Abstract
Although sequence-based studies show that basal-like features lead to worse prognosis and chemotherapy-resistance compared to the classical subtype in advanced pancreatic ductal adenocarcinoma (PDAC), a surrogate biomarker distinguishing between these subtypes in routine diagnostic practice remains to be identified. We aimed to evaluate the utility of immunohistochemistry (IHC) expression subtypes generated by unsupervised hierarchical clustering based on staining scores of four markers (CK5/6, p63, GATA6, HNF4a) applied to endoscopic ultrasound-guided fine needle aspiration biopsy (EUS-FNAB) materials. EUS-FNAB materials taken from 190 treatment-naïve advanced PDAC patients were analyzed, and three IHC patterns were established (Classical, Transitional, and Basal-like pattern). Basal-like pattern (high co-expression of CK5/6 and p63 with low expression of GATA6 and HNF4a) was significantly associated with squamous differentiation histology (p < 0.001) and demonstrated the worst overall survival among our cohort (p = 0.004). IHC expression subtype (Transitional, Basal vs Classical) was an independent poor prognosticator in multivariate analysis [HR 1.58 (95% CI 1.01–2.38), p = 0.047]. Furthermore, CK5/6 expression was an independent poor prognostic factor in histological glandular type PDAC [HR 2.82 (95% CI 1.31–6.08), p = 0.008]. Our results suggest that IHC expression patterns successfully predict molecular features indicative of the Basal-like subgroup in advanced PDAC. These results provide the basis for appropriate stratification for therapeutic selection and prognostic estimation of advanced PDAC in a simplified manner.
Similar content being viewed by others
Introduction
Pancreatic ductal adenocarcinoma (PDAC) is a fatal malignancy with an overall 5-year survival rate less than 10%1,2. Most patients (80–90%) have surgically unresectable advanced-stage carcinomas at the time of diagnosis1, and combination chemotherapy including gemcitabine or fluorouracil results in median survival of less than 1 year in patients with advanced PDAC3,4.
For therapeutic and prognostic stratification, several transcriptomic subtypes based on gene expression profiling using resected PDAC tissue samples have been proposed to date5,6,7, and it is suggested that PDAC is chiefly classified into two distinct subgroups termed classical/progenitor or basal-like/squamous (poorer prognosis)8.
Recent studies demonstrated that basal-like signature is a poor prognosticator even in advanced PDAC and associated with resistance to chemotherapy9,10 as well as its single cell level heterogeneity11 and/or plasticity during clonal evolution12. In routine diagnostic practice, however, the implementation of these approaches, which require next-generation sequencers and/or other high-technology devices13, can be challenging. In addition, given that PDAC cases are often discovered in inoperable stage1 and the technical challenges of endoscopic ultrasound-guided fine needle aspiration biopsy (EUS-FNAB)14, only limited biopsy samples are available. Considering these, immunohistochemistry (IHC)-based surrogate biomarkers emerge as highly useful tools for clinicians and pathologists. Numerous studies and reviews, including The Cancer Genome Atlas (TCGA) study, have identified potential surrogate markers5,8,10,12,15,16,17,18,19. Among these markers, we selected those supported by multiple studies and evaluated by the Human Protein Atlas (https://www.proteinatlas.org/). For example, GATA 6, which was identified as a candidate surrogate marker in TCGA study8, has been analyzed for its expression levels using RNA sequencing (RNA-seq), in situ hybridization (ISH), and IHC with a computer-assisted method10,13,15. These approaches have been introduced as valuable tools to differentiate between basal-like and classical PDAC using GATA6 expression as a potential surrogate biomarker. HNF4a and p63 are also mentioned as candidate surrogate markers in the International Cancer Genome Consortium (ICGC study)5 and further supported by mouse model and/or human IHC studies16,17,18,19. While these transcription factors are the potential key molecules that can distinguish expressional subtypes, recent IHC based human studies have shown the CK5/6 (KRT5 and KRT6A) is one of the promising surrogate markers for basal-like7 and/or squamous5 subtypes12,15. Considering the commercial availability of antibodies for IHC and the Human Protein Atlas evaluation, we believe that GATA6, HNF4, CK5/6, and p63 are the most suitable IHC surrogate biomarkers. In this study, we assessed the usefulness of combining these four IHC markers in evaluating treatment-naïve advanced PDAC using EUS-FNAB samples. Our aim was to determine whether these markers could serve as viable surrogate biomarkers for predicting molecular signatures. Additionally, we evaluated the publicly available expression and single-cell data to further validate their utility.
Materials and methods
Ethics approval and consent to participate
This study was conducted with approval from the Kyorin University Institutional Review Board (approval code: R03-162). All procedures involving human participants were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was secured through an opt-out process. All participant data were anonymized and de-identified to protect privacy.
ICGC data download and analysis
Clinical information and microarray data in ICGC study were downloaded from Supplementary Tables 1 and 14 from the original manuscript5. The downloaded microarray data were used to retrieve a normalized count matrix, which was then processed to obtain counts per million (c.p.m) and log2 transformed using RSEM. This count matrix was further Z-score normalized for calculating the sum as the gene signature scores and used to generate a heat map. Gene signatures corresponding to the Squamous/Progenitor (Bailey)5 and Basal-like/Classical subtypes (Moffitt)7 were used to stratify patients into groups characterized by high or low expression levels. Differential expression profiles of all transcription factors were quantified between groups. Gene expression data from Bailey et al. were reanalyzed and organized into hierarchical clusters based on established gene signatures5 and specific markers identified in this study (TP63, KRT5, KRT6A, HNF4A, and GATA6). Our custom 5 gene signature score is defined as sum (cpm.z(TP63, KRT5, KRT6A))—sum (cpm.z(GATA6, HNF4A)). Patient samples are assigned to high or low groups based on individual signature scores, and differences between groups are shown using log2FoldChange in expression and patient survival outcome. The FANTOM5 database was used to identify transcription factors.
Single cell data download and analysis
Single-cell RNA-seq data were downloaded from EGA under accession code EGAD00010001811. Analysis was done following methods described in the original manuscript11. Marker expression and gene signatures were displayed using Seurat’s AddModuleScore functionality. Similar results were reproduced for patient 100,070 and 85,948 individually.
Case selection and histological review
After approval by Kyorin University Institutional Review Board (R03-162), the consecutive PDAC cases diagnosed by FNAB samples taken from primary lesions were retrieved from the pathology archives of Kyorin University Hospital (Tokyo). Patients whose FNAB specimens contained sufficient tumor cells for histological diagnosis were included. Intraductal papillary mucinous neoplasm (IPMN)-derived carcinoma was excluded from the cohort. Clinicopathological information including age, sex, clinical stage, treatment and outcome was retrieved from the medical records.
The histology (hematoxylin and eosin slides) and IHC scoring were reviewed by two gastrointestinal pathologists (TS and AH) who were independently blinded to molecular and clinical data. Although the evaluations of most cases were concordant, discrepant diagnostic results were unified into one category through discussion. Because EUS-FNA is not routinely performed for cystic pancreatic lesions in Japan, fragmented atypical epithelium without a definite stromal invasion is diagnosed as a carcinoma by pathologists. Tumor histology was subclassified into glandular pattern, glandular pattern with poorly differentiated component (Por comp), or glandular pattern with squamous differentiation (Sq diff). Glandular pattern was defined as ductal formation and/or fragmented tumor cells with intracytoplasmic mucin including cytoplasmic foamy cell change. Squamous differentiation was evaluated as dense, opaque eosinophilic polygonal tumor cells regardless of whether keratinization (keratohyalin granules and/or pearls) and/or intercellular bridges were observed. Poorly differentiated component was defined as tumor cell cluster harboring neither glandular pattern nor squamous differentiation. If only a small amount of tumor cells with squamous differentiation was present, it was classified as Sq diff. The same criteria were applied to Por comp.
Immunohistochemistry
IHC was performed on 4-μm-thick formalin-fixed and paraffin-embedded (FFPE) whole tissue sections. The primary antibodies used were CK5/6 (D5/16 B4, 1:50; DAKO, Glostrup, Denmark), p63 (4A4, 1:100; Zeta Corporation, Arcadia, CA, USA), GATA6 (D61E4, 1:100; Cell Signaling Technology, Danvers, MA, USA) and HNF4a (H1415, 1:1000; Perseus Proteomics, Tokyo, Japan). Heat-induced epitope retrieval was performed and the Envision system (DAKO) was used for detection. Cytoplasmic and/or membranous staining was evaluated as positive for CK5/6, and nuclear staining was regarded as positive for p63, GATA6 and HNF4a. IHC slides were evaluated semiquantitatively based on the Allread scoring system which is clinically used in hormone receptor evaluation of breast cancer20. Staining intensity score (0: negative, 1: weak, 2: moderate, 3: strong) and proportional score based on percentages of stained cells (0: 0%, 1: < 1%, 2: 1–10%, 3: 10–33%, 4: 33–67%, 5: > 67%) were evaluated independently and total score (range 0–8) was calculated as sum of these scores. In this study we subclassified total scores into 4 grades as score 0 (total score 0–2), score 1+ (total score 3–5), score 2+ (total score 6, 7), score 3+ (total score 8).
Statistical analysis
All data analysis was performed using JMP (version 17.0.0) and/or R (version 4.2.2). The difference of categorical variables was evaluated using the Fisher’s exact test. Overall survival rate, measured from the date on which FNAB specimen was obtained, was calculated using Kaplan–Meier Method, and survival differences were compared using log-rank test. The Cox proportional hazards regression method was used for multivariate analysis. All tests were two-sided. Results with p value < 0.05 were considered statistically significant.
Ethics declarations
This study protocol was approved by the Faculty of Medicine Research Ethics Committee (R03-162), Kyorin University, with waiver of informed consent.
Results
Transcription factor enrichment and representative markers in PDAC subtypes
Gene signatures for Squamous/Progenitor subtypes (Bailey)5 or Basal-like/Classical (Moffitt)7 subtypes were used to assign patients into high or low groups. Differences in the expression profiles between the two groups were calculated for all transcription factors and representative keratin markers (KRT5 and KRT6A), as shown in Fig. 1a. TP63, KRT5, and KRT6A were enriched in the Squamous and Basal-like phenotypes, whereas HNF4A and GATA6 were enriched in the Progenitor and Classical phenotypes. Gene expression data from Bailey et al. were reanalyzed and hierarchically clustered based on published gene signatures5 and markers identified in this study (TP63, KRT5, KRT6A, HNF4A, and GATA6). Our five representative markers distinguished the Squamous/Progenitor (Bailey) and Basal-like/Classical (Moffitt) subtypes almost to the level of their original criteria, as shown in Fig. 1b. In addition, these markers also reflected the clinical course, as shown in Supplementary Fig. 1.
Single-cell level tumor heterogeneity based on subtypes and subtyping markers
Single-cell data were reanalyzed to demonstrate the distribution of marker genes and known gene signatures (Patient 100,070)11 as shown in Fig. 1c. Squamous/Basal-like type tumor cells and Progenitor/Classical type tumor cells co-existed within the same PDAC patient. TP63 and KRT5/6A positive cells were enriched in Squamous/Basal-like populations, while HNF4a and GATA6 positive cells were enriched in Progenitor/Classical populations (Fig. 1c,d and Supplementary Fig. 2).
Clinicopathological features of the cohort
A total of 191 treatment naïve cases were enrolled. A patient whose FNAB material was insufficient for additional IHC analysis was excluded. A total of 190 cases were analyzed. The clinicopathologic characteristics of this study cohort are shown in Supplementary Table 1. The patients included 96 males and 94 females with a median age of 71 years (range, 40 to 87 y). 144 patients (76%) of this cohort were diagnosed with advanced stage PDAC (stage III, IV). All patients were treated (surgical resection and/or chemotherapy) after FNAB performed from primary lesions. Median follow-up was 292 days (range, 7 to 42,901). Histologically, 134 patients (71%) showed Glandular pattern, whereas 40 patients (19%) and 16 patients (10%) met the criteria of poorly differentiated component and squamous differentiation, respectively (Fig. 2a).
Histological characteristics of IHC expression types
The results of IHC scoring for each of the four markers were summarized in Table 1, and the representative images were demonstrated in Fig. 2b. Unsupervised hierarchical clustering analysis was applied to IHC expression scoring data sets of each four markers, and classified the cohort into three distinct subgroups with clear separation, termed Classical pattern (N = 85), Transitional pattern (N = 85) and Basal-like pattern (N = 20), respectively (Fig. 3a). We evaluated five clustering methods (Ward, Hybrid Ward, Average, Centroid, and Complete) using JMP 17.0.1. (Supplementary Fig. 3). The dendrograms indicated that the Ward and Hybrid Ward methods effectively differentiated the classifications, with both methods yielding identical clustering outcomes. Because of this, the Ward method was adopted for further analysis. Classical pattern tended to show high expression (score 2+ or 3+) of HNF4a and GATA6 with low expression (score 0 or 1+) of CK5/6 and/or p63, whereas Basal-like pattern generally demonstrated the opposite trend (Fig. 3b). These IHC-based three clusters were significantly associated with histological categories (p < 0.001, Fig. 3b, Table 2), Classical and Basal-like pattern were related to histological Glandular pattern and squamous differentiation, respectively (Fig. 3b and Table 2).
Prediction of IHC expression status based on histology and IHC pattern
The results showed that “CK5/6 IHC score 3+” or “CK 5/6 IHC score 2+ and p63 IHC score 2+” or “CK5/6 IHC score 2+ and Sq diff histology” is efficiently classified as IHC basal-like pattern with sensitivity 95.0% and specificity 100% (Supplementary Table 2). Although “CK5/6 IHC score 2+ and Por comp histology” provides an equivocal result for classification, simultaneous GATA6 and HNF4a IHC results contribute to more accurate classification because “GATA6 IHC score 3+ and HNF4a IHC score 3+” predicts IHC classical pattern with sensitivity 59.8%, specificity 100% (Supplementary Table 3). On the contrary, “CK5/6 IHC score 0 or p63 IHC score 0” predicts “not IHC basal-like pattern” with sensitivity 92.9%, specificity 100% (Supplementary Table 4). Thus, the combination of histological evaluation and IHC scoring of four markers seems to provide an approximate prediction of IHC classification and the associated molecular signature.
Clinical characteristics of IHC expression types
IHC expression pattern was not related to clinical stage or sex and age (Table 2). Kaplan–Meier analysis demonstrated significant worst prognosis of Basal-like pattern patients amongst all three groups (Log-rank p = 0.004, Fig. 3c). Cox proportional hazards model showed IHC expression status (Transitional, Basal-like vs Classical) was a significant prognostic variable by both univariate analysis [p = 0.002, Hazard ratio (95% CI) 1.89 (1.27–2.83)] and multivariate analysis [p = 0.047, Hazard ratio (95% CI) 1.58 (1.01–2.38)]. On the other hand, histological status (with vs without squamous differentiation) was not an independent prognostic variable by multivariate analysis [p = 0.067, Hazard ratio (95% CI) 1.77 (0.85–3.26)] (Table 3).
CK5/6 positivity in histological Glandular pattern PDAC
Within the histological Glandular pattern cases (N = 134), CK5/6 positive PDAC (N = 14) (Fig. 4a) showed significantly poorer prognosis compared to CK5/6 negative PDAC (N = 120) (Log-rank p = 0.012, Fig. 4b). CK5/6 positive status (score 1+ , 2+, 3+ vs score 0) was an independent prognostic factor by both univariate analysis [p = 0.015, Hazard ratio (95% CI) 2.44 (1.19–5.00)] and multivariate analysis [p = 0.008, Hazard ratio (95% CI) 2.82 (1.31–6.08)] among histological Glandular population (Supplementary Table 5. Nevertheless, p63 positive status revealed no significant association with prognosis in this situation on both univariate and multivariate analysis [p = 0.714, Hazard ratio (95% CI) 1.13 (0.58–2.91)] (Supplementary Table 5).
Discussion
In this study, we showed the usefulness of the representative PDAC subtyping markers (TP63, KRT5, KRT6A, HNF4a and GATA6) and its clinicopathologic meanings with treatment naïve consecutive EUS-FNAB samples. A recent study showed that CK5 and/or GATA6, along with software analysis of digital images, are useful markers for expressional subtyping15. However, we believe that pathologists can perform subtyping based on their daily IHC evaluation if our combination is adopted.
Based on several prior data as to molecular mechanism for the acquisition of basal phenotype with loss of classical gene program in PDAC progression9,10,11,12,13,16,17,21,22 and gene expression profile from our re-evaluation of TCGA data sets, we selected four representative IHC candidates (GATA6, HNF4a, CK5/6, p63) for predicting molecular signatures. IHC expression subtypes (Classical, Transitional, Basal-like) based on IHC scores for these markers demonstrated a significant association with the histological subtypes and prognosis; IHC basal-like status is significantly related to histological squamous differentiation and poorer prognosis compared to IHC classical subtype. Since the reference transcriptomic subclassification data for each case was not established through RNA-seq in this study, it is controversial whether this IHC expression classifier is directly associated with molecular subtypes. Nevertheless, given that clinicopathological characteristics of IHC basal-like status overlap with those of molecular basal-like signature described in previous studies12,23, it suggests that IHC expression classifier is a potential surrogate biomarker for detection of molecular basal-like subgroup which is of clinical importance. Previous clinical and preclinical findings demonstrated that basal-like PDAC was less sensitive to standard chemotherapy, including mFOLFIRINOX or gemcitabine-based regimens, than the classical subtype10,11,24. Additionally, a multi-institutional randomized phase II trial to evaluate molecular subtypes as biomarkers for therapy response is in progress25, indicating that IHC subtyping has the potential to contribute to prognostic stratification and personalized treatment selection.
Although recent studies demonstrated that GATA6 expression by RNA-seq, ISH and IHC with digital assistance distinguished transcriptomic classical and basal-like subtype in advanced PDAC10,13, it is difficult to implement these technologies into clinical practice because it is costly, time-consuming and is limited to access only in high-volume cancer centers. In this regard, our methodology successfully overcomes these three limitations due to applying only four commercially available IHC markers to daily diagnostic practice. As previous studies have noted13, subjective evaluation of GATA6 staining intensity by pathologists is less accurate. The combination of GATA6 and three additional IHC markers, however, is expected to improve the interpretation discrepancy between pathologists for the reasons described in results.
Recent studies including single-cell resolutions have indicated complex spatial and temporal intratumor heterogeneity of PDAC11,12. Both classical and basal-like cells co-exist within a single PDAC patient in various proportions, and basal-like expression program is acquired during disease progression. Furthermore, it has been revealed that recurrent PDAC harbored an altered genomic landscape through treatment-induced genetic bottleneck26. As the present analysis focuses on biopsied materials taken from primary lesions of treatment-naïve patients, it is possible that basal-like cells hidden in primary lesions outside the biopsy site and/or metastatic lesions are underestimated, and it is difficult to evaluate basal-like subclones developed in recurrent lesions after treatment. Indeed, the fact that patients exhibiting focal CK5/6 expression despite glandular histology (not classified as IHC basal-like pattern) are associated with poor prognosis may indicate that intratumor heterogeneity with basal-like subclones stemming from classical clones is critical for PDAC outcome12. Further validation analyses using matched EUS-FNAB samples with gene expression profiles obtained from liquid biopsy specimens that provides spatial comprehensive genetic landscape are needed to evaluate how accurately this IHC expression classifier model reflects the impact of basal-like subclones behind the biopsy site. In addition, expanding the cohort to post-treatment patients is leveraged to estimate whether a similar model is applicable to PDAC post-neoadjuvant chemotherapy.
In addition to the challenges facing the IHC classifier model mentioned above, this study has several limitations. First, we excluded cases in which the FNAB material was insufficient for pathological diagnosis and additional IHC assessment. Whether this IHC classifier applies to FNAB samples with only a small number of atypical tumor cells remains unknown. Second, the treatment information for this cohort and possible factors affecting patient prognosis, including treatment type (surgery and/or chemotherapy) and chemotherapy regimens, were not considered in the multivariate analysis. Further validation studies on this cohort, including cases where FNAB samples have few atypical cells and/or treatment factor-adjusted analyses, are needed.
In conclusion, our study shows that IHC expression classification based on combined four IHC markers (CK5/6, p63, GATA6, HNF4a) is a versatile surrogate biomarker predicting transcriptomic subgroup in treatment-naïve advanced PDAC. Although further validation analyses are needed to be applied to personalized medicine for PDAC, our results propose the basis for the utility of a prognostic stratification model using histopathological and immunohistochemical evaluation of EUS-FNAB samples.
Data availability
The datasets used and/or analyzed during this study are available from the corresponding author on reasonable request.
References
WHO Classification of Tumours Editotial Board 5th ed. WHO classification of tumours of the digestive system. Vol. 1. IARC, Lyon 2019.
Jemal, A. et al. Annual report to the nation on the status of cancer, 1975–2014, featuring survival. J. Natl. Cancer Inst. 109, djx030 (2017).
Conroy, T. et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N. Engl. J. Med. 364, 1817–1825 (2011).
Von Hoff, D. D. et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N. Engl. J. Med. 369, 1691–1703 (2013).
Bailey, P. et al. Genomic analyses identify molecular subtypes of pancreatic cancer. Nature. 531, 47–52 (2016).
Collisson, E. A. et al. Subtypes of pancreatic ductal adenocarcinoma and their differing responses to therapy. Nat. Med. 17, 500–503 (2011).
Moffitt, R. A. et al. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat. Genet. 47, 1168–1178 (2015).
Cancer Genome Atlas Research Network. Integrated genomic characterization of pancreatic ductal adenocarcinoma. Cancer Cell. 32, 185–203 (2017).
Aung, K. L. et al. Genomics-driven precision medicine for advanced pancreatic cancer: Early results from the COMPASS trial. Clin. Cancer Res. 24, 1344–1354 (2018).
O’Kane, G. M. et al. GATA6 expression distinguishes classical and basal-like subtypes in advanced pancreatic cancer. Clin. Cancer Res. 26, 4901–4910 (2020).
Chan-Seng-Yue, M. et al. Transcription phenotypes of pancreatic cancer are driven by genomic events during tumor evolution. Nat. Genet. 52, 231–240 (2020).
Hayashi, A. et al. A unifying paradigm for transcriptional heterogeneity and squamous features in pancreatic ductal adenocarcinoma. Nat. Cancer. 1, 59–74 (2020).
Duan, K. et al. The value of GATA6 immunohistochemistry and computer-assisted diagnosis to predict clinical outcome in advanced pancreatic cancer. Sci. Rep. 11, 14951 (2021).
Kandel, P. & Wallace, M. B. Advanced EUS guided tissue acquisition methods for pancreatic cancer. Cancers (Basel) 10, 54 (2018).
Kokumai, T. et al. GATA6 and CK5 stratify the survival of patients with pancreatic cancer undergoing neoadjuvant chemotherapy. Mod. Pathol. 36, 100102 (2023).
Kloesch, B. et al. A GATA6-centred gene regulatory network involving HNFs and ΔNp63 controls plasticity and immune escape in pancreatic cancer. Gut. 71, 766–777 (2022).
Brunton, H. et al. HNF4A and GATA6 loss reveals therapeutically actionable subtypes in pancreatic cancer. Cell Rep. 31, 107625 (2020).
Somerville, T. D. D. et al. TP63-mediated enhancer reprogramming drives the squamous subtype of pancreatic ductal adenocarcinoma. Cell Rep. 25, 1741–1755 (2018).
Camolotto, S. A. et al. Reciprocal regulation of pancreatic ductal adenocarcinoma growth and molecular subtype by HNF4α and SIX1/4. Gut. 70, 900–914 (2021).
Allred, D. C., Harvey, J. M., Berardo, M. & Clark, G. M. Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod. Pathol. 11, 155–168 (1998).
de Andrés, M. P. et al. GATA4 and GATA6 loss-of-expression is associated with extinction of the classical programme and poor outcome in pancreatic ductal adenocarcinoma. Gut. 72, 535–548 (2023).
Martinelli, P. et al. GATA6 regulates EMT and tumour dissemination, and is a marker of response to adjuvant chemotherapy in pancreatic cancer. Gut. 66, 1665–1676 (2017).
Kalimuthu, S. N. et al. Morphological classification of pancreatic ductal adenocarcinoma that predicts molecular subtypes and correlates with clinical outcome. Gut. 69, 317–328 (2020).
Tiriac, H. et al. Organoid profiling identifies common responders to chemotherapy in pancreatic cancer. Cancer Discov. 8, 1112–1129 (2018).
Knox, J. J. et al. PASS-01: Pancreatic adenocarcinoma signature stratification for treatment-01. J. Clin. Oncol. 40, TPS635 (2022).
Sakamoto, H. et al. The evolutionary origins of recurrent pancreatic cancer. Cancer Discov. 10, 792–805 (2020).
Acknowledgements
We would like to thank Kazunari Tanabe, Ayumi Sumiishi, Namiko Kondo and Kaoruko Kojima for their superb technical assistance.
Funding
This work was supported in part by Grant-in-Aid for Scientific Research (22K06968) and Mochida Memorial Foundation for Medical and Pharmaceutical Research.
Author information
Authors and Affiliations
Contributions
Study concept and design: J.S. and A.H. Collection of the clinical samples and data: M.T., K.-Ki., T.Y., S.K., Ko.-K., K.O., K.G., N.O., Y.S. and T.H. Collection of the pathologic data: T.S., A.H., K.-Ki., K.-Ka. and J.S. Analysis of sequencing data: A.H. and Y.J.H. Analysis of clinicopathologic data and interpretation: T.S., K.-Ki. and A.H. Technical and material support: T.H. Expert consultation: J.M. and C.I.-D. Manuscript writing and review: T.S., A.H., Y.J.H., C.I.-D. and J.S. All authors read and approved the final paper.
Corresponding authors
Ethics declarations
Competing interests
CI-D receives research support from BMS. All other authors (TS, AH, MT, K-Ki, YJH, K-Ka, TY, SK, Ko-K, KO, KG, NO, JM, YS, TH, JS) declare that they have no conflicts of interest in relation to this work.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Shibayama, T., Hayashi, A., Toki, M. et al. Combination immunohistochemistry for CK5/6, p63, GATA6, and HNF4a predicts clinical outcome in treatment-naïve pancreatic ductal adenocarcinoma. Sci Rep 14, 15598 (2024). https://doi.org/10.1038/s41598-024-65900-w
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-024-65900-w
- Springer Nature Limited