PET Imaging of Neurofibromatosis Type 1 with a Fluorine-18 Labeled Tryptophan Radiotracer.

Neurofibromatosis type 1 (NF1) is a neurocutaneous disorder. Plexiform neurofibromas (PNFs) are benign tumors commonly formed in patients with NF1. PNFs have a high incidence of developing into malignant peripheral nerve sheath tumors (MPNSTs) with a 5-year survival rate of only 30%. Therefore, the accurate diagnosis and differentiation of MPNSTs from benign PNFs are critical to patient management. We studied a fluorine-18 labeled tryptophan positron emission tomography (PET) radiotracer, 1-(2-[18F]fluoroethyl)-L-tryptophan (L-[18F]FETrp), to detect NF1-associated tumors in an animal model. An ex vivo biodistribution study of L-[18F]FETrp showed a similar tracer distribution and kinetics between the wild-type and triple mutant mice with the highest uptake in the pancreas. Bone uptake was stable. Brain uptake was low during the 90-min uptake period. Static PET imaging at 60 min post-injection showed L-[18F]FETrp had a comparable tumor uptake with [18F]fluorodeoxyglucose (FDG). However, L-[18F]FETrp showed a significantly higher tumor-to-brain ratio than FDG (n = 4, p < 0.05). Sixty-minute-long dynamic PET scans using the two radiotracers showed similar kidney, liver, and lung kinetics. A dysregulated tryptophan metabolism in NF1 mice was further confirmed using immunohistostaining. L-[18F]FETrp is warranted to further investigate differentiating malignant NF1 tumors from benign PNFs. The study may reveal the tryptophan-kynurenine pathway as a therapeutic target for treating NF1.

Spatially localized sparse approximations of deep features for breast mass characterization.

We propose a deep feature-based sparse approximation classification technique for classification of breast masses into benign and malignant categories in film screen mammographs. This is a significant application as breast cancer is a leading cause of death in the modern world and improvements in diagnosis may help to decrease rates of mortality for large populations. While deep learning techniques have produced remarkable results in the field of computer-aided diagnosis of breast cancer, there are several aspects of this field that remain under-studied. In this work, we investigate the applicability of deep-feature-generated dictionaries to sparse approximation-based classification. To this end we construct dictionaries from deep features and compute sparse approximations of Regions Of Interest (ROIs) of breast masses for classification. Furthermore, we propose block and patch decomposition methods to construct overcomplete dictionaries suitable for sparse coding. The effectiveness of our deep feature spatially localized ensemble sparse analysis (DF-SLESA) technique is evaluated on a merged dataset of mass ROIs from the CBIS-DDSM and MIAS datasets. Experimental results indicate that dictionaries of deep features yield more discriminative sparse approximations of mass characteristics than dictionaries of imaging patterns and dictionaries learned by unsupervised machine learning techniques such as K-SVD. Of note is that the proposed block and patch decomposition strategies may help to simplify the sparse coding problem and to find tractable solutions. The proposed technique achieves competitive performances with state-of-the-art techniques for benign/malignant breast mass classification, using 10-fold cross-validation in merged datasets of film screen mammograms.

Single slice thigh CT muscle group segmentation with domain adaptation and self-training.

Purpose: Thigh muscle group segmentation is important for assessing muscle anatomy, metabolic disease, and aging. Many efforts have been put into quantifying muscle tissues with magnetic resonance (MR) imaging, including manual annotation of individual muscles. However, leveraging publicly available annotations in MR images to achieve muscle group segmentation on single-slice computed tomography (CT) thigh images is challenging. Approach: We propose an unsupervised domain adaptation pipeline with self-training to transfer labels from three-dimensional MR to single CT slices. First, we transform the image appearance from MR to CT with CycleGAN and feed the synthesized CT images to a segmenter simultaneously. Single CT slices are divided into hard and easy cohorts based on the entropy of pseudo-labels predicted by the segmenter. After refining easy cohort pseudo-labels based on anatomical assumption, self-training with easy and hard splits is applied to fine-tune the segmenter. Results: On 152 withheld single CT thigh images, the proposed pipeline achieved a mean Dice of 0.888 (0.041) across all muscle groups, including gracilis, hamstrings, quadriceps femoris, and sartorius muscle. Conclusions: To our best knowledge, this is the first pipeline to achieve domain adaptation from MR to CT for thigh images. The proposed pipeline effectively and robustly extracts muscle groups on two-dimensional single-slice CT thigh images. The container is available for public use in GitHub repository available at: https://github.com/MASILab/DA_CT_muscle_seg.

Automated quantification of muscle and fat in the thigh from water-, fat-, and nonsuppressed MR images.

PURPOSE: To introduce and validate an unsupervised muscle and fat quantification algorithm based on joint analysis of water-suppressed (WS), fat-suppressed (FS), and water and fat (nonsuppressed) volumetric magnetic resonance imaging (MRI) of the mid-thigh region. MATERIALS AND METHODS: We first segmented the subcutaneous fat by use of a parametric deformable model, then applied centroid clustering in the feature domain defined by the voxel intensities in WS and FS images to identify the intermuscular fat and muscle. In the final step we computed volumetric and area measures of fat and muscle. We applied this algorithm on datasets of water-, fat-, and nonsuppressed volumetric MR images acquired from 28 participants. RESULTS: We validated our tissue composition analysis against fat and muscle area measurements obtained from semimanual analysis of single-slice mid-thigh computed tomography (CT) images of the same participants and found very good agreement between the two methods. Furthermore, we compared the proposed approach with a variant that uses nonsuppressed images only and observed that joint analysis of WS and FS images is more accurate than the nonsuppressed only variant. CONCLUSION: Our MRI algorithm produces accurate tissue quantification, is less labor-intensive, and more reproducible than the original CT-based workflow and can address interparticipant anatomic variability and intensity inhomogeneity effects.

Multi-atlas segmentation and quantification of muscle, bone and subcutaneous adipose tissue in the lower leg using peripheral quantitative computed tomography.

Accurate and reproducible tissue identification is essential for understanding structural and functional changes that may occur naturally with aging, or because of a chronic disease, or in response to intervention therapies. Peripheral quantitative computed tomography (pQCT) is regularly employed for body composition studies, especially for the structural and material properties of the bone. Furthermore, pQCT acquisition requires low radiation dose and the scanner is compact and portable. However, pQCT scans have limited spatial resolution and moderate SNR. pQCT image quality is frequently degraded by involuntary subject movement during image acquisition. These limitations may often compromise the accuracy of tissue quantification, and emphasize the need for automated and robust quantification methods. We propose a tissue identification and quantification methodology that addresses image quality limitations and artifacts, with increased interest in subject movement. We introduce a multi-atlas image segmentation (MAIS) framework for semantic segmentation of hard and soft tissues in pQCT scans at multiple levels of the lower leg. We describe the stages of statistical atlas generation, deformable registration and multi-tissue classifier fusion. We evaluated the performance of our methodology using multiple deformable registration approaches against reference tissue masks. We also evaluated the performance of conventional model-based segmentation against the same reference data to facilitate comparisons. We studied the effect of subject movement on tissue segmentation quality. We also applied the top performing method to a larger out-of-sample dataset and report the quantification results. The results show that multi-atlas image segmentation with diffeomorphic deformation and probabilistic label fusion produces very good quality over all tissues, even for scans with significant quality degradation. The application of our technique to the larger dataset reveals trends of age-related body composition changes that are consistent with the literature. Because of its robustness to subject motion artifacts, our MAIS methodology enables analysis of larger number of scans than conventional state-of-the-art methods. Automated analysis of both soft and hard tissues in pQCT is another contribution of this work.

Label efficient segmentation of single slice thigh CT with two-stage pseudo labels.

Purpose: Muscle, bone, and fat segmentation from thigh images is essential for quantifying body composition. Voxelwise image segmentation enables quantification of tissue properties including area, intensity, and texture. Deep learning approaches have had substantial success in medical image segmentation, but they typically require a significant amount of data. Due to the high cost of manual annotation, training deep learning models with limited human label data is desirable, but it is a challenging problem. Approach: Inspired by transfer learning, we proposed a two-stage deep learning pipeline to address the thigh and lower leg segmentation issue. We studied three datasets, 3022 thigh slices and 8939 lower leg slices from the BLSA dataset and 121 thigh slices from the GESTALT study. First, we generated pseudo labels for thigh based on approximate handcrafted approaches using CT intensity and anatomical morphology. Then, those pseudo labels were fed into deep neural networks to train models from scratch. Finally, the first stage model was loaded as the initialization and fine-tuned with a more limited set of expert human labels of the thigh. Results: We evaluated the performance of this framework on 73 thigh CT images and obtained an average Dice similarity coefficient (DSC) of 0.927 across muscle, internal bone, cortical bone, subcutaneous fat, and intermuscular fat. To test the generalizability of the proposed framework, we applied the model on lower leg images and obtained an average DSC of 0.823. Conclusions: Approximated handcrafted pseudo labels can build a good initialization for deep neural networks, which can help to reduce the need for, and make full use of, human expert labeled data.

Quantification of muscle, bones, and fat on single slice thigh CT.

Muscle, bone, and fat segmentation of CT thigh slice is essential for body composition research. Voxel-wise image segmentation enables quantification of tissue properties including area, intensity and texture. Deep learning approaches have had substantial success in medical image segmentation, but they typically require substantial data. Due to high cost of manual annotation, training deep learning models with limited human labelled data is desirable but also a challenging problem. Inspired by transfer learning, we proposed a two-stage deep learning pipeline to address this issue in thigh segmentation. We study 2836 slices from Baltimore Longitudinal Study of Aging (BLSA) and 121 slices from Genetic and Epigenetic Signatures of Translational Aging Laboratory Testing (GESTALT). First, we generated pseudo-labels based on approximate hand-crafted approaches using CT intensity and anatomical morphology. Then, those pseudo labels are fed into deep neural networks to train models from scratch. Finally, the first stage model is loaded as initialization and fine-tuned with a more limited set of expert human labels. We evaluate the performance of this framework on 56 thigh CT scans and obtained average Dice of 0.979,0.969,0.953,0.980 and 0.800 for five tissues: muscle, cortical bone, internal bone, subcutaneous fat and intermuscular fat respectively. We evaluated generalizability by manually reviewing external 3504 BLSA single thighs from 1752 thigh slices. The result is consistent and passed human review with 150 failed thigh images, which demonstrates that the proposed method has strong generalizability.

Discriminative Localized Sparse Approximations for Mass Characterization in Mammograms.

The most common form of cancer among women in both developed and developing countries is breast cancer. The early detection and diagnosis of this disease is significant because it may reduce the number of deaths caused by breast cancer and improve the quality of life of those effected. Computer-aided detection (CADe) and computer-aided diagnosis (CADx) methods have shown promise in recent years for aiding in the human expert reading analysis and improving the accuracy and reproducibility of pathology results. One significant application of CADe and CADx is for breast cancer screening using mammograms. In image processing and machine learning research, relevant results have been produced by sparse analysis methods to represent and recognize imaging patterns. However, application of sparse analysis techniques to the biomedical field is challenging, as the objects of interest may be obscured because of contrast limitations or background tissues, and their appearance may change because of anatomical variability. We introduce methods for label-specific and label-consistent dictionary learning to improve the separation of benign breast masses from malignant breast masses in mammograms. We integrated these approaches into our Spatially Localized Ensemble Sparse Analysis (SLESA) methodology. We performed 10- and 30-fold cross validation (CV) experiments on multiple mammography datasets to measure the classification performance of our methodology and compared it to deep learning models and conventional sparse representation. Results from these experiments show the potential of this methodology for separation of malignant from benign masses as a part of a breast cancer screening workflow.

Automated Cell Tracking Using Motion Prediction-Based Matching and Event Handling.

Automated cell segmentation and tracking enables the quantification of static and dynamic cell characteristics and is significant for disease diagnosis, treatment, drug development, and other biomedical applications. This paper introduces a method for fully automated cell tracking, lineage construction, and quantification. Cell detection is performed in the joint spatio-temporal domain by a motion diffusion-based Partial Differential Equation (PDE) combined with energy minimizing active contours. In the tracking stage, we adopt a variational joint local-global optical flow technique to determine the motion vector field. We utilize the predicted cell motion jointly with spatial cell features to define a maximum likelihood criterion to find inter-frame cell correspondences assuming Markov dependency. We formulate cell tracking and cell event detection as a graph partitioning problem. We propose a solution obtained by minimization of a global cost function defined over the set of all cell tracks. We construct a cell lineage tree that represents the cell tracks and cell events. Finally, we compute morphological, motility, and diffusivity measures and validate cell tracking against manually generated reference standards. The automated tracking method applied to reference segmentation maps produces an average tracking accuracy score ( TRA) of 99 percent, and the fully automated segmentation and tracking system produces an average TRA of 89 percent.

Spatially localized sparse representations for breast lesion characterization.

RATIONALE: The topic of sparse representation of samples in high dimensional spaces has attracted growing interest during the past decade. In this work, we develop sparse representation-based methods for classification of radiological imaging patterns of breast lesions into benign and malignant states. METHODS: We propose a spatial block decomposition method to address irregularities of the approximation problem and to build an ensemble of classifiers (CL) that we expect to yield more accurate numerical solutions than conventional whole-region of interest (ROI) sparse analyses. We introduce two classification decision strategies based on maximum a posteriori probability (BBMAP-S), or a log likelihood function (BBLL-S). RESULTS: To evaluate the performance of the proposed approach we used cross-validation techniques on imaging datasets with disease class labels. We utilized the proposed approach for separation of breast lesions into benign and malignant categories in mammograms. The level of difficulty is high in this application and the accuracy may depend on the lesion size. Our results indicate that the proposed integrative sparse analysis addresses the ill-posedness of the approximation problem, producing AUC (area under the receiver operating curve) value of 89.1% for randomized 30-fold cross-validation. CONCLUSIONS: Furthermore, our comparative experiments showed that the BBLL-S decision function may yield more accurate classification than BBMAP-S because BBLL-S accounts for possible estimation bias.

Integrative blockwise sparse analysis for tissue characterization and classification.

The topic of sparse representation of samples in high dimensional spaces has attracted growing interest during the past decade. In this work, we develop sparse representation-based methods for classification of clinical imaging patterns into healthy and diseased states. We propose a spatial block decomposition method to address irregularities of the approximation problem and to build an ensemble of classifiers that we expect to yield more accurate numerical solutions than conventional sparse analyses of the complete spatial domain of the images. We introduce two classification decision strategies based on maximum a posteriori probability (BBMAP), or a log likelihood function (BBLL) and an approach to adjusting the classification decision criteria. To evaluate the performance of the proposed approach we used cross-validation techniques on imaging datasets with disease class labels. We first applied the proposed approach to diagnosis of osteoporosis using bone radiographs. In this problem we assume that changes in trabecular bone connectivity can be captured by intensity patterns. The second application domain is separation of breast lesions into benign and malignant categories in mammograms. The object classes in both of these applications are not linearly separable, and the classification accuracy may depend on the lesion size in the second application. Our results indicate that the proposed integrative sparse analysis addresses the ill-posedness of the approximation problem and produces very good class separation for trabecular bone characterization and for breast lesion characterization. Our approach yields higher classification rates than conventional sparse classification and previously published convolutional neural networks (CNNs) that we fine-tuned for our datasets, or utilized for feature extraction. The BBLL technique also produced higher classification rates than learners using hand-crafted texture features, and the Bag of Keypoints, which is a sophisticated patch-based method. Furthermore, our comparative experiments showed that the BBLL function may yield more accurate classification than BBMAP, because BBLL accounts for possible estimation bias.

Region-based image registration for remote sensing imagery.

We propose an automatic region-based registration method for remote sensing imagery. In this method, we aim to register two images by matching region properties to address possible errors caused by local feature estimators. We apply automated image segmentation to identify the regions and calculate regional Fourier descriptors and standardized regional intensity descriptors for each region. We define a joint matching cost, as a linear combination of Euclidean distances, to establish and extract correspondences between regions. The segmentation technique utilizes kernel density estimators for edge localization, followed by morphological reconstruction and the watershed transform. We evaluated the registration performance of our method on synthetic and real datasets. We measured the registration accuracy by calculating the root-mean-squared error (RMSE) between the estimated transformation and the ground truth transformation. The results obtained using the joint intensity-Fourier descriptor were compared to the results obtained using Harris, Minimum eigenvalue, Features accelerated segment test (FAST), speeded-up robust features (SURF), binary robust invariant scalable keypoints (BRISK) and KAZE keypoint descriptors. The joint intensity-Fourier descriptor yielded average RMSE of 0.446 ± 0.359 pixels and 1.152 ± 0.488 pixels on two satellite imagery datasets consisting of 35 image pairs in total. These results indicate the capacity of the proposed technique for high accuracy. Our method also produces a lower registration error than the compared feature-based methods.

Automated skeletal tissue quantification in the lower leg using peripheral quantitative computed tomography.

OBJECTIVE: In this paper we introduce a methodology for hard and soft tissue quantification at proximal, intermediate and distal tibia sites using peripheral quantitative computed tomography scans. Quantification of bone properties is crucial for estimating bone structure resistance to mechanical stress and adaptations to loading. Soft tissue variables can be computed to investigate muscle volume and density, muscle-bone relationship, and fat infiltration. APPROACH: We employed implicit active contour models and clustering techniques for automated segmentation and identification of bone, muscle and fat at [Formula: see text], [Formula: see text], and [Formula: see text] tibia length. Next, we calculated densitometric, area and shape characteristics for each tissue type. We implemented our approach as a multi-platform tool denoted by TIDAQ (tissue identification and quantification) to be used by clinical researchers. MAIN RESULTS: We validated the proposed method against reference quantification measurements and tissue delineations obtained by semi-automated workflows. The average Deming regression slope between the tested and reference method was 1.126 for cross-sectional areas and 1.078 for mineral densities, indicating very good agreement. Our method produced high average coefficient of variation (R 2) estimates: 0.935 for cross-sectional areas and 0.888 for mineral densities over all tibia sites. In addition, our tissue segmentation approach achieved an average Dice coefficient of 0.91 over soft and hard tissues, indicating very good delineation accuracy. SIGNIFICANCE: Our methodology should allow for high throughput, accurate and reproducible automatic quantification of muscle and bone characteristics of the lower leg. This information is critical to evaluate risk of future adverse outcomes and assess the effect of medications, hormones, and behavioral interventions aimed at improving bone and muscle strength.

Measurement of fat fraction in the human thymus by localized NMR and three-point Dixon MRI techniques.

PURPOSE: To develop a protocol to non-invasively measure and map fat fraction, fat/(fat+water), as a function of age in the adult thymus for future studies monitoring the effects of interventions aimed at promoting thymic rejuvenation and preservation of immunity in older adults. MATERIALS AND METHODS: Three-dimensional spoiled gradient echo 3T MRI with 3-point Dixon fat-water separation was performed at full inspiration for thymus conspicuity in 36 volunteers 19 to 56 years old. Reproducible breath-holding was facilitated by real-time pressure recording external to the console. The MRI method was validated against localized spectroscopy in vivo, with ECG triggering to compensate for stretching during the cardiac cycle. Fat fractions were corrected for T1 and T2 bias using relaxation times measured using inversion recovery-prepared PRESS with incremented echo time. RESULTS: In thymus at 3 T, T1water = 978 ±â€¯75 ms, T1fat = 323 ±â€¯37 ms, T2water = 43.4 ±â€¯9.7 ms and T2fat = 52.1 ±â€¯7.6 ms were measured. Mean T1-corrected MRI fat fractions varied from 0.2 to 0.8 and were positively correlated with age, weight and body mass index (BMI). In subjects with matching MRI and MRS fat fraction measurements, the difference between these measurements exhibited a mean of -0.008 with a 95% confidence interval of (0.123, -0.138). CONCLUSIONS: 3-point Dixon MRI of the thymus with T1 bias correction produces quantitative fat fraction maps that correlate with T2-corrected MRS measurements and show age trends consistent with thymic involution.

Anatomical equivalence class: a morphological analysis framework using a lossless shape descriptor.

Methods of computational anatomy are typically based on a spatial transformation that maps a template to an individual anatomy and vice versa. However, important morphological characteristics are frequently not captured by this transformation, thereby leading to lossy representations. We extend this formulation by incorporating residual anatomical information, i.e., information that is not captured by the shape transformation but is necessary in order to fully and exactly reconstruct the anatomy under measurement. We, therefore, arrive at a lossless morphological representation. By virtue of being lossless, this representation allows us to represent the same anatomy by an infinite number of pairs [transformation, residual], since different residuals correspond to different transformations. We treat these pairs as members of an anatomical equivalence class (AEC), which we approximate using principal component analysis. We show that projection onto the orthogonal to the AEC subspace produces measurements that allow us to better detect morphological abnormalities by eliminating variation in the data that is irrelevant and confounds underlying subtle morphological characteristics. Finally, we show that higher classification rates between a group of normal brains and a group of brains with localized atrophy are obtained if we use nonmetric distances between AECs instead of conventional Euclidean distances between individual morphological measurements. The results confirm that this representation can improve the results compared to conventional analysis, but also highlight limitations of the current approach and point to directions of further development of this general morphological analysis framework.

Differential Aging Signals in Abdominal CT Scans.

RATIONALE AND OBJECTIVES: Changes in the composition of body tissues are major aging phenotypes, but they have been difficult to study in depth. Here we describe age-related change in abdominal tissues observable in computed tomography (CT) scans. We used pattern recognition and machine learning to detect and quantify these changes in a model-agnostic fashion. MATERIALS AND METHODS: CT scans of abdominal L4 sections were obtained from Baltimore Longitudinal Study of Aging (BLSA) participants. Age-related change in the constituent tissues were determined by training machine classifiers to differentiate age groups within male and female strata ("Younger" at 50-70 years old vs "Older" at 80-99 years old). The accuracy achieved by the classifiers in differentiating the age cohorts was used as a surrogate measure of the aging signal in the different tissues. RESULTS: The highest accuracy for discriminating age differences was 0.76 and 0.72 for males and females, respectively. The classification accuracy was 0.79 and 0.71 for adipose tissue, 0.70 and 0.68 for soft tissue, and 0.65 and 0.64 for bone. CONCLUSIONS: Using image data from a large sample of well-characterized pool of participants dispersed over a wide age range, we explored age-related differences in gross morphology and texture of abdominal tissues. This technology is advantageous for tracking effects of biological aging and predicting adverse outcomes when compared to the traditional use of specific molecular biomarkers. Application of pattern recognition and machine learning as a tool for analyzing medical images may provide much needed insight into tissue changes occurring with aging and, further, connect these changes with their metabolic and functional consequences.

Bone texture characterization for osteoporosis diagnosis using digital radiography.

We introduce texture classification techniques to effectively diagnose osteoporosis in bone radiography data. Osteoporosis is an age-related systemic bone skeletal disorder characterized by low bone mass and bone structure deterioriation that results in increased bone fragility and higher fracture risk. Therefore, early diagnosis can effectively predict fracture risk and prevent the disease. Automated diagnosis from digital radiographs is very challenging since the scans of healthy and osteoporotic subjects show little or no visual differences, and their density histograms mostly overlap. We designed a system to separate healthy from osteoporotic subjects using high-dimensional textural feature representations computed from radiographs. These features were then reduced using feature selection to obtain the more discriminant subset that was finally classified by our methods. The top performing approach yields 79.3% accuracy and 81% area under the ROC over 116 bone radiographs.

Joint level-set and spatio-temporal motion detection for cell segmentation.

BACKGROUND: Cell segmentation is a critical step for quantification and monitoring of cell cycle progression, cell migration, and growth control to investigate cellular immune response, embryonic development, tumorigenesis, and drug effects on live cells in time-lapse microscopy images. METHODS: In this study, we propose a joint spatio-temporal diffusion and region-based level-set optimization approach for moving cell segmentation. Moving regions are initially detected in each set of three consecutive sequence images by numerically solving a system of coupled spatio-temporal partial differential equations. In order to standardize intensities of each frame, we apply a histogram transformation approach to match the pixel intensities of each processed frame with an intensity distribution model learned from all frames of the sequence during the training stage. After the spatio-temporal diffusion stage is completed, we compute the edge map by nonparametric density estimation using Parzen kernels. This process is followed by watershed-based segmentation and moving cell detection. We use this result as an initial level-set function to evolve the cell boundaries, refine the delineation, and optimize the final segmentation result. RESULTS: We applied this method to several datasets of fluorescence microscopy images with varying levels of difficulty with respect to cell density, resolution, contrast, and signal-to-noise ratio. We compared the results with those produced by Chan and Vese segmentation, a temporally linked level-set technique, and nonlinear diffusion-based segmentation. We validated all segmentation techniques against reference masks provided by the international Cell Tracking Challenge consortium. The proposed approach delineated cells with an average Dice similarity coefficient of 89 % over a variety of simulated and real fluorescent image sequences. It yielded average improvements of 11 % in segmentation accuracy compared to both strictly spatial and temporally linked Chan-Vese techniques, and 4 % compared to the nonlinear spatio-temporal diffusion method. CONCLUSIONS: Despite the wide variation in cell shape, density, mitotic events, and image quality among the datasets, our proposed method produced promising segmentation results. These results indicate the efficiency and robustness of this method especially for mitotic events and low SNR imaging, enabling the application of subsequent quantification tasks.