Introduction
 Nonalcoholic fatty liver, commonly known as hepatic steatosis, is a liver inflammation caused by excessive accumulation of the fat in the liver cells (1). In fatty liver disease, hepatic accumulation of excess fat often leads to impaired normal functioning of the liver, which can cause rapid progression of liver failure, liver cirrhosis and other hepatic infections (2,3). Although the main cause of fatty liver disease is unknown, but it is suggested that the disease is closely associated with some metabolic disorders such as obesity, diabetes, and high cholesterol as well as high blood triglycerides (4). Fatty liver disease is divided into three main categories of alcoholic, and nonalcoholic fatty liver and acute fatty liver of pregnancy (AFLP), which is a rare but potentially fatal condition of liver complication (5). Alcohol and some particular drug consumption as well as some metabolic disorders can be considered as the main possible causes of hepatic steatosis. In recent years, the prevalence of obesity and metabolic disease in the population, especially among young people is considerably increased (6,7). About 10 to 20 percent of people in the United States suffer from this disease, and most cases of fatty liver are diagnosed between the ages of 50 to 60 years (7).
Fatty liver is not specific to any particular age or sex, and it may be observed even in children, especially those who have related genetic disorders. However, the incidence of the disease may rise in elderly, because fat and blood sugar is increased in these populations. Although some symptoms such as fatigue and pain in the upper-right abdomen is common in all patients; but, the most important problem is that there is no major clinical symptoms for fatty liver, and the disease can only be diagnosed by biochemical testing or ultrasound (8,9). Considering that in different pathologies the amounts and proportions of detectable metabolites varies, it is possible to determine the type of lesion and the progression of pathologies by measuring these metabolites. Major clinical examinations that are used to determine the risk of fatty liver include blood tests, imaging, and testing of liver tissue. Magnetic resonance spectroscopy (MRS) can be effectively used in diagnosis of pathological lesions, diabetes, and malignancies by detection of metabolites in the damaged area (10). Previously, such information had been obtained only through the biopsy, which is a difficult, time-consuming and sometimes dangerous procedure. In recent years, these data can be obtained in a few minutes by using MRS as a completely non-invasive and harmless technique. Findings show that MRS is a suitable technique to determine liver damages such as hepatic steatosis. In this study, we aimed to systematically review the literatures in which the MRS had been used to determine hepatic steatosis quantitatively.

Methods
Search methods
All articles in which the MRS had been used to diagnose hepatic steatosis were searched in the PubMed and Scopus as two major databases. For this purpose, these databases were searched using the key terms “magnetic resonance spectroscopy”, and “hepatic steatosis”. Following search strategy “(((magnetic resonance spectroscopy OR MR spectroscopy OR MRS)) AND (hepatic steatosis OR Steatohepatitis OR fatty liver)) AND Quantitative)” was used to find related documents in which the fatty liver disease had been quantitatively diagnosed using MRS. Google Scholar and Google search engine were searched as well with the previously described search method to include other possible relevant articles. The databases search was completed on June 2015. As well, relevant articles found through manually reference list searching of previously collected articles were also included to reduce the data loss.

Study selection and inclusion/exclusion criteria
To avoid any possible errors and misinterpretation of data in all processes of data extraction, only English articles were selected and used for further evaluation. Articles with various types of study designs were collected and included for further data collection. But, letters, case reports, editorials, conference papers, review articles and meta-analysis were excluded from further evaluation. The records were then limited to those studies, which were conducted on human. Thus, studies conducted in vitro or on animal models were subsequently excluded from further assessment. Those studies in which MRS had been used to detect pathological changes other than hepatic steatosis were also excluded from further assessment. As well, to avoid data loss in further processes of data analysis, articles with full text unavailability were omitted during article selection. No time limitation was defined for literature search. Therefore, the general inclusion criteria for article selection were all English documents in which the practicality of MRS technique had been evaluated to detect hepatic steatosis in patients with fatty liver disease. For further evaluation of the papers in detail, articles with subject or language irrelevancy, as well as documents with duplicated data were excluded by reviewing the title, keywords and abstract of selected literatures.

Data extraction
Data including the date of publication, the name of first author, the country of study, study design, method of assessment and other available critical findings were extracted and used for data synthesis based on the main purpose of this study. All other obtainable information including total number of participants and demographic data of studied populations were collected as possible regarding the previously defined inclusion and exclusion criteria. Data were categorized based on the findings of selected documents reporting the sensitivity and usefulness of MRS technique for quantitatively and correctly detection of fatty liver disease. All processes including articles selection and data extraction were performed with standard protocol by two independent reviewers according to the recommendation of PRISMA 2009 checklist (11). The authors tried to resolve any possible discrepancies before proceeding to the next step of data synthesis.

Results
Study search results
Total of 416 articles were found in database searches where 88 were in the PubMed, and 328 records were in the Scopus. By reviewing the title and abstract of articles, almost 196 articles were omitted in the first step of article selection due to subject irrelevancy. Moreover, 139 and 66 documents were also excluded in subsequent process of article selection due to language irrelevancy and duplication, respectively. By comprehensive reviewing of the articles and considering the previously defined inclusion/exclusion criteria, only 15 unique records seemed to be more relevant to the purpose of this study. By manual reference screening of the previously included articles, 7 additional documents were also included and used for data extraction. Two unique documents were also found through searching in Google Scholar with previously described search strategy. Finally, after several strict step by step processes in the selection of appropriate documents, only 10 relevant articles, which fully met the defined inclusion criteria were selected and used for data synthesis regarding the results in which the practicality and efficiency of MRS had been reported on quantitative analysis of hepatic steatosis. Figure 1 shows the strategy of literature search and processes of study selection.
the detection of liver fat content.

Conclusion
Findings of articles included in this literature review demonstrated that the magnetic resonance spectroscopy is a widely used non-invasive imaging technique that provides spectroscopic information about metabolic changes in pathologically damaged area. The results of this study also showed that MRS can be used as a reference standard method to evaluate pathologies, and other techniques such as SteatoTest, FLI, HSI and biopsies cannot replace MR spectroscopy for the assessment of liver fat content. As well, findings showed that MRS is a suitable technique for quantitative evaluation of liver steatosis.

Conflict of Interest
The authors declare no conflict of interest.