changes in the retina caused by DME might be distributed sparsely in B-scan images of the OCT volume, and OCT data is often coarsely labeled at the volume level. Hence, the DME identification task can be formulated as a multiple instance classification problem that could be addressed by multiple instance learning (MIL) techniques. Nevertheless, none of previous studies utilize unlabeled data simultaneously to promote the classification accuracy, which is particularly significant for a high quality of analysis at the minimum annotation cost. To this end, we present a novel deep semi-supervised multiple instance learning framework to explore the feasibility of leveraging a small amount of coarsely labeled data and a large amount of unlabeled data to tackle this problem. Specifically, we come up with several modules to further improve the performance according to the availability and granularity of their labels. To warm up the training, we propagate the bag labels to the corresponding instances as the supervision of training, and propose a self-correction strategy to handle the label noise in the positive bags. This strategy is based on confidence-based pseudo-labeling with consistency regularization. The model uses its prediction to generate the pseudo-label for each weakly augmented input only if it is highly confident about the prediction, which is subsequently used to supervise the same input in a strongly augmented version. This learning scheme is also applicable to unlabeled data. To enhance the discrimination capability of the model, we introduce the Student-Teacher architecture and impose consistency constraints between two models. For demonstration, the proposed approach was evaluated on two large-scale DME OCT image datasets. Extensive results indicate that the proposed method improves DME classification with the incorporation of unlabeled data and outperforms competing MIL methods significantly, which confirm the feasibility of deep semi-supervised multiple instance learning at a low annotation cost.

The Liver Tumor Segmentation Benchmark (LiTS).

In this work, we report the set-up and results of the Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with the IEEE International Symposium on Biomedical Imaging (ISBI) 2017 and the International Conferences on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2017 and 2018. The image dataset is diverse and contains primary and secondary tumors with varied sizes and appearances with various lesion-to-background levels (hyper-/hypo-dense), created in collaboration with seven hospitals and research institutions. Seventy-five submitted liver and liver tumor segmentation algorithms were trained on a set of 131 computed tomography (CT) volumes and were tested on 70 unseen test images acquired from different patients. We found that not a single algorithm performed best for both liver and liver tumors in the three events. The best liver segmentation algorithm achieved a Dice score of 0.963, whereas, for tumor segmentation, the best algorithms achieved Dices scores of 0.674 (ISBI 2017), 0.702 (MICCAI 2017), and 0.739 (MICCAI 2018). Retrospectively, we performed additional analysis on liver tumor detection and revealed that not all top-performing segmentation algorithms worked well for tumor detection. The best liver tumor detection method achieved a lesion-wise recall of 0.458 (ISBI 2017), 0.515 (MICCAI 2017), and 0.554 (MICCAI 2018), indicating the need for further research. LiTS remains an active benchmark and resource for research, e.g., contributing the liver-related segmentation tasks in http://medicaldecathlon.com/. In addition, both data and online evaluation are accessible via https://competitions.codalab.org/competitions/17094.

Trans-SVNet: hybrid embedding aggregation Transformer for surgical workflow analysis.

PURPOSE: Real-time surgical workflow analysis has been a key component for computer-assisted intervention system to improve cognitive assistance. Most existing methods solely rely on conventional temporal models and encode features with a successive spatial-temporal arrangement. Supportive benefits of intermediate features are partially lost from both visual and temporal aspects. In this paper, we rethink feature encoding to attend and preserve the critical information for accurate workflow recognition and anticipation. METHODS: We introduce Transformer in surgical workflow analysis, to reconsider complementary effects of spatial and temporal representations. We propose a hybrid embedding aggregation Transformer, named Trans-SVNet, to effectively interact with the designed spatial and temporal embeddings, by employing spatial embedding to query temporal embedding sequence. We jointly optimized by loss objectives from both analysis tasks to leverage their high correlation. RESULTS: We extensively evaluate our method on three large surgical video datasets. Our method consistently outperforms the state-of-the-arts across three datasets on workflow recognition task. Jointly learning with anticipation, recognition results can gain a large improvement. Our approach also shows its effectiveness on anticipation with promising performance achieved. Our model achieves a real-time inference speed of 0.0134 second per frame. CONCLUSION: Experimental results demonstrate the efficacy of our hybrid embeddings integration by rediscovering the crucial cues from complementary spatial-temporal embeddings. The better performance by multi-task learning indicates that anticipation task brings the additional knowledge to recognition task. Promising effectiveness and efficiency of our method also show its promising potential to be used in operating room.

Learning With Privileged Multimodal Knowledge for Unimodal Segmentation.

Multimodal learning usually requires a complete set of modalities during inference to maintain performance. Although training data can be well-prepared with high-quality multiple modalities, in many cases of clinical practice, only one modality can be acquired and important clinical evaluations have to be made based on the limited single modality information. In this work, we propose a privileged knowledge learning framework with the 'Teacher-Student' architecture, in which the complete multimodal knowledge that is only available in the training data (called privileged information) is transferred from a multimodal teacher network to a unimodal student network, via both a pixel-level and an image-level distillation scheme. Specifically, for the pixel-level distillation, we introduce a regularized knowledge distillation loss which encourages the student to mimic the teacher's softened outputs in a pixel-wise manner and incorporates a regularization factor to reduce the effect of incorrect predictions from the teacher. For the image-level distillation, we propose a contrastive knowledge distillation loss which encodes image-level structured information to enrich the knowledge encoding in combination with the pixel-level distillation. We extensively evaluate our method on two different multi-class segmentation tasks, i.e., cardiac substructure segmentation and brain tumor segmentation. Experimental results on both tasks demonstrate that our privileged knowledge learning is effective in improving unimodal segmentation and outperforms previous methods.

Unsupervised feature disentanglement for video retrieval in minimally invasive surgery.

In this paper, we propose a novel method of Unsupervised Disentanglement of Scene and Motion (UDSM) representations for minimally invasive surgery video retrieval within large databases, which has the potential to advance intelligent and efficient surgical teaching systems. To extract more discriminative video representations, two designed encoders with a triplet ranking loss and an adversarial learning mechanism are established to respectively capture the spatial and temporal information for achieving disentangled features from each frame with promising interpretability. In addition, the long-range temporal dependencies are improved in an integrated video level using a temporal aggregation module and then a set of compact binary codes that carries representative features is yielded to realize fast retrieval. The entire framework is trained in an unsupervised scheme, i.e., purely learning from raw surgical videos without using any annotation. We construct two large-scale minimally invasive surgery video datasets based on the public dataset Cholec80 and our in-house dataset of laparoscopic hysterectomy, to establish the learning process and validate the effectiveness of our proposed method qualitatively and quantitatively on the surgical video retrieval task. Extensive experiments show that our approach significantly outperforms the state-of-the-art video retrieval methods on both datasets, revealing a promising future for injecting intelligence in the next generation of surgical teaching systems.

Real-time landmark detection for precise endoscopic submucosal dissection via shape-aware relation network.

We propose a novel shape-aware relation network for accurate and real-time landmark detection in endoscopic submucosal dissection (ESD) surgery. This task is of great clinical significance but extremely challenging due to bleeding, lighting reflection, and motion blur in the complicated surgical environment. Compared with existing solutions, which either neglect geometric relationships among targeting objects or capture the relationships by using complicated aggregation schemes, the proposed network is capable of achieving satisfactory accuracy while maintaining real-time performance by taking full advantage of the spatial relations among landmarks. We first devise an algorithm to automatically generate relation keypoint heatmaps, which are able to intuitively represent the prior knowledge of spatial relations among landmarks without using any extra manual annotation efforts. We then develop two complementary regularization schemes to progressively incorporate the prior knowledge into the training process. While one scheme introduces pixel-level regularization by multi-task learning, the other integrates global-level regularization by harnessing a newly designed grouped consistency evaluator, which adds relation constraints to the proposed network in an adversarial manner. Both schemes are beneficial to the model in training, and can be readily unloaded in inference to achieve real-time detection. We establish a large in-house dataset of ESD surgery for esophageal cancer to validate the effectiveness of our proposed method. Extensive experimental results demonstrate that our approach outperforms state-of-the-art methods in terms of accuracy and efficiency, achieving better detection results faster. Promising results on two downstream applications further corroborate the great potential of our method in ESD clinical practice.

Anchor-guided online meta adaptation for fast one-Shot instrument segmentation from robotic surgical videos.

The scarcity of annotated surgical data in robot-assisted surgery (RAS) motivates prior works to borrow related domain knowledge to achieve promising segmentation results in surgical images by adaptation. For dense instrument tracking in a robotic surgical video, collecting one initial scene to specify target instruments (or parts of tools) is desirable and feasible during the preoperative preparation. In this paper, we study the challenging one-shot instrument segmentation for robotic surgical videos, in which only the first frame mask of each video is provided at test time, such that the pre-trained model (learned from easily accessible source) can adapt to the target instruments. Straightforward methods transfer the domain knowledge by fine-tuning the model on each given mask. Such one-shot optimization takes hundred of iterations and the test runtime is unfeasible. We present anchor-guided online meta adaptation (AOMA) for this problem. We achieve fast one-shot test time optimization by meta-learning a good model initialization and learning rates from source videos to avoid the laborious and handcrafted fine-tuning. The trainable two components are optimized in a video-specific task space with a matching-aware loss. Furthermore, we design an anchor-guided online adaptation to tackle the performance drop throughout a robotic surgical sequence. The model is continuously adapted on motion-insensitive pseudo-masks supported by anchor matching. AOMA achieves state-of-the-art results on two practical scenarios: (1) general videos to surgical videos, (2) public surgical videos to in-house surgical videos, while reducing the test runtime substantially.

Versatile multi-constrained planning for thermal ablation of large liver tumors.

The surgical planning of large hepatic tumor ablation remains a challenging task that relies on fulfilling multiple medical constraints, especially for the ablation based on configurations of multiple electrodes. The placement of the electrodes to completely ablate the tumor as well as their insertion trajectory to their final position have to be planned to cause as little damage to healthy anatomical structures as possible to allow a fast rehabilitation. In this paper, we present a novel, versatile approach for the computer-assisted planning of multi-electrode thermal ablation of large liver tumors based on pre-operative CT data with semantic annotations. This involves both the specification of the number of required electrodes and their distribution to adequately ablate the tumor region without damaging too much healthy tissue. To determine the insertion trajectory of the electrodes to their final position, we additionally incorporate a series of medical constraints into our optimization, which allows a global analysis where obstacles such as bones are taken into account and damage to healthy tissue is mitigated. Compared with the state-of-the-art method, our method achieves compact ablation regions without relying on assumptions on a potential needle path for optimal global search and, hence, is suitable for guiding clinicians through the planning of the tumor ablation. We also demonstrate the feasibility of our approach in various experiments of clinical data and demonstrate that our approach not only allows completely ablating the tumor region but also reducing the damage of healthy tissue in comparison to the previous state-of-the-art method.

A Multitask Deep-Learning System to Classify Diabetic Macular Edema for Different Optical Coherence Tomography Devices: A Multicenter Analysis.

OBJECTIVE: Diabetic macular edema (DME) is the primary cause of vision loss among individuals with diabetes mellitus (DM). We developed, validated, and tested a deep learning (DL) system for classifying DME using images from three common commercially available optical coherence tomography (OCT) devices. RESEARCH DESIGN AND METHODS: We trained and validated two versions of a multitask convolution neural network (CNN) to classify DME (center-involved DME [CI-DME], non-CI-DME, or absence of DME) using three-dimensional (3D) volume scans and 2D B-scans, respectively. For both 3D and 2D CNNs, we used the residual network (ResNet) as the backbone. For the 3D CNN, we used a 3D version of ResNet-34 with the last fully connected layer removed as the feature extraction module. A total of 73,746 OCT images were used for training and primary validation. External testing was performed using 26,981 images across seven independent data sets from Singapore, Hong Kong, the U.S., China, and Australia. RESULTS: In classifying the presence or absence of DME, the DL system achieved area under the receiver operating characteristic curves (AUROCs) of 0.937 (95% CI 0.920-0.954), 0.958 (0.930-0.977), and 0.965 (0.948-0.977) for the primary data set obtained from CIRRUS, SPECTRALIS, and Triton OCTs, respectively, in addition to AUROCs >0.906 for the external data sets. For further classification of the CI-DME and non-CI-DME subgroups, the AUROCs were 0.968 (0.940-0.995), 0.951 (0.898-0.982), and 0.975 (0.947-0.991) for the primary data set and >0.894 for the external data sets. CONCLUSIONS: We demonstrated excellent performance with a DL system for the automated classification of DME, highlighting its potential as a promising second-line screening tool for patients with DM, which may potentially create a more effective triaging mechanism to eye clinics.

Semi-supervised learning with progressive unlabeled data excavation for label-efficient surgical workflow recognition.

Surgical workflow recognition is a fundamental task in computer-assisted surgery and a key component of various applications in operating rooms. Existing deep learning models have achieved promising results for surgical workflow recognition, heavily relying on a large amount of annotated videos. However, obtaining annotation is time-consuming and requires the domain knowledge of surgeons. In this paper, we propose a novel two-stage Semi-Supervised Learning method for label-efficient Surgical workflow recognition, named as SurgSSL. Our proposed SurgSSL progressively leverages the inherent knowledge held in the unlabeled data to a larger extent: from implicit unlabeled data excavation via motion knowledge excavation, to explicit unlabeled data excavation via pre-knowledge pseudo labeling. Specifically, we first propose a novel intra-sequence Visual and Temporal Dynamic Consistency (VTDC) scheme for implicit excavation. It enforces prediction consistency of the same data under perturbations in both spatial and temporal spaces, encouraging model to capture rich motion knowledge. We further perform explicit excavation by optimizing the model towards our pre-knowledge pseudo label. It is naturally generated by the VTDC regularized model with prior knowledge of unlabeled data encoded, and demonstrates superior reliability for model supervision compared with the label generated by existing methods. We extensively evaluate our method on two public surgical datasets of Cholec80 and M2CAI challenge dataset. Our method surpasses the state-of-the-art semi-supervised methods by a large margin, e.g., improving 10.5% Accuracy under the severest annotation regime of M2CAI dataset. Using only 50% labeled videos on Cholec80, our approach achieves competitive performance compared with full-data training method.

Towards quantitative and intuitive percutaneous tumor puncture via augmented virtual reality.

In recent years, the radiofrequency ablation (RFA) therapy has become a widely accepted minimal invasive treatment for liver tumor patients. However, it is challenging for doctors to precisely and efficiently perform the percutaneous tumor punctures under free-breathing conditions. This is because the traditional RFA is based on the 2D CT Image information, the missing spatial and dynamic information is dependent on surgeons' experience. This paper presents a novel quantitative and intuitive surgical navigation modality for percutaneous respiratory tumor puncture via augmented virtual reality, which is to achieve the augmented visualization of the pre-operative virtual planning information precisely being overlaid on intra-operative surgical scenario. In the pre-operation stage, we first combine the signed distance field of feasible structures (like liver and tumor) where the puncture path can go through and unfeasible structures (like large vessels and ribs) where the needle is not allowed to go through to quantitatively generate the 3D feasible region for percutaneous puncture. Then we design three constraints according to the RFA specialists consensus to automatically determine the optimal puncture trajectory. In the intra-operative stage, we first propose a virtual-real alignment method to precisely superimpose the virtual information on surgical scenario. Then, a user-friendly collaborative holographic interface is designed for real-time 3D respiratory tumor puncture navigation, which can effectively assist surgeons fast and accurately locating the target step-by step. The validation of our system is performed on static abdominal phantom and in vivo beagle dogs with artificial lesion. Experimental results demonstrate that the accuracy of the proposed planning strategy is better than the manual planning sketched by experienced doctors. Besides, the proposed holographic navigation modality can effectively reduce the needle adjustment for precise puncture as well. Our system shows its clinical feasibility to provide the quantitative planning of optimal needle path and intuitive in situ holographic navigation for percutaneous tumor ablation without surgeons' experience-dependence and reduce the times of needle adjustment. The proposed augmented virtual reality navigation system can effectively improve the precision and reliability in percutaneous tumor ablation and has the potential to be used for other surgical navigation tasks.

Automatic identification of sweet spots from MERs for electrodes implantation in STN-DBS.

PURPOSE: Microelectrode recordings (MERs) are a significant clinical indicator for sweet spots identification of implanted electrodes during deep brain stimulation of the subthalamic nucleus (STN) surgery. As 1D MERs signals have the unboundedness, large-range, large-amount and time-dependent characteristics, the purpose of this study is to propose an automatic and precise identification method of sweet spots from MERs, reducing the time-consuming and labor-intensive human annotations. METHODS: We propose an automatic identification method of sweet spots from MERs for electrodes implantation in STN-DBS. To better imitate the surgeons' observation and obtain more intuitive contextual information, we first employ the 2D Gramian angular summation field (GASF) images generated from MERs data to perform the sweet spots determination for electrodes implantation. Then, we introduce the convolutional block attention module into convolutional neural network (CNN) to identify the 2D GASF images of sweet spots for electrodes implantation. RESULTS: Experimental results illustrate that the identification result of our method is consistent with the result of doctor's decision, while our method can achieve the accuracy and precision of 96.72% and 98.97%, respectively, which outperforms state-of-the-art for intraoperative sweet spots determination. CONCLUSIONS: The proposed method is the first time to automatically and accurately identify sweet spots from MERs for electrodes implantation by the combination an advanced time series-to-image encoding way with CBAM-enhanced networks model. Our method can assist neurosurgeons in automatically detecting the most likely locations of sweet spots for electrodes implantation, which can provide an important indicator for target selection while it reduces the localization error of the target during STN-DBS surgery.

Deep virtual adversarial self-training with consistency regularization for semi-supervised medical image classification.

Convolutional neural networks have achieved prominent success on a variety of medical imaging tasks when a large amount of labeled training data is available. However, the acquisition of expert annotations for medical data is usually expensive and time-consuming, which poses a great challenge for supervised learning approaches. In this work, we proposed a novel semi-supervised deep learning method, i.e., deep virtual adversarial self-training with consistency regularization, for large-scale medical image classification. To effectively exploit useful information from unlabeled data, we leverage self-training and consistency regularization to harness the underlying knowledge, which helps improve the discrimination capability of training models. More concretely, the model first uses its prediction for pseudo-labeling on the weakly-augmented input image. A pseudo-label is kept only if the corresponding class probability is of high confidence. Then the model prediction is encouraged to be consistent with the strongly-augmented version of the same input image. To improve the robustness of the network against virtual adversarial perturbed input, we incorporate virtual adversarial training (VAT) on both labeled and unlabeled data into the course of training. Hence, the network is trained by minimizing a combination of three types of losses, including a standard supervised loss on labeled data, a consistency regularization loss on unlabeled data, and a VAT loss on both labeled and labeled data. We extensively evaluate the proposed semi-supervised deep learning methods on two challenging medical image classification tasks: breast cancer screening from ultrasound images and multi-class ophthalmic disease classification from optical coherence tomography B-scan images. Experimental results demonstrate that the proposed method outperforms both supervised baseline and other state-of-the-art methods by a large margin on all tasks.

Comparative validation of multi-instance instrument segmentation in endoscopy: Results of the ROBUST-MIS 2019 challenge.

Intraoperative tracking of laparoscopic instruments is often a prerequisite for computer and robotic-assisted interventions. While numerous methods for detecting, segmenting and tracking of medical instruments based on endoscopic video images have been proposed in the literature, key limitations remain to be addressed: Firstly, robustness, that is, the reliable performance of state-of-the-art methods when run on challenging images (e.g. in the presence of blood, smoke or motion artifacts). Secondly, generalization; algorithms trained for a specific intervention in a specific hospital should generalize to other interventions or institutions. In an effort to promote solutions for these limitations, we organized the Robust Medical Instrument Segmentation (ROBUST-MIS) challenge as an international benchmarking competition with a specific focus on the robustness and generalization capabilities of algorithms. For the first time in the field of endoscopic image processing, our challenge included a task on binary segmentation and also addressed multi-instance detection and segmentation. The challenge was based on a surgical data set comprising 10,040 annotated images acquired from a total of 30 surgical procedures from three different types of surgery. The validation of the competing methods for the three tasks (binary segmentation, multi-instance detection and multi-instance segmentation) was performed in three different stages with an increasing domain gap between the training and the test data. The results confirm the initial hypothesis, namely that algorithm performance degrades with an increasing domain gap. While the average detection and segmentation quality of the best-performing algorithms is high, future research should concentrate on detection and segmentation of small, crossing, moving and transparent instrument(s) (parts).

Global guidance network for breast lesion segmentation in ultrasound images.

Automatic breast lesion segmentation in ultrasound helps to diagnose breast cancer, which is one of the dreadful diseases that affect women globally. Segmenting breast regions accurately from ultrasound image is a challenging task due to the inherent speckle artifacts, blurry breast lesion boundaries, and inhomogeneous intensity distributions inside the breast lesion regions. Recently, convolutional neural networks (CNNs) have demonstrated remarkable results in medical image segmentation tasks. However, the convolutional operations in a CNN often focus on local regions, which suffer from limited capabilities in capturing long-range dependencies of the input ultrasound image, resulting in degraded breast lesion segmentation accuracy. In this paper, we develop a deep convolutional neural network equipped with a global guidance block (GGB) and breast lesion boundary detection (BD) modules for boosting the breast ultrasound lesion segmentation. The GGB utilizes the multi-layer integrated feature map as a guidance information to learn the long-range non-local dependencies from both spatial and channel domains. The BD modules learn additional breast lesion boundary map to enhance the boundary quality of a segmentation result refinement. Experimental results on a public dataset and a collected dataset show that our network outperforms other medical image segmentation methods and the recent semantic segmentation methods on breast ultrasound lesion segmentation. Moreover, we also show the application of our network on the ultrasound prostate segmentation, in which our method better identifies prostate regions than state-of-the-art networks.