Introduction
Human T cell leukemia/lymphoma virus type 1 (HTLV-1) was isolated in 1980, by National Institutes of Health, USA as the first oncoretrovirus in human. The relation between chronic myelopathy (tropical spastic paraparesis, TSP) with unknown etiology and this retrovirus was investigated as HTLV-1 associated myelopathy (HAM) in some tropical and inter-tropical areas including Caribbean region, South and Central Africa and India,  proposed as the hybrid name HAM/TSP (1). Today, considerable number of patients with HAM/TSP has been diagnosed in regions with high prevalence of HTLV-1 including Japan, Jamaica, Trinidad, Martinique, several parts of South and Central America (Brazil, Peru, Colombia), Central and South Africa, and Iran. HAM/TSP is a chronic progressive neurologic disease that progresses slowly with different onset of initial signs and accounts for 0.2-3% of patients with HTLV-1. Various clinical characteristics and diagnostic symptoms have been reported in patients with HAM/TSP including lower extremity hyperreflexia and spasticity, sensory disturbances, urinary bladder and sexual complications, urinary incontinence, and peripheral sensory loss to light touch (2-4).
Magnetic resonance imaging (MRI) has revealed high sensitivity for detecting and controlling patients with CNS complications. In HAM/TSP patients, MRI has the ability to identify the level of brain and spinal cord abnormalities. In this study, we systematically reviewed the diagnostic importance of MRI in patients with HAM/TSP.

Methods
Literature search strategy
We conducted the current systematic review based on the PRISMA guidelines (Figure1). PubMed was searched for relevant articles with the following search terms: (spastic paraparesis OR myelopathy) AND HTLV AND (magnetic resonance imaging OR MRI). The last search was done on December 2014. Title and abstract of the articles were studied and then irrelevant articles were omitted. Therefore, the full text of the remaining articles was studied to detect any irrelevant article to our systematic review purpose. The reference list of the relevant articles was searched for reducing the possibility of missing relevant citations.

Study selection
No time limitation was inserted for the included articles and only English language articles were entered in our systematic review.  Inclusion criteria were all the case series and cohort studies, which included at least 10 patients with recognized HAM/TSP disease and investigated the accuracy and efficacy of MRI modality in detecting brain and spinal abnormalities. Exclusion criteria of our study were articles with less than 10 patients, patients with HTLV-II, patients with unknown causes of TSP, HTLV-1/myelopathy, non-English articles and case reports.  

Data extraction
Data about authors, publication date, country, patient characteristics (age, sex, duration of the disease) and result of applying MRI were extracted.

Outcome variable
We compared data obtained in each study regarding the brain lesions and atrophy level, location of the lesions, spinal cord lesions and atrophy level, relation between patients’ age and disease progress and the relation between patients’ disease duration and the presence of brain and spinal lesions.

Data synthesis
 We provided an evidence table involves all the information abstracted from eligible studies.  Results were organized based on the data obtained on brain and spinal cord lesions and the level of atrophy, the association between incidence of observed abnormalities with patient’s age and duration of disease.

Results
Search results
We identified a total of 151 articles in PubMed. We excluded all the non-relevant articles based on title, abstract and eventually the full text. The remaining relevant articles were 14 in PubMed.

Description of the included studies
All the included articles were performed in United States, Canada, Brazil, Japan, United Kingdom and Dominican Republic. The population size of the included studies ranged between 10-68 HAM/TSP patients. The mean disease duration ranged between 3-12 years.
Table 1 shows data extracted from included studies used MRI technique for HAM/TSP patients.

Discussion
Based on studies, vast majority of HTLV-1 patients are asymptomatic, who are rarely evaluated especially in non-endemic regions. In some studies, carriers were investigated with MRI and no significant atrophy and focal lesions were observed in this group of patients. Thus, this suggested further studies to reveal the importance of imaging parameters in detecting spinal cord involvement in HTLV1 carriers.  

MRI findings
Brain atrophy
Only 2/14 included articles detected brain atrophy in HAM/TSP patients using MRI technique (6,9). Based on these studies, brain atrophy was observed in 4/28 and 1/19 patients of each articles. A recent study by Puccioni-Sohler, in 2012, proposed volume reduction, reduced brain parenchymal reduction, widening of cortical sulci and fissures, ventricular enlargement, cerebellar atrophy with slight accentuation of cisternas, sylvian and frontal regions convexity reduction. Based on these results, brain atrophy symptoms does not seem to occur frequently in all HAN/TSP patients and studies with larger population are needed to investigate the prevalence of brain atrophy more accurately as a marker of HAM/TSP patients.

Brain lesions
Overall, 13/14 articles reported the incidence of brain hyperintense abnormalities in their patients mostly based on T2-weighted images and had variable frequency of 52%-89% (4,6,10,12-17). Based on mentioned articles, brain abnormalities of HAM/TSP patients were in deep cerebral and subcortical white matter, periventricular areas, semioval centre, corona radiate and they had grey matter lesions and lacunar infarcts in basal ganglia. Periventcular parts of the brain white matter are suggested as major and preferred location of the detected lesions followed by subcortical white matter lesions  (4,6,11,12,16). Based on literature, focal and non-specific identified brain lesions show that MRI technique is highly sensitive for lesions but does not have high sensitivity in HAM/TSP patients (10). In one study, it was suggested that the incidence of brain non-specific abnormalities might be the consequence of micro-degenerative process during aging phenomenon (6).

Spinal cord atrophy
According to our identified articles, only 5/14 articles investigated individuals with advanced HAM/TSP disease revealed the spinal cord atrophy in HAM/TSP patients, which was detected as thinning and diameter reduction of the cervical spine and the proximal portion of the dorsal medulla and atrophic of thoracic cord (5,6,11,12,15). Puccioni-Sohler et al suggested that incidence of both cerebral and spinal cord atrophy might be due to degeneration of parenchymal (6). In another study the 5.6% of 86 patients showed spinal cord atrophy in T2-wighted sagittal images. They suggested that spinal cord atrophy was significantly prominent in patients with longer duration of disease (12). Liu et al., in 2014, proposed thoracolumbar cord as the main involved part of cord in HAM/TSP patients by specifically evaluating spinal cord cross sectional areas (SCCSA) along the entire length as an indicator of atrophy(5). Based on high correlation between SCCSA and clinical measures of HAM/TSP patients, this index is proposed as an effective tool for estimating the atrophy level (5). Similar results obtained in another recent study by Vilchez et al., showed thoracic cord atrophy in all 10 HAM/TSP patients, which was more intense in patients with longer duration of the disease (4).

Spinal cord lesions
HAM/TSP mostly involved thoracic parts of spinal cord. Analyzed extracted data showed that 6/14 included articles have demonstrated T2-hyperintense spinal cord lesions in HAM/TSP patients considered by MRI; these cord abnormalities were associated with transient and diffuse swelling of spinal cord in some cases (4,6,7,9-11). According to the mentioned articles, spinal cord lesions were mostly in thoracic level of cord.
Yukitake et al. mentioned that 3/4 patients with spinal cord T2-hyperintense demyelinating lesions had advanced disability, pleocytosis, and high concentration of protein and blood-CSF barrier dysfunction and had to use wheelchair; the coincidence of acute inflammatory process and spinal cord lesions have been also mentioned by others (6,7). To obtain more accurate evaluation of spinal cord lesions, performing larger sample size studies on HAM/TSP patients are recommended.
Based on literature, variable percent of cases have shown spinal cord atrophy in each study, which might be due to different individual characteristics such as age, sex, disease duration and duration of symptoms.

Disease duration
The relation between disease duration and the presence of atrophy or brain and spinal lesions were investigated in some studies. The incidence rate of thoracic cord lesions and atrophy was significantly related with patients’ disease duration in some included studies; this relation was not detected in patients with cervical cord lesions(4,5). Patients with white matter lesions have had longer duration of disease; more than 2 years (12,16,17). Thoracic cord atrophy has also shown significant association with duration of the disease in HAM/TSP patients. In two included studies, no significant relation has been detected between presence and location of brain lesions and disease duration (6,7).

Age of the patients
According to some studies,  increased age could be proposed as a risk factor of the higher incidence of white matter lesions in patients with HAM/TSP. Puccioni-Sohler et al. suggested a relation between brain MRI T2 hyperintense lesions and micro-degenerative process in middle aged or older patients, which were mainly in periventricular parts (6). This correlation was also observed in some other studies (8). However Griffith et al. did not obtain any statistically significant relation between brain volume and age of the HAM/TSP patients. They only observed significant brain reduction in a 39-year old patient with 2-year disease duration, among 19 patients (9).  Kira et al. conducted 2 different studies on HAM/TSP patients and obtained similar results with Griffith et al. regarding the patients` age at the time of examination and prevalence of WM  lesions  (16,17).
Blurring due to motion, lack of contrast between the cord and the cerebrospinal fluid, are proposed as the possible artifacts of the SCCSA profiles in MRI(5).
According to the extracted data of the included articles, it could be suggested to perform more studies on HAM/TSP patients in endemic areas to increase the number of patients and the accuracy and validity of results. In asymptomatic patients, further studies are required on asymptomatic HTLV-1 patients to accurately reveal the prognostic and diagnostic validity of imaging parameters in carriers.
 
Conclusion
In HAM/TSP patients, brain atrophy is not reported frequently, however, brain lesions are more prevalent in these patients. Brain lesions and atrophy might progress in patients; thus, yearly longitudinal follow-up of patients are needed to accurately identify the brain changes over time.  It was also suggested that conventional MRI could be suggested as a sensitive diagnostic tool due to none-specific brain lesions observed in most of the studies, but not as a highly specific tool in HAM/TSP patients. The MRI technique has shown high sensitivity in detecting abnormalities but lesions seen on the MRI images are not specifically indicative of HAM/TSP pathology and could be attributed to other subclinical pathologies.

Acknowledgement
We would like to thank Clinical Research Development Center of Ghaem Hospital for their assistant in this manuscript. This study was supported by a grant from the Vice Chancellor for Research of the Mashhad University of Medical Sciences for the research project as a medical student thesis with approval number of 920120.

Conflict of Interest
The authors declare no conflict of interest.