This retrospective review of the adult CT scans that were conducted in 10 radiology departments in the city of Douala-Cameroon. Exams were randomly selected within age-gender-year strata from April 2019 and April 2020.

A total of 1287 adult patients aged > 20 years, underwent CT examinations and may provide appropriate DLP data according to body region were included in this study, 628 females (48.8 %) and 659 males (51.2%). Study describing radiation dose associated with the 8 most common CT examination protocol of abdomen/pelvis, head, cervical spine, petrous bone, sinus, lumbar spine, chest/abdomen/pelvis and chest were collected.

Patient scan details had been previously recorded in the service CT logbook. The following data were collected: For each CT machine: brand, model and detector rows. For each CT performed: Patient’s age and gender, types of CT examinations and date of completion of the examination. The technical exposure parameters and delivered doses including: number of acquisitions, console-displayed Dose Length Product values (DLP) in mGy∙cm (milliGray.centimeter) for each acquisition and the complete DLP of the examination were collected from the picture archiving and communiction system (PACS) for each patient. Other studied variables were: Used of IV contrast (IV−/IV+).

For each patient, the average DLP value for exam was calculated and the average Effective dose was calculated by conversion factors retrieeved from Shrimpton study [8] . from the International Commission on Radiological Protection (ICRP) [9] [10] . and from the American Association of Physicists in Medecine (AAPM) Task Group 23 [11] (reported in Table 1 ). The average Effective Dose (ED) doses associated with the CT-exams studied were assessed from the calculated DLP using a region- and age-specific coefficient:

$E\left(\text{mSv}\right)={\text{EDLP}}_{\text{region,age}}\times \text{DLP}$(1)

where EDLP_region,age (mSv (mGy·cm)⁻¹) is the normalized value of effective dose per dose-length product over a specific body region for a particular standard patient age [12] - [14] .

The overall lifetime attributable cancer risk (LAR) for cancer incidence and mortality was estimated for each patient, based on the effective dose, patient’s sex and age at exposure using the BIER VII preferred models [15] . This was estimated from tables 12D-1 and 12D-2, respectively, documented in the BEIR VII report schow lifetime risk estamates for cancer incidence and mortality resulting from a single dose of 0.1 Gy at several specific ages. Estimates are shown for, all cancer, leukemia, all solid cancer, and cancer of several specific sites [5] (displayed as Table 2 and Table 3 in this study). These values present the additional risk of different cancers and the total risk of all cancers for ages ranging from 0 to 80 years in both sexes for a dose of 0.1 Gy per 100,000 individuals.

Note: Number of cases per 100,000 persons exposed to a single dose of 0.1 Gy.

Note: Number of deaths per 100,000 persons exposed to a single dose of 0.1 Gy.

${\text{LAR}}_{\text{at age of exposure}}=\frac{\text{Effective dose}\left(\text{Sv}\right)}{0.1}\times \frac{{\text{LAR}}_{\left(\text{cancer}\text{incidence}\right)\text{at}\text{age}\text{of}\text{exposure}}}{100000}$(2)

${\text{LAR}}_{\text{at age of exposure}}=\frac{\text{Effective dose}\left(\text{Sv}\right)}{0.1}\times \frac{{\text{LAR}}_{\left(\text{cancer}\text{mortality}\right)\text{at}\text{age}\text{of}\text{exposure}}}{100000}$(3)

Data were recorded in a excel spreadsheet (v. 2016, Microsoft). The data collection represented a non-random sampling of all the records gathered over a designed period. The dose metrics were analyzed. We provided the mean of DLPs and Eds that were calculated for each age group and projected lifetime attributable cancer risks.

This study was authorized by the Institutional Research Ethics Committee for Human Health at the University of Douala and by the Regional Health Delegation of the littoral Region of Cameroon. In order to respect confidentiality, all the data collected was studied anonymously, the services were coded by numbers (from 1 to 10).

The number of services participants in the study, brand characteristics of the CT machines are presented in Table 4 .

The characteristic of patients include number, age and genders are shown in Table 5 . A total of 1278 patients undergoing CT examinations of abdomen/pelvis, head, cervical spine, petrous bone, sinus, lumbar spine, chest/abdomen/pelvis and chest performed. In this study CT examinations were equally performed among male (659/1287, 51.2%) and female (719/1287, 48.8%) patients.

The results show that, there exist dose variations for the same examination among CT facilities within compare to other study. These variations may result from user selections of different technical parameters as well as manufacturer-specific variations in the design of CT equipment. The dose from our study is globally higher than those in Switzerland (2010) and Iran (2018), apart from the sinus, chest and chest-abdomen-pelvic CT procedures. On the other hand, our CT dose are lower than those in Brazil (2019) and then those USA (2017), apart from the head and cervical spine protocol. Our patient-doses were in the similar ranges to those reported by Uushona et al. [16] . for African countries.

We used “effective dose” to quantify the radiation exposure associated with each CT scan, as it is one of the most commonly reported measurements. The effective dose takes into account the amount of radiation received by the exposed organs and the susceptibility of each organ to the development of cancer as a result of radiation exposure.

The average effective doses associated with the different types of procedures are between: 0.30 and 8.81 mSv. The highest effective dose was observed on the following examinations: the lumbar spine CT (8.81 mSv), chest-abdomen-pelvic CT (6.61 mSv), and abdomen-pelvic CT (6.46 mSv) ( Table 6 ).

Table 7 and Table 8 show the LAR values obtained for cancer incidence and mortality.

The LAR value describes the risk of cancer incidence and mortality due to exposures to ionizing radiation. LAR is defined as an additional cancer risk beyond the baseline cancer risk. This can be calculated for specific cancers as well as all cancers combined. The age and sex-specific LAR of all cancer incidences and mortalities for average effective doses, for each type of examination, was calculated using risk estimates from the BEIR VII preferred models. In this study, we used all cancers combined as an outcome to compare all selected CT scan types.

The risk of radiation-induced cancer varies according to the following factors: the exposure dose of the examination carried out, the age and sex of the patient. The LAR incidence and mortality from cancer in patients aged 20 to 80 who underwent CT examinations were estimated.

In our study, the LAR of the incidence of cancer from adult CT scans obtained varies between 0.52 per 100,000 scans (approximately 1 in 192307) and 145.01 per 100,000 scans (approximately 1 in 689) for the group of age 20 years and 80 years for lumbar spine and sinus CT scans. That of mortality between 67.13 per 100,000 scans (approximately 1 in 1489) and 0.45 per 100,000 scans (approximately 1 in 222,222) for the same age group for lumbar spine and sinus CT examinations. Thus, the highest LAR of cancers mortality 45.01 (male) and 61.13 (female) was observed in 20 years old who underwent a lumbar spine CT scan with an effective dose of 8.81 mSv.

The risks decreased significantly with age and were lower in men than in women, so radiation-associated cancer risks are of particular concern for younger female patients. It is precisely because the risks of cancer are so high in younger patients that special attention should be paid to optimizing CT examinations performed in younger patients 20 years of age.

The result obtained also showed that the LAR of cancer incidence is higher than the LAR of cancer mortality.

Generally, the risk of death due to cancer incidence after CT exposure is estimated to be 1 excess death per 2000 scans [21] . which was estimated to be in the same range in our study (1 excess death 1489 - 2221), based on the effective dose the higher than we observed, for the age group 20 years having undergone lumbar spine CT scan and chest-abdomen-pelvic CT scan. The LAR of mortality of cancer in all other examinations at different age groups was low in our study. The patient dose and consequent induced risk varied considerably from age group and gender for the same procedure. As we used the same model (BEIR VII) to predict the values LAR, this discrepancy may result from different exposure parameters used, resulting doses and age at exposure.

The study observed that the risk of cancer mortality ranged from low to minimal compared to the additional risk of fatal cancer attributable to the duration of life for diagnostic X-ray examinations show in Table 10 [15] . Furthermore, by comparing these low risks with others causes of human death, it can be concluded that the Lifetime Attributable Risk of all cancer’s mortality of these CT exams is in the range between the estimated Lifetime Risk of Death from Drowning as seen in Table 11 .

Keep in mind, the majority of cancers occur later in life and the average lifetime risk of dying from cancer is 25% (1 in 4). These statistics are averages and do not predict what is going to happen for a specific individual. They do not take into consideration individual risk factors including lifestyle (smoking, diet, exercise, etc.), family history (genetics) or radiation exposure.

Our first study limitation was we estimated radiation doses received by patients in clinical practice, that the DLP data were estimates generated by CT scan software, which depend on scan acquisition parameters and the effective dose was estimated using DLP and k conversion factor, where as many previous studies have assessed the dose received in idealized settings on phantoms. Study parameters applied in phantoms may differ substantially (10% - 20% difference in the value obtained from the estimated effective dose using DLP and k conversion factor and measured effective dose) from those used in actual clinical settings [23] - [25] . LAR estimations were done using data from the BEIR VII report since the study did not have access to real data from epidemiological studies of cancer in the city of Douala; therefore, the LAR values obtained in this study are just approximated values and not the precise risk values.