Brain MRI in the Decision for Liver Transplantation in Pediatric Neurological Wilson's Disease

Abstract

Background

Neurological Wilson's disease (WD) presentation in the pediatric population is rare, and liver transplantation (LT) in these patients remains controversial. The aim of the present study was to assess the role of brain magnetic resonance imaging (MRI) in predicting reversion of brain lesions and neurological outcomes in pediatric WD patients after LT.

Methods

Patients with confirmed WD (Leipzig score ≥4), disease onset in pediatric age (<18 years), neurological involvement, and submitted to LT were selected. Clinical records and pre‐ and post‐LT brain MRI were evaluated.

Results

Six patients met the pre‐defined inclusion criteria, one of whom died shortly after LT and was excluded. The indication for LT was end‐stage liver disease in two patients and neurological worsening despite optimized treatment in three patients. After LT, the neurological picture progressively improved in all patients. Pre‐LT brain MRI showed T1‐weighted hyperintensities in four patients, which quickly resolved afterward. T2‐weighted hyperintensities were observed in four patients before LT, completely resolving in one patient, stabilizing in two, and improving in one after LT. A direct correlation could not be found between clinical and neuroradiological improvement. Progressive clinical improvement was observed even in patients with irreversible brain MRI changes. Conversely, some patients with normal MRI had only slight neurological improvement.

Conclusions

The pattern of T2‐weighted hyperintensities after LT was unpredictable and did not correlate with neurological outcomes, suggesting that these changes may not entail irreversible clinical damage. Therefore, brain MRI does not seem to have prognostic value for assessing clinical response to LT.

Wilson's disease (WD) is a rare autosomal recessive metabolic disease caused by dysfunction of a copper transporter (aTP7B). ¹ , ² It is characterized by disabled copper excretion into the bile, with excess deposition in the liver, kidneys, cornea, and brain. ³ Clinical manifestations include a wide spectrum of signs and symptoms, including neurological and hepatic impairment. ⁴ WD in childhood has been reported to have a predominantly hepatic phenotype. ⁵ Nevertheless, the neurological presentation is probably underestimated in pediatric age, with >15% of WD patients developing neurological symptoms during childhood. ⁶ , ⁷ , ⁸ Neurological symptoms are thought to be secondary to copper accumulation in the brain and can be very severe, even in cases with stable chronic liver disease. ⁹ , ¹⁰

Brain magnetic resonance imaging (MRI) has an important role in the diagnosis of neurological phenotypes, also enabling assessment of disease severity and prognosis. ¹¹ Despite the ubiquitous presence of toxic copper in the brain, pathologic findings are primarily limited to the basal ganglia, thalamus, brain stem (mainly substantia nigra and pontine tegmentum), and cerebellum. ¹² Classical findings in brain MRI include T2 and T2‐fluid‐attenuated inversion recovery (FLAIR) hyperintensities (usually bilateral and symmetrical) in those locations. High‐signal intensities also ensue in basal ganglia (mainly in the globus pallidus) on T1‐weighted images and may reflect changes secondary to chronic liver disease. ¹¹

Liver transplantation (LT) is the sole resolute therapy for WD and the treatment of choice for patients who have fulminant hepatic failure or end‐stage cirrhosis not responsive to medical treatment. ¹³ , ¹⁴ The role of LT in improving neurological function in patients with stable liver disease remains controversial, although preliminary data show that this can be a rescue therapy in the event of an unfavorable evolution despite a well‐conducted chelation treatment. ¹⁵ , ¹⁶ , ¹⁷ Transplantation decisions in such cases are often difficult and made on a case‐by‐case basis, making prognostic imaging features of potentially great help in making the decision.

However, the correlation between neurological manifestations and brain MRI findings before and after LT is not fully understood. The aim of this study was to assess the role of brain MRI performed before LT in predicting clinical neurological outcomes and reversion of brain image lesions after LT.

Methods

Clinical data

Patients with confirmed WD (Leipzig score ≥4), disease onset in pediatric age (<18 years), severe neurological involvement, and submitted to LT were selected among the total suspected WD cases followed at our institution between 2003 and 2020. Demographic and clinical data, laboratory findings, and clinical outcome of all patients were retrieved from clinical records.

MRI acquisition

All exams were performed on a General Electric Signa Explorer 1.5T MR system (GE Healthcare, Waukesha, Wisconsin) with an 8‐channel phased array head coil.

Because of the extension of the period under analysis (from 2003 to 2020), acquisition parameters slightly varied over the years. Still, the following sequences were consistently acquired in all patients: axial T1‐weighted imaging (WI) (repetition time [TR], 596 ms; echo time [TE], 15 ms), axial and coronal fast spin echo (FSE) T2‐WI (TR, 6783 ms; TE, 140 ms), axial FLAIR (TR, 11,000 ms; TE, 140 ms; inversion time, 2800 ms), axial diffusion‐weighted imaging and corresponding apparent diffusion coefficient map, and axial gradient recalled echo T2*‐WI (TR, 693 ms; TE, 23 ms).

Image analysis

Pre‐ and post‐LT brain MRI were evaluated. Follow‐up brain MRI was performed between 5 and 12 months after LT. All images were analyzed by a resident neuroradiologist with 4 years of experience supervised by a senior neuroradiologist. Both were blinded to all clinical information and outcomes. MRI severity score as proposed by Dusek et al. ¹⁸ was applied to both pre and post‐LT brain MRI.

Results

Six of a total of 195 WD patients (three male and three female) met the pre‐defined inclusion criteria. One patient was excluded from the analysis for early death because of post‐LT complications. Patients' mean age at diagnosis was 12.33 years (range, 4–15), mean age at LT was 15.83 years (range, 14–18), and mean length of follow‐up was 8.4 years (range, 3–15).

All five WD patients presented neurological symptoms before LT. The indication for LT was end‐stage liver disease in two patients (group 1, patients 1 and 2) and worsening neurological picture despite optimized medical treatment with stable asymptomatic hepatic disease in the remaining three patients (group 2, patients 3 to 6). Neurological symptoms progressively improved after LT in all patients, with the two patients from group 1 becoming asymptomatic. Table 1 depicts clinical and brain MRI changes observed in each patient before and after LT.

TABLE 1.

Pre‐ and post‐liver transplantation clinical and radiological findings

	Patients	Indication for LT	Pre‐LT brain MRI	Post‐LT brain MRI	Outcome
Group 1	Patient 1	Hepatic MELD 17 (severe parkinsonism, facial dystonia; hypophonia; encephalopathy) mRankin 3 Age at symptom onset (yr): 12 Age at LT (yr): 15	2 yr after symptom onset T2 hyperintensities in caudate nucleus, putamina and mesencephalon Normal T1 signal MRI score: 6	7 mo after LT 4 yr after symptom onset T2 hyperintensities in caudate nucleus, putamina and mesencephalon Normal T1 signal MRI score: 6	Asymptomatic mRankin 0 (5 mo after LT) 3 yr of follow‐up
	Patient 2	Hepatic MELD 22 (mild hand tremor, mild cognitive dysfunction) mRankin 1 Age at symptom onset (yr): 4 Age at LT (yr): 16	12 yr after symptom onset T1 hyperintensity in globi pallidi MRI score: 0	5 mo after LT 13 yr after symptom onset Normal brain MRI MRI score: 0	Asymptomatic mRankin 0 (3 mo after LT) 8 yr of follow‐up
Group 2	Patient 3	Neurologic MELD 12 (dysarthria, dysphagia; dystonia, severe parkinsonism) mRankin 3 Age at symptom onset (yr): 13 Age at LT (yr): 17	3 yr after symptom onset T2 hyperintensities in putamina, mesencephalon and cerebellar peduncles T1 hyperintensity in globi pallidi MRI score: 8	1 yr after LT 5 yr after symptom onset Maintenance of T2 hyperintensities Regression of T1 hyperintensity in globi pallidi MRI score: 8	Improved mRankin 1 (dysarthria and mild dystonia) 15 yr of follow‐up
	Patient 4	Neurologic MELD 11 (severe generalized dystonia and spasticity; anarthria; bedridden) mRankin 5 Age at symptom onset (yr): 11 Age at LT (yr): 14	1 yr after symptom onset T2 hyperintensities in pons, mesencephalon, cerebellar peduncles, thalami, lenticular nucleus, and external capsules Thalami‐restricted DWI T1 hyperintensity in globi pallidi MRI score: 12	1 yr after LT 4 years after symptom onset Partial regression of thalami and mesencephalon T2 hyperintensities Cavitated lesion on BG Regression of T1 hyperintensity in globi pallidi MRI score: 8	Improved mRankin 4 (anarthria; generalized dystonia and spasticity persisted; able to walk with support) 13 yr of follow‐up
	Patient 5	Neurologic MELD 12 (dystonia, dysarthria) mRankin 4 Age at symptom onset (yr): 14 Age at LT (yr): 16	1 yr after symptom onset T2 hyperintensities in caudate nucleus, putamina and mesencephalon T1 hyperintensity in globi pallidi MRI score: 6	–	Died after LT a
	Patient 6	Neurologic MELD 6 mRankin 5 (oromandibular and severe axial dystonia, hypophonia, anarthria) Age at symptom onset (yr): 13 Age at LT (yr): 14	1 yr after symptom onset T2 hyperintensities in caudate nucleus, putamina, mesencephalon and cerebellar peduncles Normal T1 signal MRI score: 8	6 mo after LT 2 yr after symptom onset Marked regression of T2 hyperintensities Normal T1 signal MRI score: 1	Improved mRankin 3 (generalized dystonia (better than before LT) and anarthria) 3 yr of follow‐up

^a Excluded from further analysis because of the short follow‐up.

Abbreviations: BG, basal ganglia; DWI, diffusion‐weighted imaging; LT, liver transplantation; MELD, model for end‐stage liver disease; MRI, magnetic resonance imaging.

Brain MRI was abnormal in the five patients before LT, with MRI scores ranging from 0 to 12 (Table 1). T2‐weighted hyperintensities were present at typical locations (basal ganglia, thalamus, mesencephalon, and pons) in four patients. The pattern of T2 changes after LT was uneven and significantly varied between individuals. T2‐weighted hyperintensities completely resolved in one patient (patient 6) (Fig. 1), improved in another patient (patient 4), and remained stable in two patients (patients 1 and 3).

FIG. 1.

Wilson's disease (neurological phenotype) with complete reversal of MRI changes (patient 6). T2‐FLAIR hyperintensities in the caudate nucleus and putamina (A) and mesencephalon (B) before LT, with marked regression after LT (C,D) in a child with improved neurological signs and symptoms after LT. FLAIR, fluid‐attenuated inversion recovery; LT, liver transplantation; MRI, magnetic resonance imaging.

T1‐weighted hyperintensities in globus pallidus were present in three patients before LT (patients 2, 3, and 4, belonging to groups 1 and 2) and quickly resolved afterward (Fig. 2D–H).

FIG. 2.

Wilson's disease (neurological phenotype) with partial reversal of MRI changes (patient 4). T2‐FLAIR hyperintensities in pons, mesencephalon, caudate nucleus, lentiform nucleus, and thalami before LT (A,B,C), with marked reversion of MRI changes in pons, mesencephalon, and thalami after LT (E,F,G), and persisting (although diminished) basal ganglia in T2‐FLAIR hyperintensities (G). T1‐weighted images with hypersignal on caudate nucleus and globi pallidi before LT (D), which completely reversed after LT (H). Neurological signs and symptoms improved after LT. FLAIR, fluid‐attenuated inversion recovery; LT, liver transplantation; MRI, magnetic resonance imaging.

A direct correlation could not be found between the degree of reversion of MRI lesions and the degree of neurological clinical improvement. Despite the improvement in neurological symptoms observed in patients 1 and 3 (with patient 1 becoming completely asymptomatic), T2‐weighted‐image changes persisted after LT in both (Fig. 3). An important involution of T2‐weighted‐image hyperintensities was observed in patient 4 (Fig. 2), with only slight neurological improvement. Despite persistence of important neurological symptoms, T2 lesions completely reversed after LT in patient 6.

FIG. 3.

Wilson's disease (hepatic phenotype) with persisting MRI changes (patient 1). T2‐FLAIR hyperintensities in caudate nucleus, lentiform nucleus, thalamus, and pons before LT (A,B,C) that persisted after LT, along with global brain atrophy (E,F,G). T1‐weighted images showed no signal changes before (D) or after (H) LT. Despite persistence of imaging alterations, the child became asymptomatic after LT. FLAIR, fluid‐attenuated inversion recovery; LT, liver transplantation; MRI, magnetic resonance imaging.

Discussion

LT is considered an excellent option for the treatment of adult WD patients with severe liver disease, because it can reverse biochemical and clinical signs of the disease, with sustained benefit and long‐term survival. ¹⁹ , ²⁰ , ²¹ , ²² , ²³ In pediatric age, there are no guidelines with indications for LT in neurological WD, and the European Association for the Study of the Liver, ²⁴ American Association for the Study of Liver Diseases, ²⁵ and European Society for Pediatric Gastroenterology, Hepatology and Nutrition Committee ²⁶ consider the indication for LT in pediatric WD patients based solely on neurological impairment as merely experimental. The literature is scarce regarding children with neurological WD submitted to LT and pattern of brain MRI changes in these patients.

With a sensitivity of up to 90%, brain MRI is one of the core diagnostic tools in neurological WD. ¹² In the present neurological pediatric WD cohort, all patients had abnormal brain MRI before treatment, which agrees with the high brain MRI sensitivity reported in the literature. ²⁴ Being established that MRI is a sensitive tool to assess neurological involvement, it would be expected that the technique could be used to monitor clinical neurological behavior and brain lesion changes after LT. However, variable neurological improvement was found among patients in this cohort after LT, regardless of brain MRI lesion pattern before LT and lesion evolution pattern thereafter. Based on their experience and supported by evidence on the adult population, ¹⁷ , ²¹ , ²² , ²³ , ²⁷ , ²⁸ , ²⁹ the authors suggest that transplantation should not be denied to pediatric WD patients with neurological manifestations, given that all patients in this cohort displayed at least some degree of neurological improvement after LT. A recent systematic review by Litwin et al ³⁰ analyzed the outcomes of LT in 302 WD patients with neurological symptoms and concluded that LT should be a treatment option for neurological WD patients, especially in patients with severe neurological symptoms, not responding to anti‐copper treatment. Although the pediatric group was not specifically addressed, some of the articles reviewed included children, giving further evidence to support our results.

Regarding radiological findings before LT in cases with hepatic presentation (group 1), patient 1 presented with acute hepatic failure and T2‐weighted hyperintensities, whereas patient 2 presented with cirrhosis with portal hypertension and T1‐weighted hyperintensities on brain MRI. The authors believe that brain MRI translates the profile of liver disease evolution in these patients, with brain manganese deposition in patient 2 (because of chronic hepatic failure). The MRI score proposed by Dusek et al. ¹⁸ does not take into account T1 hyperintensity, explaining why patient 2 was assigned a score of zero. Patients with neurological presentation (group 2) also presented with brain MRI T1‐weighted changes, in addition to T2‐weighted hyperintensities, revealing the manganese deposition, along with brain changes because of copper deposition. The complete reversion of T1‐weighted hyperintensities in globus pallidus after LT further supports the chronic hepatic dysfunction affiliation of these changes, which should, therefore, revert after LT. ¹¹ Brain MRI T2‐weighted changes are thought to reflect copper toxicity, with cytotoxic edema and myelinolysis in initial stages, ³¹ and gliosis, demyelination, neuronal necrosis, and cystic degeneration later on. ³² The reversibility of these image changes after LT is debatable. In this study, T2‐weighted hyperintense lesions reversed in two patients, showing that T2 changes do not always represent irreversible damage and hence, should not preclude LT. The authors speculate that T2 changes observed in these two cases (patients 4 and 6) probably reflected edema rather than necrosis or cystic degeneration. Mocchegiani et al, ³³ Litwin et al, ³⁴ Wu et al,³⁵ and Stracciari et al ³⁶ reported cases of MRI changes that resolved after hepatic transplantation, and more recently, Dusek et al ¹⁸ also claimed that T2 hyperintensities should be reversible.

Clinical neurological improvement (patient 3) and complete neurological recovery (patient 1) not accompanied by resolution of MRI T2 changes were also observed in this study, showing that persistence of MRI changes post‐LT does not imply worse clinical outcomes. This raises the possibility that not all T2 hyperintensities are reversible and therefore, the reversibility of T2 hyperintensities in brain MRI of WD patients may not be a valid biomarker to assess response to treatment with LT. Similar findings of neurological symptom improvement without changes in the disease pathological burden in brain MRI have also been reported in the literature. ³ , ³⁷ In these cases, untreated copper intoxication probably led to neuron necrosis and irreversible cystic changes in the brain. ³⁸ Despite attempts to correlate image findings with WD pathophysiology, lack of neuropathologic‐radiologic studies makes the nature of T2 high‐signal‐intensity lesions still elusive. Although edema and demyelination can be theoretically reversible, gliosis, neuronal necrosis, and cystic degeneration are usually permanent. ¹¹ However, these different pathophysiological stages can coexist and be indistinguishable in imaging assessment. In the present cohort, pre‐ and post‐LT brain MRI findings did not correlate with neurological clinical performance, and it was not possible to distinguish reversible from irreversible T2 hyperintense lesions on pre‐LT brain MRI.

Although several factors may contribute to persistence of T2 MRI changes despite clinical improvement, the time between initial symptoms and treatment onset, as well as some degree of individual susceptibility are pointed out as the most probable. ²⁵ , ³⁹ , ⁴⁰ Therefore, although believing that the timing of LT is crucial and should take place early in the course of disease, the decision to opt for an experimental treatment with high morbidity and mortality without assurance that clinical picture and brain MRI changes are reversible is a major concern.

There is no consensus regarding the best time to perform brain MRI follow‐up after LT, but it is assumed that the later MRI follow‐up takes place, the greater will be the recovery. Brain MRI follow‐up was performed 5 months after LT in patient 1 (with complete neurological recovery despite persistence of T2‐weighted image changes), what may be too soon to assess full radiological improvement. Because this was a retrospective study, the timing of image and clinical follow‐up assessment was not standardized, which represents an important limitation in result interpretation. Although no studies have been conducted to date, the authors hypothesize that brain MRI improvement may be delayed in relation to neurological symptoms.

Conclusion

The present study provides new insights regarding the apparently poor prognostic value of brain MRI in distinguishing reversible from irreversible T2‐hyperintense lesions. Although reversal of brain MRI changes after LT has been documented, some patients still present imaging changes despite clear clinical neurological improvement. Brain MRI changes in children with neurological WD should not preclude LT, because they failed to predict clinical outcomes after treatment. Further prospective and large‐series studies focusing the predictive value of brain MRI in clinical outcomes of pediatric neurological WD after LT are warranted to better understand this emerging topic.

Author Roles

(1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript: A. Writing of the First Draft, B. Review and Critique.

C.P.: 1B, 1C, 2B, 3A.

M.J.M.: 1C.

H.P.M.: 3B.

T.T.: 3B.

E.S.: 3B.

C.R.: 3B.

M.M.: 1A, 3B.

Disclosures

Ethical Compliance Statement

This study was approved by local Ethics Committee. Informed consent was waived. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflicts of Interest

The authors declare that there are no financial disclosures or other conflicts of interest to report.

Financial Disclosures for the Previous 12 Months

No disclosures to declare.