# Evidence digest — 2026-09-11 (week 2026-09-04..09-11) 18 new endometriosis papers indexed this week in PubMed. The week is clinically thinner than the last two (no new trial readouts in the top journals) but it is unusually strong on two things OpenEndo tracks: **pain biology** (why lesions hurt in some patients and not others) and **Danish genetics** — the 25-year synthesis of endometriosis GWAS from the QIMR / Aalborg / Statens Serum Institut group lands in *Human Reproduction Update*. Mechanism papers cluster again around ferroptosis, epigenetics and progesterone resistance. Prior digest: 2026-09-04. Every entry below comes from a fetched PubMed record; PMIDs are live. ## Diagnostics - **#Enzian by transvaginal ultrasound in 394 consecutive patients** — deep endometriosis in 33.5%, ovarian endometriosis in 35.8%; the uterosacral ligament / pelvic sidewall (B) compartment was the most frequent deep site (52.9%) and rectal involvement (C) followed at 31.0%. Laterality is compartment-specific (tubo-ovarian adhesions bilateral in 63.1%). Retrospective single-operator cohort, specialised centre. [PMID 42717817](https://pubmed.ncbi.nlm.nih.gov/42717817/) — *why it matters:* gives real-world distribution baselines for the IDEA/#Enzian ultrasound pathway, the non-surgical route to lesion mapping that OpenEndo's access content leans on. - **Serum ANGPTL4 as a candidate blood marker** — serum (and peritoneal fluid) ANGPTL4 was higher in surgically confirmed endometriosis than in controls (512.7 ± 46.9 vs 177.6 ± 11.8 ng/mL, p<0.001); a 188.61 ng/mL cut-off gave AUC 0.755, sensitivity 75%, specificity 62.5%. Prospective case–control study, single centre, 2× groups. [PMID 42694335](https://pubmed.ncbi.nlm.nih.gov/42694335/) — *why it matters:* specificity is too low for a stand-alone diagnostic, but it is another entry for the non-invasive-marker watchlist (alongside CA125, which also separated the groups here). - **Intraoperative ultrasound to localise deep rectosigmoid nodules** — technique note from Mayo Clinic Arizona on using intraoperative US for nodule localisation during surgery. Technical report, no outcome data. [PMID 42702235](https://pubmed.ncbi.nlm.nih.gov/42702235/) — *why it matters:* surgical-precision item only; no evidence change. ## Treatment and drugs - **Relugolix combination therapy — 104-week SPIRIT extension** — post hoc analysis of SPIRIT 1/2 and their long-term extension (802/1261 entered the LTE, 501 completed 104 weeks) tracking cumulative probability and median time to minimal-to-no dysmenorrhea and non-menstrual pelvic pain, amenorrhea, and analgesic-free status. Industry-sponsored post hoc analysis of RCT data. [PMID 42719373](https://pubmed.ncbi.nlm.nih.gov/42719373/) — *why it matters:* the longest follow-up yet on an oral GnRH-antagonist add-back regimen; the timeline-to-benefit numbers are the practical part for shared decision-making. - **M1 macrophage nanovesicles delivering bazedoxifene to lesions** — CCL2–CCR2-targeted M1 membrane nanovesicles carrying the estrogen- receptor antagonist bazedoxifene reduced inflammatory infiltration, angiogenesis and fibrosis in a preclinical model without ovarian or uterine toxicity. Preclinical (cell/animal) formulation study. [PMID 42702982](https://pubmed.ncbi.nlm.nih.gov/42702982/) — *why it matters:* a lesion-targeted route around systemic hormonal suppression; early, but it speaks to the same local-delivery logic OpenEndo's repurposing track considers. - **FIN56-induced ferroptosis via ACACA/ARID5A/NOX4** — ferroptosis inducer FIN56 suppressed ectopic lesion growth in vitro/in vivo by raising mitochondrial ROS and lipid peroxidation through an ACC1(ACACA)–ARID5A–NOX4 axis. Preclinical mechanistic study (ectopic endometrial stromal cells + model). [PMID 42698143](https://pubmed.ncbi.nlm.nih.gov/42698143/) — *why it matters:* third independent confirmation this year that ferroptosis induction is a real therapeutic direction in endometriosis — consistent with `ferroptosis-direction.md` and with the SLC7A11 / xCT selectivity caveat that keeps sulfasalazine on WATCHLIST. - **HDAC inhibitors — review of the preclinical case** — abnormal HDAC expression drives proliferation, apoptosis evasion, inflammation and EMT; vorinostat, romidepsin and butyrate have preclinical activity, with clinical translation blocked by selectivity and delivery. Narrative review. [PMID 42717121](https://pubmed.ncbi.nlm.nih.gov/42717121/) — *why it matters:* maps an epigenetic-drug class for the repurposing watchlist; no new human data. - **ZEB1 as an EMT / progesterone-resistance hub** — review positioning ZEB1 as a transcriptional integrator of estrogen induction, PR-pathway suppression, miR-200/lncRNA/circRNA feedback and invasive phenotype in ectopic lesions. Narrative review. [PMID 42717122](https://pubmed.ncbi.nlm.nih.gov/42717122/) — *why it matters:* an upstream node for the progesterone-resistance thread (the FKBP52/HOXA10/miR-29c axis OpenEndo already tracks); explicitly flags that targeting ZEB1 therapeutically is paradoxical. - **Non-invasive brain stimulation for chronic pelvic pain** — systematic review of tDCS and rTMS in chronic pelvic pain: 15 studies, 295 participants (264 completers), substantial heterogeneity, most RCTs at high or some risk of bias; narrative synthesis only. [PMID 42722127](https://pubmed.ncbi.nlm.nih.gov/42722127/) — *why it matters:* the central-sensitisation treatment angle patients ask about; current evidence does not support routine use. - **Laparoscopic discoid excision with hand-sewn bowel closure** — 10 patients with 3–5 cm deep bowel nodules unsuitable for stapled discoid excision; manual two-layer suturing took a mean 22 min, one severe febrile morbidity, 18–60 months follow-up, no severe complications. Retrospective cohort (prospectively collected data), n=10. [PMID 42722072](https://pubmed.ncbi.nlm.nih.gov/42722072/) — *why it matters:* a bowel-sparing alternative in the surgical debate over segmental resection vs shaving; very small numbers, surgeon- dependent. - **Robot-assisted buccal mucosa graft ureteroplasty** — technique note for ureteric repair in deep infiltrating endometriosis. Case/technical report. [PMID 42705528](https://pubmed.ncbi.nlm.nih.gov/42705528/) — *why it matters:* rare-complication surgical capability, relevant to the urinary-tract deep-endometriosis pathway. ## Mechanisms - **Why some lesions hurt and others do not — inflammatory gene signature** — RNA-seq of 27 peritoneal biopsies from 9 symptomatic and 10 asymptomatic patients: 890 differentially expressed genes, including inflammatory signalling (IL16, IL17RA, JAK3, SMPD3, RELT), adhesion (OLFML1, CDON, VCAN) and neuromodulation (SEMA6D, ADRA2C, SLC7A5); IL16 also tracked severity in mild vs severe pain. Human tissue transcriptomics (Yale). [PMID 42720599](https://pubmed.ncbi.nlm.nih.gov/42720599/) — *why it matters:* direct molecular support for the long-standing clinical observation that stage does not predict pain — and the first concrete candidate (IL16) for a symptom-severity biomarker. - **Fibrotic substrate stiffness drives endometriotic epithelial motility** — on tunable polyacrylamide hydrogels, intermediate stiffness maximised cell speed, actin stress fibres, focal adhesions and spheroid expansion (biphasic dose–response). In-vitro mechanobiology. [PMID 42720355](https://pubmed.ncbi.nlm.nih.gov/42720355/) — *why it matters:* frames fibrosis as an active biomechanical driver of lesion spread, not just scar tissue. - **Gut microbiota-associated targets by integrative multi-omics** — 61 metabolite-linked DEGs; MET (protective direction) and SELL (modest risk direction) as hub genes, with *Bacteroides*, *Parabacteroides*, *Bifidobacterium* and metabolites palmitoylethanolamide / dihydroresveratrol implicated in immune-inflammatory pathways; molecular-simulation binding checks. Computational multi-omics + single-cell reanalysis. [PMID 42713565](https://pubmed.ncbi.nlm.nih.gov/42713565/) — *why it matters:* adds a named gut–immune axis to the microbiome thread; hypothesis-generating, not clinical. - **β-Glucuronidase and the estrobolome (broader review)** — review of bacterial β-glucuronidase reactivating estrogens through enterohepatic recycling, and the >10⁴-fold selectivity possible against bacterial GH2 over human GUSB; names endometriosis among the estrobolome-linked conditions. Narrative review, not endometriosis-specific. [PMID 42716136](https://pubmed.ncbi.nlm.nih.gov/42716136/) — *why it matters:* the mechanistic backdrop for the gut-estrogen hypothesis including a precision-inhibitor angle; context only, no endo data. ## Epidemiology, genetics and access - **Genetic architecture of endometriosis — 25-year synthesis** — *Human Reproduction Update* review from the QIMR Berghofer group with **Aalborg University and Statens Serum Institut co-authors**: from Treloar's 1999 twin-study heritability estimate of 51% through GWAS risk loci and target genes, to the use of large public genetic datasets for comorbidity analysis and clinical implication. Invited review. [PMID 42695929](https://pubmed.ncbi.nlm.nih.gov/42695929/) — *why it matters:* the current single best map of endometriosis genetics, and it is co-authored from the Danish registry/biobank ecosystem — directly relevant to OpenEndo's Danish-registries and Danish-data workstreams. - **Endometriosis and fertility problems in a Polish national survey** — cross-sectional online survey, 5,532 respondents who had tried to conceive; fertility problems reported by 36.6% of women; among women, endometriosis carried OR 2.82 (vs 23.0 for tubal obstruction, 5.35 for PCOS). Survey epidemiology. [PMID 42715119](https://pubmed.ncbi.nlm.nih.gov/42715119/) — *why it matters:* population-level burden figure for a European cohort; useful comparator for Danish survey/registry framing. - **Correction to the ARCHES endometriosis/type-2-diabetes cohort paper** — erratum (Diabetologia) for the 2026-08-06 ARCHES retrospective cohort study on endometriosis subtypes and type-2-diabetes risk. Erratum only; the corrected text is not summarised in the record. [PMID 42709073](https://pubmed.ncbi.nlm.nih.gov/42709073/) — *why it matters:* if the comorbidity signal is cited anywhere (OpenEndo wiki, digest archives) it should be re-checked against the corrected version. ## Notable clinical signals - **Relugolix-CT (PMID 42719373)** is the one item this week that changes what a clinician could say to a patient: 104-week data on time-to- minimal-to-no pain and analgesic-free status for an oral GnRH-antagonist add-back regimen. Read it as sponsor-run post hoc analysis of RCT data, not as an independent trial. - **Pain-molecular signature (PMID 42720599)** is not yet clinical, but it is the strongest mechanistic answer this year to "why do I hurt and my scan doesn't show stage IV?" — IL16 is the name to watch. - Nothing this week changes any OpenEndo repurposing verdict; the FIN56 ferroptosis paper (PMID 42698143) reinforces the existing direction without altering the sulfasalazine selectivity question. ## Carried over / follow-ups checked - `PMID 42693330` (exosomal miR-22-3p→p53) reappears inside this window but was already synthesized in digest-2026-09-04 — not re-analyzed. - No follow-up appeared this week to the dienogest real-world patient-reported study (PMID 42692407) flagged in the last digest as a candidate Danish wiki note. - The AMH / cystectomy thread (PMID 42675751, 42672534) produced no new papers this week; the ANGPTL4 paper notes a negative ANGPTL4–AMH correlation but is not a fertility-outcome study. --- *Sources: PubMed snapshot `weekly/2026-09-11.json` (E-utilities esearch + esummary, `endometriosis[Title/Abstract]`, datetype=pdat, 2026-09-04..2026-09-11). Synthesized by Galahad (Hermes Agent), 2026-09-11. Research infrastructure, not medical advice.*