Risk Factors for Progression in Patients Undergoing Surveillance for Pancreatic Cysts

Structured Abstract:

Objective:

To identify risk factors associated with progression of pancreatic cysts in patients undergoing surveillance.

Summary Background Data:

Previous studies of intraductal papillary mucinous neoplasms (IPMNs) rely on surgical series to determine malignancy risk and have inconsistently identified characteristics associated with IPMN progression.

Methods:

We conducted a retrospective review of 2197 patients presenting with imaging concerning for IPMN from 2010–2019 at a single institution. Cyst progression was defined as resection or pancreatic cancer development.

Results:

Median follow-up time was 84 months from presentation. The median age was 66 years, and 62% were female. Ten percent had a first-degree relative with pancreatic cancer, and 3.2% had a germline mutation or genetic syndrome associated with increased risk of PDAC. Cumulative incidence of progression was 17.8% and 20.0% at 12- and 60-months post-presentation, respectively. Surgical pathology for 417 resected cases showed non-invasive IPMN in 39% of cases and PDAC with or without associated IPMN in 20%. Only 18 patients developed PDAC after 6 months of surveillance (0.8%). On multivariable analysis, symptomatic disease (hazard ratio [HR] 1.58 [95% CI 1.25–2.01]), current smoker status (HR 1.58 [95% CI 1.16–2.15]), cyst size (HR 1.26 [95% CI 1.20–1.33]), main duct dilation (HR 3.17 [95% CI 2.44–4.11]), and solid components (HR 1.89 [95% CI 1.34–2.66]) were associated with progression.

Conclusions:

Worrisome features on imaging at presentation, current smoker status, and symptomatic presentation are associated with IPMN progression. Most patients progressed within the first year of presentation to MSKCC. Further investigation is necessary to develop personalized cyst surveillance strategies.

Mini Abstract:

The management of pancreatic cysts and identification of high-risk lesions remain challenging. In this retrospective review of 2197 patients with an imaging diagnosis of intraductal papillary mucinous neoplasms (IPMN), symptomatic disease, current smoker status, cyst size, main duct dilation, and presence of solid component were associated with IPMN progression.

Introduction

Owing to wider use of cross-sectional imaging, asymptomatic pancreatic cysts are increasingly diagnosed.^1–3 Incidental cysts are identified in nearly 3% of computed tomography (CT) scans and up to 20% of magnetic resonance imaging (MRI) scans, with an estimated 800,000 new pancreatic cysts diagnosed annually in the United States.^4–6 Approximately 80% of incidental cysts are intraductal papillary mucinous neoplasms (IPMNs),⁷ mucinous cysts with variable potential for malignant transformation.

Management of such incidental pancreatic lesions is largely based on presenting symptoms and the presence of worrisome radiologic features or high-risk stigmata. Accurate preoperative diagnosis is essential for appropriate treatment.³ Management of cystic lesions is based on the appearance on initial imaging and the application of consensus guidelines to lesions presumed to be IPMNs.⁸ The timeline of IPMN progression to carcinoma is unclear, and current guidelines recommend lifelong surveillance.⁹ Surgical series estimate risk of high-grade dysplasia or invasive disease as high as 70% for main duct IPMN (MD-IPMN) and up to 30 % for branch duct IPMN (BD-IPMN).^4,9

Surgical series likely overestimate the risk for all patients with IPMN, given the highly selected group who undergo resection based on high-risk clinical characteristics or radiologic features. Establishing an accurate denominator by including patients in surveillance and those that undergo surgical resection is critical to accurately estimate risk of pancreas cyst progression. As such, the goal of this study was to estimate the cumulative incidence of IPMN progression and identify demographic, clinical, and imaging features associated with IPMN progression in a large cohort of patients under surveillance.

Methods

Patients

With Institutional Review Board (IRB) approval, patients evaluated at Memorial Sloan Kettering Cancer Center (MSKCC) for a pancreatic cyst in gastroenterology or hepatopancreatobiliary surgery clinics from 2010–2019 were identified from a prospectively maintained database. The first imaging study report with a pancreatic cyst detected was reviewed to identify features, including cyst size, location, solid components, solid enhancement, calcifications, main duct size, atrophy, and pancreatitis. Patients were excluded if they had lesions suspicious for malignancy, including pancreatic ductal adenocarcinoma (PDAC), pancreatic neuroendocrine tumors (PNET), or metastatic disease from their known cancer or imaging features consistent with non-IPMN cystic lesions including mucinous cystadenoma, serous cystadenoma or pseudocyst. The surveillance group consisted of patients with a presumptive diagnosis of IPMN based on imaging features or endoscopic ultrasound. Cyst progression was defined a priori as changes in the cyst prompting surgical intervention or a diagnosis of pancreatic cancer given the significance to surgeons and their patients when these events occur as a consequence of failed surveillance. Progression was evaluated from presentation to August 2021.

Additionally, clinical characteristics associated with PDAC, including germline mutation status, personal and family history of malignancy, presence of jaundice or abdominal pain at presentation, and surgical pathology were retrospectively collected from the electronic medical record. Germline mutations (ATM, BRCA, PRSS1, CDKN2A, TP53, APC, CHEK2, RET, VHL, MEN1, NF1, PTEN, CDH1, MTHFR) and genetic syndromes (Lynch, Peutz-Jeghers) were selected based on the International Cancer of the Pancreas (CAPS) Consortium inclusion of these high-risk individuals in pancreatic cancer screening programs.¹⁰

Statistical Analysis

Baseline demographic and tumor characteristics are summarized using the frequency and percentage for categorical covariates, and the median and interquartile range (IQR) for continuous covariates. Cumulative incidence function was used to estimate the probability of IPMN progression from the time of the disease presentation. Patients who died during the surveillance period were considered as a competing event. Factors associated with the cumulative incidence of IPMN were evaluated using Fine and Gray regression. A multivariable competing risk regression was constructed by including factors associated with the risk of IPMN progression from the univariate analysis at P<0.05 level. Among patients who underwent resection or developed pancreatic cancer, overall survival from the time of disease progression or pancreatic cancer diagnosis was estimated using Kaplan-Meier methods and compared between degree of dysplasia on surgical pathology using the log-rank test.

All statistical analyses were performed using R Version 4.0.5 (R Foundation for Statistical Computing Vienna, Austria). All P-values were two-sided, and P <0.05 was considered statistically significant.

Results

Patient Demographics, Cyst Characteristics, and Surgical Pathology

A total of 2245 patients presented with imaging concerning for IPMN from 2010–2019, of which 2197 met the study’s inclusion/exclusion criteria and were included in our analysis. The majority (81%) of patients had their first imaging study performed at an outside institution. Median follow-up from presentation was 84 months. The study population was 62% (n=1369/2197) female, with a median age of 66 years old (IQR: 57–74) (Table 1). A germline mutation or genetic syndrome was present in 3.2% (n=70/2197), with BRCA alterations the most common, and 10% (n=219/2197) had a first-degree relative with pancreatic cancer. Imaging characteristics on the first imaging report (40% MRI, 54% CT, 6% ultrasound) are summarized in Table 1. The index cyst or largest cyst for patients with multiple lesions was a median size of 1.5 cm (IQR: 0.9–2.3). The majority of patients had a single cyst (n=1560/2197, 71%), with 47% located in the head of the pancreas (n=1037/2197) without worrisome features.

Table 1.

Clinical characteristics of patients presenting for pancreatic cyst evaluation and imaging characteristics from their first surveillance study (n=2197)

Clinical Characteristics		Imaging Characteristics
Age (years), Median (IQR)	66 (57–74)	Index cyst size (cm), Median (IQR)	1.5 (0.9–2.3)
Sex		Unknown	90
Female	1369 (62)	Index location
Male	828 (38)	Body	668 (30)
Symptoms at presentation	441 (20)	Head	1037 (47)
History of malignancy	754 (34)	Tail	492 (22)
Unknown	5	Number of lesions
History of pancreatitis	178 (8)	Multiple	637 (29)
Unknown	2	Single	1560 (71)
History of diabetes	353 (16)	Solid component	132 (6)
Unknown	5	Unknown	1
Family history of pancreatic cancer	339 (16)	Solid enhancement	30 (1)
Unknown	45	Calcifications	113 (5)
First-degree relative with pancreatic cancer	219 (10)	Unknown	2
Unknown	45	Branch duct dilation	31 (1)
Smoking history		Main duct dilation	211 (10)
Never	1399 (65)	Unknown	1
Former	570 (26)	Pancreatitis	27 (1)
Current	182 (9)	Unknown	2
Unknown	46
Genetic mutation	70 (3)

Data are n (%) unless noted.

Four hundred forty-three patients met criteria for progression, with 417 patients undergoing surgery and 26 patients developing unresectable pancreatic cancer during surveillance (Figure 1). The cumulative incidence of progression was 17.8% (95% CI: 16.2–19.4) and 20.0% (95% CI: 18.3–21.7) at 12- and 60-months post-presentation, respectively (Figure 2). Of the 417 patients who underwent resection, pathology demonstrated PDAC in 8% (n=35) of cases. Other pathologies included 11% (n=45) mucinous cystic neoplasms, 6% (n=23) neuroendocrine, and 3% (n=12) solid pseudopapillary. Other benign disease was present in 21% (n=89) including 30 serous cystadenomas, 22 retention cysts, 14 simple /epidermoid / epithelial cysts, 14 pseudocysts or pancreatitis related inflammation and fibrosis, and 9 acinar cystadenomas. Of the patients with IPMN confirmed on pathology (n=213), 34% were branch duct (n=72), 25% main duct (n=53), 23% mixed type (n=50), and 18% unspecified on pathology (n=38). Additionally, 57% (n=122) were low- to intermediate-grade dysplasia, 19% (n=41) high-grade dysplasia, and 23% (n=50) had IPMN-associated PDAC.

Figure 1.

Study flow diagram demonstrating classification of patient progression.

Abbreviations: PDAC, Pancreatic Ductal Adenocarcinoma; IPMN, Intraductal Papillary Mucinous Neoplasm; PNET, Pancreatic Neuroendocrine Tumor

Figure 2.

Cumulative incidence of disease progression was 17.8% (95% CI: 16.2–19.4) at 12 months and 20.0% (95% CI: 18.3–21.7) at 60 months post presentation.

In this cohort of 417 patients, 90-day mortality was 1.2%. Overall survival of patients with IPMN or PDAC who underwent surgery (n=248) varied by degree of dysplasia, with 91% (95% CI: 85%–96%) of patients alive at 5 years for IPMN with low-grade dysplasia, 81% (95% CI: 68%–95%) for IPMN with high-grade dysplasia, and 57% (95%CI: 47%–69%) for invasive disease (Figure 3).

Figure 3.

Overall survival for patients who underwent surgery (n=248) based on degree of dysplasia on surgical pathology (IPMN with low- to intermediate-grade dysplasia, high-grade dysplasia, and invasive disease).

Time to progression

Of the 85 patients with invasive disease on surgical pathology, resection was performed at a median of 22 days (IQR: 13–38) from presentation. Looking only at patients who were observed for at least 6 months following presentation, 55 patients progressed to surgery and only 11 (20%) showed PDAC on pathology. Similarly, of the 26 patients with PDAC who did not undergo resection, biopsy diagnosis occurred within a median of 29 days (IQR: 10–469) from presentation with only 7 patients (27%) progressing to unresectable PDAC after at least 6 months of observation. Patients with high grade dysplasia on pathology presented within a median of 28 days (IQR: 16–54), with 5 patients (12%) undergoing resection after 6 months of observation.

Patients presenting with high-risk stigmata on imaging, main duct dilation of 1 centimeter or greater (n=24) or solid component (n=111), also displayed differential progression in the first 6 months. Fifty-three percent of patients with a solid component progressed, with 93% (n=55) of those progressing in 6 months. Seventy-one percent of patients with main duct dilation of at least 1cm progressed with 88% (n=15) of those progressing within 6 months. Seventeen patients presented with both a solid component and a main duct of at least 1cm and 88% progressed, all within 6 months of presentation.

Factors Associated with Progression

In univariate analysis, patients were more likely to progress if they were male (hazard ratio [HR] 1.31 [95% CI: 1.08–1.58]), presented with symptoms (HR 2.52 [95% CI: 2.06–3.07]), had a history of pancreatitis (HR 2.25 [95% CI: 1.73–2.93]) or diabetes (HR 1.52 [95% CI: 1.21–1.9]), were current smokers (HR 2.92 [95% CI: 2.29–3.73]), or had worrisome features (i.e., solid component, calcifications, or main duct dilation5 mm or greater) on their first scan (Table 2). Symptomatic disease (HR 1.58 [95% CI 1.25–2.01]), current smoker status (HR 1.58 [95% CI 1.16–2.15]), cyst size (HR per 1-unit increase: 1.26 [95% CI 1.2–1.33]), main duct dilation (HR 3.17 [95% CI 2.44–4.11]), and solid components on imaging (HR 1.89 [95% CI 1.34–2.66]) remained significant on multivariable analysis.

Table 2.

Univariate and multivariable analysis to examine risk factors for IPMN progression

Variable	HR (95% CI)	P-value	HR (95% CI)a	P-value
Index cyst size	1.37 (1.3–1.45)	<0.001	1.26 (1.20–1.33)	<0.001
Sex: Male	1.31 (1.08–1.58)	0.005	0.94 (0.75–1.17)	0.58
Symptoms	2.52 (2.06–3.07)	<0.001	1.58 (1.25–2.01)	<0.001
History of malignancy	0.69 (0.56–0.85)	<0.001	0.95 (0.76–1.19)	0.66
History of pancreatitis	2.25 (1.73–2.93)	<0.001	1.34 (0.97–1.85)	0.073
History of diabetes	1.52 (1.21–1.90)	<0.001	1.23 (0.95–1.59)	0.11
Family history of PDAC	0.87 (0.66–1.15)	0.332	NA	NA
First-degree relative with PDAC	0.93 (0.67–1.28)	0.644	NA	NA
Mutation	0.61 (0.32–1.19)	0.149	NA	NA
Smoking status		<0.001		NA
Never	1.00		NA	NA
Former	0.94 (0.74–1.19)		0.99 (0.78–1.27)	0.96
Current	2.92 (2.29–3.73)		1.58 (1.16–2.15)	0.004
Index cyst location		0.002
Body	1.00
Head	1.19 (0.95–1.5)		0.88 (0.69–1.12)	0.29
Tail	1.56 (1.21–2.01)		1.28 (0.97–1.68)	0.077
Single cyst	1.57 (1.25–1.97)	<0.001	1.47 (1.15–1.89)	0.002
Solid component	5.08 (3.93–6.58)	<0.001	1.89 (1.34–2.66)	<0.001
Calcifications	2.15 (1.57–2.96)	<0.001	1.05 (0.68–1.64)	0.82
Main duct dilation (>5 mm)	5.13 (4.17–6.3)	<0.001	3.17 (2.44–4.11)	<0.001

Abbreviations: PDAC, Pancreatic ductal adenocarcinoma; NA, not applicable.

^aMultivariable competing risk regression constructed using variables from univariate analysis with p value <0.05 level

Discussion

In this retrospective study, we included all patients with a pancreatic cyst during a 10-year period at a single institution to determine factors associated with cyst progression. Symptomatic disease, current smoker status, cyst size, main duct dilation, and presence of solid component on imaging were significant predictors of disease progression, whereas germline mutation status and family history were not associated with progression. Cumulative incidence of progression was 17.8% at 12 months and 20.0% at 60 months, suggesting that the majority of patients underwent surgical intervention or developed PDAC within a year of presentation to MSKCC with minimal progression events thereafter.

The role of patient clinical characteristics and social factors associated with the development of PDAC have been inconsistently implicated in IPMN progression to malignancy. In a review of 362 IPMN patients, those with a family history of PDAC did not have an increase in invasive disease.¹⁰ In a separate study, family history of PDAC and personal history of any malignancy were associated with a diagnosis of MD-IPMN but not BD-IPMN.¹¹ Similarly, Nehra et al. found no difference in pathology for resected IPMN with or without a family history of PDAC; however, those with a family history were more likely to have concurrent PDAC.¹² In smaller retrospective studies, current smoking status was associated with earlier diagnosis of invasive IPMN or PDAC concomitant with IPMN^11,13 and patients were three times more likely to have malignancy on pathology if they had a smoking history.² Similarly, our study demonstrated an increased risk of surgery or development of PDAC in current smokers. With a larger cohort than prior studies, our study adds to the literature suggesting smoking status may be a modifiable risk factor in the prevention of malignant transformation in IPMN.

In a nomogram developed from our institution, solid component, weight loss, and cyst size were associated with dysplasia progression in BD-IPMN.¹⁴ For main and mixed duct IPMNs, male gender, solid component, weight loss, and history of prior malignancy were associated with increased risk of invasive disease.¹⁴ Multiple surgical series have validated the inclusion of high-risk features such as jaundice and enhancing solid component in the International Association of Pancreatology guidelines.^15–18 However, the risk estimated based on each of these features varies widely, and estimating patient-specific risk remains a challenge in the management of this disease. Marchegiani et al. found that only jaundice was significantly associated with the development of invasive cancer in patients with BD-IPMN.¹⁹ No single worrisome feature was associated with high-grade dysplasia on pathology. However, the development of additional worrisome features and progression from worrisome features to high-risk stigmata during surveillance were associated with high-grade dysplasia on pathology at the time of initial resection.¹⁹ This suggests that utilizing static measures or a single worrisome feature may not accurately predict risk of progression, and change over time should be considered.

The use of strict resection criteria still leads to potentially unnecessary resection (i.e., no malignancy or high-grade dysplasia on surgical pathology) of 40% of MD-IPMNs and 70% of BD-IPMNs.^1,20 This is consistent with our reported surgical pathology, where the majority of IPMN patients underwent surgical resection for low-grade disease. In a retrospective review of 1290 patients with pancreatic cysts at Massachusetts General Hospital, 76% underwent surgery shortly after their initial diagnosis.² Based on surgical pathology, 51% of patients had an IPMN and overall, 42% had high-grade dysplasia or invasive disease. Our study builds on existing literature demonstrating that the majority of patients will not progress during surveillance and those that do undergo surgical resection will not have high grade disease. The high incidence of progression within the first year is likely driven by referral bias at high volume centers and patients with known high-risk features presenting for further evaluation. This represents a cohort of patients that are likely common at tertiary care centers where patients with high-risk stigmata are referred once intervention is required. As such, the true cumulative incidence of progression in patients undergoing surveillance is likely lower than 20%.

To address this, the 2015 American Gastroenterological Association (AGA) guidelines included a conditional recommendation for cessation of surveillance if radiographic cyst stability at 5 years.²¹ In a multi-center retrospective review of 310 patients with cystic neoplasms on EUS with over 5 years of follow up, there was less than a 1% risk of malignant transformation and no patients without AGA high-risk features developed cancer after 5 years.²² This suggest that discontinuation of surveillance after 5 years may be justified in select patients. A similar study from MSKCC compared patients with at least 5 years of follow up with cyst stability to those with less than 5 years of follow up. Rate of progression to resection and developing cancer was significantly lower after 5 years with cumulative incidence of progression to cancer plateauing at 9 years of surveillance.²³ The decline in cancer occurrence and resection rates indicates that for the majority of patients their individual risk is small with mortality related to non-pancreatic causes to be more likely. In both studies, however, progression after 5 years was noted. Further studies are needed to assess appropriate surveillance strategies and imaging intervals in this patient population.

There are limitations of the present study. This is a retrospective study that relied on data available in the electronic medical record, involved review of imaging reports only, and does not capture the granularity of surgeon decision making. The majority of initial imaging studies were performed at outside institutions, introducing variation in imaging quality and radiologic findings. Also, the initial imaging study in our records might not reflect their first diagnosis of a pancreatic cyst; however, this reflects real clinical practice and decision-making based on the information provided at presentation. Additionally, we defined progression as only surgical resection or diagnosis of pancreas cancer. This does not include the other interventions that occur during surveillance, such as endoscopic ultrasound and more frequent imaging, which could also be indicators of cyst progression. We attempted to identify a cohort representative of all patient with IPMNs and identified a large portion of patients undergoing surgery soon after presentation. As a result, our cohort is heterogeneous and highlights that although we do not have surgical pathology for all patients, there are likely patients being followed for an IPMN who do not have one. Further in any IPMN surveillance program, there is likely a fraction of patients that do not truly have an IPMN; thus, we think this is an important aspect of the study that could also help inform patient discussions as they enter into a pancreas cyst surveillance program. Given the retrospective nature of this study, we are unable to suggest changes to current surveillance guidelines. Although we believe this study clearly highlights that a single approach to cyst surveillance is likely inappropriate and life-long surveillance is unnecessary for the majority of our patients.

In conclusion, we found cyst progression is associated with current smoker status, worrisome features, and symptoms at presentation. Even with strict resection criteria, many patients with IPMN undergo surgical resection with unclear benefit; accurately identifying individuals who will benefit from surgical resection vs. surveillance can prevent unnecessary morbidity and anxiety. Our study demonstrates the challenge in diagnosing patients with an IPMN based on radiology alone and suggest that many patients are recommended for long term surveillance due to a lesion that may not require it. Better diagnostic techniques are warranted to appropriately diagnose and subsequently allocate to a surveillance program. Continued investigation is necessary to identify modifiable risk factors and patients at the highest risk of cystic malignant transformation to better inform surveillance strategies and management.