Letter to the Editor: Beyond cyst size - a time-risk-state framework for dynamic surveillance de-escalation of branch duct intraductal papillary mucinous neoplasms

Abstract

For incidentally detected branch duct intraductal papillary mucinous neoplasms, the long-term surveillance strategy faces a difficult trade-off among clinical benefit, healthcare burden, and patient quality of life. Using a 17-year real-world cohort, the article by Hopley et al published in the World Journal of Gastroenterology showed that most clinically relevant lesion transformations and surgical interventions occur within the first 2 years; thereafter, the incremental benefit declines sharply. In this letter, we propose a framework for dynamic surveillance de-escalation. It has three dimensions: Time, risk, and patient status. The key innovation is the explicit incorporation of the patient’s overall health trajectory together with cyst biological behavior as decision drivers. At the 2-year landmark, patients with a cyst size < 30 mm and a serum carbohydrate antigen 19-9 level < 43 kU/L should transition from intensive tumor screening to chronic disease-style follow-up aimed at preventing overtreatment.

Key Words: Branch duct intraductal papillary mucinous neoplasm; Surveillance de-escalation; Risk stratification; Time-risk-state framework; Patient-centeredness; Overtreatment

Core Tip: This letter proposes a three-dimensional dynamic surveillance de-escalation framework for branch duct intraductal papillary mucinous neoplasms. After 2 years of stability, patients with a cyst size < 30 mm and a serum carbohydrate antigen 19-9 level < 43 kU/L can safely transition from intensive tumor screening to chronic disease-style follow-up aimed at preventing overtreatment.

TO THE EDITOR

With the widespread use of imaging, incidental pancreatic cysts are increasingly being detected[1]. Branch duct intraductal papillary mucinous neoplasms, because of their malignant potential, have been regarded as an ideal window for the early detection of pancreatic ductal adenocarcinoma[2]. However, the long-term surveillance recommended by current guidelines leads to many low-risk individuals undergoing repeated examinations for years, with high cumulative costs and psychological distress[3,4].

A large observational study by Hopley et al[5] published in the World Journal of Gastroenterology followed 1191 patients for up to 17 years. That study challenged the traditional paradigm. Two key findings stood out. First, surgical interventions cluster heavily in the first 2 years. Second, later findings are very rare. These findings became the foundation for our proposed “time-risk-state” dynamic de-escalation framework.

TWO KEY FINDINGS

First, the “window of opportunity”: Why a benign finding at 2 years seems to predict long-term low risk. In fact, over 88% of surgical resections (including resectable pancreatic ductal adenocarcinoma) occur in the first 2 years, after which resectability for malignant lesions drops from 82% to just 9%[5]. Molecular studies suggest that an aggressive state might be “preset” early and that pancreatic ductal adenocarcinoma can develop independently of the original cyst[6,7]. Thus, 2-year stability likely identifies a naturally indolent subgroup of patients.

The second key finding is the “patient decline” paradox. The median age at the start of surveillance was 71 years. Five years later, many patients are approaching 80 years of age with increasing comorbidities and declining performance status. Even if an early malignant change is detected, many patients no longer have the physiological reserve for curative surgery. Competing risk analysis provides hard data: In patients aged ≥ 75 years with low-risk branch duct intraductal papillary mucinous neoplasms and a Charlson comorbidity index > 3, the 5-year cumulative incidence of death from non-pancreatic cancer is 15.93%, whereas that of pancreatic malignancy is only 1.6%[8]. In this population, death from noncancer causes is nearly 10 times more likely than death from intraductal papillary mucinous neoplasm-associated malignant transformation. Age ≥ 75 years is an independent predictor (hazard ratio: 4.15) and is even stronger than a Charlson comorbidity index > 3 (hazard ratio: 3.61)[8]. Consequently, continued surveillance provides little benefit for such patients.

NATURE OF THE PROBLEM

Current risk stratification systems (Fukuoka[9] and Kyoto[10] guidelines) rely on static cyst morphology and a few biochemical markers, ignoring both time and patient status. Two-year stability provides strong negative predictive value, and aging is an independent outcome modifier. Decision-making should shift from “What more can we detect?” to “What more can we do for this patient?”—a move from disease-centered to patient-centered medicine.

PROPOSED FRAMEWORK

We propose a “time-risk-state” three-dimensional dynamic surveillance de-escalation framework with three decision nodes (Figure 1).

Figure 1 Proposed dynamic surveillance de-escalation framework for branch duct intraductal papillary mucinous neoplasms. The framework is organized around three decision nodes based on time, risk, and patient status: (1) Years 0-2: Intensive monitoring to detect high-risk lesions; (2) At 2 years: De-escalation for stable cysts (< 30 mm) with a carbohydrate antigen 19-9 (CA19-9) level < 43 kU/L; and (3) Years 2-5 and beyond: Gradual transition to chronic disease-style follow-up, with discontinuation of routine imaging considered for patients aged ≥ 75 years or those with a Charlson comorbidity index (CCI) ≥ 3. BD-IPMN: Branch duct intraductal papillary mucinous neoplasms; MRI/EUS: Magnetic resonance imaging/endoscopic ultrasound.

Node 1 (0-2 years): Standard risk confirmation

The first decision node consists of guideline-recommended intensified monitoring [magnetic resonance imaging/endoscopic ultrasound plus carbohydrate antigen 19-9 (CA19-9) measurement every 6 months to 12 months]. The goal is to capture rapidly progressing “high-risk” subtypes. The endpoint is “stability” or an “event.”

Node 2 (at 2 years)

For patients without high-risk or worrisome features, formal surveillance de-escalation is recommended when the cyst size remains stable at < 30 mm and the serum CA19-9 level remains persistently < 43 kU/L (area under the curve 0.78 for high-grade dysplasia/malignancy[5]). The absolute risk of subsequent malignant transformation requiring intervention is < 1%[5].

Node 3 (2-5 years and beyond)

The surveillance intensity should transition from intensive cancer surveillance to chronic disease follow-up. From 2 to 5 years, low-frequency, non-invasive monitoring is performed (e.g., magnetic resonance imaging every 2 years, annual CA19-9 measurement). Beyond 5 years, for patients who remain stable and are aged ≥ 75 years or have a Charlson comorbidity index of ≥ 3, formal discontinuation of routine imaging may be considered, switching to symptom-triggered follow-up. This aligns with the 2024 Kyoto guideline suggestion[10] and is supported by Levink et al[11] and Marchegiani et al[12]. We summarize the main differences between the traditional static model and our proposed dynamic framework in Table 1.

Table 1 Core differences between the traditional static model and the proposed time-risk-state dynamic framework.

Dimension	Traditional static model	Proposed three-dimensional dynamic framework
Risk rating	One-time, fixed; based on baseline cyst morphology	Dynamic, adjustable; redefined after 2 years of stable follow-up
Surveillance intensity	Uniform for all patients (magnetic resonance imaging/endoscopic ultrasound every 6-12 months)	Differentiated by risk: Intensified (0-2 years) → low-frequency (2-5 years) → possible discontinuation (≥ 5 years)
Core decision variables	Cyst size, mural nodules, MPD diameter, CA19-9 level	Three-dimensional: Cyst biology + temporal stability (2-year NPV) + patient status (age and CCI)
Endpoint for discontinuation	“Surveillance until patient unfit for surgery” (passive, vague)	Clear, active thresholds: Stable for ≥ 5 years, age ≥ 75 years or CCI ≥ 3, SIR not significantly different from that of the general population

CA19-9: Carbohydrate antigen 19-9; CCI: Charlson comorbidity index; MPD: Main pancreatic duct diameter; NPV: Negative predictive value; SIR: Standardized incidence ratio.

PRACTICAL CONSIDERATIONS

First, the so-called “Sword of Damocles” anxiety has been well documented in patients under long-term surveillance[3], highlighting the urgent need for interventions such as shared decision-making and patient decision aids. Second, a more data-rich approach using registries such as PACYFIC may confirm the CA19-9 cutoff. Third, new biomarkers such as KRAS/GNAS mutations in cyst fluid[13] might eventually provide stronger evidence of indolence, although they remain investigational.

FUTURE PERSPECTIVE

The fundamental rationale underlying our proposed framework would remain unchanged even if emerging technologies such as artificial intelligence radiomics[14,15] or liquid biopsy[16,17] were incorporated because these technologies could further refine risk stratification, particularly for borderline cysts.

CONTROVERSIES AND CONCLUSION

In resectable cases, however, the miss rate in the study by Hopley et al[5] was only approximately 1%, a finding with which most clinicians would agree and that suggests that the majority of patients will benefit. One limitation of this evidence is the study’s single-center retrospective design. Therefore, a future multicenter validation study is vital.

Hopley et al[5] should be credited with questioning the need for lifelong surveillance. Not all pancreatic cysts requiring years of follow-up necessitate such monitoring. We suggest a time-risk-state model that replaces the existing population-based management approach with a new, patient-centered one. This shift is a necessary step toward healthcare that is accurate, compassionate, and effective, moving beyond the inflexibility of outdated paradigms.