Breast cancer incidence in mobile screening vs. in-hospital screening programmes based on 6313607 mammograms in 2387756 women in Taiwan

Background Since 2010, Taiwan has conducted a population-based breast cancer screening programme that includes mobile units to improve the screening rate. This study aimed to compare the results of mobile breast cancer screening units with those of hospital-based units and estimate the preclinical detectable phase (PCDP) transition time for Taiwanese women with breast cancer. Methods This retrospective cohort study included women aged 45 to 69 years who participated in the programme from 2010 to 2018, with at least two years of follow-up, allowing time to detect breast cancer and determine mortality status. The five-state Markov exponential regression model was used to find (1) the underlying incidence of breast cancer for each location type and (2) the sensitivity that each mammography type offers for detecting PCDP early- and late-stage asymptomatic breast cancer. These parameters shed light on the natural history of breast cancer. Results Between 2010 and 2018, 2387756 women were screened via 6313607 mammograms, 55% of which were performed in hospital-based units. The annual pre-clinical incidence rate per person was 0.0035 (95% CI=0.0035–0.0036) in hospital units, greater than that in mobile units (0.0022, 95% CI=0.0022–0.0022). The progression rate from early to late stage within the PCDP was 0.2950 (95% CI=0.2877–0.3025). The progression rates to the clinical phase from PCDP with early- and late-stage breast cancer were 0.1762 (95% CI=0.1713–0.1811) and 0.4157 (95% CI=0.4056–0.4258), leading to calculated mean sojourn times of 3.50 (95% CI=3.45–3.56) years and 2.36 (95% CI=2.30–2.42) years, respectively. Using computed radiography to collect the mammograms during PCDP with early-stage breast cancer resulted in a sensitivity of 0.503 (95% CI=0.484–0.521), whereas using digital radiography resulted in a sensitivity of 0.669 (95% CI=0.653–0.684). When used during PCDP with late-stage breast cancer, these sensitivities were 0.629 (95% CI=0.609–0.648) and 0.797 (95% CI=0.782–0.811), respectively. Conclusions The incidence of breast cancer was greater among women visiting in-hospital screening than mobile screening, and the mammogram type was the primary factor affecting sensitivity for detecting asymptomatic breast cancer.