The Grown Throat

The Weight of a Missing Tube

The condition is called oesophageal atresia. The oesophagus grows in two disconnected segments during foetal development, and the gap between them can be too wide to bridge surgically. In roughly 8% of cases, the gap exceeds three centimetres. These are classified as long-gap oesophageal atresia, and they are the ones that break the surgical playbook.

For shorter gaps, surgeons can stretch the two ends toward each other and stitch them together. For long gaps, the options are harder. The Foker procedure uses traction sutures to stimulate the oesophagus to grow, pulling the two ends gradually closer over days. Jejunal interposition replaces the missing section with a piece of small intestine, repurposing tissue that was designed for a different job. Gastric pull-up drags the stomach upward to meet the remaining oesophagus. All of these work, in the sense that the child can eat. None of them are an oesophagus.

A child who receives a jejunal interposition has a food pipe made of intestinal tissue. It doesn't contract in the same peristaltic waves. It doesn't resist acid the same way. It doesn't grow at the same rate as the child's body. The surgery saves lives, and the trade-offs accumulate over a lifetime.

Professor Paolo De Coppi, a paediatric surgeon at Great Ormond Street Hospital, has spent years looking at those trade-offs. The question his team at UCL's Great Ormond Street Institute of Child Health has been working toward is whether you could build a replacement that isn't a workaround. An actual oesophagus, grown from the patient's own cells, that swallows and grows.

In March 2026, they published the answer in Nature Biotechnology.

What the Scaffold Remembers

The technique begins with a donor oesophagus, in this case from a pig. The pig's cells are flushed away over ten days using a precise sequence of detergents and enzymes. What remains is the extracellular matrix: the collagen scaffolding, the structural proteins, the physical architecture of the organ without any of the immunogenic material that would trigger rejection. Think of it as a building emptied of its inhabitants but with every wall, pipe, and staircase intact.

The scaffold retains something that no synthetic tube can replicate. The oesophagus has two distinct muscle layers, an inner circular layer and an outer longitudinal layer, that contract in coordinated waves to push food downward. The architecture of those layers, the angles of the fibres, the channels where blood vessels ran, the attachment points for nerves, all of that geometry is preserved in the matrix. The scaffold doesn't just have the right shape. It has the right instructions, encoded in its physical structure, for new cells to follow.

The scaffold is then repopulated with the recipient's own cells. Muscle progenitor cells are taken from a small biopsy, expanded in the laboratory, and injected directly into the scaffold. These aren't stem cells in the embryonic sense. They're adult progenitor cells, already committed to becoming muscle, coaxed into multiplying and then placed into an environment that tells them exactly where to go and what to become.

The cells migrate along the collagen fibres. They find the muscle layers. They orient themselves correctly, circular in the inner layer, longitudinal in the outer. Over weeks, they form functional muscle tissue that can contract. The scaffold's architecture does the teaching.

What Happened When It Was Implanted

The De Coppi team created 2.5-centimetre oesophageal grafts and implanted them into eight recipient pigs, replacing a full section of each animal's oesophagus. The pigs were young and still growing, which was the point: any replacement for a child's missing oesophagus has to work not just at the moment of surgery but as the body changes over months and years.

The results were striking. All eight animals recovered and developed working swallowing muscles. The grafts squeezed food downward toward the stomach in coordinated peristaltic waves. Full integration of the engineered tissue occurred within three months. No immunosuppression was needed at any point, because the implant was built from the recipient's own cells on a matrix stripped of all donor immunogens.

Previous work in regenerative medicine had demonstrated individual components: decellularisation of oesophageal tissue, cell seeding on scaffolds, short-term function in small animal models. What made this study different was that it combined the entire pipeline, from scaffold preparation to cell expansion to surgical implantation to long-term functional assessment, in a large animal model that mimics the growth dynamics of a child. The grafts didn't just work on the day of surgery. They grew with the animal.

That last point matters more than any benchmark. A static implant that functions at six months but cannot expand as the child doubles in size would be another surgical workaround. What De Coppi's team showed is that the tissue remodels: the cells continue to divide, the scaffold is gradually replaced by the body's own extracellular matrix, and the organ scales. It behaves like native tissue because, in every biological sense that matters, it has become native tissue.

What the Scaffold Promises

The immediate clinical path leads to children with long-gap oesophageal atresia. The team at GOSH is working toward human trials, though the timeline depends on regulatory approvals and the scaling of the cell-expansion process. A baby born today with a gap too wide to bridge surgically would still receive one of the existing procedures. But the trajectory is clear: within years, not decades, there may be an option that gives that child an actual oesophagus rather than a rearrangement of other tissues.

The broader implications reach beyond the food pipe. The technique, decellularisation plus autologous cell seeding plus growth in the recipient, is not organ-specific. The scaffold's genius is that it encodes the organ's architecture in its physical structure, and that encoding is preserved through the decellularisation process. If the approach works for the oesophagus, a tube with two coordinated muscle layers and a specialised lining, there is reason to believe it could work for other tubular organs: trachea, ureter, intestinal segments. Each would require its own cell types and its own validation, but the platform is transferable.

There's something quietly radical about the underlying logic. Traditional organ transplantation is a supply problem: too few donors, too many patients, and a lifetime of immunosuppressive drugs for those who receive a transplant. The De Coppi approach sidesteps all three constraints. The scaffold comes from an animal, available in unlimited supply. The cells come from the patient, eliminating rejection. And the organ grows with the body, meaning a single procedure could last a lifetime.

A 40-gram tube. That's what the team grew. It swallows, it contracts, it fights acid, it scales. It is, in every way that matters to the child who needs it, an organ. Not a prosthesis. Not a graft made from something else. An organ, built from the body's own cells on a scaffold that remembered how to be an oesophagus even after every cell it once contained was gone.