Table 1.

Brain metastases from breast cancer MRI findings.

First author (year)	Number of Patients	MRI Technique(s)	Anatomical Location(s)	MRI Findings
Xue (2023) 20	4	T1‐w and T2‐w	Superficial parenchyma lesions and deep lesions	Following LITT, the majority of tumours exhibited an increase in volume on the MRI scan taken after 30 days. Additionally, if the EVR is greater than 40% on the 30‐day MRI, it could suggest a possibility of tumour recurrence at a later stage
Young (2023) 21	34	CE MRI	Frontal, temporal, perietal, occipital and cerebellum	The HER2 status of BC BMs was significantly associated with lesion contour and composition on MRI
Reibelt (2022) 22	15	3D T1‐w MPRAGE	L. cerebellum WM, R. pallidum, L. thalamus, L. choroid plexus, L. Lat. ventricle and total GM	Subcortical volume changes after radiotherapy are a sensitive indicator of neuroanatomical modifications and brain atrophy
Young (2021) 23	38	T1‐w and T1‐w post	Frontal, temporal, perietal, occipital and cerebellum	The prospective validation of MRI enhancement for detecting HER2 overexpression in BC BMs is warranted
Santos (2020) 24	147	T1‐w, T1‐w post, T2‐w, DWI and T2*‐w	Hippocampal and hydrocephalus	The prognosis of the disease and certain imaging features of the BM can be predicted based on the BC subtype, but not their distribution
Zhang (2019) 25	3	T1‐w, T1‐w post, T2‐w, DWI, ADC and FLAIR	NA	ADC‐based texture analysis can be used to distinguish between solitary BM and GBM, and it is recommended that the ROI be placed on the solid portion when calculating the ADC‐based texture metrics
Mayinger (2019) 26	851	3D T1‐w	Bilateral and central/median subcortical structures	The size of the brain's substructure varied significantly between the patients The volumetric size of the fourth ventricle could be a useful diagnostic marker in the future
Kniep (2019) 27	37	T1‐w, T1‐w post and FLAIR	NA	Using a machine‐learning classifier trained on quantitative features of routine MR images of the brain, we found high discriminatory accuracy in predicting the tumour type of BMs
Ortiz‐Ramón (2018) 28	17	3D T1‐w	NA	After image quantisation with the appropriate number of grey levels, the texture features of volumetric MRI can be used to distinguish between BMs and different primary cancers
Skogen (2018) 29	5	DTI	Intracranial lesions without visible haemorrhage, multiple lesions and infratentorial lesions	Radiologists can differentiate between GBM and single BM by evaluating the heterogeneity around the tumour By having a better idea of the tumour's severity, doctors can better prepare for surgical removal and care for the patient afterwards
Muto (2018) 30	13	DSC	Temporal lobe	DSC is a clinically useful method for distinguishing between tumour recurrence, tumour necrosis and pseudoprogression in patients with cerebral metastases A cut‐off value of 2.1 for rCBV was found to be the most accurate and consistent measurement
Kyeong (2017) 31	100	3D T1‐w	HER2‐positive type: occipital, temporal lobes and cerebellum / Luminal type: frontal, occipital lobes and cerebellum	Hot spots of triple‐negative metastases are evenly distributed in the brain, whereas BMs of HER2 positive and luminal type occur predominantly in the occipital lobe and cerebellum When BMs from the TNBC were compared with other types, they were found more frequently in the frontal lobe, limbic region and parietal lobe Different types of BC often show a variety of BM distribution patterns
Kesler (2017) 32	74	rsfMRI, T2*‐w GRE, HR3D IR, FSPGR GRE T1‐w and DTI	R. inferior parietal lobe, R. middle inferior orbital frontal gyrus, R. medial superior frontal gyrus, R. inferior and middle frontal gyri, bilateral postcentral gyri, R. precuneus, L. inferior temporal gyrus, L. middle occipital gyrus, R. parietal lobule, R. cuneus, R. superior temporal gyrus and R. inferior temporal gyrus.	For the BC group, structural and functional clustering was found to be significantly inversely related, while functional clustering was found to be significantly positively related to the Hurst exponent a Greater cognitive impairment was associated with greater overlap between structural and functional connectome clustering The significance of structural and functional connectome properties is a potential biomarker of general neurologic deficit
Bette (2017) 33	13	T2‐FLAIR	NA	With previously resected BMs, the FLAIR signal intensity of the fluid inside the resection cavity increases
Fan (2017) 34	13	T1‐w FLAIR and DWI (before enhancement) and T1‐w FLAIR	R. occipital lobe	Gadobutrol (0.1 mmol/kg body weight) enhanced MR using a 3 T T1‐w FLAIR sequence, leading to more central BMs and more metastases overall. Postcontrast MRI in patients with BMs is recommended to be delayed for 7 min in clinical practice
Franceschi (2016) 35	38	T1‐w post and SWI	NA	The occurrence of bleeding was shown to be rare in micro‐metastases using SWI and contrast‐enhanced high‐resolution T1‐w imaging, but prevalent in larger metastases independent of the initial cause, melanoma versus BC Haemorrhage was more prevalent in melanoma than in original BC, especially in the larger metastases that developed from both of these cancers
Kesler (2015) 36	36	DTI and T1‐w	L. corpus callosum, bilateral inferior longitudinal fasciculus, L inferior fronto‐occipital fasciculus, and bilateral temporal and frontal lobe white matter	The BC group had a significantly greater number of streamline fibres than the controls, but these streamlines were shorter and had a lower mean FA
Yeh (2015) 37	62	T1‐w, T1‐w post, T2‐w fast spin echo (FSE), FLAIR and DWI	Parietal lobe, R. frontal lobe and L. cerebellum	TNBC patients were more likely to have cystic necrotic BMs visible on MR images There are unique MRI characteristics of patients with TNBC BMs that aid in the evaluation of newly developed BMs
Quattrocchi (2014) 38	42	T1w, T1‐w post and T2‐w FLAIR	Parieto‐occipital lobes and cerebellum	The large proportion of patients with BC and the high number of cerebellar metastases influenced the results in the non‐lung cancer group Patients showed significant clusters in the cerebellum when WMHs was absent. In contrast, no distribution shift was observed in the presence of WMH, suggesting that WMH does not have any effect on the brain distribution of BM in these patients
Huang (2010) 39	17	Multivoxel 2D‐CSI MRS and DCE (n = 21 BC)	NA	HIF‐1 can be responsible for the association between choline metabolism and tumour perfusion in BMs
Hakyemez (2010) 40	5	T2‐w and PWI for rCBV	NA	PWI is used as a supportive method in cases where conventional images fail to differentiate metastatic lesions from gliomas Intra‐tumoural rCBV measurements do not contribute significantly to this differentiation. However, as peritumoural oedema displays low rCBV ratios in metastatic lesions and higher rCBV ratios in high‐grade gliomas, it can be used to differentiate these two lesions Mass and rCBV ratios of the oedema surrounding the tumour prior to operation in solitary masses proved to be useful for differentiating metastases from high‐grade gliomas
Takeda (2008) 41	13	2D T1‐w SE, 3D MPRAGE T1‐w post, 2D T1‐w SE post and 2D T2‐w SE	Posterior fossa, middle fossa and supratentorial	BM detection using 3D MPRAGE was more effective than using 2D SE
Kremer (2003) 42	2	T1‐w, T2‐w, T2*‐w, R2* and DCE	NA	Large and solitary necrotic metastases can be indistinguishable from high‐grade astrocytomas A high rCBV level can indicate a hypervascular lesion
Geijer (2002) 43	1	T1‐w, T2‐w, DWI and ADC	Lateral wall of the posterior horn of the L. lateral ventricle and falx cerebri	In BC metastases, the lesions were surrounded by oedema Standard DWI protocols show significant overlap between the characteristics of common BMs and those of subacute and acute ischemic lesions

2D T1‐w SE, two‐dimensional T1‐weighted spin echo; 2D T2‐w SE, two‐dimensional T2‐weighted spin echo; 3D T1‐w MPRAGE, 3D T1‐weighted magnetisation prepared rapid acquisition gradient echo; ADC, apparent diffusion coefficient; BC, breast cancer; BMs, brain metastases; CBV, cerebral blood volume; CE MRI, contrast‐enhanced MRI; CE, contrast‐enhanced; DCE, dynamic contrast‐enhanced; DSC, dynamic susceptibility contrast; DTI, diffusion tensor imaging; DWI, diffusion‐weighted imaging; EVR, enhanced volume ratio; FLAIR, fluid‐attenuated inversion recovery; FSPGR, fast spoiled gradient recalled; GBM, glioblastoma multiforme; HER2, human epidermal growth factor receptor 2; HIF‐1, hypoxia‐inducible factor‐1; HR3D IR high‐resolution 3D inversion recovery; L., Left; LAVA, liquid attenuation inversion recovery; LITT, laser interstitial thermal therapy; MPRAGE, magnetisation prepared rapid acquisition gradient echo; MRI, magnetic resonance imaging; MRS, magnetic resonance spectroscopy; Multivoxel 2D‐CSI MRS, multivoxel two‐dimensional chemical shift Imaging magnetic resonance spectroscopy; PWI, perfusion‐weighted imaging; R., right; R2*, transverse relaxation rate; rCBV, relative cerebral blood volume; rsfMRI, resting‐state functional MRI; SWI, susceptibility‐weighted imaging; T1‐w post, T1‐weighted postcontrast; T1‐w, T1‐weighted; T2*‐w, T2‐star weighted; T2‐w, T2‐weighted; TNBC, triple‐negative breast cancer; WMH, white matter hyperintensities.

^a Hurst exponent measures a time series' memory, with values >0.5 showing trends and <0.5 indicating mean reversion.