Expert Review of Medical Devices

Value of colposcopy with dynamic spectral imaging in the detection and evolution of high-grade cervical lesions

To analyze the performance of colposcopy assisted with dynamic spectral imaging (CC-DSI) compared to conventional colposcopy (CC) in diagnosing cervical intraepithelial neoplasia 2+ (CIN2+). In total, 973 consecutive women were referred for colposcopy between 2012 and 2016 and were examined simultaneously by CC and CC-DSI; 877 were eligible. Comparisons between CC and CC-DSI were performed against the histological diagnosis performed by either punch biopsy or loop electrosurgical excision procedure (LEEP). In final histology, 494 women had no CIN, 250 had CIN1 and 133 had CIN2+. The sensitivity to identify women with CIN2+ was significantly higher for CC-DSI compared to CC for the entire group and in the subgroup of 675 women referred with ASCUS or LSIL. In women with HPV16/18 infections, the sensitivity of CC increased with the addition of DSI from 53% to 79% ( CC-DSI improves the ability to detect cervical lesions compared to CC and could help predict their potential to progress.

Learning curve and factors influencing successful robot-assisted bilateral sentinel lymph node mapping in early-stage cervical cancer: an observational cohort study

To evaluate whether a learning curve affects the bilateral sentinel lymph node (SLN) detection in early-stage cervical cancer. All patients with FIGO (2018) stage IA1-IB2 or IIA1 cervical cancer who had undergone robot-assisted SLN mapping performed with a combination of preoperative technetium-99m nanocolloids (including preoperative imaging) and intraoperative blue dye were retrospectively included. Risk-adjusted cumulative sum (RA-CUSUM) analysis was used to determine if a learning curve based on bilateral SLN detection existed in this cohort. A total of 227 cervical cancer patients were included. In 98.2% of patients (223/227) at least one SLN was detected. The bilateral SLN detection rate was 87.2% (198/227). Except for age (OR 1.06 per year, 95%CI 1.02-1.09), no significant risk factors for non-bilateral SLN detection were found (e.g., prior conization, BMI or FIGO stage). The RA-CUSUM analysis showed no clear learning phase during the first procedures and cumulative bilateral detection rate remained at least 80% during the entire inclusion period. In this single-institution experience, we observed no learning curve affecting robot-assisted SLN mapping using a radiotracer and blue dye in early-stage cervical cancer patients, with stable bilateral detection rates of at least 80% when adhering to a standardized methodology.

Digital colposcopy image analysis techniques requirements and their role in clinical diagnosis: a systematic review

Colposcopy is a medical procedure for detecting cervical lesions. Access to devices required for colposcopy procedures is limited in low- and middle-income countries. However, various existing digital imaging techniques based on artificial intelligence offer solutions to analyze colposcopy images and address accessibility challenges. We systematically searched PubMed, National Library of Medicine, and Crossref, which met our inclusion criteria for our study. Various methods and research gaps are addressed, including how variability in images and sample size affect the accuracy of the methods. The quality and risk of each study were assessed following the QUADAS-2 guidelines. Development of image analysis and compression algorithms, and their efficiency are analyzed. Most of the studied algorithms have attained specificity, sensitivity, and accuracy which range from 86% to 95%, 75%-100%, and 100%, respectively, and these results were validated by the clinician to analyze the images quickly and thus minimize biases among the clinicians. This systematic review provides a comprehensive study on colposcopy image analysis stages and the advantages of utilizing digital imaging techniques to enhance image analysis and diagnostic procedures and ensure prompt consultations. Furthermore, compression techniques can be applied to send medical images over media for further analysis among periphery hospitals.