Effect of Risk Factor Score on Early Recurrence After Pancreatectomy for Invasive Pancreatic Ductal Adenocarcinoma

Background/Aim: This study aimed to identify the risk factors for early recurrence (ER) after pancreatic ductal adenocarcinoma (PDAC) resection to create a novel scoring system for ER and analyze their effect on the recurrence pattern. Patients and Methods: Sixty patients with PDAC who underwent pancreatectomy were included. The predicted risk factors for ER were analyzed. A new score defining ER was created and analyzed for recurrence pattern and prognosis. Results: Independent predictors included high CA 19-9 (≥147 U/ml), high lymph node ratio (LNR of ≥0.1277), and no adjuvant chemotherapy (AC). The 5-year overall survival rates with a score of 0, 1, and 2 were 55.8%, 11.0%, and 0%, respectively. In the moderate- risk score group, prognosis was improved by induction of AC within 58 days. Conclusion: Preoperative high CA19-9, high LNR, and no AC could be ER predictors. Induction of postoperative chemotherapy within 58 days may improve prognosis.


Patients and Methods
Patients and patient management.This study retrospectively reviewed a series of 73 consecutive patients who underwent pancreatectomy for invasive pancreatic ductal adenocarcinoma (PDAC) from January 2013 to December 2020 in Jikei University Daisan Hospital.Of the 73 patients, four with macroscopic residual tumor (R2), three with distant metastases, one mortality in hospital, four with acquired carbohydrate antigen 19-9 (CA19-9) values only during jaundice (total bilirubin of >5 mg/dl), and one without lymph node dissection were excluded from the analysis.The remaining 60 patients were included in the final analysis and were all preoperatively diagnosed with a resectable (R) or borderline resectable (BR) PDAC according to the 7 th edition of the General Rules for the Study of Pancreatic Cancer by the Japan Pancreas Society (JPS) (7) and the National Comprehensive Cancer Network (NCCN) guidelines (8) during a multidisciplinary team meeting.We retrospectively reviewed the prospectively maintained database of patients who were histologically diagnosed with PDAC.Performance status was classified using the performance status classification system of the American Society of Anesthesiologist.The Glasgow prognostic score (GPS) was constructed as previously described (9).Pathologic findings and tumor stage were based on the 7 th edition of the General Rules for the Study of Pancreatic Cancer by the JPS (7), which was compatible with the general rules of the 8 th edition of the Union for International Cancer Control (UICC) tumor, node, metastasis (TNM) classification system (10).Lymph node ratio (LNR) was determined by dividing the total number of lymph nodes harboring a metastasis by the total number of examined nodes (11).The cutoff period to differentiate between early and late recurrence (LR) was defined as 12 months according to the previous report (12).
Patient follow-up occurred at our outpatient clinic or at the outpatient clinics of our affiliated clinic, performed by either a surgeon, physician, or radiologist.In general, carcinoembryonic antigen (CEA) and CA19-9 were evaluated every three months.Patients had at least one follow-up by computed tomography of the chest, abdomen, and pelvis every six months for the first five years.Occasional additional imaging modalities performed to determine pattern of recurrence consisted of magnetic resonance imaging (MRI) and fluorodeoxyglucose positron emission tomography.Recurrence of PDAC was defined as newly detected local or distant metastases by the imaging modalities with or without increase in serum CEA or CA19-9.
Adjuvant chemotherapy (AC) was started as soon as possible in the outpatient clinic after discharge, except for patients with Eastern Cooperative Oncology Group performance status 2 or higher and those who refused AC.Prior to 2017, S-1, which is an oral anticancer agent including tegafur, gimeracil, and oteracil potassium, was the first choice for postoperative AC regimens, with gemcitabine selected in cases of allergy to S-1 or renal dysfunction.After 2018, gemcitabine plus nab-paclitaxel (GnP) was considered as the AC of choice for cases with pathologic evidence of residual tumor.
First, the following 14 preoperative factors were analyzed to predict the ER after pancreatectomy for PDAC in univariate and multivariate analyses: age, sex, preoperative biliary drainage, PDAC resectability according to the 7 th edition of the General Rules for the Study of Pancreatic Cancer by the JPS (7) and the NCCN guidelines (8), neoadjuvant therapy, maximum preoperative serum CEA value, maximum preoperative serum CA19-9 value, GPS, procedure, maximum tumor diameter, LNR, resection margin status, AC, and duration between surgery and AC.If patients had no AC, the duration between surgery and AC was defined as the duration between surgery and recurrence.The cutoff values were determined using a receiving operating characteristic (ROC) curve for the ER.
Second, the scoring system was designed using significant predictive factors.The above-selected variables were used for the scoring system.One point was added to the score for each abnormality for the respective cutoff values.Then, overall survival (OS) was compared among the groups according to the scoring system.
Finally, we examined the relationship of duration between surgery and AC with the following first recurrent regions: none, single region (local, liver, lung, peritoneum, or lymph node), and multiple regions.Then, the effect of timing of AC initiation on OS according to the score was examined.
The retrospective study was approved by the Ethics Committee of the Jikei University School of Medicine (Tokyo, Japan); 27-177 (8062), and performed according to the ethical standards of the World Medical Association Declaration of Helsinki.Need for informed consent was waived due to the retrospective nature of the study.
Statistical methods.Continuous data were expressed as the median and range, and compared using the Mann-Whitney U-test between two groups and the Kruskal-Wallis test among three groups.Categorical data were compared using the chi-square test.Univariate factors predictive of the ER (p<0.10) were entered into a logistic regression model to identify independent conversion predictors.Multivariate analysis was performed using a stepwise backward procedure.The survival curve was developed using the Kaplan-Meier method and differences were evaluated using the log-rank test.A pvalue of <0.05 was considered statistically significant in all analyses.

Results
Characteristics of the clinical patients.The characteris¬tics of 60 patients are summarized in Table I, wherein 28 were male and 32 were female.The median age was 72-years  Univariate and multivariate ER predictors.Univariate analysis of the relationship between patient characteristics and the scoring system.These three variables, preoperative maximum serum CA19-9 of ≥147 U/ml, LNR of ≥0.1277, and no AC, were used in designing the scoring system.A score of 3 was not assigned to any patients.Figure 3 shows the comparison of OS rates according to the score.
The OS rate of the patients with a score of 1 was also significantly lower than that of patients with a score of 0 (p=0.009,HR=2.868, 95%CI=1.361-6.827).By using the scoring system, OS after pancreatectomy for PDAC was significantly stratified.

Univariate analysis between the relation of the timing of AC
initiation and the first recurrent regions according to the scoring system.We examined the impact of timing of AC initiation on the first recurrent regions according to the remaining two variables except for AC (preoperative maximum CA19-9 of ≥147 U/ml and LNR of ≥0.1277).
Table III shows that AC initiated ≤58 postoperative days significantly suppressed multiregional metastases in comparison with AC initiated >58 postoperative days in the group with both preoperative maximum CA19-9 of ≥147 U/ml and LNR of ≥0.1277, which was the high-risk group, and either preoperative maximum CA19-9 ≥147 U/ml or LNR ≥0.1277, which was the moderate-risk group (p=0.035 and p=0.017, respectively).There was no significant difference in the number of metastatic regions in the timing of AC initiation after surgery in the group with both preoperative maximum CA19-9 of <147 U/ml and LNR of <0.1277, which was the low-risk group (p=0.416).Additionally, we compared the OS rates according to the timing of AC initiation postoperatively.The timing of AC initiation was based on the previously obtained cutoff value from the ROC curve, which was 58 days.The OS rates of patients in the low-and high-risk groups were not significantly different regarding the timing of AC initiation postoperatively (Figure 4A, p=0.737 and Figure 4C, p=0.700).However, the OS rate of the moderate-risk group with AC initiated >58 days postoperatively was significantly worse than that of the group initiated ≤58 days postoperatively (Figure 4B, p=0.026,HR=2.752, 95%CI=1.194-10.92).The OS rates of the 24 patients in the moderate-risk group were not significantly different between 16 patients with preoperative maximum CA19-9 of ≥147 U/ml and eight patients with LNR of ≥0.1277 (p=0.968).

Discussion
Our study indicated that the independent predicted factors for ER of ≤12 months after pancreatectomy for PDAC were preoperative maximum CA19-9 of ≥147 U/ml, LNR of ≥0.1277, and no AC.OS after pancreatectomy for PDAC was significantly stratified using the scoring system with three significant predicted factors.The subgroup classification using the two factors (preoperative maximum CA19-9 and LNR) except for AC revealed that early AC initiation in the moderate-risk group of ER not only suppressed recurrence of multiregional metastasis but also improved the prognosis.
The originality of this paper is the risk factors scoring and the use of these scores to stratify the prognosis.Furthermore, we analyzed the effect of risk factors on the recurrence pattern and examined the prognosis improvement of the population by treatment-related risk factors evaluated using tumor-related risk factors.Some past reports revealed that the preoperative risk factors for ER included a Charlson age-comorbidity index of ≥4, tumor size of >3.0 cm, lymph node enlargement or resectability on CT, p53 expression, CA19-9, and neutrophilto-lymphocyte ratio (12)(13)(14).The postoperative risk factors for ER included poor tumor differentiation grade, microscopic lymphovascular invasion, perineural invasion, LNR, CA19-9, UICC TNM stage, no AC, and no adjuvant chemoradiotherapy (12,15,16).The risk factors for ER in this study were consistent with previous reports.
Patients with ER of PDAC have a very poor prognosis (13), and there is a need for improvement measures.Previous studies have reported that the liver was the most common location of metastasis in ER (2), and had a poorer prognosis than other metastatic sites (16).Similarly, this study revealed that liver metastasis was most often observed as an ER site.
Past reports included neoadjuvant chemotherapy, chemoradiotherapy, or AC as measures to improve the prognosis of ER that could be managed by oncologists (12,15,16).In this study, neoadjuvant therapy was performed in four BR and two R cases, and the rate of BR patients with neoadjuvant therapy cases was significantly higher than that of R patients (p<0.001;data are not shown).The number of cases that received neoadjuvant therapy was small.Probably, these did not affect the postoperative ER suppression in the univariate analysis.Conversely, AC was a significant ER inhibitor in this study.However, to our knowledge, no previous study has examined the improved prognosis of the population scored for ER factors according to the timing of AC.The benefit of AC to patients with pancreatic cancer is accepted, but its optimal timing postoperatively remains under investigation.Prior studies suggested no benefit to  early AC initiation but only compared OS between arbitrarily assigned time periods (17,18).Conversely, Ma reported that patients with stage I to II pancreatic cancer who commenced AC within 28-59 days after pancreatectomy had improved survival outcomes compared with those with AC before 28 days or after 59 days (19).In this study, only patients in the moderate-risk group with either CA19-9 of ≥147 U/ml or LNR of ≥0.1277 might obtain survival benefits by initiating AC ≤58 days after surgery.Most of our cases commenced AC within 28-59 days postoperatively although only one started postoperative adjuvant chemotherapy within 28 days postoperatively.The rate of patients with stage I to II pancreatic cancer in the moderate-risk group was 79.2% (data were not shown).AC was a significant ER inhibitor (12); however, it might be very important not only to firmly introduce AC, but also to initiate it at an optimal timingnot too early or too late, particularly, for patients in groups at risk of recurrence.
Study limitations.First, this was a retrospective study.Selection bias for the regimen of the neoadjuvant therapy and AC could arise, despite treatment strategies based on Clinical Practice Guidelines for Pancreatic Cancer 2019 from the Japan Pancreas Society (20) or Pancreatic Adenocarcinoma, Version 2.2021, NCCN Clinical Practice Guidelines in Oncology (8).This is because the decision to select the regimen of  neoadjuvant therapy and AC might vary significantly among surgeons or oncologists.Second, the number of patients who underwent pancreatectomy was small, and the survival rate might be underestimated.Therefore, large-scale multicenter studies are required to confirm the influence of other risk factors, including no neoadjuvant therapy or tumor resectability, on survival.

Conclusion
In conclusion, our findings demonstrated that preoperative maximum CA19-9 of ≥147 U/ml, LNR of ≥0.1277, and no AC could be predictors of ER of ≤12 months.OS after pancreatectomy for PDAC was significantly stratified using the scoring system with the three significant predicted factors.Early AC initiation in the moderate-risk group of ER might suppress recurrence of multiregional metastasis and improve prognosis.

Figure 1 .
Figure 1.Overall survival according to early recurrence of ≤12 months.Overall survival rate in early recurrence (ER) group was significantly worse than that in the late recurrence (LR) group.Median survival time (MST); 1.81 years in all cases; 1.00 year in the ER group; 2.63 years in the LR group.

Figure 3 .
Figure 3. Overall survival (OS) according to the scoring system.OS rate was significantly different in each group.Median survival time (MST); 2.87 years in the group with a score of 0; 1.81 years in the group with a score of 1; 0.81 years in the group with a score of 2.

Figure 2 .
Figure 2. Distribution of recurrence patterns (A) and distribution of recurrence sites at different time points (B).

Figure 4 .
Figure 4. Overall survival (OS) rates according to the timing of initiation of adjuvant therapy after surgery.OS in the group with both preoperative maximum CA19-9 of <147 U/ml and lymph node ratio (LNR) of <0.1277 (the low-risk group) (A).OS in the group with either preoperative maximum CA19-9 of ≥147 U/ml or LNR of ≥0.1277 (the moderate-risk group) (B).OS in the group with both preoperative maximum CA19-9 of ≥147 U/ml and LNR of ≥0.1277 (the high-risk group) (C).Median survival time (MST) in the patients with AC initiated ≤58 postoperative days and those withAC initiated >58 postoperative days; 2.15 years and 2.66 years in the low-risk group, respectively; 2.14 years and 0.78 years in the moderate-risk group, respectively; 1.02 years and 0.87 years in the high-risk group, respectively.

Table I .
Patient characteristics.

Table II .
Univariate and multivariate analysis for early recurrence.

Table III .
The relation of recurrent regions with duration between surgery and adjuvant chemotherapy.Carbohydrate antigen 19-9; LNR: lymph node ratio.