Prognostic values and clinical significance of POU2F3 expression in neuroendocrine carcinomas: a meta-analysis with a focus on small cell lung cancer

Small cell lung cancer (SCLC) represents a highly aggressive neuroendocrine malignancy, accounting for a significant proportion of lung cancer diagnoses [1]. Despite initial responsiveness to platinum-based chemotherapy, SCLC is notorious for rapid acquisition of resistance and a dismal prognosis [1]. Recent advancements have highlighted the importance of molecular subtyping in SCLC, particularly with the identification of four key transcriptional subtypes: ASCL1, NEUROD1, POU2F3 and YAP1 [2]. Unlike the ASCL1- and NEUROD1-driven tumours that display classic neuroendocrine features, the POU2F3 and YAP1 subtypes are recognized as a non-neuroendocrine category [2,3]. The POU2F3 subtype, in particular, is thought to originate from a unique cellular source – a specialized tuft-cell variant – and its reliance on the POU2F3 transcription factor may confer unique therapeutic vulnerabilities [2]. However, its prognostic significance remains a subject of debate [4–8].

While the majority of neuroendocrine carcinomas (NECs) develop in the lower respiratory tract, a smaller proportion, ranging from 2% to 9%, are classified as extrapulmonary NECs [9]. Extrapulmonary NECs are known to arise from various organ systems, including the gastrointestinal, hepatobiliary, pancreatic, genitourinary and gynaecologic tracts [10,11]. Extrapulmonary NECs share similarities with their pulmonary counterparts in terms of their morphological characteristics. Their diagnosis is primarily based on the assessment of their morphological features and the utilization of immunohistochemical techniques to identify neuroendocrine markers. There have been suggestions that in extrapulmonary NEC, a POU2F3 high subtype, potentially originating from tuft cells, could be identified as a distinct classification [12].

In the landscape of SCLC and extrapulmonary NECs, the POU2F3 subtype has emerged as a distinct entity, marked by varying clinical features and prognostic outcomes. However, the prognostic impact of this subtype remains unclear due to conflicting evidence. For example, some studies have reported that POU2F3-positive tumours are associated with improved survival [5,13], whereas others found improved outcomes or no significant prognostic difference [14,15]. This disparity underscores the need for more comprehensive studies to fully elucidate the clinical implications and prognostic significance of the POU2F3 subtype. Therefore, we conducted this systematic review and meta-analysis to synthesize the current evidence and clarify the prognostic value and clinical significance of POU2F3 expression in SCLC and extrapulmonary NECs.

This systematic review and meta-analysis of 16 studies involving 2011 patients provides important insights into the prognostic significance of POU2F3 expression in SCLC and extrapulmonary NECs. Our pooled analysis revealed that POU2F3 expression is not significantly associated with OS or PFS/RFS cross all NEC, but when restricted to SCLC, POU2F3 positivity demonstrated favourable association with both OS and PFS. Subgroup analyses further indicated that the prognostic value of POU2F3 may be influenced by specimen collection methods, treatment modalities and analytical approaches, with stronger associations with prognosis observed in studies using surgical specimens and chemotherapy regimen. Additionally, POU2F3 expression was significantly associated with younger patient age and male sex, but showed no correlation with smoking history, tumour stage, nodal involvement, or vascular invasion.

The prognosis of the POU2F3 subtype remains a subject of conflicting evidence. Ding et al. reported that the SCLC-P subtype was identified as an independent poor prognostic factor in patients with resectable small cell lung cancer in a multivariate analysis. In their study, which included 53 patient samples, 9 cases (17.0%) were classified as the POU2F3 subtype [4]. Similarly, Deng et al. reported that the POU2F3 subtype was associated with an unfavourable prognosis in univariate analysis (p = .038). However, when adjusted for tumour stage in a multivariate Cox analysis, the association was not statistically significant (p = .098) [23].

Conversely, other studies reported improved or neutral prognostic implications of POU2F3 expression. In a multicentre international study by Megyesfalvi et al. patients with high POU2F3-expressing tumours exhibited significantly improved overall survival (OS) in univariate analysis (p = .046). In this study, the POU2F3 subtype was present in 16% of cases; however, the findings did not achieve statistical significance in multivariate analysis (p = .08) [5]. Similarly, Jimbo et al. observed that NECs with POU2F3 expression (26% of cases) demonstrated significantly improved relapse-free survival compared to their non-POU2F3 counterparts (p = .026). However, no statistically significant difference in overall survival was found (p = .172) [13]. Zhu et al. found no significant relationship between POU2F3 expression and prognosis, although a non-significant trend toward better outcomes was noted [24].

The conflicting evidence regarding the prognostic role of POU2F3 may be partially explained by the methodological heterogeneity between studies, which our meta-analysis sought to clarify. For instance, we found a significant association between POU2F3 expression and improved survival only in studies that used surgical specimens, whereas no significance was found when both surgical and biopsy specimens were included, as well as in ES-SCLC, in which surgical resection is not considered a primary treatment option. The lack of significance in ES-SCLC is likely attributable not only to the reliance on small biopsy specimens but also to treatment-induced temporal heterogeneity, which can alter molecular subtypes through chemotherapy or disease progression [27,28]. This suggests that the limited tissue volume in biopsies may be insufficient for accurate subtyping, potentially diluting the prognostic association. POU2F3 is known for its considerable intratumoural heterogeneity in SCLC which could affect difference in prognostic association between surgical and biopsy specimen [29]. Although studies directly comparing biopsy and surgical resection specimens are lacking, one study by Handa et al. provides valuable insight [19]. This study shows a significant correlation for SCLC subtyping among paired whole slides, single TMA cores and lymph node metastases (p < .001), suggesting that biopsy specimens can be useful for detecting molecular subtypes if properly sampled; for example, avoiding necrosis [19]. In addition, our subgroup analysis revealed a strong correlation between the POU2F3-positive subtype and OS, and this association was consistent whether the immunohistochemistry was based on whole slides or TMA and whole slides. However, the finding related to TMAs should be interpreted with caution, as it is based on only two studies.

Furthermore, the choice of analytical method and positivity criteria appears to influence prognostic significance. Our analysis showed that POU2F3 positivity was linked to better overall survival when specific cut-off criteria were used (H-score >10 or 1% cutoff), but not when defined by the broader SCLC-P classification with predominant POU2F3 expression and H-score >50.

POU2F3-driven small cell carcinomas are biologically distinct, characterized by the absence of classic neuroendocrine lineage markers, including synaptophysin, chromogranin and INSM1 and the expression of tuft cell lineage markers [18,30]. This unique biology suggests that POU2F3-positive SCLC may have a different prognosis than its neuroendocrine-positive counterparts. However, the nature of the tumour microenvironment (TME) and its impact on immunotherapy response in POU2F3-positive SCLC (SCLC-P) remain controversial. On one hand, some studies suggest that low-neuroendocrine SCLC (a group that includes SCLC-P) is an immune-'hot' phenotype with high T-cell infiltration and PD-L1 expression, making it a promising candidate for immunotherapy [31,32]. This view is complicated by several factors. For instance, the immune-active SCLC-I (Inflamed) subtype, which lacks both neuroendocrine and POU2F3 markers, may confound this association [33]. On the other hand, conflicting reports indicate that a more immune-rich TME and better OS are found in SCLC subtypes expressing conventional neuroendocrine markers or low POU2F3 expression [34–36]. Further complicating the picture, a study by Velut et al. concluded that SCLC molecular subtypes do not predict immunotherapy efficacy [37].

Our findings contribute to this complex discussion. In our subgroup analysis, POU2F3 expression was associated with a favourable prognosis in patients treated with traditional chemotherapy alone. However, this prognostic benefit was lost in patients who received chemoimmunotherapy. There are two primary explanations for this discrepancy. First, the smaller number of studies in the chemoimmunotherapy group (k = 2 for OS and PFS/RFS), and this issue is compounded by the fact that POU2F3-positive tumours are relatively rare, which further limits the effective sample size and statistical power in a meta-analysis. Alternatively, this finding could imply that the addition of immunotherapy has a less favourable, or potentially even detrimental, effect on POU2F3-positive tumours, thereby negating the prognostic benefit observed with chemotherapy alone. Regardless of the cause, our results challenge the notion that POU2F3-positive SCLC is an inherently immune-responsive subtype that benefits from immunotherapy regimens.

Extrapulmonary NEC is a heterogeneous group of tumours that are morphologically and genetically similar to SCLC. For example, mutation of TP53 and RB1 was frequent in various extrapulmonary NECs and SCLCs. However, difference in clinically, genetically, and immunologically features are also found both within extrapulmonary NECs and compared with SCLC. Rb1 loss and p16 overexpression are found more often in genitourinary small cell NECs than in small cell NECs of other organs [38]. In bladder small cell NECs, TERT promoter mutation is common but not in other NECs and SCLC [39]. Moreover, Patients with genitourinary small cell NECs had better OS than those with gastrointestinal tract small cell NECs [38]. In SCLC and extrapulmonary NECs, expression level of Phospholipase C Gamma 2(PLCG2), a transmembrane signalling enzyme that produces DAG and IP3, has been strongly associated with POU2F3-positive subsets [8,18,40]. High expression of PLCG2 was associated with increased invasiveness, stem-like features and prometastatic potential in SCLC [40]. This suggests that PLCG2 could be a potential therapeutic target for POU2F3-positive SCLCs. However, PLCG2 or POU2F3 expressions were not associated with metastatic status or response to chemotherapy in bladder NECs. Moreover, POU2F3-positve NECs resembled 'immune cold' microenvironment in SCLC, which represents low density of tumour-infiltrating lymphocyte infiltration or PD-L1 expression [41]. This finding contrasts that of cervical small cell NECs, where HPV infection is the primary aetiology, in which POU2F3-positive subtype displays an 'immune hot' phenotype with active immune cell infiltration [42]. This discrepancy underscores that the prognostic and immunological characteristics of POU2F3-positive tumours are not universal but are determined by the unique biological context and aetiology of the primary organ. Therefore, it is crucial to avoid uniformly applying the SCLC paradigm when understanding and developing treatment strategies for extrapulmonary NEC, emphasizing the necessity of an organ-specific approach.

Despite our best efforts to provide valuable insights through this systematic review, several limitations must be acknowledged. First, heterogeneity among the included studies poses a significant challenge. Variations in study design, the organs involved, associated tumour types, detection methods for POU2F3 expression – including differences in immunohistochemical protocols, antibodies used and types of biomaterials – and criteria for defining POU2F3 positivity may have introduced bias and affected the comparability of results. Second, studies that did not report dichotomous POU2F3 versus non-POU2F3 expression, such as survival analysis from Qi et al. [22], were excluded. Third, the lack of standardized reporting on treatment modalities across studies may have confounded the analysis, as variations in drug regimens and surgical methods could potentially influence survival outcomes independently of POU2F3 expression. Finally, our study has a significant imbalance in the data between SCLC and extrapulmonary NECs. Our meta-analysis included fourteen studies focused on SCLC but only two study on extrapulmonary NECs. Consequently, while our findings are robust for the SCLC population, they cannot be generalized to all extrapulmonary NECs. The conclusions drawn regarding extrapulmonary NECs are therefore speculative and highlight a critical need for more research focused on this patient group.

Our study suggests that POU2F3 expression in SCLCs is associated with improved survival, but not when considering both SCLCs and extrapulmonary NECs. Additionally, POU2F3 positivity was associated with younger age and male sex, but not with combined histology or smoking. These findings highlight the distinct characteristics of POU2F3-positive tumours and emphasize the need for standardized detection methods. To clarify the role of POU2F3 as a prognostic marker and therapeutic target, future research should focus on SCLC and, critically, expand to include more studies on the diverse group of extrapulmonary NECs.