Serum amylase elevation following pancreatoduodenectomy with pancreatogastrostomy is strongly associated with major complications

Machine learning model-based prediction of postpancreatectomy acute pancreatitis following pancreaticoduodenectomy: A retrospective cohort study

BACKGROUND

The International Study Group of Pancreatic Surgery has established the definition and grading system for postpancreatectomy acute pancreatitis (PPAP). There are no established machine learning models for predicting PPAP following pancreaticoduodenectomy (PD).

AIM

To explore the predictive model of PPAP, and test its predictive efficacy to guide the clinical work.

METHODS

Clinical data from consecutive patients who underwent PD between 2016 and 2024 were retrospectively collected. An analysis of PPAP risk factors was performed, various machine learning algorithms [logistic regression, random forest, gradient boosting decision tree, extreme gradient boosting, light gradient boosting machine, and category boosting (CatBoost)] were utilized to develop predictive models. Recursive feature elimination was employed to select several variables to achieve the optimal machine algorithm.

RESULTS

The study included 381 patients, of whom 88 (23.09%) developed PPAP. PPAP patients exhibited a significantly higher incidence of postoperative pancreatic fistula (55.68% vs 14.68%, P < 0.001), grade C postoperative pancreatic fistula (9.09% vs 1.37%, P = 0.001). The CatBoost algorithm outperformed other algorithms with a mean area under the receiver operating characteristic curve of 0.859 [95% confidence interval (CI): 0.814-0.905] in the training cohort and 0.822 (95%CI: 0.717-0.927) in the testing cohort. According to shapley additive explanations analysis, pancreatic texture, main pancreatic duct diameter, body mass index, estimated blood loss, and surgery time were the most important variables based on recursive feature elimination. The CatBoost algorithm based on selected variables demonstrated superior performance, with an area under the receiver operating characteristic curve of 0.837 (95%CI: 0.788-0.886) in the training cohort and 0.812 (95%CI: 0.697-0.927) in the testing cohort.

CONCLUSION

We developed the first machine learning-based predictive model for PPAP following PD. This predictive model can assist surgeons in anticipating and managing this complication proactively.

Extracellular volume fraction derived from dual-energy CT: a potential predictor for acute pancreatitis after pancreatoduodenectomy

OBJECTIVES

To investigate the value of extracellular volume (ECV) fraction and fat fraction (FF) derived from dual- energy CT (DECT) for predicting postpancreatectomy acute pancreatitis (PPAP) after pancreatoduodenectomy (PD).

METHODS

This retrospective study included patients who underwent DECT and PD between April 2022 and September 2022. PPAP was determined according to the International Study Group for Pancreatic Surgery (ISGPS) definition. Iodine concentration (IC) and FF of the pancreatic parenchyma were measured on preoperative DECT. The ECV fraction was calculated from iodine map images of the equilibrium phase. The independent predictors for PPAP were assessed by univariate and multivariable logistic regression analysis and receiver operating characteristic (ROC) curve analysis.

RESULTS

Sixty-nine patients were retrospectively enrolled (median age, 60 years; interquartile range, 55-70 years; 47 men). Of these, nine patients (13.0%) developed PPAP. These patients had lower portal venous phase IC, equilibrium phase IC, FF, and ECV fraction, and higher pancreatic parenchymal-to-portal venous phase IC ratio and pancreatic parenchymal-to-equilibrium phase IC ratio, compared with patients without PPAP. After multivariable analysis, ECV fraction was independently associated with PPAP (odd ratio [OR], 0.87; 95% confidence interval [CI]: 0.79, 0.96; p < 0.001), with an area under the curve (AUC) of 0.839 (sensitivity 100.0%, specificity 58.3%).

CONCLUSIONS

A lower ECV fraction is independently associated with the occurrence of PPAP after PD. ECV fraction may serve as a potential predictor for PPAP after PD.

CLINICAL RELEVANCE STATEMENT

DECT-derived ECV fraction of pancreatic parenchyma is a promising biomarker for surgeons to preoperatively identify patients with higher risk for postpancreatectomy acute pancreatitis after PD and offer selective perioperative management.

KEY POINTS

PPAP is a complication of pancreatic surgery, early identification of higher-risk patients allows for risk mitigation. Lower DECT-derived ECV fraction was independently associated with the occurrence of PPAP after PD. DECT aids in preoperative PAPP risk stratification, allowing for appropriate treatment to minimize complications.

Serum amylase on postoperative day one is a strong predictor of pancreatic fistula after pancreaticoduodenectomy: a retrospective cohort

Background/aim

Early identification of patients at risk for developing postoperative pancreatic fistula (POPF) after pancreaticoduodenectomy (PD) may facilitate drain management. In this context, it was aimed to examine the efficiency of the serum amylase (SA) value on postoperative day (PoD) 1 in predicting the occurrence of POPF.

Materials and methods

A total of 132 patients who underwent PD were studied. Occurrences of POPF were classified according to the International Study Group on Pancreatic Fistula classification as a biochemical leak (BL) or clinically relevant grade b/c POPF (CR-POPF). Receiver operating characteristic analysis identified a threshold value of SA on PoD 1 associated with POPF formation.

Results

Overall, 66 (50%) patients had POPF, including 51 (38.7%) with BL and 15 with CR-POPF (11.3%). The threshold value of SA associated with the development of POPF was 120 IU/L (odds ratio [OR]: 3.20; p = 0.002). In the multivariate analysis, independent POPF risk factors were SA ≥120 IU/L, soft pancreatic texture, and high-risk pathology (i.e., duodenal, biliary, ampullary, islet cell, and benign tumors); SA ≥120 IU/L outperformed soft pancreatic texture and high-risk pathology in predicting POPF, respectively (OR: 2.22; p = 0.004 vs. OR: 1.37; p = 0.012 vs. OR: 1.35; p = 0.018). In a subset analysis according to gland texture (soft vs. hard), patients with soft pancreatic texture exhibited a significantly higher incidence of POPF (63.4% vs. 34.4%) and SA ≥120 IU/L (52.1% vs. 27.9%); SA <120 IU/L had a negative predictive value of 82.5% for developing POPF in patients with hard pancreatic texture (OR: 4.28, p = 0.028).

Conclusion

A SA value ≥120 IU/L on the day after PD, which is the strongest predictor for POPF, can be used as a biomarker of the occurrence of POPF. The advantage of SA measurement is that it can contribute to identifying suitable patients for early drain removal.

Acute pancreatitis after pancreatoduodenectomy: clinical outcomes and predictive factors analysis according to the International Study Group of Pancreatic Surgery definition

BACKGROUND

Post-pancreatectomy acute pancreatitis (PPAP) is an increasingly described complication after pancreatic resection. No uniform definition criteria were present in the literature until the recent proposal of the International Study Group of Pancreatic Surgery (ISGPS). Aim of this study is to evaluate the clinical significance of the novel ISGPS definition of PPAP.

METHODS

Patients who underwent pancreatoduodenectomy (PD) between 2006 and 2022 were enrolled. PPAP was defined and graded according to the ISGPS criteria.

RESULTS

Among 520 PDs, 120 (23%)patients developed post-operative hyperamylasemia (POH), while PPAP occurred in 63(12.1%) cases. PPAP occurrence related to a higher rate of more severe complications (48-76.1%vs118-25.8%; p < 0.0001), delayed gastric emptying (DGE) (27-42.9%vd114-24.9%; p = 0.003) and post-operative pancreatic fistula (POPF) (57-90.5%vs186-40.8%; p < 0.0001). When stratified for PPAP severity, grade B and C patients more frequently developed major complications (p < 0.0001), POPF (p < 0.0001), DGE (p = 0.02) and post-operative hemorrhage (p < 0.0001) as compared to POH. At the multivariable analysis, soft pancreatic texture (p = 0.01)and a Wirsung diameter ≤3 mm (p = 0.01) were recognized as prognostic factors for PPAP onset, while a pancreatic duct ≤3 mm was the only feature significantly influencing a more severe course of PPAP (p = 0.01).

CONCLUSION

The ISGPS classification is confirmed as a valuable method for a uniform definition and clinical course evaluation. Further studies in a prospective manner are still needed for a further confirmation.

Early and Sustained Elevation in Serum Pancreatic Amylase Activity: A Novel Predictor of Morbidity After Pancreatic Surgery

OBJECTIVE

To characterize early postoperative serum pancreatic amylase (spAMY) trends after pancreatic resections.

SUMMARY BACKGROUND DATA

A postoperative spAMY elevation is a common finding but uncertainties remain about its meaning and prognostic implications.

METHODS

Analysis of patients who consecutively underwent pancreatectomy from 2016 to 2019. spAMY activity was assessed from postoperative day (POD) 0 to 3. Different patterns of spAMY have been identified based on the spAMY standard range (10-52 U/l).

RESULTS

Three patterns were identified: (#1) spAMY values always < the lower limit of normal/within the reference range /a single increase in spAMY > upper limit of normal at any POD; (#2) Sustained increase in spAMY activity on POD 0 + 1; (#3) Sustained increase in spAMY activity including POD 1 + 2. Shifting through spAMY patterns was associated with increase morbidity (21% in #1 to 68% in #3 at POD 7; log rank < 0.001). Almost all severe complications (at least Clavien-Dindo ≥3) occurred in patients with pattern #3 (15% vs 3% vs 5% in #1 and #2 at POD 7, P = 0.006), without difference considering >3-times or >the spAMY normal limit ( P = 0.85). POPF (9% in #1 vs 48% in #3, P < 0.001) progressively increased across patterns. Pre-operative diabetes (OR 0.19), neoadjuvant therapy (OR 0.22), pancreatic texture (OR 8.8), duct size (OR 0.78), and final histology (OR 2.2) were independent predictors of pattern #3.

CONCLUSIONS

A sustained increase in spAMY activity including POD 1 + 2 (#3) represents an early postoperative predictor of overall and severe early morbidity. An early and dynamic evaluation of spAMY could crucially impact the subsequent clinical course with relevant prognostic implications.

Postoperative serum hyperamylasemia (POH) predicts additional morbidity after pancreatoduodenectomy: It is not all about pancreatic fistula

BACKGROUND

The association between postoperative serum hyperamylasaemia (POH) and morbidity has been hypothesized but rarely explored once occurring with or without (POH-exclusive) a combined postoperative pancreatic fistula (POPF).

METHODS

Analysis of patients who consecutively underwent pancreaticoduodenectomy from 2016 to 2020. POH was defined as serum amylase activity greater than the institutional upper limit of normal (52 U/L), persisting within the first 48 hours postoperatively (postoperative day [POD] 1 and 2).

RESULTS

Among 852 patients, 15.8% developed POH-exclusive. Compared with patients without POH or POPF (64.3%), they showed a significantly higher postoperative burden (Clavien-Dindo ≥II: 52.6% vs 30.8%) with increased rates of bacteraemia (12.6% vs 6%), pleural effusion (13.3% vs 5.3%), postpancreatectomy haemorrhage (13.3% vs 7.5%), postpancreatectomy acute pancreatitis (PPAP) (10.3% vs 0%), and organ site infections (18.5% vs 10.9%; all P < .05). A total of 13.8% experienced POH with POPF leading to the worse outcome. The combined occurrence of POH with POPF led to a shorter median time to morbidity (3 PODs, 95% confidence interval [CI] 2.2-3.7 vs 6 PODs, 95% CI 4.2-8; P < .001) than patients experiencing POPF-exclusive (5.9%). In all, 46.6% of POH patients developed POPF. Body mass index (BMI) (odds ratio [OR] 1.1), male sex (OR 2.1), increased drain fluid amylase on POD 1 (OR 1.001), and increased C-reactive protein (OR 1.01) were independent risk factors for POPF once POH has occurred.

CONCLUSION

POH has relevant postoperative clinical implications, independently from POPF occurrence. Developing POH with POPF leads to an earlier onset of higher postoperative burdens. Once POH is diagnosed, risk factors for additional POPF could identify patients who may benefit from additional surveillance, specific drains protocols, and preventive strategies.

Postpancreatectomy Acute Pancreatitis (PPAP): Definition and Grading From the International Study Group for Pancreatic Surgery (ISGPS)

OBJECTIVE

The ISGPS aimed to develop a universally accepted definition for PPAP for standardized reporting and outcome comparison.

BACKGROUND

PPAP is an increasingly recognized complication after partial pancreatic resections, but its incidence and clinical impact, and even its existence are variable because an internationally accepted consensus definition and grading system are lacking.

METHODS

The ISGPS developed a consensus definition and grading of PPAP with its members after an evidence review and after a series of discussions and multiple revisions from April 2020 to May 2021.

RESULTS

We defined PPAP as an acute inflammatory condition of the pancreatic remnant beginning within the first 3 postoperative days after a partial pancreatic resection. The diagnosis requires (1) a sustained postoperative serum hyperamylasemia (POH) greater than the institutional upper limit of normal for at least the first 48 hours postoperatively, (2) associated with clinically relevant features, and (3) radiologic alterations consistent with PPAP. Three different PPAP grades were defined based on the clinical impact: (1) grade postoperative hyperamylasemia, biochemical changes only; (2) grade B, mild or moderate complications; and (3) grade C, severe life-threatening complications.

DISCUSSIONS

The present definition and grading scale of PPAP, based on biochemical, radiologic, and clinical criteria, are instrumental for a better understanding of PPAP and the spectrum of postoperative complications related to this emerging entity. The current terminology will serve as a reference point for standard assessment and lend itself to developing specific treatments and prevention strategies.

Outcomes comparison of Pancreato-Gastrostomy and Isolated Jejunal Loop Pancreato-Jejunostomy following Pancreato-Duodenectomy in patients with soft pancreas and at moderate-high risk for POPF: a retrospective multicenter experience-based analysis

Following pancreatoduodenectomy (PD), the modality of pancreato-enteric continuity restoration may impact on postoperative pancreatic fistula (POPF) risk. The aim of this study is to compare, among patients with soft pancreas and at moderate/high risk for POPF, the outcomes of PD with Pancreato-Gastrostomy (PG), versus Isolated Jejunal Loop Pancreato-Jejunostomy (IJL-PJ). 193 patients with a Callery Fistula Risk Score (C-FRS) ≥ 3 operated at 3 HPB Units, two performing PG and one IJL-PJ as their preferred anastomotic technique following PD (2009-2019) were included in this study (PG = 123, IJL-PJ = 70). Primary outcomes were POPF, clinically relevant (cr-)postoperative pancreatic hemorrhage (cr-PPH), delayed gastric emptying (cr-DGE), and postoperative major complications and mortality. POPF, cr-PPH, and cr-DGE occurred in 21.8%, 17.6%, and 11.4% of patients, and did not differ significantly between PG (26%, 19.5%, and 10.6%, respectively) and IJL-PJ (17.1%, 14.3%, and 12.9%, respectively; all p > 0.05) patients. Major (Dindo ≥ 3) complication and mortality rates were 26.4% and 3.3%, respectively, and did not differ significantly between PG (29.3% and 3.8%) and IJL-PJ (21.4% and 2.9) patients (p > 0.05). A faster surgical drain and nasogastric tube removal matched a significantly shorter hospitalization among IJL-PJ patients (median LOS: 18 days versus 25 days among PG patients, p < 0.001). In conclusion, IJL-PJ and PG, when performed by surgeons specialized with the concerned anastomotic technique in patients with soft pancreas and moderate/high risk for POPF, have similar results in terms of perioperative mortality and postoperative complications both overall and specific for PD.