Organ-on-chip and lung-on-chip literature scan, 2026

Date: 2026-08-27 Scope: PubMed, publication date 2026 (lung-specific set) and 2026-02 onward (general organ-on-chip set). Purpose: survey where this literature is being published, and triage which papers carry mechanism claims curatable into dismech.

Search

Two PubMed queries, both restricted to title/abstract:

Lung set — lung-on-a-chip OR lung-on-chip OR "lung chip" OR airway-on-a-chip OR alveolus-on-a-chip OR "alveolus chip", 2026. 49 papers, all retrieved.
General set — organ-on-a-chip OR organ-on-chip OR organs-on-chips OR microphysiological, 2026-02 onward. 703 papers, first 500 retrieved and analysed.

The general query's microphysiological clause is what makes it large; that word is now standard in regulatory and toxicology writing, so the set includes many papers that are about the regulatory status of these models rather than about any disease.

Where this work is published

There is no home journal. This is the single most useful thing the scan establishes.

In the general set, 500 papers are spread across 254 distinct journals. 175 of those journals contributed exactly one paper. The twelve most frequent venues together account for only 28% of the set. The lung subset is more extreme still: 49 papers across 43 journals, with no journal contributing more than two.

Most-frequent venues, general set (2026-02 onward, n=500):

n	Journal
23	bioRxiv (preprint)
21	Lab on a Chip
15	Frontiers in Bioengineering and Biotechnology
13	Advanced Healthcare Materials
12	Biofabrication
12	Advanced Drug Delivery Reviews
11	International Journal of Molecular Sciences
9	Advanced Science
7	Small
7	Micromachines
6	Acta Biomaterialia
6	Frontiers in Toxicology

These fall into four groups, and the grouping matters more than the ranking:

Microfabrication and biomaterials — Lab on a Chip, Biofabrication, Advanced Healthcare Materials, Small, Acta Biomaterialia, Materials Today Bio. The largest group. Papers here are usually reporting a device; the disease application is a validation experiment at the end.
Toxicology and drug delivery — Advanced Drug Delivery Reviews, Frontiers in Toxicology, Toxicological Sciences, Journal of Applied Toxicology, Archives of Toxicology, International Journal of Pharmaceutics. Driven by the shift away from mandatory animal testing; these papers have real exposure–response data.
Clinical and organ-specific journals — European Respiratory Journal, Thorax, Journal of Clinical Investigation, Transplantation, Respiratory Research, American Journal of Respiratory Cell and Molecular Biology. The smallest group but the highest mechanistic yield.
Preprints — bioRxiv is the single largest venue in the general set.

Consequences for how we monitor this field. Following a journal list will not work; a keyword-based PubMed sweep is the only viable approach, and it must include bioRxiv. Roughly half the returned records (259 of 500 in the general set, 22 of 49 in the lung set) are reviews, so any automated scan needs a Review[pt] filter or it will spend most of its budget on secondary literature.

A caution the scan surfaced directly. One 2026 record in the lung set is a retraction notice for a lung-on-chip influenza/pneumococcus co-infection paper (PMID:42481340). Anything harvesting this literature automatically should check publication type before citing.

Triage: which papers carry curatable mechanism

Of the 49 lung papers, most are device or platform reports whose disease content is a demonstration rather than a finding. Eleven had a substantive, quotable mechanism claim tied to a specific disease. Those are listed below with what was done about each.

PMID	Journal	Claim	Disposition
41477823	Science Advances	Autologous iPSC alveolus-chip; macrophage-restricted ATG14 knockout raises necrosis without bacterial replication	Curated → `Tuberculosis`
40987954	Nature Biomedical Engineering	IL-1β and TNF-α act in opposite directions on the influenza cytokine storm; fibroblast CXCL12–CXCR4 axis	Curated → `Influenza`
41252215	J Clinical Investigation	Hyperphysiological strain worsens Pseudomonas infection and bacterial translocation	Curated → `Acute_Respiratory_Distress_Syndrome`
41442163	Toxicological Sciences	Chlorine gas: epithelial junction loss immediate, endothelial loss delayed to 72 h	Curated → `Acute_Respiratory_Distress_Syndrome`
42473541	Materials Today Bio	Whole cigarette smoke destroys alveolar microvasculature only when alveolar epithelium is present	Curated → `Chronic_Obstructive_Pulmonary_Disease`
42475431	J Visualized Experiments	COPD patient-derived chip shows mucus hypersecretion with barrier intact	Curated → `Chronic_Obstructive_Pulmonary_Disease`
41406599	Biofabrication	Cyclic stretch amplifies TGF-β1 fibrotic signalling; reversed by nintedanib	Curated → `Idiopathic_Pulmonary_Fibrosis`
41786071	Eur J Pharmacology	Hypoxia degrades blood-gas barrier proteins via HIF-1α/HO-1; canine + chip	Curated → new `High_Altitude_Pulmonary_Edema`
42546767	Biofabrication	Diesel particulate injures the unexposed endothelial layer — "trans-barrier propagation"	Curated → new module (see below)
42083145	Transplantation	Cold-storage IRI alveolus-chip: barrier loss plus compartment-specific adhesion-molecule shedding	Curated → new module
42204607	Respiratory Research	Isothiazolinone humidifier disinfectant disrupts the alveolar barrier	Curated → new module

The cross-cutting finding

Five of these papers, using five unrelated insults — diesel particulate, chlorine gas, cigarette smoke, isothiazolinone, and cold-storage ischemia — independently report the same thing: an exposure confined to one side of the blood-gas barrier produces measurable injury in the other, unexposed side. The chlorine paper times it (epithelial injury immediate, endothelial at 72 h). The cigarette-smoke paper controls it (vascular networks exposed to smoke without overlying epithelium stay intact).

This is a mechanism claim, not a device observation, and it is the kind of thing only a compartmentalised model can establish — you cannot expose one side of the alveolar barrier and not the other in a patient or an animal. It recurs across enough diseases to meet the module bar, so it was curated as one:

kb/modules/alveolar_capillary_barrier_failure.yaml — five nodes, trigger → amplifier → central_effector → effector → consequence, with Alveolar Epithelial and Endothelial Junctional Disruption as the key conformance target. Acute_Respiratory_Distress_Syndrome (4 nodes) and High_Altitude_Pulmonary_Edema (3 nodes) conform to it.

Not curated, and why

Reviews (22 of 49). Useful for orientation, not for evidence snippets.
Pure device and materials papers — PLGA/PCL membrane chemistry (42546767's primary contribution), LCD 3D printing (42480620), inkjet pH sensors (41590290), extracellular-vesicle isolation chips (41500804). Where these carried a disease finding it was extracted; the fabrication content has no dismech home.
PMID:42481340 — retracted.
PMID:41982094 (mechanical stretch reduces SARS-CoV-2 pseudovirus membrane fusion) — a real finding, but pseudovirus fusion on stretched AT2 cells does not map cleanly onto any existing node in Long_COVID or a COVID entry. Left for a curator who knows that entry.
PMID:42138232 (benzene VOC cumulative cytokine response, reversed by montelukast) — an air-quality risk-assessment result rather than a disease mechanism; would need an entry for benzene inhalation exposure that does not yet exist.

Curation output

Path	Change
`kb/modules/alveolar_capillary_barrier_failure.yaml`	new module, 13 evidence snippets
`kb/disorders/High_Altitude_Pulmonary_Edema.yaml`	new disorder (MONDO:0031257), 33 snippets
`kb/disorders/Tuberculosis.yaml`	iPSC alveolus-chip, 5 mechanism links incl. one `FAILS_TO_RECAPITULATE`
`kb/disorders/Influenza.yaml`	immune-competent lung-chip, 3 mechanism links
`kb/disorders/Chronic_Obstructive_Pulmonary_Disease.yaml`	2 chips, 4 mechanism links
`kb/disorders/Idiopathic_Pulmonary_Fibrosis.yaml`	alveolar array chip, 2 mechanism links
`kb/disorders/Acute_Respiratory_Distress_Syndrome.yaml`	2 chips, 4 mechanism links; 4 `conforms_to` edges to the new module

Notes for future scans

The lung-chip literature is small enough (≈50 papers/year) to read exhaustively. The general organ-on-chip literature (≈700 in six months) is not, and needs the Review[pt] filter plus a disease-term intersection to be tractable.
Negative results are unusually well reported in this field, because a chip that fails to reproduce something is publishable as a limitation. FAILS_TO_RECAPITULATE and HUMAN_MODEL_MISMATCH are the right slots and were under-used before this scan; the tuberculosis non-permissiveness result is a good worked example.
Several groups now publish patient-derived chips (COPD in 42475431 and 42473541). These are worth watching: they carry genotype into the model, which is what makes a chip usable for the subtype-level claims dismech records.