It was a cross-sectional, descriptive study, over a period of 6 months (from October 8, 2020 to March 21, 2021), carried out in the Radiology and Medical Imaging Department of the Military Hospital of Yaoundé (HMY).

The population consisted of patients admitted to the MRI department of the Yaounde Military Hospital (YMH) for pelvic or abdominal-pelvic MRI for clinical suspicion, or extension assessment of histologically confirmed cervical cancer, or for follow-up of treated cervical cancer. Patients with known renal insufficiency or contraindication to contrast media injection were excluded from our study.

The magnetic resonance imaging protocol included T2 Blade sequences in the three planes, T1 Dixon, T1 FS, diffusion, axial and coronal TSE T2 sequences on the abdomen (lymph node exploration), as well as T1FS sequences in the three planes after contrast injection. The slice thickness was 4 - 5 mm, and the field of exploration covered from the recto-sigmoid hinge to the pubic symphysis, including the iliac lymph nodes. The injected contrast medium was 50 cc of Gadolinium chelates.

We retrieved from the medical records on presenting complaints, initial clinical examination, MRI reports and images.

All MRI were reviewed by experienced radiologists.

The primary variables in this study were the pelvic MRI result. The imaging data collected were the presence of the tumor, its size, and the characterization of the tumor stage according to the IFOG classification [5] as well as lymph node involvement.

A tumor localized to the cervix was considered as stage I, a tumor with extension to parametrium and/or to the upper 1/3 of the vagina was stage II tumor, a tumor reaching the pelvic wall and/or lower 1/3 of the vagina and/or with presence of hydronephrosis was a stage III tumor, and the tumor with extension beyond the pelvis or attenuating the bladder and/or rectum was a stage IV tumor [5].

Secondary variables were socio-demographic data (age, geographical distribution, marital status, occupation), and clinical signs and symptoms.

The data collected was processed through the following software: Microsoft Excel 2016 and IBM SPSS 23. The results are expressed in tables and figures, as mean and percentage.

A total of 54 patients were included in our study, with a mean age of 50.5 years. The most represented age group was 50 - 64 years (38.90%, n = 21), as shown in Figure 1. 55.60% (n = 30) of these patients lived in rural area, 59.6% (n = 31) of them were married, and 25% (n = 15) were widows. Housewives represented 51.9% of the patients (n = 28).

The vast majority of our patients (55%, n = 30) consulted on average 6 months after the onset of symptoms.

Table 1 shows the different functional signs, which were dominated by metrorragia (92.5%, n = 50).

Figure 2 summarizes the IFOG tumor stages obtained in the 54 patients of our study population. Stage IIB was the most encountered (29.5%, n = 26).

Tumors limited to the cervix accounted for 26.7% (n = 14). Parametrium invasion was unilateral in 13.6% (n = 12) patients. Tumors involving the pelvic wall with associated hydronephrosis were present in 12.5% of cases (n = 11).

Lesions involving the bladder and/or rectum were present in 20.4% of cases (n = 18), as shown in Figure 3. No patient had metastasis at the time of examination.

Concerning lymph node involvement, the internal and external iliac chains were invaded in 33.3% (n = 18 patients), followed by the retroperitoneal areas in 20.4% (n = 11).

Figure 4 and Figure 5 present some illustrative stage of cervical cancer.

In our series, the mean age at diagnosis was 50.5 years. The age group most affected was 50 - 64 years (38.90%). This result is similar to that of Aboubacar Keita et al. who found an age of 52 years in Mali [7], and differs moderately from that of El Ghorfi in Morocco et al. who found an age of 59 years [6]. However, in general, the early onset of cancer in sub-Saharan Africa has already been highlighted by many authors [2] [8].

Most of our patients lived in rural areas (55.60%), as pointed out by Faye et al., who found a greater proportion of women living in rural areas (84.4%) in Senegal [9], the rural area is characterized by the scarcity of screening and awareness campaigns.

The major revealing sign was essentially metrorragia found in 92% of cases, which could be provoked or spontaneous. This result is similar to what Aboubacar et al. found in Mali metrorragia in 98.8% of cases [7].

According to our results, the delay of consultation is long, 55% of the patients consulted after 6 months of evolution, similar to the results of Kouskous et al. who found 57.9% in the city of Fez [10].

The MRI maximum tumor diameter in this study was 52.9 mm (30 - 79 mm) on average. This differs slightly from the results of El Ghorfi et al., who founded 61 mm [6].

Although the estimation of tumor size remains clinical, MRI is of major contribution to improve the definition with an accuracy of 95% [4].

Paramedic invasion represented 72.2% of cases. Our data are similar to those of Kouskous et al. in Fez who found 68% parametrial invasion [10].

As pointed out by Shweel et al., parametrial invasion is an important criterion in the preoperative workup [11] [12]. However, there is a risk of overestimation due to peritumoral edema in the case of voluminous involvement [11] [13] [14].

MRI has a specificity of 97%, and a negative predictive value of 100% for parametrial invasion [15].

Regarding the classification used, which was that of IFOG, we obtained the following results: stage IA (9.10%), stage IB (10.20%), stage IIA (12.50%), stage IIB (29.50%), stage IIIA (5.70%), stage IIIB (12.50%), and stage IVA (20%).

Similar proportions in severe forms were found by Ghorfi et al. (32%) [6] and Shweel et al. in Egypt (26.6%) [11].

This similarity of severe forms could be explained by the difficulties encountered in the health systems of developing countries in terms of mass screening, particularly in Cameroon where 62.9% of cervical cancers are diagnosed at an advanced stage [16].

Concerning lymph node involvement, retroperitoneal areas were invaded in 20.4% of cases, which is a poor prognostic factor, and is equivalent to extrapelvic tumor extension (stage IVB), according to some authors [11] [17].

The limitation of this study was the small size of our sample, probably due to the monocentric nature of our work.