This copy is for personal use only. To order printed copies, contact reprints@rsna.

org
675
G
AS
Revised Atlanta Classification T
R
for OI
N

Acute Pancreatitis: A Pictorial Essay1 T

ES
TI
N
Bryan R. Foster, MD A
Kyle K. Jensen, MD The 2012 revised Atlanta classification is an update of the original L
1992 Atlanta classification, a standardized clinical and radiologic
Gene Bakis, MD IM
Akram M. Shaaban, nomenclature for acute pancreatitis and associated complications
MBBCh Fergus V. Coakley, based on research advances made over the past 2 decades. Acute
MD pancreatitis is now divided into two distinct subtypes, necrotizing
pancreatitis and interstitial edematous pancreatitis (IEP), based on
Abbreviations: ANC = acute necrotic collec- the presence or absence of necrosis, respectively. The revised clas-
tion, APFC = acute peripancreatic fluid collec- sification system also updates confusing and sometimes inaccurate
tion, IEP = interstitial edematous pancreatitis,
WON = walled-off necrosis terminology that was previously used to describe pancreatic and
peripancreatic collections. As such, use of the terms acute pseudo-
RadioGraphics 2016; 36:675–687
cyst and pancreatic abscess is now discouraged. Instead, four distinct
Published online 10.1148/rg.2016150097
collection subtypes are identified on the basis of the presence of
Content Codes: pancreatic necrosis and time elapsed since the onset of
1
From the Department of Diagnostic Radiol- pancreatitis. Acute peripancreatic fluid collections (APFCs) and
ogy, Oregon Health & Science University, 3181 pseudocysts occur in IEP and contain fluid only. Acute necrotic
SW Sam Jackson Park Rd, Portland, OR
97239 (B.R.F., K.K.J., G.B., F.V.C.); and collections (ANCs) and walled-off necrosis (WON) occur only in
Department patients with necrotizing pancreatitis and contain variable amounts
of Radiology, University of Utah, Salt Lake City,
Utah (A.M.S.). Presented as an education ex-
of fluid and necrotic debris. APFCs and ANCs occur within 4
hibit at the 2014 RSNA Annual Meeting. Re- weeks of disease onset. After this time, APFCs or ANCs may either
ceived April 6, 2015; revision requested June 30 resolve or per- sist, developing a mature wall to become a
and received August 10; accepted August 31.
For this journal-based SA-CME activity, the pseudocyst or a WON, respectively. Any collection subtype may
authors, editor, and reviewers have disclosed no become infected and man- ifest as internal gas, though this occurs
relevant relationships. Address
correspondence to
most commonly in necrotic collections. In this review, the authors
B.R.F. (e-mail: [email protected]). present a practical image-rich guide to the revised Atlanta
©
RSNA, 2016
classification system, with the goal of fostering implementation of
the revised system into radiology prac- tice, thereby facilitating
SA-CME LEARNING OBJECTIVES accurate communication among clinicians and reinforcing the
radiologist’s role as a key member of a multidis- ciplinary team in
This article was corrected on March After completing this journal-based SA-CME
treating patients withactivity,
acuteparticipants
pancreatitis.
will be able to:
28, 2019.
Discuss
■©RSNA, 2016 the revised Atlanta classifica- tion system.
• radiographics.rsna.org
■ Distinguish interstitial edematous pan- creatitis from necrotizing pancreatitis
at imaging.
■ Describe the imaging appearance of pancreatitis-associated collections, in-
cluding acute peripancreatic fluid collec- tions, pseudocysts, acute necrotic collec- tions,
and walled-off necrosis.
See www.rsna.org/education/search/RG.

An earlier
incorrect version
of this article
appeared online.
 ng of pathophysiology
necessitated substantial
revision of the classification
system (2). A draft was posted
on the Pancreas Club Web site
in 2008 (3). This draft was
referrenced in several initial
Introduct
publications in the radiology
literature (4–6). After a total of
four Web-based revisions, the
Beginning in revised Atlanta classification
2007, the system was finalized in 2012
Acute and published shortly thereafter,
Pancreatitis with updates to the earlier-cited
Classificatio draft (2).
n Working
Group
polled an
internationa
l cohort of
pancreatic
experts,
includ- ing
11
pancreatic
societies,
with the
goal of
revising the
original
1992 Atlanta
classificatio
n system for
acute
pancreatitis
(1,2).
Although the
original
Atlanta
classificatio
n system
established
common
terms for
acute
pancreatitis
and related
complicatio
ns, the
nomenclatur
e proved to
be
confusing,
leading to
incorrect
use of terms
in both
clinical
practice and
research. In
addition,
advances in
imaging and
in the
understandi
676 May-June 2016 radiographics.rsna.org

recognized by the working group that previously

TEACHING POINTS used terms such as acute pseudocyst and pancreatic
■ In the 1st week after the onset of pancreatitis, imaging
findings correlate poorly with clinical severity, and imaging abscess were confusing and had fallen out of favor,
sensitivity for necrotizing pancreatitis is decreased in the first in addition to their not being entirely descriptive
few days. In addition, any local complications that are of the pathologic condition. Therefore, a specific
detected in the 1st week generally do not necessitate
intervention because treat- ment is based on supportive lexicon was needed to distinguish necrotic from
measures and management of or- gan failure. Initial nonnecrotic collections (7). Use of standardized
imaging is most useful when performed 5–7 days after reporting schemas and application of a well-
hospital admission, when local complications have
developed and pancreatic necrosis (if present) should be defined lexicon to a specific disease are thought
clearly distinguishable. to be important for effective communication
■ In the revised classification system, new definitions were and quality care and are in fact becoming more
cre- ated to clearly stratify acute pancreatitis into two common in everyday radiologic practice (8,9).
subcategories based on imaging findings: IEP and
necrotizing pancreatitis. IEP is more common and
Most important, adoption of such standardized
represents nonnecrotizing inflam- mation of the pancreas. terminology allows the radiologist to be an effec-
The entire pancreas will enhance at contrast-enhanced CT tive member of a multidisciplinary team in the
or MR imaging, with no unenhanced (necrotic) areas,
although enhancement of the gland may be less avid than
diagnosis and treatment of acute pancreatitis.
that of the normal pancreas owing to interstitial edema. IEP The purpose of this article is to serve as an im-
usually manifests with focal or diffuse pancreatic age-rich practical overview of the revised Atlanta
enlargement and is typically surrounded by wispy
peripancre- atic inflammation or a small amount of fluid. In
classification system such that the radiologist
addition, there should be no surrounding peripancreatic will have a working understanding of the system
necrotic collections in IEP, although there may be and can immediately incorporate the lexicon
surrounding fluid-containing collections. Necrotizing
pancreatitis accounts for 5%–10% of cases of acute
into clinical practice. To that end, we focus on
pancreatitis. It is important to understand that necrosis may key imaging characteristics that help differenti-
involve either the pancreatic parenchyma or the ate collections that often have a similar appear-
peripancreatic tissues and in both cases is termed
ance, describe various infectious complications,
necrotizing pancreatitis. There are three subtypes of
necrotizing pancreati- tis; the subtypes are based on the and discuss imaging pitfalls. In addition, we will
anatomic area of necrotic involvement: (a) pancreatic only, propose a reporting schema to facilitate
(b) peripancreatic only, and accurate communication.
(c) combined pancreatic and peripancreatic.
■ The revised Atlanta classification makes an important
distinc- tion between collections that contain purely fluid Acute Pancreatitis: Over-
(those en- countered in nonnecrotizing pancreatitis and IEP) view and New Diagnostic Criteria
and collec- tions that contain necrotic debris in addition Acute pancreatitis is an acute inflammatory
to fluid (those encountered in necrotizing pancreatitis).
The terms acute pseudocyst and pancreatic abscess have condition, with a range of severity as well as
been abandoned. Similarly, the use of the term pseudocyst vari- ous local and systemic complications.
in radiology reporting to describe any pancreatitis-related Gallstones and alcohol are the first and second
collection is misleading to treating physicians, since the term
implies that these collections always contain purely fluid, most com- mon causes of acute pancreatitis,
which is not the case in necrotic collections. Instead, the respectively, and additional variants occur when
revised classification includes new defi- nitions that more patients are stratified by sex. In 2009, acute
accurately describe the various types of col- lections
encountered: APFC, pseudocyst, ANC, and WON. The pancreatitis was the most common cause of
important distinctions for classifying collections correctly are hospital admission
the time course (4 weeks or >4 weeks from onset of pain) for gastrointestinal disorders in the United States,
and the presence or absence of necrosis at imaging.
with approXimately 275 000 admissions, nearly
■ If an APFC has not resolved within 4 weeks, it becomes
more organized, with development of a capsule that
double the number in the previous decade. This
manifests as an enhancing wall at contrast-enhanced CT. At increase is thought to be secondary not only to
this point, the collection is referred to as a pseudocyst, and, nationwide increases in obesity and the incidence
since there is no necrosis, it should contain only fluid with
no nonliquefied com- ponents. If there is even a small area of
of gallstones, but also to more sensitive and more
fat or soft-tissue attenu- ation in an otherwise fluid- frequently used laboratory testing (10). A majority
attenuation collection, the diagnosis is not pseudocyst but of patients have mild acute pancreatitis, which car-
WON.
ries essentially no risk of mortality. In the subset of
■ Any collection can be sterile or infected, although infection
oc- curs far more frequently in necrotic collections. Clinically,
patients with organ failure (severe disease) or in-
infec- tion is suspected in a previously stable patient who fected necrosis, however, the mortality rate reaches
30% (2,11). The revised Atlanta classification
requires that two or more of the following crite-
ria be met for the diagnosis of acute pancreatitis:
One major component of the revised
(a) abdominal pain suggestive of pancreatitis, (b)
classifi- cation system that is of particular
serum amylase or lipase level greater than three
importance to radiologists is the manner in
times the upper normal value, or (c) characteristic
which pancreatitis- associated collections (local
imaging findings (2). Contrast material–enhanced
complications) are described and named (Table
CT is most commonly used to fulfill the radiologic
1). It was well
 Table 1: Pancreatic and Peripancreatic Collections
Time after OnsetPancreatitis

Collection of Pain (wk) SubcategoryLocation

Imaging Features
APFC 4 IEPExtrapancreatic
Homogeneous, fluid attenuation, con- forms to retroperitoneal structures, no wall

ANC 4 Necrotizing Intra- and/or extra- Inhomogeneous*, nonliquefied com-

pancreatitis pancreatic ponents†, no wall
Pseudocyst >4 IEP EXtrapancreatic‡ Homogeneous, fluid filled, circum-
scribed, encapsulated with wall
WON >4 Necrotizing Intra- and/or extra- Inhomogeneous, nonliquefied compo-
pancreatitis pancreatic nents, encapsulated with wall

erences 2–4.
llection may become infected. ANC = acute necrotic collection, APFC = acute peripancreatic fluid collection, IEP = interstitial edematous pancreatitis, WON = walled
ay be homogeneous; follow-up computed tomography (CT) performed in 2nd week may help clarify status.
-appearing components or fat globules within fluid.
tent pancreatic leak or disconnected duct may lead to intrapancreatic pseudocyst.

criterion, but magnetic resonance (MR) imaging patients as having either severe or mild
is also appropriate. Although many patients will pancreati- tis on the basis of the presence or
meet the criteria for acute pancreatitis on the basis absence of or- gan failure, respectively (1).
of symptoms and laboratory results alone and may However, emerging evidence indicated that a
not require imaging initially, imaging may be per- large subset of patients with local complications
formed early in the disease course when the cause experienced substantial morbidity but little
of the disease is unclear, to look for causative mortality (11). Therefore, a third category was
factors such as choledocholithiasis and pancreatic added to the new classifica- tion—moderately
cancer. Imaging for the diagnosis of pancreatitis severe acute pancreatitis—to describe this
is also appropriate when abdominal pain suggests patient group (2). Organ failure and local
pancreatitis but the amylase or lipase level is not complications are not seen in patients with mild
elevated to the threshold value, which is often the pancreatitis, who are usually discharged within
case at delayed presentation (12). The onset of the 1st week, with very low mortality
pancreatitis is considered to coincide with the 1st (13). These patients rarely require CT for local
day of pain, not the day on which the patient pres- complications, and imaging may be useful only
ents for care or the day of hospital admission in assessing the cause of pancreatitis (eg, ultraso-
(2). nography [US] or MR
cholangiopancreatography for
Phases of Acute Pancreatitis choledocholithiasis) (12).
In pathophysiologic terms, acute pancreatitis is Moderately severe acute pancreatitis manifests
divided into early and late phases. The early phase in patients with transient organ failure lasting
occurs in the 1st week after onset, with the disease less than 48 hours and/or local or systemic
manifesting as a systemic inflammatory response. complica- tions. Systemic complications are
At this time, clinical severity and treatment are generally co- morbidities exacerbated by
mainly determined on the basis of type and degree pancreatitis, such as acute kidney injury in the
of organ failure. The late phase, which generally setting of chronic renal failure. Local
starts in the 2nd week and can last for weeks to complications include a variety of pancreatic and
months, occurs only in patients with moderately peripancreatic collections. Such collections
severe or severe pancreatitis, as defined by persis- generally develop in the 2nd week (the late
tent organ failure and by local complications (2). phase of pancreatitis) and are clinically
suspected in patients with unremitting or recur-
Grading of rent pain, a secondary peak in pancreatic enzyme
Severity of Acute Pancreatitis levels, worsening organ dysfunction, or sepsis
To improve the stratification of patients at the (2). These symptoms should prompt imaging
time of presentation, the pancreatitis severity studies such as (in order of preference) contrast-
scale was updated in the revised Atlanta clas- enhanced CT, contrast-enhanced MR imaging,
sification. The original classification categorized or unenhanced MR imaging (12).
 Severe disease is characterized by organ failure that
persists for more than 48 hours. Because
 Table 2: Modified Marshall Scoring System

Organ System Score 0 Score 1 Score 2 Score 3 Score 4

Respiratory* >400 301–400 201–300 101–200 100
Renal: serum creati- 1.4 1.5–1.8 1.9–3.5 3.6–4.9 5
nine (mg/dL)
Cardiovascular: sys->90 tolic blood<90,
pressure
responding to fluid resusci- tation
<90, not respond-<90 with pH <90 with pH
(mm Hg) ing to fluid resus- citation <7.3 <7.2

Sources.—References 2,4,14.
Note.—A score of 2 or higher indicates organ failure, with transient failure lasting less than 48 hours and persis- tent failure lasting more than 48 hours.
*Partial pressure of oxygen/fraction of inspired oxygen, or PaO2/FIO2.

Figure 1. IEP in a 28-year-old man with

alcohol-related pancreatitis. Axial con-
trast-enhanced CT image shows wispy
peripancreatic inflammation (arrows)
with normal pancreatic enhancement and
no collections.

organ failure plays a key role in determining pancreas

disease severity, an accurate definition is essential
for clinical management of acute pancreatitis.
The modified Marshall scoring system (Table
2) is endorsed in the revised Atlanta
classification as the primary method for
determining organ failure. The modified
Marshall scoring system incorporates
measurements from the respiratory,
cardiovascular, and renal systems, with a score
of 2 or higher for any system indicating organ
failure (2,14). In the 1st week after the onset of
pancreatitis, imaging findings correlate poorly
with clinical severity, and imaging sensitivity for
necrotizing pancreatitis is decreased in the first
few days (2). In addition, any local complications
detected in the 1st week generally do not neces-
sitate intervention because treatment is based on
supportive measures and management of organ
failure (15). Initial imaging is most useful when
performed 5–7 days after hospital admission,
when local complications have developed and
pancreatic necrosis (if present) should be clearly
distinguishable (2).

IEP versus
Necrotizing Pancreatitis
In the revised classification system, new defini-
tions were created to clearly stratify acute pan-
creatitis into two subcategories based on
imaging findings: IEP and necrotizing
pancreatitis. IEP
is more common and represents nonnecrotiz-
ing inflammation of the pancreas. The entire
pancreas will enhance at contrast-enhanced
CT or MR imaging, with no unenhanced
(necrotic) areas, although enhancement of the
gland may be less avid than that of the normal
owing to interstitial edema. IEP usually manifests with
focal or diffuse pancreatic enlargement and is typically
surrounded by wispy peripancreatic inflammation or a
small amount of fluid (Fig 1). In addition, there should
be no surrounding peri- pancreatic necrotic collections
in IEP, although there may be surrounding fluid-
containing collec- tions (Fig 2).
Necrotizing pancreatitis accounts for 5%–10% of
cases of acute pancreatitis (2). It is important to
understand that necrosis may involve either the
pancreatic parenchyma or the peripancreatic tissues
and in both cases is termed necrotizing pancreatitis.
There are three subtypes of necro- tizing pancreatitis;
the subtypes are based on
the anatomic area of necrotic involvement: (a)
pancreatic only, (b) peripancreatic only, and (c)
combined pancreatic and peripancreatic. The latter
subtype is the most common, accounting for 75% of
cases. The combined subtype dem- onstrates
nonenhancing pancreatic parenchyma, as well as
nonenhancing heterogeneous peripan- creatic
collections, and typically accumulating in the lesser
sac and anterior pararenal space (Fig 3a).
Peripancreatic necrosis alone, in which the pancreas
enhances normally but the peripan- creatic tissues
show necrosis, with collections containing variable
amounts of fluid and nonliq-
 Figure 2. IEP in a 43-year-old
man. Axial contrast-enhanced
CT image shows peripan-
creatic inflammation (black
arrow) and a homogeneous
fluid-attenuation collection
in the left anterior pararenal
space (white arrow), a finding
that is consistent with APFC.

lack arrow), and a large ANC in the lesser sac (white arrows). (b) Image of peripancreatic necrosis alone in an 18-year-old man shows a large, complex, heterogeneou

contrast-enhanced CT performed 5–7 days later

is more accurate for the diagnosis of necrotizing
uefied components, occurs in 20% of cases (Fig pancreatitis (2). In general, routine contrast-
3b). Pancreatic necrosis alone is the least com- enhanced CT is adequate for diagnosis, and
mon subtype, occurring in 5% of cases, and lacks pancreatic protocol multiphase imaging (arterial
peripancreatic collections (Fig 3c). When imaging and portal phase) is typically unnecessary. If renal
is performed within the first few days of disease failure prevents administration of intravenous
onset, necrosis may not be detected because the contrast material, nonenhanced MR imaging is
pancreas can appear edematous and globally preferred to nonenhanced CT, although limited
or focally hypoenhancing and, as such, may be availability and the difficulty of imaging criti-
indistinguishable from IEP. In these cases, repeat cally ill patients are significant limitations of MR
 Figure 4. IEP with APFCs on axial contrast-enhanced CT images in a 49-year-old man. (a) One day after
the onset of pain, IEP is seen with a small, homogeneous, fluid-attenuation collection in the left anterior
pararenal space (arrow). The pancreas demonstrates normal enhancement. (b–d) At 10 days after onset of pain,
multiple homogeneous fluid collections (*) can be seen in the lesser sac (b) and peripancreatic (c) and bilateral
anterior pararenal spaces (d), findings consistent with APFCs. Arrow on c = normally enhancing pancreatic
tail.

imaging, and nonenhanced CT may be easier weeks from onset of pain) and the presence or
in practice (12). absence of necrosis at imaging (Table 1) (2).

Pancreatic and APFC and Pseudocyst

Peripancreatic Collections APFCs occur during the first 4 weeks and
The revised Atlanta classification makes an are present only in patients with IEP. Because
important distinction between collections that the pathogenesis involves inflammation
contain purely fluid (those encountered in IEP) without
and collections that contain necrotic debris in necrosis, APFCs contain only fluid and are visu-
addition to fluid (those encountered in necrotiz- alized as homogeneous fluid-attenuation collec-
ing pancreatitis). The terms acute pseudocyst and tions that lack a wall and tend to conform to the
pancreatic abscess have been abandoned. Similarly, retroperitoneal spaces (Fig 4). APFCs are always
the use of the term pseudocyst in radiology report- peripancreatic in location. If a similar-appearing
ing to describe any pancreatitis-related collection is seen within the pancreatic paren-
collection is misleading to treating physicians, chyma, it is by definition an ANC, and the diag-
since the term implies that these collections nosis is no longer IEP but necrotizing
always contain purely fluid, which is not the case pancreatitis (2). Most APFCs resolve
in necrotic collections. Instead, the revised spontaneously, and drainage should be not be
classification includes new definitions that more performed because of the risk of infecting an
accurately describe the various types of otherwise sterile collec- tion (16). Just as some
collections encoun- tered: APFC, pseudocyst, cases of IEP and necrotiz- ing pancreatitis can be
ANC, and WON. The important distinctions for difficult to distinguish at contrast-enhanced CT
classifying collections correctly are the time during the 1st week, it may also be difficult to
course (4 weeks or >4 distinguish between an APFC and an ANC. In
general, repeat imaging at
Figure 5. Pseudocyst in a 36-year-old man. Axial CT
image obtained 6 weeks after onset of gallstone
pancre- atitis shows a large homogeneous fluid
collection in the lesser sac (*) with no nonliquefied Figure 7. Pseudocyst in a 55-year-old woman with
components and a thick enhancing wall (arrows), disconnected duct syndrome. Coronal two-
findings that are consis- tent with pseudocyst. dimensional MR cholangiopancreatographic image
shows a large bilobed peripancreatic and
intrapancreatic pseudocyst (*) that communicates
with the pancreatic tail duct (ar- row), which is mildly
dilated. Note the disconnection of the tail duct from
the ampulla owing to segmental pancreatic body
necrosis.

tography owing to superior contrast resolution

(Fig 7) (2). These ductal connections can seal off,
often leading to cyst resolution. Pseudocysts de-
velop in fewer than 10% of cases of IEP (16). A
pseudocyst is typically peripancreatic in location,
although it can, on rare occasions, be intrapan-
creatic in cases of prior necrosectomy with a
persistent pancreatic duct leak into the necrosec-
tomy bed. In such cases, a pseudocyst forms be-
cause of disconnected duct syndrome, whereby a
Figure 6. Pseudocyst in a 48-year-old woman. Coro- viable pancreatic tail remains after necrosectomy
nal T2-weighted MR image obtained 8 weeks after or pancreatic body necrosis, with the pseudocyst
onset of pancreatitis shows a large homogeneous col-
forming as a result pancreatic juice leakage from
lection in the lesser sac ( *) with uniformly hyperintense
fluid signal and no necrotic components, findings the disconnected duct (Fig 7) (2,17).
that are consistent with pseudocyst.
ANC and WON
ANCs are present within the first 4 weeks of
2 weeks better shows the internal symptom onset and are poorly organized necrotic
heterogeneity of an ANC (2). collections that occur only in necrotizing pancre-
If an APFC has not resolved after 4 weeks, it atitis. ANCs are often found in the lesser sac and
becomes more organized and develops a capsule pararenal spaces and may extend into the
that manifests as an enhancing wall at contrast- pancreas within areas of parenchymal necrosis.
enhanced CT. At this point, the collection is They are often multiple, with a loculated
referred to as a pseudocyst (Fig 5), and, since appearance, and may extend inferiorly as far as the
there is no necrosis, it should contain only fluid pelvic sidewalls.
with no nonliquefied components. If there is even ANCs typically demonstrate a variable amount
a small area of fat or soft-tissue attenuation in an of fluid and can be distinguished from APFCs
otherwise fluid-attenuation collection, the diag- by the presence of nonliquefied debris, such as
nosis is not pseudocyst but WON. At MR imag- solid-appearing components or fat globules within
ing, pseudocysts are uniformly hyperintense on the fluid (Figs 3, 4, 8). When intravenous contrast
T2-weighted images, with no solid components material is contraindicated (eg, in a patient with
or debris in the fluid (Fig 6). Pseudocysts may acute renal failure), the presence of fat attenuation
have a connection to the pancreatic ductal sys- within a collection at nonenhanced CT is help-
tem, which is best seen at MR cholangiopancrea- ful for identifying necrosis and diagnosing ANCs
(Fig 9). In the early phase of pancreatitis, differ-
entiating between an APFC and an ANC can be
ous fluid attenuation, and contain nonliquefied debris, including fat globules (arrows). (b) On day 15, the collection (*) has a more organized appearance, and it is ea

Figure 9. ANC in a 41-year-old woman with

acute pancreatitis. Nonenhanced CT image
obtained 9 days after symptom onset
shows poor contrast resolution between
pancreas and peripancreatic collections,
such that ne- crosis in the pancreas is
uncertain. However, the presence of small
fat globules (arrows) in the peripancreatic
collections and the overall heterogeneity are
consistent with necrotizing pancreatitis with
an ANC.

difficult, and the diagnosis of necrosis may be un-

certain. Imaging in the 2nd week is usually helpful
for distinguishing an APFC from an ANC. Any
peripancreatic collection associated with known
pancreatic parenchymal necrosis should be
termed an ANC, even if it is homogeneous and
contains no nonliquefied debris (2). necrotic fat and/or pancreatic tissue, which are
After 4 weeks, an ANC typically develops a well demonstrated at both contrast-enhanced CT
thick mature wall, at which point the collection and MR imaging as nonliquefied debris within
is referred to as WON. Like pseudocysts, a the fluid (Figs 3c, 10). WON may be confined to
WON contains fluid and has a thick enhancing the pancreatic parenchyma but more commonly
wall. occurs in the peripancreatic space and can often
Unlike pseudocysts, however, WON contains
Figure 10. Evolution of necrotizing pancreatitis during 2 months in a 48-year-old woman. (a) Week 1: Axial contrast-
enhanced CT image shows a necrotic pancreatic neck (*). (b) Week 2: Axial contrast-enhanced CT image shows a
new heterogeneous necrotic peripancreatic collection (arrow) that is inseparable from the pancreatic necrotic
collection (*); both findings are consistent with an ANC. (c) Week 3: Axial nonenhanced T2-weighted MR image
better shows the contents of the ANC (*), including hyperintense fluid and nonliquefied debris, including necrotic
pancreatic neck and body (arrow). Note the developing partial wall. (d) Week 5: Axial contrast-enhanced CT
image shows maturation of the wall and a more round appearance of what is now referred to as WON (*). (e)
Week 6: Endoscopic image from the gastric body, obtained during cystogastrostomy and debridement, shows
necrotic debris (arrow) from the WON.
(f) Week 7: Postprocedure axial contrast-enhanced CT image shows successful decompression of the WON.
Multiple double-pigtail stents traversing the cystogastrostomy (arrows) can be seen.

occur in both locations, with a coalescent collec- older than 4 weeks. Therefore, MR imaging is a
tion extending from the lesser sac into a portion valuable alternative to contrast-enhanced CT for
of parenchyma (Fig 10) (2). There is evidence planning interventions because it allows
that MR imaging outperforms CT, with higher determi- nation of the amount of necrotic debris
interreader agreement, in the assessment of the that must be removed by means of more
ratio of fluid to necrotic debris in collections aggressive inter- ventions (Fig 11) (18).
 Figure 11. Necrotizing pancreatitis in a 47-year-old man in week 16 of persistent symptoms. (a)
Axial nonenhanced CT image shows pancreatic WON involving nearly the entire pancreas (arrows);
at least half of its volume appears to contain fluid. (b) Subsequent axial fat-saturated T2-weighted
MR image shows the WON (arrows) containing mostly nonliquefied debris and pancreatic necrosis,
with little fluid present. Such a collection would respond poorly to percutaneous or endoscopic
drainage and would require more aggressive therapy.

Infection
and Local Complications debated. At some institutions, fine-needle aspira-
Any collection can be sterile or infected, although tion has fallen out of favor in recent years, partly
infection occurs far more frequently in necrotic due to a shift in preference toward early conser-
collections (2). Clinically, infection is suspected vative management with percutaneous drainage,
in a previously stable patient who experiences which may delay or even obviate surgical inter-
decompensation with signs of infection (19). vention. If percutaneous drainage is performed,
The only imaging finding of an infected collec- culture of the fluid can be performed at the
tion is the presence of gas within the collection. same time (19,22). Another argument against
Wall enhancement is not a reliable indicator of fine- needle aspiration is that one must consider
infection, since it is invariably present in mature the substantial false-negative (sampling too
collections (pseudocyst and WON). An infected early) and false-positive (contamination) results
pseudocyst still lacks solid components that, if of the procedure (25% and 15%, respectively).
present, should instead lead to the diagnosis of At some institutions, fine-needle aspiration is
infected WON. The gas often appears as multiple thought to be helpful when clinical signs and
small bubbles scattered throughout the collec- imaging find- ings are confusing or complicated
tion owing to the complex nature of necrotic and the ben- efits of diagnosis of infection
collections (Figs 8, 12) (2). Infected collections outweigh the risk of iatrogenic introduction of
can also manifest with gas bubbles due to a infection (22).
pancreatic-enteric fistula, which can occasionally In addition to infection, vascular complica-
be seen when necrotic collections erode through tions are common, occurring in a quarter of
the bowel wall, most commonly in the colon and patients with acute pancreatitis, and can cause
duodenum (Fig 13) (20). In one series, an enteric substantial morbidity and mortality. Two separate
fistula occurred in 4% of patients hospitalized pathophysiologic processes lead to vascular com-
for acute pancreatitis (21). Gas within the plications. First, inflammatory reactions can lead
pancre- atic duct can also mimic gas within a to splenic vein thrombosis, the most common
pancreatic collection but generally has an vascular complication (23). Second, pancreatic
identifiable linear distribution and typically enzymes can cause vessel erosion and lead to ei-
occurs in the clinical context of recent ther spontaneous arterial hemorrhage or pseu-
endoscopic pancreatography or pancreatic duct doaneurysm of (in order of decreasing
stent placement (Fig 14). frequency) the splenic, gastroduodenal, and
The use of imaging-guided fluid collection pancreaticoduo- denal arteries (24). A detailed
aspiration or necrotic tissue fine-needle discussion of these complications is beyond the
aspira- tion to help diagnose infection prior to scope of this article, as their evaluation is not
invasive therapeutic necrosectomy has both specified under the revised Atlanta classification.
advantages and disadvantages, and the topic is We direct the reader to other excellent
still widely publications that describe these complications
(25).
Figure 12. Necrotizing pancreatitis in a 37-year-old woman. (a) Axial contrast-enhanced CT image
obtained in week 3 shows the pancreatic tail (*) and a peripancreatic ANC containing nonliquefied debris
with foci of fat attenuation (arrows). (b) Axial contrast-enhanced CT image obtained in week 6 because the
patient experi- enced decompensation and was readmitted shows organization of the collection (*) with
multiple new foci of gas, findings that are consistent with infected WON.

Figure 14. Necrotizing pancreatitis at week 3 in a

59-year-old man. Axial contrast-enhanced CT image
Figure 13. Necrotizing pancreatitis in a 74-year-old obtained 3 weeks after disease onset shows necrosis
woman. Axial contrast-enhanced CT image obtained of pancreatic neck, body, and tail (*), as well as
at 5 weeks shows peripancreatic WON with multiple peripan- creatic necrosis. A small focus of gas (white
foci of gas (*). A large fistula is seen to the distal arrow) in the pancreatic duct in the necrotic
transverse colon and contains gas and fluid (arrows), a pancreatic body is secondary to recent stent
finding that explains the development of gas in the placement in the duct (black arrow) and thus does
WON. not represent an infected ANC.

Management Implications recent studies in which open necrosectomy has

Various interventions are available for managing been reanalyzed have found low mortality rates
local complications of acute pancreatitis, with a and results comparable to those achieved with
range of invasiveness and, therefore, morbidity; minimally invasive techniques (27). While exact
these include percutaneous drainage, endoscopic treatment regimens vary among institutions,
cystogastrostomy, endoscopic débridement (Fig collections generally do not warrant
10), and surgical necrosectomy (19). Newer ap- intervention unless there are persistent
proaches have focused on a combined “step-up” symptoms, inability to maintain nutrition, or
approach in which percutaneous drainage cath- signs of infection (28). By implementing the
eters are placed in necrotic collections, followed revised Atlanta classifica- tion, the radiologist is
by minimally invasive débridements along the able to help the care team prescribe the
catheter tract if the patient fails to improve within appropriate therapy according to
72 hours (26). Although these minimally invasive
the type of collection. For instance, in the case
techniques have been replacing open surgical
of a lesser sac pseudocyst necessitating drainage,
débridement, the previous standard of care, more
the gastroenterologist may create a
cystogastrostomy with the use of endoscopic US, a
procedure with
Figure 15. Decision tree for re-
porting according to the revised
Atlanta classification.

a high success rate (29). However, if the collec- team members recognize the expertise
tion represents WON, adequate drainage may and the standardization of terminology.
not be achieved with this approach owing to
the nonliquefied components and, therefore,
neces-
sitating more aggressive débridement through
the endoscope as a first-line therapy (Fig 11).

Suggestions for Reporting

When reviewing a patient’s history, it is important
to note the time elapsed since the 1st day of ab-
dominal pain, as this defines the time course used
to stratify local complications. A statement about
the presence or absence of necrosis should be
made, and the location (peripancreatic,
pancreatic, or both) and amount of necrotic
gland (<30% or
>30%) should be documented. Local complica-
tions should be described in terms of location
(lesser sac, anterior pararenal space, transverse
mesocolon, etc), size, appearance, and presence or
absence of a mature wall. Specifically, the
contents of local complications should be
described either as homogeneous with fluid
attenuation (APFC
or pseudocyst) or as having nonliquefied necrotic
components (ANC or WON). Finally, the col-
lection should be specifically named according
to the revised Atlanta classification lexicon,
with the subtype of pancreatitis and the
number of weeks since the onset of abdominal
pain (<4 or
4 weeks) taken into account (Fig 15). We sug-
gest that the findings be succinctly summarized
in the impression section of the report (eg, IEP or
necrotizing pancreatitis with type and location of
collection) and a statement that the revised Atlanta
classification was used so that multidisciplinary
 Conclusion
Incorporation of the revised Atlanta classification system
into everyday practice updates and stan- dardizes
terminology, which facilitates accurate documentation of
the range of imaging findings in acute pancreatitis. It is
important to remember that pancreatitis-related
collections are not always
fluid filled, and evaluation for nonliquefied compo- nents is
essential for differentiating collections that contain only
fluid (APFCs and pseudocysts) from those that contain
necrotic nonliquefied debris (ANCs and WON). In
general, imaging findings combined with the time course
of the disease allow clear differentiation between the
collections and en- able stratification among different
treatment plans, facilitating the radiologist’s seamless
integration into a multidisciplinary team of
gastroenterologists, intensivists, interventionalists, and
surgeons.

References
1. Bradley EL 3rd. A clinically based classification system for acute
pancreatitis: summary of the International Symposium on Acute
Pancreatitis, Atlanta, Ga, September 11–13, 1992. Arch Surg
1993;128(5):586–590.
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(1): 102–111.
3. The Pancreas Club, Inc. Revision of the Atlanta classifica- tion of
acute pancreatitis. Available at: http://pancreasclub
.com/wp-content/uploads/2011/11/AtlantaClassification.pdf.
Published April 9, 2008. Accessed March 28, 2015.
4. Thoeni RF. The revised Atlanta classification of acute pan-
creatitis: its importance for the radiologist and its effect on
treatment. Radiology 2012;262(3):751–764.
5. Bollen TL. Imaging of acute pancreatitis: update of the revised
Atlanta classification. Radiol Clin North Am 2012;50(3): 429–
445.
6. Sheu Y, Furlan A, Almusa O, Papachristou G, Bae KT. The
revised Atlanta classification for acute pancreatitis: a CT imag- ing
guide for radiologists. Emerg Radiol 2012;19(3):237–243.
 7. Bollen TL, Besselink MG, van Santvoort HC, Gooszen
study using the revised Atlanta classification 2012.
HG, van Leeuwen MS. Toward an update of the Atlanta
Abdom Imaging 2015;40(6):1608–1616.
classifi- cation on acute pancreatitis: review of new and
19. Freeman ML, Werner J, van Santvoort HC, etal.
abandoned terms. Pancreas 2007;35(2):107–113.
Interventions for necrotizing pancreatitis: summary of a
8. Dunnick NR, Langlotz CP. The radiology report of the
multidisciplinary consensus conference. Pancreas
future: a summary of the 2007 Intersociety Conference. J
2012;41(8):1176–1194.
Am Coll Radiol 2008;5(5):626–629.
20. Tüney D, Altun E, Barlas A, Yegen C. Pancreatico-colonic
9. Ellenbogen PH. BI-RADS: revised and replicated. J Am
fistula after acute necrotizing pancreatitis: diagnosis with
Coll Radiol 2014;11(1):2.
spiral CT using rectal water soluble contrast media. JOP
10. Yadav D, Lowenfels AB. The epidemiology of pancreatitis
2008;9(1):26–29.
and pancreatic cancer. Gastroenterology 2013;144(6):
21. Kochhar R, Jain K, Gupta V, et al. Fistulization in the GI
1252–1261.
tract in acute pancreatitis. Gastrointest Endosc
11. Petrov MS, Windsor JA. Classification of the severity of
2012;75(2): 436–440.
acute pancreatitis: how many categories make sense? Am
22. van Baal MC, Bollen TL, Bakker OJ, et al. The role of
J Gastroenterol 2010;105(1):74–76.
routine fine-needle aspiration in the diagnosis of infected
12. Baker ME, Nelson RC, Rosen MP, et al. ACR Appropriate-
necrotizing pancreatitis. Surgery 2014;155(3):442–448.
ness Criteria acute pancreatitis. Ultrasound Q
23. Türkvatan A, Erden A, Türkoğlu MA, Seçil M, Yüce G.
2014;30(4): 267–273.
Imaging of acute pancreatitis and its complications. II.
13. Singh VK, Bollen TL, Wu BU, et al. An assessment of
Complications of acute pancreatitis. Diagn Interv
the severity of interstitial pancreatitis. Clin Gastroenterol
Imaging 2015;96(2):161–169.
Hepatol 2011;9(12):1098–1103.
24. Memiş A, Parildar M. Interventional radiological treatment
14. Marshall JC, Cook DJ, Christou NV, Bernard GR,
in complications of pancreatitis. Eur J Radiol
Sprung CL, Sibbald WJ. Multiple organ dysfunction score:
2002;43(3): 219–228.
a reliable descriptor of a complex clinical outcome. Crit
25. Shyu JY, Sainani NI, Sahni VA, et al. Necrotizing
Care Med 1995;23(10):1638–1652.
pancre- atitis: diagnosis, imaging, and intervention.
15. Spanier BW, Nio Y, van der Hulst RW, Tuynman HA,
RadioGraphics 2014;34(5):1218–1239.
Dijkgraaf MG, Bruno MJ. Practice and yield of early CT
26. van Santvoort HC, Besselink MG, Bakker OJ, et al. A step-
scan in acute pancreatitis: a Dutch observational
up approach or open necrosectomy for necrotizing
multicenter study. Pancreatology 2010;10(2-3):222–228.
pancreatitis. N Engl J Med 2010;362(16):1491–1502.
16. Lenhart DK, Balthazar EJ. MDCT of acute mild
27. Madenci AL, Michailidou M, Chiou G, Thabet A,
(nonnecro- tizing) pancreatitis: abdominal complications
Fernández- del Castillo C, Fagenholz PJ. A contemporary
and fate of fluid collections. AJR Am J Roentgenol
series of patients undergoing open debridement for
2008;190(3):643–649.
necrotizing pancreatitis. Am J Surg 2014;208(3):324–
17. Sandrasegaran K, Tann M, Jennings SG, et al.
331.
Disconnec- tion of the pancreatic duct: an important but
28. Wu BU, Banks PA. Clinical management of patients with
overlooked complication of severe acute pancreatitis.
acute pancreatitis. Gastroenterology 2013;144(6):1272–
RadioGraphics 2007;27(5):1389–1400.
1281.
18. Kamal A, Singh VK, Akshintala VS, et al. CT and MRI
29. Varadarajulu S, Christein JD, Tamhane A, Drelichman
assessment of symptomatic organized pancreatic fluid
ER, WilcoX CM. Prospective randomized trial compar-
collec- tions and pancreatic duct disruption: an interreader
ing EUS and EGD for transmural drainage of pancreatic
variability
pseudocysts (with videos). Gastrointest Endosc 2008;68(6):
1102–1111.
This journal-based SA-CME activity has been approved for AMA PRA Category 1 CreditTM. See www.rsna.org/education/search/RG
 ER
1
R
A
T

Errata
May-June 2019 • Volume 39 • Number 3

Originally published in: Originally published in:

RadioGraphics 2018; 38 (6):1609-1616 • https:// RadioGraphics 2016; 36(3):675–687 • https://doi.
doi.org/10.1148/rg.2018180031 org/10.1148/rg.2016150097
Rules and Regulations Relating to Roles of Non- Revised Atlanta Classification for Acute Pancreatitis:
physician Providers in Radiology Practices A Pictorial Essay
C. Matthew Hawkins Bryan R. Foster, Kyle K. Jensen, Gene Bakis, Akram
Erratum in: M. Shaaban, Fergus V. Coakley
RadioGraphics 2019:39(3):XXXX • https://doi. Erratum in:
org/10.1148/rg.2019XXXXX RadioGraphics 2019:39(3):XXXX •
Second paragraph, second column, p 1610: The https://doi. org/10.1148/rg.2019XXXXX
sixth sentence should read as follows: “The authors of a Note, Table 1: The abbreviation for IEP should read as
review article emphasized that employing RAs increases follows: “IEP = interstitial edematous pancreatitis.”
physician supervision in nearly all clinical settings and
that billing for services provided by RAs without ap- Originally published in:
propriate supervision violates the False Claims Act.” RadioGraphics 2012; 32(2):437–451 • https://doi.
org/10.1148/rg.322115032
Originally published in: Peritoneal and Retroperitoneal Anatomy and Its
RadioGraphics 2018; 32(2):1421–1440 • https://doi. Relevance for CrossSectional Imaging
org/10.1148/rg.2018180041 Temel Tirkes, Kumaresan Sandrasegaran, Aashish
Radiation Dose Reduction at Pediatric CT: Use of A. Patel, Margaret A. Hollar, Juan G. Tejada, Mark
Low Tube Voltage and Iterative Reconstruction Tann, Fatih M. Akisik, John C. Lappas
Yasunori Nagayama, Seitaro Oda, Takeshi Nakaura, Erratum in:
Akinori Tsuji, Joji Urata, Mitsuhiro Furusawa, RadioGraphics 2019:39(3):XXXX • https://doi.
Daisuke Utsunomiya,Yoshinori Funama, Masafumi org/10.1148/rg.2019XXXXX
Kidoh,Yasuyuki Yamashita
First sentence, p 447: The first sentence should read
Erratum in: as follows: “The retroperitoneum is divided into three
RadioGraphics 2019:39(3):XXXX • distinct compartments: the posterior pararenal space,
https://doi. org/10.1148/rg.2019XXXXX bounded by the transversalis fascia posteriorly; the
First sentence, third full paragraph, p 1425: The anterior pararenal space, bounded by the parietal
sentence should read as follows: “Radiation output is peri- toneum anteriorly; and the perirenal space,
proportional to the square of tube voltage but shows bounded by the perirenal fascia (Fig 12a, 12b).”
a linear relationship with tube current.”