Skip to content

Molecular and genomic features of uterine adenosarcoma

TL;DR — Targeted sequencing of 20 samples from 18 patients found no mutation-count difference between tumours with and without sarcomatous overgrowth, but a large copy-number difference (mean 24.6 vs 5 gene-level CNVs, p=0.0002); recurrent events were MDM2/CDK4 amplification (5/18, 28%), MYBL1 amplification (4/18, 22%), PIK3CA/AKT/PTEN pathway alteration (13/18, 72%), and ATRX mutation in SO-associated cases (3/18) (Howitt 2015, PMID 25231023). Subsequent series added TERT promoter mutation/amplification, recurrent BAP1 homozygous deletion (~25%), DICER1 mutation in a minority, and TP53 pathway alteration concentrated in high-grade tumours (Geyer 2017, PMID 28267263; Momeni Boroujeni 2022, PMID 36138078; Hodgson 2017, PMID 28834809). A 2026 series identified recurrent ESR1::NCOA3/2 fusions in 12 adenosarcoma-like uterine neoplasms, but the authors regard their relationship to conventional adenosarcoma and UTROSCT as unresolved (Agaimy 2026, PMID 41555057). None is a validated standalone diagnostic test; all molecular evidence comes from small, selected series.

Howitt 2015 — the targeted-panel baseline

DNA from 20 samples of Müllerian adenosarcoma (18 subjects: 12 without SO, 6 with SO; two patients contributed paired typical and SO areas) was sequenced on a 275-gene exon panel plus 91 introns for rearrangements (Howitt 2015, PMID 25231023). Mean mutation count was 9.7 (range 3–14) with SO versus 9.6 (5–16) without — not different. Mean gene-level CNVs were 24.6 with SO versus 5 without (p=0.0002). Recurrent findings:

Alteration Frequency Notes
MDM2 and CDK4 amplification 5/18 (28%) Accompanied by focal CDK4 and MDM2 and diffuse HMGA2 IHC
MYBL1 amplification 4/18 (22%) Predominantly in SO
PIK3CA/AKT/PTEN pathway 13/18 (72%) Heterogeneous members
TP53 mutation 2 cases, both with SO Uncommon overall
ATRX mutation 3/18 (17%) All associated with SO
Chromosomal rearrangements 0/18 None identified on this panel

Howitt’s companion cytogenetic paper on 21 samples (14 successful karyotypes) found chromosome 8 abnormalities in 5 of 7 tumours with noncomplex clonal aberrations, including 8q13 rearrangements in two (Howitt 2017, PMID 26974998). Two tumours (one SO, one metastasis) were extremely aneuploid.

The authors’ own conclusion: specific molecular or IHC tools for diagnosis were lacking, and confirmation of the recurrent alterations would be needed before any became a marker (Howitt 2015, PMID 25231023). The literature search rerun 2026-09-01 found no prospectively validated diagnostic molecular or IHC assay for this histology.

Copy-number complexity, TERT, and the phyllodes comparison

Geyer compared 19 adenosarcomas with 22 breast phyllodes tumours (6 benign, 6 borderline, 10 malignant) (Geyer 2017, PMID 28267263). Phyllodes more often had MED12 mutations, TERT promoter mutations and bona fide cancer-gene mutations; adenosarcomas had a higher rate of MDM2/CDK4 and TERT gene amplifications. Wnt-pathway genes were enriched in both adenosarcomas and benign/borderline phyllodes. Morphologic similarity is therefore only partly genomic similarity; MED12, the signature of uterine leiomyoma and of many phyllodes tumours, is not a driver of adenosarcoma in this comparison.

High-grade disease: TP53, ATRX, genome instability

Hodgson sequenced 9 high-grade adenosarcomas (409-gene panel). TP53 pathway alterations were identified in 7/9 (78%); p53 IHC correlated with mutation status. Mean copy-number variations were 28.8 per tumour; frequently involved genes included CDK4, MDM2, GNAS, SGK1 and DICER1. Six of 9 patients developed rapid recurrence (Hodgson 2017, PMID 28834809). The authors proposed high-grade adenosarcoma as a distinct subset with driver TP53 pathway alteration, and recommended reporting any high-grade component even without sarcomatous overgrowth.

Momeni Boroujeni profiled 27 adenosarcomas, enriched for high-grade. High-grade tumours had more SO (71% vs 10%) and exclusive heterologous elements; all deaths were high-grade. Genetic alterations specific to high-grade disease included TP53 mutations (n=4) and MDM2 and CCNE1 amplifications (n=2 each). ATRX frameshift mutations were found in 2 patients with high-grade recurrences after a primary low-grade adenosarcoma, and ATRX deletion in 1 high-grade tumour with an adjacent low-grade component — a possible transformation mechanism, n=3 (Momeni Boroujeni 2022, PMID 36138078). Fraction of genome altered was higher in high-grade than low-grade (p=0.001). Across the whole cohort: BAP1 homozygous deletion (n=4), DICER1 mutations (n=4), ARID1A mutations (n=3), TERT promoter mutation (n=2) plus amplification (n=1), PI3K and MAPK pathway alterations, one ESR1-NCOA3 fusion, one MLH1 homozygous deletion. BAP1 IHC loss in 6/24 (25%), including all four deletion cases. In 196 gynecologic mesenchymal neoplasms, BAP1 homozygous deletion was found only in adenosarcoma (p=0.0003) (Momeni Boroujeni 2022, PMID 36138078). That last finding is the closest thing in this literature to a relatively specific molecular marker, and it still covers only a minority of cases.

DICER1 — overlap with embryonal rhabdomyosarcoma, not a diagnostic positive

de Kock compared centrally reviewed uterine tumours originally thought to be ERMS or adenosarcoma: consensus ERMS n=19, adenosarcoma n=27, no consensus n=18. DICER1 alterations: 18/19 (95%) ERMS, 7/27 (26%) adenosarcomas (p<0.001), 4/18 (22%) no-consensus. Germline DICER1 alteration in 6/12 ERMS patients tested versus 0/6 adenosarcoma patients (de Kock 2020, PMID 31900434). Median age 30 years (range 2.5–69) for ERMS versus 57.5 (27–82) for adenosarcoma. The authors’ operational conclusion: DICER1 testing is useful in the differential only when negative, because a negative result makes ERMS unlikely; a positive result does not confirm ERMS. Patients with uterine ERMS should be referred for DICER1 germline testing; that recommendation does not extend to typical adenosarcoma on this evidence (de Kock 2020, PMID 31900434; Apellaniz-Ruiz 2021, PMID 33135284).

Atypical polyps share some copy-number events

Chapel’s 21 molecularly profiled atypical uterine polyps: chr 12q13–15 gain/amplification in 5/21 (24%), chr 6q25.1 gain in 9/20 (45%), no significant CNVs in 7/21 (33%), mean tumour mutational burden 3.1 mutations/Mb (range 0.76–8.4), pathogenic point mutations in 12/20 (60%) (Chapel 2022, PMID 34675347). None developed adenosarcoma over 150 months median follow-up. The molecular overlap with early adenosarcoma is real; the clinical course in this series was not.

Recurrent ESR1 fusions define an unresolved adenosarcoma-like subset

Agaimy and colleagues assembled 12 uterine neoplasms with adenosarcoma-like morphology and ESR1 fusions, then reviewed four previously reported cases (16 total). The study series contained ESR1::NCOA3 in 10, ESR1::NCOA2 in one, and ESR1::MAMLD1 in one; 9/12 were low-grade, 3/12 high-grade, and all lacked heterologous elements. Follow-up was available for only six patients: one lung metastasis at 52 months, one abdominopelvic recurrence after more than 20 years, and four disease-free at 9–55 months. The authors explicitly leave unresolved whether these tumours belong to conventional adenosarcoma, UTROSCT, or a separate fusion-defined group (Agaimy 2026, PMID 41555057). This is a 12-case molecular series, not validation of a diagnostic assay.

What is not diagnostically usable

  • Howitt found no rearrangements on a 30-gene intron panel (Howitt 2015, PMID 25231023), and Momeni Boroujeni reported one ESR1-NCOA3 fusion (PMID 36138078). The 2026 Agaimy series now establishes recurrence of ESR1::NCOA3/2 in an adenosarcoma-like subset, but its nosological specificity and diagnostic performance are untested (Agaimy 2026, PMID 41555057).
  • MDM2/CDK4 co-amplification is shared with well-differentiated/dedifferentiated liposarcoma and is not specific.
  • BAP1 loss is relatively specific among gyn mesenchymal tumours in one series of 196, and still only ~25% sensitive (Momeni Boroujeni 2022, PMID 36138078).
  • TP53/p53 IHC identifies a high-grade subset; it does not diagnose adenosarcoma.
  • PIK3CA/AKT/PTEN pathway alteration in ~70% (Howitt 2015, PMID 25231023; Nathenson 2016, PMID 27718181) has not been translated into a trial of a PI3K-pathway agent in this histology.

Open questions

  • Does ATRX loss mark the low-grade → high-grade transition, or is it a passenger in already-complex genomes? n=3 observations (Momeni Boroujeni 2022, PMID 36138078).
  • Can BAP1 IHC be used as a diagnostic adjunct in the atypical-polyp versus adenosarcoma differential? No diagnostic-accuracy validation study was identified in the PubMed search rerun 2026-09-01.
  • Do ESR1::NCOA3/2-positive adenosarcoma-like tumours belong to adenosarcoma or UTROSCT, and can methylation or other epigenetic profiling resolve the classification? Current evidence is a 12-case series plus four literature cases (Agaimy 2026, PMID 41555057).
  • Is the PIK3CA/AKT/PTEN pathway a therapeutic target in adenosarcoma? Frequency 13/18 (72%) in Howitt’s selected series (PMID 25231023); the ClinicalTrials.gov searches rerun 2026-09-01 identified no PI3K-pathway-selected adenosarcoma trial.

References

  1. Howitt BE, et al. Targeted genomic analysis of Müllerian adenosarcoma. J Pathol. 2015;235:37-49. PMID 25231023
  2. Howitt BE, et al. Involvement of Chromosome 8 in Müllerian Adenosarcoma. Int J Gynecol Pathol. 2017;36:24-30. PMID 26974998
  3. Geyer FC, et al. Genetic analysis of uterine adenosarcomas and phyllodes tumors of the breast. Mol Oncol. 2017;11:913-926. PMID 28267263
  4. Hodgson A, et al. High-grade Müllerian Adenosarcoma: Genomic and Clinicopathologic Characterization of a Distinct Neoplasm With Prevalent TP53 Pathway Alterations and Aggressive Behavior. Am J Surg Pathol. 2017;41:1513-1522. PMID 28834809
  5. Momeni Boroujeni A, et al. Mullerian adenosarcoma: clinicopathologic and molecular characterization highlighting recurrent BAP1 loss and distinctive features of high-grade tumors. Mod Pathol. 2022;35:1684-1694. PMID 36138078
  6. de Kock L, et al. Significantly greater prevalence of DICER1 alterations in uterine embryonal rhabdomyosarcoma compared to adenosarcoma. Mod Pathol. 2020;33:1207-1219. PMID 31900434
  7. Apellaniz-Ruiz M, McCluggage WG, Foulkes WD. DICER1-associated embryonal rhabdomyosarcoma and adenosarcoma of the gynecologic tract. Genes Chromosomes Cancer. 2021;60:217-233. PMID 33135284
  8. Chapel DB, et al. Atypical uterine polyps show morphologic and molecular overlap with mullerian adenosarcoma but follow a benign clinical course. Mod Pathol. 2022;35:106-116. PMID 34675347
  9. Nathenson MJ, et al. Uterine Adenosarcoma: a Review. Curr Oncol Rep. 2016;18:68. PMID 27718181
  10. Agaimy A, et al. ESR1::NCOA2/3 fusions in uterine neoplasms with adenosarcoma-like morphology: clinicopathologic and molecular features of 12 cases and review of the literature. Virchows Arch. 2026. PMID 41555057