Keyvan Farahani

The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)

In this paper we report the set-up and results of the Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) organized in conjunction with the MICCAI 2012 and 2013 conferences. Twenty state-of-the-art tumor segmentation algorithms were applied to a set of 65 multi-contrast MR scans of low- and high-grade glioma patients - manually annotated by up to four raters - and to 65 comparable scans generated using tumor image simulation software. Quantitative evaluations revealed considerable disagreement between the human raters in segmenting various tumor sub-regions (Dice scores in the range 74%-85%), illustrating the difficulty of this task. We found that different algorithms worked best for different sub-regions (reaching performance comparable to human inter-rater variability), but that no single algorithm ranked in the top for all sub-regions simultaneously. Fusing several good algorithms using a hierarchical majority vote yielded segmentations that consistently ranked above all individual algorithms, indicating remaining opportunities for further methodological improvements. The BRATS image data and manual annotations continue to be publicly available through an online evaluation system as an ongoing benchmarking resource.

Challenges in image-guided therapy system design

System development for image-guided therapy (IGT), or image-guided interventions (IGI), continues to be an area of active interest across academic and industry groups. This is an emerging field that is growing rapidly: major academic institutions and medical device manufacturers have produced IGT technologies that are in routine clinical use, dozens of high-impact publications are published in well regarded journals each year, and several small companies have successfully commercialized sophisticated IGT systems. In meetings between IGT investigators over the last two years, a consensus has emerged that several key areas must be addressed collaboratively by the community to reach the next level of impact and efficiency in IGT research and development to improve patient care. These meetings culminated in a two-day workshop that brought together several academic and industrial leaders in the field today. The goals of the workshop were to identify gaps in the engineering infrastructure available to IGT researchers, develop the role of research funding agencies and the recently established US-based National Center for Image Guided Therapy (NCIGT), and ultimately to facilitate the transfer of technology among research centers that are sponsored by the National Institutes of Health (NIH). Workshop discussions spanned many of the current challenges in the development and deployment of new IGT systems. Key challenges were identified in a number of areas, including: validation standards; workflows, use-cases, and application requirements; component reusability; and device interface standards. This report elaborates on these key points and proposes research challenges that are to be addressed by a joint effort between academic, industry, and NIH participants.

Assessing the barriers to image-guided drug delivery

Imaging has become a cornerstone for medical diagnosis and the guidance of patient management. A new field called image-guided drug delivery (IGDD) now combines the vast potential of the radiological sciences with the delivery of treatment and promises to fulfill the vision of personalized medicine. Whether imaging is used to deliver focused energy to drug-laden particles for enhanced, local drug release around tumors, or it is invoked in the context of nanoparticle-based agents to quantify distinctive biomarkers that could risk stratify patients for improved targeted drug delivery efficiency, the overarching goal of IGDD is to use imaging to maximize effective therapy in diseased tissues and to minimize systemic drug exposure in order to reduce toxicities. Over the last several years, innumerable reports and reviews covering the gamut of IGDD technologies have been published, but inadequate attention has been directed toward identifying and addressing the barriers limiting clinical translation. In this consensus opinion, the opportunities and challenges impacting the clinical realization of IGDD-based personalized medicine were discussed as a panel and recommendations were proffered to accelerate the field forward.

Advancing The Cancer Genome Atlas glioma MRI collections with expert segmentation labels and radiomic features

Gliomas belong to a group of central nervous system tumors, and consist of various sub-regions. Gold standard labeling of these sub-regions in radiographic imaging is essential for both clinical and computational studies, including radiomic and radiogenomic analyses. Towards this end, we release segmentation labels and radiomic features for all pre-operative multimodal magnetic resonance imaging (MRI) (n=243) of the multi-institutional glioma collections of The Cancer Genome Atlas (TCGA), publicly available in The Cancer Imaging Archive (TCIA). Pre-operative scans were identified in both glioblastoma (TCGA-GBM, n=135) and low-grade-glioma (TCGA-LGG, n=108) collections via radiological assessment. The glioma sub-region labels were produced by an automated state-of-the-art method and manually revised by an expert board-certified neuroradiologist. An extensive panel of radiomic features was extracted based on the manually-revised labels. This set of labels and features should enable i) direct utilization of the TCGA/TCIA glioma collections towards repeatable, reproducible and comparative quantitative studies leading to new predictive, prognostic, and diagnostic assessments, as well as ii) performance evaluation of computer-aided segmentation methods, and comparison to our state-of-the-art method.

LUNGx Challenge for computerized lung nodule classification: reflections and lessons learned.

Challenges, in the context of medical imaging, are valuable in that they allow for a direct comparison of different algorithms designed for a specific radiologic task, with all algorithms abiding by the same set of rules, operating on a common set of images, and being evaluated with a uniform performance assessment paradigm. The variability of system performance based on database composition and subtlety, definition of “truth,” and scoring metric is well-known;1–3 challenges serve to level the differences across these various dimensions. The medical imaging community has hosted a number of successful thoracic imaging challenges that have spanned a wide range of tasks,4,5 including lung nodule detection,6 lung nodule change, vessel segmentation,7 and vessel tree extraction.8 Each challenge presents its own unique set of circumstances and considerations; however, important common themes exist. Future challenge organizers (and participants) could benefit from an open discussion of successes achieved, pitfalls encountered, and lessons learned from each completed challenge.

LUNGx Challenge for computerized lung nodule classification

Abstract. The purpose of this work is to describe the LUNGx Challenge for the computerized classification of lung nodules on diagnostic computed tomography (CT) scans as benign or malignant and report the performance of participants’ computerized methods along with that of six radiologists who participated in an observer study performing the same Challenge task on the same dataset. The Challenge provided sets of calibration and testing scans, established a performance assessment process, and created an infrastructure for case dissemination and result submission. Ten groups applied their own methods to 73 lung nodules (37 benign and 36 malignant) that were selected to achieve approximate size matching between the two cohorts. Area under the receiver operating characteristic curve (AUC) values for these methods ranged from 0.50 to 0.68; only three methods performed statistically better than random guessing. The radiologists’ AUC values ranged from 0.70 to 0.85; three radiologists performed statistically better than the best-performing computer method. The LUNGx Challenge compared the performance of computerized methods in the task of differentiating benign from malignant lung nodules on CT scans, placed in the context of the performance of radiologists on the same task. The continued public availability of the Challenge cases will provide a valuable resource for the medical imaging research community.

Development of a Research Agenda for Interventional Oncology: Proceedings from an Interdisciplinary Consensus Panel

INTERVENTIONAL specialties including interventional radiology have been responsible for the innovation, validation, and dissemination of numerous procedures and technologies that have impacted the practice of medicine. The universal endorsement of such established techniques as coronary and peripheral angioplasty (1,2), transjugular intrahepatic portosystemic shunts (3), and uterine fibroid embolization (4,5) has been predicated on the translational research skills of numerous scientists investigating minimally invasive techniques, allowing the advancement of initial conceptual hypotheses toward research trials and subsequently into clinical practice. Within the past decade, there has been evolutionary research and clinical interest in the ability to combine minimally invasive interventional radiologic techniques with multiplanar and functional imaging as a tool to guide organ-specific cancer treatments (6–10). These techniques, collectively referred to as image-guided intervention (IGI) or interventional oncology (IO), have the theoretic appeal of allowing enhanced target-specific therapy for patients with limited or single-organ–dominant disease while potentially minimizing the expectant risks and systemic effects of alternative therapies. In addition, locoregional therapies may also play an important role by providing adjuvant or neoadjuvant therapies to palliate the effects of a malignancy, potentiate other treatments, or prevent untoward effects of systemically targeted therapies, ie, limit organ-specific toxicity with use of catheter-based methods to protect the organ at risk. Concomitant with the growth of IGI, numerous areas of potential further investigation have surfaced. Broadly speaking, several of the more prominent future research issues include (i) systematic assessment of the role of existing and emerging molecular and functional imaging modalities to guide or assess interventions; (ii) use of combinations of existing therapies; (iii) evaluation of existing, novel, or nascent biologic and pharmaceutical agents for catheter-directed regional delivery; and (iv) use of IGI for palliation or ancillary treatment of malignancy-associated processes or paraneoplastic phenomena. To further develop these topics and to identify a research agenda for IO, an interdisciplinary meeting of prominent experts in the fields of interventional radiology, cancer research, and medical oncology was convened in September 2002. The meeting was sponsored by the National Cancer Institute (NCI), National Institute of Biomedical Imaging and Bioengineering, American Cancer Society, American Association of Physicists in Medicine, Cardiovascular and Interventional Radiology Research and Education Foundation, and American College of Radiology Imaging Network. This report summarizes the results of that meeting.

Guest Editorial: LUNGx Challenge for computerized lung nodule classification: reflections and lessons learned

Challenges, in the context of medical imaging, are valuable in that they allow for a direct comparison of different algorithms designed for a specific radiologic task, with all algorithms abiding by the same set of rules, operating on a common set of images, and being evaluated with a uniform performance assessment paradigm. The variability of system performance based on database composition and subtlety, definition of “truth,” and scoring metric is well-known;1–3 challenges serve to level the differences across these various dimensions. The medical imaging community has hosted a number of successful thoracic imaging challenges that have spanned a wide range of tasks,4,5 including lung nodule detection,6 lung nodule change, vessel segmentation,7 and vessel tree extraction.8 Each challenge presents its own unique set of circumstances and considerations; however, important common themes exist. Future challenge organizers (and participants) could benefit from an open discussion of successes achieved, pitfalls encountered, and lessons learned from each completed challenge.

Safety of Pulsed High Intensity Focused Ultrasound for Enhanced Drug and Gene Delivery

For a limited range of exposure parameters, pulsed high intensity focused ultrasound (HIFU) has been shown to increase the delivery of certain systemically administered macromolecular diagnostic and therapeutic agents in mice. The mechanism for the enhanced delivery has not been demonstrated definitively and, in principle, can include thermal, cavitational, and non‐cavitation mechanical effects. The sonicated tissue has no damage on histology. As a step towards clinical translation, the safety of this technique needs to be assessed in a clinically relevant manner. In this study, the safety of pulsed HIFU is evaluated with near real‐time phase shift magnetic resonance (MR) thermometry and anatomic MR imaging using rabbits as subjects. MR guidance enables pulsed HIFU enhanced delivery to be implemented safely from a thermal standpoint. Although the effects of pulsed HIFU are not seen on anatomic MR images, they may be detected on MR sequences sensitive to permeability, diffusion, and elasticity. Such work t...

Characterization of Pulsed High Intensity Focused Ultrasound for Enhanced Drug and Gene Delivery

Within a certain range of parameters, pulsed high intensity focused ultrasound (HIFU) has been shown to increase the delivery of systemically administered drugs and plasmid DNA in tumors in mice. The sonicated tissue is not damaged by light microscopy. The mechanism for the enhanced delivery has not been shown conclusively and can include thermal, cavitational, and non‐cavitation mechanical effects. In order to assess the effects of pulsed HIFU in a manner that allows for clinical translation, pulsed HIFU is performed within a magnetic resonance (MR) scanner. In this work, the thermal effect is evaluated with phase‐shift MR thermometry in ex vivo chicken muscle. The thermal effect is small at the most common exposure parameters. In the future, non‐thermal effects like permeability, diffusion, and elasticity changes will be evaluated with dynamic contrast enhanced MRI, diffusion‐weighted MRI, and MR elastography. If changes in permeability, diffusion, and shear modulus are associated with pulsed HIFU enhan...