Across the many clinical trials examining the use of stem cells (SCs) in the treatment of SCI, there are a few key technical considerations. The first is the method of injection. Intravenous and intrathecal injection have relatively few technical considerations since IV injection and lumbar puncture are routine procedures. Intramedullary injection of stem cells, however, poses a greater challenge. The spinal cord generally oscillates with respiration and cardiac pulsation. Studies in animals show that the pulsatile nature of the cord is increased following increases in blood pressure and that it is not due to cerebrospinal fluid based wave transmission. While respiratory oscillation causes greatest spinal cord movement in the thoracic spine, the genuine spinal cord pulsation is thought to be driven by the blood flow through radicular arteries[7]. This spinal cord pulsation may pose an issue with intramedullary injection techniques. In general there are two canonical ways of achieving intramedullary spinal cord injections: The free hand technique, and a stereotactic technique utilizing a bed mounted frame. It is important to note that the stereotactic technique, as it is mounted to the bed, does not necessarily improve stability in relation to spinal cord pulsation. A recent study by Levi et al[8] described the use of a free hand technique. Specifically, to achieve an injection depth of 3-5 mm they marked the injection needle using a rongeur or silicone tip and had the surgeon stabilize the needle using two hands anchored at the edge of the surgical field. The injection time, which was a maximum of 3 min and 30 s per injection, required the use of two hands for stabilization[8].

Although no serious complications occurred in this trial or others that used a free-hand technique[9,10], the damage to the spinal cord may be obfuscated by the presence of preexisting pathology. Moreover, the need for stabilization may limit the maximum injection time and thereby the amount of stem cells able to be transplanted. Curtis et al[11] described a novel technique using a floating cannula that is able to address these theoretical issues. This cannula is attached to an XYZ manipulator mounted directly on a patient's vertebral column. The cannula is able to move with pulsation of the spinal cord making long term injections feasible with minimal damage to the existing spinal cord tissue[11].

Whether the preservation of existing spinal cord tissue is technically necessary is up for debate. While the previously described studies focused on chronic SCI, a study by Xiao et al[12] in acute SCI patients completely excised the necrotic spinal cord around the SCI lesion. The authors placed a collagen scaffold impregnated with human umbilical cord MSCs in this area and showed nerve conduction across the SCI lesion plus functional improvement from American Spinal Injury Association Impairment Scale (ASIA) A to ASIA C in two patients[12]. This study was built on previous work by Lima et al[13], where necrotic scar tissue was excised and an olfactory mucosal autograft was placed. Importantly, this radical technique of laminectomy, scar excision and mucosal autograft replacement in chronic cervical and thoracic SCI led to marked functional improvement. In the earlier study by Lima et al[13], 28% of patients had recovery of bladder sensation or voluntary anal contraction[13], while in the more recent study 55% of patients had ASIA improvement of at least 1 grade[14]. Importantly, this second study involved intensive rehabilitation following autograft transplantation. It is important to note that olfactory mucosa grafts have been associated with spinal masses pointing to their increased regenerative potential but also to their increased side effect profile[15,16]. As a technical consideration, excision of scar tissue may allow for increased regeneration of neural white matter tracts and may be an important technical consideration even with other injection techniques that do not use a graft substrate or scaffold.

Although some studies of SCI show improvement following a single stem cell injection[11,17], studies with multiple injections spaced out over time seem to have greater improvements in outcomes[8,18]. A comparison of two trials conducted by the same group, Oh et al[10] and Park et al[9], further illustrated this point. The original study, by Park et al[9], was a Phase I single-arm study and included the injection of BM-MSCs derived from iliac crest grafts into both the intramedullary and subdural space with additional injections into the thecal space using lumbar puncture at 4 wk and 8 wk following the initial operation. Six of 10 patients showed motor improvement and 3 showed gradual improvement in activities of daily living[9]. The follow-up study, by Oh et al[10], was a Phase III clinical trial and the study authors used the same initial injection treatment paradigm with no follow-up injections. The one time injection yielded poor functional improvement (12.5% with improved motor outcomes) compared to the multiple injection protocol (60% with motor improvement or improvement on Activities of Daily Living). Although limited, these data point to the importance of optimizing chronicity of SC injection in SCI patients. Importantly, while 5/16 patients had diffusion tensor imaging magnetic resonance imaging changes showing tracts spanning the SCI level in the single injection Phase III clinical trial compared to 7/10 showing such changes in the pilot study[10]. It is difficult to associate the differences solely due to the presence of multiple treatments. Multiple other studies have shown improvements in ASIA scores, motor function and urodynamics with only a single stem-cell treatment[19,20]. Furthermore, understanding the number of cells transplanted and its effect on outcomes is more difficult. Various trials have transplanted cell numbers ranging from 1 × 10⁵ to 40 × 10⁷. There was no consistent relationship between cell number and outcome over the trials reported. It remains to be seen whether a multiple treatment vs single treatment protocol could improve functional recovery in a large-scale clinical trial.

Generally, SCI is characterized based on chronicity into acute, subacute and chronic phases. While the time course for each period is highly variable, the acute phase generally spans days post-injury, the subacute phase weeks post-injury and the chronic phase months post-injury. Although some studies on chronic SCI show improvement following stem cell injection[8,11], studies in the acute or subacute SCI population seem to show a more dramatic return of function[10,21,22]. A study by Yoon et al[23] stratified patients into acute (< 13 d), subacute (14 d to 8 wk) and chronic (> 8 wk) groups with all patients receiving intramedullary injection of BM-MSCs with systemic granulocyte macrophage colony stimulating factor treatment. Notably, 30.4% of the acute and subacute treated patients had improved ASIA grade (ASIA A to Asia B/C), while none of the chronic patients showed any improvement[23]. This was not limited to intramedullary cell transplantation. Bansal et al[22] demonstrated that lumbar puncture for delivery of BM-MSCs led to ASIA grade improvement in 6/10 patients who were less than 6 mo from injury with no patients over 6 mo from injury achieving functional improvement[22].

A study by Shin et al[20], used human fetal tissue-derived neural stem cell progenitor cells and included both a control group and an intramedullary cell transplantation group. Notably, while 26% of patients in the transplantation group had ASIA grade improvement, only 6.6% had improvement in the control group[20]. This disparity between transplantation and control group patients was also seen by Karamouzian et al[24]. These authors injected purified BM-MSCs via lumbar puncture and noted that 45 patients in the cell transplantation group had ASIA improvement compared to 15% in the control group (P = 0.09)[24]. This helps to answer the major critique that some degree of functional improvement occurs in the acute/subacute period naturally and may explain the functional improvements seen in the acute to subacute cell transplantation studies. Also, the fact that both studies with control groups used different stem cell types and different methods of injection (lumbar puncture vs intramedullary injection) also highlights the importance of the treatment window vs mechanism of treatment. This is further supported by the studies carried out by Bansal et al[22] and Yoon et al[23]. These studies had patients in multiple treatment windows, with intramedullary or lumbar puncture delivered BM-MSCs and noted that patients in the acute to subacute period from injury achieved greater functional improvement[22,23]. In fact, of the studies that focused on the acute/subacute SCI population, only one out of seven studies showed no motor or functional improvement with the rest having a subset of patients that had ASIA grade improvement. The single study with no functional improvement notably used a novel form of stem cells derived from peripheral Schwann cells and included subjects from 4-7 wk following injury-generally on the upper limit as compared to other acute/subacute studies[25].