Cervical lesion assessment using real‐time microendoscopy image analysis in Brazil: The CLARA study

• Published in: International journal of cancer Vol. 149; no. 2; pp. 431 - 441
• Main Authors: Hunt, Brady, Fregnani, José Humberto Tavares Guerreiro, Brenes, David, Schwarz, Richard A., Salcedo, Mila P., Possati‐Resende, Júlio César, Antoniazzi, Márcio, Oliveira Fonseca, Bruno, Santana, Iara Viana Vidigal, Macêdo Matsushita, Graziela, Castle, Philip E., Schmeler, Kathleen M., Richards‐Kortum, Rebecca
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 15.07.2021; Wiley Subscription Services, Inc
• Subjects: Adult; Aged; Algorithms; Biopsy; Brazil; Cancer; Cervical cancer; cervical cancer prevention; Cervical Intraepithelial Neoplasia - diagnostic imaging; Cervix; Colposcopy; Computer Systems; deep learning; diagnostic imaging; Female; high‐resolution microendoscopy; Humans; Hysteroscopy - instrumentation; Image processing; Medical research; Microtechnology; Middle Aged; Neural networks; Neural Networks, Computer; Point-of-Care Systems; point‐of‐care; Prospective Studies; Radiographic Image Interpretation, Computer-Assisted - methods; Sensitivity and Specificity; Young Adult; Brazil; high-resolution microendoscopy; deep learning; point-of-care; diagnostic imaging; cervical cancer prevention
• DOI: 10.1002/ijc.33543
• ISSN: 0020-7136; 1097-0215

We conducted a prospective evaluation of the diagnostic performance of high‐resolution microendoscopy (HRME) to detect cervical intraepithelial neoplasia (CIN) in women with abnormal screening tests. Study participants underwent colposcopy, HRME and cervical biopsy. The prospective diagnostic performance of HRME using an automated morphologic image analysis algorithm was compared to that of colposcopy using histopathologic detection of CIN as the gold standard. To assess the potential to further improve performance of HRME image analysis, we also conducted a retrospective analysis assessing performance of a multi‐task convolutional neural network to segment and classify HRME images. One thousand four hundred eighty‐six subjects completed the study; 435 (29%) subjects had CIN Grade 2 or more severe (CIN2+) diagnosis. HRME with morphologic image analysis for detection of CIN Grade 3 or more severe diagnoses (CIN3+) was similarly sensitive (95.6% vs 96.2%, P = .81) and specific (56.6% vs 58.7%, P = .18) as colposcopy. HRME with morphologic image analysis for detection of CIN2+ was slightly less sensitive (91.7% vs 95.6%, P < .01) and specific (59.7% vs 63.4%, P = .02) than colposcopy. Images from 870 subjects were used to train a multi‐task convolutional neural network‐based algorithm and images from the remaining 616 were used to validate its performance. There were no significant differences in the sensitivity and specificity of HRME with neural network analysis vs colposcopy for detection of CIN2+ or CIN3+. Using a neural network‐based algorithm, HRME has comparable sensitivity and specificity to colposcopy for detection of CIN2+. HRME could provide a low‐cost, point‐of‐care alternative to colposcopy and biopsy in the prevention of cervical cancer. What's new? Mobile cervical cancer screening programs that travel to rural areas can improve access to screenings, but it remains difficult for women who screen positive to access diagnostic follow‐up. In this prospective study, the authors evaluated a low‐cost, point of care diagnostic alternative to colposcopy. High‐resolution microendoscopy (HRME) was shown to be equal in sensitivity and specificity to colposcopy for detection of high grade cervical neoplasias. The images collected were used to train and evaluate a neural network‐based algorithm to classify the neoplasias, improving diagnostic performance. Using HRME as an alternative to colposcopy could improve preventive care for cervical cancer.