P53 immunolabeling in EUS‐FNA biopsy can predict low resection rate and early recurrence in resectable or borderline resectable pancreatic cancer treated with neoadjuvant therapy

KRAS, P16, TP53, and SMAD4/DPC4 mutations are common in pancreatic ductal adenocarcinoma (PDAC). The study aimed to evaluate the association between gene mutations in pre‐treatment endoscopic ultrasound‐guided fine needle aspiration (EUS‐FNA) samples and clinical outcomes of patients with PDAC.


| INTRODUCTION
Pancreatic ductal adenocarcinoma (PDAC), a lethal malignant disease, is the fourth leading cause of cancer-related death worldwide. 1 Less than 8% of patients diagnosed with PDAC survive more than 5 years after their initial diagnosis. 2 Although surgical resection is a curative treatment option, the resection rate remains very low (15%-20%). 3 Even after surgical resection with curative pathological findings, an estimated 80%-85% of patients experience recurrent cancer, with a median survival of 20-24 months. 4 Thus, a multidisciplinary approach, which includes surgery, is required in the treatment of PDAC. Recently, neoadjuvant chemotherapy has become a standard procedure prior to pancreatectomy. 4,5 Additionally, numerous studies reported the promising effects of neoadjuvant chemotherapy (NAC) and chemoradiotherapy (NACRT) for resectable (R) and borderline resectable (BR) PDAC 6,7,8 NACRT and NAC for R or BR PDAC have been used at our institution since 2009, and we have already observed better outcomes, which have already been published. 9 However, neoadjuvant therapy is not universally employed in all patients with PDAC. In fact, a certain percentage of patients suffered from rapid tumor spread or early postoperative recurrence despite NACRT or NAC. 6,10 Thus, precision medicine is needed to further ameliorate the treatment outcomes of patients with R and BR PDAC.
Whole-exome sequencing of tissue samples of PDAC was conducted, which revealed that there were only four frequently mutated genes in this carcinoma, namely, KRAS,iCDKN2A/p16, TP53, and DMAD4/DPC4. 11 Several studies described that immunohistochemical (IHC) labeling of p16, p53, and Smad4/Dpc4 reflected the genetic statuses of CDKN2NA/p16, 12 TP53, 13,14 and SMAD4/ DPC4. 15,16 Our previous study 17 was done to analyze the relationship between the clinicopathological characteristics and CDKN2A/p16, TP53, and DMAD4/DPC4 mutations because the KRAS gene was mutated in virtually all PDAC patients. The mutation statuses of CDKN2A/p16, TP53, and SMAD4/DPC4 were determined by means of IHC using paraffin-embedded samples of excised primary carcinomas from 106 patients who underwent pancreatectomy. The study concluded that genetic alterations in these three genes and their accumulation were strongly associated with the malignant behavior of PDAC. However, our previous study was limited clinically as it did not explore resected samples after pancreatectomy.
Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) has emerged as an effective and safe diagnostic tool for PDAC. 18,19 Importantly, specimens can be obtained with EUS-FNA to examine their genetic information at the time of the first diagnosis. Therefore, precision medicine for PDAC patients might be possible based on the IHC statuses of CDKN2A/p16, TP53, and SMAD4/DPC4 in EUS-FNA specimens before treatment. The aim of this study was to evaluate the association between the three major genetic mutations (P16, TP53, and SMAD4/DPC4) associated with PDAC and their malignant behavior. The genetic mutations were examined with EUS-FNA, while the malignant behavior of the samples was evaluated through the pathological findings and prognosis of patients with PDAC who underwent NACRT or NAC.

| Patient selection and characteristics
The patient selection flowchart is shown in Figure 1. PDAC was pathologically diagnosed in a total of 227 patients. Tissue samples were obtained via EUS-FNA during the study period (July 2014 to July 2019). A total of 14 patients were excluded due to the absence or small size of the tissue sample, which could not be evaluated with IHC. A total of 213 patients (93.8%) were classified into R or BR (n = 102) or unresectable PDAC (n = 111) according to the resectability criteria of National Comprehensive Cancer Network Guidelines version 1.2021. IHC labeling with p16, p53, and Smad4/Dpc4 of EUS-FNA samples was assessed in all 213 patients. Another set of 11 patients was excluded because they terminated their therapy for PDAC due to their relatively old age, poor performance statuses, or past medical history. Moreover, seven patients who underwent upfront surgery without NACRT or NAC were excluded from this study. Finally, 84 patients with R or BR PDAC were included in the present study. This study F I G U R E 1 Patient selection and treatment. R, resectable; BR, borderline resectable; PDAC, pancreatic ductal adenocarcinoma; NACRT, neoadjuvant chemoradiotherapy; NAC, neoadjuvant chemotherapy was approved by the Kagawa University Review Board (NoH25-095).

| Tissue sampling
In EUS-FNA, PDAC tissue sampling was performed two to three times with a 22-or 25-gauge FNA needle in all enrolled patients. In patients who underwent pancreatectomy, the gene statuses of the extracted specimens were also evaluated via IHC.

| Immunohistochemistry
Paraffin-embedded samples of PDAC, which were obtained with EUS-FNA and/or pancreatectomy were immunostained for p16, p53, and Smad4/Dpc4. EUS-FNA samples were immunostained immediately as routine practice during the study period. At least three different formalin-fixed paraffin-embedded (FFPE) slides of the extracted specimens that were obtained via pancreatectomy were immunostained for each of the three antibodies (p16, p53, and Smad4/Dpc4) to evaluate the heterogeneity of the PDAC samples. Heterogeneity was evaluated in the whole single EUS-FNA specimen, yielding maximum tumor cells in each case with IHC. Tissue specimens were fixed in 10% buffered formaldehyde, routinely processed for embedding in paraffin, and sectioned at 4 μm. Immunohistochemical labeling was carried out using a BenchMark ULTRA automated immunohistochemical slide staining system (Ventana Medical System, Inc.) as previously described. 17,20 Anti-human p16 monoclonal antibody (CINtec E6H4 p16 ink4a, ready-to-use), anti-human p53 monoclonal antibody (Roche DO-7102364, ready-to-use), and antihuman Smad4/Dpc4 monoclonal antibody (SANTA CRUZ B-8: sc-7966, diluted 1:100) were used for immunostaining. The relationship between IHC statuses and mutation was evaluated according to previous studies. 17 Briefly, islet cells were the internal controls for positive p16 immunolabeling. p16-labeled specimens were regarded as positive, which indicated that the gene was present. Contrastingly, non-labeled specimens were considered negative, which suggested that the gene might have deletion or inactivating mutation (Figure 2a,b). As for p53 immunolabeling, scattered acinar and ductal cells with nuclear labeling were typically present in the adjacent normal tissue (Figure 2c). p53 immunolabeling was considered abnormal when the carcinoma cells showed a virtual absence of immunolabeling compared with adjacent normal tissue (immunolabeling in <5% of carcinoma cells), which suggested the presence of an intragenic deletion such as nonsense or frameshift mutation (Figure 2d-1), or showed robust nuclear accumulation of immunolabeled protein in ≥30% of carcinoma cells compared with adjacent normal cells (Figure 2d-2). In the evaluation of Smad4/Dpc4, normal acinar cells, ductal cells, islet cells, and stromal cells in each case served as internal controls for positive immunolabeling. Immunohistochemical labeling of Smad4/Dpc4 was scored as intact (positive), which indicated the presence of an intact gene, or as lost (negative), which indicated a deletion or an inactivating mutation of the gene (Figure 2e,f). The statuses of immunolabeling were evaluated individually by two pathologists.

| Neoadjuvant chemotherapy and neoadjuvant chemoradiotherapy
All 84 patients with R or BR PDAC received NACRT or NAC. NACRT was indicated for R or BR with venous invasion (BR-V) PDAC according to our two clinical trials during the study period. 9,21 In this study, hypofractionated external-beam radiation (30 Gy in 10 fractions) with concurrent S-1 (60 mg/m 2 /day) was administered 5 days per week for 2 weeks in 27 patients. Subsequently, the standard external-beam radiation (50 Gy in 25 fractions) with concurrent S-1 (60 mg/m 2 /day) was administered 5 days per week for 5 weeks in 44 patients since April 2016. In contrast, 13 patients with arterial invasion (BR-A) PDAC underwent NAC. The NAC regimen included a combination of gemcitabine and nanoparticle albumin-bound paclitaxel (nab-PTX) or a combination of fluorouracil, leucovorin, irinotecan, and oxaliplatin (FOLFIRINOX) for 2 or 3 months as standard NAC. Additionally, two patients with BR-A underwent chemoradiotherapy after NAC.

| Surgery
Pancreaticoduodenectomy (PD) or total pancreatectomy (TP) with regional lymphadenectomy was carried out according to the Whipple procedure. Reconstruction after PD was performed using the modified Child's method. Distal pancreatectomy (DP) was done using the radical antegrade modular pancreatosplenectomy procedures with regional lymphadenectomy. Although the reconstruction procedure after PD was changed from the Kakita method to the modified Blumgart method in 2016, the procedure for resection and lymph node dissection was not changed during the present study period.

| Data collection
Data were retrospectively collected from the patients' medical records.

| Statistical analyses
Frequency distributions were compared using the Mann-Whitney U-test. The survival curve was estimated using the Kaplan-Meier method, and differences were tested using the log-rank test. The Cox proportional-hazards models were generated for the univariate analysis. Overall survival   Table 2 summarizes the data of the relationship between genetic statuses (p16, p53, or Smad4/Dpc4) and clinicopathologic factors in R or BR PDAC (n = 84) patients. The genetic statuses of the three major genes were evaluated with EUS-FNA. An abnormal p53 was significantly associated with a lower resection rate compared with normal p53 labeling (21.3% vs. 0%, p = .0167). All 13 patients with an unresectable tumor had abnormal p53 labels.

| Patient outcomes in resected cases
The outcomes of the patients who underwent resection are shown in  Table 4 and Table S1 summarize the data of the relationships between p16, p53, or Smad4/Dpc4 and pathological and oncological variables in resected R or BR cases (n = 70). The statuses of p16, p54, and Smad4/Dpc4 were evaluated with EUS-FNA and surgically excised specimens. Pathological tumor regression by NAC or NACRT was not statistically associated with IHC labeling with p16, p53, and Smad4/Dpc4 (p = .16, p = .26, p = .72, respectively, Table 4). This was determined by analyzing the EUS-FNA specimens. Only p16 loss was significantly associated with a higher local response to NAC or NACRT in the analysis of resected specimens (p = .05, Table S1). Contrastingly, early recurrence within 6 months after pancreatectomy was more frequently seen in patients with abnormal p53 and Smad4 loss compared with those with normal p53 and intact Smad4/Dpc4 (25.5% vs. 4.3%, p = .03; 30.0% vs. 10.0%, p = .03, respectively). Other pathological and survival parameters were not significantly associated with IHC labeling of p16, p53, and Smad4/Dpc4 in EUS-FNA and resected samples. Table 5 shows the results of the univariate analysis using the Cox proportional-hazards models of the clinicopathological variables and genomic alterations in relation to OS in resected samples from patients who underwent pancreatectomy (n = 70). A high accumulation of fluorodeoxyglucose-positron emission tomography (FDG-PET) (SUV max ≥ 10.6), differentiation (moderate or poor), portal vein invasion, and arterial invasion were associated with poor prognosis (p = .003, p = .01, p = .003, and p = .002, respectively). Among genetic mutations, an abnormal p53 expression was associated with a shorter OS with a marginal significance (p = .07). In addition, the combination of p16 loss and abnormal p53, the  .50

| Association between genetic mutations and pathological/oncological variables in the analyses of EUS-FNA and surgically excised specimens
Note: Data are presented as n

EUS-FNA and resected samples
The statuses and changes in IHC labeling (p16, p53, Smad4/Dpc4) of EUS-FNA and resected samples are shown in Table S2. A total of 60 (85.7%) cases showed concordance in the p16 IHC labeling status between EUS-FNA and resected samples. There were 46 (65.7%) and 50 (71.4%) cases that showed concordance in the p53 and Smad4/Dpc4 IHC labeling statuses, respectively, between EUS-FNA and resected samples. Table S3 shows the relationship between IHC status (p16, p53, and Smad4/ Dpc4) of EUS-FNA biopsy and resected samples and the pathological treatment effect of neoadjuvant therapy. The discordance of p53 IHC status was associated with a significantly higher therapeutic effect compared with the concordance of p53 (p = .04). The discordance of p16 and Smad4/Dpc4 IHC status was not associated with the treatment effect of neoadjuvant therapy (p = .26 and p = .94, respectively).

| DISCUSSION
IHC-detected expression of the three major genes, particularly TP53, in EUS-FNA biopsy specimens before treatment was associated with the malignant behavior of R or BR PDAC in patients who received neoadjuvant therapy followed by pancreatectomy. Whole-exome sequencing of 24 PDAC samples was performed in 2008. 11 Identification of protein-determining exons of 20 661 genes led to the finding that PDACs had an average of 63 genomic alterations, the majority of which were point mutations. Moreover, the genetic landscape of PDAC had only four frequently mutated genes, which were KRAS, CDKN2A/p16, TP53, and SMAD4/DPC4. These four genes play an important role in pancreatic carcinogenesis.
Our previous study analyzed the relationship between clinicopathological findings and mutations of CDKN2A/p16, TP53, and DMAD4/DPC4 among the four genes because the KRAS gene was mutated in virtually all PDAC patients. The statuses of CDKN2A/p16, TP53, and SMAD4/DPC4 were determined via IHC using paraffin-embedded primary carcinoma samples from 106 patients who underwent pancreatectomy. The study concluded that genetic alterations of these three genes and their accumulation were strongly associated with the malignant behavior of PDAC. Other studies reported that genetic mutations were related to the biological malignant potential and behavior of PDAC. 22,23 However, several previous studies, including our previous study, explored only PDAC specimens during pancreatectomy or autopsy. Therefore, an assessment of the malignant potential of individual PDAC samples was only possible after surgery.
NAC is often necessary and effective. 3,5 Advanced research in adjuvant and neoadjuvant therapies, including chemotherapy, radiotherapy, and chemoradiotherapy, has focused on its therapeutic effects in PDAC. 6,7,24 We also performed a prospective phase II study of neoadjuvant S-1 with concurrent hypofractionated radiotherapy and reported its tolerability and efficacy in patients with R and BR PDAC. 9 Recently, a randomized phase II/III trial that examined the effect of NAC with gemcitabine and S-1 (Prep-02/JSAP05) (clinical trial information: UMIN000009634) reported an improvement in the treatment outcomes and prognosis of patients with R-PDAC compared with those who underwent upfront surgery. 8 Based on the efficacy of NAC in the planned resection of patients with PDAC, as determined by clinical trials, neoadjuvant therapy has become one of the standard treatments for R and BR PDAC. Therefore, evaluation of the malignant behavior of the tumor by examining EUS-FNA biopsy specimens before treatment can enable precision medicine and facilitate adequate treatment planning. EUS-FNA has emerged as an effective and safe diagnostic technique for PDAC. 18,19 Recent studies, including meta-analyses, reported that EUS-FNA was highly accurate in the pathological diagnosis of solid pancreatic lesions (sensitivity, 89%; specificity, 96%-100%). 25,26 Furthermore, EUS-FNA enabled the procurement of tumor specimens at the time of the first diagnosis before treatment. Sweeney et al. 27 reported that IHC staining with S100p, Smad4, and Imp3 on cell blocks procured from a pancreatic EUS-FNA procedure was sensitive, highly specific, and useful for the diagnosis of PDAC. In addition, several papers have described that EUS-FNA was useful for obtaining adequate samples for genomic analysis, especially for the next-generation sequencing (NGS). 28,29 Gleeson et al. 28 reported a multigene mutation concordance evaluation between EUS-FNA and surgical pathology specimens in the same case. Of the paired samples, 83.3% had an absolute multigene mutational concordance in the NGS. 28 However, several papers described that 60%-70% of EUS-FNA samples were inadequate for the NGS for the diagnosis of PDAC. 30,31 Furthermore, NGS of tissue samples is generally not easy to perform in routine clinical practice in many hospitals. On the other hand, immunostaining is a standard procedure in many clinical facilities with a pathological division. Therefore, genetic evaluation by IHC will remain an important practice for the time being. In this study, we assessed the genetic alterations in PDAC specimens that were obtained via EUS-FNA and IHC labeling. There were only a few reports that described the predictive value of IHC labeling of EUS-FNA specimens. The profile and incidence of each genetic alteration were similar to those of our previous study, in which we used surgically resected materials. 17 The present study demonstrated that an abnormal labeling of p53, as assessed with EUS-FNA, was associated with a lower resection rate and an early recurrence in R or BR PDAC cases treated with neoadjuvant therapy ( Table S3). All of the unresected cases (n = 13) with a noncurative factor detected before or during surgery showed an abnormal p53 labeling. In addition, an abnormal labeling of p53 and Smad4/Dpc4 in EUS-FNA samples was associated with early recurrence within 6 months after pancreatectomy. Therefore, in patients with abnormal labeling of p53 in EUS-FNA specimens, occult distant metastasis or dissemination should be considered even if the progression stage of PDAC was diagnosed as R or BR at first diagnosis based on imaging examinations.
Mutations in the TP53 tumor suppressor gene were the most common genetic lesions in cancer. Weissmueller 32 reported that a mutant TP53 induced PDAC metastasis through cell-autonomous platelet-derived growth factor receptor beta signaling. 32 Additionally, a potential association between mutant TP53 in PDAC and resistance to NACRT or adjuvant chemotherapy with S1 could also exist. Patients with abnormal p53 labeling in EUS-FNA specimens might be candidates for more enhanced systemic chemotherapy before surgery. The combination of loss of p16 and abnormal labeling of p53, the combination of loss of Smad4/Dpc4, and loss of p16 in EUS-FNA specimens were associated with a poor prognosis, with significant differences in resected cases (p = .04, p = .03, respectively). These results affirmed our previous study that the accumulation of three genetic alterations was strongly associated with the malignant behavior of PDAC. Additionally, abnormal labeling of p53 in EUS-FNA samples was associated with a poor prognosis, with a marginally significant difference in resected cases (p = .07). Contrastingly, there were no statistical associations between individual and combinations of three major immunohistochemical labeling, p16, p53, and Smad4/Dpc4, in terms of OS in R or BR PDAC resected samples. These IHC evaluations of resected samples differed from those observed in our previous study (Table S4). 17 Abnormal immunolabeling of p53 was detected in 81.1% of participants. Loss of p16 and Samd4/Dpc4 immunolabeling was identified in 67.0% and 60.4% of participants, respectively. Three major mutations in the resected samples were regarded as predictive biomarkers of PDAC in the patients undergoing surgery first, meaning without neoadjuvant therapy, in our previous study. Therefore, these discrepancies in the evaluations of IHC in the resected samples were thought to be affected by the differences in the treatment strategy for R or BR PDAC in the present and previous studies. Additionally, some clinical characteristics that were recognized as poor prognostic factors of PDAC, such as lymph node metastasis, surgical margin, and completion of adjuvant chemotherapy, did not show significant differences in OS. These results of the statistical analysis seemed to be influenced by neoadjuvant therapy or by the small number of cases.
The IHC evaluations of p16, p53, and Smad4/Dpc4 in EUS-FNA samples were valuable compared with those in resected samples from patients who received NAC or NACRT followed by pancreatectomy.
To the best of our knowledge, the present study is the first to compare the genetic alteration profiles between EUS-FNA samples and resected samples in patients who received NAC or NACRT. For p16, IHC results were matched in 85.7% of the cases. For p53 and Smad4/Dpc4, IHC results were concordant in 65.7% and 71.4% of the cases, respectively. The discrepancy in the genetic alteration status between EUS-FNA samples and resected samples could have been due to NAC or NACRT and/or might be influenced by intratumoral heterogeneity. Changes of IHC status between EUS-FNA samples and resected samples, (pre and post NAT) were presumed to be influenced by NAT. Moreover, the discordance of p53 IHC status was associated with a significantly higher therapeutic effect compared with the concordance of p53 (p = .04).
Additional investigations related to this discordance and novel relationship between changes of p53 IHC status and treatment effect of NAT are needed. Intratumoral heterogeneity is a fundamental problem in cancer research including PDAC. Our previous report described that intratumoral heterogeneity of p16, p53, and Smad4/Dpc4 was observed to be 0%, 1.0%, and 5.7%, respectively, in the resected specimens. 17 The intratumoral heterogeneity was evaluated again in the present study. The IHC status of whole single EUS-FNA specimens yielding maximum tumor cells was almost homogeneous in each case.
Although it can be difficult to fully assess the heterogeneity of the entire tumor by EUS-FNA, our results of concordance between EUS-FNA samples and resected samples are also novel findings and are extremely important for future studies and clinical applications.
One aim of the present study was to evaluate whether the genetic alteration profile in PDAC, which was assessed using EUS-FNA samples, could predict the therapeutic effect of neoadjuvant therapy. A precise prediction of the effects of NAC or NACRT could facilitate the advancement of precision medicine in PDAC. However, the present study did not indicate a definite therapeutic strategy for precision medicine.
Several limitations of this study should be mentioned. First, the number of patients included in the study may not be sufficient to draw firm conclusions based on statistical analysis. Additionally, the multivariate analysis did not show any statistical differences in this study. The insufficient number of patients in the present study could have affected the results. However, we hope that the results of this study will be verified using multivariate analysis in future large-scale studies. Second, NACRT and NAC protocols for R or BR-PDAC could not be integrated in this retrospective study. The influence of NAT on the clinical outcomes of the present study should be considered. We hope that the proposed consequences of this study will be verified by future prospective large-scale research with integration of the NAT protocol.
In conclusion, IHC-detected expression of the three major genes (P16, TP53, and SMAD4/DPC4) in EUS-FNA biopsy specimens before treatment was associated with the malignant behavior of R or BR PDAC compared with the IHC in resected specimens at pancreatectomy in patients who received neoadjuvant therapy for R or BR PDAC. In particular, abnormal labeling of p53 was associated with a lower resection rate and an early recurrence in R or BR PDAC cases treated with neoadjuvant therapy following pancreatectomy.