Aim the shunt tube where the fluid is
A digital twin built from a patient’s CT picks a better entry point and depth than the usual fixed Keen’s-point mark — so more of the tube floats freely in cerebrospinal fluid instead of getting buried in tissue.
Why it matters
When a brain shunt tube clogs, patients often need another operation; freehand aiming from a fixed scalp mark leaves the tip floating freely in fluid in only about one case in ten. This study shows that choosing entry and depth from each person’s own CT — still executed freehand with a ruler — put the tip free in fluid in 10 of 11 adults, versus 22 of 213 in older records.
The paper
Feasibility of a Patient-Specific Digital Twin for Ventriculoperitoneal Shunt Trajectory Planning
Kritsanavis Chongsrid, Vich Yindeedej, Supaporn Konmun, Xulei Yang & Si Yong Yeo. arXiv:2610.20074 [cs.CV], submitted 14 Oct 2026. Read on arXiv · PDF
In plain language: catheter obstruction is the commonest cause of shunt failure. The team rebuilt skull, scalp, ear and ventricles from preoperative CT with deterministic geometry, searched every entry around Keen’s point for the perpendicular path with the longest stretch inside the ventricle under safety limits, and handed the surgeon two ruler measurements from the ear plus two depths. In an IDEAL Stage 2a series of 11 consecutive adults, freehand execution scored Hayhurst Grade 1 in 10 of 11 versus 22 of 213 historically (p = 9×10⁻⁹), with median entry deviation 2.5 mm, angular deviation 5.9°, 1.4 cm more intraventricular catheter than Keen’s point, and no intraoperative complication. The technique itself was unchanged; the gain came from choosing entry and depth per patient. Whether that lowers revisions is the next-stage hypothesis.
Try it
Illustrative side-view toy: fixed Keen’s-point aiming versus a patient-specific entry and depth. Numbers lean on the paper’s Grade-1 rates (22/213 vs 10/11) and +1.4 cm intraventricular length. Not surgical guidance and not the authors’ software.
The maths in this page
The digital twin searches candidate entries around Keen’s point and keeps the perpendicular trajectory that maximises intraventricular catheter length under safety constraints. This toy approximates that idea on a 2-D side view:
Hayhurst Grade 1 in the paper means the tip sits free in CSF. In the toy, Grade 1 is awarded when the tip lies inside the ventricle with a small clearance from the wall; otherwise the tip is treated as buried (not Grade 1).
- Ventricular size: scales the ellipse from small to dilated; larger ventricles make even Keen’s aiming more forgiving.
- Keen’s offset error: shifts the fixed landmark trajectory so the tip often clips cortex or ventricle wall — echoing the ~1-in-10 free-tip rate.
- Planned entry offset & depth: stand in for the paper’s per-patient ruler plan; the default (+8 mm, 6.5 cm) is shaped so the optimized path typically gains about +1.4 cm of IV length versus Keen’s, matching the reported median gain.
- Study numbers (not toy): Grade 1 in 10 of 11 vs 22 of 213 historically (p = 9×10⁻⁹); median entry deviation 2.5 mm, angular deviation 5.9°; no intraoperative complication.
This is an illustrative geometry toy grounded in published rates and the +1.4 cm IV-length gain — not the authors’ CT pipeline, not a planner, and not advice for surgery.
How to use it
Start with a mid-size ventricle and a few millimetres of Keen’s error: watch the orange fixed path skim or miss the fluid space while the green plan keeps a longer stretch inside. Shrink the ventricle to see how a fixed landmark fails more often, then nudge planned offset and depth to recover Grade 1 and the extra centimetre-plus of tube in fluid. Toggle the digital-twin path off to compare Keen’s alone. The takeaway from the paper: the freehand technique stayed the same; choosing entry and depth from each patient’s CT is what moved Grade 1 from roughly one in ten to ten of eleven — whether that cuts revisions is still an open next-stage question.