A diverse range of WNT-activated tumors have been recognized, such as BA, DF, SPN, hepatocellular adenoma, deep penetrating nevi, medulloblastoma, and endometrial carcinomas. The primary cause of abnormal WNT pathway activation is typically associated with the CTNNB1 gene mutation encoding the β-catenin protein. However, alterations in the APC and AXIN1 have been implicated. The phosphorylation site located in exon 3 of the CTNNB1 gene is a prominent mutation hotspot [7, 8].

Due to morphological diagnostic difficulties, immunostaining is often used to identify these lesions, especially in needle aspiration specimens. While β-catenin is a traditional marker for diagnosing WNT pathway tumors, its evaluation remains challenging. Previous studies exploring the relationship between β-catenin expression and CTNNB1 gene mutation in various WNT-activated tumors have yielded diverse conclusions. The prevailing view once was that only nuclear β-catenin staining was significant. However, recent studies have indicated its inconsistencies. For example, Kim et al. reported eight patients with endometrial carcinoma exhibiting β-catenin cytoplasmic labeling despite CTNNB1 gene mutation [9]. Explanations for the lack of strict association between the CTNNB1 gene mutation and β-catenin nuclear staining have been provided in the literature. Hagen et al. reported that membranous accumulation of β-catenin can activate WNT pathway transcription [10]. In the current study, a few patients with BA and SPN showed exclusive membranous β-catenin staining. (Figure 1E and F) Another study by Tirbulo et al. suggested the WNT signaling pathway could be independent of the nuclear aggregation of β-catenin protein [11]. Kafri proposed that the rate of change in the nuclear accumulation of β-catenin, rather than the total quantity, correlated better with the Wnt pathway [12]. The variable pattern of β-catenin, while posing challenges in evaluation, holds unique value. For instance, cytoplasmic expression may indicate a worse prognosis in renal carcinoma [13]. However, for the majority of WNT pathway tumors, nuclear β-catenin staining almost invariably predicts CTNNB1 gene mutation. A definitive cut-off value remains elusive, and different conclusions exist. Fattet et al. suggested that extensive and nuclear β-catenin labeling is necessary for CTNNB1 gene mutation, while Wang et al. opined that even a minimal percentage of the β-catenin nuclear pattern should be considered significant [2, 14]. β-catenin exhibited diffuse and strong staining in almost all patients with DF, indicative of a relatively narrower spectrum of CTNNB1 mutation in DF.

β-catenin, acting as a cadherin-binding protein, plays a crucial role in cell adhesion. However, upon forming a complex with members of the T-cell factor/LEF family of proteins in the cell nucleus, it transforms into a transcriptional activation factor [15]. While it has been recognized that LEF1 tends to associate with β-catenin in the nucleus to exert its function, instances where LEF1 acts independently to activate the WNT pathway have been observed [16, 17].

In this study, a lower β-catenin expression ratio was observed in BA compared to LEF1. Similar to β-catenin, accentuated LEF1 staining was observed in myoepithelial cells surrounding the tumor nests, with clear nuclear expression. Patients categorized as negative for β-catenin displayed cytoplasmic or membranous staining in approximately 20–40% of tumor cells, with two patients showing no β-catenin immunoreactivity yet exhibiting approximately 30% LEF1 nuclear expression. Approximately 15% of adenoid cystic carcinomas and pleomorphic adenomas may display β-catenin nuclear staining [18, 19]. LEF1 is expressed in other salivary gland tumors, but its expression in BA is significantly higher [16, 20]. Our research demonstrated that all BAs harbored different ranges of LEF1 nuclear staining. So LEF1 has a distinct advantage in the diagnosis of BA.

The mutation in CTNNB1 (at codon 41 or 45 in exon 3) or APC can lead to the activation or dysregulation of the WNT pathway, which is the main pathogenic mechanism of DF [21]. The translocation of β-catenin protein from the cytoplasm to the nucleus due to CTNNB1 gene mutation makes the β-catenin nuclear pattern meaningful [22]. The relationship between DF, β-catenin nuclear staining, and CTNNB1 gene mutation is intricate, and we cannot simply assume that nuclear β-catenin positivity equals CTNNB1 gene mutation. Yamada et al. reported two patients without β-catenin nuclear staining that had CTNNB1 gene mutation with a distinct cytoplasmic “dotted” pattern [21]. β-catenin cytoplasmic staining was associated with DF, mainly related to the 41 A CTNNB1 mutation [22,23,24,25]. In this study, there was a fortunate coincidence that a patient with DF that occurred in the nasal cavity exhibited unique dotted ring staining. Koike et al. described that β-catenin nuclear-negative DFs harbored CTNNB1 mutation, while the mutation could not be identified in nuclear β-catenin expression cases [22]. A relevant explanation has been provided, indicating that S45 phosphorylation is the initial process of β-catenin degradation and that the S45F mutation may prevent the first step of phosphorylation. The subsequent degree of phosphorylation may vary depending on the specific codon of the mutation, leading to different staining patterns of β-catenin in different patients [26]. Additionally, Yamada et al. confirmed that different clones of β-catenin have diagnostic differences in DF, and clone β-catenin 14 had low specificity for the diagnosis of DF [21]. Lower expression of β-catenin is associated with the CTNNB1 mutation in codon 45 (45 F) and is more prone to recurrence [21]. Researchers proposed that high nuclear β-catenin staining was associated with higher 5-year survival [27]. Goto et al. reported that, using clone β-catenin 14 antibodies with a 10% cut-off value, patients with scar often exhibited β-catenin nuclear staining [28]. Other soft tissue tumors and scar tissues may harbor CTNNB1 mutations or exhibit nuclear β-catenin expression [27]. Ng et al. reported varying degrees of β-catenin nuclear staining in various tumors, such as solitary fibrous tumors, endometrial stromal sarcomas, synovial sarcomas, fibrosarcomas, and clear cell sarcomas [29]. Amary et al. reported that approximately 72% of fibromatous lesions exhibited nuclear β-catenin expression [30]. While using the marker β-catenin may seem challenging due to various interfering factors and low specificity, it possesses advantages. Previous studies showed that the sensitivity of LEF1 in the diagnosis of DF was lower than that of β-catenin, and the specificity of LEF1 was poor. Zou et al. revealed that 14 patients with scar showed LEF1 expression, but only one patient showed weak positivity for β-catenin [31]. However, only applying the marker β-catenin is not perfect; in this study, one patient with DF occurring in the nasal cavity, a relatively rare location, showed that the diagnostic performance of β-catenin was inferior to that of LEF1. The case exhibited a focal β-catenin dotted-like cytoplasmic pattern, while LEF1 showed strong nuclear staining. The diagnosis of spindle cell tumors in soft tissues is difficult, and no perfect markers are available. Despite the difficulties, the CTNNB1 gene mutation still provides high specificity for the diagnosis of DF, and the nuclear β-catenin pattern has suggestive value. However, the expression of the β-catenin protein is not stable, and for patients with β-catenin negative, the joint use of LEF1 can improve the diagnostic rate of DF.

Our study demonstrated β-catenin nuclear and cytoplasmic staining in SPN, consistent with previous reports, with a few patients showing pure nuclear expression [32, 33]. Certain patients exhibited membrane staining. However, LEF1 in SPN showed diffuse nuclear labeling, and unlike β-catenin, which harbored diffuse membranous staining in normal pancreatic tissues, LEF1 was negative. The positive rate of β-catenin and LEF1 in SPN were similar and were indiscriminate with known conclusions [34, 35]. Other pancreatic tumors, such as pancreatic neuroendocrine tumors, ductal adenocarcinoma, and pancreatoblastomas, may exhibit β-catenin staining; still, they mainly show focal nuclear/faint cytoplasmic staining or membranous labeling, and pancreatoblastomas present β-catenin staining mainly confined to squamoid corpuscles. Additionally, 25% of acinar cell carcinoma exhibit alteration of the β-catenin signaling pathway, resulting in diffuse β-catenin cytoplasmic/nuclear staining [36,37,38]. LEF1 expression has been described in other pancreatic tumors, but the conclusions are not consistent, possibly due to the limitations in sample capacity, antibody selection, and cut-off value. For example, Singhi et al. demonstrated that LEF1 is negative in acinar cell carcinomas and ductal adenocarcinomas, whereas McHugh et al. reported a weak-to-moderate LEF1 nuclear pattern in pancreatic acinar cell carcinomas and ductal adenocarcinomas [33, 34]. LEF1 expression in pancreatic neuroendocrine tumors is almost always negative; however, occasionally, patient with positive expression have been reported [33, 36]. LEF1 expression in pancreatoblastomas is mainly confined to squamoid corpuscles [32, 33, 36]. Therefore, conclusions from the related literature showed that LEF1 and β-catenin have slightly lower specificity in the diagnosis of SPN; however, they still have value due to their diffuse expression pattern differing from other pancreatic tumors. In our study, four patients identified as β-catenin negative expression in SPN showed diffuse and strong LEF1 nuclear staining. Thus, in comparison with β-catenin, LEF1 has higher sensitivity, harbors clean, clear staining, and is easy to evaluate the tumor border, making it a better marker for diagnosing SPN. However, LEF1 was positively stained in normal T and pro-B lymphocytes, which should be considered in needle biopsy specimens.

To date, there have been few studies on the expression of β-catenin and LEF1 in WNT pathway tumors. In this study, we shared limited 80 tumor cases, including 26 BAs, 30 DFs and 24 SPNs. Combined use of β-catenin and LEF1 significantly increased the diagnostic ratio in BA, DF, SPN and all 80 cases by 46.16%, 3.33%, 16.67% and 21.25%, respectively. Only using LEF1, there was no statistically significant difference (P = 0.14). However, the combined application of β-catenin and LEF1 in diagnosing the 80 WNT pathway tumors showed a distinct significance (P = 0.001). Our perspective is relatively novel, and the goal was not to correlate the expression of LEF1 and β-catenin proteins with CTNNB1 gene mutation. Although molecular technology is advancing, morphology and immunohistochemistry remain the basis and core of pathological diagnosis. We used a simple and economical method to screen for these three rare WNT pathway tumors, which can quickly exclude morphologically similar tumors and make an accurate diagnosis. This is particularly important for patients with limited financial capacities or hospitals that have not yet implemented the corresponding molecular technologies. Meanwhile, our findings may have positive implications for the diagnosis of other related WNT signaling pathway tumors. However, our study was limited by insufficient tumor types to demonstrate the diagnostic performance of the combined use of LEF1 and β-catenin in WNT-activated tumors. It is well-known that the smaller sample is accompanied by relatively large random errors, resulting in insufficient statistical significance. So we hope that more data could be available. Meanwhile, the diagnostic value of β-catenin and LEF1 in other WNT pathway tumors merits exploration. For instance, the CTNNB1 gene mutation can result in 90% WNT-activated medulloblastoma, which can induce alterations in the amino acid residues at the phosphorylation site of β-catenin, and LEF1 can form a transcriptional complex with β-catenin in the nucleus. Consequently, LEF1 is also an excellent auxiliary marker for diagnosing WNT-activated medulloblastoma. Multi-institutional collaboration, the increased number cases, and conducting CTNNB1 gene testing can render the conclusion more convincing.