OBJECTIVES : Current (2018-2019) estimated cervical cancer incidence in Brazil is 15.43/100.000 women yearly, but there is wide regional variation (25.62-9.97) with lowest incidence in the Southeast. Each new case caries high economic/social burden and population-based screening is ideal. Data from current strategies in the private scenario are scarce despite fundamental in the process of achieving the WHO goal (4/100.000).

METHODS : 41,202 women were assisted by a single private laboratory in the years 2017-2018 by conventional cytology screening, with adequate quality indicators. Of those, 218 were submitted to biopsy/surgery because of abnormal screening results (population 1). In parallel, tissue from another 208 women without previous known cytology were seen in the same laboratory (population 2). Precancerous lesions (CIN) grade 1 and 2/3 were considered low (LSIL) and high grade (HSIL).

RESULTS : Population 1 (in house screening): In the previous 3 years, 38.9% of women were screened once, 57% 2-3 times and 4% ≥4 times, with alteration (≥ASC-US) in 5.4%, 3,7% and 13.4% of exams in each subgroup, respectively. In the 218 biopsies, low grade lesions predominated (LSIL/HSIL=1.16), mostly in young women (85% <24-39y), while high grade lesions tended to occur a decade later (81.7% 25-49y). Invasive cancer was present in 6 (2.8%) patients (4/6 in 30-49y and 2/6 in >70y). Population 2 (unknown screening): high grade lesions predominated (LSIL/HSIL=0.67) and 22 (10.5%) patients had invasive cancer (2 relapses). All diagnoses showed a wider age distribution, including low grade lesions (which lost concentration in younger age) and the 20 cases of new invasive cancer (ranging from 25->70y).

CONCLUSIONS : Adequate screening enrollment was associated with earlier disease detection, leading to a cancer incidence (7.3/100.000) below the regional rate. Access to HR-HPV test is crucial to reach the ideal scenario, targeting the high risk population and providing additional opportunity for wider screening enrollment through self sampling.