Leiomyosarcoma: 2026 Research Update

A 2026 research update on uterine and soft-tissue leiomyosarcoma covering cell signalling, metabolism, immune landscape, resistance mechanisms, clinical trials, and current standard care.

This update covers uterine leiomyosarcoma (uLMS) and leiomyosarcoma arising elsewhere in soft tissue (ST-LMS).
They share important biology but are not one uniform disease: primary site, molecular subtype, hormone-receptor status and immune context can all matter.

The short version

The clearest recent clinical advance is the phase III LMS-04 trial. In previously untreated, unresectable or metastatic LMS, doxorubicin plus trabectedin for six cycles followed by trabectedin maintenance improved median progression-free survival from 6 to 12 months and median overall survival from 24 to 33 months compared with doxorubicin alone. Toxicity was greater, so suitability remains individual. [3, 4]

Research increasingly divides LMS into biological subgroups. Recurring themes include disruption of TP53 and RB1 cell-cycle control, PTEN loss and PI3K–AKT–mTOR activation, DNA-repair abnormalities, alternative telomere maintenance, MYC/BET programs, variable hormone signalling, metabolic plasticity and distinct immune states. [1, 2, 5, 6]

Key signalling pathways

System	What researchers see	Why it matters	Refs
TP53 and RB1	Loss of major cell-cycle checkpoints permits damaged cells to continue dividing and contributes to genomic instability.	Important disease drivers, but lost tumour suppressors are difficult to target directly.	5
PTEN–PI3K–AKT–mTOR	PTEN loss removes a brake on growth, survival and metabolism.	Strong preclinical rationale; feedback and heterogeneity can permit escape.	2, 6, 7
DNA repair and telomeres	Some uLMS has homologous-recombination deficiency; ATRX/DAXX abnormalities may support alternative telomere lengthening.	Supports biomarker-led study of PARP, ATR and related approaches—not treatment of all LMS as HR-deficient.	2, 6
MYC/BET transcription	MYC amplification or BET dependence occurs in a subset.	Matched patient-derived models have responded to BET inhibition; still preclinical.	6
Smooth-muscle lineage	MYOCD/SRF and related differentiation programs vary across molecular subtypes.	May improve classification and target selection, but is not yet a routine treatment selector.	1, 6
Hormone signalling	A subset, mainly uterine LMS, expresses oestrogen and/or progesterone receptors.	Endocrine therapy may suit selected ER/PR-positive, lower-volume or slower-growing disease.	8, 9
Angiogenesis/stroma	VEGF/VEGFR, PDGFR, AXL and stromal signals can promote growth and immune exclusion.	Supports anti-angiogenic treatment and combination research.	2

Metabolic map

Glucose and central carbon metabolism: LMS and related sarcoma models show altered glycolysis, glutamate metabolism and TCA-cycle activity, supplying energy and biosynthetic material. [10, 11]
Glutamine: a 2024 primary soft-tissue sarcoma mouse study—not LMS-only—found that radiation increased glutamine use and that glutaminase inhibition increased radiosensitivity in vivo. [12]
Arginine and ASS1: ASS1-low sarcomas can depend on extracellular arginine. Resistance may develop through ASS1 restoration or uptake and digestion of extracellular proteins through macropinocytosis and autophagy. [13, 14]
Redox defence and ferroptosis: glutathione, NADPH and lipid-peroxide defences can support stress survival. Ferroptosis is promising laboratory research, but no validated LMS-specific ferroptosis therapy exists. [15]
Autophagy: nutrient and treatment stress can trigger cellular recycling that supports survival. Direct, prospective LMS evidence remains limited. [16]

The practical message is metabolic flexibility: blocking one fuel route may drive use of another. These findings do not establish a therapeutic diet or supplement protocol. [12, 13]

Immune map

LMS is neither uniformly immune "hot" nor uniformly "cold." Immune infiltration is often limited and single-agent PD-1 treatment has shown little activity in unselected LMS, but biologically distinct immune subgroups exist. [2, 17]

Component	Current picture	Refs
CD8 T cells	Often sparse or excluded, but present in a more inflamed subset. Weak antigen presentation and low tumour mutational burden can limit recognition.	2, 17
B cells and tertiary lymphoid structures	A subset contains organised B- and T-cell aggregates. TLS are being studied as biomarkers of coordinated local immunity.	18
Tumour-associated macrophages	Frequently prominent, often with a suppressive CD163-positive/M2-like phenotype. Macrophage-directed targets remain investigational.	17
PD-1/PD-L1	PD-L1 is variable and insufficient by itself to select most patients. A small uLMS nivolumab study reported no objective responses.	2, 17
PTEN/PI3K	PTEN loss can support immune exclusion. A resistant metastasis in an exceptional pembrolizumab responder acquired biallelic PTEN loss and fewer immune infiltrates.	19
Myeloid and regulatory suppression	Macrophages, regulatory T cells, angiogenesis, cytokines and metabolic competition can weaken T-cell function.	17

Cell and animal research

Matched uLMS patient-derived xenografts responded to olaparib, BET inhibition or PI3K inhibition when the corresponding HRD, MYC/BET or PTEN/PIK3CA vulnerability was present. [6]
Dual PI3K/mTOR inhibition reduced growth in four of five uLMS patient-derived xenografts and shrank two; phosphorylated S6 was proposed as a pathway-activity marker. [7]
Glutaminase inhibition radiosensitised primary soft-tissue sarcomas in mice after radiation increased glutamine use; this is not yet routine LMS evidence. [12]
Arginine-starved sarcoma cells survived in vivo by scavenging extracellular proteins; blocking macropinocytosis or autophagy restored sensitivity in models. [13]

Animal models have major limits. Conventional xenografts often use immunodeficient mice and cannot reproduce the full human immune system, organ environment, microbiome or treatment history. [20]

Treatment resistance

Resistance is usually a network problem rather than one mutation. Treatment can select pre-existing resistant clones, while surviving cells alter drug transport, DNA repair, apoptosis, metabolism and interactions with the tumour microenvironment. [2, 21]

Drug efflux: doxorubicin-selected MES-SA sarcoma cells developed multidrug resistance with lower intracellular drug accumulation. This is a laboratory model, not a validated patient biomarker. [21]
DNA repair: trabectedin response depends partly on nucleotide-excision and homologous-recombination repair biology; adaptation can permit escape. [22]
Immune escape: PTEN loss, reduced immune infiltration and reduced expression of recognised neoantigens were found in a pembrolizumab-resistant uLMS metastasis. [19]
Metabolic escape: arginine-deprived tumours may restore ASS1 or scavenge extracellular proteins through macropinocytosis and autophagy. [13, 14]
Mixed-sarcoma models have linked doxorubicin resistance with TP53 abnormalities, copy-number complexity and WNT, Hedgehog and TGF-beta-related programs. These are research leads, not established LMS prescribing biomarkers. [23]

Clinical-trial news

Practice-changing positive result: LMS-04 provides randomised evidence that first-line doxorubicin–trabectedin followed by trabectedin maintenance improves progression-free and overall survival over doxorubicin alone in advanced LMS. [3, 4]

PARP reality check: a small phase II uLMS study of olaparib plus temozolomide reported a 27% response rate and median progression-free survival of 6.9 months. The subsequent randomised Alliance A092104 trial in a biomarker-unselected population did not improve PFS over investigator's choice and stopped for futility. Better HRD selection remains a research need. [24, 25]

Trials to watch: the phase III SaLuDo trial is comparing lurbinectedin plus doxorubicin with doxorubicin alone as first-line metastatic LMS treatment. Recruitment has completed and results are expected in 2027; efficacy should not be assumed before results are reported. [26, 27]

Current standard care — brief

Localised LMS: complete specialist surgery with clear margins is the foundation. Radiation is used selectively for local-control risk; adjuvant chemotherapy has no universal role. [28, 29]
Completely resected early uterine LMS: observation remains common because routine adjuvant systemic treatment has not shown a clear survival benefit. [30]
Unresectable or metastatic first line: doxorubicin plus trabectedin followed by maintenance has the strongest LMS-specific randomised survival evidence for a suitable fit patient. Doxorubicin alone remains reasonable when combination toxicity or access is a concern. [3, 4]
Later lines: options can include trabectedin, pazopanib, gemcitabine/docetaxel or other gemcitabine regimens, dacarbazine and clinical trials; sequence is individual. [29, 31]
Selected ER/PR-positive uLMS: an aromatase inhibitor or other endocrine approach may suit lower-volume or more indolent disease. [8, 9]
Selected limited metastases: surgery, stereotactic radiation or thermal ablation may be considered within multidisciplinary care. [28]

Australian note: In Australia, trabectedin is TGA-approved and PBS-listed for unresectable or metastatic liposarcoma or LMS after prior anthracycline treatment. eviQ lists a 24-hour trabectedin regimen every 21 days in that setting. Access to first-line doxorubicin–trabectedin may not automatically match overseas practice, so current PBS and hospital access should be checked with the sarcoma team. [32, 33]

What to ask after progression

Has the pathology been reviewed by a specialist sarcoma pathologist?
Would biopsy of a progressing site add useful information?
Is broad molecular profiling appropriate, including rare tumour-agnostic markers?
Has ER/PR testing been considered for uterine LMS?
Is progression limited enough for surgery, focused radiation or ablation?
Is an LMS-specific or biomarker-selected clinical trial available?

Important caution: "Pathway present" does not mean "drug will work." Many LMS abnormalities are losses of tumour suppressors rather than easily druggable targets, and different tumour regions may differ. Cell and mouse activity often fails to translate into patient benefit. No diet, supplement or off-label metabolic intervention should be presented as 'treating LMS' on the basis of preclinical findings. Potential interactions with current treatment, the person's liver function, their blood counts and tumor biology all require review.

Linked references

Withaferin A, HCQ and Autophagy