Patrick Petignat

PURPOSE Thermal ablation (TA) has become the conventional method for treatment of precancerous cervical lesions in low-resource settings. After TA, both the squamocolumnar junction (SCJ) and the transformation zone (TZ) may be subject to change. Our aim was to evaluate SCJ and TZ variability after TA. METHODS Study data were collected in a large prospective trial of a cervical cancer screening campaign in Cameroon. For each patient, two sets of cervical photos (native and with acetic acid) were taken before and 6-12 months after TA. The SCJ and TZ were evaluated independently by three observers according to the WHO nomenclature. When discordances were observed between the type of TZ and SCJ selected by each observer, a corrected TZ was established on the basis of the SCJ categorization. Interobserver agreement for TZ interpretation was evaluated using Cohen's kappa coefficient for agreement between two observers and Fleiss' kappa between three observers. RESULTS Fifty consecutive participants were included in the analysis. Seventy-six percent were interpreted as TZ1-2, and 24% as TZ3 before TA. In 56% of cases, the entire SCJ could not be entirely visualized after TA, thus being recategorized as TZ3. Interobserver agreement was fair for diagnosis before TA (Kappa coefficient, 0.34; 95% CI, 0.27 to 0.45) and moderate for diagnosis after TA (Kappa coefficient, 0.40; 95% CI, 0.30 to 0.50). After TA, 36% progressed from TZ1-2 to TZ3, with a moderate interobserver agreement (Kappa coefficient, 0.44; 95% CI, 0.40 to 0.54). CONCLUSION We observed a shift of the SCJ into the endocervical canal after TA. A significant proportion of participants had TZ 3 after treatment, raising the question of visual inspection with acetic acid's applicability as a first-line follow-up examination method after TA.

Visual inspection with acetic acid (VIA) is one of the methods recommended by the World Health Organization for cervical cancer screening. VIA is simple and low-cost; it, however, presents high subjectivity. We conducted a systematic literature search in PubMed, Google Scholar and Scopus to identify automated algorithms for classifying images taken during VIA as negative (healthy/benign) or precancerous/cancerous. Of the 2608 studies identified, 11 met the inclusion criteria. The algorithm with the highest accuracy in each study was selected, and some of its key features were analyzed. Data analysis and comparison between the algorithms were conducted, in terms of sensitivity and specificity, ranging from 0.22 to 0.93 and 0.67 to 0.95, respectively. The quality and risk of each study were assessed following the QUADAS-2 guidelines. Artificial intelligence-based cervical cancer screening algorithms have the potential to become a key tool for supporting cervical cancer screening, especially in settings where there is a lack of healthcare infrastructure and trained personnel. The presented studies, however, assess their algorithms using small datasets of highly selected images, not reflecting whole screened populations. Large-scale testing in real conditions is required to assess the feasibility of integrating those algorithms in clinical settings.