Radiomics workflow definition & challenges - German priority program 2177 consensus statement on clinically applied radiomics.

OBJECTIVES: Achieving a consensus on a definition for different aspects of radiomics workflows to support their translation into clinical usage. Furthermore, to assess the perspective of experts on important challenges for a successful clinical workflow implementation. MATERIALS AND METHODS: The consensus was achieved by a multi-stage process. Stage 1 comprised a definition screening, a retrospective analysis with semantic mapping of terms found in 22 workflow definitions, and the compilation of an initial baseline definition. Stages 2 and 3 consisted of a Delphi process with over 45 experts hailing from sites participating in the German Research Foundation (DFG) Priority Program 2177. Stage 2 aimed to achieve a broad consensus for a definition proposal, while stage 3 identified the importance of translational challenges. RESULTS: Workflow definitions from 22 publications (published 2012-2020) were analyzed. Sixty-nine definition terms were extracted, mapped, and semantic ambiguities (e.g., homonymous and synonymous terms) were identified and resolved. The consensus definition was developed via a Delphi process. The final definition comprising seven phases and 37 aspects reached a high overall consensus (> 89% of experts "agree" or "strongly agree"). Two aspects reached no strong consensus. In addition, the Delphi process identified and characterized from the participating experts' perspective the ten most important challenges in radiomics workflows. CONCLUSION: To overcome semantic inconsistencies between existing definitions and offer a well-defined, broad, referenceable terminology, a consensus workflow definition for radiomics-based setups and a terms mapping to existing literature was compiled. Moreover, the most relevant challenges towards clinical application were characterized. CRITICAL RELEVANCE STATEMENT: Lack of standardization represents one major obstacle to successful clinical translation of radiomics. Here, we report a consensus workflow definition on different aspects of radiomics studies and highlight important challenges to advance the clinical adoption of radiomics. KEY POINTS: Published radiomics workflow terminologies are inconsistent, hindering standardization and translation. A consensus radiomics workflow definition proposal with high agreement was developed. Publicly available result resources for further exploitation by the scientific community.

Cooperative AI training for cardiothoracic segmentation in computed tomography: An iterative multi-center annotation approach.

PURPOSE: Radiological reporting is transitioning to quantitative analysis, requiring large-scale multi-center validation of biomarkers. A major prerequisite and bottleneck for this task is the voxelwise annotation of image data, which is time-consuming for large cohorts. In this study, we propose an iterative training workflow to support and facilitate such segmentation tasks, specifically for high-resolution thoracic CT data. METHODS: Our study included 132 thoracic CT scans from clinical practice, annotated by 13 radiologists. In three iterative training experiments, we aimed to improve and accelerate segmentation of the heart and mediastinum. Each experiment started with manual segmentation of 5-25 CT scans, which served as training data for a nnU-Net. Further iterations incorporated AI pre-segmentation and human correction to improve accuracy, accelerate the annotation process, and reduce human involvement over time. RESULTS: Results showed consistent improvement in AI model quality with each iteration. Resampled datasets improved the Dice similarity coefficients for both the heart (DCS 0.91 [0.88; 0.92]) and the mediastinum (DCS 0.95 [0.94; 0.95]). Our AI models reduced human interaction time by 50 % for heart and 70 % for mediastinum segmentation in the most potent iteration. A model trained on only five datasets achieved satisfactory results (DCS > 0.90). CONCLUSIONS: The iterative training workflow provides an efficient method for training AI-based segmentation models in multi-center studies, improving accuracy over time and simultaneously reducing human intervention. Future work will explore the use of fewer initial datasets and additional pre-processing methods to enhance model quality.

Quantitative evaluation of the influence of multiple MRI sequences and of pathological tissues on the registration of longitudinal data acquired during brain tumor treatment.

Registration methods facilitate the comparison of multiparametric magnetic resonance images acquired at different stages of brain tumor treatments. Image-based registration solutions are influenced by the sequences chosen to compute the distance measure, and the lack of image correspondences due to the resection cavities and pathological tissues. Nonetheless, an evaluation of the impact of these input parameters on the registration of longitudinal data is still missing. This work evaluates the influence of multiple sequences, namely T1-weighted (T1), T2-weighted (T2), contrast enhanced T1-weighted (T1-CE), and T2 Fluid Attenuated Inversion Recovery (FLAIR), and the exclusion of the pathological tissues on the non-rigid registration of pre- and post-operative images. We here investigate two types of registration methods, an iterative approach and a convolutional neural network solution based on a 3D U-Net. We employ two test sets to compute the mean target registration error (mTRE) based on corresponding landmarks. In the first set, markers are positioned exclusively in the surroundings of the pathology. The methods employing T1-CE achieves the lowest mTREs, with a improvement up to 0.8 mm for the iterative solution. The results are higher than the baseline when using the FLAIR sequence. When excluding the pathology, lower mTREs are observable for most of the methods. In the second test set, corresponding landmarks are located in the entire brain volumes. Both solutions employing T1-CE obtain the lowest mTREs, with a decrease up to 1.16 mm for the iterative method, whereas the results worsen using the FLAIR. When excluding the pathology, an improvement is observable for the CNN method using T1-CE. Both approaches utilizing the T1-CE sequence obtain the best mTREs, whereas the FLAIR is the least informative to guide the registration process. Besides, the exclusion of pathology from the distance measure computation improves the registration of the brain tissues surrounding the tumor. Thus, this work provides the first numerical evaluation of the influence of these parameters on the registration of longitudinal magnetic resonance images, and it can be helpful for developing future algorithms.

Anisotropic 3D Multi-Stream CNN for Accurate Prostate Segmentation from Multi-Planar MRI.

BACKGROUND AND OBJECTIVE: Accurate and reliable segmentation of the prostate gland in MR images can support the clinical assessment of prostate cancer, as well as the planning and monitoring of focal and loco-regional therapeutic interventions. Despite the availability of multi-planar MR scans due to standardized protocols, the majority of segmentation approaches presented in the literature consider the axial scans only. In this work, we investigate whether a neural network processing anisotropic multi-planar images could work in the context of a semantic segmentation task, and if so, how this additional information would improve the segmentation quality. METHODS: We propose an anisotropic 3D multi-stream CNN architecture, which processes additional scan directions to produce a high-resolution isotropic prostate segmentation. We investigate two variants of our architecture, which work on two (dual-plane) and three (triple-plane) image orientations, respectively. The influence of additional information used by these models is evaluated by comparing them with a single-plane baseline processing only axial images. To realize a fair comparison, we employ a hyperparameter optimization strategy to select optimal configurations for the individual approaches. RESULTS: Training and evaluation on two datasets spanning multiple sites show statistical significant improvement over the plain axial segmentation (p<0.05 on the Dice similarity coefficient). The improvement can be observed especially at the base (0.898 single-plane vs. 0.906 triple-plane) and apex (0.888 single-plane vs. 0.901 dual-plane). CONCLUSION: This study indicates that models employing two or three scan directions are superior to plain axial segmentation. The knowledge of precise boundaries of the prostate is crucial for the conservation of risk structures. Thus, the proposed models have the potential to improve the outcome of prostate cancer diagnosis and therapies.

Automated density-based counting of FISH amplification signals for HER2 status assessment.

BACKGROUND: Automated image analysis can make quantification of FISH signals in histological sections more efficient and reproducible. Current detection-based methods, however, often fail to accurately quantify densely clustered FISH signals. METHODS: We propose a novel density-based approach to quantifying FISH signals. Instead of detecting individual signals, this approach quantifies FISH signals in terms of the integral over a density map predicted by Deep Learning. We apply the density-based approach to the task of counting and determining ratios of ERBB2 and CEN17 signals and compare it to common detection-based and area-based approaches. RESULTS: The ratios determined by our approach were strongly correlated with results obtained by manual annotation of individual FISH signals (Pearson's r = 0.907). In addition, they were highly consistent with cutoff-scores determined by a pathologist (balanced concordance = 0.971). The density-based approach generally outperformed the other approaches. Its superiority was particularly evident in the presence of dense signal clusters. CONCLUSIONS: The presented approach enables accurate and efficient automated quantification of FISH signals. Since signals in clusters can hardly be detected individually even by human observers, the density-based quantification performs better than detection-based approaches.

Evaluation of deep learning methods for parotid gland segmentation from CT images.

The segmentation of organs at risk is a crucial and time-consuming step in radiotherapy planning. Good automatic methods can significantly reduce the time clinicians have to spend on this task. Due to its variability in shape and low contrast to surrounding structures, segmenting the parotid gland is challenging. Motivated by the recent success of deep learning, we study the use of two-dimensional (2-D), 2-D ensemble, and three-dimensional (3-D) U-Nets for segmentation. The mean Dice similarity to ground truth is ∼ 0.83 for all three models. A patch-based approach for class balancing seems promising for false-positive reduction. The 2-D ensemble and 3-D U-Net are applied to the test data of the 2015 MICCAI challenge on head and neck autosegmentation. Both deep learning methods generalize well onto independent data (Dice 0.865 and 0.88) and are superior to a selection of model- and atlas-based methods with respect to the Dice coefficient. Since appropriate reference annotations are essential for training but often difficult and expensive to obtain, it is important to know how many samples are needed for training. We evaluate the performance after training with different-sized training sets and observe no significant increase in the Dice coefficient for more than 250 training cases.

Deep learning nuclei detection: A simple approach can deliver state-of-the-art results.

BACKGROUND: Deep convolutional neural networks have become a widespread tool for the detection of nuclei in histopathology images. Many implementations share a basic approach that includes generation of an intermediate map indicating the presence of a nucleus center, which we refer to as PMap. Nevertheless, these implementations often still differ in several parameters, resulting in different detection qualities. METHODS: We identified several essential parameters and configured the basic PMap approach using combinations of them. We thoroughly evaluated and compared various configurations on multiple datasets with respect to detection quality, efficiency and training effort. RESULTS: Post-processing of the PMap was found to have the largest impact on detection quality. Also, two different network architectures were identified that improve either detection quality or runtime performance. The best-performing configuration yields f1-measures of 0.816 on H&E stained images of colorectal adenocarcinomas and 0.819 on Ki-67 stained images of breast tumor tissue. On average, it was fully trained in less than 15,000 iterations and processed 4.15 megapixels per second at prediction time. CONCLUSIONS: The basic PMap approach is greatly affected by certain parameters. Our evaluation provides guidance on their impact and best settings. When configured properly, this simple and efficient approach can yield equal detection quality as more complex and time-consuming state-of-the-art approaches.

Auditory display for fluorescence-guided open brain tumor surgery.

PURPOSE: Protoporphyrin (PpIX) fluorescence allows discrimination of tumor and normal brain tissue during neurosurgery. A handheld fluorescence (HHF) probe can be used for spectroscopic measurement of 5-ALA-induced PpIX to enable objective detection compared to visual evaluation of fluorescence. However, current technology requires that the surgeon either views the measured values on a screen or employs an assistant to verbally relay the values. An auditory feedback system was developed and evaluated for communicating measured fluorescence intensity values directly to the surgeon. METHODS: The auditory display was programmed to map the values measured by the HHF probe to the playback of tones that represented three fluorescence intensity ranges and one error signal. Ten persons with no previous knowledge of the application took part in a laboratory evaluation. After a brief training period, participants performed measurements on a tray of 96 wells of liquid fluorescence phantom and verbally stated the perceived measurement values for each well. The latency and accuracy of the participants' verbal responses were recorded. The long-term memorization of sound function was evaluated in a second set of 10 participants 2-3 and 7-12 days after training. RESULTS: The participants identified the played tone accurately for 98% of measurements after training. The median response time to verbally identify the played tones was 2 pulses. No correlation was found between the latency and accuracy of the responses, and no significant correlation with the musical proficiency of the participants was observed on the function responses. Responses for the memory test were 100% accurate. CONCLUSION: The employed auditory display was shown to be intuitive, easy to learn and remember, fast to recognize, and accurate in providing users with measurements of fluorescence intensity or error signal. The results of this work establish a basis for implementing and further evaluating auditory displays in clinical scenarios involving fluorescence guidance and other areas for which categorized auditory display could be useful.

Data-Driven Discovery of Immune Contexture Biomarkers.

Background: Features characterizing the immune contexture (IC) in the tumor microenvironment can be prognostic and predictive biomarkers. Identifying novel biomarkers can be challenging due to complex interactions between immune and tumor cells and the abundance of possible features. Methods: We describe an approach for the data-driven identification of IC biomarkers. For this purpose, we provide mathematical definitions of different feature classes, based on cell densities, cell-to-cell distances, and spatial heterogeneity thereof. Candidate biomarkers are ranked according to their potential for the predictive stratification of patients. Results: We evaluated the approach on a dataset of colorectal cancer patients with variable amounts of microsatellite instability. The most promising features that can be explored as biomarkers were based on cell-to-cell distances and spatial heterogeneity. Both the tumor and non-tumor compartments yielded features that were potentially predictive for therapy response and point in direction of further exploration. Conclusion: The data-driven approach simplifies the identification of promising IC biomarker candidates. Researchers can take guidance from the described approach to accelerate their biomarker research.

Different patterns of longitudinal brain and spinal cord changes and their associations with disability progression in NMO and MS.

OBJECTIVE: To investigate the longitudinal spinal cord and brain changes in neuromyelitis optica (NMO) and multiple sclerosis (MS) and their associations with disability progression. PATIENTS AND METHODS: We recruited 28 NMO, 22 MS, and 20 healthy controls (HC), who underwent both spinal cord and brain MRI at baseline. Twenty-five NMO and 20 MS completed 1-year follow-up. Baseline spinal cord and brain lesion loads, mean upper cervical cord area (MUCCA), brain, and thalamus volume and their changes during a 1-year follow-up were measured and compared between groups. All the measurements were also compared between progressive and non-progressive groups in NMO and MS. RESULTS: MUCCA decreased significantly during the 1-year follow-up in NMO not in MS. Percentage brain volume changes (PBVC) and thalamus volume changes in MS were significantly higher than NMO. MUCCA changes were significantly different between progressive and non-progressive groups in NMO, while baseline brain lesion volume and PBVC were associated with disability progression in MS. MUCCA changes during 1-year follow-up showed association with clinical disability in NMO. CONCLUSION: Spinal cord atrophy changes were associated with disability progression in NMO, while baseline brain lesion load and whole brain atrophy changes were related to disability progression in MS. KEY POINTS: • Spinal cord atrophy progression was observed in NMO. • Spinal cord atrophy changes were associated with disability progression in NMO. • Brain lesion and atrophy were related to disability progression in MS.

Semi-automated delineation of breast cancer tumors and subsequent materialization using three-dimensional printing (rapid prototyping).

OBJECTIVE: Three-dimensional (3D) printing has become widely available, and a few cases of its use in clinical practice have been described. The aim of this study was to explore facilities for the semi-automated delineation of breast cancer tumors and to assess the feasibility of 3D printing of breast cancer tumors. METHODS: In a case series of five patients, different 3D imaging methods-magnetic resonance imaging (MRI), digital breast tomosynthesis (DBT), and 3D ultrasound-were used to capture 3D data for breast cancer tumors. The volumes of the breast tumors were calculated to assess the comparability of the breast tumor models, and the MRI information was used to render models on a commercially available 3D printer to materialize the tumors. RESULTS: The tumor volumes calculated from the different 3D methods appeared to be comparable. Tumor models with volumes between 325 mm3 and 7,770 mm3 were printed and compared with the models rendered from MRI. The materialization of the tumors reflected the computer models of them. CONCLUSION: 3D printing (rapid prototyping) appears to be feasible. Scenarios for the clinical use of the technology might include presenting the model to the surgeon to provide a better understanding of the tumor's spatial characteristics in the breast, in order to improve decision-making in relation to neoadjuvant chemotherapy or surgical approaches. J. Surg. Oncol. 2017;115:238-242. © 2016 Wiley Periodicals, Inc.

Analyzing myocardial torsion based on tissue phase mapping cardiovascular magnetic resonance.

BACKGROUND: The purpose of this work is to analyze differences in left ventricular torsion between volunteers and patients with non-ischemic cardiomyopathy based on tissue phase mapping (TPM) cardiovascular magnetic resonance (CMR). METHODS: TPM was performed on 27 patients with non-ischemic cardiomyopathy and 14 normal volunteers. Patients underwent a standard CMR including late gadolinium enhancement (LGE) for the assessment of myocardial scar and ECG-gated cine CMR for global cardiac function. TPM was acquired in short-axis orientation at base, mid, and apex for all subjects. After evaluation by experienced observers, the patients were divided in subgroups according to the presence or absence of LGE (LGE+/LGE-), local wall motion abnormalities (WM+/WM-), and having a preserved (≥50%) or reduced (<50%) ejection fraction (EF+/EF-). TPM data was semi-automatically segmented and global LV torsion was computed for each cardiac time frame for endocardial and epicardial layers, and for the entire myocardium. RESULTS: Maximum myocardial torsion was significantly lower for patients with reduced EF compared to controls (0.21 ± 0.15°/mm vs. 0.36 ± 0.11°/mm, p = 0.018), but also for patients with wall motion abnormalities (0.21 ± 0.13°/mm vs. 0.36 ± 0.11°/mm, p = 0.004). Global myocardial torsion showed a positive correlation (r = 0.54, p < 0.001) with EF. Moreover, endocardial torsion was significantly higher than epicardial torsion for EF+ subjects (0.56 ± 0.33°/mm vs. 0.34 ± 0.18°/mm, p = 0.039) and for volunteers (0.46 ± 0.16°/mm vs. 0.30 ± 0.09°/mm, p = 0.004). The difference in maximum torsion between endo- and epicardial layers was positively correlated with EF (r = 0.47, p = 0.002) and age (r = 0.37, p = 0.016) for all subjects. CONCLUSIONS: TPM can be used to detect significant differences in LV torsion in patients with reduced EF and in the presence of local wall motion abnormalities. We were able to quantify torsion differences between the endocardium and epicardium, which vary between patient subgroups and are correlated to age and EF.

Principles and methods for automatic and semi-automatic tissue segmentation in MRI data.

The development of magnetic resonance imaging (MRI) revolutionized both the medical and scientific worlds. A large variety of MRI options have generated a huge amount of image data to interpret. The investigation of a specific tissue in 3D or 4D MR images can be facilitated by image processing techniques, such as segmentation and registration. In this work, we provide a brief review of the principles and methods that are commonly applied to achieve superior tissue segmentation results in MRI. The impacts of MR image acquisition on segmentation outcome and the principles of selecting and exploiting segmentation techniques tailored for specific tissue identification tasks are discussed. In the end, two exemplary applications, breast and fibroglandular tissue segmentation in MRI and myocardium segmentation in short-axis cine and real-time MRI, are discussed to explain the typical challenges that can be posed in practical segmentation tasks in MRI data. The corresponding solutions that are adopted to deal with these challenges of the two practical segmentation tasks are thoroughly reviewed.

Fast and accurate identification of fat droplets in histological images.

BACKGROUND AND OBJECTIVE: The accurate identification of fat droplets is a prerequisite for the automatic quantification of steatosis in histological images. A major challenge in this regard is the distinction between clustered fat droplets and vessels or tissue cracks. METHODS: We present a new method for the identification of fat droplets that utilizes adjacency statistics as shape features. Adjacency statistics are simple statistics on neighbor pixels. RESULTS: The method accurately identified fat droplets with sensitivity and specificity values above 90%. Compared with commonly-used shape features, adjacency statistics greatly improved the sensitivity toward clustered fat droplets by 29% and the specificity by 17%. On a standard personal computer, megapixel images were processed in less than 0.05s. CONCLUSIONS: The presented method is simple to implement and can provide the basis for the fast and accurate quantification of steatosis.

Differential patterns of spinal cord and brain atrophy in NMO and MS.

OBJECTIVE: To investigate spinal cord and brain atrophy in neuromyelitis optica (NMO), and its relationship with other MRI measurements and clinical disability, compared with patients with multiple sclerosis (MS) and healthy controls (HC). METHODS: We recruited 35 patients with NMO, 35 patients with MS, and 35 HC, who underwent both spinal cord and brain MRI. Mean upper cervical cord area (MUCCA), brain parenchymal fraction (BPF), gray matter fraction (GMF), white matter fraction (WMF), and spinal cord and brain lesion loads were measured and compared among groups. Multivariate associations between MUCCA and brain volume measurement and clinical variables were assessed by partial correlations and multiple linear regression. RESULTS: Patients with NMO showed smaller MUCCA than HC (p = 0.004), and patients with MS had a trend of smaller MUCCA compared to HC (p = 0.07), with no significant difference between the patient groups. Patients with NMO showed lower BPF than HC, and patients with MS had lower BPF and GMF than patients with NMO. In NMO, MUCCA was correlated with Expanded Disability Status Scale score (EDSS), number of relapses, and total spinal cord lesion length, while in MS, MUCCA was correlated with WMF and EDSS. MUCCA was the only independent variable for predicting clinical disability measured by EDSS in NMO (R(2) = 0.55, p < 0.001) and MS (R(2) = 0.17, p = 0.013). CONCLUSION: NMO showed predominately spinal cord atrophy with mild brain atrophy, while MS demonstrated more brain atrophy, especially in the gray matter. MUCCA is the main MRI-derived parameter for explaining clinical disability in NMO and MS, and may serve as a potential biomarker for further clinical trials, especially in NMO.

Cervical spinal cord volume loss is related to clinical disability progression in multiple sclerosis.

OBJECTIVE: To examine the temporal evolution of spinal cord (SC) atrophy in multiple sclerosis (MS), and its association with clinical progression in a large MS cohort. METHODS: A total of 352 patients from two centres with MS (relapsing remitting MS (RRMS): 256, secondary progressive MS (SPMS): 73, primary progressive MS (PPMS): 23) were included. Clinical and MRI parameters were obtained at baseline, after 12 months and 24 months of follow-up. In addition to conventional brain and SC MRI parameters, the annualised percentage brain volume change and the annualised percentage upper cervical cord cross-sectional area change (aUCCA) were quantified. Main outcome measure was disease progression, defined by expanded disability status scale increase after 24 months. RESULTS: UCCA was lower in SPMS and PPMS compared with RRMS for all time points. aUCCA over 24 months was highest in patients with SPMS (-2.2% per year) and was significantly higher in patients with disease progression (-2.3% per year) than in stable patients (-1.2% per year; p=0.003), while annualised percentage brain volume change did not differ between subtypes (RRMS: -0.42% per year; SPMS -0.6% per year; PPMS: -0.46% per year) nor between progressive and stable patients (p=0.055). Baseline UCCA and aUCCA over 24 months were found to be relevant contributors of expanded disability status scale at month-24, while baseline UCCA as well as number of SC segments involved by lesions at baseline but not aUCCA were relevant contributors of disease progression. CONCLUSIONS: SC MRI parameters including baseline UCCA and SC lesions were significant MRI predictors of disease progression. Progressive 24-month upper SC atrophy occurred in all MS subtypes, and was faster in patients exhibiting disease progression at month-24.

Mean upper cervical cord area (MUCCA) measurement in long-standing multiple sclerosis: relation to brain findings and clinical disability.

BACKGROUND: The majority of patients with multiple sclerosis (MS) present with spinal cord pathology. Spinal cord atrophy is thought to be a marker of disease severity, but in long-disease duration its relation to brain pathology and clinical disability is largely unknown. OBJECTIVE: Our aim was to investigate mean upper cervical cord area (MUCCA) in patients with long-standing MS and assess its relation to brain magnetic resonance imaging (MRI) measures and clinical disability. METHODS: MUCCA was measured in 196 MS patients and 55 healthy controls using 3DT1-weighted cervical images obtained at 3T MRI. Clinical disability was measured using the Expanded Disability Status Scale (EDSS), Nine-Hole-Peg test (9-HPT), and 25 feet Timed Walk Test (TWT). Stepwise linear regression was performed to assess the association between MUCCA and MRI measures, and between MUCCA and clinical disability. RESULTS: MUCCA was smaller (mean 11.7%) in MS patients compared with healthy controls (72.56±9.82 and 82.24±7.80 mm2 respectively; p<0.001), most prominently in male patients. MUCCA was associated with normalized brain volume, and number of cervical cord lesions. MUCCA was independently associated with EDSS, TWT, and 9-HPT. CONCLUSION: MUCCA was reduced in MS patients compared with healthy controls. It provides a relevant marker for clinical disability in long-standing disease, independent of other MRI measures.

Segmentation-based partial volume correction for volume estimation of solid lesions in CT.

In oncological chemotherapy monitoring, the change of a tumor's size is an important criterion for assessing cancer therapeutics. Measuring the volume of a tumor requires its delineation in 3-D. This is called segmentation, which is an intensively studied problem in medical image processing. However, simply counting the voxels within a binary segmentation result can lead to significant differences in the volume, if the lesion has been segmented slightly differently by various segmentation procedures or in different scans, for example due to the limited spatial resolution of computed tomography (CT) or partial volume effects. This variability limits the sensitivity of size measurements and thus of therapy response assessments and it can even lead to misclassifications. We present a fast, generic algorithm for measuring the volume of solid, compact tumors in CT that considers partial volume effects at the border of a given segmentation result. The algorithm is an extension of the segmentation-based partial volume analysis proposed by Kuhnigk for the volumetry of solid lung lesions , such that it can be applied to inhomogeneous lesions and lesions with inhomogeneous surroundings. Our generalized segmentation-based partial volume correction is based on a spatial subdivision of the segmentation result, from which the fraction of tumor for each voxel is computed. It has been evaluated on phantom data, 1516 lesion segmentation pairs (lung nodules, liver metastases and lymph nodes) as well as 1851 lung nodules from the LIDC-IDRI database. The evaluations of our algorithm show a more accurate estimation of the real volume and its ability to reduce inter- and intra-observer variability significantly for each entity. Overall, the variability (interquartile range) for phantom data is reduced by 49% ( p ≪ 0.001) and the variability between different readers is reduced by 28% ( p ≪ 0.001). The average computation time is 0.2 s.

On the evaluation of segmentation editing tools.

Efficient segmentation editing tools are important components in the segmentation process, as no automatic methods exist that always generate sufficient results. Evaluating segmentation editing algorithms is challenging, because their quality depends on the user's subjective impression. So far, no established methods for an objective, comprehensive evaluation of such tools exist and, particularly, intermediate segmentation results are not taken into account. We discuss the evaluation of editing algorithms in the context of tumor segmentation in computed tomography. We propose a rating scheme to qualitatively measure the accuracy and efficiency of editing tools in user studies. In order to objectively summarize the overall quality, we propose two scores based on the subjective rating and the quantified segmentation quality over time. Finally, a simulation-based evaluation approach is discussed, which allows a more reproducible evaluation without the need for human input. This automated evaluation complements user studies, allowing a more convincing evaluation, particularly during development, where frequent user studies are not possible. The proposed methods have been used to evaluate two dedicated editing algorithms on 131 representative tumor segmentations. We show how the comparison of editing algorithms benefits from the proposed methods. Our results also show the correlation of the suggested quality score with the qualitative ratings.

A hybrid method towards automated nipple detection in 3D breast ultrasound images.

In clinical work-up of breast cancer, nipple position is an important marker to locate lesions. Moreover, it serves as an effective landmark to register a 3D automated breast ultrasound (ABUS) images to other imaging modalities, e.g., X-ray mammography, tomosynthesis or magnetic resonance imaging (MRI). However, the presence of speckle noises caused by the interference waves and variant imaging directions poses challenges to automatically identify nipple positions. In this work, a hybrid fully automatic method to detect nipple positions in ABUS images is presented. The method extends the multi-scale Laplacian-based method that we proposed previously, by integrating a specially designed Hessian-based method to locate the shadow area beneath the nipple and areola. Subsequently, the likelihood maps of nipple positions generated by both methods are combined to build a joint-likelihood map, where the final nipple position is extracted. To validate the efficiency and robustness, the extended hybrid method was tested on 926 ABUS images, resulting in a distance error of 7.08±10.96 mm (mean±standard deviation).