Purpose

Recent studies have reported an association between hyperintensity of the dentate nucleus (DN) and globus pallidus (GP) on T1-weighted images (T1WI) and a history of gadolinium-based contrast agent (GBCA) administration. Increased signal intensity of the DN on T1WI has been reported to be positively correlated with previous exposure to linear GBCAs, but not to macrocyclic GBCAs (1–3). Several reports have focused on pediatric patients (4–8). However, there is no report of the phenomenon in a Japanese population. Therefore, this study aimed to evaluate whether an association exists between a T1-signal increase of the DN or GP on T1WI and previous administration of linear GBCAs in pediatric patients.

Methods

This retrospective study was approved by the institutional review board; the requirement for informed patient consent was waived. The inclusion criteria for participant enrollment were: (i) at least five consecutive administrations of linear GBCAs (gadodiamide or gadopentetate dimeglumine) followed by brain MRI, (ii) no history of administration of other or unknown GBCAs, and (iii) brain MRI using a 1.5-T scanner. A total of 42 patients were assigned to the “Linear GBCAs ≥ 5 administrations” group. Additionally, 42 patients with a history of 1–4 consecutive administrations of linear GBCAs and 42 with no history of GBCA administration (“Linear GBCAs 1–4 administrations” group and “No GBCAs administration” group) were included. Age and sex distributions were matched among the 3 groups. The final study included 126 patients (male:female, 72:54; median age, 16 (range 4–18) years) (Table 1).

Results

The DN-to-pons ratio increased as the number of linear GBCAs administrations increased (P < 0.0001–0.0063, Fig. 1). The GP-to-thalamus ratio of the “Linear GBCAs ≥ 5 administrations” group was significantly higher than that of the “No GBCAs administration” and the “Linear GBCAs 1–4 administrations” groups (P < 0.0001, Fig. 1). The GP-to-thalamus ratio of the “Linear GBCAs 1–4 administrations” group did not differ significantly from that of the “No GBCAs administration” group (P = 1.0000, Fig. 1). The number of previous administrations of linear GBCAs was positively correlated with the DN-to-pons ratio and the GP-to-thalamus ratio (Fig. 2). The interobserver ICC for measurement of the DN-to-pons ratio was excellent (0.8236). That of the GP-to-thalamus ratio was good (0.6738). Figs. 3 and 4 present representative clinical cases. [Discussion] The results of this study indicate that hyperintensity of the DN and GP on T1WI is associated with previous administration of linear GBCAs in pediatric patients. These findings are consistent with previous reports (6–8). Flood et al. reported that a signal increase of the GP was not observed (6). Our study included patients with a history of gadodiamide administration, which is the least thermodynamically stable of the GBCAs. This may affect the discrepancy.

Conclusion

High signal intensity of the DN and GP on T1WI is associated with previous administration of linear GBCAs in pediatric patients.

Acknowledgements

No acknowledgement found.

References

1) Kanda T, et al. Radiology. 2014;270:834-841. 2) Radbruch A, et al. Invest Radiol 2015;50:805-810. 3) Cao Y, et al. AJR Am J Roentgenol 2016;206:414-419. 4) Ramalho J, et al. J Magn Reson Imaging. Radiology 2015;276:836-844. 5) McDonald JS, et al. JAMA pediatr 2017;171:705-707. 6) Flood TF, et al. Radiology 2017;282:222-228. 7) Roberts DR, et al. AJNR Am J Neuroradiol 2016;37:2340-2347. 8) Hu HH, et al. Pediatr Radiol 2016;46:1590-1598.