Scoring Systems
Currently available scoring systems are shown in Table 4.1. Of these, SIRS and the Bedside Index of Severity in Acute Pancreatitis (BISAP) are the most commonly used and easy to calculate (Table 4.1). Using SIRS criteria to risk‐stratify has advantages over BISAP score because it does not need a chest X‐ray to calculate, and its predictive performance is fairly comparable to BISAP. One in two patients meeting all four criteria for SIRS on admission required transfer to the intensive care unit (ICU) within the first week in a prospective study of 252 patients with AP [23]. Other systems are more cumbersome to calculate (APACHE‐II) or rely on data availability over 48 hours (Ranson). The Harmless Acute Pancreatitis Score aims to identify patients predicted to have a mild course, with high sensitivity and specificity [77,78]. The Pancreatitis Activity Scoring System (PASS) score (see Table 4.1) is the most recently developed score [26]. While originally designed and developed to measure “activity” of AP, the authors also recently tested its predictive function [27].
Table 4.3 Components of computed tomography severity index (CTSI) and modified CTSI.
| CTSI |
| CT grade |
| Normal: 0 points Focal or diffuse enlargement: 1 point Intrinsic change or fat stranding: 2 points One ill‐defined fluid collection: 3 points Multiple fluid collections or gas: 4 points |
| Necrosis score None: 0 points One‐third of pancreas: 2 points Half of pancreas: 4 points More than half of pancreas: 6 points |
| CTSI: Severe >6 points (CT grade + necrosis) |
| Modified CTSI |
| CT grade |
| Normal: 0 points Intrinsic pancreatic abnormalities with/without inflammatory changes in peripancreatic fat: 2 points Pancreatic or peripancreatic fluid collection or peripancreatic fat necrosis: 4 points |
| Necrosis score None: 0 points <30%: 2 points ≥30%: 4 points |
| Extrapancreatic complications (one or more of pleural effusion, ascites, vascular complications, parenchymal complications or gastrointestinal involvement): 2 points |
| Modified CTSI: 0–2 mild; 4–6, moderate; 8–10, severe |
Limitations and Future of Current Scoring Systems and Predictive Markers
While many scoring systems and markers have been recognized and validated, most have only modest accuracy (see Table 4.1) [6]. For practicing clinicians, a severity prediction tool should ideally help change management of a patient based on the prognostic forecast given by the tool. An excellent example of a useful prediction tool is the Model for End‐Stage Liver Disease (MELD) score in patients with cirrhosis [79,80]. MELD score had long been used as a tool to assign priorities for liver transplants, and appropriately risk‐stratify cirrhotic patients undergoing surgery and transjugular intrahepatic portosystemic shunt [79,80]. In contrast, it is not known if any of the existing acute pancreatitis severity prediction tools influence clinical management.
Given the abundance of scoring systems, the next priority is to examine which of the systems directly impact management in real clinical settings. While SIRS score represents the most promising candidate for such practical use, there is lack of data on which is most useful at a clinical practice level. Even in patients with severe pancreatitis as defined by Revised Atlanta Classification, many could be managed on a regular nursing floor if the end‐organs do not require inotropes, mechanical ventilation, or renal replacement therapy. In a large multicenter prospective cohort study, 28% of severe AP patients were managed on the regular nursing floor [81]. In this context, some practical end points could include “impending” need for ICU admission or organ support requirement, need for a full admission from the emergency room, progression to multisystem organ failure, and early death. Almost all existing scoring systems predict in‐hospital mortality. While relevant, they do not consider where in the course of disease the death occurs.
Artificial Intelligence and Biomarkers: the Future?
While an extensive number of cytokines, adipokines and chemokines have been tested, none are routinely available, and none outperform existing clinical scoring systems. The most notable cytokines are IL‐1β, IL‐6, TNF‐α, resistin, visfatin, angiopoietin‐2, and MCP‐1 (see Table 4.2). Table 4.2 is a list of the most promising biomarkers that have been tested for their predictive performance in AP patients, but it is not exhaustive as many other markers have been associated with AP severity [82,83]. In the future, combining biomarkers with clinical scoring systems can be examined to see if prediction performance can be enhanced.
Machine learning algorithms are increasingly tested to aid clinical decision‐making [84–87]. Similarly, in pancreatology, deep learning could be integrated into patients’ electronic medical records to build more accurate models that will continue to refine themselves automatically with time. For example, using machine learning models the risk of “under‐triaging” patients in the emergency room was significantly less than the human triage system [85]. Deep learning could also be employed to incorporate cytokine, chemokine, and adipokine data as well as clinical data to develop better prediction tools that clinicians can use.
References
1 1 Lee PJ, Papachristou GI. New insights into acute pancreatitis. Nat Rev Gastroenterol Hepatol 2019; 16(8):479–496.
2 2 Banks PA, Bollen TL, Dervenis C, et al. Classification of acute pancreatitis – 2012: revision of the Atlanta classification and definitions by international consensus. Gut 2013; 62:102–111.
3 3 Kadiyala V, Suleiman SL, McNabb‐Baltar J, et al. The Atlanta Classification, Revised Atlanta Classification, and Determinant‐Based Classification of acute pancreatitis: which is best at stratifying outcomes? Pancreas 2016; 45:510–515.
4 4 Dellinger EP, Forsmark CE, Layer P, et al. Determinant‐based classification of acute pancreatitis severity: an international multidisciplinary consultation. Ann Surg 2012; 256:875–880.
5 5 Koutroumpakis E, Wu BU, Bakker OJ, et al. Admission hematocrit and rise in blood urea nitrogen at 24 h outperform other laboratory markers in predicting persistent organ failure and pancreatic necrosis in acute pancreatitis: a post hoc analysis of three large prospective databases. Am J Gastroenterol 2015; 110:1707–1716.
6 6 Mounzer R, Langmead CJ, Wu BU, et al. Comparison of existing clinical scoring systems to predict persistent organ failure in patients with acute pancreatitis. Gastroenterology 2012; 142:1476–1482.
7 7 He W‐H, Zhu Y, Zhu Y, et al. Comparison of multifactor scoring systems and single serum markers for the early prediction of the severity of acute pancreatitis. J Gastroenterol Hepatol 2017; 32:1895–1901.
8 8 Cho JH, Kim TN, Chung HH, Kim KH. Comparison of scoring systems in predicting the severity of acute pancreatitis. World J Gastroenterol 2015; 21:2387–2394.
9 9 Ikeura T, Horibe M, Sanui M, et al.