TY - JOUR AU - Jeong, Ki Hyeon AU - Henao, Ricardo AU - Park, Christine AU - Jiang, Simon AU - Nicholas, Matilda AU - Chen, Suephy AU - Kheterpal, Meenal PY - 2023/8/31 TI - Image Quality Assessment Using a Convolutional Neural Network for Clinical Skin Images JO - iproc SP - e49534 VL - 9 KW - teledermatology KW - image quality assessment KW - deep learning KW - convolutional neural network KW - artificial intelligence KW - AI N2 - Background: The quality of the images received for teledermatology evaluation is often suboptimal, with up to 50% of patients providing images that are poorly lit, off-center, or blurry. To ensure a similar level of care to in-person consultations, high-quality images are essential. Objective: The aim of this study is to develop an image quality analysis tool to assess patient- and primary care physician (PCP)?derived images using a deep learning model leveraging multiple instance learning and ordinal regression for model predictions. Methods: The data set used for this study was acquired from patient-derived images submitted to the Department of Dermatology, Duke University, between August 21, 2018, and December 31, 2019, and PCP-derived images between March 1, 2021, and June 30, 2022. Seven dermatology faculty members with a designation of professor, associate professor, and assistant professor evaluated 400 images each, and 2 dermatology residents evaluated 400 images, assuring that each image had 4 different quality labels. We used a pretrained model VGG16 architecture, further fine-tuned by updating weights based on the input data. The images were taken with cell phones (patients) or cameras (PCPs) in RGB scale, with the resolution being 76 pixels per inch for both height and width, and the average pixel size of the image being 2840×2793 (SD 986×983; 1471 inch2, SD 707 inch2). The optimal threshold was determined using the Youden index, which represents the best trade-off between sensitivity and specificity and balance the number of true positives and true negatives in the classification results. Once the model predicts the rank, the ordinal labels are transformed to binary labels by using a majority vote as the goal is to distinguish between 2 distinct categories (good vs bad quality) and not predict quality as a continuous variable. Results: Based on the Youden index, we achieved a positive predicted value of 0.906, implying that the model will predict 90% of the good-quality images as such, while 10% of the poor-quality images are predicted as being of good quality to enhance clinical utility, with an area under the receiver operating characteristic curve (AUC) for the test set at 0.885 (95% CI 0.838-0.933) and sensitivity, specificity, and negative predictive value (NPV) of 0.829, 0.784, and 0.645, respectively. Further evaluation on independent validation consisting of 300 images from patients and 150 images from PCPs revealed AUCs of 0.864 (95% CI 0.818-0.909) and 0.902 (95% CI 0.85-0.95), respectively. The sensitivity, specificity, positive predicted value, and NPV for the 300 images were 0.827, 0.800, 0.959, and 0.450, respectively. Conclusions: This study shows a practical approach to improve image quality for clinical decision-making. While patients and PCPs may have to capture additional images (due to lower NPV), this is offset by the reduced workload and improved efficiency of clinical teams due to the receipt of higher-quality images. Additional images can also be useful if all images (good or poor) are transmitted to medical records. Future studies need to focus on real-time clinical validation of our results. Conflicts of Interest: None declared. UR - https://www.iproc.org/2023/1/e49534 UR - http://dx.doi.org/10.2196/49534 UR - http://www.ncbi.nlm.nih.gov/pubmed/ ID - info:doi/10.2196/49534 ER - TY - JOUR AU - Sodhi, Anmol PY - 2023/8/17 TI - Teledermatology in India During the Peri?COVID-19 Outbreak Period: Advantages, Shortcomings, and Challenges JO - iproc SP - e49488 VL - 9 KW - teledermatology KW - COVID-19 KW - India KW - health care KW - outbreak N2 - Background: Telemedicine is defined as the use of electronic information and communication technologies for health care professionals to provide care to patients. Although available since the pre?COVID-19 era, a huge surge in teledermatology consultations occurred during the COVID-19 outbreak. As access to health care became limited and difficult due to repeated lockdowns, teledermatology helped us provide health care to our patients. Moreover, as dermatology is a visual field, it was even more suitable for teleconsultations. Objective: The objectives of this study were to investigate the advantages, shortcomings, and challenges of teledermatology in India during the peri?COVID-19 outbreak period. Methods: This was a single-center, retrospective, observational study conducted at a tertiary care hospital in India. Teledermatology consultation data from April 1, 2020, till September 2021 (18 months) were included. All modes, including real-time (RT) video, asynchronous store and forward (SAF), and hybrid, were used to conduct teledermatology consultations. Statistical analyses were performed using SPSS software (IBM Corp). Results: During these 18 months, a total of 4280 patients took teledermatology consultations at our center. The mean age of the patients was 34.19 years, with most of them (36.4%) in the age group of 31-40 years. The patient population comprised a mix of urban (55%) and rural (45%) individuals. Overall, 70% of consultations were conducted in the SAF mode; hybrid mode, 16%; and RT video consultations, 14%. Diagnosis was established in 89.1% of the cases, and the most common diagnosis was superficial fungal infection (28%), followed by eczema (16%) and acne (8.6%). Hospital visits were required in the remaining 10.9% of cases for the following reasons: lack of clear pictures and technical errors (5.57%). Additional diagnostic tests were required in 1.3% of cases, physical examination in 1.05% of cases, and 0.39% of patients had life-threatening conditions requiring hospitalization. The advantages of teledermatology include decreased need for hospital visits among 89.1% of patients, which played a very important role in decreasing overcrowding. Also, this helped us provide expert health care to the rural population of India. Owing to shortcomings including the lack of good-quality pictures (4.2%; more so in SAF teleconsultations) and technical errors (1.37%), teledermatology cannot be used to manage life-threatening conditions (0.39%), and, in particular, RT video consultations are more time-consuming (14%). Challenges faced by dermatologists during teledermatology consultations were mainly operational, such as the lack of good internet access leading to interrupted consultations (1.37%), poor quality of pictures (4.2%), and difficulty in extracting history in cases of SAF consultations. Conclusions: Teledermatology serves as a triage platform and helps reduce hospital visits. It helps to cater to the rural population, which otherwise has limited access to health care. Some technical challenges are the dependence of teledermatology on pictures and information sent by the patient for establishing the diagnosis. Also, sometimes patients faced difficulty in conveying problems clearly to the doctors. Because of the ease and advantages, several dermatologists have continued to use teledermatology along with the physical consultations in the post?COVID-19 era. With a few advancements, teledermatology will certainly remain a successful and useful model for consultations, more so for catering to the population with the lack of access to specialist services. Conflicts of Interest: None declared. UR - https://www.iproc.org/2023/1/e49488 UR - http://dx.doi.org/10.2196/49488 UR - http://www.ncbi.nlm.nih.gov/pubmed/ ID - info:doi/10.2196/49488 ER - TY - JOUR AU - Tirado-Pérez, José-Pablo AU - Oakley, Amanda PY - 2023/8/17 TI - Surgical Excision Margins in Primary Care and Plastic Surgery for Keratinocytic Cancers Diagnosed via Teledermatology: Retrospective Observational Cross-Sectional Study JO - iproc SP - e49466 VL - 9 KW - skin cancer KW - tumour recurrence KW - tumor KW - basal cell carcinoma KW - squamous cell carcinoma N2 - Background: The incidence of keratinocytic cancers is increasing. In New Zealand, surgical treatment of skin cancers is often undertaken in primary care. In the Waikato district, general practitioners (GPs) are encouraged to confirm diagnoses via teledermatology. Histological examination should confirm clear surgical margins to reduce tumor recurrence. International guidelines recommend a lateral margin of ?3 mm for basal cell carcinomas (BCCs) and ?4 mm for squamous cell carcinomas (SCCs). Objective: This study aimed to determine lateral and deep margins in keratinocytic cancer excisions performed by GPs (in a private setting) and plastic surgeons (in a private or public setting) after a teledermatologist had confirmed that excision was necessary. Demographic, clinical, and histological features were recorded. Methods: A retrospective observational cross-sectional study was conducted. The sample in the electronic dermatology referral database included keratinocyte cancers recommended for excision from March to May 2022. Results: Histological reports revealed that excision was complete in 186 of 201 confirmed cases of keratinocyte cancer. The lateral margins of resection were considered in 10 tumors and deep margins in 8 tumors. All incomplete excisions were carried out by GPs, 11 of which were on the head and neck. There were 133 BCCs, 100 of which were excised by a GP, 3 by a private plastic surgeon, and 30 by a public hospital surgeon. In total, 52 BCCs were present on the head and neck (25 excised by GPs, 25 by hospital plastic surgeons, and 2 by private plastic surgeons) and 81 were present on other sites (75 excised by GPs, 5 by hospital plastic surgeons, and 1 by a private plastic surgeon). Lateral margins were considered in 9 cases (of which 5 cases involved head and neck tumors). The minimum distance from the tumor to the lateral margin was <3 mm in 80 tumors: 64 were excised by a GP, 2 by private plastic surgeons, and 14 by hospital plastic surgeons. This distance was ?3 mm in 44 tumors (27 excised by GPs, 1 by a private plastic surgeon, and 16 by hospital plastic surgeons). These data show significant adherence to surgical margin recommendations among plastic surgeons compared to that among GPs (odds ratio 2.873, CI 1.274-6.477; P=.009). There were 68 SCCs: 57 were excised by a GP, 2 by a private plastic surgeon, and 9 by a public hospital surgeon. In total, 21 SCCs were on the head and neck (14 excised by GPs, 6 by hospital plastic surgeons, and 1 by a private plastic surgeon) and 47 were on other sites (43 excised by GPs, 3 by hospital plastic surgeons, and 1 by a private plastic surgeon). Lateral margins were considered in 1 head and neck SCC case and were not reported in others. The minimum distance from the tumor to the lateral margin was <4 mm in 35 cases: 31 were excised by a GP, 1 by a private plastic surgeon, and 3 by a hospital plastic surgeon. This distance was ?4 mm in 31 cases (24 excised by GPs, 1 by a private plastic surgeon, and 6 by hospital plastic surgeons). These data do not show significant difference in adherence to surgical margin recommendations between GPs and plastic surgeons (P>.05). Conclusions: Complete resection reduces the risk of recurrence of keratinocytic tumors. GPs in our study were less likely than specialist surgeons to respect surgical margin recommendations established in international guidelines for managing keratinocytic cancer. Conflicts of Interest: None declared. UR - https://www.iproc.org/2023/1/e49466 UR - http://dx.doi.org/10.2196/49466 UR - http://www.ncbi.nlm.nih.gov/pubmed/ ID - info:doi/10.2196/49466 ER - TY - JOUR AU - Silva, Cristiana AU - Vasconcellos, Cidia AU - Souza, Barreto Murilo AU - Fernandes, Dumet Juliana AU - Rego, Pedreira Vitoria Regina de Almeida PY - 2023/8/1 TI - The Effectiveness of Blended Learning for Dermatology Undergraduate Medical Students JO - iproc SP - e49651 VL - 9 KW - dermatology KW - medical education KW - undergraduate medical education KW - distance education KW - e-learning KW - blended learning KW - hybrid course N2 - Background: Novel internet-based applications and associated technologies have influenced all aspects of our society, ranging from areas of commerce and business to entertainment and health care. Education is no exception. In this context, this study was designed to evaluate the impact of a dermatology e-learning program on the academic performance of medical students. Objective: We aimed to develop a dermatology blended learning course for undergraduate medical students and compare the knowledge gained by students who took this course to those who attended traditional classes. Methods: This prospective study evaluated the performance of fourth-semester medical students from the Federal University of Bahia, Brazil. A total of 129 students were selected and divided into 2 groups. The first group (n=57) attended traditional classes and used printed material (books and handouts). The second group (n=72) took our e-learning course and used an e-book as a supplement in a hybrid setting comprising online plus traditional learning. Each course was evaluated with multiple-choice, paper-based tests that were administered at the beginning and end of the course. Results: Although the precourse tests did not show any difference between the traditional and hybrid groups (mean 2.74, SD 1.25 vs mean 3.2, SD 1.22), students attending the hybrid course had better final term grades (mean 8.18, SD 1.26) than those who attended traditional classes (mean 7.11, SD 1.04). This difference was statistically significant (P<.05). Conclusions: The results suggest that the performance of undergraduate students who took a course supplemented with e-learning material was superior to those who attended a traditional course alone. Conflicts of Interest: None declared. UR - https://www.iproc.org/2023/1/e49651 UR - http://dx.doi.org/10.2196/49651 UR - http://www.ncbi.nlm.nih.gov/pubmed/ ID - info:doi/10.2196/49651 ER - TY - JOUR AU - Oh, Dennis PY - 2023/8/1 TI - Implementation of Teledermatology for Veterans in the United States JO - iproc SP - e49532 VL - 9 KW - teledermatology KW - implementation science KW - veteran N2 - Conflicts of Interest: None declared. UR - https://www.iproc.org/2023/1/e49532 UR - http://dx.doi.org/10.2196/49532 UR - http://www.ncbi.nlm.nih.gov/pubmed/35059592 ID - info:doi/10.2196/49532 ER -