Mediastinal lymph node detection and station mapping on chest CT using spatial priors and random forest.

PURPOSE: To develop an automated system for mediastinal lymph node detection and station mapping for chest CT. METHODS: The contextual organs, trachea, lungs, and spine are first automatically identified to locate the region of interest (ROI) (mediastinum). The authors employ shape features derived from Hessian analysis, local object scale, and circular transformation that are computed per voxel in the ROI. Eight more anatomical structures are simultaneously segmented by multiatlas label fusion. Spatial priors are defined as the relative multidimensional distance vectors corresponding to each structure. Intensity, shape, and spatial prior features are integrated and parsed by a random forest classifier for lymph node detection. The detected candidates are then segmented by the following curve evolution process. Texture features are computed on the segmented lymph nodes and a support vector machine committee is used for final classification. For lymph node station labeling, based on the segmentation results of the above anatomical structures, the textual definitions of mediastinal lymph node map according to the International Association for the Study of Lung Cancer are converted into patient-specific color-coded CT image, where the lymph node station can be automatically assigned for each detected node. RESULTS: The chest CT volumes from 70 patients with 316 enlarged mediastinal lymph nodes are used for validation. For lymph node detection, their system achieves 88% sensitivity at eight false positives per patient. For lymph node station labeling, 84.5% of lymph nodes are correctly assigned to their stations. CONCLUSIONS: Multiple-channel shape, intensity, and spatial prior features aggregated by a random forest classifier improve mediastinal lymph node detection on chest CT. Using the location information of segmented anatomic structures from the multiatlas formulation enables accurate identification of lymph node stations.

A new 2.5D representation for lymph node detection using random sets of deep convolutional neural network observations.

Automated Lymph Node (LN) detection is an important clinical diagnostic task but very challenging due to the low contrast of surrounding structures in Computed Tomography (CT) and to their varying sizes, poses, shapes and sparsely distributed locations. State-of-the-art studies show the performance range of 52.9% sensitivity at 3.1 false-positives per volume (FP/vol.), or 60.9% at 6.1 FP/vol. for mediastinal LN, by one-shot boosting on 3D HAAR features. In this paper, we first operate a preliminary candidate generation stage, towards -100% sensitivity at the cost of high FP levels (-40 per patient), to harvest volumes of interest (VOI). Our 2.5D approach consequently decomposes any 3D VOI by resampling 2D reformatted orthogonal views N times, via scale, random translations, and rotations with respect to the VOI centroid coordinates. These random views are then used to train a deep Convolutional Neural Network (CNN) classifier. In testing, the CNN is employed to assign LN probabilities for all N random views that can be simply averaged (as a set) to compute the final classification probability per VOI. We validate the approach on two datasets: 90 CT volumes with 388 mediastinal LNs and 86 patients with 595 abdominal LNs. We achieve sensitivities of 70%/83% at 3 FP/vol. and 84%/90% at 6 FP/vol. in mediastinum and abdomen respectively, which drastically improves over the previous state-of-the-art work.

2D view aggregation for lymph node detection using a shallow hierarchy of linear classifiers.

Enlarged lymph nodes (LNs) can provide important information for cancer diagnosis, staging, and measuring treatment reactions, making automated detection a highly sought goal. In this paper, we propose a new algorithm representation of decomposing the LN detection problem into a set of 2D object detection subtasks on sampled CT slices, largely alleviating the curse of dimensionality issue. Our 2D detection can be effectively formulated as linear classification on a single image feature type of Histogram of Oriented Gradients (HOG), covering a moderate field-of-view of 45 by 45 voxels. We exploit both max-pooling and sparse linear fusion schemes to aggregate these 2D detection scores for the final 3D LN detection. In this manner, detection is more tractable and does not need to perform perfectly at instance level (as weak hypotheses) since our aggregation process will robustly harness collective information for LN detection. Two datasets (90 patients with 389 mediastinal LNs and 86 patients with 595 abdominal LNs) are used for validation. Cross-validation demonstrates 78.0% sensitivity at 6 false positives/volume (FP/vol.) (86.1% at 10 FP/vol.) and 73.1% sensitivity at 6 FP/vol. (87.2% at 10 FP/vol.), for the mediastinal and abdominal datasets respectively. Our results compare favorably to previous state-of-the-art methods.

An advanced communication skills course for fourth-year, post-clerkship students.

OBJECTIVE: A novel five-module advanced communication skills course entitled "Doctor-Patient Relationships" was planned and implemented in 2000-01 at the University of British Columbia (UBC). The course was part of the final four-month component of the new MD undergraduate program: Effective Skills for Medical Practice. The goals of the communication skills course were to (1) address problems experienced by the students so far; (2) address deficiencies in achieving the UBC exit competencies; (3) help the students pass the Medical Council of Canada examinations, in particular objectives related to the Considerations of the Legal, Ethical, and Organizational aspects of the practice of medicine (CLEO); and (4) help students prepare for their roles beyond undergraduate medicine (residency, independent practice). DESCRIPTION: The course was developed by an interdisciplinary team (family practice, pathology, pediatrics, psychiatry, surgery) with input from students. The broad strengths and weaknesses of their communication skills training were identified by seven third-year medical students who kept logs over the course of their clinical clerkships to document their learning of communication skills. Analysis of these logs plus feedback meetings with the students revealed attitudinal and skills issues that needed to be addressed in the new course. The goals and principles of the course were in part agreed upon by focus groups with students, attended by faculty observers, to ensure their relevance to students. The first module "Beyond the Mask: Surviving and Thriving in Residency Training" is designed to focus students' attention on the personal relevance of developing excellence in communication skills in preparation for residency training. It includes a video of residents talking about their experiences of communication problems to trigger reflection and discussion. In the remaining four modules the students are required to put communication skills together with their medical knowledge. Each module includes pre-readings, video demonstrations (in sessions 4 and 5), practice with standardized patients (total of 14 scenarios) and structured feedback from SPs, students, and tutor. The themes of the sessions are "Dealing with Emotionally Challenging Patient Situations (informing about bad news), "Compliance and Patient Information," "Informed Consent and Shared Decision Making," and "Difficult Physician-Patient Encounters." Each module lasts two hours. The course was implemented for 120 students, facilitated by 14 tutors (seven to eight students per group). DISCUSSION: Student involvement in many different ways provided an important reality check and made us think about how to present the new course so that it was relevant and interesting to students. Attention to student input was a major contributor to the good evaluations given the course. Students rated the course highly: the relevance of the weekly themes was rated 4.21 on a five-point scale; the effectiveness of the SP interviews, 4.10; the effectiveness of the group discussion and feedback, 4.18; and overall course effectiveness in enhancing communication skills, 3.91. The tutors also rated the course highly, and the students rated the tutors highly. Minor changes will be made to the course next year based on the specific suggestions for improvement, which were identified.