Automated Prediction of Kidney Failure in IgA Nephropathy with Deep Learning from Biopsy Images.

BACKGROUND AND OBJECTIVES: Digital pathology and artificial intelligence offer new opportunities for automatic histologic scoring. We applied a deep learning approach to IgA nephropathy biopsy images to develop an automatic histologic prognostic score, assessed against ground truth (kidney failure) among patients with IgA nephropathy who were treated over 39 years. We assessed noninferiority in comparison with the histologic component of currently validated predictive tools. We correlated additional histologic features with our deep learning predictive score to identify potential additional predictive features. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: Training for deep learning was performed with randomly selected, digitalized, cortical Periodic acid-Schiff-stained sections images (363 kidney biopsy specimens) to develop our deep learning predictive score. We estimated noninferiority using the area under the receiver operating characteristic curve (AUC) in a randomly selected group (95 biopsy specimens) against the gold standard Oxford classification (MEST-C) scores used by the International IgA Nephropathy Prediction Tool and the clinical decision supporting system for estimating the risk of kidney failure in IgA nephropathy. We assessed additional potential predictive histologic features against a subset (20 kidney biopsy specimens) with the strongest and weakest deep learning predictive scores. RESULTS: We enrolled 442 patients; the 10-year kidney survival was 78%, and the study median follow-up was 6.7 years. Manual MEST-C showed no prognostic relationship for the endocapillary parameter only. The deep learning predictive score was not inferior to MEST-C applied using the International IgA Nephropathy Prediction Tool and the clinical decision supporting system (AUC of 0.84 versus 0.77 and 0.74, respectively) and confirmed a good correlation with the tubolointerstitial score (r=0.41, P<0.01). We observed no correlations between the deep learning prognostic score and the mesangial, endocapillary, segmental sclerosis, and crescent parameters. Additional potential predictive histopathologic features incorporated by the deep learning predictive score included (1) inflammation within areas of interstitial fibrosis and tubular atrophy and (2) hyaline casts. CONCLUSIONS: The deep learning approach was noninferior to manual histopathologic reporting and considered prognostic features not currently included in MEST-C assessment. PODCAST: This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2022_07_26_CJN01760222.mp3.

One DAG to Rule Them All.

In this paper, we present novel strategies for optimizing the performance of many binary image processing algorithms. These strategies are collected in an open-source framework, GRAPHGEN, that is able to automatically generate optimized C++ source code implementing the desired optimizations. Simply starting from a set of rules, the algorithms introduced with the GRAPHGEN framework can generate decision trees with minimum average path-length, possibly considering image pattern frequencies, apply state prediction and code compression by the use of Directed Rooted Acyclic Graphs (DRAGs). Moreover, the proposed algorithmic solutions allow to combine different optimization techniques and significantly improve performance. Our proposal is showcased on three classical and widely employed algorithms (namely Connected Components Labeling, Thinning, and Contour Tracing). When compared to existing approaches -in 2D and 3D-, implementations using the generated optimal DRAGs perform significantly better than previous state-of-the-art algorithms, both on CPU and GPU.

Evaluation of the Classification Accuracy of the Kidney Biopsy Direct Immunofluorescence through Convolutional Neural Networks.

BACKGROUND AND OBJECTIVES: Immunohistopathology is an essential technique in the diagnostic workflow of a kidney biopsy. Deep learning is an effective tool in the elaboration of medical imaging. We wanted to evaluate the role of a convolutional neural network as a support tool for kidney immunofluorescence reporting. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: High-magnification (×400) immunofluorescence images of kidney biopsies performed from the year 2001 to 2018 were collected. The report, adopted at the Division of Nephrology of the AOU Policlinico di Modena, describes the specimen in terms of "appearance," "distribution," "location," and "intensity" of the glomerular deposits identified with fluorescent antibodies against IgG, IgA, IgM, C1q and C3 complement fractions, fibrinogen, and κ- and λ-light chains. The report was used as ground truth for the training of the convolutional neural networks. RESULTS: In total, 12,259 immunofluorescence images of 2542 subjects undergoing kidney biopsy were collected. The test set analysis showed accuracy values between 0.79 ("irregular capillary wall" feature) and 0.94 ("fine granular" feature). The agreement test of the results obtained by the convolutional neural networks with respect to the ground truth showed similar values to three pathologists of our center. Convolutional neural networks were 117 times faster than human evaluators in analyzing 180 test images. A web platform, where it is possible to upload digitized images of immunofluorescence specimens, is available to evaluate the potential of our approach. CONCLUSIONS: The data showed that the accuracy of convolutional neural networks is comparable with that of pathologists experienced in the field.

Spaghetti Labeling: Directed Acyclic Graphs for Block-Based Connected Components Labeling.

Connected Components Labeling is an essential step of many Image Processing and Computer Vision tasks. Since the first proposal of a labeling algorithm, which dates back to the sixties, many approaches have optimized the computational load needed to label an image. In particular, the use of decision forests and state prediction have recently appeared as valuable strategies to improve performance. However, due to the overhead of the manual construction of prediction states and the size of the resulting machine code, the application of these strategies has been restricted to small masks, thus ignoring the benefit of using a block-based approach. In this paper, we combine a block-based mask with state prediction and code compression: the resulting algorithm is modeled as a Directed Rooted Acyclic Graph with multiple entry points, which is automatically generated without manual intervention. When tested on synthetic and real datasets, in comparison with optimized implementations of state-of-the-art algorithms, the proposed approach shows superior performance, surpassing the results obtained by all compared approaches in all settings.

Dynamic Optical Coherence Tomography in Dermatology.

Optical coherence tomography (OCT) represents a non-invasive imaging technology, which may be applied to the diagnosis of non-melanoma skin cancer and which has recently been shown to improve the diagnostic accuracy of basal cell carcinoma. Technical developments of OCT continue to expand the applicability of OCT for different neoplastic and inflammatory skin diseases. Of these, dynamic OCT (D-OCT) based on speckle variance OCT is of special interest as it allows the in vivo evaluation of blood vessels and their distribution within specific lesions, providing additional functional information and consequently greater density of data. In an effort to assess the potential of D-OCT for future scientific and clinical studies, we have therefore reviewed the literature and preliminary unpublished data on the visualization of the microvasculature using D-OCT. Information on D-OCT in skin cancers including melanoma, as well as in a variety of other skin diseases, is presented in an atlas. Possible diagnostic features are suggested, although these require additional validation.

Optimized block-based connected components labeling with decision trees.

In this paper, we define a new paradigm for eight-connection labeling, which employs a general approach to improve neighborhood exploration and minimizes the number of memory accesses. First, we exploit and extend the decision table formalism introducing OR-decision tables, in which multiple alternative actions are managed. An automatic procedure to synthesize the optimal decision tree from the decision table is used, providing the most effective conditions evaluation order. Second, we propose a new scanning technique that moves on a 2 x 2 pixel grid over the image, which is optimized by the automatically generated decision tree. An extensive comparison with the state of art approaches is proposed, both on synthetic and real datasets. The synthetic dataset is composed of different sizes and densities random images, while the real datasets are an artistic image analysis dataset, a document analysis dataset for text detection and recognition, and finally a standard resolution dataset for picture segmentation tasks. The algorithm provides an impressive speedup over the state of the art algorithms.

Colour clusters for computer diagnosis of melanocytic lesions.

BACKGROUND: To overcome subjectivity and variability in the interpretation of dermoscopic images, image analysis programs, enabling the numerical description of melanocytic lesion images, have been developed. OBJECTIVES: Our aim was to assess a method for the description of colours in melanocytic lesion images, based on the subdivision of image colours into red, green and blue clusters. METHODS: Melanomas and naevi of the test set were described by means of 23 colour clusters previously selected by a training set comprising 369 melanocytic lesion images. The diagnostic performance obtained by this automated method was compared to sensitivity and specificity of diagnosis of 4 dermatologists. RESULTS: Colour cluster values significantly differed between melanomas and naevi. Moreover, sensitivity and specificity values of computer diagnosis were similar to those achieved by the dermatologists. CONCLUSION: Our image analysis program based on the assessment of one single parameter has the diagnostic accuracy of dermatologists employing dermoscopy on a regular basis.

Asymmetry in dermoscopic melanocytic lesion images: a computer description based on colour distribution.

Digital dermoscopy improves the accuracy of melanoma diagnosis. The aim of this study was to develop and validate software for assessment of asymmetry in melanocytic lesion images, based on evaluation of colour symmetry, and to compare it with assessment by human observers. An image analysis program enabling numerical assessment of asymmetry in melanocytic lesions, based on the evaluation and comparison of CIE L*a*b* colour components (CIE L*a*b* is the name of a colour space defined by the Commission Internationale de l'Eclairage) inside image colour blocks, was employed on the recorded lesion images. Clinical evaluation of asymmetry in dermoscopic images was performed on the same image set employing a 0-1 scoring system. Asymmetry judgement was expressed by the clinicians for 12.8% of benign naevi, 44.7% of atypical naevi and 64.2% of malignant melanomas, whereas the computer identified as asymmetric 6.3%, 33.3% and 82.2%, respectively. Numerical parameters referring to malignant melanomas were significantly higher, both with respect to benign naevi and atypical naevi. The numerical parameters produced could be effectively employed for computer-aided melanoma diagnosis.

Microscopic in vivo description of cellular architecture of dermoscopic pigment network in nevi and melanomas.

OBJECTIVE: To characterize the microscopic aspects of the dermoscopic pigment network in vivo, by means of confocal scanning laser microscopy. DESIGN: Confocal imaging was performed on melanocytic lesions characterized by pigment network at dermoscopy. Some confocal architectural and cytologic features, as observed at the dermoepidermal junction, were morphologically described and quantified by means of a dedicated program. SETTING: University medical department. STUDY POPULATION: We studied confocal images of 15 melanomas, 15 dermoscopic atypical nevi, and 15 common nevi. MAIN OUTCOME MEASURES: Features referring to aspect, size, regularity, homogeneity, and infiltration of dermal papillae and to cellular size, regularity, and atypia were described by 2 observers on confocal images. Mean dermal papillary diameter, mean cell area, and shape irregularity were quantified by drawing papillae and cell contours on confocal images and measured with the use of a computer program. RESULTS: Pigment network in melanomas consisted of large basal cells that circumscribed small to medium-sized dermal papillae with marked cellular atypia, sometimes infiltrating dermal papillae. On the other hand, common acquired nevi were characterized by lack of atypical cells and edged dermal papillae. Atypical nevi presented intermediate characteristics between clearly benign and malignant lesions. CONCLUSION: Cellular atypia was the most sensitive feature for melanoma diagnosis, whereas the presence of nucleated cells infiltrating dermal papillae was the most specific one.

Colors in atypical nevi: a computer description reproducing clinical assessment.

BACKGROUND/PURPOSE: Atypical nevi (AN) share some dermoscopic features with early melanoma (MM), and computer elaboration of digital images could represent a useful support to diagnosis to assess automatically colors in AN, and to compare the data with those referring to clearly benign nevi (BN) and MMs. METHODS: An image analysis program enabling the numerical description of color areas in melanocytic lesions was used on 459 videomicroscopic images, referring to 76 AN, 288 clearly BN and 95 MMs. RESULTS: Black, white and blue-gray were more frequently found in AN than in clearly BN, but less frequently than in MMs. Color area values significantly differed between the three groups. CONCLUSION: The clinical-morphological interpretation of the numerical data, based on the mathematical description of the aspect and distribution of different color areas in different lesion types may contribute to the characterization of AN and their distinction from MMs.

In vivo confocal scanning laser microscopy of pigmented Spitz nevi: comparison of in vivo confocal images with dermoscopy and routine histopathology.

BACKGROUND: Spitz nevus is a benign melanocytic lesion sometimes mistakenly diagnosed clinically as melanoma. OBJECTIVE: Our aim was to evaluate in vivo reflectance-mode confocal scanning laser microscopy (CSLM) aspects of globular Spitz nevi and to correlate them with those of surface microscopy and histopathology. METHODS: A total of 6 Spitz nevi, with globular aspects on epiluminescence observation, were imaged with CSLM and subsequently excised for histopathologic examination. RESULTS: A close correlation among CSLM, epiluminescence, and histopathologic aspects was observed. Individual cells, observed in high-resolution confocal images, were similar in shape and dimension to the histopathologic ones. Lesion architecture was described on reconstructed CSLM images. Melanocytic nests corresponded to globular cellular aggregates at confocal microscopy and to globules at epiluminescence observation. Melanophages were clearly identified in the papillary dermis both by confocal microscopy and histopathology. CONCLUSION: In vivo CSLM enabled the identification of characteristic cytologic and architectural aspects of Spitz nevi, correlated with histopathology and epiluminescence microscopy observation.

Automated description of colours in polarized-light surface microscopy images of melanocytic lesions.

The aim of this study was to develop a computerized method for the identification and description of colour areas in melanocytic lesion images based on an approach mimicking the human perception of colours. A colour palette comprising six colour groups (black, dark brown, light brown, blue-grey, red and white) was created by selecting single colour components within melanocytic lesion images acquired using a digital videomicroscope, and was implemented in the image analysis program. For each colour region, the area, the distance from the lesion centroid, the spread, the colour area distribution in the internal and the external part of the lesion, and asymmetries were assessed on 604 melanocytic lesion images in our image database. Black, white and blue-grey colour areas were detected more frequently in melanomas compared with naevi. Moreover, significant differences in colour descriptors were observed for each colour group, showing that colour areas are more unevenly distributed in melanomas compared with naevi. Using a discriminant analysis approach, the extension of dark, white and blue-grey areas and some descriptors of the distribution of the colour areas were identified as the most relevant colour parameters for differentiating between benign and malignant lesions. In conclusion, our automatic procedure breaks down the image into the colour areas used in the clinical examination process, and also supplies a description of their extension and distribution, with parameters that correlate with the clinical concepts of regularity and homogeneity.

Automated extraction and description of dark areas in surface microscopy melanocytic lesion images.

BACKGROUND: Identification of dark areas inside a melanocytic lesion (ML) is of great importance for melanoma diagnosis, both during clinical examination and employing programs for automated image analysis. OBJECTIVE: The aim of our study was to compare two different methods for the automated identification and description of dark areas in epiluminescence microscopy images of MLs and to evaluate their diagnostic capability. METHODS: Two methods for the automated extraction of 'absolute' (ADAs) and 'relative' dark areas (RDAs) and a set of parameters for their description were developed and tested on 339 images of MLs acquired by means of a polarized-light videomicroscope. RESULTS: Significant differences in dark area distribution between melanomas and nevi were observed employing both methods, permitting a good discrimination of MLs (diagnostic accuracy = 74.6 and 71.2% for ADAs and RDAs, respectively). CONCLUSIONS: Both methods for the automated identification of dark areas are useful for melanoma diagnosis and can be implemented in programs for image analysis.

A new algorithm for border description of polarized light surface microscopic images of pigmented skin lesions.

The aim of this study was to provide mathematical descriptors for the border of pigmented skin lesion images and to assess their efficacy for distinction among different lesion groups. New descriptors such as lesion slope and lesion slope regularity are introduced and mathematically defined. A new algorithm based on the Catmull-Rom spline method and the computation of the gray-level gradient of points extracted by interpolation of normal direction on spline points was employed. The efficacy of these new descriptors was tested on a data set of 510 pigmented skin lesions, composed by 85 melanomas and 425 nevi, by employing statistical methods for discrimination between the two populations.