Search

Segment anything model (SAM) demonstrates strong generalization ability on natural image segmentation. However, its direct adaption in medical image segmentation tasks shows significant performance drops with inferior accuracy and unstable results. It may also requires an excessive number of prompt points to obtain a reasonable accuracy. For segmenting 3D radiological CT or MRI scans, a 2D SAM model has to separately handle hundreds of 2D slices. Although quite a few studies explore adapting SAM into medical image volumes, the efficiency of 2D adaption methods is unsatisfactory and 3D adaptation methods only capable of segmenting specific organs/tumors. In this work, we propose a comprehensive and scalable 3D SAM model for whole-body CT segmentation, named CT-SAM3D. Instead of adapting SAM, we propose a 3D promptable segmentation model using a (nearly) fully labeled CT dataset. To train CT-SAM3D effectively, ensuring the model's accurate responses to higher-dimensional spatial prompts is crucial, and 3D patch-wise training is required due to GPU memory constraints. For this purpose, we propose two key technical developments: 1) a progressively and spatially aligned prompt encoding method to effectively encode click prompts in local 3D space; and 2) a cross-patch prompt learning scheme to capture more 3D spatial context, which is beneficial for reducing the editing workloads when interactively prompting on large organs. CT-SAM3D is trained and validated using a curated dataset of 1204 CT scans containing 107 whole-body anatomies, reporting significantly better quantitative performance against all previous SAM-derived models by a large margin with much fewer click prompts. Our model can handle segmenting unseen organ as well. Code, data, and our 3D interactive segmentation tool with quasi-real-time responses will be made publicly available.

Establishing dense anatomical correspondence across distinct imaging modalities is a foundational yet challenging procedure for numerous medical image analysis studies and image-guided radiotherapy. Existing multi-modality image registration algorithms rely on statistical-based similarity measures or local structural image representations. However, the former is sensitive to locally varying noise, while the latter is not discriminative enough to cope with complex anatomical structures in multimodal scans, causing ambiguity in determining the anatomical correspondence across scans with different modalities. In this paper, we propose a modality-agnostic structural representation learning method, which leverages Deep Neighbourhood Self-similarity (DNS) and anatomy-aware contrastive learning to learn discriminative and contrast-invariance deep structural image representations (DSIR) without the need for anatomical delineations or pre-aligned training images. We evaluate our method on multiphase CT, abdomen MR-CT, and brain MR T1w-T2w registration. Comprehensive results demonstrate that our method is superior to the conventional local structural representation and statistical-based similarity measures in terms of discriminability and accuracy., Comment: Accepted by CVPR2024

The absence of adequately sufficient expert-level tumor annotations hinders the effectiveness of supervised learning based opportunistic cancer screening on medical imaging. Clinical reports (that are rich in descriptive textual details) can offer a "free lunch'' supervision information and provide tumor location as a type of weak label to cope with screening tasks, thus saving human labeling workloads, if properly leveraged. However, predicting cancer only using such weak labels can be very changeling since tumors are usually presented in small anatomical regions compared to the whole 3D medical scans. Weakly semi-supervised learning (WSSL) utilizes a limited set of voxel-level tumor annotations and incorporates alongside a substantial number of medical images that have only off-the-shelf clinical reports, which may strike a good balance between minimizing expert annotation workload and optimizing screening efficacy. In this paper, we propose a novel text-guided learning method to achieve highly accurate cancer detection results. Through integrating diagnostic and tumor location text prompts into the text encoder of a vision-language model (VLM), optimization of weakly supervised learning can be effectively performed in the latent space of VLM, thereby enhancing the stability of training. Our approach can leverage clinical knowledge by large-scale pre-trained VLM to enhance generalization ability, and produce reliable pseudo tumor masks to improve cancer detection. Our extensive quantitative experimental results on a large-scale cancer dataset, including 1,651 unique patients, validate that our approach can reduce human annotation efforts by at least 70% while maintaining comparable cancer detection accuracy to competing fully supervised methods (AUC value 0.961 versus 0.966).

In the realm of medical 3D data, such as CT and MRI images, prevalent anisotropic resolution is characterized by high intra-slice but diminished inter-slice resolution. The lowered resolution between adjacent slices poses challenges, hindering optimal viewing experiences and impeding the development of robust downstream analysis algorithms. Various volumetric super-resolution algorithms aim to surmount these challenges, enhancing inter-slice resolution and overall 3D medical imaging quality. However, existing approaches confront inherent challenges: 1) often tailored to specific upsampling factors, lacking flexibility for diverse clinical scenarios; 2) newly generated slices frequently suffer from over-smoothing, degrading fine details, and leading to inter-slice inconsistency. In response, this study presents CycleINR, a novel enhanced Implicit Neural Representation model for 3D medical data volumetric super-resolution. Leveraging the continuity of the learned implicit function, the CycleINR model can achieve results with arbitrary up-sampling rates, eliminating the need for separate training. Additionally, we enhance the grid sampling in CycleINR with a local attention mechanism and mitigate over-smoothing by integrating cycle-consistent loss. We introduce a new metric, Slice-wise Noise Level Inconsistency (SNLI), to quantitatively assess inter-slice noise level inconsistency. The effectiveness of our approach is demonstrated through image quality evaluations on an in-house dataset and a downstream task analysis on the Medical Segmentation Decathlon liver tumor dataset., Comment: CVPR accepted paper

Radiologists highly desire fully automated versatile AI for medical imaging interpretation. However, the lack of extensively annotated large-scale multi-disease datasets has hindered the achievement of this goal. In this paper, we explore the feasibility of leveraging language as a naturally high-quality supervision for chest CT imaging. In light of the limited availability of image-report pairs, we bootstrap the understanding of 3D chest CT images by distilling chest-related diagnostic knowledge from an extensively pre-trained 2D X-ray expert model. Specifically, we propose a language-guided retrieval method to match each 3D CT image with its semantically closest 2D X-ray image, and perform pair-wise and semantic relation knowledge distillation. Subsequently, we use contrastive learning to align images and reports within the same patient while distinguishing them from the other patients. However, the challenge arises when patients have similar semantic diagnoses, such as healthy patients, potentially confusing if treated as negatives. We introduce a robust contrastive learning that identifies and corrects these false negatives. We train our model with over 12,000 pairs of chest CT images and radiology reports. Extensive experiments across multiple scenarios, including zero-shot learning, report generation, and fine-tuning processes, demonstrate the model's feasibility in interpreting chest CT images., Comment: Accepted by CVPR 2024

Item does not contain fulltext

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https:// learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https:// learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https:// learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https:// learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https:// learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https:// learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Estimating displacement vector field via a cost volume computed in the feature space has shown great success in image registration, but it suffers excessive computation burdens. Moreover, existing feature descriptors only extract local features incapable of representing the global semantic information, which is especially important for solving large transformations. To address the discussed issues, we propose SAMConvex, a fast coarse-to-fine discrete optimization method for CT registration that includes a decoupled convex optimization procedure to obtain deformation fields based on a self-supervised anatomical embedding (SAM) feature extractor that captures both local and global information. To be specific, SAMConvex extracts per-voxel features and builds 6D correlation volumes based on SAM features, and iteratively updates a flow field by performing lookups on the correlation volumes with a coarse-to-fine scheme. SAMConvex outperforms the state-of-the-art learning-based methods and optimization-based methods over two inter-patient registration datasets (Abdomen CT and HeadNeck CT) and one intra-patient registration dataset (Lung CT). Moreover, as an optimization-based method, SAMConvex only takes $\sim2$s ($\sim5s$ with instance optimization) for one paired images.

Radiotherapists require accurate registration of MR/CT images to effectively use information from both modalities. In a typical registration pipeline, rigid or affine transformations are applied to roughly align the fixed and moving images before proceeding with the deformation step. While recent learning-based methods have shown promising results in the rigid/affine step, these methods often require images with similar field-of-view (FOV) for successful alignment. As a result, aligning images with different FOVs remains a challenging task. Self-supervised landmark detection methods like self-supervised Anatomical eMbedding (SAM) have emerged as a useful tool for mapping and cropping images to similar FOVs. However, these methods are currently limited to intra-modality use only. To address this limitation and enable cross-modality matching, we propose a new approach called Cross-SAM. Our approach utilizes a novel iterative process that alternates between embedding learning and CT-MRI registration. We start by applying aggressive contrast augmentation on both CT and MRI images to train a SAM model. We then use this SAM to identify corresponding regions on paired images using robust grid-points matching, followed by a point-set based affine/rigid registration, and a deformable fine-tuning step to produce registered paired images. We use these registered pairs to enhance the matching ability of SAM, which is then processed iteratively. We use the final model for cross-modality matching tasks. We evaluated our approach on two CT-MRI affine registration datasets and found that Cross-SAM achieved robust affine registration on both datasets, significantly outperforming other methods and achieving state-of-the-art performance.

State-of-the-art deep learning-based registration methods employ three different learning strategies: supervised learning, which requires costly manual annotations, unsupervised learning, which heavily relies on hand-crafted similarity metrics designed by domain experts, or learning from synthetic data, which introduces a domain shift. To overcome the limitations of these strategies, we propose a novel self-supervised learning paradigm for unsupervised registration, relying on self-training. Our idea is based on two key insights. Feature-based differentiable optimizers 1) perform reasonable registration even from random features and 2) stabilize the training of the preceding feature extraction network on noisy labels. Consequently, we propose cyclical self-training, where pseudo labels are initialized as the displacement fields inferred from random features and cyclically updated based on more and more expressive features from the learning feature extractor, yielding a self-reinforcement effect. We evaluate the method for abdomen and lung registration, consistently surpassing metric-based supervision and outperforming diverse state-of-the-art competitors. Source code is available at https://github.com/multimodallearning/reg-cyclical-self-train., Comment: accepted at MICCAI 2023

The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bottlenecks faced by the community in tackling the research questions posed. To shed light on the status quo of algorithm development in the specific field of biomedical imaging analysis, we designed an international survey that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences (80 competitions in total). The survey covered participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics. A median of 72% challenge participants took part in the survey. According to our results, knowledge exchange was the primary incentive (70%) for participation, while the reception of prize money played only a minor role (16%). While a median of 80 working hours was spent on method development, a large portion of participants stated that they did not have enough time for method development (32%). 25% perceived the infrastructure to be a bottleneck. Overall, 94% of all solutions were deep learning-based. Of these, 84% were based on standard architectures. 43% of the respondents reported that the data samples (e.g., images) were too large to be processed at once. This was most commonly addressed by patch-based training (69%), downsampling (37%), and solving 3D analysis tasks as a series of 2D tasks. K-fold cross-validation on the training set was performed by only 37% of the participants and only 50% of the participants performed ensembling based on multiple identical models (61%) or heterogeneous models (39%). 48% of the respondents applied postprocessing steps.

Registration of pre-operative and follow-up brain MRI scans is challenging due to the large variation of tissue appearance and missing correspondences in tumour recurrence regions caused by tumour mass effect. Although recent deep learning-based deformable registration methods have achieved remarkable success in various medical applications, most of them are not capable of registering images with pathologies. In this paper, we propose a 3-step registration pipeline for pre-operative and follow-up brain MRI scans that consists of 1) a multi-level affine registration, 2) a conditional deep Laplacian pyramid image registration network (cLapIRN) with forward-backward consistency constraint, and 3) a non-linear instance optimization method. We apply the method to the Brain Tumor Sequence Registration (BraTS-Reg) Challenge. Our method achieves accurate and robust registration of brain MRI scans with pathologies, which achieves a median absolute error of 1.64 mm and 88\% of successful registration rate in the validation set of BraTS-Reg challenge. Our method ranks 1st place in the 2022 MICCAI BraTS-Reg challenge., Comment: BraTS-Reg workshop

Affine registration is indispensable in a comprehensive medical image registration pipeline. However, only a few studies focus on fast and robust affine registration algorithms. Most of these studies utilize convolutional neural networks (CNNs) to learn joint affine and non-parametric registration, while the standalone performance of the affine subnetwork is less explored. Moreover, existing CNN-based affine registration approaches focus either on the local misalignment or the global orientation and position of the input to predict the affine transformation matrix, which are sensitive to spatial initialization and exhibit limited generalizability apart from the training dataset. In this paper, we present a fast and robust learning-based algorithm, Coarse-to-Fine Vision Transformer (C2FViT), for 3D affine medical image registration. Our method naturally leverages the global connectivity and locality of the convolutional vision transformer and the multi-resolution strategy to learn the global affine registration. We evaluate our method on 3D brain atlas registration and template-matching normalization. Comprehensive results demonstrate that our method is superior to the existing CNNs-based affine registration methods in terms of registration accuracy, robustness and generalizability while preserving the runtime advantage of the learning-based methods. The source code is available at https://github.com/cwmok/C2FViT., Comment: Accepted by CVPR2022

Registration of pre-operative and post-recurrence brain images is often needed to evaluate the effectiveness of brain gliomas treatment. While recent deep learning-based deformable registration methods have achieved remarkable success with healthy brain images, most of them would be unable to accurately align images with pathologies due to the absent correspondences in the reference image. In this paper, we propose a deep learning-based deformable registration method that jointly estimates regions with absent correspondence and bidirectional deformation fields. A forward-backward consistency constraint is used to aid in the localization of the resection and recurrence region from voxels with absence correspondences in the two images. Results on 3D clinical data from the BraTS-Reg challenge demonstrate our method can improve image alignment compared to traditional and deep learning-based registration approaches with or without cost function masking strategy. The source code is available at https://github.com/cwmok/DIRAC., Comment: Accepted by MICCAI2022

Registration of longitudinal brain MRI scans containing pathologies is challenging due to dramatic changes in tissue appearance. Although there has been progress in developing general-purpose medical image registration techniques, they have not yet attained the requisite precision and reliability for this task, highlighting its inherent complexity. Here we describe the Brain Tumor Sequence Registration (BraTS-Reg) challenge, as the first public benchmark environment for deformable registration algorithms focusing on estimating correspondences between pre-operative and follow-up scans of the same patient diagnosed with a diffuse brain glioma. The BraTS-Reg data comprise de-identified multi-institutional multi-parametric MRI (mpMRI) scans, curated for size and resolution according to a canonical anatomical template, and divided into training, validation, and testing sets. Clinical experts annotated ground truth (GT) landmark points of anatomical locations distinct across the temporal domain. Quantitative evaluation and ranking were based on the Median Euclidean Error (MEE), Robustness, and the determinant of the Jacobian of the displacement field. The top-ranked methodologies yielded similar performance across all evaluation metrics and shared several methodological commonalities, including pre-alignment, deep neural networks, inverse consistency analysis, and test-time instance optimization per-case basis as a post-processing step. The top-ranked method attained the MEE at or below that of the inter-rater variability for approximately 60% of the evaluated landmarks, underscoring the scope for further accuracy and robustness improvements, especially relative to human experts. The aim of BraTS-Reg is to continue to serve as an active resource for research, with the data and online evaluation tools accessible at https://bratsreg.github.io/.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https://learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Recent deep learning-based methods have shown promising results and runtime advantages in deformable image registration. However, analyzing the effects of hyperparameters and searching for optimal regularization parameters prove to be too prohibitive in deep learning-based methods. This is because it involves training a substantial number of separate models with distinct hyperparameter values. In this paper, we propose a conditional image registration method and a new self-supervised learning paradigm for deep deformable image registration. By learning the conditional features that are correlated with the regularization hyperparameter, we demonstrate that optimal solutions with arbitrary hyperparameters can be captured by a single deep convolutional neural network. In addition, the smoothness of the resulting deformation field can be manipulated with arbitrary strength of smoothness regularization during inference. Extensive experiments on a large-scale brain MRI dataset show that our proposed method enables the precise control of the smoothness of the deformation field without sacrificing the runtime advantage or registration accuracy., Comment: Early accepted by MICCAI2021

Diffeomorphic deformable image registration is crucial in many medical image studies, as it offers unique, special features including topology preservation and invertibility of the transformation. Recent deep learning-based deformable image registration methods achieve fast image registration by leveraging a convolutional neural network (CNN) to learn the spatial transformation from the synthetic ground truth or the similarity metric. However, these approaches often ignore the topology preservation of the transformation and the smoothness of the transformation which is enforced by a global smoothing energy function alone. Moreover, deep learning-based approaches often estimate the displacement field directly, which cannot guarantee the existence of the inverse transformation. In this paper, we present a novel, efficient unsupervised symmetric image registration method which maximizes the similarity between images within the space of diffeomorphic maps and estimates both forward and inverse transformations simultaneously. We evaluate our method on 3D image registration with a large scale brain image dataset. Our method achieves state-of-the-art registration accuracy and running time while maintaining desirable diffeomorphic properties.

Diffeomorphic deformable image registration is crucial in many medical image studies, as it offers unique, special features including topology preservation and invertibility of the transformation. Recent deep learning-based deformable image registration methods achieve fast image registration by leveraging a convolutional neural network (CNN) to learn the spatial transformation from the synthetic ground truth or the similarity metric. However, these approaches often ignore the topology preservation of the transformation and the smoothness of the transformation which is enforced by a global smoothing energy function alone. Moreover, deep learning-based approaches often estimate the displacement field directly, which cannot guarantee the existence of the inverse transformation. In this paper, we present a novel, efficient unsupervised symmetric image registration method which maximizes the similarity between images within the space of diffeomorphic maps and estimates both forward and inverse transformations simultaneously. We evaluate our method on 3D image registration with a large scale brain image dataset. Our method achieves state-of-the-art registration accuracy and running time while maintaining desirable diffeomorphic properties.

Diffeomorphic deformable image registration is crucial in many medical image studies, as it offers unique, special features including topology preservation and invertibility of the transformation. Recent deep learning-based deformable image registration methods achieve fast image registration by leveraging a convolutional neural network (CNN) to learn the spatial transformation from the synthetic ground truth or the similarity metric. However, these approaches often ignore the topology preservation of the transformation and the smoothness of the transformation which is enforced by a global smoothing energy function alone. Moreover, deep learning-based approaches often estimate the displacement field directly, which cannot guarantee the existence of the inverse transformation. In this paper, we present a novel, efficient unsupervised symmetric image registration method which maximizes the similarity between images within the space of diffeomorphic maps and estimates both forward and inverse transformations simultaneously. We evaluate our method on 3D image registration with a large scale brain image dataset. Our method achieves state-of-the-art registration accuracy and running time while maintaining desirable diffeomorphic properties.

Deep learning-based methods have recently demonstrated promising results in deformable image registration for a wide range of medical image analysis tasks. However, existing deep learning-based methods are usually limited to small deformation settings, and desirable properties of the transformation including bijective mapping and topology preservation are often being ignored by these approaches. In this paper, we propose a deep Laplacian Pyramid Image Registration Network, which can solve the image registration optimization problem in a coarse-to-fine fashion within the space of diffeomorphic maps. Extensive quantitative and qualitative evaluations on two MR brain scan datasets show that our method outperforms the existing methods by a significant margin while maintaining desirable diffeomorphic properties and promising registration speed., Comment: Paper accepted by MICCAI 2020

Diffeomorphic deformable image registration is crucial in many medical image studies, as it offers unique, special properties including topology preservation and invertibility of the transformation. Recent deep learning-based deformable image registration methods achieve fast image registration by leveraging a convolutional neural network (CNN) to learn the spatial transformation from the synthetic ground truth or the similarity metric. However, these approaches often ignore the topology preservation of the transformation and the smoothness of the transformation which is enforced by a global smoothing energy function alone. Moreover, deep learning-based approaches often estimate the displacement field directly, which cannot guarantee the existence of the inverse transformation. In this paper, we present a novel, efficient unsupervised symmetric image registration method which maximizes the similarity between images within the space of diffeomorphic maps and estimates both forward and inverse transformations simultaneously. We evaluate our method on 3D image registration with a large scale brain image dataset. Our method achieves state-of-the-art registration accuracy and running time while maintaining desirable diffeomorphic properties., Comment: Accepted to CVPR 2020

[Background[ Epidemiological observations have demonstrated that ambient fine particulate matter with dp < 2.5 μm (PM2.5) as the major factor responsible for the increasing incidence of lung cancer in never-smokers. However, there are very limited experimental data to support the association of PM2.5 with lung carcinogenesis and to compare PM2.5 with smoking carcinogens. [Methods] To study whether PM2.5 can contribute to lung tumorigenesis in a way similar to smoking carcinogen 4-methylnitrosamino-l-3-pyridyl-butanone (NNK) via 15-lipoxygenases (15-LOXs) reduction, normal lung epithelial cells and cancer cells were treated with NNK or PM2.5 and then epigenetically and post-translationally examined the cellular and molecular profiles of the cells. The data were verified in lung cancer samples and a mouse lung tumor model. [Results] We found that similar to smoking carcinogen NNK, PM2.5 significantly enhanced cell proliferation, migration and invasion, but reduced the levels of 15-lipoxygenases-1 (15-LOX1) and 15-lipoxygenases-2 (15-LOX2), both of which were also obviously decreased in lung cancer tissues. 15-LOX1/15-LOX2 overexpression inhibited the oncogenic cell functions induced by PM2.5/NNK. The tumor formation and growth were significantly higher/faster in mice implanted with PM2.5- or NNK-treated NCI-H23 cells, accompanied with a reduction of 15-LOX1/15-LOX2. Moreover, 15-LOX1 expression was epigenetically regulated at methylation level by PM2.5/NNK, while both 15-LOX1 and 15-LOX2 could be significantly inhibited by a set of PM2.5/NNK-mediated microRNAs. [Conclusion] Collectively, PM2.5 can function as the smoking carcinogen NNK to induce lung tumorigenesis by inhibiting 15-LOX1/15-LOX2.

Objectives: With the availability of several epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs), sequential therapy could potentially render EGFR mutation-positive non-small cell lung cancer a chronic disease in some patients. In this retrospective analysis of EGFR mutation-positive (Del19/L858R) patients receiving first-line afatinib in LUX-Lung 3, 6, and 7, we assessed uptake of, and outcomes following, subsequent therapies including the third-generation EGFR TKI, osimertinib. Methods: Post-progression therapy data were prospectively collected during follow-up. Molecular testing of tumours at progression/discontinuation of afatinib was not mandatory. Duration of subsequent therapies, and survival following osimertinib, were calculated with Kaplan–Meier estimates. Results: Among 553 patients who discontinued first-line afatinib, second-, third- and fourth-line therapy was administered in 394 (71%), 265 (48%), and 156 (28%) patients. The most common post-progression therapy was platinum-based chemotherapy (46%). Thirty-seven patients received subsequent osimertinib, 10 as second-line treatment. Median progression-free survival on afatinib in these 37 patients was 21.9 months. Median duration of osimertinib therapy was 20.2 months; median overall survival was not reached after a median follow-up of 4.7 years. Conclusions: Most patients treated with first-line afatinib received subsequent therapy. Although limited by sample size, enrichment, and a retrospective nature, data from patients who received sequential afatinib and osimertinib are encouraging, warranting further investigation.

Purpose: In the randomized phase IIb LUX-Lung 7 trial, afatinib significantly improved progression-free survival (PFS) and time-to-treatment failure vs gefitinib in patients with treatment-naïve epidermal growth factor receptor mutation-positive non-small cell lung cancer. We report post hoc analyses of tolerability-guided dose adjustment for afatinib and summarize the clinical characteristics of patients who continued afatinib/gefitinib beyond initial radiological progression in LUX-Lung 7. Methods: Patients received afatinib 40 mg/day or gefitinib 250 mg/day until investigator-assessed progression or beyond if beneficial. In case of selected treatment-related adverse events (TRAEs), the afatinib dose could be reduced by 10-mg decrements to minimum 20 mg (only dose interruptions were permitted with gefitinib). Results: All randomized patients were treated (afatinib, n = 160; gefitinib, n = 159). Sixty-three patients had afatinib dose reduction (< 40 mg/day; 47 within first 6 months). Dose reduction decreased TRAE incidence/severity (before vs after; all grade/grade 3: 100.0%/63.5% vs 90.5%/23.8%). There was no evidence of significant difference in PFS for patients who received < 40 mg/day vs ≥ 40 mg/day for the first 6 months [median: 12.8 vs 11.0 months; hazard ratio 1.34 (95% confidence interval 0.90–2.00)]. Twenty-four and 26 patients continued afatinib and gefitinib, respectively, beyond progression in target lesions; median time from nadir of target lesion diameters to initial progression was 6.7 months and 5.6 months. Of these patients, ~ 70% had objective response or non-complete response/non-progressive disease in non-target lesions at initial progression. Conclusions: Protocol-defined dose adjustment of afatinib may allow patients to remain on treatment longer, maximizing clinical benefit even in the presence of radiological progression.

Cancer cooperative groups have historically played a critical role in the advancement of non-small-cell lung cancer therapy. Representatives from cooperative groups worldwide convene at the International Lung Cancer Congress annually. The International Lung Cancer Congress had its 17th anniversary in the summer of 2016. The present review highlights the thoracic malignancy studies discussed by presenters. The included studies are merely a sample of the trials of thoracic malignancies ongoing globally.

Epidermal growth factor receptor (EGFR)-mutation-positive non-smallcell lung cancer (NSCLC) is incurable, despite high rates of response to EGFR tyrosine kinase inhibitors (TKIs). We investigated receptor tyrosine kinases (RTKs), Src family kinases and focal adhesion kinase (FAK) as genetic modifiers of innate resistance in EGFR-mutation-positive NSCLC. We performed gene expression analysis in two cohorts (Cohort 1 and Cohort 2) of EGFR-mutation-positive NSCLC patients treated with EGFR TKI. We evaluated the efficacy of gefitinib or osimertinib with the Src/FAK/Janus kinase 2 (JAK2) inhibitor, TPX0005 in vitro and in vivo. In Cohort 1, CUB domain-containing protein-1 (CDCP1) was an independent negative prognostic factor for progression-free survival (hazard ratio of 1.79, p=0.0407) and overall survival (hazard ratio of 2.23, p=0.0192). A two-gene model based on AXL and CDCP1 expression was strongly associated with the clinical outcome to EGFR TKIs, in both cohorts of patients. Our preclinical experiments revealed that several RTKs and non-RTKs, were up-regulated at baseline or after treatment with gefitinib or osimertinib. TPX-0005 plus EGFR TKI suppressed expression and activation of RTKs and downstream signaling intermediates. Co-expression of CDCP1 and AXL is often observed in EGFR-mutation-positive tumors, limiting the efficacy of EGFR TKIs. Co-treatment with EGFR TKI and TPX-0005 warrants testing.

The February 1, 2017, article by Spigel et al, entitled "Results From the Phase III Randomized Trial of Onartuzumab Plus Erlotinib Versus Erlotinib in Previously Treated Stage IIIB or IV Non-Small-Cell Lung Cancer: METLung" (J Clin Oncol 10.1200/JCO.2016.69.2160), was published with an error. Under Affiliations, Luis Paz-Ares' information should have included University of Seville.This has been corrected as of February 7, 2018. The authors apologize for the error.

Introduction: In ret proto-oncogene (RET)-rearranged lung cancers, data on the frequency of brain metastases and, in particular, the outcomes of multikinase inhibitor therapy in patients with intracranial disease are not well characterized. Methods: A global, multi-institutional registry (cohort A, n = 114) and a bi-institutional data set (cohort B, n = 71) of RET-rearranged lung cancer patients were analyzed. Patients were eligible if they had stage IV lung cancers harboring a RET rearrangement by local testing. The incidence of brain metastases and outcomes with multikinase inhibitor therapy were determined. Results: The frequency of brain metastases at the time of diagnosis of stage IV disease was 25% (95% confidence interval [CI]: 18%-32%) in all patients from both cohorts. The lifetime prevalence of brain metastasis in stage IV disease was 46% (95% CI: 34%-58%) in patients for whom longitudinal data was available. The cumulative incidence of brain metastases was significantly different (p = 0.0039) between RET-, ROS1-, and ALK receptor tyrosine kinase (ALK)-rearranged lung cancers, with RET intermediate between the other two groups. Although intracranial response data was not available in cohort A, the median progression-free survival of multikinase inhibitor therapy (cabozantinib, vandetanib, or sunitinib) in patients with brain metastases was 2.1 months (95% CI: 1.3-2.9 months, n = 10). In cohort B, an intracranial response was observed in 2 of 11 patients (18%) treated with cabozantinib, vandetanib (+/- everolimus), ponatinib, or alectinib; the median overall progression-free survival (intracranial and extracranial) was 3.9 months (95% CI: 2.0-4.9 months). Conclusions: Brain metastases occur frequently in RET-rearranged lung cancers, and outcomes with multikinase inhibitor therapy in general are suboptimal. Novel RET-directed targeted therapy strategies are needed. (C) 2018 International Association for the Study of Lung Cancer. Published by Elsevier Inc. Al

Introduction In LUX-Lung 3 and LUX-Lung 6, afatinib significantly improved progression-free survival (PFS) versus chemotherapy in patients with tumors harboring common epidermal growth factor receptor (EGFR) mutations (Del19/L858R) and significantly improved overall survival (OS) in patients with tumors harboring Del19 mutations. Patient-reported outcomes stratified by EGFR mutation type are reported. Patients and Methods Lung cancer symptoms and health-related quality of life (QoL) were assessed every 21 days until progression using the EORTC Quality of Life Core Questionnaire C30 and its lung cancer-specific module, LC13. Analyses of cough, dyspnea, and pain were prespecified and included analysis of percentage of patients who improved on therapy, time to deterioration of symptoms, and change over time. Global health status (GHS)/QoL was also assessed. Analyses were conducted for all patients with tumors harboring Del19 or L858R mutations and were exploratory. Results Compared with chemotherapy, afatinib more commonly improved symptoms of, delayed time to deterioration for, and was associated with better mean scores over time for cough and dyspnea in patients with Del19 or L858R mutations. All three prespecified analyses of pain showed a trend favoring afatinib over chemotherapy. In both Del19 and L858R mutations, afatinib was also associated with improvements in GHS/QoL. Longitudinal analyses demonstrated statistically significant improvements in GHS/QoL for afatinib over chemotherapy for patients with tumors harboring Del19 mutations or L858R mutations. Conclusions These exploratory analyses suggest first-line afatinib improved lung cancer-related symptoms and GHS/QoL compared with chemotherapy in patients with non-small-cell lung cancer with tumors harboring common EGFR mutations, with benefits in both Del19 and L858R patients. When considered with OS (Del19 patients only) and PFS benefits, these findings substantiate the value of using afatinib over chemothera

Cancer cooperative groups have historically played a critical role in the advancement of non-small-cell lung cancer therapy. Representatives from cooperative groups worldwide convene at the International Lung Cancer Congress annually. The International Lung Cancer Congress had its 17th anniversary in the summer of 2016. The present review highlights the thoracic malignancy studies discussed by presenters. The included studies are merely a sample of the trials of thoracic malignancies ongoing globally.

Epidermal growth factor receptor (EGFR)-mutation-positive non-smallcell lung cancer (NSCLC) is incurable, despite high rates of response to EGFR tyrosine kinase inhibitors (TKIs). We investigated receptor tyrosine kinases (RTKs), Src family kinases and focal adhesion kinase (FAK) as genetic modifiers of innate resistance in EGFR-mutation-positive NSCLC. We performed gene expression analysis in two cohorts (Cohort 1 and Cohort 2) of EGFR-mutation-positive NSCLC patients treated with EGFR TKI. We evaluated the efficacy of gefitinib or osimertinib with the Src/FAK/Janus kinase 2 (JAK2) inhibitor, TPX0005 in vitro and in vivo. In Cohort 1, CUB domain-containing protein-1 (CDCP1) was an independent negative prognostic factor for progression-free survival (hazard ratio of 1.79, p=0.0407) and overall survival (hazard ratio of 2.23, p=0.0192). A two-gene model based on AXL and CDCP1 expression was strongly associated with the clinical outcome to EGFR TKIs, in both cohorts of patients. Our preclinical experiments revealed that several RTKs and non-RTKs, were up-regulated at baseline or after treatment with gefitinib or osimertinib. TPX-0005 plus EGFR TKI suppressed expression and activation of RTKs and downstream signaling intermediates. Co-expression of CDCP1 and AXL is often observed in EGFR-mutation-positive tumors, limiting the efficacy of EGFR TKIs. Co-treatment with EGFR TKI and TPX-0005 warrants testing.

Introduction: In ret proto-oncogene (RET)-rearranged lung cancers, data on the frequency of brain metastases and, in particular, the outcomes of multikinase inhibitor therapy in patients with intracranial disease are not well characterized. Methods: A global, multi-institutional registry (cohort A, n = 114) and a bi-institutional data set (cohort B, n = 71) of RET-rearranged lung cancer patients were analyzed. Patients were eligible if they had stage IV lung cancers harboring a RET rearrangement by local testing. The incidence of brain metastases and outcomes with multikinase inhibitor therapy were determined. Results: The frequency of brain metastases at the time of diagnosis of stage IV disease was 25% (95% confidence interval [CI]: 18%-32%) in all patients from both cohorts. The lifetime prevalence of brain metastasis in stage IV disease was 46% (95% CI: 34%-58%) in patients for whom longitudinal data was available. The cumulative incidence of brain metastases was significantly different (p = 0.0039) between RET-, ROS1-, and ALK receptor tyrosine kinase (ALK)-rearranged lung cancers, with RET intermediate between the other two groups. Although intracranial response data was not available in cohort A, the median progression-free survival of multikinase inhibitor therapy (cabozantinib, vandetanib, or sunitinib) in patients with brain metastases was 2.1 months (95% CI: 1.3-2.9 months, n = 10). In cohort B, an intracranial response was observed in 2 of 11 patients (18%) treated with cabozantinib, vandetanib (+/- everolimus), ponatinib, or alectinib; the median overall progression-free survival (intracranial and extracranial) was 3.9 months (95% CI: 2.0-4.9 months). Conclusions: Brain metastases occur frequently in RET-rearranged lung cancers, and outcomes with multikinase inhibitor therapy in general are suboptimal. Novel RET-directed targeted therapy strategies are needed. (C) 2018 International Association for the Study of Lung Cancer. Published by Elsevier Inc. Al

There is a common belief that the successful training of deep neural networks requires many annotated training samples, which are often expensive and difficult to obtain especially in the biomedical imaging field. While it is often easy for researchers to use data augmentation to expand the size of training sets, constructing and generating generic augmented data that is able to teach the network the desired invariance and robustness properties using traditional data augmentation techniques is challenging in practice. In this paper, we propose a novel automatic data augmentation method that uses generative adversarial networks to learn augmentations that enable machine learning based method to learn the available annotated samples more efficiently. The architecture consists of a coarse-to-fine generator to capture the manifold of the training sets and generate generic augmented data. In our experiments, we show the efficacy of our approach on a Magnetic Resonance Imaging (MRI) image, achieving improvements of 3.5% Dice coefficient on the BRATS15 Challenge dataset as compared to traditional augmentation approaches. Also, our proposed method successfully boosts a common segmentation network to reach the state-of-the-art performance on the BRATS15 Challenge.

Purpose In addition to prospective trials for non-small-cell lung cancers (NSCLCs) that are driven by less common genomic alterations, registries provide complementary information on patient response to targeted therapies. Here, we present the results of an international registry of patients with RET-rearranged NSCLCs, providing the largest data set, to our knowledge, on outcomes of RET-directed therapy thus far. Methods A global, multicenter network of thoracic oncologists identified patients with pathologically confirmed NSCLC that harbored a RET rearrangement. Molecular profiling was performed locally by reverse transcriptase polymerase chain reaction, fluorescence in situ hybridization, or next-generation sequencing. Anonymized data-clinical, pathologic, and molecular features-were collected centrally and analyzed by an independent statistician. Best response to RET tyrosine kinase inhibition administered outside of a clinical trial was determined by RECIST v1.1. Results By April 2016, 165 patients with RET-rearranged NSCLC from 29 centers across Europe, Asia, and the United States were accrued. Median age was 61 years (range, 29 to 89 years). The majority of patients were never smokers (63%) with lung adenocarcinomas (98%) and advanced disease (91%). The most frequent rearrangement was KIF5B-RET (72%). Of those patients, 53 received one or more RET tyrosine kinase inhibitors in sequence: cabozantinib (21 patients), vandetanib (11 patients), sunitinib (10 patients), sorafenib (two patients), alectinib (two patients), lenvatinib (two patients), nintedanib (two patients), ponatinib (two patients), and regorafenib (one patient). The rate of any complete or partial response to cabozantinib, vandetanib, and sunitinib was 37%, 18%, and 22%, respectively. Further responses were observed with lenvantinib and nintedanib. Median progression-free survival was 2.3 months (95% CI, 1.6 to 5.0 months), and median overall survival was 6.8 months (95% CI, 3.9 to 14.3 months). Co

BACKGROUND Osimertinib is an epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKI) that is selective for both EGFR-TKI sensitizing and T790M resistance mutations in patients with non–small-cell lung cancer. The efficacy of osimertinib as compared with platinum-based therapy plus pemetrexed in such patients is unknown. METHODS In this randomized, international, open-label, phase 3 trial, we assigned 419 pa - tients with T790M-positive advanced non–small-cell lung cancer, who had disease progression after first-line EGFR-TKI therapy, in a 2:1 ratio to receive either oral osimertinib (at a dose of 80 mg once daily) or intravenous pemetrexed (500 mg per square meter of body-surface area) plus either carboplatin (target area under the curve, 5 [AUC5]) or cisplatin (75 mg per square meter) every 3 weeks for up to six cycles; maintenance pemetrexed was allowed. In all the patients, disease had progressed during receipt of first-line EGFR-TKI therapy. The primary end point was investigator-assessed progression-free survival. RESULTS The median duration of progression-free survival was significantly longer with osimer - tinib than with platinum therapy plus pemetrexed (10.1 months vs. 4.4 months; haz - ard ratio; 0.30; 95% confidence interval [CI], 0.23 to 0.41; P<0.001). The objective response rate was significantly better with osimertinib (71%; 95% CI, 65 to 76) than with platinum therapy plus pemetrexed (31%; 95% CI, 24 to 40) (odds ratio for objec - tive response, 5.39; 95% CI, 3.47 to 8.48; P<0.001). Among 144 patients with metas - tases to the central nervous system (CNS), the median duration of progression-free survival was longer among patients receiving osimertinib than among those receiv - ing platinum therapy plus pemetrexed (8.5 months vs. 4.2 months; hazard ratio, 0.32; 95% CI, 0.21 to 0.49). The proportion of patients with adverse events of grade 3 or higher was lower with osimertinib (23%) than with platinum therapy plus pe

Purpose The phase III OAM4971g study (METLung) examined the efficacy and safety of onartuzumab plus erlotinib in patients with locally advanced or metastatic non–small-cell lung cancer selected by MET immunohistochemistry whose disease had progressed after treatment with a platinum-based chemotherapy regimen. Patients and Methods Patients were randomly assigned at a one-to-one ratio to receive onartuzumab (15 mg/kg intravenously on day 1 of each 21-day cycle) plus daily oral erlotinib 150 mg or intravenous placebo plus daily oral erlotinib 150 mg. The primary end point was overall survival (OS) in the intent-to-treat population. Secondary end points included median progression-free survival, overall response rate, biomarker analysis, and safety. Results A total of 499 patients were enrolled (onartuzumab, n = 250; placebo, n = 249). Median OS was 6.8 versus 9.1 months for onartuzumab versus placebo (stratified hazard ratio [HR], 1.27; 95% CI, 0.98 to 1.65; P = .067), with a greater number of deaths in the onartuzumab arm (130 [52%] v 114 [46%]). Median progression-free survival was 2.7 versus 2.6 months (stratified HR, 0.99; 95% CI, 0.81 to 1.20; P = .92), and overall response rate was 8.4% and 9.6% for onartuzumab versus placebo, respectively. Exploratory analyses using MET fluorescence in situ hybridization status and gene expression showed no benefit for onartuzumab; patients with EGFR mutations showed a trend toward shorter OS with onartuzumab treatment (HR, 4.68; 95% CI, 0.97 to 22.63). Grade 3 to 5 adverse events were reported by 56.0% and 51.2% of patients, with serious AEs in 33.9% and 30.7%, for experimental versus control arms, respectively. Conclusion Onartuzumab plus erlotinib did not improve clinical outcomes, with shorter OS in the onartuzumab arm, compared with erlotinib in patients with MET-positive non–small-cell lung cancer.

Background - In the Phase III LUX-Lung 3/6 (LL3/LL6) trials in epidermal growth factor receptor (EGFR) mutation-positive lung adenocarcinoma patients, we evaluated feasibility of EGFR mutation detection using circulating cell-free DNA (cfDNA) and prognostic and predictive utility of cfDNA positivity (cfDNA+). Methods - Paired tumour and blood samples were prospectively collected from randomised patients. Mutations were detected using cfDNA from serum (LL3) or plasma (LL6) by a validated allele-specific quantitative real-time PCR kit. Results - EGFR mutation detection rates in cfDNA were 28.6% (serum) and 60.5% (plasma). Mutation detection in blood was associated with advanced disease characteristics, including higher performance score, number of metastatic sites and bone/liver metastases, and poorer prognosis. In patients with common EGFR mutations, afatinib improved progression-free survival vs chemotherapy in cfDNA+ (LL3: HR, 0.35; P=0.0009; LL6: HR, 0.25; P<0.0001) and cfDNA− (LL3: HR, 0.46; P<0.0001; LL6: HR, 0.12; P<0.0001) cohorts. A trend towards overall survival benefit with afatinib was observed in cfDNA+ patients. Conclusions - Plasma cfDNA is a promising alternative to biopsy for EGFR testing. Detectable mutation in blood was associated with more advanced disease and poorer prognosis. Afatinib improved outcomes in EGFR mutation-positive patients regardless of blood mutation status.

Purpose In addition to prospective trials for non-small-cell lung cancers (NSCLCs) that are driven by less common genomic alterations, registries provide complementary information on patient response to targeted therapies. Here, we present the results of an international registry of patients with RET-rearranged NSCLCs, providing the largest data set, to our knowledge, on outcomes of RET-directed therapy thus far. Methods A global, multicenter network of thoracic oncologists identified patients with pathologically confirmed NSCLC that harbored a RET rearrangement. Molecular profiling was performed locally by reverse transcriptase polymerase chain reaction, fluorescence in situ hybridization, or next-generation sequencing. Anonymized data-clinical, pathologic, and molecular features-were collected centrally and analyzed by an independent statistician. Best response to RET tyrosine kinase inhibition administered outside of a clinical trial was determined by RECIST v1.1. Results By April 2016, 165 patients with RET-rearranged NSCLC from 29 centers across Europe, Asia, and the United States were accrued. Median age was 61 years (range, 29 to 89 years). The majority of patients were never smokers (63%) with lung adenocarcinomas (98%) and advanced disease (91%). The most frequent rearrangement was KIF5B-RET (72%). Of those patients, 53 received one or more RET tyrosine kinase inhibitors in sequence: cabozantinib (21 patients), vandetanib (11 patients), sunitinib (10 patients), sorafenib (two patients), alectinib (two patients), lenvatinib (two patients), nintedanib (two patients), ponatinib (two patients), and regorafenib (one patient). The rate of any complete or partial response to cabozantinib, vandetanib, and sunitinib was 37%, 18%, and 22%, respectively. Further responses were observed with lenvantinib and nintedanib. Median progression-free survival was 2.3 months (95% CI, 1.6 to 5.0 months), and median overall survival was 6.8 months (95% CI, 3.9 to 14.3 months). Co

To complement the existing treatment guidelines, ESMO organises consensus conferences to focus on specific issues. The 2nd ESMO Consensus Conference on Lung Cancer included 35 experts who met to address several questions on non-small-cell lung cancer (NSCLC). Recommendations were made with reference to grade of recommendation and level of evidence. This paper focuses on locally advanced disease

This manuscript provides recommendations for the diagnosis, treatment and follow-up of early stage non-small cell lung cancer developed during the 2nd ESMO Consensus Conference on Lung Cancer in May 2013