Pancreatic cystic lesions are being increasingly detected due to advancements in and the widespread use of cross-sectional imaging techniques such as computed tomography (CT) and magnetic resonance imaging (MRI) [1,2]. Most pancreatic cystic lesions are benign, exhibiting indolent behavior and a favorable prognosis, particularly when they are small and lack worrisome features. However, certain types of pancreatic cysts, including intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms, carry a risk of malignant transformation [3,4]. Therefore, accurate diagnosis and long-term surveillance are essential for incidentally detected pancreatic cysts.

Current guidelines, including those from the American Gastroenterological Association (AGA) and the International Association of Pancreatology, recommend surveillance using CT, MRI, and endoscopic ultrasound (EUS) for asymptomatic pancreatic cysts [5,6]. However, each modality has inherent limitations: CT involves radiation exposure, MRI may be costly and less accessible, and EUS is invasive. Given that surveillance may be required for several years [7], there is a growing need for imaging alternatives that are more accessible, repeatable, and cost-effective, particularly for cysts at low to moderate risk of malignancy.

Transabdominal ultrasound (TAUS) offers a potential supplementary option for pancreatic cyst surveillance, as it is noninvasive, widely available, relatively inexpensive, and easy to perform. However, its role in pancreatic cyst surveillance remains limited because of poor acoustic penetration related to the deep location of the pancreas and interference from bowel gas, both of which compromise the acoustic window [8,9]. Consequently, TAUS is not currently recommended as an imaging modality for pancreatic cyst surveillance in international guidelines [5,6]. Nonetheless, previous studies have reported that TAUS can be useful for monitoring pancreatic cysts, particularly for detecting larger cysts (≥1 cm) and when correlated with prior cross-sectional imaging [10,11]. However, the ability of TAUS to identify high-risk features, such as high-risk stigmata or worrisome features, in comparison with cross-sectional imaging modalities, including CT and MRI, remains under investigation.

Therefore, the purpose of this study was to evaluate the clinical utility of TAUS in pancreatic cyst surveillance by assessing its concordance with follow-up contrast-enhanced CT and MRI findings and by determining the malignancy risk associated with TAUS-detected features.

This study demonstrated the potential role of TAUS in pancreatic cyst surveillance by showing a high identification rate (88.7%) and strong concordance with CT/MRI, particularly for the identification of high-risk cysts. Identification rates were significantly influenced by sex, BMI, and cyst location. Notably, 74.4% of cysts classified as high-risk on TAUS were similarly classified on CT/MRI, and the presence of multiple worrisome features on TAUS further increased this likelihood. In addition, malignancy was more frequently observed in high-risk cysts identified on TAUS (12.5% vs. 1.5%), supporting the potential value of TAUS for risk stratification.

Current guidelines recommend long-term surveillance for unresected pancreatic cysts because of their malignant potential [4,6,13]. MRI is generally preferred to minimize cumulative radiation exposure [6], but it is costly and less accessible for routine long-term monitoring. Although TAUS is not currently recommended in surveillance guidelines, its noninvasiveness, cost-effectiveness, and wide availability make it an attractive supplementary option, particularly for patients requiring repeated follow-up when image quality is expected to be adequate. Previous studies have reported TAUS detection rates of 81.8%–86.7% for pancreatic cysts ≥1 cm, especially when prior correlative imaging was available [10,11,14]. Jeon et al. [10] found that the detection rate improved from 49.2% to 86.7% when TAUS was performed after correlative imaging. Similarly, Sun et al. [14] reported an 82.2% detection rate for cysts ≥1 cm, compared with 40.9% for cysts of any size. In the present study, TAUS was performed not for initial detection but for re-identification or focused assessment in patients with known pancreatic cystic lesions, and this study demonstrated a comparable or higher identification rate of 88.7% (464/523), likely attributable to the inclusion of only cysts ≥1 cm and the availability of prior imaging in most patients (85%, 445/523). Consistent with previous studies [10,11], the identification rate was lower for tail-located cysts than for non-tail locations (68.2% vs. 92.7%), likely reflecting anatomical challenges related to overlying bowel gas and adipose tissue. Notably, Jeon et al. [10] reported that detection of tail-located cysts improved from 37.9% to 67.8% when TAUS was performed with prior correlative imaging. Therefore, comprehensive scanning techniques and careful review of prior imaging are essential when using TAUS. When adequate visualization remains challenging despite these efforts, alternative surveillance strategies should be considered in consultation with the referring clinician.

Although previous studies have established the feasibility of TAUS for detecting pancreatic cysts [10,11], effective imaging surveillance also requires accurate assessment of imaging features, particularly high-risk stigmata and worrisome features. Few studies have specifically evaluated the ability of TAUS to identify these risk features. In this study, 18.5% (86/464) of cysts detected on TAUS were categorized as high-risk, and 74.4% (64/86) of these were consistently classified as high-risk on follow-up CT/MRI. These findings suggest that TAUS may contribute meaningfully to risk stratification when high-risk features are clearly visualized. Notably, the presence of multiple worrisome features on TAUS was strongly associated with high-risk classification on CT/MRI, consistent with prior reports highlighting the predictive value of feature multiplicity [15–17]. Collectively, these findings suggest that pancreatic cysts classified as high-risk on TAUS, particularly those demonstrating multiple risk features, warrant closer evaluation and more intensive follow-up.

Discrepancies between TAUS and CT/MRI classification were observed in 14.2% (66/464) and 17.0% (79/464) of cases for reviewers 1 and 2, respectively. Most discordant cysts exhibited only a single high-risk stigmata or worrisome feature (84.1%–90.9%), suggesting that the presence of multiple features improves classification consistency. The most common TAUS finding among cysts classified as high-risk on TAUS but low-risk on CT/MRI was thickened cyst walls or septa. Thickened cyst walls or septa are considered ambiguous features, as measurement methods and cutoff values predictive of malignancy have not yet been established [6], and reported interobserver agreement is low (κ=0.17–0.29) [18–20]. Moreover, TAUS often has difficulty distinguishing complex internal cyst structures caused by debris or reverberation artifacts, even with the use of harmonic imaging [21]. In such cases, assessment of indirect signs of malignancy, such as MPD dilatation or abrupt caliber change, is also important and may support more accurate risk classification. In this study, nonenhancing lesions, such as mucin plugs or debris, were misinterpreted as mural nodules ≥5 mm on TAUS in four cases, leading to false-positive classifications. Because TAUS cannot evaluate lesion enhancement, any solid-appearing component identified on TAUS should prompt confirmatory imaging with contrast-enhanced CT or MRI to ensure accurate diagnosis. This approach may reduce misclassification and support appropriate clinical decision-making. Conversely, among cysts classified as low-risk on TAUS, 11.6% (44/378) and 15.1% (57/378) were reclassified as high-risk on CT/MRI, likely because of underestimation of cyst size or MPD caliber on TAUS. This pattern suggests that the relatively lower spatial resolution of TAUS, combined with susceptibility to bowel gas interference, can result in inaccurate morphologic assessment and false-negative classifications. These limitations underscore that the clinical utility of TAUS is substantially higher when clear visualization of the pancreas and cystic lesions can be achieved.

This study had several limitations. First, its retrospective design inherently introduces a risk of selection bias, as patient inclusion was based on existing records rather than a predefined protocol. Second, the study population included patients with an interval of less than 1 year between TAUS and contrast-enhanced CT or MRI, with a median interval of 349 days (range, 0 to 365 days). Natural interval-related changes in cyst morphology during this period may have contributed to some of the observed classification discrepancies, representing a potential source of interval bias. This limitation is inherent to retrospective study designs and underscores the need for future prospective validation with shorter imaging intervals. Third, interobserver agreement for TAUS findings was not assessed. Given the operator-dependent nature of TAUS, this represents an important limitation of the present study, and future investigations should address interobserver variability to improve consistency and generalizability of TAUS-based assessments. Fourth, the malignancy risk analysis was limited by the small number of pathologic confirmations, as most cases relied on imaging follow-up. Although this reflects real-world clinical practice, in which many pancreatic cystic lesions are benign and managed noninvasively, larger cohorts with histopathologic confirmation are needed to more robustly validate the association between TAUS features and malignancy risk. Fifth, some high-risk features were identified through retrospective review of stored TAUS images rather than from the original reports, introducing potential retrospective bias despite blinding to CT/MRI findings. Because TAUS reports are not fully structured, cyst size, mural nodules, and MPD diameter were typically documented by the operator, whereas certain findings, such as thickened septa, were often omitted and required retrospective assessment in 34.2% of cases. Future studies incorporating prospective and standardized documentation of TAUS features would help minimize this limitation. Sixth, this study did not compare high-risk features between prior CT or MRI and TAUS findings, which limits the ability to directly evaluate TAUS performance relative to previously identified imaging features.

TAUS may serve as a useful noninvasive imaging tool for pancreatic cyst surveillance, particularly in patients with favorable acoustic conditions, such as lower BMI and non-tail cyst location. TAUS features correlated well with CT/MRI in identifying high-risk pancreatic cysts, and the presence of multiple worrisome features on TAUS was associated with an increased likelihood of high-risk classification on subsequent imaging.