We developed an independent and original model to predict mean age at diagnosis of BC diagnosis in this study. The link between mean age of BC diagnosis and geographic regions was similar to that of previous studies [12–14]. Therefore, the mean age of BC diagnosis prediction model in this study with a high explanatory coefficient reflecting the characteristics of the five continents can help develop and implement BC prevention programs in all countries, especially LMIC countries. Mammography screening programs typically target women within the age range of 50–74 years [11] by IARC screening test protocol report. Cancer registries operating in low- and middle-income settings may face particular challenges to following international registration standards [9, 15]. For example, a lack of coverage by pathology laboratories or difficulty accessing diagnosis records reduces the percentage of microscopically verified cases [16–18], and results in postponement of the incidence date as determined according to the ENCR recommendations, which define the incidence date as the date of first histological or cytological confirmation of malignancy [19–21].
Bidoli previously highlighted the statistically significant relationship between population age and median BC age (of note, median age and mean age of BC diagnosis are very similar, Table 1). They calculated a 42% R2 and claimed that a one-year increase in the median age of the population increased by half a year the median onset age of BC [10]. In our study, 69% percent of the variance of the mean age at diagnosis of BC was explained by the variance of continent-based population ageing through their linear relationship: the mean age of the population increased by 0.35 a year the mean onset age of BC. The bivariate approach is an important method for predicting mean age at diagnosis of BC. Random effects models are a useful tool for both exploratory analyses and prediction problems and is particularly pertinent in our study because incidence rate trends of BC are different between continents [22, 23]. In addition, the mean age of BC diagnosis also showed a similar trend [13]. Between continents, a wide variety of factors, race, nutrition, reproductive factors, and genetic factors, should be considered [24–26]. These well identified risk factors may be included in further studies [25, 27]. If national production levels, such as GDPPC, are one of the important factors affecting the BC incidence, it also usually impacts BC screening cancer [4, 28]. Figure 3A shows real difference for BC age in the different continents supporting individualized BC screening age onset. One could argue that difference in is the terms of BC age is actually caused by difference in screening age onset in continents, but the magnitude of mean BC age difference is over screening age onset difference. The robust mean age of BC diagnosis prediction model can be selected for BC prevention, considering intercontinental differences and will be a sufficient basis for the development and implementation of screening test protocols.
However, our study has several weaknesses that must be acknowledged. First, the most recent year of data was 2012. We plan to extend validation study year with CI5 data that will be released next year. Second, this study was conducted with a population approach and not an individualized approach, without considering all known risk factors for BC. Third, there are various races on continents, and the characteristics of each race should be considered, and the model of research needs to be continuously expanded. Finally, there is a small number of registrations on a particular continent, a lack of control for population-based data registrations, and a lack of data continuity. This part can be said to be characteristic of LMIC countries, especially Africa, and due to this, there are great limitations in predicting all continent. Data from the countries in question cannot be verified through several reliable public data sites, and it is conducted on an NGO or private medical basis, which is highly restricted in securing data. Through this study, we were able to confirm the importance of population-based cancer registry in a particular continent, and compare BC incidence and incidence age according to continental characteristics.
The strengths of this study are as follow. First, we used CI5 volume and summary datasets from IARC, which are reliable data collected from the population-based canister registry of each region. This validated use of data increases the reliability of the values analyzed and calculated in this study and proves to be a reliable result. As for the statistical approach, statistical verification through construction set and validation set confirmed significant results, and the statistical approach was an appropriate study. From this we have identified a simple but unprecedented measure of mean age at diagnosis of BC prediction formula, which is simply applicable to all continents in the study. In addition, if applied to countries with insufficient or low data collection capacity, it will greatly help to decide appropriate screening targets and protocols in LMIC countries. BC screening tests will need to be introduced in many LMIC countries, and at the same time, a system that can collect population-based data should be introduced. BC diagnostic age predictions may help specifically, taking into account life expectancy evolution.
We have developed an unprecedented and robust model based on population age and continents to predict the average age of BC in the population. This tool can be used as an important basis for implementing BC screening in countries, especially LMICs, without prevention programs, and for promoting system development for BC-related population-based data collection, ultimately helping prevent BC and reduce mortality.