Setting up a liver model programme

The problem is usually posed as a capability, not an experiment: “we need a human liver model”. Behind it sit two different pressures. Either a compound has failed on liver signal and nobody can say whether the finding is human-relevant, or a programme has decided it wants hepatic toxicity data earlier and cheaper than the animal study currently provides.

Both lead to the same trap. The routes available differ by a factor of forty in per-experiment cost and by a factor of ten in setup time, and the expensive one is the one that gets specified first because it is the one people have read about.

This page is the staged route out of that. The short version: buy a $4,280 ready-to-use spheroid plate before you buy anything else, because it is the only line in this category with a published price, it needs no instrument, and it will tell you within eight weeks whether the expensive routes are worth pursuing.

Step 1 — Name the question, because it selects the model

Four routes exist and they are not substitutes for one another.

The single most useful question to put to the requester: does the toxicity you are worried about appear after one dose or after repeated dosing over days to weeks? Acute is 2D. Repeated is 3D. That one answer eliminates half the catalogue.

Two published anchors are worth knowing because they are the reason the 3D route exists at all. A 2016 characterisation of primary human hepatocyte spheroids established them as a system for drug-induced liver injury, liver function and disease work. A multicentre study directly comparing hepatic 2D sandwich cultures against 3D spheroids for long-term toxicity applications did the head-to-head. And a 2017 study assessed the utility of spherical human liver microtissues specifically for predicting clinical DILI.

Read those for what they establish about the model class. Do not read the reported sensitivity and specificity figures as transferable to your compound set — they are properties of the compound sets used in those studies.

Step 2 — The only published price in the category

This is the fact that should shape your first purchase, and it is unusual enough to state plainly: almost nobody in the liver model category publishes a price, and the one supplier who does publishes a complete, self-contained unit.

Three things follow.

The $4,280 is the complete unit, not a tissue price with a medium premium bolted on. The bundle is the microtissues and the appropriate medium at 250 mL. There is no separate cell line to add, because the cells are inside the spheroids. That is why it is directly comparable against nothing else in this market — every other route quotes you a component.

The per-microtissue figure is $44.58, and the plate underneath it is $0.74. Read those two numbers together. The plastic is 1.7 percent of the delivered cost. If you are considering building your own spheroids on Akura plates, the plate is not what you are saving on; the hepatocytes and the aggregation labour are, and primary human hepatocytes have no public price at any supplier we surveyed.

Primary human hepatocytes cannot be costed from public sources. We checked. If your design requires them, that line is quote-only and you should get the quote before you build a budget around it rather than after. The full supplier picture is in liver model suppliers compared.

On the perfused chip route

Published list prices exist for chip consumables — AIM Biotech’s idenTx 40 Plate is $2,220.00 per box of 5, which is $444.00 per plate and $11.10 per chip, and the organiX Plate is $410.00. Those are real, live-verified list prices captured 2026-09-01.

The instrument is a different matter, and we are going to be blunt about the evidence. Federal procurement records give us $20,900.00 per 24-pack of Liver Human Emulation System quad liver chip kits — roughly $870.83 per kit unit — DERIVED from a 2023 FDA award for a stated 5 each. But two awards describing an identical two-instrument-plus-two-module configuration differ by 48 percent, so those records establish only an envelope of roughly $147,000 to $354,000 for a Human Emulation System installation, not a price.

An envelope that wide is not a budget. It is a reason to get a quote. The full derivation, and the equivalent figures across platforms, are in organ-on-chip platform costs.

One published study is worth reading before anyone signs a capital request: a 2022 performance assessment and economic analysis of a human Liver-Chip for predictive toxicology. Whatever you conclude from its economics, it is the only published attempt at the calculation and it will sharpen the internal argument in both directions.

Step 3 — The staged budget

The structure below defers the expensive, irreversible decisions until cheap experiments have justified them. Each stage has an explicit go/no-go.

This is a MODELLED budget. Every unit price is real, sourced and dated. Every quantity and stage boundary is our assumption. It is not a quotation.

Stage 1 — Prove the assay works, in your hands, on a plate reader

What Stage 1 buys you: 96 wells, one microtissue per well, stable for up to four weeks at high viability with consistent albumin secretion and stable CYP activity over that period, per the supplier’s own specification. The supplier guarantees at least 95 percent of microtissues available on arrival, confirmed by optical scan of each plate at packing.

What it needs from you: an incubator, a plate reader, and pipetting discipline. No capital request. No instrument procurement. No method development on cell aggregation.

Go/no-go: if you cannot generate a clean dose-response with an acceptable control separation on this plate, the problem is your assay or your compounds, and buying a chip platform will not fix it. Stop here and fix that first.

Stage 2 — Establish routine throughput

The most valuable line in that table is the cheapest one. Do the rehearsals — pipetting, medium exchange technique, plate handling, readout timing — on a cell line before spending $4,280 on a plate you will use to learn how to pipette into a microwell. That advice is uncomfortable and it is the single highest-return item on this page.

Go/no-go: if throughput is now the limit and the biology is answering your questions, you do not need a chip. If the questions that remain are specifically about flow, shear or inter-organ interaction, proceed to Stage 3.

Stage 3 — Only now, evaluate perfusion

Why this ordering saves money. Stage 1 and Stage 2 together are roughly $21,800 of MODELLED materials — less than a sixth of the low end of the instrument envelope alone. If Stage 1 fails, you have spent $4,348 finding out. If Stage 2 answers the programme’s questions, you never spend the capital at all. And if you do proceed to Stage 3, you arrive with two years of internal data, a working assay and a much stronger negotiating position, because you can tell the vendor exactly what the chip has to beat.

The comparison across all three formats — static, insert-based and perfused — is in organ-on-chip vs Transwell vs static culture.

Step 4 — The timeline

The scheduling fact that catches people: liver microtissues are gated by the supplier’s published production schedule, at 2-6 weeks. They are not a stock item. If the experiment has a date, place the order against a named production week and confirm it in writing. See realistic lead times for how this behaves across the category.

The corollary for Stage 2: order plate N+1 while you are running plate N. Sequential ordering turns four campaigns into eight months of mostly waiting.

Step 5 — What breaks

Comparing a bundle against a component. The $4,280 plate includes the cells and the medium. A perfused-plate quote does not, and a primary-hepatocyte quote definitely does not. Comparing those headline numbers directly makes the bundle look expensive when it is the only complete unit in the comparison. This is point three of the quote normalisation worksheet and the liver category is where it bites hardest.

Buying plates when you needed microtissues. Akura 96 plates at $0.74 per well and a liver microtissue plate at $44.58 per microtissue are not alternatives. The first is empty plastic. If you buy the plastic you still have to source hepatocytes, and primary human hepatocytes have no public price at any supplier we surveyed.

Assuming the 384 format is a straight scale-up. Akura 384 plates are $0.36 per well against $0.74 for the 96 — genuinely cheaper per well — but the microtissue product itself is specified and priced on the 96 format, with the 384 human product available on request and no public price. A format change is a re-quote, not a multiplication. See scaling from pilot to screening.

Reading published DILI prediction performance as your own. The sensitivity and specificity figures in the cited literature belong to the compound sets those studies used. Your compounds are different, your concentrations are different, and your endpoint may be different. Use the papers to justify the model class; do not put their percentages in your own protocol.

Forgetting the species panel exists. Dog microtissues at $9,140 and mouse at $9,680 are published, live figures. If your problem is a cross-species discrepancy — animal signal, unclear human relevance — running the human and the animal microtissue side by side is a directly purchasable experiment, and it is the experiment that most often resolves the argument. Note that both animal bundles are two plates rather than one, which is why the totals are roughly double at a similar per-microtissue price.

Skipping mycoplasma testing because you bought a finished product. If you culture anything yourself in the same incubator, test it. An ATCC PCR kit is $167.00 with a 3-5 business day turnaround, which is trivial against a $4,280 plate sharing an incubator with a contaminated flask.

The specification to send

Model and question

• Route: 2D sandwich / 3D spheroid / perfused chip / cell line
• Question: acute cytotoxicity / repeat-dose DILI / metabolism and clearance / flow-dependent biology
• Exposure duration and number of dosing events
• Species required: human only / human plus [dog, mouse]

If sourcing finished microtissues

• Cell composition: hepatocyte donor count and pooling, and which non-parenchymal cell types are present and from how many donors
• Stated functional stability window, and the functional markers it is stated against (albumin secretion, CYP activity)
• Medium configuration required: TOX / AF / INF or supplier equivalent, and the volume included
• Number of plates and the well format
• Arrival guarantee: percentage of microtissues available on arrival, how it is verified, and the remedy if it is not met
• Production week and ship day, in writing

If sourcing cells to build your own

• Cell type: primary human hepatocytes / iPSC-derived hepatocytes / HepaRG
• Quantity, and whether the count is total at freeze or viable post-thaw
• Lot: entire quantity from one lot, reserved
• Plateability specification and the method it is measured by
• Aggregation plate: format, well count, cavity geometry
• Media, supplement and coating: included or excluded, quoted separately if excluded

If evaluating a chip platform

• Instrument, module, controller and software licence, quoted as separate lines
• Consumable price per plate and per chip, at one box and at annual volume
• Whether the plate is supplied empty or pre-seeded, and the price of each
• Minimum viable configuration to run one experiment, stated explicitly
• Training and installation, in or out
• Repeat-order consumable price with the instrument already in place

Commercial, all routes

• Certificate of analysis and the QC panel included at this price
• Remedy for material failing its release specification on arrival, and the claim window
• Use rights: internal research, fee-for-service, or commercial derivative
• Repeat-order price at your projected annual volume

Where we fit

We put that specification to named suppliers and normalise the responses onto one basis — per microtissue, per well, per experiment — which in this category means adding the cells back into the quotes that excluded them and pulling the instrument out of the quotes that buried it.

We do not manufacture and we do not resell any platform, which is why we can tell you that the first purchase should be a $4,280 plate rather than a $200,000 instrument, that the rehearsals belong on a cell line, and that if the remaining question is not specifically about flow then the perfusion stage should never be reached at all.