Evaluation of optimal strategies for breast cancer screening in Ghana: A simulation study based on a continuous tumor growth model

Abstract

Mammographic breast cancer screening plays a crucial role in detecting small tumors, which can prevent the progression of the disease and reduce the risk of breast cancer mortality. This study aimed to evaluate optimal strategies for a breast cancer screening program in Ghana. A continuous growth model was employed to evaluate the natural history of breast cancer in Ghana, from its onset to detection. We estimated tumor growth rates and the age at which symptomatic detection occurs using the maximum likelihood estimation method based on clinical data from the National Center of Radiotherapy and Nuclear Medicine at Korle Bu Teaching Hospital. Our results revealed that biennial screening provided a better trade-off between interval cancers and overdiagnosis than annual or triennial intervals. The simulation results for early screening under biennial intervals showed an average detection age of 47 years for unscreened individuals (control group) and 46 years for those screened (intervention group). While the screening approach (50-69 years) with biennial screening proved more reliable than other strategies, the early screening approach (30-65 years with biennial screenings) provided certain advantages in detection for the Ghanaian population. Our findings highlight the importance of early detection and advocate for the systematic adoption of mammography in Ghana and other low- and middle-income countries, contributing to enhanced breast cancer screening and patient treatment plans, as well as informing policy development.

Fig 1. Age at onset distribution.

Results reveal a median age of onset of 38 years (mean: 36.57 ${\displaystyle \pm }$ 11, interquartile range: 29–46). Interval cancers appear to onset later, with a median age of 40 years (mean: 40 ${\displaystyle \pm }$ 10, interquartile range: 33–48), compared to screen-detected cases, which have a median age of 37 years (Mean: 36 ${\displaystyle \pm }$ 11 and interquartile range: 28–45).

Fig 2. Tumor size distribution.

Interval cancers were larger, with a median size of 26 mm (mean = 27 $\pm$ 12, interquartile range = 17–37). The median tumor size for screen-detected cases remained smaller at 12 mm (mean = 14 $\pm$ 10, interquartile range = 9–18).

Fig 3. Tumor size distribution.

The median tumor size for interval cancers was 35 mm (mean = 41 $\pm$ 26, inter quartile range = 21–56). The median tumor size for screen detected cases was 13 mm (mean = 15 $\pm$ 10, inter quartile range = 9–18).

Fig 4. Tumor volume doubling time.

The estimated median tumor doubling times are 168 days (mean: 192 ${\displaystyle \pm }$ 137, interquartile range: 81–278) for interval cases and 309 days (mean: 329 ${\displaystyle \pm }$ 208, interquartile range: 162–469) for screen-detected cases.

Fig 5. Tumor presence time for interval cancers.

The median tumor presence time for interval cancers was 3.9 years, with a mean of 6.46 years ($\pm$6.95) and interquartile range of 1.68 to 8.87 years.