Late fusion of deep and shallow features to improve discrimination of actinic keratosis from normal skin using clinical photography.

BACKGROUND: Actinic keratosis (AK) is a common premalignant skin lesion that can potentially progress to squamous cell carcinoma. Appropriate long-term management of AK requires close patient monitoring in addition to therapeutic interventions. Computer-aided diagnostic systems based on clinical photography might evolve in the future into valuable adjuncts to AK patient management. The present study proposes a late fusion approach of color-texture features (shallow features) and deep features extracted from pre-trained convolutional neural networks (CNN) to boost AK detection accuracy on clinical photographs. MATERIALS AND METHODS: System uses a sliding rectangular window of 50 × 50 pixels and a classifier that assigns the window region to either the AK or the healthy skin class. 6010 and 13 915 cropped regions of interest (ROI) of 50 × 50 pixels of AK and healthy skin, respectively, from 22 patients were used for system implementation. Different support vector machine (SVM) classifiers employing shallow or deep features and their late fusion using the max rule at decision level were compared with the McNemar test and Yule's Q-statistic. RESULTS: Support vector machine classifiers based on deep and shallow features exhibited overall competitive performances with complementary improvements in detection accuracy. Late fusion yielded significant improvement (6%) in both sensitivity (87%) and specificity (86%) compared to single classifier performance. CONCLUSION: The parallel improvement of sensitivity and specificity is encouraging, demonstrating the potential use of our system in evaluating AK burden. The latter might be of value in future clinical studies for the comparison of field-directed treatment interventions.

Variation of cultured skin microbiota in mothers and their infants during the first year postpartum.

BACKGROUND/OBJECTIVES: The establishment of newborn skin flora depends on the ongoing skin maturation and the existence of potential microbial colonizers within the environment of the infant during a period of intense mother-infant physical interaction. This longitudinal study assessed culturable skin bacteria in the mother-infant dyad during the first year of life. METHODS: A total of 17 mother-infant dyads were swabbed within 24 hours postpartum and at 3, 6, 9, and 12 months. Skin swabbing was performed on two anatomical areas per individual (mothers: chest-abdomen; infants: forehead-buttocks) and were incubated in five different solid culture media to optimize yield. Isolated bacterial species were identified to genus or species level using the API system (BioMeriéux, Marcy l'Etoile, France). RESULTS: A total of 444 microbial strains were isolated belonging to 22 genera: 6 "frequent" (isolated from > 5% samples: S aureus, Proteus, Klebsiella, Pseudomonas, Enterobacter, and Enterococcus) and 16 "infrequent." Isolated genera per individual peaked at 6 months postpartum for mothers and infants (P < 0.05). Enterobacter, Enterococcus, Klebsiella, and Pseudomonas isolation rates varied significantly as a function of sampling time contrary to the rather constant isolation rates of Proteus and S aureus. The rates of concordant isolation of the same microbial species within the mother-infant dyad tended to drop from birth to the end of the first year postpartum. CONCLUSIONS: Distinct variations in the isolation rates of skin commensals from specific anatomical sites of the mother-infant dyad indicate bidirectional microbial transmission. Increasing skin flora individuality of the growing infant was recorded, manifested by declining rates of concordant isolation of the same microbial species from mother and her infant.

Gray Hausdorff distance measure for medical image comparison in dermatology: Evaluation of treatment effectiveness by image similarity.

INTRODUCTION: In clinical dermatology, the stabilization of the overall skin condition can be in many cases the earliest qualitative measure of the effectiveness of the therapeutic intervention. Subjective image comparisons, that offer empirical 'qualitative' judgments of degrees of image similarities, are traditionally employed by the involved physicians. OBJECTIVES: To quantify, by means of an image similarity metric, the degree of stabilization of an expanding skin disease, and to identify the situation of 'no further change' of the skin condition of the patient, providing thus the physician with an early, objective measure of the efficacy of the used therapy. METHODS: For treatment assessment, a variant of gray Hausdorff distance metric was employed to compare images of lesional skin segments of a patient, taken at different time points during a therapeutic course. Prior to image comparison, an effective preprocessing scheme was adapted to constrain wide pose and light variations. The proposed similarity algorithm was tested on raw clinical image data sets of patients diagnosed with toxic epidermal necrolysis, a life-threatening condition with rapid evolution. Fine tuning of algorithm's parameters was optimized using Precision-Recall curves. RESULTS: Proposed image comparison method resulted in a high-degree of image similarity (about 96%) between pictures taken at second and fifth day of hospitalization. Current similarity results substantiate a significant agreement between the computer-treatment assessment, by means of image comparison, and the corresponding clinical experts' review of skin condition. CONCLUSION: Objective evidence of 'no further change' situation may provide (a) intuitive clinical decision support to dermatologists in assessing aggressive skin conditions, where the timely evaluation of treatment response is of vital importance and (b) a versatile end-point measure for corresponding therapeutic clinical trials.

The Promise of Semantic Segmentation in Detecting Actinic Keratosis Using Clinical Photography in the Wild.

AK is a common precancerous skin condition that requires effective detection and treatment monitoring. To improve the monitoring of the AK burden in clinical settings with enhanced automation and precision, the present study evaluates the application of semantic segmentation based on the U-Net architecture (i.e., AKU-Net). AKU-Net employs transfer learning to compensate for the relatively small dataset of annotated images and integrates a recurrent process based on convLSTM to exploit contextual information and address the challenges related to the low contrast and ambiguous boundaries of AK-affected skin regions. We used an annotated dataset of 569 clinical photographs from 115 patients with actinic keratosis to train and evaluate the model. From each photograph, patches of 512 × 512 pixels were extracted using translation lesion boxes that encompassed lesions in different positions and captured different contexts of perilesional skin. In total, 16,488 translation-augmented crops were used for training the model, and 403 lesion center crops were used for testing. To demonstrate the improvements in AK detection, AKU-Net was compared with plain U-Net and U-Net++ architectures. The experimental results highlighted the effectiveness of AKU-Net, improving upon both automation and precision over existing approaches, paving the way for more effective and reliable evaluation of actinic keratosis in clinical settings.

Image Perceptual Similarity Metrics for the Assessment of Basal Cell Carcinoma.

Efficient management of basal cell carcinomas (BCC) requires reliable assessments of both tumors and post-treatment scars. We aimed to estimate image similarity metrics that account for BCC's perceptual color and texture deviation from perilesional skin. In total, 176 clinical photographs of BCC were assessed by six physicians using a visual deviation scale. Internal consistency and inter-rater agreement were estimated using Cronbach's α, weighted Gwet's AC2, and quadratic Cohen's kappa. The mean visual scores were used to validate a range of similarity metrics employing different color spaces, distances, and image embeddings from a pre-trained VGG16 neural network. The calculated similarities were transformed into discrete values using ordinal logistic regression models. The Bray-Curtis distance in the YIQ color model and rectified embeddings from the 'fc6' layer minimized the mean squared error and demonstrated strong performance in representing perceptual similarities. Box plot analysis and the Wilcoxon rank-sum test were used to visualize and compare the levels of agreement, conducted on a random validation round between the two groups: 'Human-System' and 'Human-Human.' The proposed metrics were comparable in terms of internal consistency and agreement with human raters. The findings suggest that the proposed metrics offer a robust and cost-effective approach to monitoring BCC treatment outcomes in clinical settings.

Treatment of toxic epidermal necrolysis with the combination of infliximab and high-dose intravenous immunoglobulin.

BACKGROUND: Therapeutic evidence for toxic epidermal necrolysis (TEN) is indicative for high-dose intravenous immunoglobulin yet inconclusive for corticosteroids. OBJECTIVE: To describe the combination of corticosteroids, infliximab and a high-dose intravenous immunoglobulin course for TEN. PATIENTS AND METHODS: In three patients (SCORTEN survival probabilities: 41.7%, 64.2%, 41.7%) disease control was evaluated by (a) employing quantitative image analysis to measure progression of skin detachment and (b) patients' outcome (complete re-epithelization). Published cases of TEN treatments with infliximab were retrieved from PubMed. RESULTS: Within 48 h skin disease progression was arrested in all patients. Two patients were discharged after 3 weeks without any sequels from skin or conjunctivae. One patient passed away on the ninth day, however with noticeably improved skin (mortality rate: 33% observed vs. 50% expected). A PubMed search retrieved five TEN patients treated successfully with infliximab. CONCLUSION: The described combination presents a feasible therapeutic alternative for TEN that warrants further evaluation.

Application of a burn scar assessment tool for the evaluation of visual scarring following immunocryosurgery for facial basal cell carcinoma

Background: The therapeutic interventions for basal cell carcinoma (BCC) impact facial cosmesis. Objectives: Our aim was to assess the aesthetic burden of facial BCC treatment by evaluating the extent of the tumour site visually based on clinical images before and after immunocryosurgery; a minimally invasive combination treatment of topical imiquimod and cryosurgery. Materials & Methods: A three-item (texture, height, colour) burn scar scale was independently applied by four physicians (two dermatologists and two plastic surgeons) on archival semi-standardized clinical images of facial BCC before and one year after immunocryosurgery. The score assessments were compared using non-parametric statistical tests; internal consistency (reliability) and inter-rater agreement were assessed using Cronbach's α and Gwet's AC2, respectively. Results: Images (before and one year after treatment) of 27 BCCs from 26 patients (15 males) were analysed. The reliability was good (α>0.80) for all items before surgery and for colour after surgery. The inter-rater agreement was acceptable (AC2>0.70) for all items except for height pre-treatment (AC2 = 0.482). Based on averaging of all raters, cosmesis improved significantly after treatment at all tumour sites (p<0.001, Wilcoxon signed ranks test). The main limitations of the study are the inclusion of cases from a single centre and its retrospective nature. Conclusion: The evaluated burn scar scale can be reliably used to compare the extent of the BCC site visually before therapy and after resultant scars have formed. In this way, the impact of therapeutic interventions on cosmesis (from tumour to scar) can be evaluated, particularly for less aggressive facial skin tumours, such as most BCC.

Methylprednisolone Plus Low-Dose Methotrexate for Bullous Pemphigoid-A Single Center Retrospective Analysis.

Monomodal systemic glucocorticoids remain the mainstay of treatment for bullous pemphigoid (BP). In this retrospective, single-arm study, we evaluated the feasibility (efficacy and tolerability) of the combination of methylprednisolone and low-dose (up to 12.5 mg/week) methotrexate (MP + MTX) for BP. At week 12, 53/55 (96.4%) patients initiated on MP + MTX during a five-year period (potential follow up time: ≥4 years) remained on treatment. At this time-point, BP remission was achieved in all compliant patients (including n = 24 cases of dipeptidyl peptidase-4 inhibitors-associated BP; 12-week remission rate: 100% [95% CI: 91.9-100.0%]; mean time to remission: 29.5 days, SEM: 2.3 days) at a mean cumulative MP dose to disease control of 678.4 mg (SEM = 49.4 mg). Eight patients relapsed during follow up (10.81 [95% CI: 5.16-21.72] relapses/100 person years, py), and seven manifested a severe adverse event (6.80 [95% CI: 3.00-14.28] severe adverse events/100 py); however, 73.4% (±7.9%) had suffered neither a relapse nor a SAE at the three-years follow up. Continuing low dose MP intake (≤8 mg/day) beyond week 12 in combination with MTX minimized the risk of a feasibility limiting event (p = 0.013). Conclusively, the combination of methylprednisolone with methotrexate is a promising, safe, and efficient modality for BP patients, which enables rapid glucocorticoid tapering.

Characterizing Malignant Melanoma Clinically Resembling Seborrheic Keratosis Using Deep Knowledge Transfer.

Malignant melanomas resembling seborrheic keratosis (SK-like MMs) are atypical, challenging to diagnose melanoma cases that carry the risk of delayed diagnosis and inadequate treatment. On the other hand, SK may mimic melanoma, producing a 'false positive' with unnecessary lesion excisions. The present study proposes a computer-based approach using dermoscopy images for the characterization of SΚ-like MMs. Dermoscopic images were retrieved from the International Skin Imaging Collaboration archive. Exploiting image embeddings from pretrained convolutional network VGG16, we trained a support vector machine (SVM) classification model on a data set of 667 images. SVM optimal hyperparameter selection was carried out using the Bayesian optimization method. The classifier was tested on an independent data set of 311 images with atypical appearance: MMs had an absence of pigmented network and had an existence of milia-like cysts. SK lacked milia-like cysts and had a pigmented network. Atypical MMs were characterized with a sensitivity and specificity of 78.6% and 84.5%, respectively. The advent of deep learning in image recognition has attracted the interest of computer science towards improved skin lesion diagnosis. Open-source, public access archives of skin images empower further the implementation and validation of computer-based systems that might contribute significantly to complex clinical diagnostic problems such as the characterization of SK-like MMs.

Advanced keratinocyte skin cancer is a tumor with considerable disease burden and aggressiveness.

In a 2013 study published in this Journal, Dacosta Byfield et al. used MediCare data to extract reliable estimations of the incidence (I = 6.16) and prevalence (P = 10.31) rates of advanced keratinocyte skin cancer (aKSC) per 100,000 US population. These data unmask a considerable disease burden of aKSC (tumor stages ≥ 3) compared to the corresponding projected SEER predictions in 2019 of all invasive cases (tumor stages ≥ 1). According to its incidence, aKSC ranks 19th out of 29 major SEER registered neoplasms and has an average disease duration of 1.67 years, which is the second shortest disease duration next only to pancreatic carcinoma. Furthermore, in support of the high disease aggressiveness of aKSC and using a calibration approach, we calculated a mortality estimate of 4.64 per 100,000 and a 5-year survival rate of 21.8% for this tumor, which corresponds to positions of 13th and 5th out of 29 cancers among the SEER tracked malignancies, respectively. Taken together, these data indicate a considerable disease burden and biologic aggressiveness of aKSC.

A Guttate Psoriasis That Tends to Spare Three Tattoos: A Macrophage Liaison.

Induction of new psoriasis sites was reported in only a small amount of psoriasis patients undergoing tattooing, despite the intuitive belief that tattoo trauma might awaken the disease due to the isomorphic phenomenon of Koebner. In this case report, we discuss a patient who presented with a remarkable sparing of his three tattoo sites during a guttate psoriasis flare-up that was unrelated to tattooing. The spatial concordance of tattoo and psoriasis lesions was analyzed on clinical pictures of tattoo sites taken during the psoriasis episode. For the quantification of the spatial distribution of the psoriasis lesions, Voronoi diagrams were generated, and coefficients of variation and the two-sample t-test were employed to compare the distributions of Voronoi patch sizes in different settings. Compared to skin areas without tattoos, a tattoo introduced a higher variation in the sizes of the Voronoi patches centered around psoriasis lesions. Based on our findings, we would like to discuss the possible role of macrophages as the key cellular link in the complex pathophysiologic relationship between tattooing/tattoo and psoriasis. Taking into account the relationship of autophagy and psoriasis lesions, we propose the hypothesis that tattoos represent a "psoriasis-hostile" tissue environment pertained by a population of LAP active M2-polarized macrophages. Further clinical studies of the relationship of psoriasis lesions to the tattooed skin are needed and may provide important insights into the role of macrophages in the pathogenesis of psoriasis.

Improving accuracy in astrocytomas grading by integrating a robust least squares mapping driven support vector machine classifier into a two level grade classification scheme.

Grading of astrocytomas is an important task for treatment planning; however, it suffers from significantly great inter-observer variability. Computer-assisted diagnosis systems have been propose to assist towards minimizing subjectivity, however, these systems present either moderate accuracy or utilize specialized staining protocols and grading systems that are difficult to apply in daily clinical practice. The present study proposes a robust mathematical formulation by integrating state-of-art technologies (support vector machines and least squares mapping) in a cascade classification scheme for separating low from high and grade III from grade IV astrocytic tumours. Results have indicated that low from high-grade tumours can be correctly separated with a certainty as high as 97.3%, whereas grade III from grade IV tumours with 97.8%. The overall performance was 95.2%. These high rates have been a result of applying the least squares mapping technique to features prior to classification. A significant byproduct of least squares mapping is that the number of support vectors of the SVM classifiers dropped dramatically from about 80% when no mapping was used to less than 5% when mapping was used. The latter is a clear indication that the SVM classifier has a greater potential to generalize well to new data. In this way, digital image analysis systems for automated grading of astrocytomas are brought closer to clinical practice.

Design of a multi-classifier system for discriminating benign from malignant thyroid nodules using routinely H&E-stained cytological images.

A multi-classifier diagnostic system was designed for distinguishing between benign and malignant thyroid nodules from routinely taken (FNA, H&E-stained) cytological images. To construct the multi-classifier system, several combination rules and different mixtures of ensemble classifier members, employing morphological and textural nuclear features, were comparatively evaluated. Experimental results illustrated that the classifier combination k-NN/PNN/Bayesian and the majority vote rule enhanced significantly classification accuracy (95.7%) as compared to best single classifier (PNN: 89.6%). The proposed system was designed with purpose to be utilized in daily clinical practice as a second opinion tool to support cytopathologists' decisions, when a definite diagnosis is difficult to be obtained.

Computerized analysis of digital subtraction angiography: a tool for quantitative in-vivo vascular imaging.

The purpose of our study was to develop a user-independent computerized tool for the automated segmentation and quantitative assessment of in vivo-acquired digital subtraction angiography (DSA) images. Vessel enhancement was accomplished based on the concept of image structural tensor. The developed software was tested on a series of DSA images acquired from one animal and two human angiogenesis models. Its performance was evaluated against manually segmented images. A receiver's operating characteristic curve was obtained for every image with regard to the different percentages of the image histogram. The area under the mean curve was 0.89 for the experimental angiogenesis model and 0.76 and 0.86 for the two clinical angiogenesis models. The coordinates of the operating point were 8.3% false positive rate and 92.8% true positive rate for the experimental model. Correspondingly for clinical angiogenesis models, the coordinates were 8.6% false positive rate and 89.2% true positive rate and 9.8% false positive rate and 93.8% true positive rate, respectively. A new user-friendly tool for the analysis of vascular networks in DSA images was developed that can be easily used in either experimental or clinical studies. Its main characteristics are robustness and fast and automatic execution.

A Case Series of Bowen's Disease Treated with the Combination of Cryosurgery and Ingenol Mebutate and Followed Up with Optical Coherence Tomography.

Bowen's disease (BD) is a relatively rare in situ squamous cell carcinoma (SCC) with a limited potential of becoming invasive. Ingenol mebutate (IM) was relatively successful for the treatment of BD lesions in small case series. Optical coherence tomography (OCT) is a promising method for the diagnosis of cutaneous keratinocytic carcinomas, including BD. Herein we report the treatment of BD with the combination of cryosurgery and IM and the application of OCT imagining in treatment monitoring. Patients treated within a period of 12 months are retrospectively compiled. Treatment consisted of a mild cryosurgery session (liquid N2, open spray, and 2 freeze-thaw cycles of 15 sec each) of a field including the BD lesion and a 0.5cm rim and IM application for 4 consecutive days starting at the cryosurgery day. Four patients (3 females; average age: 76.5 years) with 4 lesions (20-70mm maximal diameter; average 36.2mm) were included. Healing was excellent and no relapse was observed at 12 months' follow-up. Baseline OCT revealed a disarranged, thickened epidermis, while a normally layered epidermis overlying a hyperreflective dermis was present after treatment. Conclusively, the combination of cryosurgery followed by IM is a feasible, effective treatment for BD that should be evaluated in further studies.

Brain activation during repeated imagining of chocolate consumption: a functional magnetic resonance imaging study.

OBJECTIVE: To assess brain activation during mental visualization of eating chocolate. DESIGN: Twenty-one subjects were included. FMRI was acquired with a single-shot, multislice, gradient echo-planar sequence, while subjects were performing two specific imaginary tasks. RESULTS: Activation of motor-associated brain areas was observed during both mental visualization tasks. Increased activation of the right dorsolateral prefrontal cortex, the thalamus, the postcentral gyrus and the left anterior cingulate cortex, and the precuneus was observed during imagining eating chocolate. CONCLUSIONS: Repeated imagination of chocolate consumption results in activation of brain areas associated with hedonic effects of food and satiety and inhibition of orexigenic areas.