No specific molecular profile (NSMP) endometrial cancer: from residual category to clinical decision driver

For years, No Specific Molecular Profile (NSMP) endometrial cancer has been regarded as a residual category, defined by the absence of polymerase-epsilon (POLE) mutations, mismatch repair deficiency (MMRd) or p53 abnormalities (p53abn). In clinical practice, this translated into management driven by traditional clinicopathological factors such as stage, grade and lymphovascular space invasion (LVSI). While molecular classification has reshaped endometrial cancer care, NSMP remained the least informative and most heterogeneous subgroup.

This paradigm is now shifting. NSMP tumours are not simply a molecular ‘default’ category, but a biologically diverse group that can be stratified into subsets. Among biomarkers explored within NSMP, oestrogen receptor (ER) expression has consistently demonstrated reproducibility and clinical relevance. Multiple retrospective studies have shown that ER-positive NSMP tumours are associated with more favourable outcomes than ER-negative disease, supporting its role as a robust discriminator.1–5

The issue of ER cut-off should not be viewed as a limitation undermining ER assessment, but as the next methodological step in its implementation. While Jamieson et al adopted a >1% threshold and identified an exceptionally low-risk grade 1–2/ER-positive NSMP population, both Vermij et al and the ENDOESTRO analysis support ER <10% as a clinically meaningful definition of ER-low/negative disease for prognostic risk stratification.1 6 7 Clinically, this threshold avoids grouping borderline ER-low tumours (1%–9%) with clearly ER-high tumours; biologically, low ER expression may indicate attenuated hormone-receptor signalling, although a sharp 1%–10% boundary remains unproven. Thus, a 10% nuclear staining threshold currently represents the most pragmatic cut-off. Semi-quantitative systems such as the Allred score may further refine ER assessment, but remain promising rather than prospectively validated.8 Although progesterone receptor expression may reflect preserved hormone-receptor signalling, available NSMP-specific data suggest that its prognostic value largely tracks with ER expression and does not add independent prognostic information beyond ER-based stratification.1

A major turning point has been the evolution from ER-based stratification alone towards an integrated ER-and-grade approach. Large retrospective series and institutional datasets support a pragmatic two-tiered model distinguishing low-risk NSMP tumours (ER-positive and grade 1–2) from high-risk NSMP tumours (ER -negative and/or grade 3).6 9 The magnitude of outcome differences observed with this combined stratification appears greater than that achieved by ER status alone, suggesting that NSMP tumours should be viewed as clinically distinct entities.

This evolution is reflected in clinical taxonomy. This interpretation is consistent with recent position papers and with the 2025 European Society of Gynaecological Oncology (ESGO)–European Society for Radiotherapy and Oncology (ESTRO)–European Society of Pathology (ESP) guidelines, which increasingly treat NSMP ER-negative disease as a high-risk subgroup.10 11 High-risk NSMP tumours display recurrence patterns approaching those of p53abn carcinomas, reinforcing the notion that they represent a biologically aggressive subset rather than an intermediate-risk extension of ER-positive disease.6 Nevertheless, because subgroup-specific treatment data remain retrospective or derived from cohorts not originally designed around NSMP ER/grade subgroups, prospective validation remains essential to define treatment escalation or de-escalation.

The clinical implications are evident in the 2025 ESGO–ESTRO–ESP guidelines, which incorporate ER status—and, increasingly in practice, its integration with grade—into risk stratification.11 ER-positive NSMP tumours generally align with low–intermediate-risk categories, whereas ER-negative and/or grade 3 tumours are classified within high-intermediate or high-risk groups, leading to different adjuvant strategies. However, the strength of evidence supporting subgroup-specific recommendations remains low (IV,C), underscoring the need for caution while these concepts await prospective validation.11 12

Classical clinicopathological variables remain essential. LVSI carries independent prognostic significance and should be interpreted alongside, not replaced by, ER/grade-based NSMP stratification. Thus, a low-grade, ER-positive NSMP tumour with substantial LVSI or deep myometrial invasion should not be considered equivalent to an otherwise similar tumour without these features, and may still warrant nodal staging interpretation and/or adjuvant treatment escalation. Conversely, ER-negative status identifies an adverse NSMP phenotype even in the presence of otherwise favourable clinicopathological features. Depth of myometrial invasion and cervical involvement likewise remain relevant.13 14

Further refinement of NSMP classification provides clearer biological context for novel prognostic biomarkers. Isolating a more homogeneous low-risk NSMP patient group enables accurate evaluation of markers with modest predictive impacts. L1CAM stands out as the most robust candidate: it mediates cell adhesion, tumour invasion and LVSI to drive poor survival outcomes, with the postoperative radiation therapy in endometrial carcinoma (PORTEC) trial data validating its incremental prognostic utility.15–17 L1CAM-positive low-risk NSMP tumours may merit treatment escalation, yet this clinical hypothesis lacks definitive clinical verification (figure 1). In contrast, CTNNB1 and ARID1A mutations stay exploratory. Conflicting cohort results, non-standardised detection assays and the absence of matched targeted therapies collectively hinder their routine adoption in clinical decision-making.18–20

In advanced and recurrent disease, NSMP heterogeneity is clinically relevant but not yet directly actionable in the same way as MMR status. Most pivotal trials stratify patients by MMR status rather than complete molecular classification, and first-line strategies for advanced/recurrent mismatch repair-proficient disease are not routinely diversified according to ER/grade-defined NSMP subgroups. Still, ER-positive/low-grade NSMP tumours are the most plausible candidates for endocrine therapy or endocrine-based combinations, whereas ER-negative and/or grade 3 tumours may reflect a less hormone-dependent phenotype requiring conventional systemic approaches.21 In the adjuvant setting, the refining adjuvant treatment in endometrial cancer based on molecular features (RAINBO)/NSMP-ORANGE prospectively evaluates ER-positive NSMP disease by comparing radiotherapy plus progestin maintenance with chemoradiation.22 Future trials should avoid treating NSMP as a single residual category and incorporate ER status and grade as stratification variables.

Despite these advances, challenges remain. Evidence supporting ER- and grade-based stratification remains largely retrospective and limited; ER thresholds, interlaboratory reproducibility and reporting standardisation require harmonisation and equitable implementation requires infrastructure, training and reimbursement models. Prospective evidence confirms high concordance between biopsy-based and surgical molecular classification, supporting early risk-adapted treatment decisions.23

Taken together, the reframing of NSMP endometrial cancer represents one of the most clinically meaningful developments in contemporary gynaecological oncology. What was once considered a residual category is now emerging as a central component of biologically driven decision-making. The transition from ER-based stratification to an integrated ER-and-grade model strengthens clinical relevance while highlighting gaps in evidence and implementation. The challenge ahead is not only to validate these emerging concepts, but also to determine how decisively they should influence treatment algorithms in routine practice.