Fig 1.
Conventional MRI; T1-MPRAGE without contrast, T1-MPRAGE with Gadolinium and APTw in color for subject 25 (MET; top) and subject 17 (LGG; middle) and subject 18 (HGG; bottom). The ROI placement for each subject is depicted. The APTw signal intensity for the high grade glioma and metastasis subject is distinctly higher than for the low grade glioma subject in the middle. Subject 25 MET (misdiagnosed by reader 2 and reader 3 as HGG, correctly diagnosed by APTw), Subject 17 LGG (correctly diagnosed by all). Subject 18 HGG (misdiagnosed by the initial radiological assessment as LGG, correctly diagnosed by APTw and all readers)”.
Fig 2.
Enlargement of Fig 1 for visualization of ROI-placement.
T1-MPRAGE without contrast, T1-MPRAGE with Gadolinium and APTw in color for subject 25 (MET; top) and subject 17 (LGG; middle) and subject 18 (HGG; bottom).
Fig 3.
Conventional MRI; T1-MPRAGE without contrast, T1-MPRAGE with Gadolinium, FLAIR and APTw in color for subject one (HGG; top) and subject 3 (LGG; bottom) The APTw signal intensity for the high grade glioma subject at the top is distinctly higher than for the low grade glioma subject at the bottom. Also notice the increased APTw signal in blood vessel regions. Subject 1 HGG (misdiagnosed by reader 1 and 3 as MET, correctly diagnosed by APTw) and subject 3 LGG (correctly diagnosed by all)”.
Fig 4.
Four slices through the lesion of subject 5 with a high grade glioma (HGG).
Notice the coherence between the increasing hypointensity on T1-MPRAGE (a-d), hyperintensity on FLAIR (e-h), increased hyperintensity on T2w MRI (i-l) and increased APTw signal (m-p). Also notice the increased APTw signal in blood vessel regions. Gd-enhancement was studied with T1w; T1-MPRAGE. Hyperintensities on T2 FLAIR and T2 turbo spin-echo of conventional pre-contrast protocols were correlated to T1w-Gd enhancement where possible, otherwise hypointensity.
Fig 5.
T1-MPRAGE without contrast, T1-MPRAGE with Gadolinium, FLAIR, T2w and APTw in subject 9 with a low grade glioma (a-e) and subject 10 with a high grade glioma (f-j). Of note is the increased APTw signal in the high grade glioma (j) compared to the low grade glioma (e). on APTw. Also notice the increased APTw signal in blood vessel regions of all patients. Subject 9 LGG (correctly diagnosed by all) and subject 10 HGG (misdiagnosed by the initial radiological classification as MET, correctly diagnosed by APTw).
Fig 6.
Z-spectra and magnetization transfer ratio asymmetry spectra for subjects; subject 13 (HGG) and 9 (LGG) within tumor and in contralateral normal appearing white matter.
Table 1.
Patient demographics, histopathology with corresponding APTw signal intensity and gene mutation status.
Table 2.
APTw signal mean, max, min and range for HGG, LGG and MET.
Table 3.
Mann-Whitney U test for distinguishing low grade glioma and high grade glioma using APTw signal.
Fig 7.
AUC, 95% CI, sensitivity and specificity with cut off values reported in Table 4.
Table 4.
ROC analysis for distinguishing HGG and LGG using APTw signal intensity.
Fig 8.
Mean, max and range APTw signal combined with logistic regression.
AUC, 95% CI, Sensitivity and specificity with cut off values reported in Table 4. The combined model mislabelled subjects 3 and 17 as they were labelled HGG in the model but are histologically verified LGG, also subject 7 was mislabelled as a LGG whereas it is histologically a Glioblastoma, Table 1.
Table 5.
Classification of brain lesions into HGG/LGG/MET at primary presentation of disease and follow-up review with 3 readers and quantified intralesional mean APTw signal with cut off >+2.0% and qualitative assessment of T1w Gd enhancement.