Filip Malmberg

Evaluation of prostate segmentation algorithms for MRI: the PROMISE12 challenge

Prostate MRI image segmentation has been an area of intense research due to the increased use of MRI as a modality for the clinical workup of prostate cancer. Segmentation is useful for various tasks, e.g. to accurately localize prostate boundaries for radiotherapy or to initialize multi-modal registration algorithms. In the past, it has been difficult for research groups to evaluate prostate segmentation algorithms on multi-center, multi-vendor and multi-protocol data. Especially because we are dealing with MR images, image appearance, resolution and the presence of artifacts are affected by differences in scanners and/or protocols, which in turn can have a large influence on algorithm accuracy. The Prostate MR Image Segmentation (PROMISE12) challenge was setup to allow a fair and meaningful comparison of segmentation methods on the basis of performance and robustness. In this work we will discuss the initial results of the online PROMISE12 challenge, and the results obtained in the live challenge workshop hosted by the MICCAI2012 conference. In the challenge, 100 prostate MR cases from 4 different centers were included, with differences in scanner manufacturer, field strength and protocol. A total of 11 teams from academic research groups and industry participated. Algorithms showed a wide variety in methods and implementation, including active appearance models, atlas registration and level sets. Evaluation was performed using boundary and volume based metrics which were combined into a single score relating the metrics to human expert performance. The winners of the challenge where the algorithms by teams Imorphics and ScrAutoProstate, with scores of 85.72 and 84.29 overall. Both algorithms where significantly better than all other algorithms in the challenge (p<0.05) and had an efficient implementation with a run time of 8min and 3s per case respectively. Overall, active appearance model based approaches seemed to outperform other approaches like multi-atlas registration, both on accuracy and computation time. Although average algorithm performance was good to excellent and the Imorphics algorithm outperformed the second observer on average, we showed that algorithm combination might lead to further improvement, indicating that optimal performance for prostate segmentation is not yet obtained. All results are available online at http://promise12.grand-challenge.org/.

Intracranial volume estimated with commonly used methods could introduce bias in studies including brain volume measurements

In brain volumetric studies, intracranial volume (ICV) is often used as an estimate of pre-morbid brain size as well as to compensate for inter-subject variations in head size. However, if the estimated ICV is biased by for example gender or atrophy, it could introduce errors in study results. To evaluate how two commonly used methods for ICV estimation perform, computer assisted reference segmentations were created and evaluated. Segmentations were created for 399 MRI volumes from 75-year-old subjects, with 53 of these subjects having an additional scan and segmentation created at age 80. ICV estimates from Statistical Parametric Mapping (SPM, version 8) and Freesurfer (FS, version 5.1.0) were compared to the reference segmentations, and bias related to skull size (approximated with the segmentation measure), gender or atrophy were tested for. The possible ICV related effect on associations between normalized hippocampal volume and factors gender, education and cognition was evaluated by normalizing hippocampal volume with different ICV measures. Excellent agreement was seen for inter- (r=0.999) and intra- (r=0.999) operator reference segmentations. Both SPM and FS overestimated ICV. SPM showed bias associated with gender and atrophy while FS showed bias dependent on skull size. All methods showed good correlation between time points in the longitudinal data (reference: 0.998, SPM: 0.962, FS: 0.995). Hippocampal volume showed different associations with cognition and gender depending on which ICV measure was used for hippocampal volume normalization. These results show that the choice of method used for ICV estimation can bias results in studies including brain volume measurements.

Bisphenol A exposure increases liver fat in juvenile fructose-fed Fischer 344 rats

BACKGROUND Prenatal exposure to bisphenol A (BPA) has been shown to induce obesity in rodents. To evaluate if exposure also later in life could induce obesity or liver damage we investigated these hypothesises in an experimental rat model. METHODS From five to fifteen weeks of age, female Fischer 344 rats were exposed to BPA via drinking water (0.025, 0.25 or 2.5 mg BPA/L) containing 5% fructose. Two control groups were given either water or 5% fructose solution. Individual weight of the rats was determined once a week. At termination magnetic resonance imaging was used to assess adipose tissue amount and distribution, and liver fat content. After sacrifice the left perirenal fat pad and the liver were dissected and weighed. Apolipoprotein A-I in plasma was analyzed by western blot. RESULTS No significant effects on body weight or the weight of the dissected fad pad were seen in rats exposed to BPA, and MRI showed no differences in total or visceral adipose tissue volumes between the groups. However, MRI showed that liver fat content was significantly higher in BPA-exposed rats than in fructose controls (p=0.04). BPA exposure also increased the apolipoprotein A-I levels in plasma (p<0.0001). CONCLUSION We found no evidence that BPA exposure affects fat mass in juvenile fructose-fed rats. However, the finding that BPA in combination with fructose induced fat infiltration in the liver at dosages close to the current tolerable daily intake (TDI) might be of concern given the widespread use of this compound in our environment.

A graph-based framework for sub-pixel image segmentation

Many image segmentation methods utilize graph structures for representing images, where the flexibility and generality of the abstract structure is beneficial. By using a fuzzy object representation, i.e., allowing partial belongingness of elements to image objects, the unavoidable loss of information when representing continuous structures by finite sets is significantly reduced, enabling feature estimates with sub-pixel precision. This work presents a framework for object representation based on fuzzy segmented graphs. Interpreting the edges as one-dimensional paths between the vertices of a graph, we extend the notion of a graph cut to that of a located cut, i.e., a cut with sub-edge precision. We describe a method for computing a located cut from a fuzzy segmentation of graph vertices. Further, the notion of vertex coverage segmentation is proposed as a graph theoretic equivalent to pixel coverage segmentations and a method for computing such a segmentation from a located cut is given. Utilizing the proposed framework, we demonstrate improved precision of area measurements of synthetic two-dimensional objects. We emphasize that although the experiments presented here are performed on two-dimensional images, the proposed framework is defined for general graphs and thus applicable to images of any dimension.

Efficient algorithm for finding the exact minimum barrier distance

Abstract The minimum barrier distance, MBD, introduced recently in [1] , is a pseudo-metric defined on a compact subset D of the Euclidean space R n and whose values depend on a fixed map (an image) f from D into R . The MBD is defined as the minimal value of the barrier strength of a path between the points, which constitutes the length of the smallest interval containing all values of f along the path. In this paper we present a polynomial time algorithm, that provably calculates the exact values of MBD for the digital images. We compare this new algorithm, theoretically and experimentally, with the algorithm presented in [1] , which computes the approximate values of the MBD. Moreover, we notice that every generalized distance function can be naturally translated to an image segmentation algorithm. The algorithms that fall under such category include: Relative Fuzzy Connectedness, and those associated with the minimum barrier, fuzzy distance, and geodesic distance functions. In particular, we compare experimentally these four algorithms on the 2D and 3D natural and medical images with known ground truth and at varying level of noise, blur, and inhomogeneity.

A 3d live-wire segmentation method for volume images using haptic interaction

Designing interactive segmentation methods for digital volume images is difficult, mainly because efficient 3D interaction is much harder to achieve than interaction with 2D images To overcome this issue, we use a system that combines stereo graphics and haptics to facilitate efficient 3D interaction We propose a new method, based on the 2D live-wire method, for segmenting volume images Our method consists of two parts: an interface for drawing 3D live-wire curves onto the boundary of an object in a volume image, and an algorithm for connecting two such curves to create a discrete surface.

Sub-pixel Segmentation with the Image Foresting Transform

The Image Foresting Transform (IFT) is a framework for image partitioning, commonly used for interactive segmentation. Given an image where a subset of the image elements (seed-points) have been assigned user-defined labels, the IFT completes the labeling by computing minimal cost paths from all image elements to the seed-points. Each image element is then given the same label as the closest seed-point. In its original form, the IFT produces crisp segmentations, i.e., each image element is assigned the label of exactly one seed-point. Here, we propose a modified version of the IFT that computes region boundaries with sub-pixel precision by allowing mixed labels at region boundaries. We demonstrate that the proposed sub-pixel IFT allows properties of the segmented object to be measured with higher precision.

Intracranial volume normalization methods: considerations when investigating gender differences in regional brain volume.

Intracranial volume (ICV) normalization of regional brain volumes (v) is common practice in volumetric studies of the aging brain. Multiple normalization methods exist and this study aimed to investigate when each method is appropriate to use in gender dimorphism studies and how differences in v are affected by the choice of method. A new method based on weighted ICV matching is also presented. Theoretical reasoning and simulated experiments were followed by an evaluation using real data comprising 400 subjects, all 75 years old, whose ICV was segmented with a gold standard method. The presented method allows good visualization of volume relation between gender groups. A different gender dimorphism in volume was found depending on the normalization method used for both simulated and real data. Method performance was also seen to depend on the slope (B) and intercept (m) from the linear relation between v and ICV (v=B·ICV+m) as well as gender distribution in the cohort. A suggested work-flow for selecting ICV normalization method when investigating gender related differences in regional brain volume is presented.

Shape and volume of craniofacial cavities in intentional skull deformations

Intentional cranial deformations (ICD) have been observed worldwide but are especially prevalent in preColombian cultures. The purpose of this study was to assess the consequences of ICD on three cranial cavities (intracranial cavity, orbits, and maxillary sinuses) and on cranial vault thickness, in order to screen for morphological changes due to the external constraints exerted by the deformation device. We acquired CT-scans for 39 deformed and 19 control skulls. We studied the thickness of the skull vault using qualitative and quantitative methods. We computed the volumes of the orbits, of the maxillary sinuses, and of the intracranial cavity using haptic-aided semi-automatic segmentation. We finally defined 3D distances and angles within orbits and maxillary sinuses based on 27 anatomical landmarks and measured these features on the 58 skulls. Our results show specific bone thickness patterns in some types of ICD, with localized thinning in regions subjected to increased pressure and thickening in other regions. Our findings confirm that volumes of the cranial cavities are not affected by ICDs but that the shapes of the orbits and of the maxillary sinuses are modified in circumferential deformations. We conclude that ICDs can modify the shape of the cranial cavities and the thickness of their walls but conserve their volumes. These results provide new insights into the morphological effects associated with ICDs and call for similar investigations in subjects with deformational plagiocephalies and craniosynostoses.

Visualization and haptics for interactive medical image analysis: image segmentation in cranio-maxillofacial surgery planning

A central problem in cranio-maxillofacial (CMF) surgery is to restore the normal anatomy of the skeleton after defects, e.g., trauma to the face. With careful pre-operative planning, the precision and predictability of the craniofacial reconstruction can be significantly improved. In addition, morbidity can be reduced thanks to shorter operation time. An important component in surgery planning is to be able to accurately measure the extent of anatomical structures. Of particular interest are the shape and volume of the orbits (eye sockets). These properties can be measured in 3D CT images of the skull, provided that an accurate segmentation of the orbits is available. Here, we present a system for interactive segmentation of the orbit in CT images. The system utilizes 3D visualization and haptic feedback to facilitate efficient exploration and manipulation of 3D data.