April 2025 — Monthly Briefing¶
1. What was read¶
736 research articles from 31 of the 33 journals on the roster.
A further 218 records (23% of the 954 this month's harvest returned) carry no abstract and are not research: news, correspondence, editor highlights and obituaries. PubMed types them Journal Article and nothing else, so the publication-type filter cannot see them — across the whole index they are 22.7% of it and 35% of general, and 3,801 of them are JAMA. Every count below is over the full harvest and is inflated by this; the figure above is not.
| General | Cancer | Immune | Bioinfo | Sysbio |
|---|---|---|---|---|
| 445 | 186 | 84 | 215 | 24 |
Largest contributors: Nature 105, Science 91, Nucleic Acids Res 83, JAMA 82, J Immunother Cancer 50, Blood 39.
2. Net change on the thread curve¶
| Thread | This month | Cumulative | Status |
|---|---|---|---|
agentic-ai |
— | +9 / −1 | forming |
cell-therapy |
+9 / −1 | +169 / −48 | forming |
ctdna-mrd |
+5 | +53 / −22 | forming |
foundation-models |
+5 / −3 | +92 / −62 | splitting |
immunotherapy-resistance |
+7 / −1 | +180 / −65 | forming |
interception |
+8 | +100 / −21 | forming |
method-credibility |
+1 / −14 | +13 / −374 | crisis |
neoantigen-vaccines |
+3 | +65 / −28 | forming |
neuro-immune-microbiome-axis |
+11 | +140 / −14 | forming |
next-gen-checkpoints |
+3 | +57 / −35 | splitting |
single-cell-3d-genome |
+2 | +56 / −17 | forming |
suppressive-population-inversion |
+3 / −4 | +41 / −65 | contested |
Status is derived by rule, not asserted. Fewer than 5 cards → too-early; then refutes ≥ 2×supports → crisis; refutes > supports → contested; supports ≥ 2×refutes → forming; else splitting.
3. The evidence that moved it¶
Nature Medicine declares a benchmarking crisis in biomedical machine learning (40200055). This is the single most consequential card in the first half of 2025, and it is the origin of a claim the corpus later had to correct: April was recorded as the high-water mark of negative clinical-AI evidence on the strength of "null randomized trials plus a benchmarking-crisis editorial". The editorial is real. The null randomized trials never existed — re-reading every month of 2025 under one rule found zero titles stating a null for any AI intervention.
And in the same month, the audit reaches the models themselves. 40251685 (Genome Biol): zero-shot evaluation reveals limitations of single-cell foundation models. Read with the benchmarking-crisis editorial, April 2025 is when the methods literature began auditing the thing it had spent the previous year building — eight months before this corpus recorded that as having happened.
"Virtual cell" is already a Nature Biotechnology title
(40229360, virtual cells for
predictive immunotherapy) — a comment proposing virtual-cell models for
predicting immunotherapy response, in print two months before the Cell
challenge usually credited with naming the concept. An earlier version of this
sentence added "and seven months before the phrase reaches the methods journals",
which encoded an arrival date data/threads.yml has since retracted; no interval
replaces it. Read the April title as a bound on the months that were read, not as
a date for the vocabulary. This corpus has made ten claims of that shape and has
retracted ten.
An intracellular checkpoint reaches first-in-human. 40315882 (Lancet Oncol): CRISPR-Cas9 CISH-edited T cells in metastatic colorectal cancer. Note the shape — the target is not blocked with an antibody, it is deleted in the cell being infused. That manoeuvre becomes the class's main survival route by 2026.
4. Refutations¶
(a) Landing on a tracked thread¶
40249828 (Sci Immunol): CCR8-positive decidual regulatory T cells maintain maternal-fetal immune tolerance in early pregnancy. CCR8 is the marker chosen for effector-Treg depletion precisely because it is supposed to be tumour-restricted. This is the most direct card in the corpus against that premise, and the question it opens — whether the chosen surface marker separates the subset that should go from the one that should stay — is the one every depletion programme downstream of it has to answer.
40139188: metallophilic marginal zone macrophages cross-prime protective CD8 immunity against blood-borne tumours, matching cDC1s at cross-priming and depending on Batf3 rather than on handover from cDC1s. A macrophage population whose depletion would remove antitumour immunity rather than release it.
Against it, 40280128 is the thread's clearest support: Treg depletion promotes myeloid activation and glioblastoma response to anti-PD-1. Both directions are real, which is the point of carding stance rather than topic.
Corrected 2026-08-27
Three further cards were filed here as inversions and none of them is one.
40202509 was listed as "neutrophil MYO1F is required for the response to checkpoint blockade", read as a suppressive population being necessary for immunotherapy. The abstract runs the other way: tumour-associated neutrophils are a significant barrier, MYO1F is what restrains their suppression, and tumour-derived TGF-beta1 silencing it is what makes them suppressive. That is the canonical account, and the card now supports the thread.
40306958 was listed as tumour-infiltrating mast cells acting as a chemoimmunotherapy enhancer, which reads as a primary finding. It is a JITC commentary on Zhang et al. in Cancer Cell, a paper this corpus does not hold. The inversion it reports may well be real; this record is not evidence for it, and carding the echo rather than the study is how this corpus opened two threads on the wrong paper.
39786370 was filed as the adrenergic axis "cutting the other way in a clinical setting", against an intervention built on it. No such intervention was tested. Carvedilol and metoprolol are cardiac drugs, and the finding is that blocking beta2/beta3 — but not beta1 — signalling delays platelet engraftment and reduces survival after allogeneic transplant, in two patient cohorts and in mice. Neural control of a haematopoietic-immune process, demonstrated in humans, is the thread's mechanism claim rather than its contradiction. The setting is transplant engraftment, not a tumour microenvironment, and the card now supports.
(b) Method credibility¶
40205439: the standard high-mitochondrial-content quality filter depletes viable, metabolically altered malignant cells from cancer single-cell studies. A default QC step that systematically removes the cells of interest.
40197327: missing cell types in single-cell references distort bulk deconvolution. 40240607 (Nature): there is low overlap between transcription factor DNA binding and regulatory targets — which undercuts binding-derived regulatory networks generally.
Corrected 2026-08-27
40082610 was listed here as a third method-credibility refutation — "feature selection method choice changes integration and querying performance" — on the strength of a title that names its object rather than its verdict. The verdict runs the other way: benchmarking feature selection across integration, query mapping, label transfer and unseen-population detection, the paper states that it "reinforce[s] common practice by showing that highly variable feature selection is effective for producing high-quality integrations", and its output is guidance on how many features to take and when to take them batch-aware. This is the audit genre with the sign running the unexpected way: an audit that clears the method. It belongs on the supporting side.
5. Open items¶
Nothing carried forward.
6. Meetings¶
No abstract book published this month. The abstract layer runs 2023-02 to 2026-08 and is seasonal by nature — meetings are annual, so most months are empty here and that is the expected shape, not a gap in the harvest.
7. Regulatory¶
| Event | Count |
|---|---|
| trial-first-posted | 555 (of which 496 genuinely oncology) |
| ema-authorisation | 5 |
| fda-approval | 3 |
3 FDA actions — dated from each notice's own prose, not the table's posting date, which disagrees on 7 of 190 rows and once by enough to move an approval into the wrong month:
- 2025-04-08 — FDA approves nivolumab with ipilimumab for unresectable or metastatic MSI-H or dMMR colorectal cancer
- 2025-04-11 — FDA approves nivolumab with ipilimumab for unresectable or metastatic hepatocellular carcinoma
- 2025-04-23 — FDA approves penpulimab-kcqx for non-keratinizing nasopharyngeal carcinoma
3 EMA oncology authorisations:
- 2025-04-04 — Datroway
- 2025-04-23 — Lynozyfic
- 2025-04-25 — Trabectedin Accord
8. Trade press¶
26 on-topic of 364 items across 3 feeds (n-stat 348, n-aacr-blog 12, n-acir 4). Headlines are selected by subject keyword, not recency — most of what these feeds carry is health policy and industry finance the corpus does not read.
- 2025-04-30 — Pembrolizumab’s Perioperative Power in Head and Neck Cancer —
n-aacr-blog - 2025-04-30 — STAT+: Will the FDA approve the first new drug to help people quit smoking in 20 years? —
n-stat - 2025-04-30 — Treg depletion in brain tumors enhances antitumor responses —
n-acir - 2025-04-28 — Eye on AI: Improving Cancer Care With Artificial Intelligence —
n-aacr-blog - 2025-04-28 — STAT+: HPV vaccine news ‘that will save millions of lives’ from cancer, and other AACR news —
n-stat - 2025-04-28 — The case against cancer surveillance — for some —
n-stat - 2025-04-25 — STAT+: Akeso wins Chinese approval for cancer drug positioned to rival Merck’s Keytruda —
n-stat - 2025-04-24 — STAT+: Caribou delays readout, discontinues programs, betting on off-the-shelf CAR-T therapies —
n-stat - 2025-04-23 — STAT+: CRISPR’s CEO got a big payday —
n-stat - 2025-04-22 — New Treatment Strategies for Esophageal Cancer —
n-aacr-blog - 2025-04-22 — Advanced cancers returned to pre-pandemic levels, according to a reassuring report —
n-stat - 2025-04-21 — STAT+: FDA approves blockbuster Dupixent for chronic hives —
n-stat - … and 14 more on-topic items.