Enzyme Replacement Therapy for CLN2 Disease: MRI Volumetry Shows Significantly Slower Volume Loss Compared with a Natural History Cohort ======================================================================================================================================== * Pritika Gaur * Paul Gissen * Asthik Biswas * Kshitij Mankad * Sniya Sudhakar * Felice D’Arco * Angela Schulz * Jens Fiehler * Jan Sedlacik * Ulrike Löbel ## Abstract **BACKGROUND AND PURPOSE:** Neuronal ceroid lipofuscinoses are a group of neurodegenerative disorders. Recently, enzyme replacement therapy (ERT) was approved for neuronal ceroid lipofuscinosis type 2 (CLN2), a subtype of neuronal ceroid lipofuscinoses. The aim of this study was to quantify brain volume loss in CLN2 disease in patients on ERT in comparison with a natural history cohort using MRI. **MATERIALS AND METHODS:** Nineteen (14 female, 5 male) patients with CLN2 disease at 1 UK center were studied using serial 3D T1-weighted MRI (follow-up time, 1–9 years). Brain segmentation was performed using FreeSurfer. Volume measurements for supratentorial gray and white matter, deep gray matter (basal ganglia/thalami), the lateral ventricles, and cerebellar gray and white matter were recorded. The volume change with time was analyzed using a linear mixed-effects model excluding scans before treatment onset. Comparison was made with a published natural history cohort of 12 patients (8 female, 4 male), which was re-analyzed using the same method. **RESULTS:** Brain volume loss of all segmented brain regions was much slower in treated patients compared with the natural history cohort. For example, supratentorial gray matter volume in treated patients decreased by a mean of 3% (SD, 0.74%) (*P* < .001) annually compared with an annual volume loss of a mean of 16.8% (SD, 1.5%) (*P* < .001) in the natural history cohort. **CONCLUSIONS:** Our treatment cohort showed a significantly slower rate of brain parenchymal volume loss compared with a natural history cohort in several anatomic regions. Our results complement prior clinical data that found a positive response to ERT. We demonstrate that automated MRI volumetry is a sensitive tool to monitor treatment response in children with CLN2 disease. ## ABBREVIATIONS: CLN2 : neuronal ceroid lipofuscinosis type 2 ERT : enzyme replacement therapy ICV : intracerebro-ventricular NCL : neuronal ceroid lipofuscinosis TPP1 : tripeptidyl peptidase 1 SUMMARY #### PREVIOUS LITERATURE: CLN2 is 1 of 13 genetically distinct subtypes of human neuronal ceroid lipofuscinoses described to date. Recently, recombinant human tripeptidyl peptidase 1 (cerliponase α), an enzyme-replacement therapy, has been developed and was approved in 2017 for the treatment of CLN2 disease after a clinical trial confirmed a significantly reduced decline in motor and language function in treated patients compared with historic controls. Automated MRI volumetry has been used successfully to quantify brain atrophy in a natural history study of patients with CLN2 and CLN3. #### KEY FINDINGS: Automated MRI volumetry showed that brain atrophy is significantly slower in patients with CLN2 treated with intrathecal enzyme replacement compared with the natural history cohort. The difference in atrophy was most marked for the supratentorial gray matter with an annual 13.8% reduction in volume loss. #### KNOWLEDGE ADVANCEMENT: MRI volumetry is a useful, unbiased tool for the assessment of treatment-related response in CLN2 trials. The method is feasible despite the presence of signal drop-out from the intracerebro-ventricular reservoir by evaluating the contralateral hemisphere only. Neuronal ceroid lipofuscinoses (NCLs) are a group of lysosomal storage disorders characterized by excessive accumulation of lipofuscin in neuronal and extraneuronal tissues1 and neurodegeneration. Currently, 13 genetically distinct subtypes of human NCLs have been identified with variable ages of onset, such as congenital, infantile, late-infantile, juvenile, or adult. The genes affected in NCL encode lysosomal enzymes and other proteins linked to lysosomal functions.2 Ceroid lipofuscinosis type 2 (CLN2) disease is caused by pathogenic variants in the *CLN2* gene, resulting in a deficiency of the lysosomal enzyme tripeptidyl peptidase 1 (TPP1). TPP1 is a serine protease; a deficiency in which results in lysosomal accumulation of a mixture of proteins and lipids.3 Most patients with CLN2 disease present with the classic late-infantile form of the disease. Delay in language development is usually noticed first, followed by the onset of seizures and ataxia between 2 and 4 years of age and eventually psychomotor, language, and visual decline.4,5 An increasing number of patients with CLN2 present with an atypical phenotype that is characterized by later onset and slower neuroregression as well as an incomplete combination of symptoms. The atypical patients with CLN2 are more difficult to recognize clinically because of disease variability, and the time from disease onset to diagnosis is usually longer. On MRI, the natural disease is characterized by progressive infratentorial and supratentorial gray matter atrophy with associated white matter signal abnormalities.6⇓-8 Because the disease mainly affects neurons, white matter signal changes are likely secondary and hypothesized to be related to Wallerian degeneration and gliosis.9,10 Thalamic T2 hypointensity and volume loss have been attributed to the accumulation of saposin and glial fibrillary acid protein containing hypertrophic astrocytes.11,12 Preclinical studies were performed in TPP1-deficient mouse and dog models and eventually on children, in whom the intrathecal and intracerebro-ventricular (ICV) administration of TPP1 halted the neuropathologic features of the disease and slowed disease progression.13⇓-15 In 2017, an enzyme replacement therapy (ERT), cerliponase alfa, was approved for treatment of CLN2 disease by the US FDA and European Medicines Agency after almost 2 decades of research and scientific knowledge advancement. The clinical trials were conducted in 4 international centers (Hamburg, London, Rome, and Columbus) and showed remarkable slowing in the expected rate of clinical decline based on a visual and motor clinical rating score.5 In 2019, cerliponase alfa was approved for reimbursement in the UK for patients diagnosed with CLN2 as part of the managed access agreement.16 The drug is delivered directly into the intrathecal compartment via an ICV reservoir. As part of the managed access agreement, MRI is performed yearly to assess disease progression or treatment-related changes. Recently, MRI volumetry has been suggested as a viable biomarker for monitoring disease progression in neuronal ceroid lipofuscinoses, and the good correlation of volume loss has been described on MRI with the clinical scores.5,6,17,18 In addition, in a CLN2 miniswine model, MRI brain volumetry has been shown to be highly sensitive to early disease detection.19 The purpose of this study was to compare anatomic regional MRI brain volumes of the patients with CLN2 treated with ERT at Great Ormond Street Hospital in London, UK, with a previously published natural history cohort6 and to determine if the reported clinical effect of therapy in the article by Schulz et al15 can also be detected using MRI on long-term follow-up. ## MATERIALS AND METHODS Twenty-eight patients with CLN2 disease were enrolled in treatment studies at Great Ormand Street Hospital for Children at the time of the imaging analysis. For this retrospective analysis, institutional review board approval was waived. Three patients without available follow-up imaging, 3 patients with an atypical form of CLN2 disease, and 3 patients with extensive imaging artifacts were excluded from the analysis, leaving 19 patients (14 female, 5 male) (Online Supplemental Data). Patient details are summarized in the Table. Patients commenced treatment at different time points, typically at ages 2–5 years. Patients did not receive other treatment apart from antiseizure medication. | Demographics (Age at diagnosis/sex) | ERT Start Age (Baseline MRI 1–2wks Earlier) | CLN2 (Motor + Language) Score at ERT Start | Latest CLN2 (Motor + Language) Score (Age) | Phenotype | Mutation | |:-----------------------------------:| ------------------------------------------- | ------------------------------------------ | ------------------------------------------ | ------------------- | -------------------------- | | 4y8m/F | 4y10m | 1 + 0 | 0 + 0 (11y5m) | classical | c89 + 5G>A/ c.509-1G>C | | 4y4m/F | 4y7m | 1 + 2 | 0 + 0 (10y3m) | classical | c.509-1G>C/ c.509-1G>C | | 2y2m/F | 4 y | 3 + 3 | 2 + 2 (9 y) | classical | c.509-1G>C/ c.509-1G>C | | 4y1m/Fa | 4y5m | 3 + 2 | 2 + 2 (8 y) | classical | c.1094G>A/ c.622C>T | | 4y3m/M | 4y5m | 2 + 0 | 1 + 0 (10y9m) | classical | c.1052C>T/ c.1052C>T | | 4y2m/F | 4y7m | 1 + 1 | 0 + 0 (14 y) | classical | c.509-1G>C/ c.509-1G>C | | 3y11m/Ma | 4y4m | 2 + 2 | 1 + 2 (13y3m) | classical | c.1266 G > C/ c.1266 G > C | | 3y7m/F | 3y11m | 2 + 2 | 0 + 0 (13 y) | classical | c.509-1G>C/ c.622C>T | | 0y15m/Fa | 1y9m | 3 + 3 | 3 + 3 (8 y) | classical | c89 + 5G>A/ c.509-1G>C | | 3y5m/Fa | 3y11m | 3 + 2 | 1 + 2 (6y4m) | classical | c.509-1G>C/ c.622C>T | | 4y5m/Ma | 4y9m | 1 + 1 | 1 + 1 (8y3m) | classical | c.1678–1679 del/ c.622C>T | | 7y11m/M | 8y5m | 2 + 2 | 1 + 2 (12 y) | Atypical phenotypeb | c511G>C/ c.622C>T | | 4y2m/Ma | 4y4m | 2 + 1 | 1 + 1 (7y5m) | classical | c379C>T/ c.509-1G>C | | 4y5m/Fa | 4y8m | 2 + 2 | 1 + 0 (6y8m) | classical | c.509-1G>C/ c.509-1G>C | | 4y5m/F | 4y7m | 2 + 2 | 1 + 0 (6y11m) | classical | c17 + 1del/ c.509-1G>C | | 4y0m/Ma | 4y2m | 2 + 2 | 1 + 1 (6y4m) | classical | c1525C>T/ c.509-1G>C | | 13y1m/M | 13y2m | 2 + 2 | 2 + 2 (13y10m) | Atypical phenotypeb | c.1340G>A/ c.509-1G>C | | 10y8m/F | 10y9m | 2 + 2 | 2 + 2 (11y5m) | Atypical phenotypeb | c.1340G>A/ c.509-1G>C | | 4y10m/Ma | 4y11m | 2 + 2 | 1 + 1 (5y10m) | classical | c.509-1G>C/ c.509-1G>C | | 3y10m/Fa | 3y11m | 2 + 2 | 1 + 2 (6 y) | classical | c.509-1G>C/ c.509-1G>C | | 3y1m/Fa | 3y2m | 3 + 0 | 3 + 0 (5 y) | classical | c.509-1G>C/ c.509-1G>C | | 4y3m/F | 4y5m | 2 + 2 | 2 + 2 (5 y) | classical | c.622C>T/c.1645G>A | * aThese patients were included in the Baseline-to-1-year-follow-up MRI subgroup analysis. * bPatients were excluded from our analysis as they had atypical phenotypes. Patient characteristics including genetic mutations and motor and language scores at the point of commencement of ERT versus the latest scores Although patients with CLN2 with atypical phenotypes who had evidence of progressive CNS disease were treated with cerliponase alfa in our center, they had unpredictable disease progression that does not match that of the natural history cohort; hence, they were excluded from this analysis (Table). The patients with unavailable follow-up imaging were enrolled just before the imaging analysis, so no follow-up imaging was yet done or follow-up imaging was not commenced for other reasons. Serial 3D T1-weighted MRIs of our treated patients were available for baseline and follow-up times, ranging from 1 to 9 years. The follow-up period was variable because patients started treatment between 2014 and 2022, with some patients still under follow-up to date. We compared the brain volume changes of patients under treatment (1- to 9-year follow-up scans only) with those of a previously published natural history cohort of 12 patients (8 female, 4 male; observation age range, 2–10 years).6 To compare our results, we re-analyzed the natural history data using the same statistical methods used for our treatment cohort and recorded the underlying gene mutation (Online Supplemental Data and Table). We also specifically analyzed the brain volume change for the first year, including the baseline scan done short before (1−2 weeks) treatment onset and all available follow-up scans within the first year that the patient was under treatment. Eight patients were excluded from this subset because no usable 3D data sets were available for analysis at baseline and within the first year of follow-up (Online Supplemental Data). In addition, MRI results were correlated with the clinical motor-language score. ### Imaging MRI was performed on 1.5T (*n* = 96) or 3T (*n* = 10) scanners with most scans obtained on 1.5T systems. Most children required sedation using general anesthesia to obtain adequate image quality. The imaging protocol included conventional MRI sequences (FLAIR, T2WI, DWI) and a 3D T1-weighted MPRAGE (*n* = 47) or a T1WI gradient-recalled echo (*n* = 59) sequence. The 3D T1-weighted sequences were used for volumetric analysis. The MPRAGE sequence parameters were the following: TR = 150 −2300 ms, TE = 2.5–2.7 ms, T1 = 900−1000 ms, flip angle = 8°; and the gradient-recalled echo parameters were the following: TR = 11−14 ms, TE = 5−7 ms, flip angle = 15°. Both sequences were acquired in a sagittal orientation with a matrix size of 256 × 256 with 150−210 slices and 1-mm isotropic voxels. One hundred six scans were available for analysis from 19 patients. We excluded 7 scans from 2 patients due to extensive artifacts from the ICV reservoir. In addition, some of the patients did not have usable 3D T1 volumetric data sets at certain imaging time points; therefore, 15 scans from a total of 10 patients were excluded from analysis. A total number of 84 good-quality MRI scans were available for our treatment cohort. ### Data Processing The FreeSurfer Software Suite (Version 6.0; [http://surfer.nmr.mgh.harvard.edu](http://surfer.nmr.mgh.harvard.edu))20 was used for brain segmentation, running the automated recon-all command with the watershed atlas option for robust skull-stripping. The segmentation (ie, the aseg.mgz files) of all scans that were 3D volumetric T1 sequences were visually checked (Fig 1) for any major segmentation errors (obvious gray and white matter segmentation errors, inclusion of the dural venous sinuses and/or skull/soft-tissue). No manual segmentation correction was performed (examples of baseline and follow-up scans are in the Online Supplemental Data). ![FIG 1.](http://www.ajnr.org/https://ajnr-sso.highwirestaging.com/content/ajnr/early/2024/10/24/ajnr.A8408/F1.medium.gif) [FIG 1.](http://www.ajnr.org/content/early/2024/10/24/ajnr.A8408/F1) FIG 1. Example of an overlay of FreeSurfer segmentation on a 3D T1 MPRAGE image. Segmentation of supratentorial cortical gray matter (red), supratentorial white matter (green and white), and cerebellar gray (orange) and white matter (yellow) is depicted. In addition, deep gray matter (basal ganglia/thalami) was segmented and analyzed. Note, the susceptibility artifacts from the Ommaya reservoir over the right hemisphere. The contralateral hemisphere was used for data analysis. The volume measurements of the segmented brain regions were extracted from the resulting aseg.stats files of each scan for supratentorial gray and white matter, deep gray matter (basal ganglia/thalami), lateral ventricles (including the choroid plexus), and cerebellar gray and white matter. These are the same regions analyzed in the earlier published natural history cohort.6 ### Statistical Analysis The volume changes were first analyzed between the baseline scans 1−2 weeks before treatment start and all scans acquired during the first year of follow-up. The volume changes were also analyzed only for the 1- to 9-year follow-up scans, excluding the pretreatment baseline scans. Furthermore, the available clinical scores (Table, sum of motor and language scores) were correlated with the supratentorial gray matter volumes of the MRI scans closest to the clinical assessments. Data analysis was in R 4.0.4 ([http://www.r-project.org/](http://www.r-project.org/)) using the linear mixed-effects model package lme4, Version 1.1–33 ([https://github.com/lme4/lme4](https://github.com/lme4/lme4))21 and the partial correlation package (ppcor; [https://cran.r-project.org/web/packages/ppcor/ppcor.pdf](https://cran.r-project.org/web/packages/ppcor/ppcor.pdf)).22 The logarithm of the volume measurements was calculated to linearize the exponentially-appearing volume loss across time (ie, patient age) for the analysis. A linear mixed-effects model was calculated using the syntax of the lme4 package as shown below, to account for the nonindependent multiple follow-up measurements of each patient. ![Formula][1] The ppcor package cannot account for the dependency of follow-up measurements of the same patient; therefore, all measurements were treated as independent for the correlation analysis. Due to the susceptibility artifacts caused by the ICV reservoir used for drug delivery, only the contralateral hemisphere was included in the analysis of the treatment cohort, assuming that the volume change with patient age is the same for both hemispheres. The lme4 formula for analyzing the natural history cohort was, therefore, modified to account for a potential effect between the hemispheres as well an interaction with age as follows: ![Formula][2] There was no statistically significant association or interaction with age of the brain volume change between the hemispheres. This feature allowed us to compare the age effect of the modified analysis of both hemispheres of the natural history cohort with the analysis of the contralateral hemisphere of the treatment cohort, which could be either side. The volume measurements across time of both hemispheres of the natural history cohort were also analyzed with the partial correlation analysis to control for potential correlations between the hemispheres. However, no statistically significant correlation was found between the hemispheres. The remaining correlation coefficients with age of that partial correlation analysis of the natural history cohort were used for comparison with the treatment cohort. ## RESULTS After exclusion of patients as described above, the treatment cohort consisted of 19 patients with CLN2 (14 female, 5 male) who were compared with the published natural history cohort of 12 patients (8 female, 4 male). The age distribution was similar among both cohorts. Overall, more female patients were included in both cohorts, but this finding was even more prevalent in the treatment cohort. The volume loss between the baseline scans and first year follow-up for our treatment cohort (Online Supplemental Data, Fig 2, gray symbols and lines) compared with the natural history cohort (Online Supplemental Data, Fig 2, yellow symbols and lines) was slightly lower for the supratentorial cortical gray and white matter as well as for the cerebellar gray matter, but the basal ganglia/thalami and cerebellar white matter showed a higher rate of volume loss. The lateral ventricles showed a much higher volume increase in the first year on treatment compared with those in the natural history cohort. The Pearson correlation coefficients are similar to those in the natural history cohort, except for the supratentorial cortical white matter, which was close to zero. ![FIG 2.](http://www.ajnr.org/https://ajnr-sso.highwirestaging.com/content/ajnr/early/2024/10/24/ajnr.A8408/F2.medium.gif) [FIG 2.](http://www.ajnr.org/content/early/2024/10/24/ajnr.A8408/F2) FIG 2. Brain volumes of patients with CLN2 treated with ERT. A reduced rate of cerebral volume loss is demonstrated across the supratentorial and infratentorial white and gray matter in our patients (*blue triangles*) compared with a natural history cohort (*yellow dots,* average of both hemispheres). *Gray triangles* depict MRI scans of our treatment cohort before ERT. A slower rate of increase in the volume of the lateral ventricles is also shown. Statistical analysis of brain volume changes correlated with age in the treated (left and middle columns) and untreated patients with CLN2 (right column) is shown in the Online Supplemental Data. The treatment cohort has been analyzed for the baseline-to-year follow-up (left columns) and change during follow-up only (middle columns). All segmented brain regions of the treatment cohort during the follow-up period excluding the baseline scans (Online Supplemental Data) showed a much slower rate of volume loss compared with the natural history cohort. The reduction rates of brain volumes in our treatment cohort (compared with the natural history cohort) per year were the following: supratentorial cortical gray matter, 3% (versus 16.8%), supratentorial white matter, 0.23% (versus 6.3%), cerebellar gray matter, 1% (versus 10.1%), cerebellar white matter, 3.8% (versus 17.2%), and basal ganglia/thalami, 4.4% (versus 12.5%). The lateral ventricles showed a slower rate of increase in volume of 3.5% annually (versus 25.9%). A significant change of volume loss between our cohort and the natural history cohort was observed for supratentorial cortical gray matter (13.8% difference), cerebellar white matter (13.4%), and cerebellar gray matter (9.1%). However, the difference was greatest for the lateral ventricles in which the difference in the volume increase was reduced by 22.4%. Of these regions, the supratentorial cortical gray matter showed the lowest error of the estimated volume change rate, making it the most reliable. The reduced Pearson correlation coefficients compared with the natural history cohort also demonstrated the effectiveness of treatment (Online Supplemental Data). Most important, on treatment, when there is minimal-to-no volume loss, as is the case for the supratentorial white matter region, the volume change rate and the Pearson correlation coefficient are close to zero and, therefore, do not reach statistical significance, with *P* values > .05. The correlation analysis of the supratentorial gray matter volume with the available clinical scores (Fig 3) showed a high correlation with *r* = 0.71 (*P* < .001). Three patients showed a volume loss with no change of the clinical scores (vertical lines, patients C1, D2, and G1), and 1 patient showed a decline of the clinical scores but a slight increase in volume (patient A2). ![FIG 3.](http://www.ajnr.org/https://ajnr-sso.highwirestaging.com/content/ajnr/early/2024/10/24/ajnr.A8408/F3.medium.gif) [FIG 3.](http://www.ajnr.org/content/early/2024/10/24/ajnr.A8408/F3) FIG 3. Correlation analysis between the clinical scores (motor and language) and cortical GM volumes. These were compared for the closest MRI scan date from the date of clinical score assessment (Table). The clinical scores and volume measurements are highly correlated with each other; cortical GM volume can be used as a surrogate marker for the clinical scores. ## DISCUSSION This study analyzed segmented MRI brain volumes of patients with CLN2 on ERT. In comparison with a natural history cohort of untreated patients, brain volume loss was significantly reduced in all brain regions in the follow-up period with the patients already 1 year on treatment, which included supratentorial gray and white matter, infratentorial gray and white matter, and basal ganglia/thalami. The strongest difference in volume decline with age compared with the natural history cohort was observed for supratentorial gray matter, with an annual 13.8% reduction in volume loss. Volume in the cerebellar white matter was reduced by 13.4% and by 9.1% in the cerebellar gray matter. Basal ganglia/thalami and supratentorial white matter showed 7.7% and 6.1% differences, respectively. Most interestingly, a strong difference between the 2 cohorts was also seen for the size of the lateral ventricles where the difference in annual volume increase was 22.4%. Our results suggest that MRI volumetry is useful for quantitative treatment monitoring. Artifacts on imaging limit accurate volumetric assessment; all the patients had ICV reservoirs that are used for ERT administration and were a source of susceptibility artifacts. We were able to demonstrate, by using statistical analysis of the natural history cohort, that it is possible to rely on 1 hemisphere only for data analysis. As previously shown,6 volumetry is a sensitive marker of disease progression as marked by continued volume loss despite scoring a zero on the clinical rating score. Correlation analysis between the clinical scores (motor and language) and cortical GM volumes (Fig 3) revealed that clinical scores and volume measurements are highly correlated with each other; cortical GM volume can be used as a surrogate marker for the clinical scores. Overall, the volume loss is correlated with a decline in clinical scores. Additionally, our current data show continued volume loss despite persistent high clinical scores (patients C1, D2, and G1) (Fig 3), indicating that even in early stages of disease in which there is no clinical progression, MRI is already primed and more sensitive to disease progression and treatment response. Our study adds new knowledge to prior literature by demonstrating that MRI is a more sensitive biomarker compared with clinical scores through regional volumetric analysis as well as encapsulating imaging changes prior to clinical deterioration. Our results are also in agreement with the initial clinical trial of patients with CLN2 on ERT23, with slowed rate of motor and language function decline. Schulz et al5 demonstrated a slowed decline of clinical scores measuring motor and speech decline. In addition, among the treated patients, the annual rate of loss of total gray matter volume over a 96-week period was 6.7%, with larger decreases seen during the first year of treatment compared with the remainder of the period of analysis.5 In our cohort, gray matter volumes showed slightly less reduction annually with 3%. Previous studies in nontreated patients have also demonstrated that MRI volumetry provides a sensitive biomarker for monitoring disease progression in CLN2 disease.6,24 The most rapid volume loss has been observed for the cerebrum, followed by the thalamus and cerebellum, dropping out of the normal range between 6 months to 3 years of age.24 Therefore, supratentorial and infratentorial gray matter showed the strongest difference in annual decline between treated and nontreated cohorts. The annual decrease in supratentorial cortical gray matter in untreated patients of 16.8% was reduced to only 3%. Most interestingly, cerebellar white matter, which showed an annual volume loss of 17.2% in untreated patients, also decreased to only 3.8% in the ERT cohort. This is an interesting finding because the loss of axons and myelin sheaths in the cerebellar and cerebral white matter is believed to be a secondary effect of the loss of granular and Purkinje cells in the cerebellum and loss of cortical neurons.9 It may be that the rate of cerebellar white matter volume loss was much improved in the ERT cohort compared with the natural history cohort because in the latter, there is not only direct secondary Wallerian degeneration via loss of axons through cerebellar gray matter loss but also secondary diaschisis of the white matter tracts secondary to interruption of the corticopontine cerebellar tracts, resulting in hypometabolism of the cerebellum.9,10 There were a few limitations to our study. The imaging parameters and scanners of the data sets varied for our patients due to the retrospective nature of the data collection. This issue can lead to variability of segmentation and, therefore, statistical analysis (Online Supplemental Data). Another limitation was the use of variable 1.5T and 3T scanners, which can lead to subtle differences in the detection of atrophy and thus can affect overall volume prediction, with 3T being better at detecting small changes in volume.25 However, we expect little effect on our results because most scans were performed on 1.5T scanners. Among other limitations were the signal drop-out from the ICV; however, this was countered by using only the contralateral hemisphere, as there was no statistically significant difference between the two hemispheres. Thus, the patients were also included in the protocol at different stages during their disease, introducing a variability in baseline neurodegeneration at the point of treatment commencement. For future directions, it would be prudent to standardize imaging protocols and perform scans on equivalent Tesla MR scanners. Furthermore, it would be interesting to analyze the inherent T1- and T2-weighted image intensities or quantitative parameters such as MR spectroscopy, quantitative susceptibility mapping, or myelin water imaging to assess more subtle changes of disease progression and treatment response. ## CONCLUSIONS MRI volumetry is an effective tool to quantitatively evaluate disease progression and assess the efficacy of treatment, demonstrating higher sensitivity than current clinical biomarkers. Our results have analyzed the rate of volume decline in several structures of the brain; this analysis is novel compared with prior research in this field. The patients on ERT showed a significant reduction in cerebral volume loss in comparison with the natural history cohort. ## Acknowledgments We express sincere thanks to the NCL clinic at the University Medical Center Hamburg-Eppendorf, Hamburg, Germany. ## Footnotes * J. Sedlacik and U. Löbel contributed equally to this work. * [Disclosure forms](https://www.ajnr.org/sites/default/files/additional-assets/Disclosures/November%202024/0248.pdf) provided by the authors are available with the full text and PDF of this article at [www.ajnr.org](http://www.ajnr.org). ## References 1. 1.Nita DA, Mole SE, Minassian BA. Neuronal ceroid lipofuscinoses. Epileptic Disord 2016;18:73–88 doi:10.1684/epd.2016.0844 pmid:27629553 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1684/epd.2016.0844&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=27629553&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 2. 2.Kaminiów K, Kozak S, Paprocka J. Recent insight into the genetic basis, clinical features, and diagnostic methods for neuronal ceroid lipofuscinosis. Int J Mol Sci 2022;23:5729 doi:10.3390/ijms23105729 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.3390/ijms23105729&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=35628533&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 3. 3.Sleat DE, Donnelly RJ, Lackland H, et al. Association of mutations in a lysosomal protein with classical late-infantile neuronal ceroid lipofuscinosis. Science 1997;277:1802–05 doi:10.1126/science.277.5333.1802 pmid:9295267 [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEzOiIyNzcvNTMzMy8xODAyIjtzOjQ6ImF0b20iO3M6Mzg6Ii9ham5yL2Vhcmx5LzIwMjQvMTAvMjQvYWpuci5BODQwOC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 4. 4.Mukherjee AB, Appu AP, Sadhukhan T, et al. Emerging new roles of the lysosome and neuronal ceroid lipofuscinoses. Mol Neurodegener 2019;14:4 doi:10.1186/s13024-018-0300-6 pmid:30651094 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1186/s13024-018-0300-6&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=30651094&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 5. 5.Schulz A, Ajayi T, Specchio N, et al; CLN2 Study Group. Study of intraventricular cerliponase alfa for CLN2 disease. N Engl J Med 2018;378:1898–907 doi:10.1056/NEJMoa1712649 pmid:29688815 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1056/NEJMoa1712649&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=29688815&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 6. 6.Löbel U, Sedlacik J, Nickel M, et al. Volumetric description of brain atrophy in neuronal ceroid lipofuscinosis, 2: supratentorial gray matter shows uniform disease progression. AJNR Am J Neuroradiol 2016;37:1938–43 doi:10.3174/ajnr.A4816 pmid:27231226 [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpuciI7czo1OiJyZXNpZCI7czoxMDoiMzcvMTAvMTkzOCI7czo0OiJhdG9tIjtzOjM4OiIvYWpuci9lYXJseS8yMDI0LzEwLzI0L2FqbnIuQTg0MDguYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 7. 7.D'Incerti L. MRI in neuronal ceroid lipofuscinosis. Neurol Sci 2000;21:S71–73 doi:10.1007/s100720070043 pmid:11073231 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1007/s100720070043&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=11073231&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 8. 8.Aydın K, Havali C, Kartal A, et al. MRI in CLN2 disease patients: subtle features that support an early diagnosis. Eur J Paediatr Neurol 2020;28:228–36 doi:10.1016/j.ejpn.2020.07.009 pmid:32855042 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/j.ejpn.2020.07.009&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=32855042&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 9. 9.Vanhanen SL, Raininko R, Santavuori P, et al. MRI evaluation of the brain in infantile neuronal ceroid-lipofuscinosis., Part 1: postmortem MRI with histopathologic correlation. J Child Neurol 1995;10:438–43 doi:10.1177/088307389501000603 pmid:8576552 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1177/088307389501000603&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=8576552&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1995TD66300003&link_type=ISI) 10. 10.Biswas A, Krishnan P, Amirabadi A, et al. Expanding the neuroimaging phenotype of neuronal ceroid lipofuscinoses. AJNR Am J Neuroradiol 2020;41:1930–36 doi:10.3174/ajnr.A6726 pmid:32855186 [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpuciI7czo1OiJyZXNpZCI7czoxMDoiNDEvMTAvMTkzMCI7czo0OiJhdG9tIjtzOjM4OiIvYWpuci9lYXJseS8yMDI0LzEwLzI0L2FqbnIuQTg0MDguYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 11. 11.Vanhanen SL, Raininko R, Autti T, et al. MRI evaluation of the brain in infantile neuronal ceroid-lipofuscinosis. Part 2: MRI findings in 21 patients. J Child Neurol 1995;10:444–50 doi:10.1177/088307389501000604 pmid:8576553 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1177/088307389501000604&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=8576553&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1995TD66300004&link_type=ISI) 12. 12.Autti T, Joensuu R, Aberg L. Decreased T2 signal in the thalami may be a sign of lysosomal storage disease. Neuroradiology 2007;49:571–78 doi:10.1007/s00234-007-0220-6 pmid:17334752 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1007/s00234-007-0220-6&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=17334752&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=000247656500007&link_type=ISI) 13. 13.Chang M, Cooper JD, Sleat DE, et al. Intraventricular enzyme replacement improves disease phenotypes in a mouse model of late infantile neuronal ceroid lipofuscinosis. Mol Ther 2008;16:649–56 doi:10.1038/mt.2008.9 pmid:18362923 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1038/mt.2008.9&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=18362923&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=000254929600007&link_type=ISI) 14. 14.Vuillemenot BR, Kennedy D, Cooper JD, et al. Nonclinical evaluation of CNS-administered TPP1 enzyme replacement in canine CLN2 neuronal ceroid lipofuscinosis. Mol Genet Metab 2015;114:281–93 doi:10.1016/j.ymgme.2014.09.004 pmid:25257657 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/j.ymgme.2014.09.004&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=25257657&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 15. 15.Schulz A, Specchio N, de Los Reyes E, et al. Safety and efficacy of cerliponase alfa in children with neuronal ceroid lipofuscinosis type 2 (CLN2 disease): an open-label extension study. Lancet Neurol 2024;23:60–70 doi:10.1016/S1474-4422(23)00384-8 pmid:38101904 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/S1474-4422(23)00384-8&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=38101904&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 16. 16.Batten Disease Family Association. [http://www.bdfa-uk.org.uk/research-and-resources/brineura-for-cln2/-](http://www.bdfa-uk.org.uk/research-and-resources/brineura-for-cln2/-). Accessed February 25, 2024 17. 17.Dyke JP, Sondhi D, Voss HU, et al. Brain region-specific degeneration with disease progression in late infantile neuronal ceroid lipofuscinosis (CLN2 disease). AJNR Am J Neuroradiol 2016;37:1160–69 doi:10.3174/ajnr.A4669 pmid:26822727 [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpuciI7czo1OiJyZXNpZCI7czo5OiIzNy82LzExNjAiO3M6NDoiYXRvbSI7czozODoiL2FqbnIvZWFybHkvMjAyNC8xMC8yNC9ham5yLkE4NDA4LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 18. 18.Hochstein JN, Schulz A, Nickel M, et al. Natural history of MRI brain volumes in patients with neuronal ceroid lipofuscinosis 3: a sensitive imaging biomarker. Neuroradiology 2022;64:2059–67 doi:10.1007/s00234-022-02988-9 pmid:35699772 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1007/s00234-022-02988-9&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=35699772&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 19. 19.Knoernschild K, Johnson HJ, Schroeder KE, et al. Magnetic resonance brain volumetry biomarkers of CLN2 Batten disease identified with miniswine model. Sci Rep 2023;13:5146 doi:10.1038/s41598-023-32071-z pmid:36991106 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1038/s41598-023-32071-z&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=36991106&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 20. 20.Fischl B. FreeSurfer. Neuroimage 2012;62:774–81 doi:10.1016/j.neuroimage.2012.01.021 pmid:22248573 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/j.neuroimage.2012.01.021&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=22248573&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=000306390600031&link_type=ISI) 21. 21.Bates D, Mächler M, Bolker B, et al. Fitting linear mixed-effects models using lme4. J Stat Soft 2015;67:1–48 doi:10.18637/jss.v067.i01 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.18637/jss.v067.i01&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=23757445&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 22. 22.Kim S. ppcor: an r package for a fast calculation to semi-partial correlation coefficients. Commun Stat Appl Methods 2015;22:665–74 doi:10.5351/CSAM.2015.22.6.665 pmid:26688802 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.5351/CSAM.2015.22.6.665&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=26688802&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 23. 23.Nickel M, Simonati A, Jacoby D, et al. Disease characteristics and progression in patients with late-infantile neuronal ceroid lipofuscinosis type 2 (CLN2) disease: an observational cohort study. Lancet Child Adolesc Health 2018;2:582–90 doi:10.1016/S2352-4642(18)30179-2 pmid:30119717 [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/S2352-4642(18)30179-2&link_type=DOI) [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=30119717&link_type=MED&atom=%2Fajnr%2Fearly%2F2024%2F10%2F24%2Fajnr.A8408.atom) 24. 24.Baker EH, Levin SW, Zhang Z, et al. MRI brain volume measurements in infantile neuronal ceroid lipofuscinosis. AJNR Am J Neuroradiol 2017;38:376–82 doi:10.3174/ajnr.A4978 pmid:27765741 [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpuciI7czo1OiJyZXNpZCI7czo4OiIzOC8yLzM3NiI7czo0OiJhdG9tIjtzOjM4OiIvYWpuci9lYXJseS8yMDI0LzEwLzI0L2FqbnIuQTg0MDguYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 25. 25.Chow N, Hwang KS, Hurtz S, et al; Alzheimer’s Disease Neuroimaging Initiative. Comparing 3T and 1.5T MRI for mapping hippocampal atrophy in the Alzheimer’s Disease Neuroimaging Initiative. AJNR Am J Neuroradiol 2015;36:653–60 doi:10.3174/ajnr.A4228 pmid:25614473 [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpuciI7czo1OiJyZXNpZCI7czo4OiIzNi80LzY1MyI7czo0OiJhdG9tIjtzOjM4OiIvYWpuci9lYXJseS8yMDI0LzEwLzI0L2FqbnIuQTg0MDguYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) * Received March 14, 2024. * Accepted after revision June 26, 2024. * © 2024 by American Journal of Neuroradiology [1]: /embed/graphic-2.gif [2]: /embed/graphic-3.gif