Recent comprehensive genomic analyses have clearly demonstrated that various changes in the tumor genomic/epigenomic profile and microenvironment occur during successive tumor evolution, resulting in different clinicopathological features with intra-tumoral heterogeneity among primary, metastatic, and recurrent tumors (11–14). Changes that occur during tumor evolution might determine the potential for drug efficacy and resistance (11, 14). In various studies on chemotherapeutic treatment, however, unresectable locally advanced PCs with/without distant metastases have been lumped together with postoperative recurrent PCs as treatment targets (9, 15–18).
Various chemotherapeutic regimens have been developed and examined for efficacy in PCs, one of the most aggressive type of malignant tumors with high mortality rate. The MPACT trial demonstrated that GnP offered significant survival benefits for both primary pancreatic and metastatic lesions, making it one of the important treatments for patients with PC (9). This trial, however, included patients with both locally advanced PC with/without distant metastases and recurrent cancers. Considering the tumor genomic/epigenomic evolution as well as microenvironmental changes, locally advanced PC with metastases might exhibit different therapeutic efficacy and resistance compare to recurrent cancer. These finding prompted us to examine the predictors of therapeutic efficacy of GnP in patients with recurrent tumors.
Multivariate analysis conducted in the current study identified only histological differentiation as an independent predictor for the efficacy. Recent advancements in diagnostic modalities have allowed us to obtain pathological information in inoperable patients with PC. In fact, there have been many reports of preoperative histological diagnosis by EUS-FNA (19–21); however, precise diagnosis of histological differentiation is sometimes difficult given the small amount of specimens. Moreover, the necessity to include histological types in analysis during a diagnostic biopsy to tailor treatment remains unproved (11, 22, 23). Conversely, all patients who developed recurrent cancer were able to exhibit more accurate and comprehensive histological information, including tumor heterogeneity, from resected specimens. Considering these findings, pathological tumor differentiation can be used to determine the chemotherapeutic strategy for patients with recurrent PCs.
Studies on other gastrointestinal cancers have highlighted the relationship between histological types and prognosis and recurrence patterns (24–27). However, about GnP in patients with recurrent PC, it has not been sufficiently analyzed. We also examined whether recurrence patterns differ depending on the histological type (28), which might result in a difference in therapeutic effect. However, the current study found no difference in recurrence pattern according to histological differentiation.
Increasing evidence has suggested that the presence of an ongoing systemic inflammatory response (15, 29–37) and nutritional data (38, 39), ECOG performance status (15), and CA19-9 (15) can be used as predictors of poor outcomes in patients with cancer. However, such factors were not identified as predictors of therapeutic effect of GnP in our study. This might be connected with the early diagnosis of recurrence in our cohort, resulting in a relatively good condition.
Currently, the strongest regimens for patients with PCs include GnP and FOLFIRINOX (16–18), which have been recommended by international guidelines as first-line treatments for patients with metastatic PCs (40, 41). However, which treatment would perform better as first-line therapy remains controversial. Whether histological types or subtypes should be considered when tailoring treatment remains to be demonstrated (11). PC has been reported to have high heterogeneity within the primary and metastatic tumors, which could possibly be associated with drug resistance (11, 28, 42). In the future, circulating tumor cells and circulating tumor DNA may provide real-time and more accurate predictors for sensitivity and/or resistance to therapies (43–48).
Some limitations of the current study are to be represented. First, this was a retrospective, single-center study with a relatively small sample. Additionally, no histological diagnosis has been made for metastatic lesions. However, no reports have focused on predictors of the therapeutic efficacy of GnP in patients with recurrent PC after pancreatectomy. Accordingly, our finding showed that histological type can be useful for selecting anti-cancer agents in clinical practice.