Society: SSAT

Background
Pancreatic ductal adenocarcinoma (PDAC) exists in several morphological subtypes differing in prognostic significance. However, to date, a clinico-morphological correlation of these subtypes in the context of neoadjuvant therapy (NAT) has not been performed. The aim of this study was 1) to investigate the frequency of the different PDAC morphologies in patients undergoing radical intent pancreatectomy after NAT; and 2) to determine the prognostic impact of the presence of a secondary morphology in the primary tumor.

Methods
All patients who underwent pancreatic resection after NAT for PDAC (2013-2019) at one academic institution were enrolled. All pathological samples were included in toto and reviewed by experienced pathologists. The presence of a secondary morphology in the primary tumor specimen was determined according to a morphological cut-off ≥10%. Tumor regression grade (TRG) was classified according to the MDACC Scoring system. The clinico-pathological characteristics and the survival of the cohort were studied by means of conventional statistical analyses.

Results
Among the 401 included patients 205 (51,5%) received Folfirinox, 134 (33,7%) gemcitabine/nab-paclitaxel. The median follow-up was 28.0 months, and the median disease specific survival (DSS) was 29.7 months. The median DSS associated to the principal tumor morphologies and their relative frequencies is shown in the Table. Gland forming PDAC with conventional morphology (n=167, 41,6%) was the most frequent subtype. Overall, no significant difference in DSS was observed. After pairwise comparison, the papillary morphology shown a significant higher survival rate compared to other less frequent subtypes (cribiform, p<0.019; gyriform, p<0.008; micropapillary, p<0.048; and adenosquamous, p<0.006). Overall, 247 (61,1%) displayed only a single principal tumor morphology, while 154 (38,4%) presented an additional secondary morphology in the primary tumor. PDACs harboring a secondary tumor morphology shown a significantly more advanced pathological profile and a higher TRG, as well as significantly shorter DSS and recurrence free survival (RFS) (Figure). At multivariable Cox regression, the presence of a secondary tumor morphology was independently associated with worse DSS (HR 1.881, 95% CI 1.384-2.557, p<0.001) and RFS (HR 1.635, 95% CI 1.230-2.175, p<0.001).

Conclusion
In patients receiving pancreatectomy after NAT, the presence of a secondary morphology in the primary tumor is frequent, occurring in over one third of the cases. This feature is associated with a less favorable pathological profile and a higher TRG, and represents an independent predictor of shorter DSS and RFS. Based on these findings, including a detailed morphological description in pancreatectomy pathology reports might provide valuable prognostic information and possibly help post-surgical decision-making.

Background: Surgical resection is necessary for the curative treatment of periampullary malignancies. Many patients will undergo endoscopic retrograde cholangiopancreatography (ERCP) prior to surgery, for obstructive jaundice or diagnostic purposes. Post-ERCP pancreatitis (PEP) is one of the most common complications of this procedure, but its impact on postoperative outcomes is not well studied. We hypothesize that patients who experience PEP will experience worse postoperative outcomes.

Methods: All patients with periampullary malignancies who underwent surgical resection between 2017-2020 at a single, high-volume institution were reviewed from a prospectively maintained database. Post-ERCP pancreatitis was defined as clinically significant pancreatitis requiring post-procedure or prolonged admission, as outlined by Cotton et al (1991). Groups were compared with Mann-Whitney U-tests for continuous variables and chi-squared or Fisher’s exact tests for categorical variables. Multivariable analysis was performed with logistic regression.

Results: Four hundred fifty-five patients underwent surgical resection for periampullary malignancy in the studied time frame, of which 317 patients underwent preoperative ERCP: 237(74.8%) for pancreatic cancer, 51(16.1%) ampullary cancer, 22(6.9%) distal cholangiocarcinoma, 4(1.3%) duodenal cancer, and 2(0.9%) pancreatic neuroendocrine tumors. A total of 27(8.8%) patients developed post-ERCP pancreatitis. Groups were comparable in demographics, comorbidities, clinical stage and tumor resectability. There was no significant difference in frequency of neoadjuvant therapy (NAT) (p=0.16). PEP was associated with greater estimated blood loss during surgery (300[300] vs 500[550] mL, p=0.03). There was no significant difference in operative time, post-operative length of stay, 30-day readmission rate and 30- and 90-day mortality rate. While overall complication rates did not differ between groups (p=0.12), PEP patients experienced higher rates of complications Clavien-Dindo class III or above (10.7% vs 33.3%, p<0.01), including clinically relevant postoperative pancreatic fistulas (CR-POPF) (7.9% vs 25.9%, p<0.01). On multivariable analysis, PEP remained independently associated with CR-POPF after adjusting for gland texture, duct diameter, EBL, and pathology (OR 4.88, 95% CI: 1.62–14.68, p<0.01), as well as class III or higher complications after adjusting for age, EBL, pathology, and other factors (OR 6.79, 95% CI: 2.22–18.89, p<0.01).

Conclusions: Patients with periampullary malignancies who develop PEP are at higher risk for major complications after surgery, including clinically relevant postoperative pancreatic fistulas. Post-ERCP pancreatitis should be considered a strong risk factor for postoperative morbidity and CR-POPF, suggesting that PEP patients may require alternative fistula mitigation approaches.

Background:Multidrug adjuvant therapy following pancreatectomy has yielded substantial improvement in the prognosis of pancreatic cancer patients with localized disease, establishing a new treatment paradigm. However, out of controlled experimentalsettings, the proportion of patients accessing modern chemotherapy regimens is largely unknown
Methods:A prospective, observational study was conducted. All consecutive patients receiving primary curative surgery for pancreatic ductal carcinoma (Jan 2019 - Jul 2022) were enrolled(#NCT03788382). The primary aim is to define actual adjuvant treatment utilization and its association with baseline and perioperative patient characteristics. Medical oncologist charts were retrieved along with close patient follow-up. A precision-based approach was used to calculate sample size
Results:317 patients underwent pancreatectomy, among which 237 (74.8%) received subsequent adjuvant therapy after a median of eight weeks (IQR 6-10) after surgery. Among such, Gemcitabine alone and FOLFIRINOX were employed in 42% and 38% of cases, respectively, followed by Gemcitabine-based(16%) and other regimes (5%). Main reasons for chemotherapy omission were postoperative failure-to-thrive (39%), baseline comorbidities (20%), and physician’s decision (21%). Patient refusal and early disease recurrence also accounted for 20%. The likelihood of postoperative therapy omission steadily increased with age up to 50% for individuals older than 80. It varied across geographical areas, being twice as high for inhabitants of Northern Italy regions compared to Central and Southern areas (30 vs 15%).Older age (OR 1.10,95%CI 1.011-1.14), family history of pancreatic cancer (OR 2.46,95%CI 1.33-4.56) and developing postoperative pancreatic fistula (OR 2.54,95%CI 11.05-6.18) were primary determinant of attrition after surgery, whereas no pathological parameter influenced adjuvant therapy initiation.Adjuvant FOLFIRINOX utilization increased tenfold over the study period (from 4.1 to 44.2%).Patients receiving such a regimen were significantly younger (median 65 vs.74 years old,OR 0.86,95%CI 0.81-0.90; p<0.001) and displayed more advanced N stage (vs.
N0: N1 OR 3.10; N2 OR 2.87), while alcohol abuse (OR 0.26,95%CI 0.09-0.78) and developing severe complications (Clavien-Dindo≥3; OR 0.24,95%CI 0.07-0.85) were associated with FOLFIRINOX omission. No difference was evident in time to chemotherapy initiation between FOLFIRINOX and other schemes
Discussion:This study provides a contemporary, real-world snapshot depicting a limited utilization of adjuvant therapy following curative resection for localized pancreatic cancer. Despite FOLFIRINOX being increasingly employed in this setting, one out of four patients still fails to receive any postoperative chemotherapy, mostly due to postoperative complications, and most patients are treated with suboptimal regimens

Objective: To compare the effectiveness of radiologic and biochemical response evaluation after neoadjuvant FOLFIRINOX chemotherapy to predict survival outcome and investigate the efficacy of adjuvant chemotherapy in patients with non-metastatic pancreatic ductal adenocarcinoma.
Summary Background Data: To maximize the effect of surgery, the treatment paradigm is shifting from upfront surgery to neoadjuvant chemotherapy with advances in chemotherapy. Therefore, neoadjuvant FOLFIRINOX chemotherapy has been increasingly used, however, the response evaluation methods are still inconsistent and the benefits of adjuvant chemotherapy are also unclear.
Methods: 160 non-metastatic pancreatic ductal adenocarcinoma patients who underwent curative-intent pancreatectomy after at least 4 cycles of neoadjuvant FOLFIRINOX chemotherapy between 2012 and 2020 were identified. Patients with a normalized CA 19-9 level after neoadjuvant chemotherapy were defined as biochemical responders. Patients with complete or partial response according to Response Evaluation Criteria in Solid Tumors were defined as radiologic responders. Survival analysis was performed using Kaplan-Meier estimates.
Results: Of 160 patients, 58 (36.3%) were identified as biochemical and radiologic responder (BR+/RR+), 31 (19.4%) as only biochemical responder (BR+/RR-), 17 (10.6%) as only radiologic responder (BR-/RR+), 44 (27.5%) as non-responder (BR-/RR-). The 2-year overall survival rates were 82.0%, 77.3%, 64.7%, 45.5% for BR+/RR+, BR+/RR-, BR-/RR+, and BR-/RR-, respectively (P = 0.008). The differences of 2-year overall survival rates between patients who completed adjuvant chemotherapy and those who stopped or did not receive adjuvant chemotherapy were 20.4% (88.8% vs. 68.4%, P < 0.001), 53.6% (91.1% vs. 37.5%, P < 0.001), 58.3% (83.3% vs. 25.0%, P = 0.002), 49.4% (62.7% vs. 13.3%, P < 0.001) in BR+/RR+, BR+/RR-, BR-/RR+, and BR-/RR-, respectively.
Conclusions: Biochemical response to neoadjuvant FOLFIRINOX chemotherapy has a greater impact on prognosis than radiologic response, and radiologic response plays an important role only in biochemical non-responders. Completion of adjuvant chemotherapy is important in all patients regardless of response to neoadjuvant FOLFIRINOX chemotherapy, especially except for BR+/RR+, the efficacy was greater.

Overall survival according to the biological and radiological response to neoadjuvant FOLFIRINOX

Overall survival according to the biological and radiological response to neoadjuvant FOLFIRINOX

Overall survival according to the completion of adjuvant chemotherapy<br /> A. Biochemical and radiologic responders<br /> B. Biochemical only responders<br /> C. Radiologic only responders<br /> D. Non-responders

Overall survival according to the completion of adjuvant chemotherapy
A. Biochemical and radiologic responders
B. Biochemical only responders
C. Radiologic only responders
D. Non-responders

Background and Aim:

Disconnected pancreatic duct syndrome (DPDS) is characterized by complete transection of the main pancreatic duct by central pancreatic necrosis, leading to discontinuity between viable secreting pancreatic tissue upstream and the gastrointestinal tract. Manifestations include high output external pancreatic fistula, symptomatic pseudocyst and recurrent acute pancreatitis/chronic pancreatitis impacting quality of life. Standard surgical management for persistent DPDS involves resection of the upstream gland (distal pancreatectomy), but may result in brittle insulin dependent post surgical diabetes. Islet cell autotransplantation is an increasingly available adjunctive measure to reduce risk of of diabetes that has been shown to improve quality of life (QOL) after total pancreatectomy, but has not been reported after DP for DPDS. We report our single center experience on DPIAT as a novel management for DPDS in necrotizing pancreatitis (NP).

Methods:
All NP patients with DPDS undergoing DPIAT from 2005-22 were included in the study. Baseline demographics, indications for DPIAT, HbA1c, opioid requirements and quality of life (QOL) metrics at baseline and at 6 months, insulin requirement and graft function at 6 months were recorded. Patients were classified as insulin independent or dependent, and as having islet graft function or graft failure (meal-stimulated C-peptide <0.6 ng/mL).
Results:
Among 676 patients with NP managed during this period, 10 patients [males 6, median (IQR) age-47.5 (35-51) years] who underwent DPIAT for DPDS were included in the study. Management of NP involved endoscopic approach in 5, percutaneous drainage in 1, open necrosectomy in 4. The site of disconnection was in the neck 6 (60%), body 3(30%) and tail 1 (10%). Indications for DPIAT were recurrent pancreatitis in 6 (60%), chronic pancreatitis in 3 (30%) and fistula 1(10%). 1 had diabetes prior to surgery and 8/10 (80%) patients were on long standing opioids. Median (IQR) islet equivalents/kg infused intraportally was 1224 (705-2818). No periprocedural complications were reported. 2 patients did not have long term follow up. At 6 months after DPIAT, only 2/10 (20%) required opioids and the majority showed improvement in QOL. All patients had preserved islet graft function with median (IQR) C-peptide of 6.4 (5.5-9.9), median (IQR) HbA1c of 6.05 (5.8-6.6) with 8/10 (80%) requiring baseline insulin and 6/10 (60%) required pancreatic enzymes.
Conclusion:
DPIAT for DPDS is associated with improvement in pain and QOL with discontinuation of opioids in most. Although majority required baseline insulin, all patients had graft function which could potentially lower the HbA1c and long term risk of microvascular complications. Long term data in larger cohort is needed to validate our conclusions.

Introduction: The advantages of neoadjuvant chemotherapy (NAC) over upfront surgery in patients with resectable pancreatic adenocarcinoma (PDAC) remain under debate. In this study, we aim to evaluate the impact of NAC and its dependence on receiving adjuvant therapy on the oncologic and survival outcomes of patients with resectable PDAC.

Methods: All patients with resectable PDAC (defined by AHPBA/SSO/SSAT consensus guidelines) who underwent oncologic resection at a single, high-volume institution between Jan 2017 and Feb 2020 were retrospectively reviewed. Groups were compared using chi-squared or Mann-Whitney U-test. Kaplan-Meier and Cox proportional-hazards regression were used for survival analysis.

Results: Out of 228 patients with resectable PDAC, 93 (40.8%) had neoadjuvant chemotherapy and 135 (59.2%) underwent upfront surgery (US). Patients who received NAC were younger (NAC vs US, med[IQR]: 67.5[12.7] vs. 80.0[13.7] yrs). Groups were similar in comorbidities except for COPD (3.2% vs 10.4%, p=0.04). Patients with NAC had larger tumors at diagnosis (T2 disease: 58.7% vs 39.0%, T3 disease: 16.3% vs 11.4%, p<0.01), but similar clinical nodal staging (p=0.89). The med[IQR] duration of NAC was 2.30[0.96] mos. Surgery type, approach, duration, and EBL were similar between groups. NAC was associated with more node negative disease on pathology (56.7% vs 43.3%, p<0.01) and lower 30-d readmission rates (5.4% vs 13.3%, p=0.05). Groups were comparable in postoperative complications such as pancreatic fistula, DGE, and organ space infection, as well as 90-d mortality (4.3% vs 9.6%, p=0.13), and similar proportions went on to reach adjuvant chemotherapy (77.5% vs 75.0%, p=0.70). NAC was associated with higher rates of one-year survival (88.5% vs 58.3%, p<0.01) and better overall survival (med (95% CI): 31.7 (24.2 – 39.3) vs 15.3 (11.5 – 19.0) mos, p<0.01). When considering patients with adequate records on the course of systemic therapy (n=174), the survival benefit of NAC is lost in patients who do not receive adjuvant chemotherapy (Figure, p<0.01). This difference remained significant after adjusting for differences between groups, clinical stage, and other factors (NAC w/ vs w/o adjuvant, HR (95% CI): 0.36 (0.15 – 0.86), p=0.02).

Conclusion: In resectable PDAC, neoadjuvant chemotherapy was associated with improved overall survival and more node negative disease after surgery. However, the survival benefit is lost if patients do not receive adjuvant chemotherapy. This supports further investigating the potential role of total neoadjuvant chemotherapy in resectable pancreatic cancer.

Introduction:
NCCN guidelines for the management of pancreatic ductal adenocarcinoma (PDAC) recommend pancreatic resection 4-8 weeks after completion of neoadjuvant therapy. Increasing data suggests the benefit of neoadjuvant therapy in PDAC, but few studies have focused on the impact of time to surgery in the setting of neoadjuvant care. Therefore, in this analysis we aim to identify the optimal time between completion of neoadjuvant therapy and definitive surgery.

Methods
Patients with PDAC who underwent single modality neoadjuvant therapy (SMNT) or total neoadjuvant therapy (TNT) prior to pancreatectomy at a single NCI designated cancer center between 2010 and 2020 were included. Where TNT was defined as both neoadjuvant chemotherapy followed by chemoradiation therapy, and SMNT was defined as either neoadjuvant chemotherapy or chemoradiation therapy. Patients were evaluated by treatment type and by time to surgery: ≤4 weeks, 5 to 8 weeks, or >8 weeks, after completion of neoadjuvant therapy. Logistic regression models were used to analyze the relationship between time to surgery and outcome variables, while Cox proportional hazards models were used for survival analyses.

Results:
121 patients who received TNT (n = 68) or SMNT (n = 53) prior to pancreatic resection met inclusion criteria. As shown in figure 1, among all patients those who underwent surgery in ≤4 weeks demonstrated significantly worse disease-free survival (DFS) relative those who had surgery in >8 weeks (12.4 vs 19.5 months, p=0.03). Similarly as shown in figure 2, among patients who received TNT those in both the 5-8 week and >8-week windows demonstrated significantly longer DFS than those in the ≤4-week group (11.3 vs 18.7 months p=0.05, and 11.3 vs 19.5 months p=0.01, respectively). For the entire cohort, those who had surgery after 8 weeks also demonstrated significantly lower rates of fibrosis relative to ≤4 weeks (p=0.013) and 5-8 weeks (p=0.043).

Conclusion:
Patients who underwent pancreatic resection ≤4 weeks after completion of neoadjuvant therapy exhibited significantly shorter DFS, and higher rates of fibrosis relative to patients who underwent surgery after longer intervals. These initial findings suggest that longer intervals to surgery after neoadjuvant therapy may offer significant survival benefits, though additional investigations are needed.

Figure 1: Kaplan-Meier plot depicting Disease-Free Survival between time to surgery groups

Figure 1: Kaplan-Meier plot depicting Disease-Free Survival between time to surgery groups

Figure 2: Kaplan-Meier plot depicting Disease-Free Survival among patients who recieved total neoadjuvant therapy by time to surgery group

Figure 2: Kaplan-Meier plot depicting Disease-Free Survival among patients who recieved total neoadjuvant therapy by time to surgery group

Introduction
Cancer care coordination (CCC) is an integral component of health care delivery. Neoadjuvant therapy (NT) is increasingly used prior to surgery for most localized gastrointestinal (GI) and hepatopancreatobiliary (HPB) cancers. Although NT necessitates a multidisciplinary approach, there is little existing literature on the quality of CCC during NT. The objective of this study was to characterize patient perceptions of CCC during NT using a mixed methods approach.

Methods
A cross-sectional analysis of patients with GI/HPB cancers receiving NT were enrolled as part of a prospective longitudinal cohort study to evaluate their real-time experience using a customized smartphone application. Enrolled patients completed the Cancer Care Coordination Questionnaire for Patients (CCCQ-P), a 20-item validated measure of care coordination. Items were scored on a 5-point Likert scale and subsections on communication (13 questions) and navigation (7 questions) were calculated with higher scores signifying better cancer care coordination. Univariate linear regression was used to calculate the impact of fragmented care and other factors on perceived CCC. Semi-structured interviews were conducted among a convenience sample of patients (n=5) using an interview script developed via preliminary survey results, evidence synthesis, and expert opinion.

Results
Among 82 participants, mean age was 61 years and 68% were male. The most common malignant diagnoses included rectal (44%), pancreas (29%), and esophagus (16%). Mean length of NT treatment was 3.3 months; 72% of patients underwent surgical resection following NT. Overall (mean 76.6 out of 100), communication subsection (48.6 out of 65) and navigation subsection (28.0 out of 35) CCCQ-P scores suggested overall positive perceptions of care coordination. Specific items rated lowest by respondents included whether providers asked patient how they were coping with treatment (3.3) and how visits with other providers were going (3.3) (Figure). On a scale from 1-10, perceived care coordination and quality of care was rated as 8.5 and 9.0, respectively. Fragmented care during NT was not associated with worse CCCQ-P scores (β=1.15, p=0.66). Qualitative analysis of patient interviews highlighted the need for better coordination among physicians before communicating the plan to patients, as well as the importance of providers summarizing plans to patients in verbal and written form.

Conclusions
Successful completion of neoadjuvant therapy requires significant care coordination among patients and healthcare professionals. In this cross-sectional analysis of a prospective cohort study, patient perceptions of CCC during NT were generally positive. Future research should focus on optimizing all aspects of care delivery in order to maximize completion of NT and receipt of surgical resection.

INTRODUCTION:
We previously demonstrated that risk-stratified pancreatectomy care pathways are associated with decreased opioid use, earlier drain removal, and decreased index hospitalization length of stay (LOS). However, readmission rates have remained constant despite these interventions. The purpose of this study was to evaluate reasons for readmission and identify opportunities to refine postoperative care to reduce readmission rates.

METHODS:
We performed a single-institution retrospective cohort study of consecutive patients who underwent pancreatectomy from October 2016 - April 2022. Complications were prospectively graded in a biweekly faculty and advanced practice provider meeting using the ACCORDION system. Reasons, timing, and treatments associated with 90-day readmissions were analyzed by further review of the electronic health record. Primary reason for readmission was classified, in order of priority, as technical, infectious, medical/metabolic, and other.

RESULTS:
849 patients underwent 541 (64%) pancreatoduodenectomies, 285 (34%) distal pancreatectomies, and 23 (3%) other resections. There were 84% (713/849) open and 16% (136/849) minimally invasive procedures. 25% (212/849) of all patients were readmitted. Among the 212 readmitted patients, readmission occurred on median day 8 (interquartile range, IQR 3-17) after discharge with LOS of median 4 days (IQR 2-8). Approximately 26% (56/212) were early readmissions (within 3 days of discharge), and 26% (55/212) required multiple readmissions. 90% (191/212) of readmissions were for complications related to surgery. Among patients readmitted for surgical complications, the primary reason was technical in 48% (92/191), infectious in 19% (37/191), metabolic/medical in 26% (50/191) and other in 6% (12/191) patients. 71% (91/129) of patients with technical or infectious complications required interventional procedures. Of patients readmitted for metabolic/medical reasons, 82% (41/50) received no intervention or only required additional consultations (most commonly nutrition, pain management, and endocrine). Median LOS was longer for technical/infectious vs. metabolic/medical readmissions (5 vs. 4 days, p=0.025). Of the 53 early readmissions due to surgical complications, 64% (34/53) required invasive intervention, while 30% (16/53) required no intervention or only additional consultations.

CONCLUSION:
Technical and infectious complications account for approximately two-thirds of surgical readmissions after pancreatectomy, but one-fourth are readmitted for metabolic/medical reasons. Of these, only one-third require invasive interventions or parenteral nutrition. Adding capacity for same-day imaging, in-clinic consultations, intravenous fluid administration, and urgent care centers distinct from emergency rooms may be practical measures for reducing readmissions after pancreatectomy.

<b>Table 1: </b>Interventions by primary reason for re-admission. Patients may have multiple interventions. Abbreviations: IR= interventional radiology, TPN= total parenteral nutrition, LOS= length of stay, IQR= interquartile range.

Table 1: Interventions by primary reason for re-admission. Patients may have multiple interventions. Abbreviations: IR= interventional radiology, TPN= total parenteral nutrition, LOS= length of stay, IQR= interquartile range.

<b>Figure 1: </b>Interventions by primary reason for re-admission. Patients may have multiple interventions. Interventions in bar graphs listed in order they appear in the legend. Abbreviations: IR= interventional radiology, TPN= total parenteral nutrition.

Figure 1: Interventions by primary reason for re-admission. Patients may have multiple interventions. Interventions in bar graphs listed in order they appear in the legend. Abbreviations: IR= interventional radiology, TPN= total parenteral nutrition.

Background: Drain amylase has been thoroughly investigated as a predictor of postoperative pancreatic fistula (POPF) after pancreaticoduodenectomy (PD). The use of drain amylase on day 1 (DA-D1) after PD to predict occurrence of POPF has been controversial with various cutoffs proposals. The purpose of this study is to evaluate the optimal DA-D1 level to predict clinically relevant postoperative pancreatic fistula (CR-POPF) in a large population-based cohort.

Methods: The National Surgical Quality Improvement Program (NSQIP) database between 2014-2020 was queried for patients who underwent elective PD for benign and malignant indications. Patients with the following data were included in the analysis: sex, body mass index (BMI), DA-D1, preoperative bilirubin, duct size, gland texture, length of stay (LOS), duration of surgical drain, and post-operative complications to determine development of POPF and CR-POPF per International Study Group of Pancreatic Fistula guidelines. Receiver operative curve (ROC) and Youden’s index were used to assess the predictive performance and optimal cutoff points for DA-D1 to predict CR-POPF. Our proposed DA-D1 value was then confirmed with a backward stepwise multivariable logistic regression to determine hazard ratios (HR) for CR-POPF and applied using conditional logistics regression to subgroups based on modified fistula risk score (FRS) as developed from the 2012 NSQIP pancreatectomy database study.

Results: A total of 47,275 cases were identified, of which 6,087 met inclusion criteria. The mean age was 66.7±12.5 years, mean BMI 27.4±6.0, and 47.4% were females. Majority of patients underwent open PD (n=5,156, 84.7%). Mean DA-D1 was 2,897±8,636 U/L and the median duration pf drain was 5 days. POPF was documented in 877 (14.4%) patients; 544 (8.9%) were CR-POPF. ROC for DA-D1 for any POPF had an AUC of 0.807 (95% CI 0.793-0.822), and an AUC of 0.779 for CR-POPF (95% CI 0.759-0.798). Youden’s index for the CR-POPF ROC coordinates was determined at a 77.6% sensitivity and 66.3% specificity, which correspond to DA-D1 values of ≥720 U/L as an optimal cutoff. CR-POPF was higher for patients with DA-D1 ≥720 U/L (HR 4.6; p=0.001), male sex (HR 1.37; p=0.001), high BMI (HR 1.57; p<0.001), small duct size of 3-6mm (HR 1.41; p=0.041) and <3mm (HR=1.75; p=0.004), and soft gland (HR 1.55; p<0.001). Patients with a negligible, low, intermediate, and high FRS had respectively 2%, 5%, 10%, and 19% rate of CR-POPF. Comparison of CR-POPF rates in each FRS category based on DA-D1 cut-off of 720 U/L is demonstrated in Figure 1.

Conclusion: A level of 720 U/L DA-D1 after elective PD is a clinically useful predictor of CR-POPF for patients in all FRS categories. For patients with low to intermediate FRS, further research is needed to affirm DA-D1<720 U/L as a part of enhanced recovery after surgery protocols for safe early drain removal.

<b>Figure 1</b>: Comparison of CR-POPF rates in each FRS category based on the DA-D1 cutoff of 720 U/L. <b>DA-D1</b>: Drain Amylase on Day 1; <b>CR-POPF</b>: Clinically Relevant Postoperative Pancreatic Fistula; <b>FRS</b>: Fistula Risk Score.

Figure 1: Comparison of CR-POPF rates in each FRS category based on the DA-D1 cutoff of 720 U/L. DA-D1: Drain Amylase on Day 1; CR-POPF: Clinically Relevant Postoperative Pancreatic Fistula; FRS: Fistula Risk Score.

Kathryn Stackhouse

Appearances