OPTIMIZING SKIN CANCER DETECTION IN THE USA HEALTHCARE SYSTEM USING DEEP LEARNING AND CNNS

Аннотация

INTRODUCTION accuracy significantly varies among health Background and Motivation professionals; some depend on subjective visual Skin cancer is the most common cancer in the USA, examination, which is prone to human error. with millions of new cases reported each year. The Advanced imaging techniques such as dermoscopy two main types of skin cancer include aggressive, have made the diagnosis more accurate, but their life-threatening melanoma and less lethal, though effectiveness greatly relies on the clinician's potentially very morbid if left unattended, non- expertise (Saleh et al., 2023). melanoma types: basal cell carcinoma and Al Amin et al. (2023), states that recent squamous cell carcinoma. According to the developments in AI, especially deep learning, give American Cancer Society, that means an over 99% promising solutions to these challenges. Deep five-year survival rate for early-stage melanoma, learning is a subset of artificial intelligence; a class against just 27% in the case of late-stage detection of such algorithms called convolutional neural (Rahman et al., 2023). This further signifies the all- networks has achieved great success in image important question of diagnosis in time and with recognition tasks, medical imaging included. By high accuracy. Despite its importance, early skin learning patterns in large datasets, CNNs can cancer detection faces various systemic barriers in classify skin lesions with accuracy comparable to, the US. These range from a shortage of or sometimes exceeding, that of dermatologists. dermatologists and limited access to care in rural These AI-driven solutions, when integrated into and underserved areas to high costs related to the USA health system, might change the face of diagnostic procedures. Moreover, diagnostic early diagnosis, reduce disparity in healthcare, and https://www.theamericanjournals.com/index.php/tajmspr 93 THE USA JOURNALS THE AMERICAN JOURNAL OF MEDICAL SCIENCES AND PHARMACEUTICAL RESEARCH (ISSN – 2689-1026) VOLUME 06 ISSUE12 result in improved patient outcomes (Bowmik et of life. Accordingly, public health campaigns al., 2023; Dutta et al., 2024). emphasizing prevention and early detection raise the chances of prognosis being exceedingly better. Objectives However, at every turn, disparities in dermatologic The principal aim of this research project is to care add to the stress-especially in rural versus devise, curate, and propose a deep-learning CNN underserved areas (Ghosh et al., 2024). methodology for skin cancer detection in the USA. Current Diagnostic Techniques of Skin and The specific objectives of this research are: To Their Limitations apply deep learning methodologies for the detection of skin cancer using dermoscopic images, 1. Visual Inspection by Clinicians maintaining sensitivity and specificity at high As per Jaber & Akbas (2024), visual inspection values. To develop a model of a convolutional remains the first line of defense in diagnosing skin neural network adapted to classify the different cancer. Clinicians evaluate lesions based on their types of skin cancers, especially melanoma, BCC, asymmetry, border irregularity, color variation, and SCC. This includes the performance evaluation diameter, and evolution using the ABCDE criteria. of the model on standard metrics of accuracy, Though easy and cheap, it is highly dependent on a precision, recall, and F1-score, with further clinician's experience and expertise. Studies consideration regarding clinical applications demonstrate big variability in diagnostic skills, within the US healthcare system. particularly among non-specialist providers. This LITERATURE REVIEW subjectivism often may lead to the under-sighting of a suspicious case or unnecessary biopsy. According to Islam et al. (2023), Skin Cancer and its Detection in the USA Skin cancer is the most 2. Dermatoscopy diagnosed cancer in the United States, with Dermoscopy is a non-invasive method that allows millions of individuals receiving a skin cancer the professional, using polarized light and diagnosis every year. The main types of skin cancer magnification, to visualize structures subsurface in are basal cell carcinoma, squamous cell carcinoma, the skin that may go unrecognized by the naked and melanoma. BCC and SCC are usually referred eye. Thus, this technique is considered one of the to as non-melanoma skin cancers and are usually finest methods for diagnosing melanoma and highly curable if they are left early. Melanoma, distinguishing malignancies from benign lesions though rare, becomes much more dangerous (Musthafa et al., 2023). Still, dermoscopy itself because it has great tendencies for metastasis. The does require extensive training and can be highly American Cancer Society estimates that in the USA, related to an operator's learning curve concerning melanoma is the cause of more deaths than any diagnostic accuracy. Being only available to a very other form of skin cancer, with 97,610 new cases restr