Abstract Background Antibody-drug conjugates (ADCs) have emerged as a promising therapeutic class for solid tumors, especially in breast cancer (BC), with a focus on agents targeting HER2 and Trop2. Currently, two ADC drugs targeting HER2: Trastuzumab emtansine (T-DM1) and fam-trastuzumab deruxtecan-nxki (T-DXd), and two targeting Trop2: Sacituzumab govitecan (SG) and datopotomab deruxtecan-dlnk (Dato-DXd) have gained approval by the FDA in treating metastatic BC, with additional ADCs in development. Despite the advances, challenges remain in patient selection including identifying the HER2 “ultra-low” patient subgroup, characterizing HER2/Trop2 co-expression, and ensuring reliable evaluation of biomarkers on the cell membrane surface. Here, we present the fully validated Opal™ multiplex-immunofluorescence (mIF) BC-ADC panel, leveraging the IVD-grade Akoya PhenoImager HT imaging system. The BC-ADC panel provides simultaneous assessment of BC standard-of-care biomarkers with Ki67 and hormone receptor (HR), ADC-specific targets HER2 and TROP2, together with a proprietary membrane marker cocktail facilitating specific biomarker membrane evaluation. Methods A proprietary membrane cocktail, comprised of multiple membrane-specific biomarkers, was developed to specifically label and visualize cell membranes. The membrane cocktail performance supporting accurate cell definition and subcellular localization was assessed against ground truth boundaries created by board-certified pathologists. Biomarkers TROP2, HER2, Ki67, ER, PR and the membrane cocktail, plus DAPI were tested for staining order effects and spectral unmixing, then developed into an single slide Opal™ 6-color multiplex immunofluorescence panel. HER2-positive/HER2-negative breast cancer tissue was utilized for assay development and verification. Board-certified pathologists evaluated panel validation according to the pre-set acceptance criteria including: absence of spectral crosstalk or umbrella effect by antibody drop-test; assay robustness by inter/intra-day triplicate precision studies, and biomarker sensitivity and specificity confirmed by equivalency between “gold-standard” IHC and mIF. HER2-negative breast cancer samples utilized for testing contained samples spanning HercepTest score 0+ (null or ultra-low), score 1+ (low), and score 2+ (low). Assay HER2 signal sensitivity was evaluated by pathologists to differentiate HER2-negative scores qualitatively. Results The membrane cocktail consistently delineates cell membrane in both tumor and stromal compartments. The finalized mIF assay met the predefined validation criteria as assessed by board-certified pathologists. Antibody drop-testing revealed no umbrella effect or spectral crosstalk was observed between channels. The precision study demonstrated high reproducibility in inter-day comparison and intra-day comparison. Strong equivalency of each biomarker was observed between IHC and the mIF panel. Pathologist review confirmed the high sensitivity in differentiating HER2 expression levels across HER2-negative cases: null/ultra-low/low. Conclusion The validated 5-plex/6-color BC-ADC mIF panel demonstrated robust analytical performance: high sensitivity and specificity across all biomarkers, especially in HER2-negative samples, stratifying HER2 low and ultra-low subgroups. The mIF panel showed reliable isolation of markers and robust assay performance. Results also suggested that the custom membrane cocktail can enhance the subcellular evaluation of biomarkers. The study will be supplemented by an independent patient verification cohort to support forthcoming clinical-utility studies. Citation Format: Y. Liu, M. Parrella, S. Schwaegerle, S. Berry, C. Betts, M. Landers. Development and Validation of TROP2/HER2-low/Ki67/HR multiplex-immunofluorescence panel with membrane identification: enabling improved Antibody Drug Conjugate therapy selection [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS4-01-26.

Carotid plaque classification based on ultrasound echogenicity and quantification of plaque burden are crucial in stroke risk assessment. In this work, we propose a framework that leverages the synergy between classification and segmentation by sharing plaque location information to enhance the performance of both tasks. Our cascaded framework integrates a ResNet-based classifier (Masked-ResNet-DS) with MedSAM, a medically adapted version of the Segment Anything Model for joint classification and segmentation of carotid plaques from 2D ultrasound images. Ground truth boundaries are used to guide region-specific feature pooling in the classifier, helping it focus on plaques during training. Since ground truth boundaries are unavailable at inference, we introduce a two-iteration strategy: the first generates a class activation map (CAM), which is then used for focused pooling in the second iteration to predict plaque type. The CAM is also used as a prompt to guide MedSAM for segmentation. To ensure accurate localization, the CAM is supervised during training using a Dice loss against the segmentation ground truth. Masked-ResNet-DS achieves a mean F1-score of 96.7% in plaque classification, at least 3.2% higher than competing methods. Ablation studies confirm that ground truth-based pooling and CAM supervision both improve classification. CAM-guided MedSAM achieves a Dice similarity coefficient (DSC) of 86.6%, outperforming U-Net and nnU-Net by 5.9% and 3.6%, respectively. In addition, CAM prompts improve MedSAM's DSC by 2.2%. By sharing plaque location between classification and segmentation, the proposed method improves both tasks and provides a more accurate tool for stroke risk stratification.

3D segmentation and volumetry of vestibular schwannomas (VSs) is a more accurate method to determine tumor growth on serial imaging, but manual annotation is time-consuming to implement in routine clinical practice. We evaluated and compared 5 deep learning-based segmentation models (nnUNet [Base, ResEncL], U-Mamba, UNETR, and MedSAM) for 3D VS segmentation and volumetry, and we examined the robustness to acquisition heterogeneity and generalization on an external cohort. Our refined internal data set consisted of T1-contrast-enhanced images, including 2692 scans (n =383 patients) for training and 277 scans (n = 97 patients) for testing. Post-model training and validation, performance was evaluated on both internal and a publicly available external test set (n = 241) using the Dice similarity coefficient, maximum distance between the predicted and ground truth boundaries (Hausdorff distance), surface-to-surface (S2S) distance, and relative volume error (RVE). A subanalysis of the model performance was also performed to evaluate the impact of tumor volumes and data set heterogeneity. The median Dice score on the external test set varied between 0.899 and 0.927 with U-Mamba achieving the highest performance, followed by nnUNet (Base and ResEncL). For these top 3 models, the median Hausdorff distance was 3.59 mm, while the 95th percentile Hausdorff distance was 1.6 mm. The S2S distance was <1 mm, and the median RVE (%) varied between 0.07 and 0.08. The median Dice scores were lower, 0.848-0.85, for smaller tumors (<200 mm3) and higher for tumors of >400 mm3 (median Dice score, 0.925-0.932). Models based on convolutional neural networks, transformer networks, and foundational models show robust performance for VS segmentation. Given the consistently high performance and self-optimizing frameworks of convolutional neural network-based models (U-Mamba, nnUNet), these may be more suitable for clinical applications.

BackgroundSimultaneous non‐contrast angiography and intraplaque hemorrhage (SNAP) imaging allows multi‐contrast MR images with large longitudinal coverage to be acquired in a single scan. With vessel wall boundaries available, vulnerable plaque components can be detected automatically from SNAP images. However, since SNAP imaging has not been previously used for vessel wall identification, vessel wall boundaries were required to be segmented from conventional multi‐contrast MRI first before registering to SNAP images. This registration process is not only time‐consuming but also prone to errors, potentially compromising subsequent plaque component analysis.PurposeWe aim to develop a model that directly segments the vessel wall from SNAP images, thereby eliminating the need for registration from another modality. The proposed model mitigates label noise arising from boundary misregistration.MethodsThe proposed framework has a student‐mean teacher architecture, trained in two phases: (i) a warm‐up phase, in which the model was trained by well‐registered manual segmentations and minimizes Dice loss between predictions and manual labels and (ii) a fine‐tuning phase, in which the model was trained by both well‐registered and misaligned manual segmentations. This phase involves adversarial training with the fast gradient sign method (FGSM) and a novel surrogate label generator. The generator produced surrogate ground truth boundaries for each misaligned image by computing a weighted sum of the manual segmentation and the pseudo‐label, generated through selective hardening of predicted probabilities from the student and mean teacher models. The sum of the adversarial training loss and the Dice loss between the manual and predicted segmentations was minimized to obtain the final segmentation result. During inference, the averaged probability maps from the student and mean teacher models were used to assign voxels to their most probable class. This study utilized 129 image volumes (1474 axial slices), of which 74 volumes (810 axial slices) were well‐registered and 55 volumes (664 axial slices) were misaligned. Training involved 110 volumes (55 well‐registered and 55 misaligned), while validation and testing sets comprised 9 and 10 well‐registered volumes, respectively.ResultsThe proposed method outperformed existing noisy label learning methods when trained by the same set of misaligned segmentations. Results demonstrate our method's superiority with Dice similarity coefficient of 73.51±11.08%, 90.76±10.21%, and 90.10±10.38% for the vessel wall, lumen, and outer wall segmentations, respectively.ConclusionThe proposed segmentation framework effectively integrates noisy and reliable labels to produce accurate vessel wall segmentations directly from SNAP images. By eliminating the need for manual segmentation and inter‐modality registration, this approach facilitates more detailed plaque component analysis with reduced interslice distance across a longer arterial segment.

Quantification of carotid atherosclerosis is important in monitoring patients at risk of cardiovascular events and in evaluating therapies. High-resolution 3D carotid magnetic resonance imaging (MRI) has been developed to provide extended coverage of the carotid arteries. However, the extended coverage poses a challenge as several hundreds of 2D axial images are required to be segmented for analysis. We propose a multi-dimensional hybrid framework that requires only a sparse set of manual segmentation. Dense surrogate ground truth boundaries required to train the framework are automatically generated by propagating the sparse manual segmentation using the proposed region of interest (ROI) U-Net. Furthermore, the Point U-Net was developed to generate surrogate ground truth for carotid branches without manual segmentation. The proposed framework leverages the advantages of 3D and 2D convolution neural networks (CNNs) to segment the outer wall and lumen from 3D MRI. The 3D multiscale U-Net provides a rough outer wall segmentation, which serves as the ROI to guide outer wall and lumen segmentation by the 2D ROI U-Net. The 3D Multiscale U-Net localizes the ROI automatically, bypassing the need for manual ROI identification. The 3D Multiscale U-Net was further improved by a 3D inception module installed at the bottleneck and the novel loss functions that promote longitudinal continuity and minimize the overlap of the internal and external carotid arteries. Extensive evaluation on the publicly available Carotid Artery Vessel Wall Segmentation challenge dataset shows that our approach outperforms the top-ranked solution in the challenge and state-of-the-art segmentation methods.

Medical image auto-segmentation techniques are basic and critical for numerous image-based analysis applications that play an important role in developing advanced and personalized medicine. Compared with manual segmentations, auto-segmentations are expected to contribute to a more efficient clinical routine and workflow by requiring fewer human interventions or revisions to auto-segmentations. However, current auto-segmentation methods are usually developed with the help of some popular segmentation metrics that do not directly consider human correction behavior. Dice Coefficient (DC) focuses on the truly-segmented areas, while Hausdorff Distance (HD) only measures the maximal distance between the auto-segmentation boundary with the ground truth boundary. Boundary length-based metrics such as surface DC (surDC) and Added Path Length (APL) try to distinguish truly-predicted boundary pixels and wrong ones. It is uncertain if these metrics can reliably indicate the required manual mending effort for application in segmentation research. Therefore, in this paper, the potential use of the above four metrics, as well as a novel metric called Mendability Index (MI), to predict the human correction effort is studied with linear and support vector regression models. 265 3D computed tomography (CT) samples for 3 objects of interest from 3 institutions with corresponding auto-segmentations and ground truth segmentations are utilized to train and test the prediction models. The five-fold cross-validation experiments demonstrate that meaningful human effort prediction can be achieved using segmentation metrics with varying prediction errors for different objects. The improved variant of MI, called MIhd, generally shows the best prediction performance, suggesting its potential to indicate reliably the clinical value of auto-segmentations.

Medical image segmentation is a critical step in many imaging applications. Automatic segmentation has gained extensive concern using a convolutional neural network (CNN). However, the traditional CNN-based methods fail to extract global and long-range contextual information due to local convolution operation. Transformer overcomes the limitation of CNN-based models. Inspired by the success of transformers in computer vision (CV), many researchers focus on designing the transformer-based U-shaped method in medical image segmentation. The transformer-based approach cannot effectively capture the fine-grained details. This paper proposes a dual encoder network with transformer-CNN for multi-organ segmentation. The new segmentation framework takes full advantage of CNN and transformer to enhance the segmentation accuracy. The Swin-transformer encoder extracts global information, and the CNN encoder captures local information. We introduce fusion modules to fuse convolutional features and the sequence of features from the transformer. Feature fusion is concatenated through the skip connection to smooth the decision boundary effectively. We extensively evaluate our method on the synapse multi-organ CT dataset and the automated cardiac diagnosis challenge (ACDC) dataset. The results demonstrate that the proposed method achieves Dice similarity coefficient (DSC) metrics of 80.68% and 91.12% on the synapse multi-organ CT and ACDC datasets, respectively. We perform the ablation studies on the ACDC dataset, demonstrating the effectiveness of critical components of our method. Our results match the ground-truth boundary more consistently than the existing models. Our approach gains more accurate results on challenging 2D images for multi-organ segmentation. Compared with the state-of-the-art methods, our proposed method achieves superior performance in multi-organ segmentation tasks. Graphical Abstract The key process in medical image segmentation.

Chromatin structural domains, or topologically associated domains (TADs), are a general organizing principle in chromatin biology. RNA polymerase II (RNAPII) mediates multiple chromatin interactive loops, tethering together as RNAPII-associated chromatin interaction domains (RAIDs) to offer a framework for gene regulation. RAID and TAD alterations have been found to be associated with diseases. They can be further dissected as micro-domains (micro-TADs and micro-RAIDs) by clustering single-molecule chromatin-interactive complexes from next-generation three-dimensional (3D) genome techniques, such as ChIA-Drop. Currently, there are few tools available for micro-domain boundary identification. In this work, we developed the MCI-frcnn deep learning method to train a Faster Region-based Convolutional Neural Network (Faster R-CNN) for micro-domain boundary detection. At the training phase in MCI-frcnn, 50 images of RAIDs from Drosophila RNAPII ChIA-Drop data, containing 261 micro-RAIDs with ground truth boundaries, were trained for 7days. Using this well-trained MCI-frcnn, we detected micro-RAID boundaries for the input new images, with a fast speed (5.26 fps), high recognition accuracy (AUROC = 0.85, mAP = 0.69), and high boundary region quantification (genomic IoU = 76%). We further applied MCI-frcnn to detect human micro-TADs boundaries using human GM12878 SPRITE data and obtained a high region quantification score (mean gIoU = 85%). In all, the MCI-frcnn deep learning method which we developed in this work is a general tool for micro-domain boundary detection.

Pedestrian behavior recognition in the driving environment is an important technology to prevent pedestrian accidents by predicting the next movement. It is necessary to recognize current pedestrian behavior to predict future pedestrian behavior. However, many studies have recognized human visible characteristics such as face, body parts or clothes, but few have recognized pedestrian behavior. It is challenging to recognize pedestrian behavior in the driving environment due to the changes in the camera field of view due to the illumination conditions in outdoor environments and vehicle movement. In this paper, to predict pedestrian behavior, we introduce a position-information added two-stream convolutional neural network (CNN) with multi task learning that is robust to the limited conditions of the outdoor driving environment. The conventional two-stream CNN is the most widely used model for human-action recognition. However, the conventional two-stream CNN based on optical flow has limitations regarding pedestrian behavior recognition in a moving vehicle because of the assumptions of brightness constancy and piecewise smoothness. To solve this problem for a moving vehicle, the binary descriptor dense scale-invariant feature transform (SIFT) flow, a feature-based matching algorithm, is robust in moving-pedestrian behavior recognition, such as walking and standing, in a moving vehicle. However, recognizing cross attributes, such as crossing or not crossing the street, is challenging using the binary descriptor dense SIFT flow because people who cross the road or not act the same walking action, but their location on the image is different. Therefore, pedestrian position information should be added to the conventional binary descriptor dense SIFT flow two-stream CNN. Thus, learning biased toward action attributes is evenly learned across action and cross attributes. In addition, YOLO detection and the Siamese tracker are used instead of the ground-truth boundary box to prove the robustness in the action- and cross-attribute recognition from a moving vehicle. The JAAD and PIE datasets were used for training, and only the JAAD dataset was used as a testing dataset for comparison with other state-of-the-art research on multitask and single-task learning.

This paper proposes a novel active boundary loss for semantic segmentation. It can progressively encourage the alignment between predicted boundaries and ground-truth boundaries during end-to-end training, which is not explicitly enforced in commonly used cross-entropy loss. Based on the predicted boundaries detected from the segmentation results using current network parameters, we formulate the boundary alignment problem as a differentiable direction vector prediction problem to guide the movement of predicted boundaries in each iteration. Our loss is model-agnostic and can be plugged in to the training of segmentation networks to improve the boundary details. Experimental results show that training with the active boundary loss can effectively improve the boundary F-score and mean Intersection-over-Union on challenging image and video object segmentation datasets.

Active contours model (ACM) has been extensively used in computer vision and image processing. In recent studies, convolutional neural networks (CNNs) have been combined with ACM replacing the user in the process of contour evolution and image segmentation to eliminate limitations associated with ACM dependence on energy functional parameters and initialization. However, prior studies did not aim for automatic initialization, which is addressed in this article. In addition to manual initialization, current methods are highly sensitive to the initial location and fail to delineate borders accurately. We propose a fully automatic image segmentation method to address problems of manual initialization, insufficient capture range, and poor convergence to boundaries, in addition to the problem of assignment of energy functional parameters. We train two CNNs, one of which generating ACM weighting parameters and the other generating a ground truth mask to extract distance transform (DT) and an initialization circle. DT is used to form a vector field pointing from each pixel of the image towards the closest ground truth boundary point. Vector magnitudes are equal to the Euclidean distance between each pixel and the closest ground truth boundary point. We evaluate our method on four publicly available datasets, including two building instance segmentation datasets, i.e., Vaihingen and Bing huts, and two mammography image datasets, INBreast and DDSM-BCRP. Our approach achieves state-of-the-art results in mean Intersection over Union (mIoU), Dice similarity coefficient and Boundary F-score (BoundF) with the values of 92.33%, 92.44%, and 86.57% for Vaihingen dataset, and 87.12%, 86.86%, and 66.91% for Bing huts dataset. We obtained the Dice similarity coefficient values of 94.23% and 90.89% for the INBreast and DDSM-BCRP, respectively.

Edge supervision and multi-scale cost volume for stereo matching

Breast cancer is the leading type of cancer diagnosed in women nowadays and for breast screening, mammography is preferred to detect and diagnose the cancer by detecting the masses with the help of Computer-aided Diagnosis (CAD) system. It helps to assist radiologists in getting accurate diagnosis. An approach is proposed to effectively detect the masses in mammographic breast cancer images by using Modified Histogram based adaptive thresholding (MHAT) method. The algorithm has been tested over with 100 mammographic images and the experimental results show that the detection method has a sensitivity of 98.3% at 0.78 false positives with accuracy of 99% per image. We evaluated the performance of our MHAT algorithm by comparing with respect to the ground-truth boundary drawn by an expert radiologist. The results are clinically relevant, according to the radiologists who evaluated the results.

Exploiting forced alignment of time-reversed data for improving HMM-based handwriting segmentation

Rupture of atherosclerotic plaques in the carotid artery has been implicated in 20% of strokes. 3D ultrasound (US) imaging is emerging as an attractive method to quantify plaque burden and track changes in plaque longitudinally over time. However, plaque segmentation from US images is challenging because of poor boundary contrast and shadowing. The objective of this study is to develop and evaluate a semiautomatic segmentation algorithm with a novel stopping criterion for segmenting outer wall boundary (OWB) and lumen intima boundary (LIB) of common, internal, and external carotid artery from 3D US images for quantifying the vessel wall volume (VWV). 3D US image volumes were acquired from ten subjects with asymptomatic carotid stenoses. Volumes were acquired using a mechanically scanned linear probe, and the reconstructed volume consisted of 21 slices acquired at an interslice distance of 1 mm. The authors used distance regularized level set method with edge-based energy, region-based energy, smoothness energy, and a novel stopping criterion to segment the LIB and OWB of carotid artery. The algorithm was initialized by six user-selected points on the LIB and OWB in seven 2D cross-sectional slices in each volume. An ellipse fitting and a stopping boundary-based energy is proposed to smooth the OWB contour and to stop leaking of the evolving contour, respectively. The algorithm was compared against ground truth boundaries generated from manual segmentations. The dice similarity coefficient (DSC), Hausdorff distance (HD), and modified HD (MHD) were used as error metrics. The authors' proposed stopping boundary energy-based stopping criterion was compared with percentage change of area and change of the MHD between evolving contours at successive iterations stopping criteria. The performance of the proposed algorithm was better than other two stopping criteria and yielded mean of LIBDSC = 88.78%, OWBDSC = 94.81%, LIBMHD = 0.26 mm, OWBMHD = 0.25 mm, LIBHD = 0.74 mm, and OWBHD = 0.80 mm. The Bland-Altman plot and correlation coefficient (r = 0.99) indicated a high agreement between ground truth and algorithm-generated boundaries. The coefficient of variation (COV) and minimum detectable change of the VWV are 5.2% and 57.2 mm(3) (5.18% of mean VWV), calculated from repeated measurements of the VWV by algorithm. The mean absolute distance between corresponding points of the algorithm-generated and the ground truth boundaries was 0.25 mm. The authors have developed a semiautomatic segmentation algorithm for measuring the VWV of the carotid artery using 3D US images with reduced operator interaction and computational time and higher reproducibility using a commercially available 3D US transducer. Their method is a step forward toward routine longitudinal monitoring of 3D plaque progression.

Interstitial Lung Disease (ILD) encompasses a wide array of diseases that share some common radiologic characteristics. When diagnosing such diseases, radiologists can be affected by heavy workload and fatigue thus decreasing diagnostic accuracy. Automatic segmentation is the first step in implementing a Computer Aided Diagnosis (CAD) that will help radiologists to improve diagnostic accuracy thereby reducing manual interpretation. Automatic segmentation proposed uses an initial thresholding and morphology based segmentation coupled with feedback that detects large deviations with a corrective segmentation. This feedback is analogous to a control system which allows detection of abnormal or severe lung disease and provides a feedback to an online segmentation improving the overall performance of the system. This feedback system encompasses a texture paradigm. In this study we studied 48 males and 48 female patients consisting of 15 normal and 81 abnormal patients. A senior radiologist chose the five levels needed for ILD diagnosis. The results of segmentation were displayed by showing the comparison of the automated and ground truth boundaries (courtesy of ImgTracer™ 1.0, AtheroPoint™ LLC, Roseville, CA, USA). The left lung's performance of segmentation was 96.52% for Jaccard Index and 98.21% for Dice Similarity, 0.61 mm for Polyline Distance Metric (PDM), -1.15% for Relative Area Error and 4.09% Area Overlap Error. The right lung's performance of segmentation was 97.24% for Jaccard Index, 98.58% for Dice Similarity, 0.61 mm for PDM, -0.03% for Relative Area Error and 3.53% for Area Overlap Error. The segmentation overall has an overall similarity of 98.4%. The segmentation proposed is an accurate and fully automated system.

A range of applications in visual science rely on accurate tracking of the human pupil’s movement and contraction in response to light. While the literature for independent contour detection and fitting of the iris-pupil boundary is vast, a joint approach, in which it is assumed that the pupil has a given geometric shape has been largely overlooked. We present here a global method for simultaneously finding and fitting of an elliptic or circular contour against a dark interior, which produces consistently accurate results even under non-ideal recording conditions, such as reflections near and over the boundary, droopy eye lids, or the sudden formation of tears. The specific form of the proposed optimization problem allows us to write down closed analytic formulae for the gradient and the Hessian of the objective function. Moreover, both the objective function and its derivatives can be cast into vectorized form, making the proposed algorithm significantly faster than its closest relative in the literature. We compare methods in multiple ways, both analytically and numerically, using real iris images as well as idealizations of the iris for which the ground truth boundary is precisely known. The method proposed here is illustrated under challenging recording conditions and it is shown to be robust.

Estimation of the breast skin-line in mammograms using multidirectional Gabor filters

We present a thorough quantitative evaluation of four image segmentation algorithms on images from the Berkeley Segmentation Database. The algorithms are evaluated using an efficient algorithm for computing precision and recall with regard to human ground-truth boundaries. We test each segmentation method over a representative set of input parameters, and present tuning curves that fully characterize algorithm performance over the complete image database. We complement the evaluation on the BSD with segmentation results on synthetic images. The results reported here provide a useful benchmark for current and future research efforts in image segmentation.

Accurate estimation of the breast skin-line is an important prerequisite for both enhancement and analysis of mammograms for computer-aided detection of breast cancer. In our proposed system, an initial estimate of the skin-line is first computed using a combination of adaptive thresholding and connected-component analysis. This skin-line is susceptible to errors in the top and bottom portions of the breast region. Using the observation that the Euclidean distance from the edge of the stroma to the actual skin-line is usually uniform, we develop a novel dependency approach for estimating the skin-line boundary of the breast. In the proposed dependency approach, the constraints are first developed between the stroma edge and the initial skin-line boundary using the Euclidean distance. These constraints are then propagated to estimate the upper and lower skin-line portions. We evaluated the performance of our skin-line estimation algorithm by comparing the estimated boundary with respect to the ground-truth boundary drawn by an expert radiologist. We adapted a metrics for error measurement: the polyline distance measure (PDM). As part of our protocol, we compared the results of our dependency approach methodology with those of a deformable model strategy (proposed by Ferrari et al. in Med Biol Eng Comput 42(2):210–208, 2004). On a dataset of 82 images from the MIAS database, the dependency approach yielded a mean error (μ) of 3.28 pixels with a standard deviation (σ) of 2.17 pixels using the PDM. In comparison, the deformable model strategy (Ferrari et al. in Med Biol Eng Comput 42(2):210–208, 2004) yielded μ = 4.92 pixels with σ = 1.91 pixels. The improvement is statistically significant. The results are clinically relevant, according to the radiologists who evaluated the results.