Skip to content

July 2025 — Monthly Briefing

1. What was read

1,049 research articles from 32 of the 33 journals on the roster.

A further 282 records (21% of the 1,331 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
553 195 99 456 28

Volume swing against the median month: bioinfo +68%

Check the source before reading anything into this. Two benign causes account for most of these. bioinfo every January carries the Nucleic Acids Research database issue and runs high. And several journals simply deposit nothing in some months: across the 44 months read, 4 of 33 journals are absent from at least one, worst being Nat Mach Intell (20 months missing, 81 papers in total) and Bioinformatics (4 months missing, 2,905 papers in total). This month is missing Nat Mach Intell. Queried directly, PubMed returns zero for those journal-months too — the holes are in the source, reproduced exactly by re-running the harvest. So a swing here is not evidence that a field went quiet, and is not usually evidence that the pipeline broke either.

Largest contributors: Nucleic Acids Res 204, Nature 132, JAMA 128, Bioinformatics 123, Science 110, Genome Biol 45.

2. Net change on the thread curve

Thread This month Cumulative Status
agentic-ai +1 +12 / −1 forming
cell-therapy +2 / −4 +187 / −58 forming
ctdna-mrd +2 +65 / −22 forming
foundation-models +5 / −2 +102 / −72 splitting
immunotherapy-resistance +6 / −1 +207 / −74 forming
interception +1 +117 / −22 forming
method-credibility +2 / −17 +18 / −415 crisis
neoantigen-vaccines +1 / −3 +76 / −33 forming
neuro-immune-microbiome-axis +5 / −2 +157 / −17 forming
next-gen-checkpoints +2 / −3 +66 / −38 splitting
single-cell-3d-genome +2 +65 / −17 forming
suppressive-population-inversion +2 / −5 +49 / −75 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

Zero of June's five nominated checkpoint targets reappear. Not one of Siglec-G/10, CMTM4, STS2, Piezo1 or TTLL12 is named anywhere in the month. The observed half-life of a novel surface checkpoint name in this corpus is under a month for the genuinely new ones; older validated names resurface to fill the slot. What July has instead is LILRB4, galectin-9, PILRalpha-CD99, TIM-3, CD30, HPK1, ICOS and B7-H3 — a fresh list with no members in common with the last one.

Both CD47 phase 3 readouts are in this month (ENHANCE-2 and ENHANCE-3, Blood, 31 July). Neither title states a direction, which is why neither is carded here: the corpus knows these programmes ended, but it does not know it from these rows, and a card that recorded the outcome would be recording something the reader inferred. The direction is settled in the 2026-07 briefing, from a JCO phase 3 report that states it.

A ceiling, stated as a number. 40662806: deep-learning sequence-based protein-interaction predictors plateau at 0.65 accuracy under unbiased evaluation. With June's two — simple controls exceeding the best deep learning algorithms for perturbation prediction, and conservation plus solvent accessibility carrying almost all of variant effect — that is three such results in eight weeks, a full year before the corpus recorded the pattern as emerging.

Corrected 2026-08-27

A fourth was counted here: 40605002, read as "plain Z-scores outperform purpose-built methods for CRISPR paralog synthetic-lethality screens". That is the title, not the result. Across four in4mer Cas12a screens, Z-transformed and rescaled dLFC identify synthetic lethal pairs comparably — both more consistently than the unscaled dLFC in use — and the Z-score's only stated edge is that it needs no training set of known interactors. Convenience, not superiority.

The card stays on method-credibility, because the finding that analysis pipelines for these screens vary considerably is intact. It is not a baseline-beats-purpose-built result, and it should not have been counted toward a run of them.

4. Refutations

(a) Landing on a tracked thread

40707132 (JITC) is the cleanest surrogate-versus-endpoint inversion in the corpus, and it is stated in one title: constitutive IL-7 signalling promotes CAR-NK survival in the solid tumour microenvironment but impairs tumour control. Persistence is the field's standard proxy for efficacy in engineered cell therapy.

40668889: perturbing S100A8/A9 — a canonical MDSC axis and an active drug target — blunts antitumour immunity by promoting protumorigenic myelopoiesis. 40628269: the costimulatory molecule ICOS limits memory-like properties of exhausted PD-1+ CD8 T cells. ICOS agonism was in first-in-human trials at the time. 40624238: IL-17A-secreting gamma-delta T cells promote resistance to CDK4/6 inhibitors — immune-mediated resistance to a drug class that is not an immunotherapy.

40645662: shared epitopes create safety and efficacy concerns across several cancer vaccines. Shared-antigen vaccines are the off-the-shelf alternative to personalization, and this says the shortcut has its own failure mode.

40193715 (Blood): post-immunotherapy lineage switch is an emergent form of relapse in leukaemia — the tumour changes lineage rather than losing the antigen, which no antigen-escape model predicts.

(b) Method credibility

40794862 (NAR): errors are prevalent in lab-made plasmids worldwide. The most widely shared physical reagent in molecular biology, audited at scale.

40682824: untargeted CUT&Tag reads enrich at accessible chromatin and correlate genome-wide with the G4-targeted signal, so calling peaks against a matched untargeted library slightly improves precision — at the cost of excluding real G4-forming sequences. The recommendation is that published G4 maps be read with that control in mind and their interpretation handled with care. An earlier version of this entry said the maps "may be accessibility maps"; the paper reports a colocalisation and a modest precision-for-recall trade-off, and does not call them artefacts. The dispute continues into November with a commentary and a rebuttal, which is what a live disagreement looks like and is worth more than a quiet correction.

40560601: structural gene annotation choice materially changes orthology inference. Filed with the reference-genome cards from October and November: the substrate, not the algorithm, is what decides the answer.

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 613 (of which 549 genuinely oncology)
ema-authorisation 11
fda-approval 2

2 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:

  • 2025-07-02 — FDA grants accelerated approval to sunvozertinib for metastatic non-small cell lung cancer with EGFR exon 20 insertion mutations
  • 2025-07-02 — FDA grants accelerated approval to linvoseltamab-gcpt for relapsed or refractory multiple myeloma

5 EMA oncology authorisations:

  • 2025-07-17 — Bomyntra
  • 2025-07-17 — Ezmekly
  • 2025-07-17 — Aucatzyl
  • 2025-07-18 — Itovebi
  • 2025-07-23 — Blenrep

8. Trade press

38 on-topic of 327 items across 3 feeds (n-stat 312, n-aacr-blog 10, n-acir 5). 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.

  • 2025-07-31 — How respiratory diseases could influence cancer — n-stat
  • 2025-07-30 — STAT+: Viral infections in the lung, like Covid-19, can reactivate dormant cancer cells — n-stat
  • 2025-07-30 — New nanobody bispecific antibody–drug conjugates induced enhanced ICB activity — n-acir
  • 2025-07-29 — From the Bench, Summer 2025: In Vivo Treatment Spies to Counter Cancer’s Cunning — n-aacr-blog
  • 2025-07-29 — STAT+: Cellular ‘trash bins’ might be a key to reviving multi-cancer early detection tests — n-stat
  • 2025-07-29 — Opinion: STAT+: U.S. must develop plan to protect clinical trials as national security asset — n-stat
  • 2025-07-25 — Examining the Rise of Early-onset Incidence Across Cancer Types — n-aacr-blog
  • 2025-07-25 — STAT+: Europe moves to reject embattled Duchenne muscular dystrophy gene therapy — n-stat
  • 2025-07-24 — STAT+: For many Duchenne families, halt to gene therapy is heartbreak upon heartbreak — n-stat
  • 2025-07-24 — A mother’s perspective on Sarepta’s gene therapy halt — n-stat
  • 2025-07-24 — STAT+: Why Sarepta may not survive its current gene therapy crisis — n-stat
  • 2025-07-23 — Editors’ Picks, July 2025: U.S. Cancer Mortality Hotspots, Chemotherapy-induced Hearing Loss, and More — n-aacr-blog
  • … and 26 more on-topic items.