Central post-stroke pain due to injury of the medial lemniscus in a patient with medullary infarction.

Central post-stroke pain (CPSP), which is neuropathic pain associated with a brain lesion following a stroke, was first described by Dejerine and Roussy in 1906 in a thalamic infarct patient (Dejerine and Roussy, 1906; Watson and Sandroni, 2016). Since then, many studies have suggested that the pathophysiological mechanism underlying CPSP is activated by injury to the somatosensory nervous system (Watson and Sandroni, 2016). There are two major pathways within the somatosensory nervous system: the spinothalamic tract (STT), which is chiefly responsible for pain and touch sensations, and the medial lemniscus (ML) pathway, which is mainly involved in proprioception and tactile discrimination (Naidich et al., 2009). Many studies have suggested that STT injury is a major pathogenetic mechanism of CPSP, but a few studies have suggested that ML injury is related to the development of CPSP (Kim and Choi-Kwon, 1999; Seghier et al., 2005; Kim, 2007; Hong et al., 2010; Jang et al., 2017).

Diffusion tensor tractography (DTT), the results of which are derived from diffusion tensor imaging (DTI), has made possible the three-dimensional reconstruction of neural tracts such as the STT and the ML pathway in the human brain (Jang and Kwon, 2013). Several DTT-based studies have demonstrated that STT injury is associated with the pathophysiological mechanism underlying CPSP in stroke patients (Hong et al., 2010; Jang et al., 2017). Hong et al. (2010) reported that STT injury was related to the presence of CPSP in 30 patients with intracerebral hemorrhage. Subsequently, Jang et al. (2017) reported that CPSP was due to STT damage in five patients with cerebral infarction. However, there are no DTT-based studies regarding the relationship between CPSP and ML injury. Therefore, in the current study, we used DTT to show ML injury in a patient with medullary infarction affecting both pyramids and both ML areas.

A 67-year-old male patient was recruited from a university hospital and he presented with quadriparesis and numbness of the right hand due to a medullary infarction that affected both pyramids and both ML areas and was related to the presence of vertebral artery stenosis (). His treatment included conservative management and rehabilitation. Six weeks after stroke onset, he was relocated to the department of rehabilitation at our university hospital. At that time, he presented with quadriparesis (motor weakness of all four extremities [Fair/Good] and gait disturbance [Functional Ambulation Category: 3.5, full score: 5]) and pain in the right upper and lower limbs and the left hand (Frese et al., 1987; Bijur et al., 2001; Cunha et al., 2002). The characteristics and severity of his pain were assessed as follows: continuous pain without allodynia or hyperalgesia and numbness (visual analog scale scores: 3 [right upper and lower limbs] and 2 [left hand]) (Bijur et al., 2001). Brain magnetic resonance images obtained 6 weeks after onset showed leukomalactic lesions in both pyramids and both ML areas in the medulla (). Motor-evoked potentials were obtained from the tibialis anterior; latency and amplitudes for the right hemisphere were 31.1 ms and 1.3 mV, respectively, while those for the left hemisphere were 32.3 ms and 1.0 mV, respectively. In addition, somatosensory-evoked potentials were obtained from the posterior tibial nerve. The P37 and N45 latencies of the right hemisphere were 42.8 and 50.3 ms, respectively, while those for the left hemisphere were 37.7 and 48.2 ms, respectively. An electromyographic study showed no evidence of radiculopathy or peripheral neuropathy. A normal subject (68-year-old male) was recruited from the same university hospital. Signed informed consent was obtained from the patient and a healthy subject, and the study protocol was approved by the Institutional Review Board of Yeungnam University Hospital (YUMC 2019-06-032) on June 21, 2019.

Results of diffusion tensor tractography for the medial lemniscus and spinothalamic tract of a patient with medullary infarction.

(A) Diffusion weighted images at onset and T2-weighted magnetic resonance (MR) images at 6 weeks after onset show medullary infarctions in both pyramids and medical lemniscus areas (blue arrow). (B) Results of diffusion tensor tractography for the medial lemniscus pathway and spinothalamic tract at 6 weeks after onset. The left medial lemniscus pathway is thinner (orange arrows) than the right medial lemniscus pathway and those of a normal subject spinothalamic tract reveals a normal configuration in both hemispheres compared with those of a normal subject (68-year-old male) who recruited the same university hospital. (C) The medical lemniscus pathway and spinothalamic tract of a normal subject.

At 6 weeks after infarct onset, DTI data were acquired using a six-channel head coil on a 1.5T Philips Gyroscan Intera (Philips, Ltd., Best, the Netherlands) with 32 gradients. Imaging parameters were as follows: acquisition matrix = 96 × 96; reconstructed to matrix = 192 × 192; field of view = 240 × 240 mm2 ; repetition time = 10,398 ms; echo time = 72 ms; parallel imaging reduction factor = 2; echo-planar imaging factor = 59; b = 1000 s/mm2 ; and slice thickness = 2.5 mm. Probabilistic tractography as provided in the default tractography option in Brain Diffucion Software of Oxford Centre (Oxford University, Oxford, UK) for Functional Magnetic Resonance Imaging was used to perform fiber tracking. For ML and STT reconstruction, seed regions of interest (ROIs) were positioned in accordance with the reported anatomy at the level of the medulla (ROI for ML: medio-posterior to the pyramid in the medulla, ROI for STT: posterior to the inferior olivary nucleus and anterior to the inferior cerebellar peduncle in the medulla) (Jang and Kwon, 2013). The target ROI was placed on the area of the ventro-postero-lateral nucleus of the thalamus and the primary somatosensory cortex (Jang and Kwon, 2013).

At 6 weeks post-onset, DTT was able to fully reconstruct the STT in both hemispheres. In contrast, DTT reconstruction revealed that the left ML pathway was thinner than the right ML pathway ().

In the current study, we investigated the state of the STT and the ML pathway in a patient with CPSP in the right upper and lower limbs and the left hand following medullary infarction affecting both pyramids and both ML areas. At 6 weeks post-onset, DTT revealed that the left ML pathway was thinner than the right pathway, suggesting an injury of the left pathway due to the infarction in the left ML area. In contrast, the right ML and both STTs were shown to have normal configurations. Based on those results, the patient’s CPSP in the right upper and lower limbs might be at least partially ascribed to the infarction in the left ML area, whereas the CPSP of the left hand might be related to the infarction in the right ML area, even though the right ML revealed a normal configuration. The difference in CPSP severity between that of both right limbs and that of the left hand only corroborates the difference in the DTT results of the left and right ML pathways.

Several studies have used DTT to demonstrate that STT injury was associated with the development of CPSP (Seghier et al., 2005; Hong et al., 2010; Jang et al., 2017). Regarding the ML, a few studies showed that a lesion related to the ML was associated with the presence of CPSP (Kim and Choi-Kwon, 1999; Kim, 2007). Kim and Choi-Kwon (1999) examined 55 patients in two groups (lateral medullary infarction (LMI) group: 41 patients; medial medullary infarction (MMI) group: 14 patients) in an investigation into sensory sequelae on the face, body, and limbs. The members of the LMI group showed numbness (10 patients), burning (9 patients), and cold (7 patients) sensations in the face, and cold (18 patients), numbness (15 patients), and burning (12 patients) sensations in the body/limbs. In contrast, the MMI group showed numbness (9 patients), squeezing (3 patients), and cold (1 patient) sensations in the body/limbs (Kim and Choi-Kwon, 1999). The authors also suggested that the sensory sequelae in the LMI group frequently had a delayed onset (up to 6 months), whereas there was immediate onset of sensory sequelae in the MMI group (Kim and Choi-Kwon, 1999). In 2007, Kim investigated a 55-year-old male patient with CPSP following both LMI (first infarct) and MMI (second infarct). After the initial LMI, which included the STT, the patient had CPSP that, gradually improved. However, his CPSP was aggravated because of the occurrence of the secondary MMI, which included the ML (Kim, 2007). Aside from those studies, to our best knowledge, this case report is the first to demonstrate an ML injury in a patient with CPSP after medullary infarction using DTT. However, limitations of this study should be considered. First, although the CRPS mainly occurred because of injury of the somatosensory nervous system, it should be also considered with various factors. Second, DTT results can be affected by the DTI scanning conditions; moreover, analysis of DTT results is operator dependent. Third, both false-positive and false-negative DTT results can occur throughout the white matter of the brain due to the presence of partial volume effects and complex fiber configurations including crossing or ‘kissing’ fibers.

In conclusion, using DTT, we demonstrated ML pathway injury in a patient with CPSP following medullary infarction. Our results suggest that DTT-based evaluation of the ML pathway can be helpful in elucidating the pathophysiological mechanism of CPSP in stroke patients. However, because this is a single case report, further studies involving larger numbers of patients should be conducted to substantiate the results described herein.