Lipid Resonance on In Vivo Proton MR Spectroscopy: Value of Other Metabolites in Differential Diagnosis.

In vivo proton MR spectroscopy ((1)H-MRS) can non-invasively provide biochemical information at the same examination as conventional magnetic resonance imaging (MRI). Lipid resonance (0.9-1.5 ppm) is a marker of cell membrane breakdown and tissue necrosis, but its diagnostic significance has not been well described. We retrospectively analyzed spectra to study the different pathological conditions in patients with abnormal lipid resonance. All patients with neurological diseases showing lipid resonance on (1)H-MRS (1.5T) in a tertiary hospital over two years were retrospectively analyzed. (1)H-MRS was performed using the single voxel PRESS technique (TR/TE=3000/144 ms, eight excitations). Spectra were analysed for the presence of NAA (2.0 ppm), creatine (3.0 ppm), choline (3.2 ppm), acetate (1.92 ppm), succinate (2.4 ppm), cytosolic amino acids (0.9 ppm), lactate (1.3 ppm) and lipid (0.9-1.5 ppm) peaks. Ninety-two spectra from 69 patients (38 males, 31 females; aged 9 to 89 years) were analyzed. The final diagnosis was infective (n= 33), (tuberculoma n=17, pyogenic abscess n= 8, fungal abscess n= 3, sterile abscess n= 3, tubercular abscess n= 2), neoplastic (n= 21) (glial tumors n= 9, metastasis n= 8, lymphoma n= 4), and other (n= 15) abnormalities (subacute and chronic stroke n= 6, postictal edema n= 4, multiple sclerosis n= 2, Erdhiem Chester disease n= 2, Rosai Dorfmann disease n= 1). Succinate and acetate were detected only in pyogenic abscesses (2/4 cases), but amino acids were present in both pyogenic (4/8) and fungal (3/3) abscesses. Choline was seen not only in neoplasms (18) but also in tuberculomas (11/17), but was consistently absent in the abscesses. Lactate was present in glioblastoma (7/9), pyogenic (3/8) tubercular (2/2) and fungal (3/3) abscess. Isolated lipid resonance was found in Erdheim Chester disease (2/2) of the orbit, and lipid and choline was seen in Rosai Dorfmann's disease (1/1). Brain lesions containing lipid on (1)H-MRS could be differentiated by the presence of succinate and acetate in pyogenic abscess, and amino acids in pyogenic/fungal abscesses. Choline was seen in neoplasms and in tuberculomas, but not in the abscesses. Thus, the presence of a lipid peak, when combined with features on other MR pulse sequences and available clinical data can help arrive at a specific diagnosis. (1)H-MRS should not be interpreted in isolation: it should always be correlated with conventional imaging features, and performing (1)H-MRS in isolation remains an important pitfall.