분지형 췌관내유두상점액종양의 장기 예후에 관한 연구: 단일 3차 의료기관에서의 경험
Clinical Outcomes over Several Years for Suspected Branch-Duct Intrapapillary Mucinous Neoplasms of the Pancreas: A Single Referral Center Experience
Article information
Abstract
배경/목적
본 연구에서는 췌장 낭종을 가진 환자들 중 임상적으로 분지형 췌관내유두상점액종양(branch-duct intraductal papillary mucinous neoplasm)으로 여겨지는 환자에서 췌장암 및 우려되는 특징의 누적 발생률을 조사하였다.
방법
분지형 췌관내유두상점액종양으로 추적관찰한 환자 177명의 자료를 후향적으로 분석하였다. 영상 검사를 통해 주췌관과의 소통이 확인된 췌장 낭종을 ‘분지형 췌관내유두상 점액종양 의심’으로 정의하였고, 주췌관과의 명확한 소통이 없는 경우를 ‘분지형 췌관내유두상점액종양 추정’으로 정의하였다.
결과
환자 중 30.5%가 분지형 췌관내유두상점액종양 의심으로 분류되었다. 진단 시의 중위 연령은 64세였으며, 낭종 크기의 중앙값은 13 mm였다. 진단 시 19명(10.7%)의 환자가 우려되는 특징을 보였으며, 고위험 인자(high risk stigmata)를 보인 환자는 없었다. 70.5개월의 추적 기간 동안 3명의 환자에게서 췌장암이 발생하였으며 96개월 시점의 누적 발생률은 4.0% (95% 신뢰 구간[confidence interval, CI], 0.98-10.76%)였다. 진단 시 우려되는 특징이 없는 158명의 환자에서 우려되는 특징의 누적 발생률은 24개월 시점에 5.1% (95% CI, 2.37-9.28%)였고, 96개월 시점에 11.2% (95% CI, 6.34-17.53%)였다. 우려되는 특징의 발생은 분지형 췌관내유두상점액종양 추정 군(p=0.102)과 75세 미만 환자(p=0.463)에서 더 낮았으나, 이러한 차이는 통계적으로 유의하지 않았다. 특히, 75세 이상 환자 군에서는 2년이 지난 이후 우려되는 특징은 발생하지 않았다.
결론
저위험 췌장 낭종 환자에서 췌장암 및 우려되는 특징의 누적 발생률은 현저히 낮았다. 또한, 첫 2년 동안 변화가 없는 고령 환자의 경우 감시를 중단하는 것을 고려할 수 있다.
Trans Abstract
Background/Aim
This study evaluated the cumulative incidence of pancreatic cancer and worrisome features (WFs) in patients with pancreatic cysts (PCs) clinically considered as branch-duct intrapapillary mucinous neoplasm (BD-IPMN).
Methods
Data from 177 patients followed for BD-IPMN were retrospectively analyzed. Suspected BD-IPMN was defined as PCs with confirmed communication with the main pancreatic duct (MPD) via imaging studies, while presumed BD-IPMN referred to PCs lacking clear MPD communication.
Results
Among the patients, 30.5% were categorized as suspected BD-IPMN. The median age at diagnosis was 64 years, with a median cyst size of 13 mm. At diagnosis, 19 patients (10.7%) exhibited WFs, and none of the patients had high-risk stigmata. Over a follow-up period of 70.5 months, pancreatic cancer developed in three patients, resulting in a cumulative incidence of 4.0% (95% confidence interval [CI], 0.98 to 10.76%) at 96 months. For the 158 patients without initial WFs, the cumulative incidence of WFs was 5.1% (95% CI, 2.37 to 9.28%) at 24 months and 11.2% (95% CI, 6.34 to 17.53%) at 96 months. The development of WFs was lower in the presumed BD-IPMN group (p=0.102) and among patients aged under 75 (p=0.463), though these differences were not statistically significant. Notably, the incidence of WFs plateaued after two years in the elderly cohort.
Conclusions
The cumulative incidence of pancreatic cancer and WFs in patients with low-risk PCs was notably low. Furthermore, discontinuing surveillance may be considered for elderly patients, especially those with no changes within the first two years of observation.
INTRODUCTION
The incidence of pancreatic cysts (PCs) has been rapidly increasing, attributed to the growing prevalence of an aging population and the more widespread use of advanced imaging techniques in recent years [1]. Among the various types of PCs, the branch-duct intrapapillary mucinous neoplasm (BD-IPMN) is particularly noteworthy due to its malignant potential [2]. Despite this potential, most patients with BD-IPMN undergo surveillance rather than surgery, as the cumulative incidence of pancreatic cancer has been reported to be relatively low [3-9].
Guideline from International Association of Pancreatology (IAP) emphasizes the importance of identifying worrisome features (WFs) and high-risk stigmata (HRS) in making surgical decisions [2]. Similarly, the American College of Gastroenterology (ACG) guideline advocates for the evaluation of these risk factors [10]. Marchegiani et al. [11] have suggested that the presence of stable WFs indicates the lowest risk of malignancy, whereas the development of additional WFs or HRS is associated with high-grade dysplasia. Consequently, the current management strategy for BD-IPMN involves surveillance, focusing on changes in WFs or HRS [2,10,11].
However, two significant issues exist. Firstly, accurately diagnosing of BD-IPMN with imaging studies is challenging, especially when the cysts are very small [1]. The imaging findings of BD-IPMN can range from an isolated small PC to diffuse multifocal cysts [12]. Among various imaging characteristics, the most critical finding for diagnosis is the communication with the main pancreatic duct (MPD) [2]. Yet, confirming this communication is challenging in real-world clinical settings. As a result, many patients with PCs that do not demonstrate communication with the MPD are clinically managed as BD-IPMN cases, owing to the malignant potential of BD-IPMN and diagnostic uncertainty [7,13]. Secondly, there is the issue of determining the appropriate age limit for follow-up. Current guidelines suggest that patients with BD-IPMN should be followed-up until they are not surgical candidates [2,10,14,15]. However, such surveillance can lead to a reduction of quality of life due to increased anxiety and stress among patients [8].
Therefore, this study aimed to investigate the cumulative incidence of pancreatic cancer and the occurrence of WFs in patients with PCs, who were clinically followed-up as having BD-IPMN. Additionally, this study sought to compare the clinical outcomes based on the confirmation of communication with the MPD and patient age.
METHODS
1. Study population
The study population consisted of 580 patients diagnosed with PCs who presented at Daegu Catholic University Medical Center between January 2015 to December 2017. Eligibility for inclusion was determined based on diagnostic codes (K862, D377, D136, and C253) from the 8th revision of Korean Standard Classification of Diseases (Fig. 1). The exclusion criteria were as follows: 1) patients who were not followed up for at least 2 years, 2) those who, after a review of medical records, were found not to have PC, 3) patients who underwent surgery within a year of diagnosis, and 4) those whose cyst size was less than 5 mm.
Flowchart of the study population. PCN, pancreatic cystic neoplasm; BD-IPMN, branch-duct intrapapillary mucinous neoplasm; SCN, serous cystic neoplasm. *Others pancreatic cysts included 7 main-duct IPMN, 3 SCN, 2 mucinous cystic neoplasm, 2 mixed IPMN, and 1 solitary pseudopapillary neoplasm. †Others included 3 who underwent surgery within 3 months of diagnosis and 1 with a history of previously diagnosed pancreatic cancer.
2. Study design
This retrospective cohort study was conducted at a single tertiary care center. The data collected from medical records included demographic profile, serum carbohydrate antigen 19-9 (CA 19-9), number of cysts, cyst size and location, imaging findings, and surgical resection of the PC. This study was performed in compliance with the ethical guidelines of the revised Helsinki Declaration of 2013, and the study protocol was reviewed and approved by the Institutional Review Board of Daegu Catholic University Medical Center (CR-23-158). As this study was a retrospective analysis, the need for informed consent was waived.
3. Definitions
Classification for PCs followed Mukewar et al. [13], distinguishing between suspected BD-IPMN, presumed BD-IPMN, and indeterminate BD-IPMN/serous cystic neoplasm (SCN) based on specific criteria. A cyst was defined as a suspected BD-IPMN when communication with the MPD was confirmed via imaging study. Presumed BD-IPMN was defined as cysts lacking clear MPD communication but were followed-up as BD-IPMN. Indeterminate BD-IPMN/SCN was referred to a cluster of microcysts or macrocysts without MPD communication and a central scar. WFs and HRS were defined according to the IAP guideline [2]. HRS are characterized by the following: 1) obstructive jaundice in a patient with a cystic lesion of the head of the pancreas, 2) enhancing mural nodule >5 mm, 3) MPD >10 mm. WFs include: 1) cyst >3 cm, 2) enhancing mural nodule <5 mm, 3) thickened/enhancing cyst walls, 4) main duct size 5-9 mm, 5) abrupt change in the caliber of the pancreatic duct with distal pancreatic atrophy, 6) lymphadenopathy, 7) increased serum level of CA19-9, 8) cyst growth rate >5 mm/2 years, or a history of pancreatitis. The duration of follow-up was defined as the interval from the initial diagnosis of PCs to either the diagnosis of pancreatic cancer, surgical resection of the cyst, or the date of the last follow-up. Furthermore, new onset diabetes mellitus (DM) was defined as the diagnosis of DM within two years prior to the identification of PCs.
4. Follow-up strategy
Follow-up was in line with IAP guidelines and tailored by attending physicians. All patients underwent computed tomography (CT) scans, while magnetic resonance imaging (MRI) and endoscopic ultrasonography (EUS) were performed based on clinical judgment. Patients developing WFs or HRS were recommended for surgical resection.
5. Statistical analysis
Statistical analysis was conducted using R statistical software (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria), with the ‘tidycmprsk’ package for competing risk analysis. The chi square or Fisher’s exact test was used to compare categorical variables. Competing risk analysis was applied to estimate cumulative incidence of pancreatic cancer or WFs, considering surgical resection as a competing event. Cumulative incidence was described with 95% confidence interval (CI). Statistical significance was defined as a p-value of <0.05 (two-tailed).
RESULTS
1. Baseline characteristics of patients with PCs
A total of 177 patients were followed-up during a median period of 70.5 months. The baseline characteristics of these patients are described in Table 1. The median age at diagnosis was 64 years, with males comprising 40.7% (n=72) of the cohort. The median Charlson’s Comorbidity Index was 2. Fifty-four patients (30.5%) were classified as having suspected BD-IPMN, 121 patients (68.4%) as presumed BD-IPMN, and two patients (1.1%) as indeterminate BD-IPMN/SCN. The proportion of patients with DM was 21.5%. All patients underwent CT scan. Fifty-six patients (31.6%) had EUS, and 76 patients (42.9%) underwent MRI.
Baseline characteristics of patients with pancreatic cysts (n=177)
The pancreatic body was the most common location for PCs (48.0%), followed by pancreatic head/uncinate (39.5%), pancreatic neck (14.7%), and pancreatic tail (14.7%). One hundred thirty-nine patients (78.5%) had a single cyst at diagnosis. Twenty-seven patients (15.3%) had two cysts, and 11 patients (6.2%) had three or more cysts. The median largest diameter of cyst was 13 mm. None of the patients had HRS at diagnosis, while 19 patients (10.7%) presented with WFs.
2. Occurrence of WFs and HRS
Among the 177 patients, 19 patients (10.7%) exhibited WFs at the time of diagnosis, and an additional 15 patients (8.5%) developed WFs during the follow-up period. The specific characteristics of these WFs are shown in Table 2. At diagnosis, the most prevalent was a cyst size greater than 3 cm (n=13, 7.3%), followed by an elevated serum CA 19-9 level (n=6, 3.4%), episodes of pancreatitis (n=2, 1.1%), and MPD diameter ranging from 5 mm to 9 mm (n=2, 1.1%). Conversely, during the follow-up, the most frequent newly emerged WF was cystic growth rate exceeding 5 mm over 2 years (n=13, 7.3%), followed by cyst size greater than 3 cm (n=4, 2.3%), episodes of pancreatitis (n=1, 0.6%), and MPD diameter ranging from 5 mm to 9 mm (n=1, 0.6%). The median largest cyst size was 13.0 mm at diagnosis, increasing to 14.0 mm at the final examination. Notably, cyst size growth was observed in 87 patients (49.2%), with a growth of over 5 mm documented in 41 patients (23.2%).
Characteristics of worrisome features and high-risk stigmata in the patients (n=177)
3. Cumulative incidence of pancreatic cancer and WFs
During the follow-up, pancreatic cancer developed in three patients. The detailed characteristics of these patients are shown in Table 3. The intervals from the diagnosis of PCs to the diagnosis of pancreatic cancer were 76.3, 66.0, and 84.9 months for each case, respectively. At the time of pancreatic cancer diagnosis, their ages were 84, 71, and 77 years, respectively. One patient belonged to the suspected BD-IPMN group and was the only one with WFs. Although all three patients had potentially resectable disease, they opted against additional evaluation and treatment, and therefore tissue confirmation was not attempted. The cumulative incidence of pancreatic cancer was 4.0% (95% CI, 0.98 to 10.76%) at 96 months (Fig. 2A). Additionally, during the follow-up period, 15 patients (8.5%) developed new WFs. The cumulative incidence rates of WFs in patients who did not present with WFs at the initial diagnosis of PCs were 5.1% (95% CI, 2.37 to 9.28%) at 24 months and 11.2% (95% CI, 6.34 to 17.53%) at 96 months (Fig. 2B).
Characteristics of pancreatic cancer development in patients with pancreatic cysts
Cumulative incidence of pancreatic cancer and worrisome features. (A) Cumulative incidence of pancreatic cancer in patients with clinically suspected branch-duct intrapapillary mucinous neoplasm. (B) Cumulative incidence of worrisome features in patients without worrisome features at diagnosis.
4. Cumulative incidence of WFs by age and classification
The analysis of the cumulative incidence of WFs categorized by patient age and confirmation of MPD communication is presented in Fig. 3. The analysis did not show a statistically significant difference in cumulative incidence by age, with a hazard ratio (HR) of 1.75 (95% CI, 0.394-7.740; p=0.463). For the group aged 75 or younger, the cumulative incidence was 4.9% (95% CI, 2.17 to 9.39%) at 24 months and 10.7% (95% CI, 5.83 to 17.33%) at 96 months. Conversely, for the group older than 75 years of age, the cumulative incidence was 6.3% (95% CI, 0.36 to 25.48%) at 24 months and 12.5% (95% CI, 1.90 to 33.57%) at 96 months.
Cumulative incidence of worrisome features. (A) Cumulative incidence of worrisome features categorized by age. (B) Cumulative incidence of worrisome features categorized by confirmation of main pancreatic duct communication. BD-IPMN, branch-duct intrapapillary mucinous neoplasm.
Regarding the classification of PCs, there was a trend toward a higher cumulative incidence of WFs in the suspected BD-IPMN group, with an HR of 2.31 (95% CI, 0.847-6.300; p=0.102). In the presumed BD-IPMN group, the cumulative incidence was 3.6% (95% CI, 1.19 to 8.41%) at 24 months and 9.1% (95% CI, 4.03 to 16.69%) at 96 months. In comparison, the suspected BD-IPMN group had a cumulative incidence of 8.5% (95% CI, 2.68 to 18.69%) at 24 months and 16.1% (95% CI, 6.91 to 28.73%) at 96 months.
5. Clinical features of resected PCs
During the follow-up period, surgery was performed on 15 patients. The clinical characteristics of these resected PCs are detailed in Table 4. The most common reason for surgery was an increase in cyst size, followed by patient’s preference, and the occurrence of pancreatitis. There was no difference in the reasons for surgery between the two groups. Among the patients with suspected BD-IPMN, six out of seven (85.7%) were definitively diagnosed with IPMN after surgery, and one patient was diagnosed with mucinous cystic neoplasm. In contrast, of the eight patients initially classified as having presumed BD-IPMN, only two (25.0%) were confirmed to have IPMN following histological examination (p=0.054). Finally, there was no high-grade dysplasia or invasive carcinoma confirmed in any of the patients.
Clinical features of resected cysts
DISCUSSION
IPMNs are clinically significant due to their potential as precursors of pancreatic cancer, typically following the classic ‘adenoma-carcinoma sequence’ [16]. The WFs and HRS outlined by the IAP guideline are recognized as significant predictors of malignant transformation.2 Similarly, the ACG guidelines identify comparable high-risk characteristics for mucinous PCs, such as cyst size, pancreatic duct dilatation, and the presence of mural nodules [10].
Despite these risks, not all patients diagnosed with BD-IPMN undergo surgical intervention, given the relatively low rates of malignancy transformation, particularly in the absence of such risk factors. In this study, the cumulative incidence of pancreatic cancer was observed to be 4.0% at 96 months. Within our cohort, no patients exhibited HRS, and only 10.7% had WFs at the time of diagnosis. This result is in line with previous research on low-risk patients [7,9,17]. Ciprani et al. [17] reported that a cumulative incidence of pancreatic cancer at 0.94% at five years and 2.6% at eight years among patients with PCs smaller than 15 mm, lacking WFs and HRS. Similarly, Yoshioka et al. noted a cumulative incidence of 2.8% at five years for patients with low-risk PCs [7]. A recent Korean study by Ahn et al. [9] reported that pancreatic cancer developed in 1.8% of patients within a 10-year follow-up period who had no WFs at diagnosis. Such evidence supported that long-term risk of malignancy remains low in patients with low-risk PCs.
BD-IPMN is defined as PCs with a size greater than 5mm and confirmed communication with the MPD [2]. However, verifying this communication presents challenges in real-world settings. To address this, Mukewar et al. introduced the concepts of suspected and presumed BD-IPMN, focusing on clear communication with MPD and clinical management [13]. In their study, suspected BD-IPMN was associated with a higher risk of developing pancreatic cancer compared to presumed BD-IPMN [13].
In this study, pancreatic cancer developed in three cases: one within the suspected BD-IPMN group and two within the presumed BD-IPMN group. Due to the limited number of cases, a statistical comparison between the group was not feasible. Therefore, we focused on comparing the cumulative incidence of WFs in patients who did not have WFs at diagnosis. This comparative analysis revealed a higher, though not statistically significant, between the suspected and presumed BD-IPMN groups showed a higher, cumulative incidence of WFs in the suspected BD-IPMN group, aligning with outcomes reported in prior research [13].
Furthermore, we evaluated the cumulative incidence of WFs according to age, specifically comparing groups aged over 75 to those aged 75 or under. The findings suggested a relatively higher cumulative incidence of WFs among the elderly, although the difference was not statistically significant. Interestingly, the incidence appeared to plateau after two years in the elderly cohort. Currently, there is no consensus on the optimal duration for surveillance. The American Gastroenterological Association (AGA) guideline recommend discontinuing surveillance if no significant changes in the characteristics of PCs are observed after five years [15]. And recently, the revised IAP guideline recommended that surveillance may be discontinued for patients with cysts smaller than 20 mm showing no morphological changes and no WFs after 5 years of surveillance [18]. Conversely, other guidelines, including those from ACG and European guideline, do not specify a particular time cutoff [10,14]. Additionally, the process of surveillance itself can negatively impact patients’ perceptions of their health and contribute to psychological distress[8]. A recent study suggested that a stable observation period of seven years could justify discontinuing surveillance[4]. Our findings support this approach, advocating for a reevaluation of the need for continued surveillance in cases that remain stable, particularly among elderly patients.
In this study, surgery was performed on 15 patients. postoperatively, 85.7% of the PCs in the suspected BD-IPMN group were confirmed as IPMN, in stark contrast to only 25.0% in the presumed BD-IPMN group. These outcomes are in close alignment with those from a previous study, which reported confirmation rates of 95.7% for suspected BD-IPMN versus 43.8% for presumed BD-IPMN [13]. This discrepancy underscores a critical insight: the absence of communication with the MPD in imaging studies significantly lowers the probability of a cyst being classified as BD-IPMN. Nonetheless, this does not categorically rule out the possibility of BD-IPMN. Accordingly, our study’s findings advocate for the continued surveillance of all PCs, suggesting, however, that the intensity of such monitoring might be reduced in cases of presumed BD-IPMN.
There were several limitations in this study. First, its retrospective, single-center design may limit the generalizability of the findings. Furthermore, the follow-up schedule was determined based on individual physician discretion, potentially introducing variability in surveillance intensity and duration. Second, the identification of pancreatic cancer was confined to only three cases, none of which were verified through pathological examination, possibly affecting the accuracy of cancer diagnosis. Third, not all patients underwent EUS or MRCP, which are known for their high sensitivity in detecting communication with MPD, possibly leading to an underestimation of such features. Despite these limitations, the study benefits from an extensive review of a large patient cohort without WFs at diagnosis over a prolonged period.
In conclusion, this study has demonstrated that the cumulative incidence of pancreatic cancer in patients with PCs, who lacked HRS at diagnosis, was not high. Furthermore, among patients who initially lacked WFs, elderly patients predominantly developed most WFs within the first 2 years. Therefore, while our findings support the continuous surveillance of all PCs, they also suggest considering a less intensive monitoring strategy, specifically for elderly patients.
Notes
Conflict of Interest
Jimin Han is currently serving as an Editor-in-Chief in Editorial Board of the Korean Journal of Pancreas and Biliary Tract; however, Jimin Han was not involved in the peer reviewer selection, evaluation, or decision process of this article. Han Taek Jeong has no potential conflicts of interest to declare.
AUTHOR CONTRIBUTIONS
Conceptualization: JMH; Data curation: HTJ; Formal analysis: HTJ, JMH; Investigation: HTJ, JMH; Methodology: JMH, HTJ; Supervision: JMH; Writing-original draft: HTJ; Writing-review & editing: HTJ, JMH.