We established seven PDCO lines from nine samples obtained from four patients, despite variable cancer types and pathological results, using the same culture medium containing growth factors and Wnt signaling pathway activators. The PDCOs made from such small amounts of samples were subjected to H&E and IHC staining, scRNA-seq, and DNA cancer panel analysis, showing the characteristics of carcinoma, regardless of the pathology results.
In this study, a notable achievement was the successful generation of organoids using small tissue samples obtained through ERCP and EUS-FNA/B. A recent meta-analysis also reported the comparability of patient-derived tumor organoids between EUS-guided biopsies and surgical specimens [12]. Compared to the meta-analysis that reported establishment success rates of 60%, 36%, and 62% from EUS-guided biopsies, percutaneous biopsies, and surgical specimens, respectively, our study reported that the overall establishment success rates were 77.8% and 75.0% from EUS-FNA/B and ERCP biopsies (except for one surgical sample). The high success rates of biopsies indicate that endoscopic tissue acquisition is a comparable method for the establishment of PDCOs with surgical samples, despite the inherent challenge posed by limited sample amounts.
We introduced two criteria to determine the success of PDCO establishment. One was the "expansion capability," which assessed whether the organoids could be continuously cultured while maintaining viability up to passage 5. Despite the high initial success rate in primary cultures, the ability to maintain viability during long-term culture was considered separately. When the criterion ‘capable of culturing for more than 5 passages’ is applied, the success rate of establishing organoids decreases [22]. The other criterion was the “thawing test,” which assessed whether the organoids-maintained viability when re-cultured after freezing. These two criteria are crucial for determining whether established organoids can be continuously utilized for genetic analysis, drug sensitivity testing, and other purposes, thereby determining the actual value of the PDCO establishment. This study generated and analyzed PDCOs from a limited number of patients, therefore, the correlation between the pathological results of the biopsy samples and organoid establishment remains to be determined. We were able to establish organoids from samples judged as pathologically negative; although all pathologically negative samples were successful in the initial stage of primary culture, some organoids encountered issues with expansion or viability after thawing. Further attempts to establish PDCOs from more patient-derived negative pathological samples are required to clarify this.
In this study, PDCOs without a pathological diagnosis of cancer showed the same results as those with a pathological diagnosis using H&E and IHC staining (Fig. 1 and Fig. 2), DNA mutation analysis (Fig. 3), and scRNA-seq (Fig. 4). Our results are similar with previous studies showing that PDCOs can capture the characteristics of the original tumor and serve as a tool for personalized medicine [23, 24]. Notably, the diagnosis of one patient changed from pancreatic cancer to GBC. The radiologic diagnosis was initially synchronous pancreatic cancer and GBC; however, the surgical specimen revealed that the GBC invaded the surrounding lymph nodes and focal pancreas. The cell type of this patient was intracholecystic tubulopapillary neoplasm with an associated adenocarcinoma, which is under the heading of “papillary adenocarcinoma” [25]. Notably, this study suggested that PDCO cells derived from a patient with GBC were enriched in the C0-cluster and could be distinguished from the malignant ductal cells of PDCOs derived from other patients with pancreatic cancer (Fig. 5G). We performed cluster-specific marker discovery analysis, which is fundamental for scRNA-seq data analysis of the C0-cluster, and found that three genes, VCAN, AQP3, and FGF19, were significantly upregulated in the C0-cluster (Fig. 5Hh– 5J). The VCAN transcript, translated into the CSPG2 protein, has been reported as a marker for the metastasis of various carcinomas, including bladder carcinoma [26, 27]. In addition, aquaporin 3 (AQP3) is an important regulator of the inflammatory response and a marker that can identify the effects of gallbladder damage [28, 29]. Furthermore, FGF19 can promote the progression of GBC [30]. Considering these results, the detection of specific expression patterns of cells in PDCO derived from patients with GBC suggests the possibility that it can be employed for diagnosing GBC and/or predicting prognosis, where accurate diagnosis is practically difficult.
Our study has several limitations. First, the inclusion of GBC samples, originally thought to be pancreatic cancer, led to new findings; however, the sample size was too small for generalization. Second, the acquisition methods varied and included ERCP-guided forceps biopsy, EUS-FNA/B, and surgery. However, we did not find a different finding based on the sample acquisition method used in this study. We believe that these practical limitations can be naturally resolved through larger-sample experiments in the near future.
In conclusion, our study highlights the potential of PDCOs as valuable diagnostic and research tools in oncology, particularly in scenarios in which only small tissue samples are available. The consistency of the results obtained from PDCOs, regardless of the underlying pathology, holds promise for advancing our understanding of cancer biology and improving patient care through personalized treatment approaches.