Backbond
A product by Optagon LabsETH ZürichUniversity of California, Berkeley
Medical devices

Design SuperiorBiomaterials Fast WithPhysics-Driven Simulation

A design assistant that co-optimizes every requirement at once, finding the right material chemistry, structure and processing parameters.

Create thousands of digital twins, simulate mechanics, degradation, transport and tissue response in the body, then test only the most promising ones.

See the platform
Today's R&D challenge

Trial and error, balancing many conflicting requirements

Open the pores and the strength goes with them. Stiffen it and the body walls it off. Slow the resorption and the ingrowth stalls. Then whatever works still has to be mouldable on the tooling you own, pass biocompatibility, and come out of sterilization unchanged.

Porous polymer implant
for soft-tissue repair
01

The project

A new R&D project is defined.

TARGET IngrowthStrengthFibrosisShelf lifeHandling
02

The targets

Target properties are defined, and they conflict. Fixing one ruins another.

03

The search

Millions of candidate designs, because every composition can be built in many structures. With prototype tooling and bench time, only a few are ever made.

ManufacturabilityBiocompatibilitySterilization
04

The risk

A working design is only the start. These three are targets of their own, optimised for last when the design is hardest to change. Any one can end the program.

The current build and test process

(up to several years)

Requirements

Prototype tooling

Bench test

Biocompat fails

Redesign

Animal study

Fibrosis too high

Redesign

Sterilization fails

Redesign

V&V and launch

What Backbond does

Backbond co-optimizes for all the targets simultaneously

The same four steps, answered. Backbond hands the team a short list of designs that already hit every target and already survive manufacturing, biocompatibility and sterilization.

Target
Ingrowth
Define a Number
Strength
Define a Number
Fibrosis
Define a Number
Shelf life
Define a Number
Handling
Define a Number
01

Define

The product, the targets it must hit, and the parameters to change: pore size, porosity, wall thickness, texture pitch.

02

Simulate

Backbond simulates all possible combinations and evaluates every one.

DesignIngrowthStrengthFibrosisComplianceShelf lifeHandling 1Validate2Confident3Validate4Validate5Confident
03

Rank

The candidate designs are ranked by promise and confidence, and those are not the same thing.

Technical properties checkManufacturabilityBiocompatibilitySterilization
04

Test

The team proceeds to the bench with the most promising designs. By default, these satisfy all targets and constraints.

The Backbond process

(the redesign loops removed)

Requirements

Simulation + focused bench and animal validation

V&V and launch

Simulation removes the redesign loops, and those are where the years go. The submission clock is unchanged, and no software shortens it.

Under the hood

How Backbond simulates a design candidate

Four levels of physics decide what a design actually does, and Backbond reasons about how all four interplay to set every spec.

Design space, comprised of many versions of the design
The candidate, dissected, and where it has to survive
Micro-textured facehexagonal wells, 9 µm pitch, 1.4 µm deepsets fibrosisOpen-pore body180–320 µm pores, 68% porositysets ingrowthDense reinforcement220 µm wall, oriented fibresets strengthCyclic load0.4 MPa, ~8,000 cycles / daySubcutaneous pocketfascial plane, 37 °C, pH 7.4, 24 months
What sets the final properties
Layer physicsmolecules inside one layer
Surface physicswhere two layers meet
Morphologythe shape of the whole part
Environmentthe place it has to survive
Final product properties
Candidate 344
Reasoning chain 1 / 250
LayerPLGA 75:25 hydrolyses in the pocketthe acid it sheds speeds its own breakdown···
Surface9 µm wells resist macrophage spreadingfewer fused giant cells at the face···
Morphology68% porosity admits vesselsthe same porosity drops modulus below fascia···
Environment0.4 MPa cyclic load at the rimmicromotion reopens the wound response···
Together they predict
Fibrotic capsule62 µm ± 15%
Ingrowth at 12 wk74% ± 12%
Strength at 12 mo41 N ± 20%
Use cases

A few example use cases

Every one is the same shape: a handful of targets that fight each other, and one design that has to hit them all.

01Fibrotic encapsulation

The implant walls off in a thick capsule and stops working. Find the surface topography, polymer and modulus that suppress it, while holding strength, sterilizability and cost.

02Porosity against strength

The pores that let tissue in take the strength out. Find the pore size, gradient and strut geometry that hold both, while keeping mouldability and imaging clearance.

03Resorption schedule

The scaffold disappears before the tissue can carry load. Find the copolymer ratio, wall thickness and crystallinity that slow it, while holding ingrowth, local pH and shelf life.

04Burst release

A drug-eluting coating dumps its dose in the first day. Find the matrix, loading and layer thickness that flatten the curve, while holding adhesion, total dose and sterilization survival.

05Sterilization damage

EtO and gamma change the polymer the design depends on. Find the material and cure that survive both, while holding modulus, degradation rate and residual limits.

06Stiffness mismatch

The device is stiffer than the tissue and erodes into it. Find the material and cross-section that match compliance, while holding fatigue life, deliverability and radiopacity.

07Coating delamination

The coating cracks off after repeated flexing. Find the primer, thickness and cross-link density that hold adhesion, while keeping lubricity, elution profile and biocompatibility.

08Shelf-life ageing

The device drifts out of spec before it is ever opened. Find the polymer, barrier and packaging that hold it to expiry, while keeping flexibility, sterility and cost.

09Calcification

The leaflet stiffens with calcium and fails early. Find the treatment chemistry and surface state that resist it, while holding durability, thrombogenicity and crimp recovery.

10Grade discontinued

The medical-grade resin is delisted. Find a replacement and process window that hold every validated spec, without reopening biocompatibility or sterilization.

What it makes possible

The future of the device R&D organization

How Optagon Labs imagines the device company of the future, drawn from work with hundreds of R&D teams: one brain coordinating specialised agents that work in parallel, running development end to end across the portfolio.

One central brain

Coordinating work across every function

New products

Designing into a category the company has never built in

a first resorbable line

Putting a known material into a form it has never survived

a moulded part from a cast polymer

Designing for a demand that arrived faster than the lab could

a same-day-discharge variant
Benchmark matching

Reverse-engineering a competitor’s device and beating it on one axis

a rival’s 5× fibrosis claim

Matching a predicate’s performance from a different material

equivalence without their polymer

Hitting a customer’s target spec from their brief alone

a scaffold at reference modulus
Tissue interface

Designing the surface that decides how the body answers

wells that stop the capsule forming

Raising integration without changing the bulk material

twice the ingrowth, same polymer

Holding the release profile when the geometry changes

thinner wall, same elution curve
Degradation in service

Designing for a life the current material cannot reach

24 months under cyclic load

Protecting a part from what its own breakdown does to it

acid autocatalysis in the core

Holding strength through a process that destroys it

modulus through EtO and ageing
Process & scale-up

Designing processing for a new part while preserving customer machinery

porous structure on existing tooling

Adapting a design to a different plant’s process and equipment

machined to moulded

Correcting a design that works at bench and fails at volume

porosity drift across the cavity
Redesign under change

Finding a new material that holds spec after a cost shift

holding modulus after a resin swap

Redesigning fast when a supplier discontinues a grade

life after a medical-grade delisting

Absorbing lot variability without losing spec

molecular weight drift between lots
Materials restriction

Removing a material the regulator or the market rejects

a PFAS-free low-friction surface

Replacing a plasticiser at the same mechanical performance

DEHP-free, same flexibility

Meeting a restriction across a whole platform

one pass, every SKU
Regulatory variants

Redesigning one device to clear a second market’s rules

a 510(k) device for EU MDR

Adjusting to a different standard’s test method

the same part against a new fatigue spec

Substituting to locally qualified materials without redesigning

a domestic resin, same spec
No precedent

Designing where the company has no data and no history

a first drug-device combination

Standing up a new application from the physics and the literature alone

a texture that holds through sterilization