Sage Journals: Discover world-class research

Abstract

Intraductal papillary mucinous neoplasms (IPMNs) are a very common incidental finding during patient radiological assessment. These lesions may progress from low-grade dysplasia (LGD) to high-grade dysplasia (HGD) and even pancreatic cancer. The IPMN progression risk grows with time, so discontinuation of surveillance is not recommended. It is very important to identify imaging features that suggest LGD of IPMNs, and thus, distinguish lesions that only require careful surveillance from those that need surgical resection. It is important to know the management guidelines and especially the indications for surgery, to be able to point out in the report the findings that suggest malignant degeneration. The imaging tools employed for diagnosis and risk assessment are Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) with contrast medium. According to the latest European guidelines, MRI is the method of choice for the diagnosis and follow-up of patients with IPMN since this tool has a highest sensitivity in detecting mural nodules and intra-cystic septa. It plays a key role in the diagnosis of worrisome features and high-risk stigmata, which are associated with IPMNs malignant degeneration. Nowadays, the main limit of diagnostic tools is the ability to identify the precursor of pancreatic cancer. In this context, increasing attention is being given to artificial intelligence (AI) and radiomics analysis. However, these tools remain in an exploratory phase, considering the limitations of currently published studies. Key limits include noncompliance with AI best practices, radiomics workflow standardization, and clear reporting of study methodology, including segmentation and data balancing. In the radiological report it is useful to note the type of IPMN so as the morphological features, size, rate growth, wall, septa and mural nodules, on which the indications for surveillance and surgery are based. These features should be reported so as the surveillance time should be suggested according to guidelines.

Keywords

pancreatic cancer intraductal papillary mucinous neoplasms magnetic resonance imaging risk assessment radiomics

Introduction

Intraductal papillary mucinous neoplasms (IPMNs) are a very common incidental finding on Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) performed for other reasons.¹ These lesions are more frequent in men than women, at an age over 65, and account for 21%-33% of all cystic lesions of the pancreas.² IPMNs are neoplasms of the pancreas caused by epithelial cell hyperplasia of the pancreatic ducts that produce mucin.³ IPMNs, together with mucinous cystic neoplasm (MCN), are the cystic lesions with the highest risk of malignant degeneration.⁴ Although in some cases an IPMN may remain stable throughout the patient’s life, these lesions may progress from low-grade dysplasia (LGD) to high-grade dysplasia (HGD) and even pancreatic cancer (Figure 1). The risk of IPMN progression growing with time, so discontinuation of surveillance is not recommended.¹ Therefore, it is very important to be able to identify imaging features that suggest malignant degeneration of IPMNs, and thus, distinguish lesions that only require careful surveillance from those that need surgical resection.

Figure 1.

Man with main duct degenerated IPMN. In A (T2-W sequence in coronal plane), the arrow shows mucin plug and thickened wall. In B (T2-W sequence with FS in axial plane), yellow arrow shows liver mucinous metastasis at VI hepatic segment.

In this paper, we would like to focus on the aspects and characteristics of IPMNs that are useful to radiologist diagnosis and patient risk assessment evaluation.

Guidelines of the Management of IPMN: What the Radiologist is Concerned With

It is important to know the management guidelines and especially the indications for surgery, to identify and to report the findings that suggest malignant degeneration.⁵

The features correlated to HGD or invasive carcinoma (IC) in IPMN are identified as high-risk stigmata (HRS) and worrisome features (WF). The first are very strong prognosticators of HGD/IC but with low specificity. The guidelines suggest prudent surgical indication only considering HRS features since critical elements are clinical condition of patient, life expectancy, so as patient preference. Regarding to HRS, those include: (1) obstructive jaundice in a patient with head of the pancreas cystic lesion; (2) solid component or an enhancing mural nodule ≥5 mm; (3) main pancreatic duct (MPD) ≥10 mm, and (4) suspicious or positive results of cytology. With regard to WF, they include: (1) acute pancreatitis, (2) high serum level of CA19- 9, (3) new onset or acute exacerbation of DM within the past year, (4) cyst lesion ≥30 mm, (5) enhancing mural nodule <5 mm, (6) thickened/enhancing cyst walls, (7) MPD ≥5 mm and <10 mm, (8) abrupt change in caliber of pancreatic duct with distal pancreatic atrophy, (9) lymphadenopathy, and (10) cystic growth rate ≥2.5 mm/year.^6–9

Several guidelines are proposed, with different indications on surveillance and surgical management. The most recent and prevalent are those of the American Gastroenterological Association (AGA) guidelines (2015)⁶; the 2017 revised version of International Association of Pancreatology guidelines (Fukuoka guidelines)⁷; and the latest 2018 version of the guidelines of the European Study Group on Cystic Tumours of the Pancreas⁸ (Table 1). The radiologists should be interested above all in morphological features and findings that can be assessed by imaging.⁹ The Fukuoka and the European guidelines define either the presence of enhancing mural nodule ≥5 mm than the MPD ≥10 mm as absolute suggestions for surgery. Whereas for AGA, MPD dilatation ≥5 mm, but in association with a solid mass, is sufficient for surgery. The relative indications for surgery in the Fukuoka and European guidelines are described below (Table 1). AGA does not provide relative indications for surgery unlike the other 2. According to European based guidelines, for IPMN patients without indications for surgical treatment, routine follow-up should be considered: in the first year, a 6-month follow-up and annually thereafter.⁸

Table 1.

Indications for Surgical Resection in IPMN as Outlined by Current Guidelines.

Guidelines	Absolute Indications for Surgery	Relative Indications for Surgery
American gastroenterology association (AGA) (2015)	• MPD ≥5 mm (on MRI and EUS) and solid component
American gastroenterology association (AGA) (2015)	• Cytology positive for malignancy
International consensus fukuoka guidelines (2017)	• Enhancing mural nodule (≥5 mm)	• MPD dilatation between 5 and 9 mm
	• MPD dilatation ≥10 mm	• Enhancing mural nodule (<5 mm)
	• Cytology suspicious or positive for malignancy	• Cyst diameter ≥30 mm g Growth rate ≥5 mm over 2 years
	• Jaundice (tumor-related)	• Abrupt change in diameter of MPD with distal atrophy
		• Lymphadenopathy
		• Thickened or enhancing cyst walls
		• Increased levels of serum CA19-9
		• Acute pancreatitis (caused by IPMN)
European study group on cystic tumors of the pancreas (2018)	• Enhancing mural nodule (≥5 mm)	• MPD dilatation between 5 and 9.9 mm e Enhancing mural nodule (<5 mm)
	• MPD dilatation ≥10 mm	• Cyst diameter ≥40 mm g Growth rate ≥5 mm per year
	• Solid mass	• Increased levels of serum CA19-9 (>37 U/mL)
	• Jaundice (tumor- related)	• New- onset diabetes mellitus
	• Positive cytology for malignancy or HGD	• Acute pancreatitis (caused by IPMN)
International evidence-based kyoto guidelines for the management of IPMB (2023)	• Enhancing mural nodule (≥5 mm)
	• MPD dilatation ≥10 mm
	• Cytology suspicious or positive for malignancy

During this revision has been published the latest revised (2023) International evidence-based Kyoto guidelines for the management of IPMB.⁹ In the revised edition 5 topics are critical: amendment of HRS and WF, follow-up of non-resected lesions, follow-up of resected lesions, amendment of pathological features so as the assessment of molecular markers in cyst fluid. The working group does not suggest endoscopic ultrasound (EUS) in every patient, however, the imaging features obtained by EUS/contrast-enhanced EUS (CE-EUS) and cytology, obtained by EUS-guided fine needle aspiration (EUS-FNA) may be included into the HRS and WF evaluation, since EUS has now been considered as a crucial diagnostic tool to assess IPMN, especially for HGD/IC lesion.^9,10

Diagnosis: The role of MRI and magnetic resonance cholangiopancreatography

The imaging tools employed for diagnosis and risk assessment are EUS, computed tomography (CT) and Magnetic Resonance Imaging (MRI) with contrast medium.¹⁰ According to the European guidelines, MRI is the method of choice for the diagnosis and follow-up of IPMN patients,^8–11 since the diagnostic accuracy is 40%-95%.⁸ Using cholangiopancreatography (MRCP) sequences, the pancreatic biliary system can be better detected so that it is easy to identify the communication branch with the MPD.¹² To increase the diameter of the pancreatic ducts and make them more evident, the use of the hormone secretin, Secretin-enhanced MRCP (S-MRCP)) may be useful, but this still has controversial indications in the detection of IPMN.¹³ MRI, using in addition to conventional protocol MRCP sequences, has a higher soft-tissue contrast resolution than CT, allowing to detect single and multiple PCNs, including multifocal branch of the main duct (BD)-IPMN and mixed type (Montana-IPMN).¹⁴ In addition, this tool has a high sensitivity in detecting mural nodules and intra-cystic septa.¹ It plays a key role in the diagnosis of worrisome features and high-risk stigmata, which are associated with malignant degeneration of IPMNs. MRI is less invasive and more tolerated by the patient than endoscopy ultrasound (EUS).^11–15 Moreover, as it does not use radiation, it is preferred over CT for follow-up.^16,17

With regard to MRI protocols for patient surveillance, there is no definitive agreement.¹ The main critical point is the use of contrast medium,¹ in fact, several abbreviated or short protocol (SP) have been proposed.¹ Several authors reserved conventional protocol at the first diagnosis and SP, including axial and coronal SSFSE T2-W sequences, 2D and 3D single-shot MRCP and 3 D T1-W spoiled gradient echo, for surveillance.¹ No consensus there is on DWI utility, although several researches suggested the use of DWI to early detection of pancreatic cancer.¹

EUS has been proposed as a problem solving in pancreatic cistyc neoplasm (PCN) with features that should be considered for surgical resection.¹ According to Fukuoka guidelines, EUS is suggested in patients with HRS and WF features.⁷ While, the AGA guidelines suggests EUS-FNA only for cysts with 2 HRS or if significant changes develop during surveillance.⁶ According to latest international evidence-based Kyoto guidelines,⁹ EUS, including EUS-FNA and CE-EUS, is suggested for IPMN with high suspicious of HGD/IC and only if in the clinical setting it is available. Therefore, EUS-FNA to obtain cytological and molecular marker is an option, and should not be done when HRS are recognizable in others diagnostic tools.⁹

The current pancreatic cancer screening programs are based on the use of diagnostic tests. A combination of EUS and MR/MRCP executed annually is suggested.¹ Computed tomography with dedicated protocol is suggested only if the patient cannot undergo EUS or MR/MRCP.¹

With regard to time of follow-up, according to the Fukuoka guidelines, high risk patients including those with HGD at the surgical margins, non-intestinal IPMN, or patients with PC family history should repeat imaging at least every 6 months.⁷ For other IPMNs, is proposed follow-up every 6-12 months. The AGA guidelines proposed post surgical follow-up only for patients with HGD/IC using MRI every 2 years.⁶

According to International evidence-based Kyoto guidelines,⁹ surveillance should be adjusted considering the cyst size, the growth rate and the time to progression. For branch duct IPMN (BD-IPMN) misuring <20 mm, >20 mm and <30 mm, and >30 mm a follow-up time of 18, 12, and 6 months after an initial surveilance of 6 months is suggested. In patients with cysts <20 mm, without morphological changes and WF, after 5-year, surveillance may be discontinued. The discontinuation should not be appropriate for younger patients and for patients with familial or genetic risk.⁹

Report:Wwhat the radiologist should consider

In the radiological report it is useful to note the type of IPMN and the morphological features on which the indications for surveillance and surgery are based.^8,18,19 In Table 2 we report an example of structured radiological report.

Table 2.

Radiological Report.

IPMN	Yes/No
First diagnosis
Follow-up (clarify the time of follow-up)
Number of lesion
Uni-locular
Multi-locular
Type
MD-IPMN
BD-IPMN
MT-IPMN
Localization (if IPMN multi-locular for each lesion clarify)
Head
Uncinate process
Body
Tail
Main duct dilation
>5 mm and <9 mm
>10 mm
Cyst size (if IPMN multi-locular for each lesion clarify)
<20 mm
>20 mm and <30 mm
>30 mm
Enhancing mural nodule (if IPMN multi-locular for each lesion clarify)
< 5 mm
≥5 mm
Solid component (mm) (if IPMN multi-locular for each lesion clarify)
Thickened/enhancing cyst walls (mm) (if IPMN multi-locular for each lesion clarify)
Quick change in MPD caliber with distal pancreatic atrophy
Lymphadenopathy
Cystic growth rate ≥2.5 mm/year (if IPMN multi-locular for each lesion clarify)
New onset or acute exacerbation of MPD within the past year
Acute pancreatitis

TYPES of IPMN

Depending on the morphology and the duct involved, they can be of 3 types: - of the MD (MD-IPMN), - of a branch of the MD (BD-IPMN) - and of a mixed type (Montana-IPMN).

MD-IPMNs are identified as a dilatation of MD (>5 mm), either segmental or along its entire length, without a concomitant cystic lesion or additional cause of ductal obstruction. It is the type that most frequently undergoes malignant degeneration (62%).⁷

BD-IPMN is dilatation (>5 mm) of the branching ducts with uni- or multi-locular appearance (‘grape cluster’).²⁰ Typically, they are multifocal and located preferentially at the head (70%) or uncinate process of the pancreas.^7,21,22 Communication with the pancreatic ducts is a specific aspect. However, failure to detect this does not exclude the diagnosis of BD-IPMN. Morphologically, BD-IPMNs can show several patterns: (a) uni-locular; (b) multi-locular with macro-cystic aspect and few septa; (c) micro-cystic characteristic defined a grape-fruit growth, in which multiple septa divide several fluid spaces; (d) pleomorphic, in with we can find a mixture of these types. The ability of these lesions to produce mucin can sometimes lead to mucin plugs or balls within the cyst or in the pancreatic duct. These often present as non-enhancing mural nodules on the dependent surface of the cyst and should not be confused with solid component. BD-IPMN has a lower mean risk of degeneration into invasive carcinoma or high-grade dysplasia than MD-IPMN (14%–48%).⁷

MT-IPMN is shown as dilatation of the MD associated with 1 or more cyst-like growths. The risk of MT-IPMN cancer involvement is comparable to MD-IPMN.²³

Location

BD-IPMN are located preferentially at the head (70%) or uncinate process of the pancreas.^7,21 But they can occur anywhere in the pancreatic paranchyma. Specifying the location of the lesion in the report is especially useful to facilitate localisation during EUS and for surgical planning.

Size and Rate Growth

Given the importance of cyst size, this must be reported and performed in a reproducible and consistently manner. The guidelines take into account only 1 measurement, calculated on the largest diameter of the formation. This should be measured in the plane in which it appears largest, which may also be oblique to the plane of the scan. It has been seen that the best plane in which to measure the cyst is the coronal plane in T2-weighted MRI sequences.^24,25 Alternatively, the axial plane or MRCP sequences can be used. The measurement should be taken from outer wall to outer wall and the connection of the cyst to the main pancreatic duct must not be included in it.²⁵ In the case of multiple lesions, as is often in BD-IPMN, the largest cyst is chosen as the reference. To reduce inter-observational variability, it may be useful to indicate in the report the slice number, the plane and the sequence or scan in which the diameter was measured. This also facilitates the reproducibility of the measure in the follow-up comparison.

In a cyst with a diameter of less than 5 mm, the presence of a carcinoma is rare and does not require further diagnostic investigation.²⁴ A retrospective analysis by Das et al, however, recommends follow-up of cysts smaller than 5 mm.²⁴ Now, there is no agreement on the optimal interval for IPMNs surveillance. The 2017 IAP guidelines, identify a cyst ≥3 cm as a “WF”, which requires further investigation with EUS, and consider surgery for young patients.^7,21 While the 2018 European guidelines, include in the relative criteria for resection of BD-IPMNs, a diameter ≥4 cm.⁸ Size comparison with previous imaging examinations is imperative. Cyst growth rate >5 mm, in 2 years for IAP and in 1 year for European Society, is a relative indication for surgery, since it is correlated with IC in IPMN.

Wall and Septa

After defining the size of the cyst, its appearance should be assessed, based on its shape and the presence of septa. Presentation patterns may be: grape-like with cyst-by-cyst formation (with numerous thin septa or due to a group of clustered microcysts), macrocystic (which may be uni- or multi-locular with a few septa), finger-like (cylindrical dilatation of the side branches) or pleomorphic (mix of the other patterns).²⁶ Of the septa and wall, the thickness, enhancement after contrast medium administration (Figure 2) and irregularity should be analysed. The cut-off for defining thickened walls has not been defined, while for septa it is 2 mm.

Figure 2.

Woman with mixed degenerated IPMN. In A, cholangiopancreatography MRI, the arrow shows ‘grape cluster'. In B (T1-W after contrast medium during portal phase, in axial plane), the arrow shows enhanced cystic wall.

Kim et al demonstrated in a meta-analysis that the presence of septa >2 mm (Figure 3) correlates with the risk of malignant transformation.²⁷

Figure 3.

Man with main duct degenerated IPMN, during MRI assessment (A: cholangiopancreatography MRI) In B, T2-W sequence in coronal plane, the arrow shows thickened septa which after contrast agent (C: T1-W sequence during portal phase in coronal plane) have contrast enhancement.

And thickened and enhanced cystic wall is a “worrisome feature”, and therefore requires further study with EUS.⁷ However, these aspects should be considered together with the other alarming features to raise suspicion of degeneration. Especially in MRI, there may be limitations that can be misleading, such as low spatial resolution and artefacts. In particular, the presence of mural nodules should be investigated, as we will see below.

Mural Nodules

In all types of IPMN, it is important to identify wall outgrowths that underlie solid masses. When can we talk of a nodule? The definition of a mural nodule is a papillary outgrowth within the dilated main duct or cyst ≥3 mm.²¹ Mural nodules, regardless of size, are suggestive of solid masses and the bigger ones are a significant predictor of malignancy (Figure 4). If this finding is < 5 mm then it requires further investigation with EUS, if it is ≥ 5 mm then surgical evaluation is recommended. Crucial is the evaluation of post-contrast CT and MRI images. Evaluation of solid nodules is performed on T1-weighted sequences after contrast medium, and venous phase is preferred. Indeed, the lack of enhancement should raise the suspicion that this is not a mural nodule but just mucin plug or a conglomerate of thin septa. Mucin plugs are generally more frequent than nodules and are due to dehydration of secreted fluid. these should be reported by radiologists, since they might grow and obstruct the lumen of the ducts resulting in possible pancreatitis. Even more recommended is to report the finding if in doubt between a conglomerate of thin septa or mural nodule. This is so because both entities require short-term follow-up or further diagnostic investigation by EUS and fluid sampling.^6–8

Figure 4.

Men with main duct degenerated IPMN, during MRI assessment. In T2-W sequence (A: axial Plane), the arrow shows mural nodule that after contrast agent (B: T1-W during portal phase in axial plane) has contrast enhancement.

Multifocal BD-IPMN

Cystic lesions in BD-IPMN are multiple in 41% of cases.²⁷ There is no consensus in the literature as either multiplicity predisposes to malignant degeneration or not. In multifocal BD-IPMN, each lesion should be assessed separately.⁸ The most alarming cyst(s) should be described in detail in the report, selected on the basis of size and the presence of WF or HRS. In these cases, it is important to indicate whether they are distributed more in 1 portion of the pancreas (head, uncinate processus or tail) or not, to direct the eventual surgical approach (total or partial pancreasectomy).

Main duct dilation

MD-IPMNs are identified as a dilatation of the MD (>5 mm), focal or diffuse. The presence of MDP dilatation alone causes problems in differential diagnosis. The diagnosis of MD-IPMN is possible when other causes of dilatation of MPD as a stone or stenosis are excluded. In the absence of these findings, it also allows us to make the differential diagnosis between MD-IPMN and chronic pancreatitis, especially when this type of IPMN is associated with atrophy of the distal parenchyma. If the dilatation of the MDP is abruptly cut, in the absence of a history of pancreatitis, the presence of an invasive carcinoma causing ductal obstruction must be suspected. Furthermore, it should be remembered that a 5 mm dilatation of the isolated MDP with remaining sections of normal calibre (3 mm), may be completely normal, especially if already present at a previous check-up.²⁸

A calibre greater than 5 mm is an absolute indication for surgery for the American Gastroenterology Association (AGA),⁶ while greater than 10 is for IAP⁷ and European guidelines,⁸ which consider the range between 5 and 9 a relative indication. Therefore, it is very important to be as accurate as possible when measuring. Because of its oblique course, it is useful to perform the measurement in MRCP sequences that allow the entire MDP to be visualised.

Radiomics

Radiomics and artificial intelligence (AI) are progressively expanding fields and, in recent years, are becoming a powerful tool for radiologists.^29,30

Radiomics is the extraction of a high number of quantitative features from medical images, including several phases: image acquisition (eg, RX, US, CT, MRI, and ibrid imaging); segmentation (aumotatic, semi-automatic and manual); generation of features; development of database; analysis of database and radiomics signature.¹⁰ Images used for radiomic analysis are collected from different hospitals or data centers; thus, these images are usually obtained using different parameters and protocols and reconstructed with different software. The differences may bring unexpected influences on the radiomic model. Segmentation is critical because the subsequent feature data are generated from the segmented volumes. It is challenging because many tumors have indistinct borders. Generation of features refers to the extraction of semantic features such as dimension, necrosis, margin, location or extraction of non semantic features such as shape, hystogram or texture.¹⁰ In the past years, texture analysis has gained attention in medical applications and has been proved to be a significant computer-aided diagnostic tool 10]. There is not a strict definition of an image texture but it can be described as the spatial arrangement of patterns that provides the visual appearance of coarseness, randomness, smoothness, etc. Texture analysis describes a wide range of techniques for quantification of gray-level patterns and pixel inter-relationships within an image providing a measure of heterogeneity. It has been shown that different image areas exhibit different textural patterns that are sometimes imperceptible to the human eye.¹⁰ Applications of texture analysis in medical imaging include classification and segmentation of tissues and lesions. Texture analysis applications involve a process that consists of 6 steps: image acquisition, region of interest (ROI) definition, ROI pre-processing, feature extraction, feature selection, and classification. None of these steps is specific, and the methods have to be chosen according to the application.¹⁰ Manual definition of ROIs is still considered the gold standard in many applications, and it is the chosen option over automatic methods.

Radiomics can be performed with as few as 100 patients, although larger data sets provide more power. Radiomics could potentially aid cancer detection, diagnosis, assessment of prognosis, prediction of response to treatment, monitoring of disease status.¹⁰ Radiomic is designed to be used in decision support of precision medicine, using standard of care images that are routinely acquired in clinical practice, it presents a cost-effective and highly feasible addition for clinical decision support. Also, this analysis non-invasively characterize the overall tumor accounting for heterogeneity, interrogating the entire tumor allows the expression of microscopic genomic and proteomics patterns in terms of macroscopic image-based features

In the recent literature, there are several studies on the application of radiomics to pancreatic tumors.³¹ These involve the development of models applicable to both CT and MR.^32–38 The application in clinical practice of these tools, can support radiologists in determining certain decisive characteristics in the workflow of PC patients, such as malignant degeneration, response to therapy, etc.^39–42 In a large study of 408 IPMN patients, Tobaly et al evaluated the ability of CT-based radiomic analysis in predicting malignant IPMN degeneration.³⁷ In their study, from a retrospective analysis of preoperative CT images, they developed a radiomics model capable of distinguishing benign from malignant IPMNs with a sensitivity of .82 and specificity of 0,74. Radiomics was also applied to MR for the characterization of IPMNs. In a multicentre study, from the interaction of clinical data and radiomic analysis of pre-operative MR images of 202 patients, Cui et al developed a radiomic nomogram model capable of differentiating low-grade from high-grade IPMNs.³⁸ Recently, an Italian study evaluated the ability of radiomics to identify degenerated IPMNs before the appearance of radiological features of malignancy (Worrisome features and high-risk stigmata).³⁹ The constructed model obtained an AUC of more than.9 for the 4 MR sequences analysed (T1W, T1W post-contrast, T2W, and Apparent Diffusion Coefficient (ADC) map). The application of these quantitative analyses offers important support to the radiologist and other specialists in identifying the best treatment for patients with IPMN, with the ultimate aim of providing customised therapy.^18,43–47

AI could have a critical effect on IPMN classification and patient risk assessment. Deep learning (DL) models could perform a critical role during the different phases of diagnostic work flow, including acquisition, reconstruction, and post-processing. Also, AI models may have the potential to automate the diagnosis and risk stratification.¹ In a 139 patients study, the addition of a DL algorithm to MRI data, permitted to find advanced neoplasia in IPMNs (sensitivity of 75%; specificity of 78%; and AUC .783).⁴⁸ However, this approach accuracy is similar to that obtained using the guidelines management, so additional and more sophisticated models are required.

There are several limits to MRI-based radiomic analyses that should be considered. First, several studies have assessed only internal dataset, and to make reproducible and to validate the results there is the necessity of external dataset. Second, MRI studies are lesser standardized than CT studies, it is due to the larger range of scanner assets, sequences and image settings. To improve this field and increase patient outcomes, the introduction of standardized protocols and multi-center datasets are critical.

Discussion

The proper management of patient at risk for pancreatic cancer is still debated.^48–51 So as, analysing the guidelines,^6–8 it is clear as the different working groups suggest different approach. The European group suggest observing all pancreatic cystic lesion, in asymptomatic as symptomatic patients. Otherwise for AGA guidelines the patient must be asymptomatic. While the Fukuoka guidelines are the main specific, since this group considers only patients with IPMN. Lacking accurate and rigorous studies comparing the different guidelines, it is not clear which is the best surveillance strategy in these patients.¹

Regarding to diagnostic management, dedicated study protocol, either in MRI/MRCP that in CT, showed similar accuracy in lesion characterization.^48–53 However, MRI/MRCP has the higher accuracy either in the detection of the communication between a cystic lesion and the pancreatic duct system, than in detection of mural nodules and internal septa.⁴⁸ Multimodality imaging evaluation should be considered for tumor staging and pre-treatment evaluation.^54–66

Regarding study protocol, considering the wide spread of published protocols, but the lack of comparative studies, there are no specific study protocols which could be suggested.^66–81 Nowadays, the main limit of diagnostic tools is the ability to identify the precursor of pancreatic cancer (PC), and since PC remains the big killer, since the number of deaths almost equal to its incidence,¹ the chance of a surgical resection as soon as possible represents the only chance of a cure.⁸¹ In this context, increasing attention is being given to AI and radiomics analysis.^82–85

Latest technical progresses have led to the increased use of AI in radiological setting. Since, a computer can evaluate larger amounts of data respect to a human brain, this tool could resolve complicated tasks in oncological setting.^74,86–92 Machine learning (ML), an AI sub-field, is a mathematical model that can realise detailed tasks through the replication of computations.^66,93–98 The large quantity of ML models developed have made possible the introduction of a new approach for quantitative imaging assessment, named radiomics.^44,99–106 Radiomics analysis allows to convert digital medical images into numbers called features, that can be analysed in a mathematical manner.^{29,46,87,107–111} These features can be correlated with clinical-pathological data and outcomes, and, hypothetically, open the way for a personalised approach, that considers that patient with that particular lesion.^{66,90,92,93,98,112–124} For several tumors, radiomic analysis has already provided a precise assessment of biology, allowing the identification of features that are correlated with outcomes.^{35,37,38,83,85,96,98–100,125–139} In the context of IPMN, AI and radiomics analisys remain in an exploratory phase, considering the limitations of currently published studies.^31–42 Key limits include noncompliance with ML best practices, radiomics workflow standardization, and clear reporting of study methodology, including segmentation and data balancing. So the lack of quality and adequate sample sizes and external validation set remains critical points to resolve.¹²⁹

Conclusions

The management of the patient with IPMN requires a multidisciplinary approach. The radiologist plays a key role in the diagnosis of IPMN and the assessment of suspicious features for malignant degeneration. MRI is the method of choice and allows the identification of findings that represent absolute or relative indications for surgery. A structured radiological report in which the radiologist report the IPMN types and the HRS and WF features to allow a proper patient management either for surgical indication either for surveillance is required. Although AI and radiomics analysis represent an interesting field, however critical issues make these applications still limited to a theoretical field.

Footnotes

Acknowledgments

The authors are grateful to Alessandra Trocino, librarian at the National Cancer Institute of Naples, Italy.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was suported by the Italian Ministry of Health “Ricerca Corrente” funds.

Ethical Statement

ORCID iD

Vincenza Granata

Data Availability Statement

Data are available at link .

References

Granata

Fusco

Setola

, et al. Risk assessment and pancreatic cancer: diagnostic management and artificial intelligence. Cancers. 2023;15(2):351. doi:10.3390/cancers15020351.

Mamone

Barresi

Tropea

Di Piazza

Miraglia

. MRI of mucinous pancreatic cystic lesions: a new updated morphological approach for the differential diagnosis. Updates Surg. 2020;72(3):617-637. doi:10.1007/s13304-020-00800-y.

Scholten

van Huijgevoort

NCM

van Hooft

Besselink

Del Chiaro

. Pancreatic cystic neoplasms: different types, different management, new guidelines. Visc Med. 2018;34(3):173-177. doi:10.1159/000489641.

Kromrey

Bülow

Hübner

, et al. Prospective study on the incidence, prevalence and 5-year pancreatic-related mortality of pancreatic cysts in a population-based study. Gut. 2018;67(1):138-145. doi:10.1136/gutjnl-2016-313127.

Kenner

Chari

Kelsen

, et al. Artificial intelligence and early detection of pancreatic cancer: 2020 summative review. Pancreas. 2021;50(3):251-279. doi:10.1097/MPA.0000000000001762.

Vege

Ziring

Jain

Moayyedi

Clinical Guidelines Committee, and American Gastroenterology Association . American gastroenterological association institute guideline on the diagnosis and management of asymptomatic neoplastic pancreatic cysts. Gastroenterology. 2015;148(4):819. doi:10.1053/j.gastro.2015.01.015.

Tanaka

Fernández-Del Castillo

Kamisawa

, et al. Revisions of international consensus Fukuoka guidelines for the management of IPMN of the pancreas. Pancreatology. 2017;17(5):738-753. doi:10.1016/j.pan.2017.07.007.

European Study Group on Cystic Tumours of the Pancreas . European evidence-based guidelines on pancreatic cystic neoplasms. Gut. 2018;67(5):789-804. doi:10.1136/gutjnl-2018-316027.

Ohtsuka

Fernandez-Del Castillo

Furukawa

, et al. International evidence-based Kyoto guidelines for the management of intraductal papillary mucinous neoplasm of the pancreas. Pancreatology. 2023;24:S1424-S3903. doi: 10.1016/j.pan.2023.12.009.

10.

Granata

Fusco

Barretta

, et al. Radiomics in hepatic metastasis by colorectal cancer. Infect Agent Cancer. 2021;16(1):39. doi:10.1186/s13027-021-00379-y.

11.

Huang

Zhang

, et al. Artificial intelligence in pancreatic cancer. Theranostics. 2022;12(16):6931-6954. doi:10.7150/thno.77949.

12.

Yang

Wang

Qiu

Ren

You

. Early screening and diagnosis strategies of pancreatic cancer: a comprehensive review. Cancer Commun. 2021;41(12):1257-1274. doi:10.1002/cac2.12204.

13.

Swensson

Zaheer

Conwell

Sandrasegaran

Manfredi

Tirkes

. Secretin-enhanced MRCP: how and why-AJR expert panel narrative review. AJR Am J Roentgenol. 2021;216(5):1139-1149. doi:10.2214/AJR.20.24857.

14.

Manuel-Vazquez

Ramia

Latorre-Fragua

, et al. Pancreatic neuroendocrine tumors and intraductal papillary mucinous neoplasm of the pancreas: a systematic review. Pancreas. 2018;47(5):551-555. doi:10.1097/MPA.0000000000001048.

15.

Granata

Fusco

De Muzio

, et al. Radiomics and machine learning analysis by computed tomography and magnetic resonance imaging in colorectal liver metastases prognostic assessment. Radiol Med. 2023;11:1310-1332. doi:10.1007/s11547-023-01710-w.

16.

Nakagawa

Yamaguchi

Ohtsuka

, et al. Usefulness of multidetector computed tomography for detecting protruding lesions in intraductal papillary mucinous neoplasm of the pancreas in comparison with single-detector computed tomography and endoscopic ultrasonography. Pancreas. 2009;38(2):131-136. doi:10.1097/MPA.0b013e31818b0040.

17.

Xiang

Zhu

Wang

Zhou

. Sixteen cases of laparoscopic central pancreatectomy for benign or low-grade malignant tumours in the pancreatic neck and proximal body. ANZ J Surg. 2024;2:888-893. doi:10.1111/ans.18893.

18.

Granata

Fusco

De Muzio

, et al. Radiomics textural features by MR imaging to assess clinical outcomes following liver resection in colorectal liver metastases. Radiol Med. 2022;127(5):461-470. doi:10.1007/s11547-022-01477-6.

19.

Ohyama

Hirotsu

Mochizuki

Kato

Omata

. Distinct genomic profiles of two small malignant lesions associated with an intraductal papillary mucinous neoplasm Co-occurring in a patient. Cureus. 2023;15(12):e51394. doi:10.7759/cureus.51394.

20.

Kucera

Perrin

Caracciolo

Schmulewitz

Kedar

. Cystic lesions of the pancreas: radiologic-endosonographic correlation. Radiographics. 2012;32(7):E283-E301. doi:10.1148/rg.327125019.

21.

Qureshi

Javed

Sarmadi

Pandol

. Artificial intelligence and imaging for risk prediction of pancreatic cancer: a narrative review. Chin Clin Oncol. 2022;11(1):1. doi:10.21037/cco-21-117.

22.

Sakorafas

Smyrniotis

Reid-Lombardo

Sarr

. Primary pancreatic cystic neoplasms revisited. Part III. Intraductal papillary mucinous neoplasms. Surgical oncology. 2011;20(2):e109-e118. doi:10.1016/j.suronc.2011.01.004.

23.

Sahora

Fernández-del Castillo

Dong

, et al. Not all mixed-type intraductal papillary mucinous neoplasms behave like main-duct lesions: implications of minimal involvement of the main pancreatic duct. Surgery. 2014;156(3):611-621. doi:10.1016/j.surg.2014.04.023.

24.

Das

Wells

Nguyen

. Incidental cystic neoplasms of pancreas: what is the optimal interval of imaging surveillance? Am J Gastroenterol. 2008;103(7):1657-1662. doi:10.1111/j.1572-0241.2008.01893.x.

25.

Dunn

Brook

, et al. Measurement of pancreatic cystic lesions on magnetic resonance imaging: efficacy of standards in reducing inter-observer variability. Abdom Radiol. 2016;41(3):500-507. doi:10.1007/s00261-015-0588-4.

26.

Kim

Park

Pyo

, et al. Imaging features to distinguish malignant and benign branch-duct type intraductal papillary mucinous neoplasms of the pancreas: a meta-analysis. Annals of surgery. 2014;259(1):72-81. doi:10.1097/SLA.0b013e31829385f7.

27.

Overbeek

Alblas

Gausman

, et al. Development of a stratification tool to identify pancreatic intraductal papillary mucinous neoplasms at lowest risk of progression. Aliment Pharmacol Ther. 2019;50(7):789-799. doi:10.1111/apt.15440.

28.

Mortelé

Rocha

Streeter

Taylor

. Multimodality imaging of pancreatic and biliary congenital anomalies. Radiographics. 2006;26(3):715-731. doi:10.1148/rg.263055164.

29.

Yao

Zhang

Demir

, et al. Radiomics boosts deep learning model for IPMN classification. Mach Learn Med Imaging. 2023;14349:134-143. doi:10.1007/978-3-031-45676-3_14.

30.

Han

Kwon

Lee

, et al. Optimal surveillance interval of branch duct intraductal papillary mucinous neoplasm of the pancreas. JAMA Surg. 2024;159:e237010. doi:10.1001/jamasurg.2023.7010.

31.

Jeong

Morse

Polk

, et al. Pancreatic cyst size measurement on magnetic resonance imaging compared to pathology. Cancers. 2024;16(1):206. doi:10.3390/cancers16010206.

32.

Gorris

van Bodegraven

Abu Hilal

, et al. Outcomes after distal pancreatectomy with or without splenectomy for intraductal papillary mucinous neoplasm: international multicentre cohort study. Br J Surg. 2024;111(1):znad424. doi:10.1093/bjs/znad424.

33.

León-Letelier

Dou

Vykoukal

, et al. Contributions of the microbiome-derived metabolome for risk assessment and prognostication of pancreatic cancer. Clin Chem. 2024;70(1):102-115. doi:10.1093/clinchem/hvad186.

34.

Spektor

Gutjahr

Lang

, et al. Immunohistochemical FAP expression reflects 68Ga-fapi PET imaging properties of low- and high-grade intraductal papillary mucinous neoplasms and pancreatic ductal adenocarcinoma. J Nucl Med. 2024;65(1):52-58. doi:10.2967/jnumed.123.266393.

35.

Faur

Lazar

Ghenciu

. Artificial intelligence as a noninvasive tool for pancreatic cancer prediction and diagnosis. World J Gastroenterol. 2023;29(12):1811-1823. doi:10.3748/wjg.v29.i12.1811.

36.

Pereira

Oldfield

Ney

, et al. Early detection of pancreatic cancer. Lancet Gastroenterol Hepatol. 2020;5(7):698-710. doi:10.1016/S2468-1253(19)30416-9.

37.

Tobaly

Santinha

Sartoris

, et al. CT-based radiomics analysis to predict malignancy in patients with intraductal papillary mucinous neoplasm (IPMN) of the pancreas. Cancers. 2020;12(11):3089. doi:10.3390/cancers12113089.

38.

Cui

Tang

, et al. Radiomic nomogram based on MRI to predict grade of branching type intraductal papillary mucinous neoplasms of the pancreas: a multicenter study. Cancer Imag. 2021;21(1):26. doi:10.1186/s40644-021-00395-6.

39.

Flammia

Innocenti

Galluzzo

, et al. Branch duct-intraductal papillary mucinous neoplasms (BD-IPMNs): an MRI-based radiomic model to determine the malignant degeneration potential. Radiol Med. 2023;128(4):383-392. doi:10.1007/s11547-023-01609-6.

40.

Crippa

Piccioli

Salandini

Cova

Aleotti

Falconi

. Treatment of branch-duct intraductal papillary mucinous neoplasms of the pancreas: state of the art. Updates Surg. 2016;68(3):265-271. doi:10.1007/s13304-016-0386-8.

41.

Granata

Fusco

Sansone

, et al. Magnetic resonance imaging in the assessment of pancreatic cancer with quantitative parameter extraction by means of dynamic contrast-enhanced magnetic resonance imaging, diffusion kurtosis imaging and intravoxel incoherent motion diffusion-weighted imaging. Therap Adv Gastroenterol. 2020:13:1756284819885052. doi:10.1177/1756284819885052.

42.

Litjens

Broekmans

JPEA

Boers

, et al. Computed tomography-based radiomics using tumor and vessel features to assess resectability in cancer of the pancreatic head. Diagnostics. 2023;13(20):3198. doi:10.3390/diagnostics13203198.

43.

Zhang

Sun

Zhao

Liu

. A prognostic and immunotherapy effectiveness model for pancreatic adenocarcinoma based on cuproptosis-related lncRNAs signature. Medicine (Baltim). 2023;102(42):e35167. doi:10.1097/MD.0000000000035167.

44.

Yuan

Shu

Peng

, et al. Prediction of postoperative liver metastasis in pancreatic ductal adenocarcinoma based on multiparametric magnetic resonance radiomics combined with serological markers: a cohort study of machine learning. Abdom Radiol (NY). 2023;49:117-130. doi:10.1007/s00261-023-04047-0.

45.

Ohno

Balduzzi

Hijioka

, et al. Association of high-risk stigmata and worrisome features with advanced neoplasia in intraductal papillary mucinous neoplasms (IPMN): a systematic review. Pancreatology. 2024;24(1):48-61. doi:10.1016/j.pan.2023.12.002.

46.

Vezakis

Gourtsoyianni

, et al. An automated prognostic model for pancreatic ductal adenocarcinoma. Genes. 2023;14(9):1742. doi:10.3390/genes14091742.

47.

Quingalahua

Al-Hawary

Machicado

. The role of magnetic resonance imaging (MRI) in the diagnosis of pancreatic cystic lesions (PCLs). Diagnostics. 202;13(4):585. doi:10.3390/diagnostics13040585.

48.

Corral

Hussein

Kandel

Bolan

Bagci

Wallace

. Deep learning to classify intraductal papillary mucinous neoplasms using magnetic resonance imaging. Pancreas. 2019;48:805-810.

49.

Sahani

Kambadakone

Macari

, et al. Diagnosis and management of cystic pancreatic lesions. AJR Am J Roentgenol. 2013;200:343-354.

50.

Chaudhari

Raman

Vuong

, et al. Pancreatic cystic lesions: discrimination accuracy based on clinical data and high resolution CT features. J Comput Assist Tomogr. 2007;31:860-867.

51.

de Jong

Nio

Mearadji

, et al. Disappointing interobserver agreement among radiologists for a classifying diagnosis of pancreatic cysts using magnetic resonance imaging. Pancreas. 2012;41:278-282.

52.

Waters

Schmidt

Pinchot

, et al. CT vs MRCP: optimal classification of IPMN type and extent. J Gastrointest Surg. 2008;12:101-109.

53.

Pilleul

Rochette

Partensky

, et al. Preoperative evaluation of intraductal papillary mucinous tumors performed by pancreatic magnetic resonance imaging and correlated with surgical and histopathologic findings. J Magn Reson Imaging. 2005;21:237-244.

54.

Leonhardt

Niesen

Pils

, et al. Patient reported outcomes of duodenum-preserving pancreatic head resection in chronic pancreatitis: high effectivity is impaired by prolonged non-surgical management. HPB (Oxford). 2023;5:S1365. doi:10.1016/j.hpb.2023.10.002.

55.

Daly

Pal

Maxwell

, et al. NCCN Guidelines® insights: genetic/familial high-risk assessment: breast, ovarian, and pancreatic, version 2.2024. J Natl Compr Canc Netw. 2023;21(10):1000-1010. doi:10.6004/jnccn.2023.0051.

56.

Nguyen

Gilad

Drogan

Eilers

Liao

Kupfer

. Risk perception and surveillance uptake in individuals at increased risk for pancreatic ductal adenocarcinoma. J Med Genet. 2023;18:109539. doi:10.1136/jmg-2023-109539.

57.

Granata

Morana

D'Onofrio

, et al. Structured reporting of computed tomography and magnetic resonance in the staging of pancreatic adenocarcinoma: a delphi consensus proposal. Diagnostics. 2021;11(11):2033. doi:10.3390/diagnostics11112033.

58.

Berbís

MÁ

Godino

Rodríguez-Comas

, et al. Radiomics in CT and MR imaging of the liver and pancreas: tools with potential for clinical application. Abdom Radiol (NY). 2023;49:322-340. doi:10.1007/s00261-023-04071-0.

59.

Granata

Grassi

Fusco

, et al. Local ablation of pancreatic tumors: state of the art and future perspectives. World J Gastroenterol. 2021;27(23):3413-3428. doi:10.3748/wjg.v27.i23.3413.

60.

Granata

Grassi

Fusco

, et al. Pancreatic cancer detection and characterization: state of the art and radiomics. Eur Rev Med Pharmacol Sci. 2021;25(10):3684-3699. doi:10.26355/eurrev_202105_25935.

61.

Jiang

Chao

Cao

, et al. Improving pancreatic cyst management: artificial intelligence-powered prediction of advanced neoplasms through endoscopic ultrasound-guided confocal endomicroscopy. Biomimetics. 2023;8(6):496. doi:10.3390/biomimetics8060496.

62.

Cholangiocarcinoma Working Group . Italian clinical practice guidelines on cholangiocarcinoma - Part II: treatment. Dig Liver Dis. 2020;52(12):1430-1442. doi:10.1016/j.dld.2020.08.030.

63.

Jung

Lee

Han

, et al. Inadequacy of the eighth edition of the American Joint Committee on Cancer pancreatic cancer staging system for invasive carcinoma associated with premalignant lesions in the pancreas: an analysis using the National Cancer Database. HPB (Oxford). 2023;26:S1365. doi: 10.1016/j.hpb.2023.11.015.

64.

Cholangiocarcinoma Working Group . Italian clinical practice guidelines on cholangiocarcinoma - Part I: classification, diagnosis and staging. Dig Liver Dis. 2020;52(11):1282-1293. doi:10.1016/j.dld.2020.06.045.

65.

Mishima

Takano

Takayashiki

, et al. Serum elastase-1 predicts malignancy in intraductal papillary mucinous neoplasm of the pancreas. Pancreatology. 2024;24(1):93-99. doi:10.1016/j.pan.2023.11.015.

66.

Zhang

. Preoperative differentiation of pancreatic cystic neoplasm subtypes on computed tomography radiomics. Quant Imag Med Surg. 2023;13(10):6395-6411. doi:10.21037/qims-22-1192.

67.

Nagayama

Ueki

Shimizu

, et al.

Is preoperative pancreatic juice cytology useful for determining therapeutic strategies for patients with intraductal papillary mucinous neoplasm of the pancreas?

J Hepatobiliary Pancreat Sci. 2023;31:183-192. doi:10.1002/jhbp.1394.

68.

Maio

Pasqualino

Bertana

, et al.

Pancreatic cancer detection with a non-contrast MR protocol: is it reliable?

Radiol Med. 2023;128(9):1035-1046. doi:10.1007/s11547-023-01680-z.

69.

Konno

Sugai

Kanoto

, et al. A retrospective preliminary study of intrapancreatic late enhancement as a noteworthy imaging finding in the early stages of pancreatic adenocarcinoma. Eur Radiol. 2023;33(7):5131-5141. doi:10.1007/s00330-022-09388-w.

70.

Thormann

Hinnerichs

Barajas Ordonez

, et al. Sarcopenia is an independent prognostic factor in patients with pancreatic cancer - a meta-analysis. Acad Radiol. 2023;30(8):1552-1561. doi:10.1016/j.acra.2022.10.025.

71.

Anai

Hayashida

Ueda

, et al. The effect of CT texture-based analysis using machine learning approaches on radiologists' performance in differentiating focal-type autoimmune pancreatitis and pancreatic duct carcinoma. Jpn J Radiol. 2022;40(11):1156-1165. doi:10.1007/s11604-022-01298-7.

72.

Sotozono

Kanki

Yasokawa

, et al. Value of 3-T MR imaging in intraductal papillary mucinous neoplasm with a concomitant invasive carcinoma. Eur Radiol. 2022;32(12):8276-8284. doi:10.1007/s00330-022-08881-6.

73.

Cao

Xia

Yao

, et al. Large-scale pancreatic cancer detection via non-contrast CT and deep learning. Nat Med. 2023;29(12):3033-3043. doi:10.1038/s41591-023-02640-w.

74.

Kierans

Gavlin

Wehrli

Flisnik

Eliades

Pittman

. Utility of gadolinium for identifying the malignant potential of pancreatic cystic lesions. Abdom Radiol (NY). 2022;47(4):1351-1359. doi:10.1007/s00261-022-03446-z.

75.

Zhuang

, et al. Acute pancreatitis in intraductal papillary mucinous neoplasm: a single-center retrospective cohort study with systematic review and meta-analysis. BMC Gastroenterol. 2023;23(1):424. doi:10.1186/s12876-023-02972-4.

76.

Donati

Casini

Cervelli

Morganti

Boraschi

. Diffusion-weighted MRI of solid pancreatic lesions: comparison between reduced field-of-view and large field-of-view sequences. Eur J Radiol. 2021;143:109936. doi:10.1016/j.ejrad.2021.109936.

77.

Zhang

Hou

Tian

Zhang

. The combination of eyeMax direct visualization system, EUS and ERCP for the precise treatment of intraductal papillary mucinous neoplasm. Rev Esp Enferm Dig. 2023;115(12):740-741. doi:10.17235/reed.2023.10110/2023.

78.

Yoshioka

Shigekawa

Ikezawa

, et al. The relationship between observation interval and prognosis in pancreatic cancer concomitant with intraductal papillary mucinous neoplasia. Pancreatology. 2024;24(1):73-77. doi:10.1016/j.pan.2023.11.005.

79.

Wang

Xie

Wei

. Pancreatic acinar-to-ductal metaplasia and pancreatic cancer. Methods Mol Biol. 2019;1882:299-308. doi:10.1007/978-1-4939-8879-2_26.

80.

Longnecker

Suriawinata

. Incidence of pancreatic intraepithelial neoplasia in an autopsy series. Pancreas. 2022;51(4):305-309. doi:10.1097/MPA.0000000000002027.

81.

Egawa

Toma

Ohigashi

, et al. Japan pancreatic cancer registry; 30th year anniversary: Japan pancreas society. Pancreas. 2012;41(7):985-992. doi:10.1097/MPA.0b013e318258055c.

82.

Young

Abrams

Ghosh

Rinaudo

JAS

Marquez

Srivastava

. Prediagnostic image data, artificial intelligence, and pancreatic cancer: a tell-tale sign to early detection. Pancreas. 2020;49(7):882-886. doi:10.1097/MPA.0000000000001603.

83.

Gao

Bai

Miao

, et al. Prediction of synchronous distant metastasis of primary pancreatic ductal adenocarcinoma using the radiomics features derived from 18F-FDG PET and MRI. Clin Radiol. 2023;78(10):746-754. doi:10.1016/j.crad.2023.06.011.

84.

Mayo

Fernández-Del Castillo

, et al. Impact of preoperative endoscopic procedures on adverse event rates after surgical resection for main-duct and mixed-type intraductal papillary mucinous neoplasms (IPMNs). Pancreatology. 2024;24(1):152-159. doi:10.1016/j.pan.2023.11.007.

85.

Lin

Zhang

, et al. Prognostic predicting model of pancreatic body tail carcinoma using clinical and CT radiomic data. Technol Cancer Res Treat. 2023;22:15330338231186739. doi:10.1177/15330338231186739.

86.

Ronot

Soyer

. Can radiomics outperform pathology for tumor grading? Diagn Interv Imaging. 2023;105:S2211-S5684. doi: 10.1016/j.diii.2023.09.001.

87.

van der Kroft

Wee

Rensen

, et al. Identifying radiomics signatures in body composition imaging for the prediction of outcome following pancreatic cancer resection. Front Oncol. 2023;13:1062937. doi:10.3389/fonc.2023.1062937.

88.

Suto

Ando

Matsukawa

, et al. Tumor-to-blood pool ratio of 18F-fluorodeoxyglucose-positron emission tomography's standardized uptake value as a useful parameter indicating malignant transformation in pancreatic branch-duct intraductal papillary mucinous neoplasm compared to the international Fukuoka guidelines: a retrospective cohort study from surgical resections. HPB (Oxford). 2024;26(2):291-298. doi:10.1016/j.hpb.2023.10.018.

89.

Granata

Fusco

Risi

, et al. Diffusion-weighted MRI and diffusion kurtosis imaging to detect RAS mutation in colorectal liver metastasis. Cancers. 2020;12(9):2420. doi:10.3390/cancers12092420.

90.

Lee

Shin

Kim

, et al. Radiomics model versus 2017 revised international consensus guidelines for predicting malignant intraductal papillary mucinous neoplasms. Eur Radiol. 2023;34:1222-1231. doi:10.1007/s00330-023-10158-5.

91.

Itoh

Omori

Seino

, et al. Gene rearrangement and expression of PRKACA and PRKACB govern morphobiology of pancreatobiliary oncocytic neoplasms. Mod Pathol. 2024;37(1):100358. doi:10.1016/j.modpat.2023.100358.

92.

Ren

Chen

Wang

. Editorial: novel circulating biomarkers and radiomics in pancreatic cancer. Front Oncol. 2023;13:1259070. doi:10.3389/fonc.2023.1259070.

93.

Yao

Zhang

Keles

Yazici

Tirkes

Bagci

. A review of deep learning and radiomics approaches for pancreatic cancer diagnosis from medical imaging. Curr Opin Gastroenterol. 2023;39(5):436-447. doi:10.1097/MOG.0000000000000966.

94.

Kenner

Abrams

Chari

, et al. Early detection of pancreatic cancer: applying artificial intelligence to electronic health records. Pancreas. 2021;50(7):916-922. doi:10.1097/MPA.0000000000001882.

95.

Ursprung

Beer

Bruining

, et al. Radiomics of computed tomography and magnetic resonance imaging in renal cell carcinoma-a systematic review and meta-analysis. Eur Radiol. 2020;30(6):3558-3566. doi:10.1007/s00330-020-06666-3.

96.

Triantopoulou

Gourtsoyianni

Karakaxas

Delis

. Intraductal papillary mucinous neoplasm of the pancreas: a challenging diagnosis. Diagnostics. 2023;13(12):2015. doi:10.3390/diagnostics13122015.

97.

Polk

Choi

McGettigan

, et al. Multiphase computed tomography radiomics of pancreatic intraductal papillary mucinous neoplasms to predict malignancy. World J Gastroenterol. 2020;26(24):3458-3471. doi:10.3748/wjg.v26.i24.3458.

98.

Huang

Chopra

Bolan

, et al. Pancreatic cystic lesions: next generation of radiologic assessment. Gastrointest Endosc Clin N Am. 2023;33(3):533-546. doi:10.1016/j.giec.2023.03.004.

99.

Chen

, et al. A deep-learning radiomics-based lymph node metastasis predictive model for pancreatic cancer: a diagnostic study. Int J Surg. 2023;109(8):2196-2203. doi:10.1097/JS9.0000000000000469.

100.

Tovar

Rosenthal

Maitra

Koay

. Potential of artificial intelligence in the risk stratification for and early detection of pancreatic cancer. Artif Intell Surg. 2023;3(1):14-26. doi:10.20517/ais.2022.38.

101.

Malcolm

Tacey

Gibbs

Lee

. Current state of radiomic research in pancreatic cancer: focusing on study design and reproducibility of findings. Eur Radiol. 2023;33(10):6659-6669. doi:10.1007/s00330-023-09653-6.

102.

Fornacon-Wood

Mistry

Ackermann

, et al. Reliability and prognostic value of radiomic features are highly dependent on choice of feature extraction platform. Eur Radiol. 2020;30(11):6241-6250. doi:10.1007/s00330-020-06957-9.

103.

Paquier

Chao

Acquisto

, et al. Radiomics software comparison using digital phantom and patient data: IBSI-compliance does not guarantee concordance of feature values. Biomed Phys Eng Express. 2022;8(6):065008. doi:10.1088/2057-1976/ac8e6f.

104.

Bettinelli

Branchini

De Monte

Scaggion

Paiusco

. Technical Note: an IBEX adaption toward image biomarker standardization. Med Phys. 2020;47(3):1167-1173. doi:10.1002/mp.13956.

105.

Pfaehler

Zhovannik

Wei

, et al. A systematic review and quality of reporting checklist for repeatability and reproducibility of radiomic features. Phys Imaging Radiat Oncol. 2021;20:69-75. doi:10.1016/j.phro.2021.10.007.

106.

de la Fuente

Chatterjee

Lui

, et al. Long-term outcomes and risk of pancreatic cancer in intraductal papillary mucinous neoplasms. JAMA Netw Open. 2023;6(10):e2337799. doi:10.1001/jamanetworkopen.2023.37799.

107.

Davis

Choubey

Langan

, et al.

Pancreatectomy for intraductal papillary mucinous neoplasm: has anything changed in North America?

HPB (Oxford). 2024;26(1):109-116. doi:10.1016/j.hpb.2023.09.001.

108.

Sacks

Shin

Braithwaite

, et al. Risk perceptions and risk thresholds among surgeons in the management of intraductal papillary mucinous neoplasms. Ann Surg. 2023;278(5):e1073-e1079. doi:10.1097/SLA.0000000000005827.

109.

Semaan

Bernard

Wong

, et al. Integrated molecular characterization of intraductal papillary mucinous neoplasms: an NCI cancer moonshot precancer atlas pilot project. Cancer Res Commun. 2023;3(10):2062-2073. doi:10.1158/2767-9764.CRC-22-0419.

110.

Sato

Nakagawa

Terashima

, et al. EphA2 proteolytic fragment as a sensitive diagnostic biomarker for very early-stage pancreatic ductal carcinoma. Cancer Res Commun. 2023;3(9):1862-1874. doi:10.1158/2767-9764.CRC-23-0087.

111.

Hamada

Oyama

Nakai

Fujishiro

. Risk stratification based on cyst size and its stability for low-risk intraductal papillary mucinous neoplasms: which size matters? Am J Gastroenterol. 2023;118(9):1705-1706. doi:10.14309/ajg.0000000000002243.

112.

Shen

Yang

, et al. A contrast-enhanced computed tomography based radiomics approach for preoperative differentiation of pancreatic cystic neoplasm subtypes: a feasibility study. Front Oncol. 2020;10:248. doi:10.3389/fonc.2020.00248.

113.

Liang

Tian

Wang

, et al. Classification prediction of pancreatic cystic neoplasms based on radiomics deep learning models. BMC Cancer. 2022;22(1):1237. doi:10.1186/s12885-022-10273-4.

114.

Jan

Ch'ang

. Selection of patients with pancreatic adenocarcinoma who may benefit from radiotherapy. Radiat Oncol. 2023;18(1):137. doi:10.1186/s13014-023-02328-y.

115.

Meng

Hang

Shi

. A nomogram based on CT texture features to predict the response of patients with advanced pancreatic cancer treated with chemotherapy. BMC Gastroenterol. 2023;23(1):274. doi:10.1186/s12876-023-02902-4.

116.

Komici

Cappuccio

Scacchi

, et al. The prevalence and the impact of frailty in hepato-biliary pancreatic cancers: a systematic review and meta-analysis. J Clin Med. 2022;11(4):1116. doi:10.3390/jcm11041116.

117.

Buondonno

Avella

Cappuccio

, et al. A hub and spoke learning program in bariatric surgery in a small region of Italy. Front Surg. 2022;9:855527. doi:10.3389/fsurg.2022.855527.

118.

Giuliani

Avella

Segreto

, et al. Postoperative outcomes analysis after pancreatic duct occlusion: a safe option to treat the pancreatic stump after pancreaticoduodenectomy in low-volume centers. Front Surg. 2021;8:804675. doi:10.3389/fsurg.2021.804675.

119.

Rocca

Brunese

Santone

, et al. Early diagnosis of liver metastases from colorectal cancer through CT radiomics and formal methods: a pilot study. J Clin Med. 2021;11(1):31. doi:10.3390/jcm11010031.

120.

Avella

Vaschetti

Cappuccio

, et al. The role of liver surgery in simultaneous synchronous colorectal liver metastases and colorectal cancer resections: a literature review of 1730 patients underwent open and minimally invasive surgery. Minerva Surg. 2022;77(6):582-590. doi:10.23736/S2724-5691.22.09716-7.

121.

Chu

Park

Soleimani

, et al. Classification of pancreatic cystic neoplasms using radiomic feature analysis is equivalent to an experienced academic radiologist: a step toward computer-augmented diagnostics for radiologists. Abdom Radiol (NY). 2022;47(12):4139-4150. doi:10.1007/s00261-022-03663-6.

122.

Anta

Martínez-Ballestero

Eiroa

García

Rodríguez-Comas

. Artificial intelligence for the detection of pancreatic lesions. Int J Comput Assist Radiol Surg. 2022;17(10):1855-1865. doi:10.1007/s11548-022-02706-z.

123.

Zhang

Qian

Chen

Liu

. Development and validation of an MRI-radiomics nomogram for the prognosis of pancreatic ductal adenocarcinoma. Front Oncol. 2023;13:1074445. doi:10.3389/fonc.2023.1074445.

124.

Søreide

Ismail

Roalsø

Ghotbi

Zaharia

. Early diagnosis of pancreatic cancer: clinical premonitions, timely precursor detection and increased curative-intent surgery. Cancer Control. 2023;30:10732748231154711. doi:10.1177/10732748231154711.

125.

Ardeshna

Cao

Rodgers

, et al. Recent advances in the diagnostic evaluation of pancreatic cystic lesions. World J Gastroenterol. 2022;28(6):624-634. doi:10.3748/wjg.v28.i6.624.

126.

Wei

Mao

Guo

Wang

. Computer-aided differentiates benign from malignant IPMN and MCN with a novel feature selection algorithm. Math Biosci Eng. 2021;18(4):4743-4760. doi:10.3934/mbe.2021241.

127.

Xie

Guo

Zhang

Wang

. Preoperative differentiation of pancreatic mucinous cystic neoplasm from macrocystic serous cystic adenoma using radiomics: preliminary findings and comparison with radiological model. Eur J Radiol. 2020;122:108747. doi:10.1016/j.ejrad.2019.108747.

128.

Machicado

Koay

Krishna

. Radiomics for the diagnosis and differentiation of pancreatic cystic lesions. Diagnostics. 2020;10(7):505. doi:10.3390/diagnostics10070505.

129.

Shur

Doran

Kumar

, et al. Radiomics in oncology: a practical guide. Radiographics. 2021;41(6):1717-1732. doi:10.1148/rg.2021210037.

130.

Dong

Yang

Tian

, et al. Pancreatic neuroendocrine tumor: prediction of tumor grades by radiomics models based on ultrasound images. Br J Radiol. 2023;96(1149):20220783. doi:10.1259/bjr.20220783.

131.

Zhang

Song

, et al. Clinical features and computed tomography radiomics-based model for predicting pancreatic ductal adenocarcinoma and focal mass-forming pancreatitis. Technol Cancer Res Treat. 2023;22:15330338231180792. doi:10.1177/15330338231180792.

132.

Zhao

, et al. The progress of PET/MRI in clinical management of patients with pancreatic malignant lesions. Front Oncol. 2023;13:920896. doi:10.3389/fonc.2023.920896.

133.

Huang

Zhou

Luo

, et al. Development and validation of a radiomics model of magnetic resonance for predicting liver metastasis in resectable pancreatic ductal adenocarcinoma patients. Radiat Oncol. 2023;18(1):79. doi:10.1186/s13014-023-02273-w.

134.

Bhasker

Kolbinger

Skorobohach

, et al. Prediction of clinically relevant postoperative pancreatic fistula using radiomic features and preoperative data. Sci Rep. 2023;13(1):7506. doi:10.1038/s41598-023-34168-x.

135.

Koch

Weitzer

Dos Santos

, et al. Multiparametric detection and outcome prediction of pancreatic cancer involving dual-energy CT, diffusion-weighted MRI, and radiomics. Cancer Imag. 2023;23(1):38. doi:10.1186/s40644-023-00549-8.

136.

Berbís

Paulano Godino

Royuela Del Val

Alcalá Mata

Luna

. Clinical impact of artificial intelligence-based solutions on imaging of the pancreas and liver. World J Gastroenterol. 2023;29(9):1427-1445. doi:10.3748/wjg.v29.i9.1427.

137.

Amaral

Oliveira

Donato

Tralhão

. Pancreatic cancer biomarkers: oncogenic mutations, tissue and liquid biopsies, and radiomics-A review. Dig Dis Sci. 2023;68(7):2811-2823. doi:10.1007/s10620-023-07904-6.

138.

Zhou

Chen

, et al. Fully automated magnetic resonance imaging-based radiomics analysis for differentiating pancreatic adenosquamous carcinoma from pancreatic ductal adenocarcinoma. Abdom Radiol (NY). 2023;48(6):2074-2084. doi:10.1007/s00261-023-03801-8.

139.

Hecht

Khatri

Morgan

, et al. Intraductal papillary mucinous neoplasm (IPMN) of the pancreas: recommendations for Standardized Imaging and Reporting from the Society of Abdominal Radiology IPMN disease focused panel. Abdom Radiol (NY). 2021;46(4):1586-1606. doi:10.1007/s00261-020-02853-4.

Risk Assessment and Radiomics Analysis in Magnetic Resonance Imaging of Pancreatic Intraductal Papillary Mucinous Neoplasms (IPMN)

Abstract

Keywords

Introduction

Guidelines of the Management of IPMN: What the Radiologist is Concerned With

Diagnosis: The role of MRI and magnetic resonance cholangiopancreatography

Report:Wwhat the radiologist should consider

TYPES of IPMN

Location

Size and Rate Growth

Wall and Septa

Mural Nodules

Multifocal BD-IPMN

Main duct dilation

Radiomics

Discussion

Conclusions

Footnotes

Acknowledgments

Declaration of Conflicting Interests

Funding

Ethical Statement

ORCID iD

Data Availability Statement

References