Adolescent Spinal Curvature Brace Optimization

The problem

Adolescent idiopathic scoliosis is a sideways curve and twist of the spine that appears in late childhood or adolescence and accelerates while the child is still growing. The usual treatment for a progressing curve is a custom brace that presses on the torso to steer the spine straighter as it grows; when bracing fails, what remains is spinal fusion surgery.

The brace itself is fitted by hand. An orthotist starts from an X-ray and body measurements, carves or generates a brace, trims it at a first fitting, then takes another X-ray with the brace on and adjusts from what it shows. Nothing in that loop measures how hard the brace presses, or where, so every adjustment rests on experience and the patient's report. Each iteration costs time and money, and a child can grow through the window in which a brace works best before the brace is right.

1. 1X-ray and body measurementsThe starting data for a new brace, taken without a brace on.
2. 2Brace made by handCarved from those measurements, or generated algorithmically and then trimmed at a first fitting.
3. 3X-ray with the brace onThe only check of how the brace is actually affecting the spine.
4. 4Adjustments by judgementNo instrument measures the pressure the brace applies, so each adjustment rests on experience and the patient's report, and the loop repeats about yearly.
5. back to the start

How it started

I went looking for a capstone that would combine medicine and engineering. Through my co-op program I had met an orthotist, and his vision of fitting braces from measured pressure rather than judgement became the project. I pitched it to classmates, and a six-personicapstone team sizeThe team's members are named in SFU's article on the project.Source: Resume, revised 2026-09-11 team formed around it, with the orthotist as our industry collaborator. SFU's School of Engineering Science wrote about the project in July 2022; the article, linked in the facts column, names the whole team and carries our photograph.

What we built

The system is an instrument for the missing measurement: the pressure between torso and brace, read continuously and shown to the orthotist while the brace is being fitted.

1. 1Sensing garmentA shirt carrying an array of force sensors, worn under the brace so each sensor sits between torso and brace.
2. 2Microcontroller firmwareReads every sensor repeatedly and averages the readings to suppress noise, then sends the digitized values to a computer over USB.
3. 3Desktop applicationConverts readings to pressure and paints them, live, as a heatmap on a three-dimensional torso model; a capture is saved as a spreadsheet file for design tools.
4. 4OrthotistReads where the brace presses hardest and adjusts the fit from measurements rather than guesswork.

• A sensing garment. A shirt carrying an array of force sensors, worn under the brace so each sensor sits between torso and brace. A later revision replaced the discrete sensors with sheets of pressure-sensitive material on a fixed grid, so the torso under the brace is covered and every reading has a known position.
• Noise-filtering firmware. The microcontroller samples every sensor repeatedly and averages the readings before sending them over USB, which suppresses noise in the analog signal. The team replaced a generic firmware library with its own, which scanned the sensors faster and made temperature drift negligible.
• A real-time heatmap on a CAD torso. A desktop application converts the readings to pressure and paints them live onto a CAD model of a torso, with a legend, so the orthotist sees where the brace presses hardest while the patient is wearing it. A capture is saved in a plain spreadsheet format that design tools such as nTopology can import. The team was explicit that the visualization validates the data and guides the fit; it does not diagnose.

The software was built with medical-device software standards in mind, IEC 62304 for the software lifecycle among them, and the team's own analysis placed the device in Health Canada's lowest-risk, non-invasive class.

Results

The team won Best Overall Project at ICAMES 2022, the International, Cultural, and Academic Meeting of Engineering Students, held at Boğaziçi University in Istanbul in May 2022. At the competition, presenting the project's potential future impact was my part. When SFU wrote about the project that July, the team's next milestone was a prototype that could be tested in clinic to begin collecting patient data.

What was proposed, not built

The decks close with extensions, and they are only that. One was a topology-optimized, printed brace with openings for ultrasound, so that check-ins could happen more often without X-rays. The other was a machine-learning model trained on the pressure data the system collects, to predict curve severity and progression and to suggest brace designs. Neither was built during the capstone; the pressure-sensing system above is what exists.

Sources

This page draws on the team's competition and final presentations and on SFU's news article. The presentations are not published here, and every figure on this page is an original drawing.