September 2025 — Monthly Briefing¶
1. What was read¶
844 research articles from 33 of the 33 journals on the roster.
A further 248 records (23% of the 1,092 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 |
|---|---|---|---|---|
| 468 | 186 | 92 | 318 | 28 |
Largest contributors: Nucleic Acids Res 118, JAMA 108, Nature 100, Science 86, Genome Biol 60, Bioinformatics 53.
2. Net change on the thread curve¶
| Thread | This month | Cumulative | Status |
|---|---|---|---|
agentic-ai |
−1 | +14 / −2 | forming |
cell-therapy |
+5 / −3 | +194 / −63 | forming |
ctdna-mrd |
+2 | +71 / −23 | forming |
foundation-models |
+7 / −1 | +115 / −77 | splitting |
immunotherapy-resistance |
+4 / −2 | +214 / −77 | forming |
interception |
+4 | +122 / −23 | forming |
method-credibility |
−15 | +18 / −444 | crisis |
neoantigen-vaccines |
+1 / −2 | +77 / −38 | forming |
neuro-immune-microbiome-axis |
+7 / −2 | +170 / −19 | forming |
next-gen-checkpoints |
+2 / −1 | +68 / −40 | splitting |
single-cell-3d-genome |
+1 | +66 / −17 | forming |
suppressive-population-inversion |
+1 / −4 | +51 / −81 | 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¶
The month's strongest result is a claim of harm, not a null. JCO published TrilynX, the placebo-controlled phase 3 of the IAP antagonist xevinapant (40902136), alongside a commentary whose title is the finding: SMAC mimetics accelerate progression of locally advanced head and neck cancer. A decade of preclinical apoptosis-sensitisation enthusiasm produced a randomized trial in which the intervention arm did worse. Failed is common; worse is rare, and worse is what makes a class dead rather than mis-targeted.
Cancer Discovery published a coordinated argument against the cell type it was
simultaneously publishing engineering papers about. Three titles in one issue —
The Dark Side of NK Cells,
cytotoxic NK cells impede checkpoint response,
and tumour-associated NK cells contribute to ICB resistance —
against a rewiring review, a multiplex-edited CAR-NK and a loop-CAR iNK in the
same month. The engineering is preclinical and phase 1; the negative biology is
from human checkpoint cohorts. This is the month that made
suppressive-population-inversion a thread rather than a recurring oddity.
Tertiary lymphoid structures contradicted themselves inside one journal. 40992783 links TLS to immunotherapy response in Merkel cell carcinoma; 40998518 finds high Treg and PMN-MDSC density in TLS to be a hallmark of fatal cervical cancer. Same month, same journal. TLS presence alone cannot be a predictive biomarker if both hold, and the composition — not the structure — is what carries the signal.
Checkpoints: the field ran out of nouns. Zero of May's five nominated targets and zero of June's five appear. September nominates no replacement roster at all. The only checkpoint content with momentum is LAG-3, and it is follow-up analysis of an agent that already cleared phase 3 (40638872). No newly nominated checkpoint entered a phase 1 anywhere.
4. Refutations¶
(a) Landing on a tracked thread¶
The NK block above, plus 40954074: MAIT cells, generally proposed as innate-like antitumour effectors, promote progression and Treg accumulation in bladder tumours. And 40929245: across 171 patients with early triple-negative breast cancer, an interferon-induced senescent CD8 subset — driven by IFN from HLA-DR+ monocytes, and marked by excessive NAD+ consumption and reduced cytotoxicity — predicts anti-PD-1 non-response. That is one subset, not the collapse of CD8 infiltration and the interferon signature as classes of predictor, which is what an earlier version of this sentence claimed. The card is also not purely negative: nicotinamide mononucleotide restored the cells' function and improved immunotherapy efficacy in patient-derived organoid–T cell co-culture and in mice. A resistance state with a named rescue is a different object from a failed biomarker.
40983343 (JITC) attacks ctDNA as a surrogate directly: the association between early ctDNA change and overall survival depends on which molecular-response cutoff and which collection timepoints are chosen. Every ctDNA-surrogate paper in this corpus now has to answer it.
Corrected 2026-08-27
40828134 (JEM, "Antibiotics unleash neuroinflammation") was filed here as a refutation on the neuro-immune thread rather than a support, on the argument that it removes the assumption that microbiome depletion is an inert experimental manipulation — which matters, because a large fraction of this thread's mouse work uses antibiotic depletion as its control arm.
The argument may be sound. This record is not what establishes it. The whole
abstract is one sentence pointing at
40627379, the gut
microbiota-derived TMAVA paper carded supports on the same thread in this
same month. One study, two cards, filed in opposite directions on one
thread — the ratio moved twice on a single finding, and the two entries
argued with each other. The commentary is now neutral and the result stands
where 40627379 put it.
(b) Method credibility¶
Fifteen cards, which is a heavy but unremarkable month on this thread — the largest single-month block in the corpus is 2024-01, with thirty. Two of the fifteen concern reagents rather than analyses, which is the more damaging kind: stable Cas9 expression alters cell growth independently of editing and AsCas12a tolerates insertions in target DNA.
Corrected 2026-09-04
This read "Eight cards, the largest single-month block in the corpus." The
count was eight because September 2025 had been carded from titles only; the
2026-09-04 abstract re-scan puts it at fifteen, which is what section 2 now
shows. The superlative was wrong on the old card set and is wrong on the new
one — method-credibility runs to 610 refutations across 44 months, and its
largest month is 2024-01 at thirty. A superlative computed over a corpus
that is still being read is a claim about the reading rather than about the
literature, and this one was never recomputed when the reading changed.
A third audits the analysis layer: 40923760 finds that existing tandem-repeat finders return markedly different results and that it is unclear which of them are correct, and offers a detector that recovers highly divergent repeats with relatively few false positives. The ubiquitous sequence reuse in its title is the biological discovery it makes with that detector; it is not the charge it lays against the published tools, and an earlier version of this entry credited them with having missed it. What the paper says about the tools is narrower and harder to answer: they disagree, and no one can say which is right.
40963104 states that scATAC-seq analysis discards finer-resolution information — in a modality most downstream methods treat as solved.
40715786 names over-reliance on AI code assistants as a threat to scientific software. It is the loudest agentic-AI paper of the month, and it is a warning.
Two further retractions of record: the CB1 agonist-bound crystal structures and live-imaging claims about kinetic genome organization.
5. Open items¶
Nothing carried forward.
6. Meetings¶
1 meeting with a written vault this month. These are read and summarised rather than harvested as abstracts, so they do not appear in the counts below.
No abstract book was deposited this month either — the vault above is built from transcripts and programmes rather than from a published abstract set. 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 | 595 (of which 536 genuinely oncology) |
| ema-authorisation | 11 |
| fda-approval | 4 |
4 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:
- 2025-09-09 — FDA approves gemcitabine intravesical system for non-muscle invasive bladder cancer
- 2025-09-10 — FDA approves selumetinib for pediatric patients 1 year of age and older with neurofibromatosis type 1 with symptomatic, inoperable plexiform neurofibromas
- 2025-09-19 — FDA approves pembrolizumab and berahyaluronidase alfa-pmph for subcutaneous injection
- 2025-09-25 — FDA approves imlunestrant for ER-positive, HER2-negative, ESR1-mutated advanced or metastatic breast cancer
3 EMA oncology authorisations:
- 2025-09-17 — Bilprevda
- 2025-09-17 — Voranigo
- 2025-09-17 — Romvimza
8. Trade press¶
33 on-topic of 346 items across 3 feeds (n-stat 333, n-aacr-blog 11, n-acir 2). 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.
- 2025-09-30 — Editors’ Picks, September 2025: Cervical Cancer Screening, Immunotherapy Breakthroughs, Metabolic Vulnerabilities, and More —
n-aacr-blog - 2025-09-25 — The Enduring Importance of Cancer Disparities Research —
n-aacr-blog - 2025-09-25 — STAT+: Crinetics wins FDA approval for acromegaly drug, faces competition with Novartis and Ipsen —
n-stat - 2025-09-25 — A gene therapy success in Huntington’s and Trump’s autism announcement —
n-stat - 2025-09-25 — STAT+: New breast cancer screening trial may answer key AI question —
n-stat - 2025-09-24 — STAT+: Anthony Letai of Dana-Farber is front-runner to lead National Cancer Institute —
n-stat - 2025-09-24 — STAT+: Pharmalitte: We’re reading about a Huntington’s gene therapy success, a low-cost HIV drug, and more —
n-stat - 2025-09-24 — STAT+: Experimental gene therapy for Huntington’s markedly slowed disease progression in key trial —
n-stat - 2025-09-24 — STAT+: Large breast cancer screening trial will be pivotal step in vetting AI for mammograms —
n-stat - 2025-09-24 — Opinion: My patient almost quit a clinical trial to save her job —
n-stat - 2025-09-23 — Looking Beyond Overall Survival: FDA-AACR Workshop Explored Novel Oncology Endpoints —
n-aacr-blog - 2025-09-22 — AACR Grantee Spotlight: Adwoa Mensah’s Mission to Improve Breast Cancer Screening in Ghana —
n-aacr-blog - … and 21 more on-topic items.