Non-functioning pituitary microadenoma in children and adolescents: Is follow-up with diagnostic imaging necessary?

Clinical evaluationMicroadenoma, probable microadenoma, and cystic lesion

At the first MRI, 74 children had a lesion and suggested radiological diagnosis, according to the primary report, of a pituitary microadenoma, probable microadenoma, or cystic lesion (Fig. 1). Their median age was 12 years, 43 (58%) were girls, and the majority (n = 48, 65%) were pubertal (Table 1). Presenting symptoms are demonstrated in Table 2. The most common symptom was precocious puberty (n = 27, 36%) and growth disturbances (n = 22, 30%). Of these 74 children, 32 had received hormonal treatment (Table 3).

Table 1 Demographics, number of MRI scans, and follow-up time in children with pituitary microadenoma, probable microadenoma (PA), cystic lesion (CL), and prolactinomaTable 2 Presenting symptoms: pituitary microadenoma, probable microadenoma (PA), cystic lesion (CL), and prolactinomaTable 3 Hormone treatment: pituitary microadenoma, probable microadenoma (PA), cystic lesion (CL), and prolactinoma

Repeated MRI was performed on 55 children, the main study population. The median number of repeated MRI in this group was 3 (2–7), with a median follow-up time of 37 months (4–189 months) (Table 1). One patient with two cystic lesions underwent surgery due to suspected pressure on the pituitary stalk. Two patients had multiple pituitary insufficiencies, one of them with congenital onset. Two patients had GH deficiency. Levels of sex hormones, gonadotropins, and IGF-1 that deviated from expected levels at a certain age were due to early or late puberty. For presenting symptoms and hormone treatment there were no significant differences between the groups diagnosed with a microadenoma/probable microadenoma/cystic lesion.

Prolactinoma

In the 12 children with prolactinoma, eight (67%) were females and their median age at the initial MRI was 16 years (Table 1). All patients were subsequently treated with dopamine agonists resulting in diminished lesion size, but one patient had surgery due to poor effect of pharmacological treatment. The median number of MRIs was 5.5 (2–11), and the median follow-up time was 75.5 months (26–136 months). The most common presenting symptom (Table 2) was galactorrhea (42%).

Clinical management consistent with the primary MRI reports and from medical records

Among those 55 children with repeated MRI, 43 patients (Fig. 1) were diagnosed with a microadenoma or probable microadenoma at first MRI, 19 had no radiologically significant pituitary lesion at the most recent MRI, five were ultimately diagnosed with a microadenoma, 14 were diagnosed with probable microadenoma, and five were diagnosed with a cystic lesion. Three of the children ultimately diagnosed with microadenoma and five of the children ultimately diagnosed with probable microadenoma were scheduled for a new MRI after the data collection. Eight children had no planned MRI and underwent only clinical follow-up. Among those with probable microadenoma, one patient was lost to follow-up because of bad compliance and two patients were lost to follow-up because of unavailable recent medical records. In 12 patients, a cystic lesion was the initial diagnosis (Fig. 1). As the result of the most recent MRI, one of these patients was diagnosed as a probable microadenoma. An additional five patients from the microadenoma/probable microadenoma group were finally diagnosed with a cystic lesion. Of the 16 patients diagnosed with a cyst at the most recent MRI, five were planned to undergo further MRI after the closure of this study. Two underwent clinical follow-up and nine had no planned follow-up. One patient with a cystic lesion ultimately diagnosed as a probable microadenoma was lost to follow-up because of bad compliance.

One child in the main study population of 55 children had a cystic lesion that increased in size. The lesion was 4 mm at the initial MRI, and a small new cyst appeared during follow-up. Five years after the first MRI, the patient developed multiple hormone deficiency. A new MRI scan showed an enlargement for one of the two cysts, which now measured 6 mm. This patient underwent transphenoidal surgery, and the microscopic examination showed a benign cyst.

In total, 19 of 74 patients underwent only one MRI (Fig. 1). These 19 patients were re-evaluated using MRI and a multidisciplinary decision was made regarding diagnosis and follow-up: seven cases were considered normal, nine were diagnosed as a cystic lesion with no need for any further follow-up, and three were diagnosed with probable microadenoma and clinical follow-up was recommended.

All 12 patients with prolactinoma had repeated MRI and a scheduled follow-up after closure of this study. At the most recent MRI, three patients no longer had visible prolactinoma after treatment with dopamine agonists (Fig. 1). In one girl with a prolactinoma, there was initially a decrease in size from 6 to 3 mm. Later, rising prolactin levels were detected due to medication side effects. A new MRI showed an increase to the original size. One patient with a prolactinoma was operated due to problems tolerating the pharmacological treatment.

Review of the lesion size at first and most recent MRI by an experienced neuroradiologist

Of the 55 patients with repeated examinations, 37 patients were scanned only on 1.5 T and eight patients only on a 3 T MRI scanner, both at first and most recent examination. In 10 patients both field strengths were used. The size of the lesions at first and most recent MRI is presented for each group (microadenoma, probable microadenoma, cystic lesion, and prolactinoma) in Fig. 2a–d. Size was defined as largest diameter in any projection – i.e., height, width, or depth. If the lesion was not visible, the size was marked as zero.

Fig. 2figure 2

Lesion size at initial and most recent MRI as measured by an experienced neuroradiologist for (a) Non-functioning pituitary microadenoma (n = 8), (b) Probable microadenoma (n = 35), (c) Cystic lesion (n = 12), (d) Prolactinoma (n = 12). Median, range, and individual size

In eight cases, the lesion was assessed primarily as a microadenoma (Fig. 2a), median size at initial MRI was 4.5 mm (range 2–6 mm), and median size at the latest MRI was 4.5 mm (range 0–6 mm). None of the eight lesions increased significantly when comparing first and last follow-up.

In 35 patients, the lesion was initially assessed as probable microadenoma (Fig. 2b). The median size for probable microadenoma at the first MRI was 3.0 mm (range 2–7 mm), and median size at the latest MRI was 3.0 mm (range 0–7 mm). In nine cases (26%), the presence of microadenoma could not be confirmed at the final MRI. After re-evaluation, none of the 35 lesions increased significantly when comparing first and last follow-up, and 10 (29%) decreased by ≥2 mm.

In 12 patients, the lesion had mainly cystic characteristics (Fig. 2c). At the initial MRI, the median size was 3.6 mm (range 1–7 mm), and the median size at the most recent MRI was 4.0 mm (range 1–7.2 mm). In all cases, a lesion in the pituitary was visible in the final scan. After re-evaluation, 11 of the 12 cystic lesions did not increase significantly when comparing the first and last examination, and none decreased significantly in size. One increased in size, as described earlier.

Of the 12 patients with a prolactinoma, the median size at initial MRI was 6.0 mm (range 4.8–8.8 mm), and the median size at the latest MRI was 3.9 mm (range 0–6.4 mm) (Fig. 2d). Three (25%) could not be detected after treatment, and seven (58%) decreased by ≥ 2 mm from the first to the last scan. In one patient, a transient increase was noted, as described earlier.

Among the children, 19 patients underwent only one MRI (Fig. 1), and 10 of these were assessed as a probable microadenoma with a median size of 2.2 mm (range 1.5–4). The remaining nine were initially assessed as cystic lesions with a median size of 2.0 mm (range 0–2.6). Two of the lesions reported as a cystic lesion in the primary report could not be visualized.

MRI re-evaluation of measurement accuracy of lesion size by two neuroradiologists

In our study, 68.4% of the scans were performed with a 1.5 T scanner and 31.6% with a 3 T scanner. Of the 269 MRI examinations the reviewers could agree that no lesion could be identified in 38 MRI scans (14%). Of the remaining 231 examinations, the reviewers reported 180 (78%) lesions in the same quarter; that is, for 51 (22%) of the MRIs, the reviewers could not agree on lesion location. At validation of the accuracy of the MRI lesion size measurement, the median of all measurements in all three dimensions and the differences in measurements between the reviewers were calculated. All examinations where both reviewers reached an agreement concerning lesion location were identified and presented as a proportion of all reviewed examinations. The median height, width, and depth of the lesion measured by both reviewers is shown in Table 4. Not all lesions could be visualized adequately in all imaging planes. Thus, some lesions were only measured in two dimensions, which explains the differences in numbers of measurements in Table 5, so volume measurements could not be obtained. In most cases, both reviewers agreed about the size of the lesions (Table 5). However, their measurements disagreed > 2 mm in 34/460 (7%) of measurements, and >3 mm in 11/460 (2%) of measurements, which both had been considered as tumour progression. The inter-observer agreement was less accurate with the high field strength 3 Tesla (3 T) (Table 6), especially for probable microadenoma (Supplemental Table 2).

Table 4 The median size in mm for lesion height, width, and depth measured by both viewers, interquartile range in parenthesis, n = 231 MRI examinationsTable 5 Difference in measurements of lesion size presented for all examinations where the two radiologists agreed (measured the same lesion)Table 6 Inter-observer agreement of two different radiologist for MRI examinations depending on the diagnosis pituitary microadenoma, probable microadenoma, cystic lesion, prolactinoma (n = 269), and on the field strength (n = 256)

In 46 of 86 patients, both reviewers and the primary report noticed the lesion within the same location in the pituitary in all examinations in the same patient. Thus, in 40 patients (46.5%) there was disagreement between the reviewers themselves and/or the primary report. In the two cases with lesion enlargement, both reviewers agreed with the primary report.

留言 (0)

沒有登入
gif