Essential updates 2018/2019: Current topics in the surgical treatment of pancreatic ductal adenocarcinoma

Abstract Pancreatic ductal adenocarcinoma (PDAC) is highly malignant. While cancers in other organs have shown clear improvements in 5‐year survival, the 5‐year survival rate of pancreatic cancer is approximately 10%. Early relapse and metastasis are not uncommon, making it difficult to achieve an acceptable prognosis even after complete surgical resection of the pancreas. Studies have been performed on various treatments to improve the prognosis of PDAC, and multidisciplinary approaches including non‐surgical treatments have led to gradual improvement. In the present literature review, we have described the significance of anatomical and biological resectability criteria, the concept of R0 resection in surgical treatment, the feasibility of minimally invasive surgery, the remarkable development of perioperative chemotherapy, the effectiveness of conversion surgery for unresectable PDAC, and ongoing challenges in PDAC treatment. We also provide an essential update on these subjects by focusing on recent trends and topics.

the progress of the cancer and responsiveness to treatment; thus, the presence of CTCs may be a potential criterion for resectability. 45

| Other biomarkers
With respect to other biomarkers, the circulating tumor DNA, 46,47 exosome, 48 and microRNA 49 levels have been reported as candidate factors for assessing the biological resectability of PDAC.
Nevertheless, they have not been established as resectability criteria to date.

| Local radiality and surgical margins
A positive surgical margin in PDAC resection is a strong indicator of poor prognosis, and the distance from the surgical margin to the tumor affects the achievement of complete resection. The prognosis after R0 resection is reported to improve gradually as the distance from the surgical margin gradually increases. [50][51][52] Therefore, the very definition of the surgical margin is changing. As per the Royal College of Pathologists (RCPath) 53

and the American Joint
Committee on Cancer (AJCC) 54 guidelines, a distance of at least 1 mm or more between the cancer cell and the resection surface is defined as R0 resection and that of 0-1 mm is defined as R1 resection; in the Union for International Cancer Control (UICC) 55 and JPS 56 guidelines, a different definition of R1 resection is adopted where the distance between the cancer cell and the resection surface is 0 mm.

| Rules for the margin distance
There is a marked difference (Table 1) in the R0 resection rate and prognosis noted between cases where resection was performed using the 0-mm rule and those where it was performed using the 1-mm rule. 51,52,[57][58][59][60][61][62][63][64][65] Overall, the R0 resection rate is lower for cases where resection was performed using the 1-mm rule than for those using the 0-mm rule. In contrast, the median survival time (MST) after R0 resection was prolonged in cases where resection was performed using the 1-mm rule. Systematic reviews 50 and meta-analyses 62 have also reported that the adoption of the 1-mm rule both reduced the R0 resection rate and prolonged the overall survival after R0 resection. The optimum cut-off margin for improving disease prognosis is reported to be ≥ 1.5 mm 50 and ≥ 2.0 mm. 59,66 It is, therefore, necessary to specify the margin rule applied when reporting the outcomes of PDAC treatment.  the major blood vessels. An analysis of data from the National Cancer Database (NCDB) also discovered improved R0 resection rates after NAT. 67,68 In addition, a meta-analysis reported that NAT for R/BR-PDAC resulted in a significant margin-negative resection and overall survival prolongation. However, margin-positive resection after NAT is associated with a poor prognosis. It is necessary to maintain an adequate and safe surgical margin even after NAT. 67

| Laparoscopic distal pancreatectomy
The operative time for laparoscopic distal pancreatectomy (LDP) is longer than that of laparotomy ( Table 2); however, LDP is also associated with significantly less blood loss, fewer complications, and shorter duration of hospital stay. [72][73][74]79 An increasing number of studies have reported on the oncological safety and long-term prognosis of LDP for PDAC. Table 2 summarizes the previously reported oncologic factors and disease prognosis associated with LDP and open distal pancreatectomy (ODP) for PDAC. [80][81][82][83][84][85][86][87][88][89][90][91][92][93][94] Propensity score matching (PSM) analysis using data from the NCDB indicated that the R0 resection rate, number of retrieved lymph nodes, and long-term prognosis were equivalent for LDP and ODP. 95 In contrast, the PSM analysis in the DIPLOMA study noted a significant difference in the R0 resection rate, postoperative chemotherapy induction rate, and MST, although the number of retrieved lymph nodes was significantly smaller with LDP. 94 The concerning issue is that both studies reported a high conversion rate of 20%-30%. In a recent meta-analysis, the R0 resection rate, postoperative chemotherapy induction rate, and overall survival rate were similar; however, a large allocation bias was noted in the degree of disease progression. Consequently, a definitive conclusion could not be drawn. 79 In the future, larger randomized controlled trials (RCTs) are required to compare LDP and ODP for PDAC. 71

| Laparoscopic pancreaticoduodenectomy
Three RCTs comparing laparoscopic pancreaticoduodenectomy (LPD) ( Table 3) and open pancreaticoduodenectomy (OPD) have been reported to date. [76][77][78] In all studies, although LPD was associated with a prolonged operative time, short-term outcomes such as complication rates, mortality rates, and costs were equivalent between the two procedures. Two single-center RCTs reported a short duration of hospital stays after LPD. 76,77 Conversely, in one multicenter RCT, the 90-day mortality associated with LPD was as high as 10% (P = .2), although the complication rate was equivalent to that of OPD. Consequently, that RCT was terminated prematurely 78 . An oncological retrospective comparison of LPD and OPD for PDAC reported that the R0 resection rate, number of retrieved lymph nodes, MST (approximately 20 months), and 5-year survival rate (20%-30%) were equivalent between the procedures. [96][97][98][99][100][101] The oncological outcomes were also comparable in the three PSM analyses. [101][102][103] In a recent meta-analysis, a significantly higher R0 resection rate and a significantly higher number of lymph node dissections were reported for LPD; however, the 5-year survival rate for LPD was equivalent to that of OPD. 104 The postoperative mortality rate for LPD was higher in the low-volume center than in the high-volume center. 97,99,105,106 The complication rate was lower in the institution with MIPD >20 cases per year or PD >20 cases per year, 107 and it was also reported that the mortality rate was lower in the institution with PD >10 cases per year. 97,101 Therefore, it is necessary to consolidate LPD patients into a highvolume center for their safety as well as to provide appropriate educational guidance to surgeons and facilities. 11,108

| Robotic pancreatectomy
Robotic surgery provides a magnified view, and extremely sophisticated three-dimensional images are associated with high operability (Table 4); therefore, robotic surgery is expected to overcome the limitations of laparoscopic surgery. However, a recent meta-analysis 109,110 and PSM analysis 111 have reported that the frequency of postoperative pancreatic fistula (POPF) and overall complication rates were equivalent between robotic and laparoscopic DP. In addition, a recent meta-analysis comparing the perioperative outcomes of robotic and laparoscopic PD reported that the perioperative outcomes were similar between the two approaches. 71,112 Table 4 shows a comparison of the oncological outcomes between robotic pancreatic surgery and laparoscopic surgery, as well as between robotic and open surgery for PDAC.
In robot-assisted distal pancreatectomy (RDP), the oncological outcomes of R0 resection rate and number of retrieved lymph nodes were comparable to those of laparoscopic and open DP. 93,[113][114][115][116] A study reported that the long-term prognosis associated with RDP, however, was significantly better than that associated with open DP. 116 In addition, the mortality rate and oncologic outcomes of robot-assisted pancreaticoduodenectomy (RPD) were comparable to those of open surgery and laparoscopic surgery. 113,[116][117][118][119][120][121] In a meta-analysis, the conversion rates of robotic PD and robotic DP were lower than those of laparoscopic surgery. 110,112 In particular, the lower emergency conversion rate is an advantage of robotic pancreatic surgery because lower emergency conversion is associated with many postoperative complications and patients that tend to present with poor prognoses. 110,112,122

| Postoperative adjuvant chemotherapy for resectable PDAC
Failure of the aggressive approach with extended lymph node (

| Neoadjuvant therapy for R/BR-PDAC
Although postoperative adjuvant chemotherapy has been effective, the actual rate of completion of courses of therapy has been limited due to postoperative complications and early recurrence after radical resection. 140 Therefore, practitioners have started to conduct preoperative adjuvant treatment for controlling potential distant metastasis, improving local curativeness, and avoiding unnecessary surgery by excluding cases with aggressive tumors. 14

| R-PDAC
Few studies have demonstrated the efficacy of NAT for R-PDAC.
In a retrospective study of PDAC resection using the National Cancer Database (NCDB), MST was found to be significantly longer in neoadjuvant chemotherapy (NAC) than in adjuvant or surgeryalone cases. 141 Another retrospective study for stage I PDAC also reported that NAC had a high R0 resection rate and a favorable prognosis. 142

| Borderline resectable PDAC
Recently, it has been reported that NAT contributed to improved R0 resection rates and extended survival of BR-PDAC patients. 161 In a multicenter retrospective analysis in Japan, it was reported  (Table 6). 69,[146][147][148][149][150]153 Accordingly, there is sufficient evidence for the effectiveness of NAT for BR-PDAC.

| Conversion surgery for initially unresectable PDAC
Overall, in 70%-80% of all PDAC patients are diagnosed as "un- The mortality rates of these conversion surgeries have been reported to be 0%-7%, and complication rates have been reported to be 14%-89%. Therefore, conversion surgery has been performed at an acceptable risk for selected patients. 35 However, most of the reports of conversion surgery for unresectable PDAC were single-center retrospective studies; therefore, the evidence of efficacy is limited. Table 7 shows the results of a recent conversion surgery ( regimen for conversion surgery. CA19-9 level is the most effective biomarker for predicting the potential for resection. To avoid early recurrence after conversion surgery and to obtain a good long-term prognosis, reduction or normalization of CA19-9 levels after TNT is a necessary requirement (see 2.2 Biomarker-based resectability criteria). Furthermore, negative FDG accumulation on PET, which is a metabolic biomarker, and a long period of chemotherapy are also advantageous for long-term survival after conversion surgery. In the future, it is necessary to continue to investigate and determine the optimal criteria for conversion surgery.

| CON CLUS ION
We reviewed the recent trends in surgical treatment for PDAC and summarized the important points. Significant advances in surgical and multimodality treatments are increasing the range of options for treating PDAC. In the future, in order to steadily improve treatment results, not only is research on new biomarkers for assessing operability and tumor dynamics desirable, but research on the development of new anti-cancer therapeutic agents and new multidisciplinary treatment methods is essential.

CO N FLI C T S O F I NTE R E S T
Authors declare no conflicts of interest for this article.