Dual-layer spectral detector computed tomography for improving the demonstration of the communication between intraductal papillary mucinous neoplasms and pancreatic duct

Introduction

With the increased use and improvement in multidetector computed tomography (MDCT) and magnetic resonance (MR) imaging technologies, the widespread adoption of advanced imaging modalities has led to more frequent identification of pancreatic cystic lesions during unrelated diagnostic workups, particularly serous cystadenomas (SCAs), intraductal papillary mucinous neoplasms (IPMNs), and mucinous cystic neoplasms (MCNs) (1,2). IPMNs refer to macroscopically identifiable neoplasms originating from the intraductal epithelium, characterized by cystic dilatation and papillary growth patterns (3,4). Histologically, IPMNs can be subcategorized into low- or moderate-grade dysplasia, considered benign, and high-grade dysplasia, associated with invasive cancers (5,6). Currently, most patients with cystic neoplasm of the pancreas undergoing resection will have the histologic diagnosis of SCA or IPMN (7). SCAs are considered benign and generally do not require resection. In contrast, IPMNs, as the most crucial pancreatic ductal adenocarcinoma precursor, are flagged for further investigation, surgical resection, and follow-up imaging (7,8). Therefore, a preoperative definite diagnosis is critical for patient management and treatment.

Modern pancreatic imaging predominantly utilizes MDCT and magnetic resonance imaging (MRI) for the comprehensive evaluation of pancreatic cystic lesions (9,10). The presence of demonstrable ductal continuity with a cystic lesion serves as a key diagnostic feature distinguishing IPMN from other pancreatic cystic neoplasms, particularly SCA and MCN (11,12). Multiplanar and 3-dimensional (3D) imaging of MDCT and MRI can aid in the assessment of ductal communication. MDCT with submillimeter slice thickness can improve image spatial resolution effectively. With faster MDCT scan acquisition speed, current generation scanners can complete comprehensive upper abdominal imaging less than 10 seconds, thus considerably reducing motion artifacts. Magnetic resonance cholangiopancreatography (MRCP) provides non-invasive visualization of both small pancreatic cysts and the main or branches of the pancreatic ductal system. Compared with MDCT, MRI is more sensitive overall to detect small pancreatic cysts <3 cm (13). In addition, MRI has the advantage of avoiding radiation exposure. Therefore, a consensus of radiologists suggested dedicated MRCP as the procedure of choice for evaluating a pancreatic lesion (14). However, there are some challenges for MRI of the upper abdomen. Irregular breathing during scanning can significantly impact image quality, potentially leading to challenges in interpretation. Further, MRI cannot be frequently performed in some regions because of its high cost. Besides, there are some contraindications for MRI that make it unsuitable for the entire population.

Dual-layer spectral detector computed tomography (DLCT) is a new computed tomography (CT) imaging technology that features a novel two-layer scintillator detector (as shown in Table S1). This advanced imaging technique generates both quantitative metrics and functional imaging data, with numerous published studies demonstrating its utility for pancreatic lesion characterization. DLCT-derived virtual monoenergetic images (VMIs) offer significant diagnostic advantages, including enhanced lesion conspicuity, artifacts reduction, and optimized radiation efficiency (15,16). Multiple clinical studies have validated DLCT’s capability to enhance pancreatic imaging through superior image quality, optimized tissue contrast, and precise tumor boundary definition (17,18). Some reports have investigated the usability of DLCT-derived VMIs at low keV for pancreatic ductal adenocarcinoma, pancreatic neuroendocrine neoplasms, and so on (19,20). However, to our knowledge, no study has specifically evaluated DLCT-derived VMIs for depicting ductal communication in IPMN, a critical discriminator for preoperative diagnosis (8,11). Furthermore, existing literature lacks direct comparison of multiple VMI energy levels (40–70 keV) against MRCP for this clinical task.

This study aimed to evaluate the image quality and the display of communication between the IPMNs and the pancreatic duct on VMIs at different keVs, compared with polyenergetic images (PEIs) and MRCP. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-2196/rc).

Methods

Patients

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Chongqing General Hospital (No. KY S2023-082-01), and the requirement for individual consent for this retrospective analysis was waived. Patients with pathologically confirmed or MRCP-diagnosed IPMN were retrospectively enrolled from Chongqing General Hospital between August 2021 and September 2023. The inclusion criteria were as follows: (I) patients confirmed by pathology or diagnosed by MRCP; (II) patients confirmed by pathology who received contrast-enhanced CT imaging within 2 weeks prior to surgical intervention; and (III) the interval between CT and MRCP ranged from 1 day to 2 weeks. The exclusion criteria were as follows: (I) those without DLCT data; (II) invasive cancer associated with IPMN; and (III) poor image quality with artifact interference: four cases were excluded from quantitative analysis due to severe partial-volume artifacts in non-dilated pancreatic ducts (<1 mm diameter).

Image acquisition and reconstruction

Unenhanced and contrast-enhanced abdominal images were acquired from all patients using the DLCT (IQon spectral CT, Philips Healthcare, Amsterdam, Netherlands) following a minimum 4-hour fasting period in this study. DLCT requires no specialized scanning protocols beyond conventional pancreatic CT. Scanning was performed ranging from the diaphragmatic apex to the inferior border of the liver with the following acquisition parameters: tube voltage, 120 kVp; tube current, automatic modulation; detector collimation, 64×0.625; helical pitch, 0.798; and rotation time, 0.5 seconds. Intravenous administration of non-ionic contrast media (Ultravist 370, Bayer, Leverkusen, Germany) was performed at a dosage of 1.5 mL/kg body weight, delivered via peripheral forearm or central venous access at an injection rate of 3–3.5 mL/s. The bolus tracking technique was applied to start arterial and portal venous phase (PVP) image acquisition with a delay of 12 seconds and 35 seconds after reaching a threshold value of 150 Hounsfield units (HU) in the descending aorta.

Following scan completion, standard 120 kVp polychromatic images were reconstructed utilizing iterative algorithm (iDose4, level 3, Philips Healthcare) for conventional image generation. Concurrently, spectral raw data were processed with proprietary spectral reconstruction algorithm (Spectral, level 3, Philips Healthcare), specifically optimized to mitigate noise amplification at low keV ranges (40 keV in particular), before transfer to the dedicated post-processing workstation (IntelliSpace Portal 10.0, Philips Healthcare) for advanced analysis. Using the spectral dataset, we generated VMIs at 40, 50, 60, and 70 keV energy levels for both arterial and PVP acquisitions. This energy range (40–70 keV) was selected based on the following principles: (I) attenuation of iodinated contrast increases exponentially as keV levels approach the k-edge of iodine (33.2 keV). Lower keV [40–50] maximizes contrast between enhanced pancreatic parenchyma and hypoattenuating cysts/ducts (19,21). (II) Although ≤40 keV maximizes contrast, image noise increases inversely with energy. The 60–70 keV range provides noise levels comparable to conventional 120 kVp PEIs, serving as a clinical reference (20,22). (III) This spectrum aligns with established dual-energy CT protocols for abdominal imaging, where 40–70 keV optimizes the contrast-to-noise ratio (CNR) for pancreatic evaluations (16,21,23). All reconstructed images maintained identical spatial resolution parameters (1 mm slice thickness with 1 mm interval spacing), matching the specifications of the standard 120 kVp PEIs.

MRI was acquired using a 3.0 T MR system (Skyra 3.0-T, Siemens Healthcare, Erlangen, Germany). MRCP acquisitions employed dual complementary approaches: breath-hold two-dimensional (2D) turbo spin-echo single-shot projections acquired at multiple angulations (−25°, 0°, 15°, 25° relative to coronal), and 3D turbo spin-echo navigator-triggered volumetric acquisition.

Quantitative analysis

Three regions of interest (ROIs) were placed by two abdominal radiologists, with 3 and 12 years of clinical experience, respectively, for quantitative analysis. The ROI of pancreatic parenchyma was at least 10 mm² and avoided the visible pancreatic duct, vessels, and artifacts. The ROI of the pancreatic lesion was at least 10 mm² and avoided the margin. The ROIs of the pancreatic duct and pancreatic lesion were as large as possible when those areas were lower than 10 mm². The post-processing workstation automatically propagated each ROI across all corresponding image series, maintaining perfect spatial registration and measurement consistency. All ROIs were determined through joint evaluation and agreement by both abdominal radiologists to ensure measurement consistency. A schematic diagram of the ROI placement and measurement is shown in Figure 1. Then, the CT attenuation and standard deviation (SD) were recorded. Quantitative image quality metrics for pancreatic structures were derived using the following equations (24).

ADparenchyma/lesion=HUparenchyma−HUlesion[1]

ADparenchyma/duct=HUparenchyma−HUduct[2]

SNRlesion=HUlesionSDlesion[3]

SNRparenchyma=HUparenchymaSDparenchyma[4]

SNRduct=HUductSDduct[5]

CNRparenchyma/lesion=HUparenchyma−HUlesion(SDparenchyma+SDlesion)÷2[6]

CNRparenchyma/duct=HUparenchyma−HUduct(SDparenchyma+SDduct)÷2[7]

Figure 1 Schematic diagram of delineating regions of interest (red arrow indicating the lesion, cyan arrow showing the duct, and yellow arrow pointing to the parenchyma). Ar, area; Av, average; HU, Hounsfield units; Perim, perimeter; SD, standard deviation.

Qualitative analysis

The same two radiologists performed blinded, randomized assessments of PEI and VMI datasets, with no knowledge of reconstruction parameters during image analysis. Both arterial phase (AP) and PVP acquisitions were evaluated independently, with PEI and VMI datasets presented in randomized sequences to prevent interpretation bias. Prior to independent evaluation, both radiologists participated in a consensus training session using 10 reference cases (excluded from the study) to standardize: interpretation of ductal communication versus partial-volume artifacts. All images were initially displayed using standard soft-tissue window settings, with subsequent reader-adjustable windowing permitted to optimize visualization of pancreatic structures (parenchyma, ducts, and lesions) (21). Five-point Likert scales were used to assess pancreatic lesion conspicuity (1, undiagnostic; 2, suboptimal; 3, average; 4, good; and 5, excellent), pancreatic lesion margin sharpness (1, unclear; 2, suboptimal; 3, average; 4, good; and 5, excellent), pancreatic duct sharpness (1, blurry; 2, suboptimal; 3, average; 4, good; and 5, sharpest), and the display of communication between pancreatic lesion and pancreatic duct (1, undiagnostic; 2, limited diagnostic; 3, average; 4, good; 5, excellent). The two radiologists’ mean scores were statistically analyzed.

MRCP image analysis

The same two radiologists independently reviewed MRCP images at a Picture Archiving and Communication System (PACS) workstation. Each reader recorded the communication between the lesions and the pancreatic duct. In cases of interobserver disagreement, final decisions were determined with consensus. The communication was defined as a channel enabling direct, continuous connection between the cystic lesion and the normal or dilated pancreatic duct, visualized as either: (I) high-signal-intensity material filling continuity (for MRCP): high-signal-intensity material displaying between the pancreatic duct and the cystic lesion on MRCP, confirmed on ≥2 consecutive oblique coronal slices or 3D maximum intensity projection (MIP) images. (II) Tract morphology (for DLCT): a visible non-enhancing tubular structure (mucin-filled channel) connecting the cyst to the duct, confirmed on ≥2 consecutive axial slices (1-mm thickness) or multiplanar reconstructions (coronal/sagittal) (as shown in Figure 2).

Figure 2 Parenchymal mass in a 50-year-old female with histologically confirmed IPMN in a surgical specimen. Both the DLCT and MRCP images clearly depict communication (indicated by the red arrows) between the lesion and the duct. (A) MIP and (B) three-dimensional turbo spin-echo MRCP. (C) Coronal and (D) axial VMI₄₀ in DLCT. DLCT, dual-layer spectral detector computed tomography; IPMN, intraductal papillary mucinous neoplasm; MIP, maximum intensity projection; MRCP, magnetic resonance cholangiopancreatography; VMI, virtual monoenergetic image.

Statistical analysis

All statistical analyses were conducted using the software SPSS 27.0 (IBM Corp., Armonk, NY, USA) following established analytical protocols. Data distribution normality was verified using the Shapiro-Wilk test. Continuous variables were expressed as the mean ± SD. The normally distributed quantitative parameters in attenuation differences (ADs), signal-to-noise ratios (SNRs), and CNRs were compared using the analysis of variance (ANOVA) test and the Bonferroni correction. Friedman test with Bonferroni correction was used to compare the quantitative parameters that did not fit the normal distribution and the qualitative parameters. Interobserver agreement for the qualitative parameters was assessed using Cohen’s kappa statistics, with established interpretation thresholds: poor (κ≤0.20), fair (κ=0.21–0.40), moderate (κ=0.41–0.60), good (κ=0.61–0.80), and excellent agreement (κ≥0.81). A probability threshold of P<0.05 defined statistical significance. The display of communication between pancreatic lesion and pancreatic duct was compared between DLCT and MRCP using McNemar’s test.

Results

Study population

A total of 74 patients were identified, and 39 patients were excluded. Finally, 35 patients were included in this cross-sectional study. Detailed inclusion and exclusion criteria are shown in Figure 3. All the IPMNs were confirmed by MRCP or histopathology. Detailed patient demographics are shown in Table 1.

Figure 3 Flow chart of the research. AD, attenuation difference; CNR, contrast-to-noise ratio; DLCT, dual-layer spectral detector computed tomography; IPMN, intraductal papillary mucinous neoplasm; MIP, maximum intensity projection; MRCP, magnetic resonance cholangiopancreatography; PEIs, polyenergetic images; SNR, signal-to-noise ratio; VMI, virtual monoenergetic image.

Quantitative image quality parameters

The results of all the quantitative analyses are shown in Table 2 and Figures S1-S3. The AD, SNR, and CNR of VMIs demonstrated progressive enhancement at lower energy levels, peaking at 40 keV. VMI_40–50 provided significantly higher AD of the pancreatic lesion, and VMI_40–60 provided significantly higher AD of the pancreatic duct for both phases compared with PEI (P<0.001). Compared with PEI, VMI_40–60 demonstrated significantly superior SNR of pancreas and CNR of pancreatic duct at both AP and PVP, and CNR of pancreatic lesion at PVP (P≤0.001); VMI_40–50 demonstrated significantly superior SNR of pancreatic lesion at both AP and PVP, CNR of pancreatic lesion at AP, and SNR of pancreatic duct at PVP (P<0.001).

Qualitative image quality parameters

Qualitative image quality was assessed using PEIs and VMI_40–70 for both AP and PVP. Table 3 summarizes the outcomes of the qualitative assessment. The scores for lesion conspicuity, lesion margin sharpness, pancreatic duct sharpness at AP, and definition of communication between the pancreatic lesion and pancreatic duct at AP and PVP in VMI_40–60, and lesion conspicuity, lesion margin sharpness, and pancreatic duct sharpness in VMI_40–70 demonstrated superior performance than those in PEIs (P<0.001). VMI₄₀ produced the highest scores. Figures 4-6 display representative cases. Good-to-excellent inter-observer agreement existed for all images in VMIs and PEIs (κ=0.743–0.813).

Figure 4 Axial DLCT images of a 59-year-old male with pancreatic IPMN (red arrow indicating the lesion, cyan arrow showing the communication tract, and yellow arrow pointing to the pancreatic duct) in the pancreatic body. (A-E) PEI and VMI at 40–70 keV during the arterial phase. (F-J) PEI and VMI at 40–70 keV during the portal venous phase. Insets in the lower-right corner display a magnified view of the communication tract. Note the improved tract conspicuity at VMI₄₀ (AD parenchyma/duct =224) compared to PEI (AD parenchyma/duct =94). AD, attenuation difference; DLCT, dual-layer spectral detector computed tomography; IPMN, intraductal papillary mucinous neoplasm; PEI, polyenergetic image; VMI, virtual monoenergetic image.

Figure 5 Coronal DLCT images of a 49-year-old female showing pancreatic IPMN (red arrows indicating the communication tract) at the pancreatic head. (A-E) PEI and VMI at 40–70 keV during the arterial phase. (F-J) PEI and VMI at 40–70 keV during the portal venous phase. Insets in the lower-right corner display a magnified view of the communication tract. Note the improved tract conspicuity at VMI₄₀ (AD parenchyma/duct =261) compared to PEI (AD parenchyma/duct =101). AD, attenuation difference; DLCT, dual-layer spectral detector computed tomography; IPMN, intraductal papillary mucinous neoplasm; PEI, polyenergetic image; VMI, virtual monoenergetic image.

Figure 6 Sagittal CT images of a 60-year-old male showing pancreatic IPMN (red arrow indicating the communication tract) in the pancreatic head. (A-E) PEI and VMI at 40–70 keV during the arterial phase. (F-J) PEI and VMI at 40–70 keV during the portal venous phase. Insets in the lower-right corner display a magnified view of the communication tract. Note the improved tract conspicuity at VMI₄₀ (AD parenchyma/duct =211) compared to PEI (AD parenchyma/duct =89). AD, attenuation difference; CT, computed tomography; IPMN, intraductal papillary mucinous neoplasm; PEI, polyenergetic image; VMI, virtual monoenergetic image.

Intermodality agreement between DLCT and MRCP

There were 30 patients who underwent both DLCT and MRCP examination. Readers identified lesion communication in 30 lesions at VMI₄₀ in DLCT. In comparison, the readers identified lesion communication in 28 lesions in MRCP (as shown in Figure 2). There was no significant difference in the display of communication of IPMNs with the pancreatic duct between DLCT and MRCP (P=0.500) (as shown in Table 4 and Table S2).

Discussion

In this study, we systematically assessed how low-energy VMIs influence both quantitative and qualitative imaging parameters of pancreatic lesion, pancreatic duct, and the communication between them in patients with IPMN. Our results at 40–70 keV reflect the fundamental trade-off in spectral CT: VMIs ≤50 keV exploit iodine’s photoelectric dominance for superior lesion-duct contrast, whereas 60–70 keV maintains diagnostic noise levels analogous to conventional CT (22). This energy spectrum is now standardized for pancreatic applications (20,23). Our study demonstrated that VMI₄₀ images provide significantly better objective and subjective image parameters compared to PEI and VMI images at higher keV [50–70]. Based on our findings, a 40 keV VMI protocol appears to be the preferable diagnostic technique for the diagnostic workup of IPMN, provided a DLCT scanner is available. VMI₄₀ provided similar communication display between the pancreatic lesion and pancreatic duct compared with MRCP.

Several studies in pancreatic disease imaging have demonstrated the diagnostic advantages of VMI over PEI, providing superior SNR and enhanced tissue contrast at comparable attenuation levels (22,23,25). However, no studies have investigated the image quality of displaying communication between the pancreatic lesion and duct in patients with IPMN. Our study discussed the quantitative and qualitative image parameters of communication between the pancreatic lesion and duct from both the AP and PVP. Our data demonstrated optimal CNR and SNR for pancreatic lesion-duct communication the assessment at 40 keV VMI from both AP and PVP. Notably, PVP VMI at low keV demonstrated a favorable diagnostic performance compared to conventional AP images in terms of lesion conspicuity, margin delineation, and visualization of communication. This may be due to the higher enhancement of pancreas in the PVP and higher contrast with non-enhanced lesions. These results suggest potential protocol simplification, as repeated multiphasic scanning may not be necessary in select patients, reducing cumulative radiation dose, contrast media requirements, and healthcare expenditures. Previous research has established that the photoelectric absorption of iodine exhibits an exponential increase near its k-edge (33.2 keV), while image noise also escalates at these lower energies. Previous investigations into pancreatic imaging have reported that the optimal subjective image parameters were VMI reconstructions at 50–70 keV (26). Although lower keV levels inherently increase image noise, DLCT technology mitigates this limitation through regional spatial frequency-based signal recombination and advanced iterative reconstruction algorithms, which preserve diagnostic clarity by reducing beam-hardening artifacts and suppressing noise across the full energy spectrum. In our investigation, both readers preferred the 40 and 50 keV. Lower keV reconstructions offer superior tissue contrast that offsets slightly elevated noise, with adaptive windowing further enhancing diagnostic clarity for net clinical benefit. Furthermore, DLCT’s unique capability to decompose low- and high-energy photons optimizes spectral separation while preserving noise consistency, avoiding the trade-offs typically seen in conventional polychromatic imaging (21). Thus, 40 keV VMIs integrate the benefits of heightened iodine contrast, artifact reduction, and clinically acceptable noise levels, explaining their superior quantitative and qualitative performance in both cystic lesion characterization and structural evaluation in this study. To facilitate the adoption of VMI in clinical workflows, an initial familiarization period should be considered to allow users to adjust to the altered SNR properties of the reconstructed images. Since CT provides superior spatial resolution and is more widely accessible than MRCP, and MRCP excels at assessing ductal connections, many institutions use both modalities equally for IPMN evaluation (26-28). In our study, the performance of MRCP was good, and the communication was depicted in 28 of 30 patients with IPMNs. The communication of 2 patients was not shown clearly on MRCP due to respiratory condition and image resolution. However, the communication was identified clearly in all patients on DLCT with VMI₄₀. In patients without a definitive diagnosis of IPMN, imaging strategies must balance diagnostic accuracy with clinical feasibility. Based on this study’s findings and existing consensus guidelines (8,27), MRCP is preferred as the first-line modality for evaluating pancreatic cystic lesions due to its lack of ionizing radiation, superior soft-tissue resolution, and multiparametric capability to characterize cyst morphology such as septations and mural nodules. However, MRCP may miss ductal communication due to limited spatial resolution, motion artifacts, bowel peristalsis, or internal metal. For patients with MRI contraindications, patients who cannot tolerate prolonged MRI protocols, or in resource-limited settings, DLCT with 40–50 keV VMI serves as a robust alternative to confirm ductal communication, as demonstrated in our study. DLCT achieves comparable performance to MRCP in visualizing pancreatic duct communication while offering rapid scan acquisition (<10 seconds) and high spatial resolution to assess lesion enhancement patterns. In cases of diagnostic uncertainty, a combined approach integrating MRCP and DLCT is recommended. For example, MRI’s T2-weighted sequences can delineate cyst fluid composition, whereas DLCT VMI at 40 keV enhances pancreatic parenchyma-to-lesion contrast, thereby improving diagnostic confidence (26,28).

Several limitations should be considered in this study. Firstly, this retrospective single-center study included a relatively small number of patients. Further large-scale multi-center prospective studies are necessary to validate our preliminary results. Secondly, the pancreatic duct of some patients is too small due to it not being dilated, resulting in a partial volume effect that affects the quantitative image parameters. However, it can also display communication clearly in the qualitative image parameters. Critically, our analysis of the excluded cases with poor image quality (n=4) revealed an important distinction (as shown in Figure S4): Quantitative analysis is not suitable for sub-millimeter ducts due to unavoidable partial-volume effects at a 1-mm slice thickness. However, qualitative diagnosis is successful, as it enhances the contrast between parenchyma and ducts sufficiently to overcome this limitation. This suggests that the clinical value of DLCT lies in its visualization capabilities rather than in absolute attenuation metrics for small structures. In addition, our study focused on distinguishing IPMNs from other cystic lesions primarily through ductal communication, without incorporating additional morphological or enhancement-based features. Although 40–50 keV VMI excels in IPMN evaluation and hypervascular lesion characterization, its utility in hypovascular or fibrotic pathologies may be limited. Future research should establish adaptive keV protocols based on lesion vascularity and anatomical context to optimize diagnostic workflows. Finally, this study included not only patients confirmed by pathology but also some patients diagnosed by MRCP. The inclusion of patients via MRCP means that partial data of this study lacks pathological confirmation. The lack of histopathology in all cases may have introduced selection bias. However, the present study was mainly concerned with exploring the display of communication. Our subgroup analysis of pathologically confirmed cases (n=12) supports the robustness of DLCT VMI₄₀, and future studies with universal pathological correlation are needed to confirm diagnostic accuracy.

Conclusions

This study demonstrates that low keV (40–50 keV) VMIs from DLCT provide improved quantitative and qualitative image quality in IPMN compared to PEI. VMI 40 keV achieves MRCP-comparable performance in assessing IPMN ductal communication. This represents an important clinical advance for spectral CT, offering a viable alternative when MRI cannot be utilized.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-2196/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-24-2196/dss

Funding: This work was supported by the Medical Research Program of the Combination of Chongqing National Health Commission and Chongqing Science and Technology Bureau (No. 2024MSXM094).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-2196/coif). X.D.Z. and H.F.C. are employees of Philips Healthcare. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Chongqing General Hospital (No. KY S2023-082-01), and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Kromrey ML, Bülow R, Hübner J, Paperlein C, Lerch MM, Ittermann T, Völzke H, Mayerle J, Kühn JP. Prospective study on the incidence, prevalence and 5-year pancreatic-related mortality of pancreatic cysts in a population-based study. Gut 2018;67:138-45. [Crossref] [PubMed]
2. Han YE, Park BJ, Sung DJ, Kim MJ, Han NY, Sim KC, Cho Y, Kim H. Dual-Layer Spectral CT of Pancreas Ductal Adenocarcinoma: Can Virtual Monoenergetic Images of the Portal Venous Phase Be an Alternative to the Pancreatic-Phase Scan? J Belg Soc Radiol 2022;106:83. [Crossref] [PubMed]
3. Allen PJ. Response to the Comment on "Intraductal Papillary Mucinous Neoplasms: Have IAP Consensus Guidelines Changed Our Approach?: Results From a Multi-institutional Study". Ann Surg 2021;274:e706-7. [Crossref] [PubMed]
4. Fong ZV, Fernández-Del Castillo C. Intraductal Papillary Mucinous Neoplasm of the Pancreas. Surg Clin North Am 2016;96:1431-45. [Crossref] [PubMed]
5. Nagtegaal ID, Odze RD, Klimstra D, Paradis V, Rugge M, Schirmacher P, Washington KM, Carneiro F, Cree IAWHO Classification of Tumours Editorial Board. The 2019 WHO classification of tumours of the digestive system. Histopathology 2020;76:182-8. [Crossref] [PubMed]
6. He X, Fan R, Sun J, Ren Y, Zhao X, Rui W, Yuan Y, Zou D. A model for predicting degree of malignancy in patients with intraductal papillary mucinous neoplasm. Front Oncol 2023;13:1087852. [Crossref] [PubMed]
7. van Huijgevoort NCM, Del Chiaro M, Wolfgang CL, van Hooft JE, Besselink MG. Diagnosis and management of pancreatic cystic neoplasms: current evidence and guidelines. Nat Rev Gastroenterol Hepatol 2019;16:676-89. [Crossref] [PubMed]
8. Ohtsuka T, Fernandez-Del Castillo C, Furukawa T, Hijioka S, Jang JY, Lennon AM, Miyasaka Y, Ohno E, Salvia R, Wolfgang CL, Wood LD. International evidence-based Kyoto guidelines for the management of intraductal papillary mucinous neoplasm of the pancreas. Pancreatology 2024;24:255-70. [Crossref] [PubMed]
9. Udare A, Agarwal M, Alabousi M, McInnes M, Rubino JG, Marcaccio M, van der Pol CB. Diagnostic Accuracy of MRI for Differentiation of Benign and Malignant Pancreatic Cystic Lesions Compared to CT and Endoscopic Ultrasound: Systematic Review and Meta-analysis. J Magn Reson Imaging 2021;54:1126-37. [Crossref] [PubMed]
10. Lee JE, Choi SY, Min JH, Yi BH, Lee MH, Kim SS, Hwang JA, Kim JH. Determining Malignant Potential of Intraductal Papillary Mucinous Neoplasm of the Pancreas: CT versus MRI by Using Revised 2017 International Consensus Guidelines. Radiology 2019;293:134-43. [Crossref] [PubMed]
11. Ippolito D, Maino C, Pecorelli A, De Vito A, Riva L, Talei Franzesi C, Sironi S. Incidental pancreatic cystic lesions: comparison between CT with model-based algorithm and MRI. Radiography (Lond) 2021;27:554-60. [Crossref] [PubMed]
12. Henninger B, Steurer M, Plaikner M, Weiland E, Jaschke W, Kremser C. Magnetic resonance cholangiopancreatography with compressed sensing at 1.5 T: clinical application for the evaluation of branch duct IPMN of the pancreas. Eur Radiol 2020;30:6014-21. [Crossref] [PubMed]
13. Lee HJ, Kim MJ, Choi JY, Hong HS, Kim KA. Relative accuracy of CT and MRI in the differentiation of benign from malignant pancreatic cystic lesions. Clin Radiol 2011;66:315-21. [Crossref] [PubMed]
14. Berland LL, Silverman SG, Gore RM, Mayo-Smith WW, Megibow AJ, Yee J, Brink JA, Baker ME, Federle MP, Foley WD, Francis IR, Herts BR, Israel GM, Krinsky G, Platt JF, Shuman WP, Taylor AJ. Managing incidental findings on abdominal CT: white paper of the ACR incidental findings committee. J Am Coll Radiol 2010;7:754-73. [Crossref] [PubMed]
15. Kolossváry M, Kellermayer M, Merkely B, Maurovich-Horvat P. Cardiac Computed Tomography Radiomics: A Comprehensive Review on Radiomic Techniques. J Thorac Imaging 2018;33:26-34. [Crossref] [PubMed]
16. Hickethier T, Byrtus J, Hauger M, Iuga AI, Pahn G, Maintz D, Haneder S, Doerner J. Utilization of virtual mono-energetic images (MonoE) derived from a dual-layer spectral detector CT (SDCT) for the assessment of abdominal arteries in venous contrast phase scans. Eur J Radiol 2018;99:28-33. [Crossref] [PubMed]
17. Lee SM, Kim SH, Ahn SJ, Kang HJ, Kang JH, Han JK. Virtual monoenergetic dual-layer, dual-energy CT enterography: optimization of keV settings and its added value for Crohn's disease. Eur Radiol 2018;28:2525-34. [Crossref] [PubMed]
18. Doerner J, Hauger M, Hickethier T, Byrtus J, Wybranski C, Große Hokamp N, Maintz D, Haneder S. Image quality evaluation of dual-layer spectral detector CT of the chest and comparison with conventional CT imaging. Eur J Radiol 2017;93:52-8. [Crossref] [PubMed]
19. Yuan J, Wang Y, Hu X, Shi S, Zhang N, Wang L, Deng W, Feng ST, Peng Z, Luo Y. Use of dual-layer spectral detector computed tomography in the diagnosis of pancreatic neuroendocrine neoplasms. Eur J Radiol 2023;159:110660. [Crossref] [PubMed]
20. Nagayama Y, Tanoue S, Inoue T, Oda S, Nakaura T, Utsunomiya D, Yamashita Y. Dual-layer spectral CT improves image quality of multiphasic pancreas CT in patients with pancreatic ductal adenocarcinoma. Eur Radiol 2020;30:394-403. [Crossref] [PubMed]
21. Hickethier T, Iuga AI, Lennartz S, Hauger M, Byrtus J, Luetkens JA, Haneder S, Maintz D, Doerner J. Virtual Monoenergetic Images From a Novel Dual-Layer Spectral Detector Computed Tomography Scanner in Portal Venous Phase: Adjusted Window Settings Depending on Assessment Focus Are Essential for Image Interpretation. J Comput Assist Tomogr 2018;42:350-6. [Crossref] [PubMed]
22. Noda Y, Tochigi T, Parakh A, Joseph E, Hahn PF, Kambadakone A. Low keV portal venous phase as a surrogate for pancreatic phase in a pancreatic protocol dual-energy CT: feasibility, image quality, and lesion conspicuity. Eur Radiol 2021;31:6898-908. [Crossref] [PubMed]
23. Laukamp KR, Tirumani SH, Lennartz S, Hokamp NG, Gupta A, Pennig L, Persigehl T, Gilkeson R, Ramaiya N. Evaluation of equivocal small cystic pancreatic lesions with spectral-detector computed tomography. Acta Radiol 2021;62:172-81. [Crossref] [PubMed]
24. Han YE, Park BJ, Sung DJ, Kim MJ, Han NY, Sim KC, Cho Y, Kim H. Dual-Layer Spectral CT of Pancreas Ductal Adenocarcinoma: Can Virtual Monoenergetic Images of the Portal Venous Phase Be an Alternative to the Pancreatic-Phase Scan? J Belg Soc Radiol 2022;106:83. [Crossref] [PubMed]
25. Beer L, Toepker M, Ba-Ssalamah A, Schestak C, Dutschke A, Schindl M, Wressnegger A, Ringl H, Apfaltrer P. Objective and subjective comparison of virtual monoenergetic vs. polychromatic images in patients with pancreatic ductal adenocarcinoma. Eur Radiol 2019;29:3617-25. [Crossref] [PubMed]
26. Choi MH, Yoon SB, Song M, Lee IS, Hong TH, Lee MA, Jung ES. Benefits of the multiplanar and volumetric analyses of pancreatic cancer using computed tomography. PLoS One 2020;15:e0240318. [Crossref] [PubMed]
27. Scherer JA, Gebhard R, Firkins SA, Shah ZK, Urbina Andersson IK, Barker SJ, Fiorillo LE, Hollander E, Shaheen N, Koay EJ, Conwell DL, Krishna SG. Lower Interobserver Reliability for Nondimensional Intracystic Features Among Abdominal Radiologists for Characterizing Intraductal Papillary Mucinous Neoplasms Using Magnetic Resonance Imaging. Pancreas 2022;51:1225-30. [Crossref] [PubMed]
28. Manfredi R, Pozzi Mucelli R. Secretin-enhanced MR Imaging of the Pancreas. Radiology 2016;279:29-43. [Crossref] [PubMed]