Two Stem Cell Approaches to Parkinson’s. Fifteen Patients. Zero Tumors. The Probability of That Happening by Chance Is 8.7%.
A Swedish embryonic stem cell trial and a Japanese induced pluripotent stem cell trial have independently produced Phase I/II safety data for lab-grown dopamine neurons transplanted into Parkinson’s patients. Combined across both studies: zero tumors, zero graft-induced dyskinesias, and preliminary evidence of graft survival in 15 subjects. We ran the binomial probability against the historical 15% GID rate and built a break-even cost model for what a one-time cell therapy could replace.
Zero. That is how many tumors were found across two independent stem cell trials for Parkinson’s disease involving 15 patients, three surgical centers, and two fundamentally different cell sources.
It is also the number of graft-induced dyskinesias, the runaway involuntary movements that torpedoed earlier fetal-tissue transplants in the early 2000s and frightened an entire research community away from cell replacement therapy for a generation. The new data suggests that the problem was never the concept but the cells themselves.
The STEM-PD trial, led by Malin Parmar at Lund University and Roger Barker at Cambridge, transplanted dopaminergic progenitors derived from the RC17 human embryonic stem cell line into eight Parkinson’s patients across two dose cohorts. Results published in Nature Medicine this month report that after 12 months, six of seven evaluable patients had substantially reduced their dopaminergic medication. No serious adverse events were observed, and MRI showed no graft overgrowth. One participant died from a pulmonary infection assessed as unrelated to the cell product. This was the first pluripotent stem cell therapy approved for Parkinson’s in Europe, and updated results were presented at the International Society for Stem Cell Research annual meeting on July 10.
Separately, a team at Kyoto University Hospital reported 24-month data from a parallel iPS cell trial in Nature in April 2025. Seven patients received bilateral transplants of dopaminergic progenitors derived from a clinical-grade induced pluripotent stem cell line, sourced from a healthy donor whose HLA type matches roughly 17% of the Japanese population. PET imaging showed a 44.7% increase in dopamine uptake in the putamen. Four of six evaluable patients improved on the standard motor assessment in the OFF state by an average of 9.5 points, a 20.4% improvement. Seventy-three adverse events were recorded, all mild or moderate, and none were classified as serious. Zero tumors.
Nobody has placed these two datasets side by side. Here is what that comparison reveals.
| Dimension | STEM-PD (Lund/Cambridge) | Kyoto iPS Trial |
|---|---|---|
| Cell source | Human embryonic stem cell line RC17 | Induced pluripotent stem cell line QHJI01s04 |
| Patients enrolled | 8 (7 evaluable at 12 months) | 7 (6 evaluable for efficacy) |
| Follow-up | 12 months | 24 months |
| Tumors | 0 | 0 |
| Graft-induced dyskinesias | 0 | 0 |
| Serious adverse events | 0 | 0 |
| Graft survival evidence | Dopamine PET signal at 6 and 12 months | 18F-DOPA Ki increased 44.7% in putamen |
| Motor improvement | Stable; medication reduced in 6 of 7 | MDS-UPDRS III OFF improved 20.4% |
| Immunosuppression duration | 12 months | 15 months (tapered from 12) |
| Scalability | Unlimited cell supply (one ESC line) | HLA-matched iPS covers 17% of Japan; global scaling requires cell banks |
| Ethical consideration | Requires embryo-derived cell line | Reprogrammed from adult donor blood; no embryo |
The GID Math Nobody Ran
Historical fetal tissue transplants carried a roughly 15% rate of graft-induced dyskinesias, based on the pivotal double-blind trials by Freed et al. (2001, NEJM) and Olanow et al. (2003, Annals of Neurology). Those runaway movements were the primary reason the field abandoned cell transplantation for Parkinson’s disease.
Combined across both new trials: zero GIDs in 15 patients. We ran the binomial probability.
P(0 in 15 | true rate = 15%) = (0.85)15 = 8.7%.
If the underlying GID risk were still 15%, you would see this clean result less than one time in eleven. That is not enough to claim statistical significance at the conventional threshold (p = 0.087 exceeds 0.05), but it is suggestive. At 20 patients with zero GIDs, the p-value drops below 0.04, and both trials are still enrolling.
Cell purity is the most probable explanation for the disappearance of graft-induced dyskinesias. The Kyoto group explicitly measured graft composition: approximately 60% dopaminergic progenitors, 40% dopamine neurons, and critically, zero serotonergic neurons. Serotonergic contamination has long been the leading suspect in GID causation because serotonin neurons can co-release dopamine in an unregulated fashion, flooding the striatum. Modern differentiation protocols appear to have eliminated the contaminant that destroyed the field two decades ago.
What Would a Cure Cost?
Parkinson’s disease costs the US economy roughly $52 billion per year across approximately one million patients. Medicare costs per patient run $32,883 to $41,934 annually, a figure that roughly doubles as the disease advances from early to late stage. Advanced patients on continuous duodenal levodopa infusion spend approximately €41,000 per year on medication alone. Deep brain stimulation, the most aggressive current surgical option, costs approximately €53,000 over five years including implantation, follow-up, and battery replacement.
Suppose a stem cell therapy replicated STEM-PD’s medication reduction at scale: substantial relief in 85% of patients lasting ten years. Annual savings per patient would fall between $25,000 and $35,000. Over a decade, that is $250,000 to $350,000. Add the avoided cost of deep brain stimulation for the subset who would have needed it, and the break-even price for a one-time cell therapy lands between $200,000 and $300,000.
That range sits within the precedent set by existing cell and gene therapies. Novartis prices Kymriah at $475,000 for CAR-T treatment of leukemia. Vertex charges $2.2 million for Casgevy, a gene-editing therapy for sickle cell disease. A Parkinson’s cell therapy at $250,000 would be a fraction of those prices while addressing a patient population measured in millions rather than thousands.
But these are trials of eight and seven patients. Phase III is years away.
The Strongest Case Against
Every cell therapy looks miraculous in Phase I. Fetal tissue transplants looked miraculous too. The open-label trials of the 1990s showed robust dopamine synthesis and clear motor improvement in patients who had suffered for years, and the neuroscience community celebrated what appeared to be a breakthrough. Then the double-blind, sham-surgery-controlled trials arrived, and the effect largely vanished.
Parkinson’s disease carries one of the largest placebo responses in all of medicine. Sham surgery alone has produced 10% to 30% improvement in motor scores across multiple controlled trials. Neither STEM-PD nor Kyoto used sham surgery controls; both were open-label. The reported medication reductions could partially reflect post-surgical expectation, the increased clinical attention that trial participants receive, and the natural motor fluctuations inherent to a disease that waxes and wanes over months.
Scaling the technology globally introduces its own formidable problems. The Kyoto trial’s HLA-matching approach covers only 17% of the Japanese population. Reaching global coverage would require cell banks with dozens of HLA-matched lines, a logistical and regulatory project that has barely begun. BlueRock Therapeutics, pursuing an autologous iPS approach under Bayer’s ownership, estimates per-patient costs between $500,000 and $1 million.
Limitations
Both trials were powered for safety, not efficacy. Every efficacy signal in this article is exploratory and should not be interpreted as evidence of clinical benefit. STEM-PD reports only 12 months of follow-up; graft maturation in the fetal transplant literature typically requires two to three years, so the current data may substantially understate or overstate the long-term picture. Kyoto has 24 months of data but discontinued immunosuppression at 15 months, and whether grafts survive long-term without ongoing immune protection remains unknown. Our break-even cost model assumes that medication reductions persist for a decade, which no data yet supports. The binomial GID calculation treats both trials as independent draws from the same patient population, ignoring meaningful differences in cell source, demographics, surgical technique, and transplantation protocols that could confound a pooled analysis.
The Bottom Line
Two fundamentally different stem cell platforms have produced simultaneous safety data in Parkinson’s disease that, for the first time, permits a direct head-to-head comparison. The combined signal is clean enough to sustain investment in larger trials. It is not clean enough to promise anyone a cure.
If you or someone in your family has Parkinson’s, the practical question is whether to pursue trial enrollment. STEM-PD operates at sites in Lund and Cambridge. The Kyoto trial recruits in Japan. In the United States, BlueRock Therapeutics (NCT04802733) is running a separate ESC-derived dopamine cell trial that has already dosed initial cohorts. All three programs list current enrollment status on ClinicalTrials.gov.
If you are tracking the field from a distance, watch for one number: 20. That is how many combined patients with zero graft-induced dyskinesias it takes to push the binomial p-value below 0.05, crossing the line from suggestive to statistically significant. We are at 15 today. Five more clean results and the safety question has a real answer.