Intramedullary mesenchymal stromal cell transplantation in humans: Procedural safety and oncological risk reassessment

Procedural safety of intramedullary and intralesional cellular delivery

Across the reviewed clinical studies summarized in Table 1, severe procedure-related adverse events or treatment-attributable neoplastic complications were not reported, whereas mild transient adverse events were described in the majority of studies.

OPEN IN VIEWER

Table 1.

Injection characteristics, delivery paradigms, and reported procedural safety outcomes across published human intramedullary and intralesional cellular transplantation studies.

Study	Cell type	Route of administration	Cell dose	Injection volume	Injection strategy	Mild/transient adverse events	Serious adverse events/neurologicaldeterioration	Follow-upduration
Chernykh et al., 2007	Bone marrow mononuclear cells	Intralesional + IV	Not standardized	Not reported	Cyst cavity injection combined with systemic infusion	None specifically reported	No allergic reactions, inflammatory complications, or procedure-related neurological deterioration reported	Mid-term clinical follow-up reported
Dai et al., 2013	Bone marrow-derived MSCs	Intralesional/Intraspinal	1 × 107 cells/ml	25 μl per injection site	Multiple parenchymal microinjections	Transient fever, headache, local discomfort	No severe injection-related complications or neurological deterioration reported	6 months
Mendonça et al., 2014	Bone marrow-derived MSCs	Intralesional (parenchymal)	5 × 106 cells/cm3 lesion volume	Lesion-dependent	MRI-based lesion-volume adjusted dosing	Mild transient adverse events reported	No serious injection-related complications or neurological deterioration reported	12 months
Curtis et al., 2018	Neural stem cells	Intramedullary	Dose-escalation protocol	Microinjection volumes	Stereotactic intramedullary injection	Mild transient postoperative events reported	No dose-limiting toxicity, severe adverse events, or procedure-related neurological deterioration reported	12 months (primary safety endpoint)
Saini et al., 2022	Bone marrow-derived MSCs	Intramedullary	6 × 106 MSCs total	1.8 ml total	Multi-site intramedullary injection	Mild transient postoperative discomfort reported	No serious injection-related adverse events or neurological deterioration reported	6 months
Kuan et al., 2023	Bone marrow-derived MSCs	Intramedullary	Not fully detailed	Not specified	Case report	No severe adverse events reported	No neoplastic or progressive intramedullary lesion reported during follow-up	Follow-up reported in case study

Intramedullary and intralesional cellular transplantation within the human spinal cord has long been perceived as inherently hazardous due to the structural fragility of neural tissue and the theoretical risk of procedure-induced neurological deterioration. However, cumulative clinical evidence increasingly challenges this assumption. Across early-phase clinical studies employing diverse cellular products, direct spinal delivery has consistently demonstrated an acceptable procedural safety profile.

As summarized in Table 1, heterogeneous cellular populations, including bone marrow-derived mononuclear cells, MSCs, and neural stem cells, have been administered using distinct injection paradigms. These approaches range from focal microinjection strategies, ⁷ lesion-volume adjusted dosing models, ⁸ bulk intramedullary injections, ⁴ and stereotactic dose-escalation designs. ³

Despite substantial heterogeneity in cell dose, injection volume, and delivery techniques (Table 1), severe injection-related complications remain remarkably uncommon. Reported adverse events are predominantly mild and transient, including low-grade fever, headache, or local discomfort. Importantly, no reproducible pattern of procedure-induced neurological deterioration has been identified across studies (Table 1).

This convergence of safety observations across divergent procedural strategies suggests that controlled spinal cellular administration may represent a biologically tolerated intervention rather than an intrinsically destabilizing insult.

Heterogeneity of dosing and injection strategies

A defining characteristic of spinal cellular transplantation studies lies in the considerable variability of dosing paradigms. Rather than reflecting methodological inconsistency, this heterogeneity may represent a biologically rational adaptation to lesion-specific variables. However, this heterogeneity also complicates interpretation of the available safety evidence. Differences in cell source, manufacturing processes, expansion protocols, cell dose, injection volume, and delivery routes may influence both biological activity and adverse event profiles. Consequently, the apparent consistency of safety findings across studies should not be interpreted as evidence of equivalence between cellular products or administration strategies. The current literature remains insufficient to determine whether specific products, doses, or delivery paradigms confer differential safety risks.

Distinct conceptual dosing strategies are evident in Table 1, including concentration-based administration, ⁷ lesion-volume normalized cell density, ⁸ empiric fixed-dose regimens, ⁴ and dose-escalation safety designs. ³ These divergent approaches reflect fundamental uncertainties regarding optimal therapeutic dose metrics in structurally heterogeneous spinal cord lesions.

Importantly, procedural safety stability across heterogeneous dosing frameworks (Table 1) reinforces the interpretation that variability in injection parameters does not inherently translate into increased procedural risk.

Long-term biological safety considerations

Beyond procedural tolerability, long-term biological safety remains a central consideration in cellular therapy discussions. Theoretical concerns surrounding tumorigenesis, ectopic tissue formation, aberrant differentiation, and delayed neurological deterioration continue to shape regulatory and translational debates.

As outlined in Table 2, no confirmed cases of treatment-related neoplasia, intramedullary mass formation, or progressive proliferative lesions attributable to MSC transplantation have been reported, although lifetime oncological surveillance data remain inherently limited. Similarly, ectopic tissue formation and progressive neurological decline causally linked to cellular transplantation remain unsubstantiated across published datasets.

OPEN IN VIEWER

Table 2.

Long-term biological safety domains reported across spinal cellular transplantation studies. Data derived from published early-phase clinical studies in spinal cord injury and intraspinal amyotrophic lateral sclerosis trials.^9–12 Follow-up durations remain limited relative to lifetime oncological risk assessment.

Safety domain	Clinical observations across published SCI and ALS intraspinal trials	Biological/theoretical considerations
Neoplastic transformation/Tumorigenesis	No confirmed cases of treatment-attributable neoplasia, secondary intramedullary mass formation, or progressive proliferative lesions reported within currently available follow-up durations (SCI and ALS Phase I–II trials).	Theoretical risk primarily associated with prolonged ex vivo expansion, potential genomic instability, and clonal selection phenomena.
Ectopic tissue formation	No radiologically or clinically confirmed ectopic tissue development within the spinal parenchyma reported.	Theoretical possibility related to multipotent differentiation capacity under permissive microenvironmental conditions.
Aberrant differentiation	No clinically significant maladaptive lineage differentiation documented in vivo.	Biological plausibility acknowledged based on stem cell plasticity; risk considered dependent on manufacturing and microenvironmental cues.
Immune or inflammatory reactions	Rare, predominantly mild and transient inflammatory responses; no persistent immune-mediated spinal pathology reported.	Autologous origin reduces immunogenic risk; allogeneic products may theoretically carry alloimmune activation potential.
Delayed neurological deterioration	No reproducible pattern of progressive neurological decline causally attributed to cellular implantation.	Distinction required between disease progression, procedural factors, and hypothetical biological effects.

This consistency across early-phase trials is further supported by isolated case reports describing intramedullary MSC implantation in complete spinal cord injury without emergence of neoplastic or progressive mass lesions during reported follow-up intervals. ¹³

Importantly, interpretation of long-term safety outcomes (Table 2) must distinguish between empirical clinical observations and theoretical biological plausibility.

From a mechanistic perspective, theoretical concerns regarding MSC-associated tumorigenesis primarily arise from observations of genomic instability during prolonged in vitro expansion. Extended culture duration may promote replicative senescence, epigenetic drift, telomere shortening, and, in rare experimental contexts, chromosomal abnormalities or clonal selection events. Such alterations have been described predominantly in preclinical models under non-clinically compliant expansion conditions. Importantly, current Good Manufacturing Practice (GMP)-compliant protocols incorporate strict passage limits, cytogenetic monitoring, and release criteria designed to mitigate these risks. To date, no clinically verified cases of malignant transformation causally attributable to transplanted MSCs have been reported to date, although available follow-up durations remain limited in intramedullary human applications.

More broadly, despite extensive cumulative clinical exposure to culture-expanded MSC products across diverse therapeutic indications worldwide, no reproducible or causally established cases of treatment-induced malignant transformation have been documented in regulated clinical settings within reported follow-up durations.

Additional safety data relevant to intramedullary exposure derive from intraspinal cellular transplantation trials conducted in amyotrophic lateral sclerosis (ALS). Although pathophysiologically distinct from traumatic spinal cord injury, these studies involved direct lumbar and cervical intramedullary injections of mesenchymal stromal cells, bone marrow mononuclear cells, or neural stem cells. Across multiple Phase I and II trials, including extended follow-up analyses, no treatment-attributable tumor formation, secondary intramedullary mass lesions, or delayed proliferative complications have been reported.^9–12

While extrapolation between neurodegenerative and traumatic contexts requires caution, these cumulative observations substantially expand documented human exposure to intraspinal cellular products without identification of a reproducible, treatment-attributable clinical oncogenic signal within currently observed follow-up intervals.

Conceptual implications for cellular therapy risk frameworks

Collectively, these findings support interpretation of spinal cellular transplantation safety within a multidimensional risk framework integrating procedural outcomes (Table 1) and long-term biological considerations (Table 2).

Procedural safety signals appear generally favorable across heterogeneous injection paradigms, while long-term biological risks remain largely theoretical within current clinical observation windows (Table 2). This dissociation between theoretical concerns and empirical safety signals highlights the need for biologically grounded risk interpretation models.

Importantly, the absence of observed tumorigenesis in clinical contexts must be interpreted within the framework of exposure time and biological plausibility. Theoretical oncogenic risk, while biologically conceivable, appears to require conditions (extended culture duration, genomic instability, immune escape) that may not be systematically present in clinically compliant MSC manufacturing protocols.

Limitations

Several limitations should be acknowledged when interpreting the findings of the present review. First, the available clinical literature remains characterized by substantial heterogeneity in study design, patient populations, cellular products, dosing strategies, and injection methodologies. Such variability inherently limits direct quantitative comparability across studies.

Second, reporting standards for injection parameters, including total cell dose, concentration, injected volume, and distribution strategies, were not uniformly applied. In several studies, critical procedural variables were incompletely described, reflecting the exploratory nature of early-phase clinical investigations.

Third, follow-up durations remain relatively limited when considering long-term biological safety outcomes, particularly with respect to theoretical risks such as tumorigenesis or delayed aberrant tissue responses. Absence of observed adverse events within current observation windows should therefore be interpreted cautiously and not as definitive exclusion of rare long-term events. Given the potentially prolonged latency of oncogenic processes, follow-up periods extending beyond 5-10 years may ultimately be required to adequately characterize very late biological risks associated with intramedullary cellular therapies.

Publication bias must also be considered, as early-phase studies reporting favorable procedural safety outcomes are more likely to reach publication than studies reporting severe adverse events or unsuccessful translational experiences.

Finally, the majority of available studies were designed primarily to assess feasibility and procedural safety rather than long-term biological endpoints. Consequently, conclusions regarding biological risk profiles remain probabilistic rather than deterministic.

Despite these limitations, the relative consistency of procedural safety observations across heterogeneous injection paradigms provides meaningful convergent evidence supporting the biological tolerability of controlled spinal cellular administration.

Future perspectives: SVF and minimally manipulated cellular therapies

Although no clinical studies have specifically evaluated intramedullary SVF injection in humans, minimally manipulated autologous cellular products have gained increasing attention in regenerative medicine as alternative approaches to culture-expanded MSC therapies. Unlike ex vivo expanded cellular products, freshly isolated SVF preserves the native cellular heterogeneity of adipose-derived tissue, including stromal progenitors, endothelial lineage cells, perivascular populations, and immune cells within a physiologically preserved cellular microenvironment. Importantly, SVF should not be conflated with mechanically derived stromal tissue products such as nanofat or stromal vascular tissue, which differ substantially in cellular composition, isolation methodology, and biological properties. ¹⁴

From a theoretical safety perspective, SVF introduces a distinct translational framework. Because preparation is performed without prolonged in vitro expansion, SVF may theoretically circumvent risks associated with prolonged ex vivo expansion, including culture-induced phenotypic drift, replicative senescence, and theoretical genomic instability. Immediate preparation and administration may additionally preserve endogenous cellular interactions while reducing manufacturing complexity. This logistical advantage may be particularly relevant in acute neurological settings, where time-dependent therapeutic windows may limit the feasibility of prolonged ex vivo expansion protocols. ¹⁵

Importantly, available clinical data across non-spinal applications have generally reported favorable short-term safety profiles for SVF-based therapies, with adverse events predominantly mild and transient. ⁶ Nevertheless, direct extrapolation to intramedullary administration remains speculative in the absence of dedicated human spinal studies.

Consequently, SVF should not currently be interpreted as an evidence-based alternative to intramedullary MSC transplantation, but rather as a potential future direction within the broader landscape of minimally manipulated cellular therapies. Further preclinical and clinical investigations will be required to determine whether preservation of physiological cellular heterogeneity may translate into distinct safety or biological properties in spinal applications.