This retrospective cross-sectional study was conducted at the imaging department of Dalal Jamm Hospital, Dakar, Senegal, between April 2021 and April 2025.

Patients who underwent gadolinium-enhanced imaging on 1.5T MRI, with histologically confirmed sellar or suprasellar masses ≥ 1 cm, were included. Exclusion criteria were: mass < 1 cm, mass without solid component, motion artifacting ROI placement, history of prior surgery or radiotherapy and incomplete DICOM data.

A total of 96 pituitary MRIs were performed during the study period, of which 54 revealed a known or newly diagnosed sellar and/or parasellar, 36 were normal and 6 examinations were lost. Among the 54 abnormal examinations, some patients were eligible for surgery. Histological results were available for 19 of them; the other patients were lost to follow-up. Of these 19 patients, 3 did not meet the inclusion criteria due to incomplete MRI in 2 patients and a lesion < 1 cm in 1 patient, leaving 16 patients included in the final analysis.

All examinations were performed using a head coil on a 1.5T HITACHI MRI, model Echelon Smart, operational since March 2021.

Figure 1. Head of the automatic dual syringe injector (A) and its control room unit (B).

A standard protocol was applied to all patients. Pre-contrast T1-weightened high-resolution (HR) spin-echo (SE) sequences were acquired in sagittal and axial planes: TR 400 - 600 ms, TE 8 - 12 ms, slice thickness 3 - 4 mm. Dynamic contrast enhancement was performed after intravenous injection of 0.1 mmol/kg gadoteric acid, at a concentration of 0.5 mmol/mL at 2 mL/s followed by 20 mL of saline flush, using an automatic double-syringe injector (Figure 1). Dynamic post-contrast T1-weightened images were acquired immediately after injection during 5 to 6 minutes. Other sequences, such as T2-weightened SE HR, Flair, Gradient Echo, and Diffusion, were performed, supplemented by acquisitions of the entire brain volume in 3D T1 SE and T1 SE after gadolinium injection in axial slices.

For this study, we examined only the ES 3D T1 HR sequences (TE at 8 - 20 ms and TR at 400 - 550 ms), ES 3D T1 throughout the entire brain volume for very large lesions, and dynamic T1-weighted sequences (TE at 8 - 20 ms and TR at 500 - 900 ms).

Signal intensity (SI) measurements were performed by one senior radiologist blinded to histology, using Myrian version 2.7.6 reading software. For each lesion, T1 signal intensity ratio and enhancement ratio were calculated.

For T1 signal intensity ratio calculation, a circular ROI was manually placed on the solid enhancing portion of the mass on pre-contrast (SI pre-lesion), avoiding cystic, necrotic or hemorrhagic areas. A reference ROI was placed on normal-appearing white matter at temporal lobe (SI pre white matter). Due to background signal suppression, noise measured in air was zero. Then, the T1 pre-injection signal intensity ratio (SIr pre) was calculated by Formula (1):

The mass was classified hypointense for a SIr pre < 0.8; hyperintense for SIr pre > 1.2 and isointense for the rest (Figure 2 and Figure 3).

For enhancement ratio calculation, a circular ROI of 10 - 25 mm² was manually placed on the solid enhancing portion of the mass, avoiding cystic, necrotic or hemorrhagic areas, on dynamic T1 images, at 0 minutes corresponding to the start of injection (SI pre gadolinium) and 4 minutes (SI post gadolinium). This time point was chosen to allow for contrast diffusion into interstitial space of solid tumor component while avoiding late-phase washout [5]. The ROIs were placed on the same dynamic sequence using the acquisition time displayed on the screen as a reference. A second, synchronously acquired sequence (T2 ES) was viewed in parallel to ensure that the second ROI was positioned in exactly the same location as the first. The enhancement ratio (ER) was calculated with Formula (2):

Enhancement intensity was classified as weak for ER < 50%, intense for ER > 100% and moderate for the rest (Figure 4).

The injected sequences, as well as other sequences such as the T2* sequence and the T2-weighted ES sequence, were sometimes used to better distinguish between tissue and non-tissue portions.

Figure 2. Sellar and suprasellar lesion on T1-weighted MRI in a 38-year-old female patient presenting with hyperprolactinemia. ROI 1 (solid portion of the lesion) = 116.9 and ROI 2 (temporal white matter) = 161.5, yielding a ROI 1/ROI 2 ratio of 0.72, consistent with a T1 hypointense lesion.

Figure 3. Sellar lesion on T1-weighted MRI in a 42-year-old male patient presenting with headache and visual disturbances. ROI 1 (solid component of the lesion) = 350.2 and ROI 2 (temporal white matter) = 415.8, yielding a ROI 1/ROI 2 ratio of 0.84, consistent with a T1 isointense lesion.

The homogeneous or heterogeneous nature of the enhancement was analyzed visually. The data were recorded in Excel 2016 software.

Figure 4. T1-weighted dynamic contrast-enhanced MRI with gadolinium injection, showing a sellar lesion with suprasellar extension, in a 37-year-old female patient presenting with visual disturbances and amenorrhea, showing a sellar and suprasellar mass. ROI pre-injection (SI pre) = 262.7 (A) and ROI 4 minutes after injection (SI post) = 321.4 (B). The calculated enhancement ratio is 22.34%, consistent with a weak enhancement.

The parameters studied were: age and sex of patients, histological diagnosis, lesion size, lesion location, T1 signal intensity of the solid portion (hypointense, isointense, hyperintense), gadolinium uptake assessed by its absence or presence, its intensity (low, moderate, or high) and its homogeneous or heterogeneous nature.

Quantitative variables were described by measures of central tendency (mean, median) and measures of dispersion (standard deviation, range). Qualitative variables were described using counts and percentages. Kruskal-Wallis H test was used for comparison of ER between histological groups due to small sample size and non-normal distribution. Post-hoc analysis was performed using Dunn’s test with Bonferroni correction. Effect size was reported as epsilon-squared with 95% confidence intervals. A p-value < 0.05 was considered significant. The significant omnibus Kruskal-Wallis result was not interpreted as evidence of pairwise differences between specific groups.

The patients had a mean age of 44 ± 19.48 years [6 years; 79 years] and the median age was 44 years. The sex ratio was 1. The distribution of patients by age and gender is illustrated in Figure 5.

Figure 5.Distribution of patients by age group and gender.

After histology, the diagnoses were 11 adenomas (68.75%), 3 craniopharyngiomas (18.75%) and 2 meningiomas (12.5%).

The adenomas had an average size of 32.54 ± 16.88 mm, with extremes of 12 mm and 57 mm and a median of 38 mm. They were all intrasellar with suprasellar extension.

Craniopharyngiomas had an average size of 54.40 ± 17.39 mm [34.9 mm; 68.3 mm] and a median size of 60 mm. Two craniopharyngiomas (66.67%) were strictly suprasellar and one craniopharyngioma (33.33%) was suprasellar with intrasellar extension.

The two meningiomas had an average size of 43.1 ± 6.08 mm and were located in the left laterosellar region in both cases.

On T1-weighted images without gadolinium injection, 14 (87.5%) of all lesions were isointense and 2 (12.50%) were hypointense.

Adenomas were isointense in 9/11 cases (81.82%) with an average Sir pre of 0.95 [0.85; 1.1] and hypointense in 2/11 cases with an average Sir pre of 0.78 (18.18%).

Craniopharyngiomas were T1 isointense in 100% of cases (3/3 cases) with an average Sir pre of 1.02 [0.94; 1.13].

Meningiomas were T1 isointense in 100% of cases (2/2 cases) with an average Sir pre of 1.05.

After injection, all lesions (16 lesions, or 100%) showed signs of gadolinium uptake. Table 1 shows the ER by histology.

Eight adenomas (72.73%) were heterogeneous, always associated with central necrotic areas with T1 hypointense (8 cases) and, in addition, hemorrhage with T1 hyperintense in 2 cases. Two craniopharyngiomas were heterogeneous, related to cystic portions and signal voids (Figure 6). Meningiomas were homogeneous.

Table 1.Enhancement ratio (ER) by histology.

Kruskal-Wallis test: p-value = 0.0066; H = 10.04; ER: enhancement ratio; SD: standard deviation.

Figure 6. T1-weighted MRI before (A) and after (B) gadolinium administration in a 17-year-old male patient with growth and visual disturbances, showing a heterogeneous craniopharyngioma. The solid component shows punctate signal voids (arrowhead), and the cystic components show no enhancement (arrows).

This study has several limitations. First, although the difference in ER between groups reached statistical significance, the sample size remains small, limited to 16 patients. These results should be considered preliminary. The small number of meningiomas and craniopharyngiomas limits the reliability of effect size estimates and prevents meaningful post-hoc pairwise comparison. This reflects both the single-center, retrospective design and the relative rarity of certain sellar and suprasellar masses, particularly craniopharyngiomas and meningiomas [6]-[8]. Second, signal intensity measurements were performed manually using circular ROIs. This approach may introduce inter- and intra-observer variability, despite standardization of ROI size and placement. Third, the single-center design using a 1.5T MRI system limits the generalizability of our findings to other field strengths [9].

Despite these limitations, this pilot study provides preliminary quantitative data on gadolinium-induced T1 signal changes in supracentimetric sellar masses imaged at 1.5T. The observed biophysical trends warrant further validation in larger, prospective, multicenter studies with standardized acquisition protocols and automated segmentation techniques.

Following quantitative analysis, 87.5% of lesions were isointense on T1-weighted images. This predominance of T1 isointensity suggests that most solid sellar masses have T1 relaxation times similar to that of white matter at 1.5T. White matter is commonly used as a reference because it has stable and easily identifiable signal. Our findings are consistent with those of Diop etal., who also reported a majority of T1 isointensity in these pituitary [10]. Similarly, other series have described T1 isointensity as the most common pattern of these lesions [11][12]. In contrast, Ohui Acko etal. and Garba etal. reported predominance of T1 hypointense sellar masses [13][14]. This difference may be related to methodological differences, as visual assessment remains subjective and operator-dependent, whereas quantitative signal measurement using ROI analysis provides more reproducible and objective data. Furthermore, differences in field strength, sequence parameters and lesion composition may also account for this variability.

From a diagnostic perspective, T1 isointensity alone is non-specific, as quantitative T1 assessment shows that isointensity is the dominant signal pattern in solid sellar lesions at 1.5T. But when combined with enhanced contrast patterns and morphological characteristics, it may help narrow the differential diagnosis of sellar masses.

Gadolinium-based contrast agents are paramagnetic T1 agents that shorten the longitudinal relaxation time T1 of surrounding tissues [1][15]. The magnitude of T1 shortening depends on tissue vascularity, capillary permeability and extracellular distribution of gadolinium [16]. In our study, acquisition time was standardized and all patients received the same dose, therefore differences in ER primarily reflect intrinsic tissue properties.

Significant differences in ER were observed between histological groups (H = 10.04; p = 0.0066).

Meningiomas showed moderate enhancement but the highest mean ER (89.2% ± 18.5%). This is consistent with their rich arterial vascularization and lack of BBB, allowing free diffusion of gadolinium into the extracellular space [17].

Macroadenomas showed the lowest enhancement with a mean ER of 30.9% ± 5.1%. This weak enhancement may be related to altered vascularization, development of an inefficient arterial blood supply and to the presence of fibrous tissue or microhemorrhages, which reduce gadolinium distribution [1][18]. Our results are in line with studies reporting moderate to low enhancement in pituitary macroadenomas [12][19]. In contrast, Garba etal. reported intense enhancement of macroadenomas based on visual assessment [14]. This discrepancy may be explained by the subjective nature of visual evaluation.

Craniopharyngiomas had inter-mediate ER values with a mean ER of 53.4% ± 11.5%. The presence of microcalcification in solid areas could lower the overall ER. This is consistent with literature describing variable enhancement patterns in craniopharyngiomas depending on their adamantinomatous or pipillary subtype [20][21].

Thus, ER appears useful for the differential diagnosis of sellar lesions.

The pattern of enhancement reflects underlying tissue architecture and vascular distribution. Homogeneous enhancement typically occurs in lesions with uniform cellularity and vascularity, while heterogeneous enhancement results from necrosis, cystic portions, hemorrhage, or calcification.

In this study, meningiomas and some macroadenomas showed homogeneous enhancement, consistent with their compact cellular structure and uniform vascular supply. In contrast, some adenomas and craniopharyngiomas demonstrated heterogeneous enhancement. For craniopharyngiomas, this is explained by their mixed components, even though signal analysis was performed on the solid portion only [20]. Focal signal voids observed in craniopharyngiomas corresponded to calcifications. Serpiginous signal voids in meningiomas usually reflect flow voids from hypervascular vessels [22].