The Effect of Polynucleotide-Hyaluronic Acid Hydrogel in the Recovery After Mechanical Skin Barrier Disruption.

BACKGROUND: The epidermal barrier acts as a defense against external agents as well as helps to maintain body homeostasis. Polynucleotides (PN), exogenous DNA fragments, promote wound repair through their stimulatory and anti-inflammatory effects. Recent findings indicate a synergistic effect of PN and hyaluronic acid (HA) combinations in regulating inflammation and promoting cell proliferation. This study aims to elucidate the effects of PN and HA on repairing the epidermal barrier following its disruption by tape stripping (TS) in a mouse model. MATERIALS AND METHODS: After disrupting the epidermal barrier using TS, a formulation containing PN (14 mg/mL) and HA (6 mg/mL) was applied. Trans-epidermal water loss (TEWL) was measured at 0, 3, 6, 24, 48, and 72 h. Mice were euthanized after the final application at 72 h, and tissue samples were analyzed for epidermal/dermal thickness, neutrophil infiltration, and filaggrin expression. RESULTS: We observed a significant reduction in TEWL in the PN+HA group compared to that in the control group (20.8 ± 0.5 vs. 43.7 ± 0.5 g/m2h at 72 h, p < 0.05), indicating an improvement in barrier function. Histological evaluation showed decreased epidermal and dermal thickening in the PN+HA group compared to that in the control group (epidermal: 29.4 ± 2.2 vs. 57.9 ± 3.5 µm; dermal: 464.8 ± 25.9 vs. 825.9 ± 44.8 µm, both p < 0.05). Additionally, neutrophil infiltration in the dermis was significantly reduced, and filaggrin protein levels were significantly higher in the PN+HA group compared to those in the control group (4.8 ± 0.4 vs. 21.1 ± 3.3 for neutrophils; 0.84 ± 0.04 vs. 0.42 ± 0.03 for filaggrin, both p < 0.05). CONCLUSION: These results suggest that PN+HA may be an effective therapeutic strategy for repairing skin barrier damage.

Prevalence, clinical features, and survival outcome trends of 627 patients with primary cutaneous lymphoma over 29 years: a retrospective review from single tertiary center in Korea.

The relative frequency of primary cutaneous lymphoma (PCL) subtypes shows wide variation across different geographical regions. This retrospective study was conducted in a tertiary referral center located in Korea to describe the relative frequency, demographics, survival outcomes, and temporal trend in PCL. A total of 627 PCL cases diagnosed between January 1994 and December 2022 were included. The majority of PCL cases (87.2%) were of T-/NK-cell lineage (CTCL), while the remaining cases (12.8%) were B-cell lineage lymphomas (CBCL). The prevalence of mycosis fungoides (MF) in CTCL increased significantly over time, while other CTCL subtypes, including primary cutaneous extranodal NK/T-cell lymphoma and subcutaneous panniculitis-like T-cell lymphoma (SPTCL), decreased in frequency. Notably, the prevalence of CD4-positive small/medium T-cell lymphoproliferative disorder showed a substantial increase over time. Primary cutaneous marginal zone lymphoma was consistently the commonest CBCL subtype. Survival analysis demonstrated that CTCL had a more favorable 5-year overall survival (OS) than CBCL. OS rate of MF, SPTCL, and primary cutaneous peripheral T-cell lymphoma, NOS improved significantly over time. This study provides comprehensive insights into the dynamic change in the relative frequency and overall survival of PCL subtypes over time.

Risk of major adverse cardiovascular events and all-cause mortality among patients with psoriatic disease treated with tumor necrosis factor-α and interleukin-12/23 inhibitors: a nationwide population-based cohort study in Korea.

PURPOSE: Few studies have investigated the impact of biologics on the risk of major adverse cardiovascular events (MACEs) among Korean patients with psoriatic diseases. We compared the risk of MACEs and all-cause mortality among patients with psoriatic disease treated with tumor necrosis factor (TNF)-α and interleukin (IL)-12/23 inhibitors in Korea. METHODS: Patients with psoriatic disease prescribed with TNF-α and IL-12/23 inhibitors since 2016 were selected from the Korean National Health Insurance Service (NHIS) Database. Follow-up data for MACEs and all-cause mortality between 2016 and 2020 were collected. A total of 2886 individuals were included, including 1987 IL-12/23 inhibitor users and 899 TNF-α inhibitor users. RESULTS: Compared with IL-12/23 inhibitor users, TNF-α inhibitor users had a higher prevalence of dyslipidemia and a significantly higher risk of all-cause mortality but not MACE. After controlling for age, female TNF-α inhibitor users had a significantly increased risk of all-cause mortality. Meanwhile, after controlling for sex, TNF-α inhibitor users aged 60 years or older demonstrated a significantly elevated risk of all-cause mortality. In conclusion, No statistically significant difference in MACE risk was observed between patients who used TNF-α and IL-12/23 inhibitors. Nevertheless, the use of IL-12/23 inhibitors, especially among older and female patients, resulted in a lower overall mortality.

Morphological characteristics of facial scars: A retrospective analysis according to scar location, onset, age, and cause.

The morphology of facial scars shows a wide variation in terms of texture and colour. To date, there are no reliable predictors of aberrant scarring. We conducted a retrospective analysis to identify factors associated with specific scar features and types. Photographs and medical records of 428 patients with facial scars were retrospectively reviewed. Patients with keloids were excluded. The mean age of the patients was 45.43 ± 23.13 years with a male-to-female ratio of 1:1.36. Atrophic scars were the most common (42.8%), followed by flat scars (38.7%) and hypertrophic scars (18.5%). Scars on the forehead were more likely to be atrophic, whereas scars on the chin/jaw and around the mouth were more likely to be hypertrophic. Hypopigmentation was significantly more common in scars located on the forehead. Redness (erythema) was significantly more common in scars located on the chin/jaw. Old scars were less likely to be erythematous, and hypertrophic. Atrophic scars were more common in younger patients. Scars caused by dermatologic conditions, such as acne, were more likely to be atrophic, whereas surgical scars had the lowest risk of being atrophic or hypertrophic. In conclusion, the location, onset, and cause of facial scars were associated with specific features of scars.

Clinicopathological characteristics and prognosis of Merkel cell carcinoma: a single-center retrospective study in Korea.

BACKGROUND: Merkel cell carcinoma (MCC) is an aggressive neuroendocrine tumor of the skin with high mortality. However, its clinical characteristics in Asian patients remain uncertain owing to its low incidence. OBJECTIVE: To analyze the clinicopathological features of MCC and identify factors associated with its prognosis. METHODS: The medical records of 62 patients with MCC were retrospectively reviewed. Data on clinical features, survival outcomes, prognostic factors, histopathology and immunohistochemical profile of the patients were collected and analyzed. Merkel cell polyomavirus status was evaluated using immunohistochemistry. RESULTS: The incidence of MCC significantly increased over time. The mean duration of follow-up was 51.2 months, with an overall 5-year survival of 80.6%. More female patients with MCC were identified than male patients (1.3:1). Approximately half of the patients had stage I disease at the time of initial presentation. The primary tumor was frequently located in the lower extremities (40.3%), followed by the head and neck (32.3%), upper extremities (22.6%), and the trunk (4.8%). Male sex was associated with poorer overall survival (p = 0.003). Post-resection adjuvant radiotherapy significantly improved the overall survival (p = 0.023). Sentinel lymph node biopsy during surgery ameliorated the progression-free survival (p = 0.036) in patients with stage I or II cancer. Lymphovascular and perineural invasion were associated with a poor prognosis. Old age, immunohistochemical profiles, and Merkel cell polyomavirus-positivity were not associated with prognosis. CONCLUSION: Post-surgical adjuvant radiotherapy and sentinel lymph node biopsy significantly improve the course of MCC.

Efficacy of Asymmetric siRNA Targeting Androgen Receptors for the Treatment of Androgenetic Alopecia.

Asymmetric small interfering RNAs (asiRNAs) that mediate RNA interference have been investigated for therapeutic use in various tissues, including skin tissue. Androgenetic alopecia (AGA) is caused by a combination of genetic factors, resulting in sensitivity to dihydrotestosterone (DHT), which binds to the androgen receptor (AR) to mediate a series of biomolecular changes leading to hair loss. This study aimed to evaluate the therapeutic potential of a cell-penetrating, AR-targeting asiRNA (cp-asiAR) for AGA treatment, which was designed to silence the AR gene. AGA mouse models were developed by stimulation with DHT, and ex vivo human scalp tissues were also used for analysis. Cp-asiAR-mediated changes in mRNA expression and protein levels of AR were assessed along with the examination of phenotypic improvements in mouse model of AGA. We also assessed downstream signaling associated with AR in primary human dermal papilla (DP) cells. Several cp-asiARs were screened for selecting the optimal sequence of AR using cell lines in vitro. A cholesterol-conjugated, chemically modified cp-asiAR candidate was optimized under passive uptake conditions in vitro. Intradermal cp-asiAR injection efficiently reduced mRNA and protein levels corresponding to AR in mouse models. Moreover, cp-asiAR injection promoted hair growth in mouse models with DHT-induced AGA. In ex vivo human hair follicle culture, the proportion of telogen hair decreased, and the mean hair bulb diameter increased in the cp-asiAR-treated group. In isolated primary human DP cells, AR expression was effectively downregulated by cp-asiAR. Furthermore, cp-asiAR attenuated DHT-mediated increases in interleukin-6, transforming growth factor-ß1, and dickkopf-1 levels. No significant toxicity was observed in DP cells after cp-asiAR treatment. Cp-asiAR treatment showed effective downregulation of AR expression and prevention of DHT-mediated alterations in the hair cycle and hair diameter, indicating its potential as a novel therapeutic option for AGA.

Evaluation of Artificial Intelligence-Assisted Diagnosis of Skin Neoplasms: A Single-Center, Paralleled, Unmasked, Randomized Controlled Trial.

TRIAL DESIGN: This was a single-center, unmasked, paralleled, randomized controlled trial. METHODS: A randomized trial was conducted in a tertiary care institute in South Korea to validate whether artificial intelligence (AI) could augment the accuracy of nonexpert physicians in the real-world settings, which included diverse out-of-distribution conditions. Consecutive patients aged >19 years, having one or more skin lesions suspicious for skin cancer detected by either the patient or physician, were randomly allocated to four nondermatology trainees and four dermatology residents. The attending dermatologists examined the randomly allocated patients with (AI-assisted group) or without (unaided group) the real-time assistance of AI algorithm (https://b2020.modelderm.com#world; convolutional neural networks; unmasked design) after simple randomization of the patients. RESULTS: Using 576 consecutive cases (Fitzpatrick skin phototypes III or IV) with suspicious lesions out of the initial 603 recruitments, the accuracy of the AI-assisted group (n = 295, 53.9%) was found to be significantly higher than those of the unaided group (n = 281, 43.8%; P = 0.019). Whereas the augmentation was more significant from 54.7% (n = 150) to 30.7% (n = 138; P < 0.0001) in the nondermatology trainees who had the least experience in dermatology, it was not significant in the dermatology residents. The algorithm could help trainees in the AI-assisted group include more differential diagnoses than the unaided group (2.09 vs. 1.95 diagnoses; P = 0.0005). However, a 12.2% drop in Top-1 accuracy of the trainees was observed in cases in which all Top-3 predictions given by the algorithm were incorrect. CONCLUSIONS: The multiclass AI algorithm augmented the diagnostic accuracy of nonexpert physicians in dermatology.

Saponins of Korean Red Ginseng May Protect Human Skin from Adipokine-Associated Inflammation and Pigmentation Resulting from Particulate Matter Exposure.

BACKGROUND: Exposure to airborne particulate matter (PM) is an ever-increasing concern worldwide. Strategies to counter the detrimental effects that follow cutaneous exposure to PM, such as induction of pigmentation, inflammation, and alterations in adipokine profile, need to be investigated further. Korean red ginseng (KRG) extracts and individual ingredients have been demonstrated to play an effective role in suppression of ROS, inflammation, and resultant skin aging. In addition, recent investigations revealed that Rg3 and Rf saponins work as antimelanogenic agents. In this study, we investigated whether saponins of KRG can protect against or reverse the PM-induced detrimental effects. METHODS: The biological effects of PM and saponins were evaluated both in vitro and ex vivo. Cell viability and intracellular ROS levels were determined in normal human epidermal melanocytes (NHMs), human epidermal keratinocytes (NHKs), and their cocultures. Experiments to demonstrate the protective properties of saponins against consequences of exposure to PM were performed. Melanin assay, quantitative real-time PCR, and Western blotting were carried out to determine the effects on melanogenesis and the implicated molecular signaling pathways. RESULTS: Exposure to PM resulted in decreased keratinocyte viability, which was coupled with augmented oxidative stress. These changes were attenuated by treatment with saponins. PM exposure resulted in increased expression of leptin, which was reduced by saponins. Moreover, PM exposure led to increased melanin production in a coculture model, which was mitigated by treatment with saponins. Treatment with saponins resulted in a decrease in matrix metalloproteinase (MMP) levels after exposure to PM. CONCLUSION: Saponins of KRG can protect the skin from the harmful effects of PM exposure by reducing levels of ROS, leptin, inflammatory cytokines, and melanin.

Efficacy and safety of a novel combined 1060-nm and 635-nm laser device for non-invasive reduction of abdominal and submental fat.

The aim of this study was to evaluate the safety and efficacy of combined 1060-nm diode laser and 635-nm low-level laser therapy (LLLT) device for non-invasive reduction of the abdominal and submental fat. Forty-two healthy subjects received single laser treatment on both the abdomen and submental area. Ultrasound images measuring the thickness of abdominal and submental fat were taken at baseline, follow-up at 4, 8, and 12 weeks after treatment. Waist circumference and body weight were also measured at all visits. Adverse events were recorded at all visits. Subjects completed a satisfaction questionnaire at the end of the trial. Twelve weeks after a single treatment with the investigational device, ultrasound images showed statistically significant (P < 0.0001) reductions in abdominal and submental fat by 18.62 and 26.4%, respectively. In addition, significant (P < 0.0001) reduction in waist circumference was observed. Ninety-six percent of subjects rated that they were satisfied. Noted side effects were transient mild to moderate tenderness which subsided within 1 to 3 weeks. No serious treatment-related adverse events were reported. The dual wavelength device combining 1060-nm diode laser with 635-nm LLLT was safe and effective for non-invasive reduction of both abdominal and submental fat.

Treatment of atopic dermatitis using non-thermal atmospheric plasma in an animal model.

Cold atmospheric plasma (CAP) has been incorporated into various fields, including promotion of cutaneous wound healing. Atopic dermatitis (AD) is a chronic cutaneous condition characterized by inflammation-induced skin wounds and impaired skin barrier function. To investigate whether CAP may improve AD using an animal model. Dermatophagoides farinae extracts (DFE)-induced murine models of AD were used in this study. The plasma-treated group received a total of 6 CAP treatments during 2 weeks, while the control group did not receive any treatment. Differences in dermatitis severity, transepidermal water loss (TEWL), serum level of immunoglobulin (Ig) E and epidermal thickness were evaluated in both groups. The dermatitis severity was significantly improved by CAP treatment. TEWL was lower in the plasma-treated group compared with the non-treated control group. Serum Ig E dropped significantly after treatment with CAP. Difference in epidermal thickness of the ear skin was not significant between the plasma-treated and non-treated groups. Localized treatment of AD with CAP decreases dermatitis severity, TEWL, and serum Ig E level. These results show CAP's potentials as a novel therapeutic modality for AD.

The Evaluation of Efficacy and Safety of A Radiofrequency Hydro-Injector Device for the Skin around the Eye Area.

In recent years, variable rejuvenation techniques, such as hyaluronic acid (HA) fillers and radiofrequency (RF) devices, have become popular. We evaluated the RF hydro-injector (RFHI) device that simultaneously delivers both a microneedle intradermal RF treatment and a HA filler injection to overcome the disadvantages of HA filler and RF devices alone. This study aimed to assess the efficacy and safety of the RFHI device for the rejuvenation of the periorbital area, including the lateral canthal lines (LCLs) and the infraorbital area. A total of 24 subjects were enrolled in this study and underwent 2 to 3 treatments using the RFHI device. The investigator's global assessment of the lateral canthal line (IGA-LCL) and the global esthetic improvement scale (GAIS) were used to evaluate the improvement in the LCL. Cutometer® was used to evaluate the skin's elasticity, and the Antera 3D image capture system® was used to evaluate the degree of wrinkles, roughness, and pore volume. At the 8 week follow-up after the first treatment session, both the IGA-LCL and GAIS showed significant improvement. The improvement in the wrinkles, roughness, and pore volume, measured by the Antera 3D image capture system®, was statistically significant. No serious adverse event was reported. This RFHI device, which delivers both microneedle intradermal RF treatment and HA filler injection, is effective and safe for periorbital rejuvenation.

Clinical characteristics and long-term outcomes of 97 Korean patients with mycosis fungoides.

BACKGROUND: Mycosis fungoides (MF) shows racial and regional differences in terms of clinical features. The availability of therapeutic options as well as drugs differs from one country to another. There are only limited data on the clinical characteristics and treatment experience of MF from South Korea. METHODS: Medical records of 97 patients with MF were retrospectively analyzed to investigate clinical features, survivals, and prognostic factors. Assessment of prognostic variables was done using univariate Cox proportional hazard models. RESULTS: Median age at time of diagnosis was 45 years. The median time from onset of skin lesion to diagnosis of MF was 36 months with a median follow-up period of 96 months. A number of clinical variants of MF were observed. Treatment mainly consisted of narrow-band UVB, systemic retinoids, methotrexate, chemotherapy, and regional radiotherapy. Complete remission was observed in 78% of patients with records on their clinical course. About 12% experienced disease progression. No clinical prognostic factor apart from TNM staging was identified. CONCLUSION: Despite delay in diagnosis, most cases of MF in Korea were diagnosed in early stages. Prognosis of our patients was more favorable than those of other geographic regions as reported in previous studies. Good response to treatment, consisting mainly of phototherapy and radiation therapy, and relatively indolent clinical behavior of disease were observed in this homogeneous cohort of Korean patients with MF.

The Efficacy and Safety of Dual-Frequency Ultrasound for Improving Skin Hydration and Erythema in Patients with Rosacea and Acne.

Inflammatory skin diseases, such as rosacea and acne, are major causes of facial erythema and accompanying skin barrier dysfunction. Several methods to restore the impaired skin barrier and improve facial erythema, such as medication, radiofrequency, laser, and ultrasound therapy were attempted. This study evaluated the efficacy and safety of dual-frequency ultrasound with impulse mode, for improving skin hydration and erythema in Asian subjects with rosacea and acne. Twenty-six subjects with facial erythema received an ultrasound treatment once per week, for 4 weeks, over both cheeks. The erythema index and transepidermal water loss (TEWL) were measured at each visit. Clinicians assessed the erythema improvement and patients evaluated their satisfaction level. The average decrease in TEWL and erythema index at 6 weeks was 5.37 ± 13.22 g·h-1·m-2 (p = 0.020) and 39.73 ± 44.21 (p = 0.010), respectively. The clinician's erythema assessment and the subject satisfaction questionnaire score significantly improved at final follow-up (p < 0.001; p = 0.003, respectively). No serious adverse effects were observed during the treatment and follow-up periods. The dual-frequency ultrasound with impulse mode appears to be effective and safe for improving skin hydration and erythema in patients with rosacea and acne.

Ursodeoxycholic Acid May Inhibit Environmental Aging-Associated Hyperpigmentation.

Extrinsic aging of the skin caused by ultraviolet (UV) light or particulate matter is often manifested by hyperpigmentation due to increased melanogenesis in senescent skin. Ursodeoxycholic acid (UDCA), which has been commonly used as a health remedy for liver diseases, is known to possess antioxidant properties. This study was done to investigate whether UDCA inhibits cellular aging processes in the cells constituting human skin and it reduces melanin synthesis. ROS, intracellular signals, IL-1α, IL-8, TNF-α, cyclooxygenase (COX)-2, type I collagen, and matrix metalloproteinases (MMPs) levels were measured in human dermal fibroblasts treated with or without UDCA after UV exposure. Melanin levels and mechanistic pathways for melanogenesis were investigated. UDCA decreased ROS, senescence-associated secretory phenotype (SASP), and proinflammatory cytokines induced by UV treatment. UDCA reduced melanogenesis in normal human melanocytes cocultured with skin constituent cells. Our results suggest that UDCA could be a comprehensive agent for the treatment of environmental aging-associated hyperpigmentation disorders.

Assessment of deep neural networks for the diagnosis of benign and malignant skin neoplasms in comparison with dermatologists: A retrospective validation study.

BACKGROUND: The diagnostic performance of convolutional neural networks (CNNs) for diagnosing several types of skin neoplasms has been demonstrated as comparable with that of dermatologists using clinical photography. However, the generalizability should be demonstrated using a large-scale external dataset that includes most types of skin neoplasms. In this study, the performance of a neural network algorithm was compared with that of dermatologists in both real-world practice and experimental settings. METHODS AND FINDINGS: To demonstrate generalizability, the skin cancer detection algorithm (https://rcnn.modelderm.com) developed in our previous study was used without modification. We conducted a retrospective study with all single lesion biopsied cases (43 disorders; 40,331 clinical images from 10,426 cases: 1,222 malignant cases and 9,204 benign cases); mean age (standard deviation [SD], 52.1 [18.3]; 4,701 men [45.1%]) were obtained from the Department of Dermatology, Severance Hospital in Seoul, Korea between January 1, 2008 and March 31, 2019. Using the external validation dataset, the predictions of the algorithm were compared with the clinical diagnoses of 65 attending physicians who had recorded the clinical diagnoses with thorough examinations in real-world practice. In addition, the results obtained by the algorithm for the data of randomly selected batches of 30 patients were compared with those obtained by 44 dermatologists in experimental settings; the dermatologists were only provided with multiple images of each lesion, without clinical information. With regard to the determination of malignancy, the area under the curve (AUC) achieved by the algorithm was 0.863 (95% confidence interval [CI] 0.852-0.875), when unprocessed clinical photographs were used. The sensitivity and specificity of the algorithm at the predefined high-specificity threshold were 62.7% (95% CI 59.9-65.1) and 90.0% (95% CI 89.4-90.6), respectively. Furthermore, the sensitivity and specificity of the first clinical impression of 65 attending physicians were 70.2% and 95.6%, respectively, which were superior to those of the algorithm (McNemar test; p < 0.0001). The positive and negative predictive values of the algorithm were 45.4% (CI 43.7-47.3) and 94.8% (CI 94.4-95.2), respectively, whereas those of the first clinical impression were 68.1% and 96.0%, respectively. In the reader test conducted using images corresponding to batches of 30 patients, the sensitivity and specificity of the algorithm at the predefined threshold were 66.9% (95% CI 57.7-76.0) and 87.4% (95% CI 82.5-92.2), respectively. Furthermore, the sensitivity and specificity derived from the first impression of 44 of the participants were 65.8% (95% CI 55.7-75.9) and 85.7% (95% CI 82.4-88.9), respectively, which are values comparable with those of the algorithm (Wilcoxon signed-rank test; p = 0.607 and 0.097). Limitations of this study include the exclusive use of high-quality clinical photographs taken in hospitals and the lack of ethnic diversity in the study population. CONCLUSIONS: Our algorithm could diagnose skin tumors with nearly the same accuracy as a dermatologist when the diagnosis was performed solely with photographs. However, as a result of limited data relevancy, the performance was inferior to that of actual medical examination. To achieve more accurate predictive diagnoses, clinical information should be integrated with imaging information.

Clinicopathological and survival analyses of primary cutaneous angiosarcoma in an Asian population: prognostic value of the clinical features of skin lesions.

BACKGROUND: Primary cutaneous angiosarcoma (CA) is a rare but aggressive tumor with a high rate of local recurrence. This study was designed to analyze the clinicopathological features of primary CA and identify factors of cutaneous manifestations associated with the prognosis of angiosarcoma. METHODS: Medical records of 55 patients with primary CA were retrospectively analyzed to investigate clinical features, survivals, and prognostic factors. Anatomical location of tumor was classified to the scalp, face, and neck, and sites outside the head and neck. RESULTS: Primary CA presented cutaneous nodules (31/55, 47.2%), patches (13/55, 23.6%), and indurated plaques (11/55, 20.0%). Nodular lesion was significantly more common in CA on the scalp compared to CA on sites outside the scalp. Histologically, tumors presenting as nodular lesions on the scalp was predominantly composed of solid sheets of large pleomorphic cells, whereas non-nodular lesions composed of tumor cells between collagen bundles forming irregular vascular spaces. Cutaneous angiosarcoma on the scalp showed a worse prognosis compared to CA on sites outside the scalp. Patients presenting clinical morphology with nodules and multiple skin lesions showed significantly reduced overall survival (OS). CONCLUSION: In primary CA, location on the scalp, morphology with nodules, and multiplicity of skin lesions significantly affected survival outcomes.

Keratinocytic Skin Cancer Detection on the Face Using Region-Based Convolutional Neural Network.

Importance: Detection of cutaneous cancer on the face using deep-learning algorithms has been challenging because various anatomic structures create curves and shades that confuse the algorithm and can potentially lead to false-positive results. Objective: To evaluate whether an algorithm can automatically locate suspected areas and predict the probability of a lesion being malignant. Design, Setting, and Participants: Region-based convolutional neural network technology was used to create 924 538 possible lesions by extracting nodular benign lesions from 182 348 clinical photographs. After manually or automatically annotating these possible lesions based on image findings, convolutional neural networks were trained with 1 106 886 image crops to locate and diagnose cancer. Validation data sets (2844 images from 673 patients; mean [SD] age, 58.2 [19.9] years; 308 men [45.8%]; 185 patients with malignant tumors, 305 with benign tumors, and 183 free of tumor) were obtained from 3 hospitals between January 1, 2010, and September 30, 2018. Main Outcomes and Measures: The area under the receiver operating characteristic curve, F1 score (mean of precision and recall; range, 0.000-1.000), and Youden index score (sensitivity + specificity -1; 0%-100%) were used to compare the performance of the algorithm with that of the participants. Results: The algorithm analyzed a mean (SD) of 4.2 (2.4) photographs per patient and reported the malignancy score according to the highest malignancy output. The area under the receiver operating characteristic curve for the validation data set (673 patients) was 0.910. At a high-sensitivity cutoff threshold, the sensitivity and specificity of the model with the 673 patients were 76.8% and 90.6%, respectively. With the test partition (325 images; 80 patients), the performance of the algorithm was compared with the performance of 13 board-certified dermatologists, 34 dermatology residents, 20 nondermatologic physicians, and 52 members of the general public with no medical background. When the disease screening performance was evaluated at high sensitivity areas using the F1 score and Youden index score, the algorithm showed a higher F1 score (0.831 vs 0.653 [0.126], P < .001) and Youden index score (0.675 vs 0.417 [0.124], P < .001) than that of nondermatologic physicians. The accuracy of the algorithm was comparable with that of dermatologists (F1 score, 0.831 vs 0.835 [0.040]; Youden index score, 0.675 vs 0.671 [0.100]). Conclusions and Relevance: The results of the study suggest that the algorithm could localize and diagnose skin cancer without preselection of suspicious lesions by dermatologists.