The growing use and exponential improvement in the quality of imaging techniques coupled with an aging population has led to a drastic increase in the incidental detection of pancreatic cystic lesions, as testified by the 14-fold increase in the incidence rate of IPMNs between 1985 and 2005 ^[1][2][3][17,18,19]. Interestingly, the mortality rate for pancreatic adenocarcinoma and IPMN-associated PDAC did not increase over the same period, suggesting a stable number of pancreatic cystic lesions now detected by improved imaging modalities rather than a net increase in incidence ^[2][18]. Nonetheless, the true prevalence of these lesions remains unclear and varies significantly between studies due to differences in the timing of the study and the age of the population included ^[4][20]. In a meta-analysis encompassing 48,860 patients from 2008 to 2018, the pooled prevalence of incidentally noted pancreatic cysts was 8%, ranging from 2% to 3% in individuals undergoing computed tomography (CT) to 45% in individuals undergoing magnetic resonance imaging (MRI) ^[5][21]. Nevertheless, it reached up to 50% in autopsy studies ^[6][22]. Of note, a non-negligible proportion of pancreatic cystic lesions, especially those with small diameters, are usually not described in imaging reports in patients without a history of pancreatic disease (69% of cystic lesions with a mean diameter of 6 mm are not reported); thus, their prevalence could be even higher ^[7][23].

The increased detection of pancreatic cysts in recent years has triggered extensive diagnostic investigations to clarify their potential risk of malignancy, leading to significant accessibility issues and lengthy waiting lists. This healthcare burden is further amplified by the low estimated malignant progression rate of these lesions, particularly the ones <1 cm, which account for most of the cases ^[8][24], resulting in a large number of patients undergoing numerous imaging follow-up studies for many years. In fact, most guidelines state that there are no good long-term data to support the safety of discontinuing surveillance after long-term stability ^[9][10]. In addition, these patients also have an increased risk of developing new IPMNs and/or PDAC elsewhere in the whole pancreatic gland (about 1% per year); thus, they must be followed even after the surgical resection of the primary cystic lesion, further increasing the patient cohort requiring repeated imaging follow-up ^[10][11][25,26].

Despite international guidelines being considered cost-effective and clinically appropriate, it is undeniable that the cumulative healthcare expenses will continue to climb as the number of patients with IPMNs participating in follow-up protocols keeps rising ^[12][27]. From this perspective, a study published in 2015 estimated that if all patients with pancreatic cysts between 40 and 79 years old underwent MRI surveillance, the median cost would be USD 9.3 billion per year. If we assume that this program was 50% effective in reducing mortality from all PDAC, then this would cost approximately USD 1 million per cancer identified. This is a very conservative estimate because it does not include the cost of endoscopic ultrasound (EUS), further imaging evaluations for staging, and surgery; furthermore, it is not updated with the current epidemiological data ^[13][28].

Additionally, MRI studies for pancreatic cysts require long examination times and, therefore, take up both human and instrumental valuable resources, reducing their availability for other scopes and potentially engulfing the healthcare systems.

Therefore, in view of the rising detection of pancreatic cystic lesions, there is a growing need for optimization of the current screening and follow-up MRI protocol, in order to reduce both healthcare costs and waiting lists while still maintaining appropriate sensibility and specificity for the detection of malignant transformation. However, as reported by a recent bibliometric analysis, despite a marked increase until 2010, the number of articles regarding IPMN management has recently been somewhat stagnant ^[14][29]. The aim of the present paper was to analyze whether the implementation of an abbreviated-MRI (A-MRI) protocol for cystic pancreatic lesion surveillance could improve healthcare economics and reduce waiting lists in clinical practice without significantly reducing diagnostic accuracy.

The Fukuoka guidelines proposed by the International Association of Pancreatology (IAP) ^[15][30] represent one the most followed guidelines in clinical practice since they are universally understood by physicians who treat pancreatic disease as well as by other referring physicians ^[16][17][31,32]. After their first publication in 2006 ^[18][33], these guidelines were revised in 2012 ^[19][34], 2017 ^[9][10], and finally, in 2022 ^[15][30] during the 26th meeting of the IAP.

According to the IAP guidelines, several predictors for malignancy must be assessed to guide further management recommendations in patients with IPMNs, based on both clinical and radiological evaluation. Among them, the parameters that can be assessed in surveillance imaging studies are considered the most helpful and significant to establish the actual risk of malignant degeneration. In particular, the presence of an enhancing mural nodule or an enhancing thickened cyst wall have proven to be highly predictive of malignancy since they have been observed in 36–70% and 65% of IPMN patients with invasive disease, respectively ^{[20][21][22][23][24]}[35,36,37,38,39]. In particular, an enhancing mural nodule ≥5 mm is considered a “high-risk stigma” and thus requires surgical resection, whereas an enhancing nodule <5 mm and an enhancing thickened wall are regarded as “worrisome features” and should mandate further evaluation with EUS before intervention. Due to the clinical relevance of these radiological findings, and particularly the importance of discriminating between neoplastic solid components and mucus plugs or debris ^[25][40], gadolinium-enhanced MRI with magnetic resonance cholangiopancreatography (MRCP) has been widely recommended by IAP as the procedure of choice for evaluating a pancreatic cyst, based on superior contrast resolution as well as the advantage of avoiding radiation exposure ^[9][26][27][10,41,42]. Consequently, it is not surprising that the majority of radiologists (70%) support the routine use of intravenous contrast for the characterization and follow-up of incidental pancreatic cystic lesions in clinical practice, as testified by a survey distributed by the Society of Abdominal Radiology Disease Focused Panel ^[17][32].

With the current emphasis on cost containment and the increasing rate of incidentally discovered pancreatic cystic lesions, several authors have questioned whether the contrast-enhanced sequences could be eliminated from MRI studies for pancreatic cystic lesions without affecting decision-making. Avoiding the administration of gadolinium through the application of an A-MRI protocol would potentially lead to a number of practical advantages. First, the additional cost of gadolinium would be eliminated ^[28][29][43,44]. Second, as the additional time related to inserting an intravenous catheter into an arm vein, connecting it to a power injector, and performing the contrast-enhanced sequences would no longer be required, the total examination time could be substantially shortened. By doing so, more slots in the daily schedule would become available and could substantially increase the workflow in the MRI suite. Moreover, the decreased time in the magnet would reduce the patient’s potential discomfort in the MRI machine due to claustrophobia and the necessity of lying in a flat position, as well as susceptibility to breathing and movement artifacts secondary to long examinations ^{[30][31][32][33]}[45,46,47,48]. Last but not least, concern about the potential risk of nephrogenic systemic fibrosis in patients with borderline renal function would be eliminated, as well as the concern for gadolinium deposition in various body tissues after multiple repeated administrations, thus further decreasing patient anxiety ^[34][35][36][49,50,51].

In 2009, Macari et al. ^[37][52] were the first to initially question whether intravenous contrast was required for MRI follow-up of cystic pancreatic lesions, reporting a retrospective analysis of surveillance MRI studies in 56 patients. The addition of the contrast-enhanced sequences resulted in the same reader interpretation compared to the unenhanced study, yielding a different management recommendation only in five cases (discordance of 4.5%). However, following consensus review, no additional findings that would specifically alter clinical decisions were identified in the contrast-enhanced images and differing recommendations were attributed to expected variations in categorizing lesions. More importantly, no cases concerning malignancy were missed using the A-MRI protocol. Similar results were also reported in the subsequent larger retrospective case review of 301 patients described by Nougaret et al. ^[28][43], where follow-up lesion assessment between the unenhanced A-MRI and the classic pancreatic protocol MRI with and without contrast administration was discordant only in 4.6% of cases. Interestingly, the discordant cases represented false-positive rather than false-negative findings and were related to reader misinterpretation of unenhanced T2-w sequences rather than a limitation of the non-enhanced technique. Here, no lesions with suspicious features for malignancy were missed on the A-MRI. In 2017, Pozzi-Mucelli et al. ^[38][53] assessed an even shorter surveillance A-MRI protocol in 154 patients consisting essentially of only T1- and T2-weighted sequences. They demonstrated no clinically significant differences in assessment cyst size, main pancreatic duct diameter, or presence of mural nodules between the proposed protocol and the longer comprehensive one, proving how clinical decision-making remained unaffected. In 2020, Kang et al. ^[39][54] were the first to report assessing the interobserver agreement and diagnostic accuracy for the presence of “worrisome features” and “high-risk stigmata” recently introduced in the latest Fukuoka revision. In their study, they concluded that there was a substantial inter-reader agreement for evaluating significant imaging features of pancreatic IPMNs using the proposed unenhanced A-MRI protocol, with high sensitivity and negative predictive value. Similarly, in 2022, Johansson et al. ^[40][55] reported that their ultra-short MRI protocol correctly allowed all true cystic mural nodules to be detected, with intra- and inter-observer agreement results comparable to the standard contrast-enhanced MRI protocol, but a significant reduction in both examination time and costs. Kierans et al. ^[41][56] also confirmed that the addition of gadolinium had no significant impact on the diagnosis of benign versus malignant pancreatic cystic lesions; moreover, they used cytopathology as the reference standard, further validating their results. Finally, Yoo et al. ^[42][57] also reported high diagnostic performance with their A-MRI protocol, although the results of subgroup analysis for patients with pathologically confirmed IPMNs revealed lower specificity due to the false-positive risk of mistaking true solid lesions with mucin plugs using unenhanced MRI. Therefore, they concluded that the high sensitivity of the A-MRI protocol regarding mural nodule detection should lead to further evaluation through full-sequence, contrast-enhanced MRI/MRCP or EUS rather than completely replacing them.

Besides demonstrating comparable efficacy in terms of sensibility and specificity, these relevant studies also confirmed the hypothesized practical advantages of adopting an A-MRI protocol in real clinical scenarios. First, the abbreviated approach resulted in significant cost savings of 61–75% per MRI examination. This would mean an estimated cost reduction by about EUR 50,000 for the potential lifelong surveillance of a 45-year-old patient with a pancreatic cystic lesion that never develops worrisome imaging features ^[28][38][40][43,53,55]. Second, the A-MRI protocol required only 5–20 min, thus providing an unquestionable reduction in the time of acquisition (from 10 to 27 min according to the standard MRI protocol used) and case reading ^{[28][38][39][40]}[43,53,54,55]. Therefore, besides improving the MRI experience for patients, the drastic reduction of scanning time would ensure the execution of roughly 50% more MRI studies, drastically reducing both waiting lists and delays in early diagnosis (Table 1) ^[43][58].

Although initial results date back to 2009 and showed promising and practical attractiveness, the most recent revision of the IAP guidelines does not specifically acknowledge the possibility of adopting an A-MRI protocol for the screening and monitoring of pancreatic cystic lesions. Actually, the authors do not even give consideration to this topic or cite the works by Macari et al. ^[37][52], Nougaret et al. ^[28][43], and Pozzi-Mucelli et al. ^[38][53], but rather strongly recommend the use of contrast enhancement in any radiologic examination ^[15][44][30,59]. Specific comments on the potential applicability of an A-MRI scanning protocol for cystic pancreatic lesion surveillance are also lacking in most of the other major clinical guidelines. For example, despite the variations in the specifics (i.e., “solid component” rather than enhancing mural nodules), the American Gastroenterological Association (AGA) Institute Guidelines ^[10][25] do not specifically address the issue of whether contrast administration is necessary for an adequate evaluation of pancreatic cysts, and an unspecified “high-quality” MRI with MRCP is proposed as the optimal imaging method in its technical review ^[45][8]. Similarly, the American College of Gastroenterology (ACG) 2018 guidelines do not specify whether mural nodules and/or solid components should present enhancement to be considered a high-risk characteristic, nor recommend a specific MRI protocol for pancreatic evaluation ^[11][26]. The American College of Radiology (ACR) promotes the execution of a standard pancreatic MRI protocol with contrast-enhanced sequences since it “may help detect enhancement within mural nodules (high-risk stigmata) and the pancreatic phase improves the ability to detect metachronous PDAC elsewhere”. Nonetheless, they at least acknowledge that routinely using contrast material for MRI follow-up is controversial ^[46][60]. Finally, despite reporting the results by Macari et al. ^[37][52] and Pozzi-Mucelli et al. ^[38][53] in evaluating a shortened protocol, the guidelines proposed in 2018 by the European Study Group concluded that “no definite MRI or CT protocol can be recommended for the diagnosis or surveillance of patients with pancreatic cystic lesions because of the wide spread of published data and the lack of dedicated comparative studies” ^[47][7]. Intriguingly, only the recent update of the guidelines from the Korean Society of Abdominal Radiology advises towards a shortened MRI protocol for surveillance of pancreatic cystic lesions, especially in patients with impaired renal function since it can provide “sufficient information equivalent to the standard MR protocol”. In particular, based on the previous works by Macari et al. ^[37][52] and Nougaret et al. ^[28][43], they suggest that a minimum of axial and coronal heavily T2-w and axial T1-w sequences should be performed ^[48][61].

The concept of “enhancement” of mural nodules was already mentioned in the second update of the IAP guidelines published in 2012 ^[19][34]. Among the cited references regarding the diagnostic work-up for cystic lesions of the pancreas ^{[49][50][51][52][53][54][55][56][57][58]}[62,63,64,65,66,67,68,69,70,71], only Ohno et al. ^[59][72] have investigated whether the “enhancement” of mural nodules can be related to a higher risk for malignancy but, interestingly, their results were based solely on contrast-enhanced EUS and CT findings. Similarly, among the works cited in the 2017 update ^[9][10], only Kang et al. ^[60][73] evaluated the enhancement of mural nodules but, again, defined them as protrusions that enhanced after the administration of a contrast agent on CT or present with blood flow on EUS, thus not providing any MRI-related data. Finally, in the latest revision published in 2023 ^[15][30], the authors discussed only whether the cut-off value for a mural nodule to be considered a high-risk stigmata should be raised to ≥10 mm. In any case, despite the apparent lack of evidence justifying the need for the administration of contrast medium during MRCP/MRI examinations for IPMNs, it is straightforward that the main problem that could arise from its absence is related to the potential risk for misinterpreting a true neoplastic mural nodule for a false nodule (mucus plugs or debris), and vice versa. Nevertheless, from the data published by Nougaret et al. ^[28][43], Pozzi-Mucelli et al. ^[38][53], Kang et al. ^[39][54], Johansson et al. ^[40][55], Kierans et al. ^[41][56], and Yoo et al. ^[42][57], this risk exists in only sporadic cases and, most importantly, concerns almost exclusively false positives, since true nodules are generally correctly depicted on the T2-w sequences.

According to the IAP guidelines, enhancement evaluation should not be limited to mural nodules but is also a feature of the cystic wall. However, also in this case, by reviewing the works cited in both the 2017 and the 2023 Fukuoka updates in support of wall enhancement evaluation ^[9][15][10,30], none of them were dedicated to the study of this specific feature ^{[50][51][54][60][61][62]}[63,64,67,73,74,75]. Furthermore, as reported by several studies ^[39][63][64][54,76,77], cyst wall evaluation shows poor inter-reader agreement even in standard contrast-enhanced MRI protocols due to its subjectivity and measurement error. Therefore, the clinical-pathological significance of an enhancing thickened cyst wall should still be considered controversial and further studies are needed before justifying the need to administer contrast to achieve its characterization.

In addition, the other “high-risk stigmata” and “worrisome features” proposed in the latest update of the Fukuoka guidelines are not affected by the administration or lack of contrast medium during MRI evaluation ^[15][30]. First, some of these predictive factors for malignancy, namely the presence of obstructive jaundice, acute pancreatitis, the new onset or acute exacerbation of diabetes within a year, and increased serum level of CA19-9, are an exclusive prerogative of clinicians; therefore, variations in imaging protocols would certainly not affect their assessment. Second, the other radiological parameters that need to be evaluated in patients with IPMNs, such as the diameter of the cyst (≥30 mm), its growth rate (≥2.5 mm/year), the size of the main pancreatic duct (>10 mm or 5–9 mm), and any possible abrupt change in its caliber, can be easily examined in unenhanced imaging. Finally, since the most reliable and reproducible finding suggestive of malignant lymphadenopathy is still size enlargement (generally based on the measurement of the short axis) ^[65][66][78,79], the lack of contrast administration would not significantly influence the diagnostic criteria for suspicious lymph nodes.

In conclusion, therefore, the impact of contrast-enhanced sequences in clinical decision-making for the surveillance of pancreatic cystic neoplasms seems quite limited, for both mural nodules and cystic walls, and thus its administration could potentially be skipped without significant practical implications.