Table 1 Data extraction table

Lober RM, 2014 [8]

Retrospective

Using DWI to stratify DIPG subsets with distinct clinical behavior

• n = 1 histone mutant / 20 DIPG

• 14 yo

• M

G3, G4 (NOS for target patient)

Pons/periventricular nodule

H3-K27M

T1wC+, DWI

• Periventricular nodular enhancement

• Baseline ADC high = 1393. Follow up ADC mixed = 1172 & 1875. Periventricular lesion: 728

Ishibashi K, 2016 [9]

Case report

To report a pediatric midline glioma with H3F3A K27M mutation before and after malignant transformation

• n = 1

• 14 yo

• M

G2 on initial biopsy; G3 on resection

Thalamus

H3-K27M

FLAIR, T1wC+

• Non-enhancing

• Tumor dissemination at 12 months after diagnosis

Aboian MS, 2017 [10]

Retrospective

Characterize imaging features of DMGs with H3 K27M mutation and determine any specific imaging features correlate with histone mutation

• n = 24 histone mutant /n = 33 DMGs

• Mean age: 108.5 months

• 17 M / 7F

Not stated

Pontine/cerebellar peduncle (11), vermis (4), subcallosal (1), thalamic (6), cervical spine (2)

H3-K27M

T2w, FLAIR, T1w, T1wC+, DWI

• 15/24: cystic components or necrosis, 4/24: oedema

• 16/24 (67%) enhanced “heterogenous or uniform”

• 5/24 multifocal, 18/24 infiltrative, 4/24 mass effect, 18/24 irregular border, 6/24 CSF based metastases, 9/24 direct cortical invasion

• No difference between imaging characteristics of histone mutant and wild type gliomas

Yoshimoto K, 2017 [11]

Case series

Examine the prevalence and clinicopathological features of H3F3A and G34R/V mutant high grade gliomas

• n = 14 histone mutant (n = 10 K27M, n = 4 G34R)

• Mean age G34R: 10.5 yo; K27M: 15 yo

• G34R: 2 M:2F; K27M: 6 M:4F

Not stated

G34R: parietal, multiple lobes, thalamus

H3-K27M, G34R and G34V

T2w, FLAIR, T1w, T1wC+, DWI

• Hyperintense

• Absent or faint then marked enhancement (2/4 cases) after tumour progression

• G34R: case 1: gliomatosis cerebri growth pattern. Case 2: high density on CT + intratumoral calcification.Central restricted diffusion. Case 3: low density on CT + intratumoral calcification, indefinite tumour margins on MRI

Lopez GY, 2017 [12]

Case series

Describing 2 cases of DMG with mosaic expression of H3.3 K27M mutant protein and its implications with regards to classification and grading.

• n = 2

• 30 yo and 69 yo

• 1 M and 1F

1) G4

2) G3

Bilateral thalamus

H3 K27M mutant, IDH wild type, ATRX (one lost, another retained)

T1wC+

• Heterogeneously enhancing (both)

Vettermann FJ, 2017 [13]

Retrospective

Whether malignant tumor progression as observed for IDH-mutated gliomas can also be found in K27M midline gliomas

• n = 14

• Median age:21 yo

• 8 M, 6F

G2–4

Thalamus (5), brain stem (4), spinal cord (2), mesial temporal (2), cerebellum (1)

H3K27M, ATRX, IDH, FGFR1

T2w, T1w, T1wC+

• 2/14 showed contrast enhancement on follow up, during progression, after 13.1 (case 2) and 3.8 m (case 10).

• FET-PET: Maximal tumor-to-background ratio (TBR) - 6/8 with follow up scans ‘significant’ increase in TBR (> 20%)

Gilbert AR, 2018 [14]

Case report

A case of pineal region DIPG in a child (2nd case of pineal region DIPG overall)

• n = 1

• 12 yo

• F

G4

Pineal gland

H3-K27M

FLAIR, T1w, T1wC+, MRS

• Hyperintense

• Heterogeneous and rim enhancing, with posterior fossa and spinal CSF dissemination on follow up

• Hypodense on CT / hydrocephalus.

• MRS: suppressed NAA peak and elevated choline, lipid, and lactate peaks

Vettermann FJ, 2018 [15]

Case series

Characterize the imaging features of H3-G34 mutant gliomas using MRI and 18F-FET PET.

• n = 8

• Median age: 27 yo

• 2 M, 6F

G3 (n = 2), G4 (n = 5), 1: NOS

Multifocal (2), rest: lobar/ thalamus/ truncus corporis callosi

H3-G34R (all)

T2w and/or FLAIR, 3D T1w, 3D T1wC+

• 4/8 cystic components, 0/8 significant perifocal oedema

• 5/8 heterogeneous enhancement

• 1/8 hemorrhage and calcification.

• 8/8 high uptake intensity on 18F-FET PET with a median TBRmax of 3.4

• Dynamic PET (7/8) short median TTP min = 12.5 minutes

D’Amico RS, 2018 [16]

Case report

Present clinical and pathological features of pineal region GBM

• n = 1 histone mutant / 8 GBM

• 38 yo

• M

G4

Pineal region

H3-K27M, IDH-wild, lost ATRX

FLAIR, DWI, T1w, T1wC+

• Enhancing

• Diffuse seeding of the ventricular system with enhancing tumor

• Hydrocephalus

Daoud EV, 2018 [17]

Case series

Characterize adult brainstem gliomas, focusing on the H3-K27M radiologic features and clinical outcome

• n = 7 histone mutant

• Median age: 41 yo

• 6 M and 1F

Low grade (n = 2), high grade (n = 5)

Midbrain tectum (1), midbrain tegmentum (1), pons (2), medulla (1), cerebellar peduncle/pons (2)

H3-K27M, ATRX-retained, p53

FLAIR, T1C+

• 1/7: T2/FLAIR hyperintense

• 2/7 exophytic, 3/7 hydrocephalus

• 6/7 enhancing: 2/7 heterogeneous, 1/7 minimal amorphous, 1/7 focal, 1/7 poor irregular, 1/7 “enhancing”

• No correlation between contrast enhancement and H3-K27M status, p = 0.1

Dormegny L, 2018 [18]

Case report

Improve accuracy of spinal cord biopsies and analyzing clinical, radiological and surgical features

• n = 1

• 21 yo

• M

G4

Dorsal thoracic cord T9–11

H3-K27M, IDH (wild-type)

T2w, T1w, T1wC+

• T2 hyperintense; 3 weeks later (no intervention) - heterogeneous

• Loss of T2: intralesional hemorrhage

• Initial - localized contrast uptake at T10; 3 weeks later (no intervention) - multiple areas of extended contrast uptake form T9 to conus (T12-L1), with nodular uptake in cauda equina

Gao Y, 2018 [19]

Case report

Report a case of DMG with histone H3-K27M mutation which simultaneously showed PNET-like appearance

• n = 1

• 51 yo

• F

Not stated

Cervical cord C5–7

H3-K27M, IDH wild type,

ATRX retained

T2w, T1w, T1wC+

• Hyperintense

• Mild enhancement

Puntonet J, 2018 [20]

Retrospective

Correlate the histological and radiological features of G34R mutant high-grade gliomas

• n = 12

• Mean age: 16 yo (range = 6–31 yo),

• 9 M and 3F

G4 (6), G3 (4),

HGG NOS (2)

All supratentorial, predominant temporal lobe (7), basal ganglia (4), all leptomeningeal contact, meningeal invasion (1), ependymal contact (9)

All H3K27M, G34R, GFAP +ve and BRAFv600e wild-type. ATRX lost (9). P53 + ve (9).

T2, FLAIR, T1wC+ (11 patients), DWI (4), PWI (2), T2*w (6)

• T2 hyperintense (7), T2 isointense (5), T2* hypointense (4/6).

• Enhancement: 4/11 intense, 5/11 minimal to moderate, 2/11 no enhancement. 3/12 nodular, 3/12 patchy, 3/12 serpiginous, 3/12 ring-like

• 3/3 restricted diffusion (ADC ~ 4800.10 − 6 mm2/s)

• 8/10 precontrast T1 hyperintensity, 4/12 no necrosis, 4/12 minimal necrosis, 4/12 high necrosis.

• 6/11 no oedema, 4/11 minimal oedema, 1/11 moderate oedema.

• 2/2 ASL-PWI hyper perfusion (relative CBF: 1.74–2.85).

• 2/2 CT hyperdense

Jung JS, 2019 [21]

Retrospective

Evaluate the imaging characteristics of DMG with H3 K27M mutation in the spinal cord + evaluate predicting the presence of H3 K27M using a machine learning–based classification model.

• n = 24 histone mutant / n = 41 spinal cord gliomas

• Mean age: 35.5 yo

• 17 M / 7F

Not stated

Axial location (18/24 central, 6/24 eccentric); Longitudinal location (8/42 cervical, 14/24 thoracic, 2/24 lumbar)

H3-K27M

T2w, T1w, T1wC+, GRE T2*w

• 21/24 hyperintense, 3/24 isointense, 1/24 hypointense; 14/24 edema

• 8/24 diffuse enhancement, 9/24 partial, 7/24 none; 5/17 irregular rim enhancement

• 6/24 hemorrhage, 4/24 necrosis, 3/24 tumoral cyst, 1/24 syringohydromyelia

• 16/24 ill-defined margin

• 10/24 CSF-based tumor spread

• Significant difference in tumoral hemorrhage between H3-K27M and wildtype, p = 0.003

• RF model: sensitivity = 45.8%, specificity = 88.2%, AUC = 0.63, 95% CI (0.45–0.80), accuracy = 63.4%

Qiu T, 2019 [22]

Retrospective

To summarize the imaging characteristics of adult H3 K27M-mutant gliomas

• n = 66 histone mutant

• Age range: 20–60 yo

• 40 M and 26F

Not stated

Thalamus (38), brainstem (6) (2 pons, 4 medulla, all dorsal); cerebellum (2); thalamus (2), whole brain (8), corpus callosum (3), hypothalamus (1), spinal cord (4) (1 in each of cervical, cervicothoracic, thoracic, and lumbar cord), cerebral hemispheres (2)

H3-K27M

T2w, FLAIR, T1w, T1wC+ (in 61/66), DWI

• 10/66 peritumoral oedema

• 11/61 none enhancing, 25/61 partial enhancement, 25/61 diffuse (≥ 50% of the whole lesion)

• 1/66 hemorrhage; all (4/4) spinal cord lesions spanned 3 or more segments

• 8/66 leptomeningeal dissemination

• 66/66 unremarkable or moderate restriction.

He P, 2019 [23]

Case report

A rare H3 K27M–mutant glioblastoma in the hypothalamus.

• n = 1

• 27 yo

• F

Not stated

Hypothalamus, intrasellar, suprasellar

H3-K27M, ATRX &p53& olig2. IDH wild-type

T2w, T1w, T1wC+

• T2 hyperintense

• CT: isodense solid mass, local recurrent mass, high rCBV (not mentioned values)

• Strongly heterogeneous enhanced solid lesion and nonenhanced cystic lesion

• Local recurrent mass: low ADC (1.2806)

Chanchotisatien A, 2019 [24]

Case report

Presenting a slow- growing thalamic glioma with H3-K27M mutation.

• n = 1

• 39 yo

• M

G2

Thalamus

H3-K27M, ATRX, IDH-wild

T2w, FLAIR, T1wC+

• T2 hyperintense

• No enhancement on followup

Chen H, 2019 [25]

Retrospective

Noninvasively identify MRI markers predictive of H3 K27 M mutational status in diffuse midline tumors.

• n = 19 / n = 38 DMGs

• Mean age: 27 yo

• 11 M and 8F

G2 (n = 4), G3 (n = 5), G4 (n = 10)

Thalamus (12), midbrain (1), pons (2), corpus callosum (2), cerebellum (1), hypothalamus (1),

H3-K27M

T2w, FLAIR, T1w, T1wC+, DWI (in 31/38)

• Shape: 12/19 round/oval, 7/19 irregular

• 4/19 solid, 15/19 cystic; 3/19 hemorrhage, 7/19 edema, multifocal 9/19

• Tumour margin: 8/19 sharp, 11/19 blurred

• 6/19 intra-tumoral rim enhancement, 1/19 marginal, 4/19 heterogeneous, 8/19 none or minimal

• DWI in H3-K27M:

• Minimal ADC = 0.734 ± 0.120, Ratio of minimal ADC = 0.972 ± 0.165,

• Peritumoral ADC = 0.937 ± 0.156, Ratio of peritumoral ADC = 1.240 ± 0.232.

• Significant difference in Histone mutant vs wild type in: Younger age (p = 0.009), Minimal ADC (p = 0.020), Peritumoral ADC (p = 0.018), Ratio of minimal ADC (p = 0.018), Ratio of peritumoral ADC (p = 0.013)

• Thresholds: Minimal ADC 0.728, Ratio of minimal ADC 0.982; Peritumoral ADC 1.004, Ratio of peritumoral ADC 1.248

• Combination of DWI parameters: sensitivity = 88.89%, specificity = 76.92%, AUC = 0.872 (0.75–1.00)

Aboian MS, 2019 [26]

Retrospective

Identify differences in imaging diffusion characteristics between H3- K27M mutant and wild-type DMGs

• n = 23

• Mean age: 8.9 yo

• 14 M and 9F

Not stated

Pons/vermis/4th ventricle (17), thalamus (5), subcallosal (1)

H3-K27M

DWI and ADC map values, FLAIR used for ROI registration

• No significant differences in ADC mean, median, minimum, and percentile values between histone mutant and wild-type gliomas

Miyazaki T, 2019 [27]

Case report

Present first case of DMG with H3-K27M mutation in a pregnant woman followed by fatal hemorrhage during the postpartum period.

• n = 1

• 26 yo

• F (pregnant)

Not stated

Thalamus/midbrain

H3K27M, IDH-wild

T2w, FLAIR T1w, T2*w, T1wC+, DWI, MRA

• T2 hyperintense, poorly circumscribed

• not enhanced

• CT: isodense, hydrocephalus

Karlowee V, 2019 [28]

Retrospective

Analyze the EZH2 expression level and revealed its association with the poor survival of patients with H3K27M mutant-positive tumors.

• n = 12

• Age range: 6–56 yo

• 8 M and 4F

Not stated

Distant recurrence (9), thalamus (9), pons (1)

H3-K27M, ATRX (4), p53 (8), EZH2 (9)

T2w, FLAIR, T1w, T1wC+, DWI

• 12/12 hyperintense, 2/12 cyst formation

• Enhancement: 5/12 partly, 2/12 homogeneously, 3/12 heterogeneously, 2/12 none

• 4/12 intratumoral hemorrhage, 9/12 dissemination + distant recurrence

• 10/12 high DWI, 1/12 iso, 1/12 low DWI

• No relationship between immunohistochemical staining results and clinical or imaging characteristics.

Giagnacovo M, 2020 [29]

Retrospective

Assess consistency between DIPG histo-molecular findings and clinical-radiological features.

• n = 19 / n = 22 DIPGs

• Mean age: 8 yo

• 10 M and 9F

G4 (n = 9), G3 (n = 8), G2 (n = 2)

Pons (at least 50% and causing expansion)

H3F3A/HIST1H3B K27M

T2w, FLAIR, T1w, T1wC+

+/− DWI, PWI, DTI, MRS

• Unsharp margins and patchy enhancement in one image for histone mutant

Chiang J, 2020 [30]

Retrospective

Identify variables that correlated with the clinical diagnosis of aDIPG (atypical) and evaluate the consistency of radiographic diagnosis of aDIPG

• n = 12 histone mutant / n = 33 aDIPGs,

• Median age: 9 yo

• 6 M and 6F

G1–4

Extra pontine extension (3), eccentric within the pons (7)

H3F3A/HIST1H3B K27M, TP53

T2w, T1w, T1wC+, DWI +/− FLAIR, T2wC+, SWI/T2*w

• 3/12 ring enhancement

• K27M: 10/12 ill-defined tumor margin, 7/12 eccentricity within pons, 3/12 extra pontine extension

• No significant difference in imaging (or clinical) features of K27M vs wild-type

Garibotto F, 2020 [31]

Retrospective

Compare the clinical behavior of DMGs H3 K27M- mutant and non-histone mutant midline HGGs in NF1 vs. non-syndromic children and to report imaging features of NF1 HGGs.

• n = 1 H3-K27M/ n = 2 NF1 DMGs; n = 11 H3-K27M / n = 16 Non-NF1 DMGs

• Age: NF1 - 11yo; Non-NF1 - range 3-16yo

• NF1: 1 F; Non-NF1: 4 M, 7F

G4 (all)

NF1: pons/midbrain (1); Non-NF1: thalamus (4), pons (4), medulla (1), diencephalon-mesencephalon junction (2)

H3.3 (H3F3A) K27M

T2w, FLAIR, T1w, T1wC+, DWI, FLAIRC+, (MRS in 2 described cases)

• One described case (NF1 H3K27M lesion):

• Irregular rim enhancement, adjacent expansile region without enhancement

• Irregular central necrotic area in ponto-mesencephalic portion, adjacent expansile portion no necrosis

• reduced diffusivity in ponto-mesencephalic portion (minimum absolute ADC value: 0.69 × 10⁻³ mm²/s)

• Leptomeningeal dissemination

• Single voxel MRS: Cho/NAA peak-height ratio: 3.44 and Cho/Cr ratio: 2.91

Tu JH, 2020 [32]

Case report

Report a case of H3 K27M mutant diffuse midline glioma with cartilaginous metaplasia

• n = 1

• 56 yo

• F

Not stated

Medulla

H3-K27M, GFAP, IDH-wild, lost ATRX, negative BRAFV600E

T2w, T1w, T1wC+

• T2 heterogeneous, cystic degeneration

• Ring enhancement

• CT: irregular calcification

Fujioka Y, 2020 [33]

Case report

Report a case of DMG, H3 K27M mutant that mimicked a hemispheric malignant glioma in an elderly patient.

• n = 1

• 66 yo

• F

Not stated

Bilateral thalamus, left hippocampus, and fronto-parietal lobes

H3-K27M, IDH wild-type, negative BRAFV600E, retained ATRX, GFAP

FLAIR, T1wC+

• Hyperintense

• Multiple isolated enhancing lesions. Enhancement disseminated to lateral ventricles

Cheng Y, 2020 [34]

Case series

Examine the prevalence and clinic pathological features of H3F3A and G34R/V mutant HGGs

• n = 3

• 15 yo, 15 yo, 28 yo

• 2 M and 1F

Not stated

Cases 1 and 2:frontal, case 3: temporal

1) & 3) H3.3 G34V,

2) H3.3 G34R,

T2w

• Variable, homogenous / heterogenous high

• Enhancement not stated

• Local recurrence in all cases (after 10 or 5 months of 1st operation)

Baroni LV, 2020 [35]

Case series

Report three brainstem tumors with an initial indolent course that later developed classical imaging and clinical features of DIPG.

• n = 3

• 3 yo, 11 months, 6 yo

• 2 M and 1F

1) Not stated,

2) G4,

3) G4

1) ponto-medullary junction, 2) medulla/pons, 3) brainstem and the pontocerebellar angle

all H3.3 K27M, 2 cases with TP 53

FLAIR, T2w

• All cases: hyperintense

• 1) Non-enhancing, 2) non-enhancing, 3) not stated

• All cases show increase in size in follow up (range 11 months to 5 years)

Babarczy K, 2020 [36]

Case report

Report one of the oldest patients having so far been reported with an immunohistochemically confirmed DMG, H3 K27M-mutant

• n = 1

• 73 yo

• F

G4

Cervical cord (C2) extending to medulla, pons, cerebral peduncles, internal capsules bilaterally and right pallidum

H3-K27M, GFAP positive, IDH-wild

T2w, FLAIR, T1w, T1wC+, DWI, MRA, STIR (spine), MRS

• FLAIR hyperintense

• No enhancement

• No diffusion restriction

• MRA: normal

• MRS: high metabolism (no values)

Lu VM, 2020 [37]

Case report

Illustrate H3 K27M mutation occurring in cortically-based diffuse gliomas not midline structures and discuss the uncertainties regarding grading and prognostic classification for such tumours.

• n = 1

• 9 yo

• F

Not stated

Pons

H3-K27M

T1wC+

• Not stated but enhancement highlighted in image

• Progression in 2 months time (intracranial/intraspinal metastases)

Su X, 2020 [38]

Retrospective

Investigate the feasibility of predicting H3 K27M mutation status using an automated machine learning

• n = 40 histone mutant / n = 100 midline gliomas

• Mean age: 23.6 yo

• 14 M and 26F

G4

Midline (not otherwise specified)

H3-K27M

T2w, FLAIR, T1w, T1wC+

• 10 important features, including 3 shape, 3 first-order, 2 GLSZM, 1 GLCM, and 1 GLDM, were included.

• Tumor shape features important in predicting H3 K27M mutation, maximum 2D diameter of the slice of H3 K27M–mutant tumors was smaller than that of wild-type (41.31 vs 59.35, P = 0.007)

• Sensitivity range of 10 models: 0.55–0.73

• Specificity range of 10 models: 0.57–0.93

• AUC range for 10 models (0.73–0.90)

Chiba K, 2020 [39]

Retrospective

Investigate the correlation between the original site of thalamic gliomas and patients’ clinical outcomes retrospectively and to determine appropriate treatment strategies.

• n = 4 Histone mutant / n = 10 pediatric thalamic gliomas

• Age range: 8–17 yo

• 2 M, 2F

G3–4

All thalamopulvinar (TP)

H3-K27M

T2w, T1w, T1wC+, DWI (2/4), methionine-PET (2/4), DTI (1/4)

• T2 hyperintense

• 1/4 heterogeneous enhancement, 1/4 mixed faint and homogeneous enhancement, 2/4 faint enhancement

• 2/4 high DWI

• The presence of H3 K27M mutation and TP location were closely related to each other (p = 0.0036)

Rodriguez Gutierrez D, 2020 [40]

Retrospective

To correlate imaging characteristics and outcome measures of pediatric patients with newly diagnosed non-brainstem HGG with pathologic and molecular data

• n = 23 H3K27M, n = 7 H3G34R / 113 gliomas

• H3K27M: mean 12.1 yo; H3G34R: mean 13.1 yo;

H3K27M: 11 M and 12F; H3G34R: 3 M and 4F

G3 (n = 21), G4 (n = 92)

Cerebral hemispheres and midline

H3F3a K27M, G34R

T2w, FLAIR, T1w, T1wC+

• Perilesional oedema: H3 K27M (10/23 none, 13/23 minor), H3 G34 (1/7 none, 3/7 minor, 2/7 moderate, 1/7 severe)

• Necrosis: H3 K27M (16/23 yes, 6/23 no), H3G34 (4/7 yes, 2/7 no)

• Enhancement: H3 K27M (14/23 strong, 1/23 moderate, 7/23 minor/none), H3 G34 (1/7 strong, 1/7 moderate, 4/7 minor/none)

• Hemorrhage: H3 K27M (7/23 yes, 15/23 no), H3 G34 (4/7 yes, 2/7 no)

• Tumor definition: H3 K27M (19/23 well-defined, 5/23 ill-defined/diffuse), H3 G34R (2 well-defined, 5 ill-defined/diffuse)

• Versus Midline wildtype, H3 K27M–mutant tumors showed more enhancement (P < 0.05) and older age

• ADC values were not significantly different between H3 K27M–mutant and Mid- line WT tumors or between H3G34R and cerebral wildtype.

• Versus cerebral wildtype, tumor definition for H3 G34R mutants was significantly different, p < 0.05

Kay MD, 2020 [41]

Case report

Report rare extra cranial metastases from glioblastoma with PNET-like components and demonstrate the utility of FDG PET/CT for revealing distant metastases from glioblastoma

• n = 1

• 17 yo

• F

G4

Temporal lobe

H3G34, IDH-wild

T2w, T1wC+, PET post resection

• T2 hyperintense

• Heterogeneous enhancement

• Presented with hematoma

• Post resection: invasion into the left greater wing of the sphenoid, leptomeningeal drop and osseous metastases

Onishi S, 2020 [42]

Case series

Describing radiological and immunostaining characteristics of H3.3 G34R-Mutant Glioma

• n = 3

• Ages: 13, 19 and 15 yo

1 M and 2F

2/3 glioblastoma, 1/3 HGG

1) frontal,

2) parietal,

3) parieto-occipital

All: H3.3 G34R mutation, IDH-wild BRAF-wildtype, retained 1p/19q

T2w, FLAIR, T1w, T1wC+, DWI, MRS, ASL-PWI

• 3/3 T2/FLAIR moderately hyperintense

• 1/3 poor enhancement 2/3 partial enhancement

• 3/3 DWI hyperintense (ADC range = 0.625–0.810)

• 3/3 mild edema.

• 2/3 low tumor blood flow on ASL

• 3/3: high choline peak, marked decrease in NAA peak, small lactate peak.

• 3/3 iso to hyperdense mass without calcification

Thust S, 2021 [43]

Retrospective

To assess anatomical and quantitative diffusion-weighted MR imaging features in H3 K27M histone-mutant diffuse midline glioma

• n = 15

• Median age: 19 yo

• 6 M and 9 F

G4

Midline brain

H3-K27M

T2w, FLAIR, T1w, T1wC+, DWI, DSC perfusion MRI (2/15)

• T2/FLAIR signal: 13/15 heterogeneous, 2/15 homogeneous. 2/15 with cysts

• Margin: 8/15 distinct, 7/15 indistinct

• Enhancement: 4/15 solid enhancement, 6/15 rim enhancing + necrosis, 3/15 none

• DWI:

• ADCmin in solid tumor = 0.84 (±0.15 SD) ADCmin/NAWM ratio = 1.097 (±0.149)

• ADCmean in solid tumor = 1.12 (±0.25) ADCmean/NAWM ratio = 1.466 (±0.299)

• Hemorrhage: 3/15 macro, 2/15 petechial

• Hydrocephalus: 6/15 requiring shunting

• 2/15 elevated rCBV ((5.9, 3.5)

• 18F-choline PET (1/15): tracer accumulation

• A significant difference (P = 0.01) between the 2nd centile of the volumetric ADC histogram and the ROI ADCmin values.

• No statistical significance between ROI ADCmin and the 5th and 10th histogram percentiles, or ROI ADCmin/NAWM ratio and the ADC5th percentile/NAWM ratio or ROI ADCmean and the histogram ADCmedian and ADCmean or ROI ADCmean/NAWM ratio and the histogram ADCmean/NAWM ratio

Picart T, 2021 [44]

Retrospective

To describe the characteristics of DHG H3G34-mutant in adults and to compare them to those of established types of adult WHO grade IV gliomas

• n = 17 H3G34R

• Mean age: 25 yo

• 11 M and 6F

Not stated

H3.3G34R: frontal (11), parietal (11), temporal (3), occipital (1), corpus callosum (3), midline (4); H3.3K27M: 33/33 midline, 2 with temporal extension

17/17 H3F3a G34R, 14/16 TP53 positive, 13/14 loss of ATRX

T2w, FLAIR, T1w, T1wC+,

MRS (9/17), DCE-PWI (8/17)

• H3.3 G34R: 9/16 poorly delineated/infiltrative, 3/16 Initial hemorrhage, 2/16 cyst, 1/16 necrosis; H3.3 K27M: 3/28 cyst, 10/28 necrosis

• H3.3 G34R: 4/15 enhancement, 6/15 faint enhancement, 5/15 none; H3.3 K27M: 8/29 enhancement

• H3.3 G34R: 12/13 restriction (8/13 focal); H3.3 K27M: 3/23 restriction

• 4/8 elevated rCBV (> 1.7). 4/9 elevated Choline/N-Acetyl-Aspartate ratio > 2 and/or of lipid/ lactate peaks

• H3 G34-mutant occurred in younger patients (p = .05), more frequently involved the parietal lobe (p < .001), more frequently presented as a hemorrhagic tumor at diagnosis (p = .03), less frequently demonstrated contrast enhancement (p = .05) and necrosis (p = .03), more frequently displayed ADC restriction (p < .001)

Cheng R, 2021 [45]

Case report

Report a pediatric patient with spinal cord H3 K27M-mutant DMG

• n = 1

• 7 yo

• F

G4

Cervical cord C2–7 intramedullary

H3-K27M

T2w, T1w, T1wC+

• T2 slightly hyperintense

• Heterogeneous enhancement

Li Q, 2021 [46]

Retrospective

MRI characteristics of brain DMG-histone mutant using radiomics

• n = 16 histone mutant / n = 30 DMGs,

• Median age: 25.5 yo

• 10 M and 6F

Not stated

10/16 thalamus, 6/16 brainstem

H3-K27M

T2w, T1w, T1wC+

• No detailed information. From single selected figure –T2 hyper intense

• Faint enhancement

• Cyst formation showed sig diff between H3K27M and WT (OR = 7.800, 95% CI 1.476–41.214; p = 0.024)

• No statistical significance in: necrosis (p = 0.191), hemorrhage (p = 0.657) and T1/T2 ratio (p = 0.689)

Kandemirli S, 2021 [47]

Retrospective

Machine learning to predict histone mutation

• n = 50 histone mutant / n = 59 histone wild type

• Median age: 10 yo mutant / 30.5 yo histone-wild

Not stated

Thalamus only (17), 12 center in thalamus (12), center in the pons (13), remaining posterior fossa structures (8)

H3K27M

T2w, T1w, T1wC+, FLAIR and ADC

• Median age in the H3K27M mutant group was significantly lower compared with the wild-type cohort

• H3K27M-mutant and wild-type tumors show no significant difference in tumor size, enhancement, internal necrotic changes, or infiltrative appearance

Kathrani N, 2022 [48]

Retrospective

Assess DWI and DSC-PWI to predict the H3K27M mutation status in DMGs non-invasively

• n = 48

• Mean age: 23 yo

• 21 M and 27F

G2–4

Thalamus (28), midbrain (6), pons (10), medulla (2), others (2)

H3-K27M

DWI, DSC (in 34/48)

• Peritumoral ADC = 1.1, nPT ADC = 1.64, min ADC = 0.76, nADC = 1.11

• rCBV = 25.17, nrCBV = 3.44, rCBF = 266.15, uncorrected nrCBV = 3.5

• Significantly lower PT ADC and nPT ADC and higher rCBV, nrCBV, rCBF, uncorrected nrCBV in histone mutant group

• Thalamic subgroup analysis: H3K27M showed significantly lower min ADC PT ADC, nADC, nPT ADC and significantly higher nrCBV, rCBF, nrCBF, uncorrected rCBV, uncorrected nrCBV, and corrected rCBV

Predicting histone mutation:

• PT ADC (cut-off = 1.245; AUC 0.6, Sn 47%, Sp 79%), nPT ADC (cut-off = 1.853; AUC = 0.6, Sn 52%, Sp 77%), nrCBV (cut-off = 1.83; AUC = 0.6, Sn 46%, Sp 76%) and uncorrected nrCBV (cut-off = 2.28; AUC = 0.7, Sn 61%, Sp 71%), In thalamic subgroup: nADC (cut-off = 1.129; AUC = 0.7, Sn 61%, Sp 75%), PT ADC (cut-off = 1.185; AUC = 0.65, Sn 68, 68%), nPT ADC (cut-off = 1.853; AUC = 0.65, Sn 64%, Sp 71%), nrCBV (cut-off = 1.83; AUC = 0.703, Sn 47%, Sp 85%), nrCBF (cut- off = 2.73; AUC = 0.74, Sn 71%, Sp 75%), and uncorrected nrCBV (cut- off = 2.27; AUC = 0.759, Sn 60%, Sp 86%)

Ikeda K, 2022 [49]

Retrospective

Establish high intensity on DWI in non-enhancing tumors (DWI-Gadolinium mismatch sign) as imaging biomarker for H3K27M DMG

• n = 6

• Median age: 23 yo (range 6–31)

• 4 M and 2F

Not stated

6/6 thalamus

H3-K27M, IDH-wild type

T2w, T2*, FLAIR, T1w, T1wC+, DWI

• T2/FLAIR hyperintense

• 6/6 enhancement

• 6/6 DWI high

• 5/6 DWI-Gd mismatch sign positive

• DWI-Gd mismatch sign and intratumoral bleeding present in thalamic gliomas than in pons with statistical significance (p = 0.046 and p = 0.0017)

Su X, 2022 [50]

Retrospective

Investigate the capacity of quantitative MRI in identifying the H3 K27M mutation status of DMG

• n = 23

• Age range = 6–47 yo

• 13 M and 10F

G4

Juvenile group: hemispheric near midline (1), diencephalon (2), brainstem (8)

adult group: diencephalon (3), brainstem (9)

H3K27M

T2, FLAIR, DWI, T1w, T1wC+, PWI, MRS, DTI

• Juvenile group: rADC_M = 1.56, rADC_15th = 1.15, rADC_25th = 1.31, rADC_50th = 1.50, rADC_75th = 1.68, rADC_max = 2.94, Ins/tCr = 1.11, NAA/tCr = 0.62, Cho/NAA = 0.96, Cho/rCr = 0.65, tNAA/tCr = 0.63, Glx/rCr = 0.89.

• Adult group: rADC_M = 1.5, rADC_15th = 1.07, rADC_25th = 1.15, rADC_50th = 1.33, rADC_75th = 1.60, rADC_max = 3.05, Ins/tCr = 0.80, NAA/tCr = 0.60, Cho/NAA = 1.11, Cho/rCr = 0.57, tNAA/tCr = 0.86, Glx/rCr = 0.97

• rADC_15th, rADC_25th, rADC_50th, and rADC_75th values were significantly lower in mutation group (p < 0.05)

• Ins/tCr values were lower in the juvenile (p = 0.003) and the adult (p = 0.025) mutation subgroups

• rMD_mean and rMD_25th/50th/75th values of the mutation group were significantly lower in the adult subgroup (p < 0.001)

• rCBV_mean/25th/50th/75th were slightly higher and the rTTP_mean/50th were slightly lower in the adult mutant subgroup; no significant differences in other PWI metrics

• No significant difference in tumor size (p > 0.05).

Hohm A, 2021 [51]

Retrospective

Describe and compare MRI of pediatric DMG with known H3 K27 mutation status including H3.1 and H3.3 K27M subgroups

• n = 52

• Age: 10.2y (range 1.25–17.75yo)

• 26 M and 26F

Not stated

19/52 thalamus/basal ganglia, 0 midbrain tectum, 27/52 pons, 5/52 spinal cord, 1 other (medulla)

H3.1 K27M & H3.3 K27M

T2w, T1w, T1wC+

• T2 hyperintense (44), homogeneous (3), predominantly homogeneous (17), predominantly inhomogeneous (23), inhomogeneous (4).

• Margins: well defined (3), moderately defined (23), ill-defined (21).

• Necrosis: yes (27); Edema: yes (8); CSF dissemination: (2)

• Enhancement: strong (19), intermediate (9), mild (7), no (12), predominantly inhomogeneous (18), inhomogeneous (10); ring enhancement (25)

• DWI: DWI hyperintense (22), restriction (5), not restricted (12)

• SWI: signal loss (7), hemorrhage: yes (18)

• H3.1 younger than H3.3 (p = 0.004)

• H3K27M more commonly arising in pons and thalamus/BG vs WT widely distributed (p = 0.001)

• H3 K27M more often T2 hyper intense (44/47, p = 0.02), T1 inhomogeneous (19/46, p = 0.02)

• Spinal tumors: There were no imaging characteristics that significantly differentiated the two molecular groups

Kim H, 2021 [52]

Retrospective/ one histone mutant case

Report 13 MMRD-associated (9 sporadic and 4 Lynch syndrome) primary brain tumors to determine clinic pathological and molecular characteristics

• n = 1

• 11 yo

• M

G4

Thalamus (right)

H3-K27M, lost ATRX, IDH-wild type

T2w, T1w, FLAIR, T1wC+ (from images)

• Multiple enhancing solid and cystic masses

Kurokawa R, 2022 [53]

Case series

Review the demographic, clinical, and neuroradiological features of DHGs-G34m in 3 original cases

• n = 3

• Ages: 16, 22 and 19 yo

• F/M/F

2/3 G4

1/3 frontal, 2/3 parietal

3/3 H3 G34R, 2/3 p53, 3/3 ATRX, 3/3 IDH wild type, BRAF -ve GFAP +ve

CT, T2w, FLAIR T1w, T2*W, T1wC+, DWI, SWI, PWI

• 3/3 T2 hyperintense, 2/3 leptomeningeal contact laterally

• 2/3 patchy enhancement, 1/3 heterogeneous enhancement

• 3/3 restricted diffusion, ADC values range = 0.53–0.8

• 2/3 hyperdense on unenhanced CT without calcification; 3/3 intratumoral hemorrhage on T2*; 1/3 elevated CBF in enhancing portion

Cheng L, 2021 [54]

Retrospective

Describe the clinical and radiological characteristics of primary spinal H3 K27M-mutant DMG and compare with the H3 K27 wild-type

• n = 28

• Age: 28.7 yo

• 19 M and 9F

G1 (0), G2 (9), G3 (10), G4 (9)

Cervical (7), cervicothoracic (4),

thoracic (11), thoracolumbar (5), holocord (1). Median involved segments = 3

H3K27M

p53 (20) ATRX loss (8),

Ki-67 >/= 20% (18), IDH-wild type (28)

T2w, T1wC+

• Enhancement: partial (12), diffuse (13); pial enhancement (24)

• Ill-defined margin (25)

• Edema (16), hemorrhage (4), cysts (4), necrosis (10), syrinx (6)

• Fewer H3K27M have syrinx vs WT (p = 0.017)

• No other imaging features showed statistical significance