pancreatic cancer detection and characterization: state of the art … · 2021. 5. 28. ·...

3684

Abstract. – Pancreatic ductal adenocarcino-ma (PDAC) represents a challenge for a multi-disciplinary oncology team. Diagnosis of PDAC remains challenging due to overlapping imag-ing features with benign lesions, notwithstand-ing great advances with computed tomography (CT) and magnetic resonance imaging (MRI). The term “Radiomics” has recently been intro-duced to define a mathematical process to ex-tract countless quantitative features from med-ical images (including each diagnostic tech-nique) with high throughput computing for di-agnosis and prediction. This article is an up-dated overview of the imaging techniques to be employed during detection and characteri-zation of pancreatic cancer diagnostic workup. Particularly, the limitations and advantages of the different imaging techniques are discussed, with a particular focus on functional imaging. This overview is the result of a self-study with-out protocol and registration number. Articles published in the English language from Janu-ary 2000 to January 2021 were included. We an-alyzed 15 papers on radiomics. The possibility of functional imaging, such as CT, MRI, and ra-diomics has revolutionized pancreatic imaging, improving the detection and characterization of the lesions and allowing a prognosis related to radiological features, favoring the process of personalized medicine.

Key Words:Pancreatic cancer, Computed tomography, Mag-

netic resonance imaging, Radiomics, Functional im-aging.

Abbreviations

ADC = apparent diffusion coefficient; AIP = autoim-mune pancreatitis; BF = blood flow; BOLD = Blood oxygenation-level dependent; BV = blood volume; CE = Contrast-enhanced; CT = computed tomography; DECT = dynamic enhanced CT; DCE = dynamic contrast-en-hanced; DECT = dual-energy CT; Dp = pseudo-diffusion coefficient; Dt = pure tissue coefficient; DKI= Diffusion Kurtosis imaging; DWI = diffusion weighted imaging; EES = extravascular extracellular space; HASTE = half-Fourier single-shot turbo spin-echo; iAUC = initial area under curve; IVIM = Intravoxel Incoherent Motion method; HIF = hypoxia-inducible factors; MD = mean value of the diffusion coefficient; MRI = magnetic res-onance imaging; MSD = maximum signal difference; MVD= microvascular density; MK = kurtosis median coefficient; PC = pancreatic cancer; PDAC = Pancreatic ductal adenocarcinoma; Fp = perfusion fraction; PPCT = pancreatic phase CT; pCT = Perfusion CT; PS = perme-ability surface; SMA = superior mesenteric artery; SMV = superior mesenteric vein; TA = Texture analysis; TTP = time to peak; VIBE = Volumetric Interpolated Breath-hold Examination; VMI = virtual monochromatic im-ages; WII = wash-in intercept; WIS = wash-in slope; WOI = wash-out intercept; WOS = wash-out slope; W = weighted.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is a challenge for a multidisciplinary oncology team. Incidence constantly increases, with an expected death rate of 81.7% among new cases in 2020

European Review for Medical and Pharmacological Sciences 2021; 25: 3684-3699

V. GRANATA1, R. GRASSI2, R. FUSCO1, R. GALDIERO1, S.V. SETOLA1, R. PALAIA3, A. BELLI3, L. SILVESTRO4, D. COZZI5, L. BRUNESE6, A. PETRILLO1, F. IZZO3

1Division of Radiology, “Istituto Nazionale Tumori IRCCS Fondazione Pascale – IRCCS di Napoli”, Naples, Italy 2Division of Radiology, “Università degli Studi della Campania Luigi Vanvitelli”, Naples, Italy 3Division of Hepatobiliary Surgical Oncology, “Istituto Nazionale Tumori IRCCS Fondazione Pascale – IRCCS di Napoli”, Naples, Italy4Abdominal Oncology Division, “Istituto Nazionale Tumori IRCCS Fondazione Pascale – IRCCS di Napoli”, Naples, Italy5Division of Radiology, “Azienda Ospedaliera Universitaria Careggi”, Florence, Italy 6Radiology Division, Università degli Studi del Molise, Via Francesco De Sanctis, Campobasso, Italy

Corresponding Author: Roberta Fusco, MD; e-mail: [email protected]

Pancreatic cancer detection and characterization: state of the art and radiomics


3685

and a 5-year survival rate of 9%1. Although many patients have locally advanced disease at diag-nosis, the only curative treatment is surgery, and systemic chemotherapy is usually the key thera-py2-7. The multidisciplinary team should make the choice concerning the resectability of pancreatic cancer following the acquisition of a complete staging7.

Diagnosis of pancreatic cancer is difficult due to overlapping imaging findings with benign le-sions8-11. However, a proper detection and char-acterization of lesions is indispensable since the prognosis is associated with tumor aggressiveness as well as a correct staging. Thus, an imaging modality that provides higher tumor detectability would be desirable8. Computed tomography (CT) has grown to be the tool of choice in the preop-erative diagnosis guiding treatment planning, as well as during follow-up12-16. Several researchers considered magnetic Resonance Imaging (MRI) to be equivalent to CT in detecting and stag-ing16. However, recent evidence recommends the addition of MRI as a diagnostic integration to identify lesions undetected by CT as well as the presence of liver metastases16. MRI is a useful diagnostic tool in oncologic patients since this of-fers morphological data by T2 weighted (W) and T1-W sequences, and functional data by diffu-sion-weighted imaging (DWI) and dynamic con-trast-enhanced (DCE)-MRI17-23, so as new tools as Blood oxygenation-level dependent (BOLD) sequences. DWI is an excellent tool for detec-tion and characterization of pancreatic lesions, rising clinical confidence, and declining false positives8. The assessment of DWI can be qual-itatively and quantitatively through the apparent diffusion coefficient (ADC) map. The ADC map is the graphical exemplification of the ratio of DW signal intensities, and their values may dif-ferentiate between benign and malignant lesions. The ADC values are correlated to the acquisition protocol and suffer from a lack of reproducibility, especially in respiratory triggering techniques, smaller size, and lesion heterogeneity8. DCE-MRI could be used to evaluate microvascular structures. DCE-MRI is acquired during venous injection of contrast agent. Within the tissue, the contrast drug passes from the intravascular to the extravascular extracellular space (EES), which causes an increase of signal on T1-W sequences. The rate of drug extravasation to EES in the neo-plastic tissue is due to vessel leakiness and blood flow. Therefore, the signal measured in DCE-MRI is linked to tissue permeability and perfusion21,23.

The signal enhancement of pancreatic perfusion can be assessed by qualitative, semi-quantita-tive or quantitative evaluation. The qualitative is based on the evaluation of the time-signal intensity curve shape; the semi-quantitative is based on the assessment of parameters that can be extracted directly from time-signal intensity curves, such as the area under the curve. The quantitative analysis needs computational-based curve algorithms using a bicompartmental model with arterial input function definition21,22.

Hypoxia is emerging as a key factor of aggres-sive tumor biology and the relative resistance to conventional as well as targeted therapies24. Compared to other solid tumors, oxygen levels in pancreatic cancer are the lowest25. Increasing evidence suggests that MRI techniques may provide a practical and more readily available alternative for hypoxia imaging25. BOLD-MRI employs paramagnetic properties of deoxyhe-moglobin and is mainly used for regional quanti-fication of oxygenation of the lesions26. Hypoxia not only induces a malignant and invasive phe-notype of pancreatic cancer (PC) cells, but also makes the cells more resistant to radiation and chemotherapy25.

The term “Radiomics” has recently been in-troduced to describe a mathematical method to extract multiple quantitative data from medical images oriented to lesion diagnosis and progno-sis. Compared to visual analysis of medical im-ages, the deep meaning of medical images from radiomics makes features uptake more efficient and objective. Radiomics is promising for tumor screening, early diagnosis, accurate grading and staging, treatment, and prognosis20. Recent stud-ies27-29 show that differentiation between PDAC and its differential diagnosis based on radiom-ic features is possible. Furthermore, radiomic features would be successfully used to predict therapeutically relevant subtypes of PDAC, so as for overall and progression-free survival of PDAC patients30,31.

The assessment of these functional parameters not only provides better detection, characteriza-tion, and staging of the PDAC, but also allows an evaluation of some prognostic features that can affect the therapy; therefore, the most appro-priate therapeutic choice can be guaranteed, in the perspective of an increasingly personalized medicine.

This article is an updated overview of the im-aging techniques to be employed during detection and characterization of pancreatic cancer diag-

V. Granata, R. Grassi, R. Fusco, R. Galdiero, S.V. Setola, et al

3686

nostic workup. Particularly, the limitations and advantages of the different imaging techniques are discussed, with a particular focus on func-tional imaging.

Materials and Methods

This overview is the result of a self-study with-out protocol and registration number.

Search Criterion We assessed several electronic databases:

PubMed (http://www.ncbi.nlm.nih.gov/pubmed), Scopus (Elsevier, http://www.scopus.com/), Web of Science (Thomson Reuters, http://apps.webof-knowledge.com/), and Google Scholar (https://scholar.goo-gle.it/). The following search criteria have been used: “Pancreatic Cancer” AND “Im-aging” AND “Detection”; “Pancreatic Cancer” AND “Imaging” AND “Characterization”; “Pan-creatic Cancer” AND “CT” AND “Detection”; “Pancreatic Cancer” AND “CT” AND “Char-acterization”; “Pancreatic Cancer” AND “MRI” AND “Detection”; “Pancreatic Cancer” AND “MRI” AND “Characterization”; “Pancreatic Cancer” AND “Functional Imaging” AND “De-tection”; “Pancreatic Cancer” AND “Functional Imaging” AND “Characterization”; “Pancreatic Cancer” AND “Radiomics” AND “Detection”; “Pancreatic Cancer” AND “Radiomics” AND “Characterization”. Moreover, the references of the found papers were evaluated for publica-tions not indexed in the electronic database. We analyzed all titles and abstracts. The inclusion criteria were based on clinical study evaluating radiological assessment of pancreatic ductal ad-enocarcinoma. Articles published in the English language from January 2000 to January 2021 were included. Exclusion criteria were studies with no sufficient reported data, case report, re-view, editorial letter or studies focused on other histological types than PDAC.

Results

We recognized 243 studies that assessed functional imaging in pancreatic cancer from January 2000 to January 2021. Ninety-five cor-responded to most criteria; 73 studies have different topics; 6 are case reports, review, or letter to editors; so, in the end, 69 articles were included (Figure 1).

Discussion

Morphological AssessmentRecent advances in radiological field and in

technology, such as CT, MRI, and hybrid nuclear imaging techniques, permit the characterization of pancreatic lesions due to morphologic, he-modynamic, and metabolic data. High-quality imaging is crucial for the detection and staging of PDAC32,33. The typical features of PDAC are a hy-povascular lesion on contrast studies. The lesion is most often located in the pancreatic head but can also be found in the body or tail of the pan-creas. Several imaging findings were detected, such as displaced calcifications in a patient with chronic calcific pancreatitis, duct-to-parenchyma ratio greater than 0.34, the double duct sign, ves-sel encasement, vessel deformity, and a superior mesenteric artery (SMA) to superior mesenteric vein (SMV) ratio of greater than 1 favor the diag-nosis of PDAC12.

A multimodality approach combines the strengths of individual imaging tools and has a synergistic effect in improving diagnostic yield.

Computed Tomography CT study acquired in the pancreatic and in

venous phases is the chosen modality at the diag-nosis and preferably no more than 4 weeks before surgery33. According to NCCN, dual-phase CT protocol is required for optimal assessment of the primary tumor, for tumor’s relationship to adja-cent arterial/venous vasculature and for detection of liver and peritoneal metastasis. The pancreas

Figure 1. Schematic representation of included and exclud-ed articles. According to our search criteria, we analyzed 51 papers that assessed CT role in PDAC; among them, 12 papers evaluated DECT and 12 papers pCT. We also ana-lyzed 27 papers for MRI; among them, 7 papers assessed DCE-MRI and 17 papers assessed DWI. We analyzed 15 papers on radiomics.


3687

has a prominent arterial supply: the highest con-trast enhancement is noticed between 35 and 45 seconds after the injection of the contrast bolus, and the peak is at nearly 40 seconds. For these reasons, the best contrast to noise ratio between hypoattenuating lesions, pancreatic parenchyma, and peripancreatic vessels is achieved in the pan-creatic phase12.

The pancreas is anatomically connected with critical vascular structures: the celiac artery, he-patic artery, superior mesenteric artery, portal vein, and superior mesenteric vein. A pancreatic phase CT (PPCT) assessment of the vascular involvement is highly predictive of the extent of vascular involvement that determines operability and overall survival16. In addition, a PPCT is also required for accurate planning and delivery of radiotherapy, which requires clear delineation of the tumor in relation to normal pancreatic paren-chyma and surrounding normal structures16,34-38.

Although the technical advances of the CT have revolutionized pancreatic imaging, there is a 10% of PDCA that remains undetected12. So, newer techniques have been introduced to increase tumor conspicuity, such as dual-energy CT (DECT) with low-energy acquisition at 80 kVp instead of the frequently applied 120 kVp (Figure 2); in addition, several researchers have assessed different techniques of contrast medium injection for improving lesion detection34,38.

DECT, which is based on simultaneously ob-taining two datasets at different energy levels, has the possibility to create virtual monochro-matic images (VMIs). These images have higher contrast at lower energy levels using an approx-imation of energy levels closer to the k-edge of iodine at 33 keV39. VMIs at lower energy levels increased the conspicuity of the pancreatic le-sion. Moreover, this technique can better depict the arteries improving contrast so that it allows

an accurate assessment of vessel involvement for surgical staging. Macari et al40 showed that at 80-kVp the difference in attenuation between tumor and normal parenchyma improves lesions conspicuity, undiagnosed at conventional 120 kVp. Marin et al38 evaluated a low-tube-voltage (80 kVp), high-tube current (675 mA) technique, showing that this can increase tumor visibility and reduce patient radiation dose. An increased lesion detection using lower keV has been later confirmed also for small lesions (<3 cm)41-43. Noda et al39 showed that the detection of pan-creatic cancer was superior on VMIs at 40-50 keV compared to that at higher energy levels, as well as that the measurements of the major and minor axes of the tumor tended to be shorter as the energy levels decreased. This was because of the greater conspicuity and edge sharpness of the PDAC on VMIs at 40 keV than those at higher energy levels. It is essential to precisely measure tumor size for the evaluation of T staging and/or treatment response assessment39.

Pancreatic cancer patients might develop bil-iary strictures that require a metallic bile duct stent. Beam hardening artifacts from metal de-vices can corrupt the image quality impeding the assessment of the pancreatic lesion. DECT also improves the image quality using metal artifact reduction algorithms43.

Regarding different techniques of contrast me-dium injection, Brook et al44 assessed the split-bo-lus technique. During the split-bolus technique, various amounts of contrast medium are admin-istered in two or three phases with a variable pause while scanning is performed only once to obtain a combined-phase image in a single scan. Split-bolus spectral CT results in pancreatic tu-mor conspicuity equal to or higher than those ob-tained with a standard technique with a reduction in the radiation dose44.

Figure 2. Dual Energy CT: An increased lesion conspicuity using lower keV (in A, 70 keV; in B 40 keV; and in C: Iodine Map).


3688

A significant dilemma is the high rate of lo-cal recurrences after surgery. Some studies45,46 have demonstrated that a standardized speci-men assessment with a R1 definition) reveals R1 resection status in as much as 70% or more of pancreatic resections45. The most vulnerable margins are shown to be mesenteric (medial) and posterior, usually due to perineural invasion CT is capable to detect the presence of perineural invasion. However, CT demonstrates some false positive results46.

Magnetic Resonance ImagingTen percent of PDACs are isoattenuating and

remain the main challenge in the detection of lesions, even further limiting the diagnostic accu-racy of CT. Because of its better contrast resolu-tion, MRI is useful in these cases, with up to 79% sensitivity for detection of isoattenuating tumors at CT47. Also, several studies48,49 have shown that MRI with MR cholangiopancreatography (Figure 3) can provide superior tumor detection and similar diagnostic performance in evaluating resectability compared with CT. Therefore, MRI challenged the role of imaging in pre-treatment assessment in the past few decades, and it is thought that MRI is superior to CT in detecting tumor extension50.

At MRI, the typical feature of PDAC is a hypointense lesion at fat-suppressed (FS) T1-W and iso-hyperintense lesion at T2-W imag-es. During contrast studies, pancreatic cancer shows hypoenhancement, more evident during pancreatic phase51. Diffusion restriction im-proves the detection of small lesions (Figure 4). DWI sequences allow the detection of pan-creatic masses with similar sensitivity to that of post-contrast sequences. Pancreatic cancer shows increased signal intensity (SI) at b >500 sec/mm2 with relatively low signal on ADC map52-55.

Concerning the detection of vascular inva-sion, MRI has a sensitivity of 83%-94% and specificity of more than 95% [49]. Zhang et al49 showed no significant difference comparing MRI with CT in diagnosing vascular inva-sion. Contrast-enhanced (ce) sequences make use of the shortened T1 relaxation times of blood after the injection of contrast media and thus achieves high signal intensity from blood with the use of 3D gradient-echo sequences. It allows the acquisition of a large volume da-ta set in a single breath-hold contrast phase. Nevertheless, CT is more available and easier to perform. A CT examination is considerably shorter than an MRI analysis (15-20 min vs. 1

Figure 3. Small adenocarcinoma of the pancreatic head: in T2-W sequences (A and B), the lesion is iso-hyperintense. In MR cholangiopancreatography (C), a ductal cutoff sign is displayed.


3689

h), and it does not constrain the patient as much as MRI49. Thus, CT is more accepted and more favored by patients, and it has become reason-ably cost-effective49.

Concerning patient management, Kim et al56 demonstrated that additional MRI resulted in changes in surgical resectability and treatment in a significant number of patients who have potentially resectable states at CT, owing to the additional detection of liver metastases, major vascular invasion, enlarged para-aortic lymph nodes, and peritoneal metastases56.

Based on the similar diagnostic accuracy, an MRI study is reserved to those patients who are not able to benefit from CT or if CT findings are inconclusive. According to NCCN, “pancreas protocol MRI with contrast can be a helpful ad-junct to CT in the staging of pancreatic cancer, particularly for characterization of CT indeter-minate liver lesions and when suspected tumors are not visible on CT or in cases of contrast allergy”57.

Functional Assessment

Perfusion CTPerfusion CT (pCT) is a radiological tool based

on contrast kinetics of tissue providing quantita-tive data on hemodynamics. pCT assesses dy-namic change in tissue iodine concentration over time, allowing to estimate specific features, such as blood flow (BF), blood volume (BV), time to peak concentration (TTP), vascular permeabil-ity surface area product (PS), and permeability (Ktrans), which can be used as surrogates for tumor vascularization, vascular immaturity, and perfusion pressure. Several researchers58-60 re-ported that pCT is a considerable tool for differen-tiating benign from malignant lesions, assessing treatment response, and defining angiogenesis. In pancreatic cancer patients, perfusion analysis is of interest since hypo-enhancement is associated with fibro-sclerotic stroma and represents an in-dependent negative predictor of survival61. Exist-ing papers for PDAC perfusion CT include a wide

Figure 4. The same patient of figure 3. No lesion is detected in DWI sequences (A and B), while diffusion restriction shows obstructive pancreatitis (arrow in A). Arterial (C) and venous phase (D) during contrast study do not show lesions.


3690

variety of technology, among which scanners are manufactured by different companies and models ranging from 64-to 640-slice. Most importantly, all were performed separately from the standard of care CT, requiring a separate scan so as IV contrast injection different from the standard of care. Notably, IV contrast injection rate for per-fusion CT has been much higher (> 5 ml/s) than what is used for routine examinations60-64. Scialpi et al43 assessed pCT imaging to detect small (≤2 cm) PDAC, showing that the quantitative analysis should increase the sensitivity of CT in the de-tection of small PDAC. Zamboni et al65 assessed the rule of qualitative analysis evaluating the morphology of the time-density curves obtained by normal parenchyma, tumors, and atrophic parenchyma. The authors showed that the curve with no washout in the last portion corresponds to adenocarcinoma; conversely, the curve with at least some washout in the last portion corre-sponds to chronic pancreatitis65. Furthermore, Delrue et al66 showed no significant differences in perfusion values between acute/chronic pancre-atitis and pancreatic adenocarcinoma. Functional imaging from CT may add useful data on tumor aggressiveness affecting treatment strategy and patient management. Perfusion features have al-ready been shown to discriminate high- and low-grade PDAC and to assess response to neo-adjuvant therapy61,62. Nishikawa et al67 evaluated the relationship between prognosis and perfusion assessment, suggesting that prognosis is related to increased perfusion in the tissue surrounding cancer.

DECT perfusion is considered one of the new promising applications of DECT. Although still at its preliminary stage, there is evidence of po-tential for improving the delineation of PDAC. Klauss et al68 demonstrated that DECT perfusion could improve the visualization of PDAC via increased iodine contrast enhancement available at 80 kVp, and in combination with the noise re-duction at 140 kVp.

Dynamic Contrast-Enhanced-MRI DCE MRI can offer data on tumor microvas-

culature, and it has been used as a biomarker for tumor response to treatment. DCE-MRI-derived parameters should be related to intratumoral mi-crovascular density (MVD), and the rationale is that the MVD count is the most consistent instru-ment for evaluating tumor angiogenesis, and it is correlated with overall survival and metastatic status8. Besides the angiogenesis required for

tumor growth, in pancreatic cancer, we found a well-developed stroma, rich in myofibroblasts representing an encouraging microenvironment for cancer cells, leading to development, inva-sion, and metastasis8. This stroma is also respon-sible for chemoresistance and radioresistance8. Quantitative data obtained by DCE-MRI reflect the underlying perfusion and/or permeability of the target tissue. Ktrans can have different phys-iologic interpretations depending on the balance between the blood flow and capillary permeabil-ity in the target tissue. If the contrast material uptake is flow limited, then Ktrans is related to the tissue perfusion. This condition is encoun-tered in normal pancreas and in tumors with high vascular permeability8. Beyond the hypothesis, the role of DCE-MRI in the pancreatic lesion is unclear. Several researchers69-71 assessed the role of DCE-MRI for the characterization of solid pancreatic diseases. Kim et al69 evaluated patients with adenocarcinoma (24 patients), with neuro-endocrine tumors (8 patients), with chronic pan-creatitis (3 patients), and with a normal pancreas (10 patients), showing that in patients with PDAC, Ktrans, kep, and iAUC (initial area under curve) values were significantly lower than in patients with a normal pancreas. Bali et al70 demonstrated that Ktrans parameters were significantly lower in malignant lesions with respect to benign. On the contrary, Zhang et al71 showed that, although there were differences in perfusion curves be-tween PDCA and mass-forming pancreatitis, this feature alone did not allow consistent diagnosis. These data are similar to the data which we found in our previous study8. We assessed semi-quanti-tative descriptors of intensity/time curves (Figure 5): maximum signal difference (MSD), TTP, the wash-in slope (WIS), the washout slope (WOS), the wash-in intercept (WII), the washout inter-cept (WOI), the WOS/WIS ratio, and the WOI/WII ratio, showing that there were no differenc-es for dynamic parameters except a statistically non-significant difference for WIS among pan-creatic tumors, peritumoral inflammatory tissue, and normal pancreatic parenchyma8.

Diffusion-Weighted Imaging-MRIDWI evaluates the microstructure of tissue by

means of the water proton mobility and cellular density evaluation. The water diffusion mobility is linked to cell density, tissue vascularity, and viscosity. Using a mono-exponential model or a bi-exponential model, it is possible to identify biomarkers for fibrosis, tumor fluid, and cell


3691

membrane integrity72-74. By means of the Intra-voxel Incoherent Motion method (IVIM), we can calculate the pure tissue coefficient (Dt) separate-ly linked exclusively to diffusion water mobility, the pseudo-diffusion coefficient (Dp) linked to miscroscopic blood mobility and the perfusion fraction (fp).

The conventional DWI analysis is based on the hypothesis that voxel water diffusion has a single component and that their distribution is Gaussian. On the contrary, water molecules dif-fusion within the tissue has non-Gaussian behav-ior75. Therefore, Jensen et al75 in 2005 reported a non-Gaussian diffusion model called Diffusion Kurtosis Imaging (DKI) (Figure 6). Considering this model, it is possible the evaluation of the

mean value of the kurtosis coefficient (MK) and the mean value of the diffusion coefficient (MD) with the correction of the non-Gaussian bias.

DWI can furnish additional data to identify focal pancreatic lesions, verifying more restricted diffusion in solid malignant tumors vs. benign inflammatory ones76,79. Lee et al76 showed that normal pancreas had higher mean ADC than pancreatic cancer or mass-forming pancreatitis. ADC and Dt of mass-forming pancreatitis were lower than those of pancreatic cancer. There-fore, the measurement of ADC and Dt may be helpful to differentiate pancreatic cancers from mass-forming pancreatitis. ADC values of PDAC was lower than normal pancreas in several stud-ies76-80. Muraoka et al78 showed that ADC values

Figure 6. Same patient of Figure 1. ADC Map (a), MK Map (b) and MD Map (c).

Figure 5. Adenocarcinoma of the pancreatic head: in a (HASTE T2-W sequence), the lesion (arrow) appears hyperintense, in b (in-phase T1-W sequence) and c (out-phase T1-W sequence) hypointense and hypovascular in d (VIBE T1-W in equilibrium phase); in e and f two time-intensity curves of dynamic perfusion study by two regions of interest.


3692

of the PDAC were linked to the tumors’ histo-pathologic composition: when the tumor reveals loose fibrosis, the ADC values can be higher than the normal pancreas; when dense fibrosis and increased cellular elements are present, the ADC values were lower than the normal pan-creas. The mean sensitivity and specificity for the PDAC detection were 96.2% and 98.6%81. Previous studies81,82 on IVIM model have shown that the reduced ADC in PDACs could be linked to a difference in perfusion fraction, which is re-duced in PDACs. Therefore, fp is the best feature among DWI-derived parameters to differentiate pancreatitis from PDACs83. In our previous inves-tigation81, we assessed ADC, IVIM parameters, and kurtosis coefficient, showing that the per-fusion-related factors of PDAC, fp and Dp, and MD of DKI differed from those seen in patients with normal parenchyma and in peritumoral tis-sue. Our results81 support the hypothesis that the kurtosis effect could have a better performance to differentiate pancreatic tumors, peritumoral inflammatory tissue, and normal pancreatic pa-renchyma.

In consideration of its great potential, DWI could be included as a complementary imaging technique in the standard MRI protocol of the pancreas84.

BOLD-MRIHypoxia is emerging as a biomarker of aggres-

sive tumor biology. Compared to other solid tu-mors, oxygen levels in PDAC are among the low-est10. Several factors contribute to the low oxygen levels, such as low microvascular density, rapid tumor growth, extensive desmoplastic stroma10. In PDAC cells, hypoxia induces the expression of a number of genes, especially tumor-promot-ing transcription factors such as the hypoxia-in-ducible factors (HIFs). Reduced oxygen levels and subsequent stabilization of HIF-1 facilitate stromal remodeling by activation of pancreatic stellate cells to produce matrix molecules. Then, hypoxia and the PDAC dense tumor stroma are interdependent and may potentiate each other in a positive feedback loop, called hypoxia-fibrosis cycle10. The assessment of the hypoxic fraction in PDAC could have a potential impact on treat-ment stratification and could improve therapy outcome10. BOLD-MRI makes use of paramag-netic properties of deoxyhemoglobin in order to regional quantification of oxygenation26.

To the best of our knowledge, no paper as-sessed the role of BOLD-MRI in PDCA. Mayer

et al10 showed that DW-MRI is well suited for the evaluation of tumor hypoxia in PDAC and could potentially be used for the identification of lesions with a high hypoxic fraction, which is at high risk for failure of radio-chemotherapy.

Radiomics AnalysisRadiomics consists of several parameters ex-

traction by images that can provide information about tumor phenotype as well as the cancer microenvironment85. Radiomics, if combined with patient outcome data, could produce evi-dence-based clinical-decision support systems. The main objective is to combine several mul-timodal quantitative data with mathematical methods to provide clear and robust parameters allowing an outcome prediction85. Radiomics of-fers outstanding benefits over qualitative imaging assessment since this is clearly limited by the subjective evaluation of radiologists. A radiomic information extension can be obtained by adding genomics data (radiogenomics); in fact, genomic markers such as microRNA expression have been shown associated with treatment response, meta-static spread, and prognosis that could offer per-sonalized and precision medicine85. In PDAC, CT is the main diagnostic tool for assessment of the local extent of disease and surgical planning. CT images can be used to extract radiomic features without extra image acquisition cost providing comprehensive information on the tumor pheno-typic and textural structure. Park et al86 assessed CT-based machine learning of radiomics features in order to distinguish autoimmune pancreatitis (AIP) from PDAC, reporting an overall accuracy of 95.2%. Similar results were found by Chu et al87 in the assessment of CT radiomics features to differentiate PDCA from normal pancreatic tissue. Several studies88-93 have shown the prog-nostic value of CT radiomic features in PDCA. A larger cohort of 161 patients90 with resected PDAC was recently analyzed to evaluate the prognostic accuracy of radiomic features com-bined with preoperative serum carbohydrate anti-gen 19.9 (CA19.9 levels) and pathology score (The Brennan score). In this study was reported that adding the pathology score to radiomic features and serum cancer antigen improves the prognos-tic power of the model90.

Iwatate et al94 examined whether the expression of p53 and PD-L1 could be predicted from CT images. They found that p53 and PD-L1 were sig-nificant independent prognostic factors. The area under the receiver operating characteristic curve


3693

for p53 and PD-L1 predictive models was 0.795 and 0.683, respectively. Radiogenomics predicted p53 mutations were significantly associated with poor prognosis, whereas the predicted abnormal expression of PD-L1 was not significant94.

Texture analysis (TA) is a form of radiomics that refers to quantitative measurements of the histogram, distribution and/or relationship of im-age pixels signal intensity. MR-TA has multiple limitations: many texture data are sensitive to multiparametric acquisition and reconstruction data (flip angle, repetition time, echo time, field-of-view, contrast, slice-thickness, and reconstruc-tion algorithms affect pixels intensity, spatial relationships, and edges)95. To minimize these ef-fects, image protocol standardization and the use of image filtration methods have been utilized95. Another major challenge with MR-TA is the volume produced data: with many texture tools generating hundreds or thousands of measure-ments. Moreover, it is difficult to understand the texture parameters meaning, and it is complicated to identify relationships between one or more texture features and a biologic outcome when the number of texture parameters exceeds the patient sample size95. Quantitative MR-TA has been eval-uated in a limited manner in PDAC. A retrospec-tive study, including 66 patients with pancreatic cancer, found that tumor size and MR-TA data were predictive of both recurrence-free survival and overall survival in univariate analysis. In contrast, only tumor size remained predictive in multivariate analysis96.

ConsiderationsThe possibility of multimodal imaging, inte-

grating the different methods with each other, allows an adequate stratification of patients based on risk and then the choice of the most suitable treatment97-127. As a standard of care imaging mo-dality, CT images can be used to obtain morpho-logical data and functional data to increase tu-mor conspicuity and characterization. Although DECT and pCT have been the subject of several studies, the data still need further verification, while the role of DCE-MRI is still unclear. Oth-erwise, a greater interest is reserved for DWI and DKI, as well as radiomics, considering that the extraction of the radiomic features without extra image acquisition cost to the healthcare system could provide comprehensive data on the tumor phenotypic and textural structure in order to allow patients outcome prediction and more personalized treatment.

Conclusions

Diagnosis of PDAC remains challenging due to overlapping imaging features with benign lesions and not-withstanding great advances with CT and MRI. However, proper detection and char-acterization of pancreatic lesions are mandatory because the prognosis is linked to tumor type and grade, as well as correct staging on imaging. The possibility of functional imaging, such as pCT, DCE-MRI, DWI, DKI, and radiomics approach-es, has revolutionized pancreatic imaging, not only improving the detection and the characteri-zation of the lesions but also allowing a prognosis related to radiological features, favoring the pro-cess versus a personalized medicine.

Conflict of InterestThe Authors declare that they have no conflict of interests.

AcknowledgementsThe authors are grateful to Alessandra Trocino, librarian at the National Cancer Institute of Naples, Italy. Moreover, for the collaboration, authors are grateful to Assunta Zazzaro, Paolo Pariante and Dr Ivano Rossi, TSRM at Radiology Di-vision, “Istituto Nazionale Tumori IRCCS Fondazione Pas-cale – IRCCS di Napoli”, Naples, Italy.

References

1) Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin 2020; 70: 7-30.

2) NCCN.org. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®) Pancreatic Ade-nocarcinoma Version 1.2020. (2020). Available at: https://www.nccn.org/professionals/physician_gls/default.aspx.

3) Tempero MA. NCCN Guidelines updates: pancreatic cancer. J Natl Compr Canc Netw 2019; 17: 603-605.

4) Alistar A, Morris BB, Desnoyer R, Klepin HD, Hos-seinzadeh K, Clark C, Cameron A, Leyendecker J, D’Agostino R Jr, Topaloglu U, Boteju LW, Bote-ju AR, Shorr R, Zachar Z, Bingham PM, Ahmed T, Crane S, Shah R, Migliano JJ, Pardee TS, Miller L, Hawkins G, Jin G, Zhang W, Pasche B. Safe-ty and tolerability of the first-in-class agent CPI-613 in combination with modified FOLFIRINOX in patients with metastatic pancreatic cancer: a sin-gle-centre, open-label, dose-escalation, phase 1 trial. Lancet Oncol 2017; 18: 770-778.

5) Biondi M, Mormile M, Banci Buonamici F, Di Gi-orgio E, Spadafora M, Tini P, Grassi R, Pirtoli L, Correale P, Cappabianca S, Grassi R. Delta-ra-diomics increases multicentre reproducibility: a phantom study. Med Oncol 2020; 37: 38.


3694

6) Addeo G, Beccani D, Cozzi D, Ferrari R, Lanzetta MM, Paolantonio P, Pradella S, Miele V. Groove pancreatitis: a challenging imaging diagnosis. Gland Surg 2019; S178-S187.

7) Granata V, Grassi R, Fusco R, Setola SV, Palaia R, Belli A, Miele V, Brunese L, Grassi R, Petrillo A, Izzo F. Assessment of ablation therapy in pan-creatic cancer: the radiologist’s challenge. Front Oncol 2020; 10: 560952.

8) Granata V, Fusco R, Sansone M, Grassi R, Maio F, Palaia R, Tatangelo F, Botti G, Grimm R, Cur-ley S, Avallone A, Izzo F, Petrillo A. Magnet-ic resonance imaging in the assessment of pan-creatic cancer with quantitative parameter ex-traction by means of dynamic contrast-enhanced magnetic resonance imaging, diffusion kurtosis imaging and intravoxel incoherent motion diffu-sion-weighted imaging. Therap Adv Gastroenter-ol 2020; 13: 1756284819885052.

9) Sun J, Yang L, Zhou Z, Zhang D, Han W, Zhang Q, Peng Y. Performance evaluation of two itera-tive reconstruction algorithms, MBIR and ASIR, in low radiation dose and low contrast dose abdom-inal CT in children. Radiol Med 2020; 125: 918-925.

10) Mayer P, Fritz F, Koell M, Skornitzke S, Bergmann F, Gaida MM, Hackert T, Maier-Hein K, Laun FB, Kauczor HU, Grenacher L, Klauß M, Stiller W. As-sessment of tissue perfusion of pancreatic cancer as potential imaging biomarker by means of Intra-voxel incoherent motion MRI and CT perfusion: correlation with histological microvessel density as ground truth. Cancer Imaging 2021 Jan 19; 21: 13.

11) Noda Y, Tochigi T, Parakh A, Kambadakone A. Simulated twin-phase pancreatic CT generated using single portal venous phase dual-energy CT acquisition in pancreatic ductal adenocarcinoma. Abdom Radiol (NY) 2021; doi: 10.1007/s00261-020-02921-9

12) Granata V, Fusco R, Catalano O, Setola SV, de Lutio di Castelguidone E, Piccirillo M, Palaia R, Grassi R, Granata F, Izzo F, Petrillo A. Multide-tector computer tomography in the pancreatic adenocarcinoma assessment: an update. Infect Agent Cancer 2016; 11: 57.

13) Schueller G, Schima W, Schueller-Weidekamm C, Weber M, Stift A, Gnant M, Prokesch R. Multi-detector CT of pancreas: effects of contrast mate-rial flow rate and individualized scan delay on en-hancement of pancreas and tumor contrast. Ra-diology 2006; 241: 441-448.

14) Kondo H, Kanematsu M, Goshima S, Miyoshi T, Shiratori Y, Onozuka M, Moriyama N, Bae KT. MDCT of the pancreas: optimizing scanning de-lay with a bolus-tracking technique for pancreat-ic, peripancreatic vascular, and hepatic contrast enhancement. AJR Am J Roentgenol 2007; 188: 751-756.

15) Marin D, Nelson RC, Barnhart H, Schindera ST, Ho LM, Jaffe TA, Yoshizumi TT, Youngblood R, Samei E. Detection of pancreatic tumors, image

quality, and radiation dose during the pancreat-ic parenchymal phase: effect of a low-tube-volt-age, high-tube-current CT technique-preliminary results. Radiology 2010; 256: 450-459.

16) Maharaj AD, Evans SM, Zalcberg JR, Ioannou LJ, Graco M, Croagh D, Pilgrim CHC, Dodson T, Goldstein D, Philip J, Kench JG, Merrett ND, Ne-ale RE, White K, Evans P, Leong T, Green SE. Barriers and enablers to the implementation of protocol-based imaging in pancreatic cancer: a qualitative study using the theoretical domains framework. PLoS One 2020; 15: e0243312.

17) Granata V, Fusco R, Avallone A, Catalano O, Pic-cirillo M, Palaia R, Nasti G, Petrillo A, Izzo F. A radiologist’s point of view in the presurgical and intraoperative setting of colorectal liver metasta-ses. Future Oncol 2018; 14: 2189-2206.

18) Granata V, Fusco R, Setola SV, Palaia R, Albino V, Piccirillo M, Grimm R, Petrillo A, Izzo F. Diffu-sion kurtosis imaging and conventional diffusion weighted imaging to assess electrochemotherapy response in locally advanced pancreatic cancer. Radiol Oncol 2019; 53: 15-24.

19) Granata V, Fusco R, Avallone A, Cassata A, Palaia R, Delrio P, Grassi R, Tatangelo F, Grazz-ini G, Izzo F, Petrillo A. Abbreviated MRI protocol for colorectal liver metastases: how the radiolo-gist could work in pre surgical setting. PLoS One 2020; 15: e0241431.

20) Fusco R, Granata V, Petrillo A. Introduction to special issue of radiology and imaging of cancer. Cancers (Basel) 2020; 12: 2665.

21) Devries AF, Griebel J, Kremser C, Judmaier W, Gneiting T, Kreczy A, Ofner D, Pfeiffer KP, Brix G, Lukas P. Tumor microcirculation evaluated by dy-namic magnetic resonance imaging predicts ther-apy outcome for primary rectal carcinoma. Can-cer Res 2001; 61: 2513-2516.

22) DeVries AF, Kremser C, Hein PA, Griebel J, Krez-cy A, Ofner D, Pfeiffer KP, Lukas P, Judmaier W. Tumor microcirculation and diffusion predict ther-apy outcome for primary rectal carcinoma. Int J Radiat Oncol Biol Phys 2003; 56: 958-996.

23) Fusco R, Granata V, Maio F, Sansone M, Petril-lo A. Textural radiomic features and time-intensi-ty curve data analysis by dynamic contrast-en-hanced MRI for early prediction of breast cancer therapy response: preliminary data. Eur Radiol Exp 2020; 4: 8.

24) Tan Z, Xu J, Zhang B, Shi S, Yu X, Liang C. Hy-poxia: a barricade to conquer the pancreatic can-cer. Cell Mol Life Sci 2020; 77: 3077-3083.

25) Mayer P, Kraft A, Witzel HR, Marnet N, Hörner N, Roth W, Heinrich S, Hackert T, Bergmann F, Kauczor HU, Klauss M, Gaida MM. Restricted water diffusion in diffusion-weighted magnetic resonance imaging in pancreatic cancer is as-sociated with tumor hypoxia. Cancers 2021; 13: 89.

26) Fusco R, Granata V, Pariante P, Cerciello V, Siani C, Di Bonito M, Valentino M, Sansone M, Bot-


3695

ti G, Petrillo A. Blood oxygenation level depen-dent magnetic resonance imaging and diffusion weighted MRI imaging for benign and malignant breast cancer discrimination. Magn Reson Imag-ing 2021; 75: 51-59.

27) Park S, Chu LC, Hruban RH, Vogelstein B, Kinzler KW, Yuille AL, Fouladi DF, Shayesteh S, Ghandi-li S, Wolfgang CL, Burkhart R, He J, Fishman EK, Kawamoto S. Differentiating autoimmune pan-creatitis from pancreatic ductal adenocarcinoma with CT radiomics features. Diagn Interv Imaging 2020; 101: 555-564.

28) Zhang Y, Cheng C, Liu Z, Wang L, Pan G, Sun G, Chang Y, Zuo C, Yang X. Radiomics analysis for the differentiation of autoimmune pancreatitis and pancreatic ductal adenocarcinoma in 18 F-FDG PET/CT. Med Phys 2019; 46: 4520-4530.

29) E L, Xu Y, Wu Z, Li L, Zhang N, Yang H, Schwartz LH, Lu L, Zhao B. Differentiation of focal-type au-toimmune pancreatitis from pancreatic ductal ad-enocarcinoma using radiomics based on multi-phasic computed tomography. J Comput Assist Tomogr 2020; 44: 511-518.

30) Kaissis G, Ziegelmayer S, Lohöfer F, Steiger K, Algül H, Muckenhuber A, Yen HY, Rummeny E, Friess H, Schmid R, Weichert W, Siveke JT, Braren R. A machine learning algorithm predicts molecular subtypes in pancreatic ductal adeno-carcinoma with differential response to gemcit-abine-based versus FOLFIRINOX chemotherapy. PLoS One 2019; 14: e0218642.

31) Kaissis GA, Ziegelmayer S, Lohöfer FK, Harder FN, Jungmann F, Sasse D, Muckenhuber A, Yen HY, Steiger K, Siveke J, Friess H, Schmid R, We-ichert W, Makowski MR, Braren RF. Image-based molecular phenotyping of pancreatic ductal ade-nocarcinoma. J Clin Med 2020; 9: 724.

32) Boursi B, Finkelman B, Giantonio BJ, Haynes K, Rustgi AK, Rhim AD, Mamtani R, Yang YX. A clinical prediction model to assess risk for pan-creatic cancer among patients with prediabetes. Eur J Gastroenterol Hepatol 2021; doi: 10.1097/MEG.0000000000002052.

33) Kulkarni NM, Soloff EV, Tolat PP, Sangster GP, Fleming JB, Brook OR, Wang ZJ, Hecht EM, Zins M, Bhosale PR, Arif-Tiwari H, Mannelli L, Kambadakone AR, Tamm EP. White paper on pancreatic ductal adenocarcinoma from society of abdominal radiology’s disease-focused pan-el for pancreatic ductal adenocarcinoma: Part I, AJCC staging system, NCCN guidelines, and borderline resectable disease. Abdom Radiol 2020; 45: 716-728.

34) Fenchel S, Fleiter TR, Aschoff AJ, van Gessel R, Brambs HJ, Merkle EM. Effect of iodine concen-tration of contrast media on contrast enhance-ment in multislice CT of the pancreas. Br J Radi-ol 2004; 77: 821-830.

35) Shinagawa M, Uchida M, Ishibashi M, Nishimu-ra H, Hayabuchi N. Assessment of pancreatic CT enhancement using a high concentration of con-trast material. Radiat Med 2003; 21: 74-79.

36) Schueller G, Schima W, Schueller-Weidekamm C, Weber M, Stift A, Gnant M, Prokesch R. Multi-detector CT of pancreas: effects of contrast mate-rial flow rate and individualized scan delay on en-hancement of pancreas and tumor contrast. Ra-diology 2006; 241: 441-448.

37) Kondo H, Kanematsu M, Goshima S, Miyoshi T, Shiratori Y, Onozuka M, Moriyama N, Bae KT. MDCT of the pancreas: optimizing scanning delay with a bolus-tracking technique for pancreatic, peri-pancreatic vascular, and hepatic contrast enhance-ment. AJR Am J Roentgenol 2007; 188: 751-756.

38) Marin D, Nelson RC, Barnhart H, Schindera ST, Ho LM, Jaffe TA, Yoshizumi TT, Youngblood R, Samei E. Detection of pancreatic tumors, image quality, and radiation dose during the pancreat-ic parenchymal phase: effect of a low-tube-volt-age, high-tube-current CT technique-preliminary results. Radiology 2010; 256: 450-459.

39) Noda Y, Goshima S, Kaga T, Ando T, Miyoshi T, Kawai N, Kawada H, Tanahashi Y, Matsuo M. Vir-tual monochromatic image at lower energy level for assessing pancreatic ductal adenocarcinoma in fast kV-switching dual-energy CT. Clin Radiol 2020; 75: 320.e17-320.e23.

40) Macari M, Spieler B, Kim D, Graser A, Megibow AJ, Babb J, Chandarana H. Dual-source dual-en-ergy MDCT of pancreatic adenocarcinoma: initial observations with data generated at 80 kVp and at simulated weighted-average 120 kVp. AJR Am J Roentgenol 2010; 194: W27-W32.

41) Mileto A, Mazziotti S, Gaeta M, Bottari A, Zimbaro F, Giardina C, Ascenti G. Pancreatic dual-source dual-energy CT: is it time to discard unenhanced imaging? Clin Radiol 2012; 67: 334-339.

42) Patel BN, Rosenberg M, Vernuccio F, Ramirez-Gi-raldo JC, Nelson R, Farjat A, Marin D. Charac-terization of small incidental indeterminate hy-poattenuating hepatic lesions: added value of single-phase contrast-enhanced dual-energy CT material attenuation analysis. AJR Am J Roent-genol 2018; 211: 571-579.

43) Mastrodicasa D, Delli Pizzi A, Patel BN. Dual-en-ergy CT of the pancreas. Semin Ultrasound CT MR 2019; 40: 509-514.

44) Brook OR, Gourtsoyianni S, Brook A, Siewert B, Kent T, Raptopoulos V. Split- bolus spectral mul-tidetector CT of the pancreas: assessment of ra-diation dose and tumor conspicuity. Radiology 2013; 269: 139-1448.

45) Esposito I, Kleeff J, Bergmann F, Reiser C, Her-pel E, Friess H, Schirmacher P, Büchler MW. Most pancreatic cancer resections are R1 resec-tions. Ann Surg Oncol 2008; 5: 1651-1660.

46) Khristenko E, Shrainer I, Setdikova G, Palkina O, Sinitsyn V, Lyadov V. Preoperative CT-based de-tection of extrapancreatic perineural invasion in pancreatic cancer. Sci Rep 2021; 11: 1800.

47) Schawkat K, Manning MA, Glickman JN, Mortele KJ. Pancreatic ductal adenocarcinoma and its variants: pearls and perils. Radiographics 2020; 40: 1219-1239.


3696

48) Park HS, Lee JM, Choi HK, Hong SH, Han JK, Choi BI. Preoperative evaluation of pancreatic cancer: comparison of gadolinium-enhanced dy-namic MRI with MR cholangiopancreatography versus MDCT. J Magn Reson Imaging 2009; 30: 586-595.

49) Zhang Y, Huang J, Chen M, Jiao LR. Pre-oper-ative vascular evaluation with computed tomog-raphy and magnetic resonance imaging for pan-creatic cancer: a meta-analysis. Pancreatology 2012; 12: 227–233.

50) Lopez Hänninen E, Amthauer H, Hosten N, Ricke J, Böhmig M, Langrehr J, Hintze R, Neu-haus P, Wiedenmann B, Rosewicz S, Felix R. Prospective evaluation of pancreatic tumors: ac-curacy of MR imaging with MR cholangiopan-creatography and MR angiography. Radiology 2002; 224: 34e41.

51) Wolske KM, Ponnatapura J, Kolokythas O, Burke LMB, Tappouni R, Lalwani N. Chronic pancreati-tis or pancreatic tumor? A problem-solving ap-proach. Radiographics 2019; 39: 1965-1982.

52) High-b value diffusion-weighted MRI for detecting pancreatic adenocarcinoma: preliminary results. High-b value diffusion-weighted MRI for detect-ing pancreatic adenocarcinoma: preliminary re-sults. AJR Am J Roentgenol 2007; 188: 409-414.

53) Wang Y, Miller FH, Chen ZE, Merrick L, Mortele KJ, Hoff FL, Hammond NA, Yaghmai V, Nikolaid-is P. Diffusion-weighted MR imaging of solid and cystic lesions of the pancreas. Radiographics 2011; 31: E47-E64.

54) Dong Y, D’Onofrio M, Hocke M, Jenssen C, Pot-thoff A, Atkinson N, Ignee A, Dietrich CF. Autoim-mune pancreatitis: imaging features. Radiology 2004; 233: 345-352.

55) Srisajjakul S, Prapaisilp P, Bangchokdee S. CT and MR features that can help to differentiate be-tween focal chronic pancreatitis and pancreatic cancer. Radiol Med 2020; 125: 356-364.

56) Kim HJ, Park MS, Lee JY, Han K, Chung YE, Choi JY, Kim MJ, Kang CM. incremental role of pan-creatic magnetic resonance imaging after staging computed tomography to evaluate patients with pancreatic ductal adenocarcinoma. Cancer Res Treat 2019; 51: 24-33.

57) Liu Y, Xu XQ, Lin XZ, Song Q, Chen KM. Pro-spective study comparing two iodine concentra-tions for multidetector computed tomography of the pancreas. Radiol Med 2010; 115: 898-905.

58) O’Malley RB, Soloff EV, Coveler AL, Cox DH, De-sai N, Busey JM, Valentin GM, Wang CL. Feasi-bility of wide detector CT perfusion imaging per-formed during routine staging and restaging of pancreatic ductal adenocarcinoma. Abdom Radi-ol (NY) 2020; doi: 10.1007/s00261-020-02786-y

59) Klotz E, Haberland U, Glatting G, Schoenberg SO, Fink C, Attenberger U, Henzler T. Technical prerequisites and imaging protocols for CT perfu-sion imaging in oncology. Eur J Radiol 2015; 84: 2359-2367.

60) Prezzi D, Khan A, Goh V. Perfusion CT imaging of treatment response in oncology. Eur J Radiol 2015; 84: 2380-2385.

61) Fukukura Y, Takumi K, Higashi M, Shinchi H, Kamimura K, Yoneyama T, Tateyama A. Con-trast-enhanced CT and diffusion-weighted MR imaging: performance as a prognostic factor in patients with pancreatic ductal adenocarcinoma. Eur J Radiol 2014; 83: 612-619.

62) D’Onofrio M, Gallotti A, Mantovani W, Crosara S, Manfrin E, Falconi M, Ventriglia A, Zamboni GA, Manfredi R, Pozzi Mucelli R. Perfusion CT can predict tumoral grading of pancreatic adenocarci-noma. Eur J Radiol 2012; 82: 227-233.

63) Hamdy A, Ichikawa Y, Toyomasu Y, Nagata M, Nagasawa N, Nomoto Y, Sami H, Sakuma H. Perfusion CT to assess response to neoadjuvant chemotherapy and radiation therapy in pancreat-ic ductal adenocarcinoma: initial experience. Ra-diology 2019; 292: 628-635.

64) Scialpi M, Cagini L, Pierotti L, De Santis F, Pusi-ol T, Piscioli I, Magli M, D’Andrea A, Brunese L, Rotondo A. Detection of small (≤2 cm) pancreat-ic adenocarcinoma and surrounding parenchy-ma: correlations between enhancement patterns at triphasic MDCT and histologic features. BMC Gastroenterol 2014; 21: 14: 16.

65) Zamboni GA, Bernardin L, Pozzi MR. Dynam-ic MDCT of the pancreas: is time-density curve morphology useful for the differential diagnosis of solid lesions? a preliminary report. Eur J Radiol 2012; 81: e381-e385.

66) Delrue L, Blanckaert P, Mertens D, Van Meer-beeck S, Ceelen W, Duyck P. Tissue perfusion in pathologies of the pancreas: assessment using 128-slice computed tomography. Abdom Imaging 2012; 37: 595-601.

67) Nishikawa Y, Tsuji Y, Isoda H, Kodama Y, Chiba T. Perfusion in the tissue surrounding pancreatic cancer and the patient’s prognosis. Biomed Res Int 2014; 2014: 648021.

68) Klauss M, Stiller W, Pahn G, Fritz F, Kieser M, Werner J, Kauczor HU, Grenacher L. Dual-ener-gy perfusion-CT of pancreatic adenocarcinoma. Eur J Radiol 2013; 82: 208-214.

69) Kim JH, Lee JM, Park JH, Kim SC, Joo I, Han JK, Choi BI. Solid pancreatic lesions: characteriza-tion by using timing bolus dynamic contrast-en-hanced MR imaging assessment–a preliminary study. Radiology 2013; 266: 185-196.

70) Bali MA, Metens T, Denolin V, Delhaye M, Demet-ter P, Closset J, Matos C. Tumoral and nontumoral pancreas: correlation between quantitative dynam-ic contrast-enhanced MR imaging and histopatho-logic parameters. Radiology 2011; 261: 456-466.

71) Zhang TT, Wang L, Liu HH, Zhang CY, Li XM, Lu JP, Wang DB. Differentiation of pancreatic carci-noma and mass-forming focal pancreatitis: qual-itative and quantitative assessment by dynam-ic contrast-enhanced MRI combined with diffu-sion-weighted imaging. Oncotarget 2017; 8: 1744-1759.


3697

72) Granata V, Fusco R, Risi C, Ottaiano A, Avallone A, De Stefano A, Grimm R, Grassi R, Brunese L, Izzo F, Petrillo A. Diffusion-weighted MRI and dif-fusion kurtosis imaging to detect RAS mutation in colorectal liver metastasis. Cancers (Basel) 2020; 12: 2420.

73) Le Bihan D, Breton E, Lallemand D, Aubin ML, Vi-gnaud J, Laval-Jeantet M. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988; 168: 497-505.

74) Le Bihan D, Breton E, Lallemand D, Grenier P, Cabanis E, Laval-Jeantet M. MR imaging of intra-voxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 1986; 161: 401-407.

75) Jensen JH and Helpern JA. MRI quantification of non-Gaussian water diffusion by kurtosis analy-sis. NMR Biomed 2010; 23: 698-710

76) Lee SS, Byun JH, Park BJ, Park SH, Kim N, Park B, Kim JK, Lee MG. Quantitative analysis of dif-fusion-weighted magnetic resonance imaging of the pancreas: usefulness in characterizing solid pancreatic masses. J Magn Reson Imaging 2008; 28: 928-936.

77) Ma C, Guo X, Liu L, Zhan Q, Li J, Zhu C, Wang L, Zhang J, Fang X, Qu J, Chen S, Shao C, Lu JP. Effect of region of interest size on ADC measure-ments in pancreatic adenocarcinoma. Cancer Im-aging 2017; 17: 13.

78) Muraoka N, Uematsu H, Kimura H, Imamura Y, Fujiwara Y, Murakami M, Yamaguchi A, Itoh H. Apparent diffusion coefficient in pancreatic can-cer: characterization and histopathological cor-relations. J Magn Reson Imaging 2008; 27: 1302-1308.

79) Ma C, Liu L, Li J, Wang L, Chen LG, Zhang Y, Chen SY, Lu JP. Apparent diffusion coefficient (ADC) measurements in pancreatic adenocarci-noma: a preliminary study of the effect of region of interest on ADC values and interobserver vari-ability. J Magn Reson Imaging 2016; 43: 407-413.

80) Fattahi R, Balci NC, Perman WH, Hsueh EC, Alkaade S, Havlioglu N, Burton FR. Pancreat-ic diffusion-weighted imaging (DWI): comparison between mass-forming focal pancreatitis (FP), pancreatic cancer (PC), and normal pancreas. J Magn Reson imaging 2009; 29: 350Y356.

81) Erturk SM, Ichikawa T, Motosugi U, Sou H, Araki T. Diffusion-weighted MR imaging in the evalua-tion of pancreatic exocrine function before and af-ter secretin stimulation. Am J Gastroenterol 2006; 101: 133Y136.

82) Lemke A, Laun FB, Klauss M, Re TJ, Simon D, Delorme S, Schad LR, Stieltjes B. Differentiation of pancreas carcinoma from healthy pancreatic tissue using multiple b-values: comparison of ap-parent diffusion coefficient and intravoxel inco-herent motion derived parameters. Invest Radiol 2009; 44: 769-775.

83) Klauss M, Lemke A, Grünberg K, Simon D, Re TJ, Wente MN, Laun FB, Kauczor HU, Delorme S, Grenacher L, Stieltjes B. Intravoxel incoherent

motion MRI for the differentiation between mass forming chronic pancreatitis and pancreatic carci-noma. Invest Radiol 2011; 46: 57-63.

84) Balci NC, Perman WH, Saglam S, Akisik F, Fatta-hi R, Bilgin M. Diffusion-weighted magnetic reso-nance imaging of the pancreas. Top Magn Reson Imaging 2009; 20: 43-47.

85) Granata V, Fusco R, Avallone A, De Stefano A, Ottaiano A, Sbordone C, Brunese L, Izzo F, Petril-lo A. Radiomics-derived data by contrast en-hanced magnetic resonance in RAS mutations detection in colorectal liver metastases. Cancers (Basel) 2021; 13: 453.

86) Park S, Chu LC, Hruban RH, Vogelstein B, Kinzler KW, Yuille AL, Fouladi DF, Shayesteh S, Ghandi-li S, Wolfgang CL, Burkhart R, He J, Fishman EK, Kawamoto S. Differentiating autoimmune pan-creatitis from pancreatic ductal adenocarcinoma with CT radiomics features. Diagn Interv Imaging 2020; 101: 555-564.

87) Chu LC, Park S, Kawamoto S, Fouladi DF, Shay-esteh S, Zinreich ES, Graves JS, Horton KM, Hruban RH, Yuille AL, Kinzler KW, Vogelstein B, Fishman EK. Utility of CT radiomics features in differentiation of pancreatic ductal adenocarci-noma from normal pancreatic tissue. AJR Am J Roentgenol 2019; 213: 349-357.

88) Eilaghi A, Baig S, Zhang Y, Zhang J, Karanico-las P, Gallinger S, Khalvati F, Haider MA. CT tex-ture features are associated with overall survival in pancreatic ductal adenocarcinoma – a quanti-tative analysis. BMC Med Imaging 2017; 17: 38.

89) Chen X, Oshima K, Schott D, Wu H, Hall W, Song Y, Tao Y, Li D, Zheng C, Knechtges P, Erickson B, Li XA. Assessment of treatment response during chemoradiation therapy for pancreatic cancer based on quantitative radiomic analysis of dai-ly CTs: an exploratory study. PLoS One 2017; 12: e0178961.

90) Attiyeh MA, Chakraborty J, Doussot A, Lang-don-Embry L, Mainarich S, Gönen M, Balachan-dran VP, D’Angelica MI, DeMatteo RP, Jarnagin WR, Kingham TP, Allen PJ, Simpson AL, Do RK. Survival prediction in pancreatic ductal adenocarci-noma by quantitative computed tomography image analysis. Ann Surg Oncol 2018; 25: 1034-1042.

91) Khalvati F, Zhang Y, Baig S, Lobo-Mueller EM, Karanicolas P, Gallinger S, Haider MA. Prog-nostic value of CT radiomic features in resect-able pancreatic ductal adenocarcinoma. Sci Rep 2019; 9: 5449.

92) Zhang Y, Lobo-Mueller EM, Karanicolas P, Gall-inger S, Haider MA, Khalvati F. Improving prog-nostic performance in resectable pancreatic duc-tal adenocarcinoma using radiomics and deep learning features fusion in CT images. Sci Rep 2021; 11: 1378.

93) Zhang Y, Lobo-Mueller EM, Karanicolas P, Gall-inger S, Haider MA, Khalvati F. CNN-based sur-vival model for pancreatic ductal adenocarcino-ma in medical imaging. BMC Med Imaging 2020; 20: 11.


3698

94) Iwatate Y, Hoshino I, Yokota H, Ishige F, Itami M, Mori Y, Chiba S, Arimitsu H, Yanagibashi H, Na-gase H, Takayama W. Radiogenomics for pre-dicting p53 status, PD-L1 expression, and prog-nosis with machine learning in pancreatic can-cer. Br J Cancer 2020; 123: 1253-1261.

95) Thomas JV, Abou Elkassem AM, Ganeshan B, Smith AD. MR Imaging texture analysis in the abdomen and pelvis. Magn Reson Imaging Clin N Am 2020; 28: 447-456.

96) Choi MH, Lee YJ, Yoon SB, Choi JI, Jung SE, Rha SE. MRI of pancreatic ductal adenocarci-noma: texture analysis of T2- weighted images for predicting long-term outcome. Abdom Radi-ol (NY) 2019; 44: 122-130.

97) Ierardi AM, Lucchina N, Petrillo M, Floridi C, Pi-acentino F, Bacuzzi A, Fonio P, Fontana F, Fu-gazzola C, Brunese L, Carrafiello G. Systematic review of minimally invasive ablation treatment for locally advanced pancreatic cancer. Radiol Med 2014; 119: 483-498.

98) Hu S, Zhang J, Zuo C, Cheng C, Liu Q, Sun G. (18)F-FDG-PET/CT findings in pancreatic me-tastasis. Radiol Med 2015; 120: 887-898.

99) Liu Y, Wang Y, Tang W, Jiang M, Li K, Tao X. Multiparametric MR imaging detects therapy efficacy of radioactive seeds brachytherapy in pancreatic ductal adenocarcinoma xenografts. Radiol Med 2018; 123: 481-488.

100) Yao X, Zeng M, Wang H, Fei S, Rao S, Ji Y. Me-tabolite detection of pancreatic carcinoma by in vivo proton MR spectroscopy at 3T: initial re-sults. Radiol Med 2012; 117: 780-788.

101) Angelelli G, Mancini M, Pignataro P, Pedote P, Scardapane A. Multidetector computed tomog-raphy in the study of pancreatic metastases. Ra-diol Med 2012; 117: 369-377.

102) Hu S, Huang W, Lin X, Wang Y, Chen KM, Chai W. Solid pseudopapillary tumour of the pancre-as: distinct patterns of computed tomography manifestation for male versus female patients. Radiol Med 2014; 119: 83-9.

103) Ferrari R, De Cecco CN, Iafrate F, Paolanto-nio P, Rengo M, Laghi A. Anatomical variations of the coeliac trunk and the mesenteric arteries evaluated with 64-row CT angiography. Radiol Med 2007; 112: 988-998.

104) Benedetti G, Mori M, Panzeri MM, Barbera M, Palumbo D, Sini C, Muffatti F, Andreasi V, Steidler S, Doglioni C, Partelli S, Manzoni M, Falconi M, Fiorino C, De Cobelli F. CT-de-rived radiomic features to discriminate histolog-ic characteristics of pancreatic neuroendocrine tumors. Radiol Med 2021; doi: 10.1007/s11547-021-01333-z

105) Reginelli A, Capasso R, Petrillo M, Rossi C, Faella P, Grassi R, Belfiore MP, Rossi G, Mu-to M, Muto P, Fiorello A, Serra N, Nizzoli R, De Filippo M, Cappabianca S, Carrafiello G, Brunese L, Rotondo A. Looking for lepidic com-ponent inside invasive adenocarcinomas ap-pearing as CT solid solitary pulmonary nodules

(SPNs): CT morpho-densitometric features and 18-FDG PET findings. Biomed Res Int 2019; 2019: 7683648.

106) Higashi M, Tanabe M, Okada M, Furukawa M, Iida E, Ito K. Influence of fat deposition on T1 mapping of the pancreas: evaluation by du-al-flip-angle MR imaging with and without fat suppression. Radiol Med 2020; 125: 1-6.

107) Sun J, Yang L, Zhou Z, Zhang D, Han W, Zhang Q, Peng Y. Performance evaluation of two itera-tive reconstruction algorithms, MBIR and ASIR, in low radiation dose and low contrast dose ab-dominal CT in children. Radiol Med 2020; 125: 918-925.

108) Ierardi AM, Lucchina N, Petrillo M, Floridi C, Pi-acentino F, Bacuzzi A, Fonio P, Fontana F, Fu-gazzola C, Brunese L, Carrafiello G. Systematic review of minimally invasive ablation treatment for locally advanced pancreatic cancer. Radiol Med 2014; 119: 483-498.

109) Hu S, Zhang J, Zuo C, Cheng C, Liu Q, Sun G. (18)F-FDG-PET/CT findings in pancreatic me-tastasis. Radiol Med 2015; 120: 887-898.

110) Liu Y, Wang Y, Tang W, Jiang M, Li K, Tao X. Multiparametric MR imaging detects therapy efficacy of radioactive seeds brachytherapy in pancreatic ductal adenocarcinoma xenografts. Radiol Med 2018; 123: 481-488.

111) Yamaga LY, Neto GC, da Cunha ML, Osawa A, Oliveira JC, Fonseca RQ, Nogueira SA, Wagner J, Funari MG. 99mTc-HYNIC-TOC increased up-take can mimic malignancy in the pancreas un-cinate process at somatostatin receptor SPECT/CT. Radiol Med 2016;121: 225-228.

112) Zhu H, Ying L, Tang W, Yang X, Sun B. Can MDCT or EUS features predict the histopatho-logical grading scheme of pancreatic neuroen-docrine neoplasms? Radiol Med 2017; 122: 319-326.

113) Colagrande S, Carbone SF, Carusi LM, Cova M, Villari N. Magnetic resonance diffusion-weight-ed imaging: extraneurological applications. Ra-diol Med 2006; 111: 392-419.

114) Yao X, Zeng M, Wang H, Fei S, Rao S, Ji Y. Me-tabolite detection of pancreatic carcinoma by in vivo proton MR spectroscopy at 3T: initial re-sults. Radiol Med 2012; 117: 780-788.

115) Hu S, Lin X, Song Q, Chen K. Solid pseudopap-illary tumour of the pancreas in children: clinical and computed tomography manifestation. Radi-ol Med 2012; 117: 1242-1249.

116) Graziani R, Frulloni L, Cicero C, Manfredi R, Am-brosetti MC, Mautone S, Pozzi Mucelli R. Bull’s-eye pattern of pancreatic-duct stones on multi-detector computed tomography and gene-mu-tation-associated pancreatitis (GMAP). Radiol Med 2012; 117: 1275-1286.

117) Camera L, Paoletta S, Mollica C, Milone F, Na-politano V, De Luca L, Faggiano A, Colao A, Salvatore M. Screening of pancreaticoduodenal endocrine tumours in patients with MEN 1: mul-


3699

tidetector-row computed tomography vs. endo-scopic ultrasound. Radiol Med 2011; 116: 595-606.

118) Hu S, Huang W, Lin X, Wang Y, Chen KM, Chai W. Solid pseudopapillary tumour of the pancre-as: distinct patterns of computed tomography manifestation for male versus female patients. Radiol Med 2014; 119: 83-89.

119) Liu Y, Xu XQ, Lin XZ, Song Q, Chen KM. Pro-spective study comparing two iodine concentra-tions for multidetector computed tomography of the pancreas. Radiol Med 2010; 115: 898-905.

120) Maurea S, Caleo O, Mollica C, Imbriaco M, Mainenti PP, Palumbo C, Mancini M, Camera L, Salvatore M. Comparative diagnostic evaluation with MR cholangiopancreatography, ultrasonog-raphy and CT in patients with pancreatobiliary disease. Radiol Med 2009; 114: 390-402.

121) Castelli F, Bosetti D, Negrelli R, Di Paola V, Zantedeschi L, Ventriglia A, Manfredi R, Mucel-li RP. Multifocal branch-duct intraductal papillary mucinous neoplasms (IPMNs) of the pancreas: magnetic resonance (MR) imaging pattern and evolution over time. Radiol Med 2013; 118: 917-929.

122) Calculli L, Pezzilli R, Brindisi C, Morabito R, Casadei R, Zompatori M. Pancreatic and extra-pancreatic lesions in patients with intraductal papillary mucinous neoplasms of the pancreas: a single-centre experience. Radiol Med 2010; 115: 442-452.

123) Ferrari R, De Cecco CN, Iafrate F, Paolanto-nio P, Rengo M, Laghi A. Anatomical variations of the coeliac trunk and the mesenteric arteries evaluated with 64-row CT angiography. Radiol Med 2007; 112: 988-998.

124) Manfredi R, Bonatti M, D’Onofrio M, Mehra-bi S, Salvia R, Mantovani W, Pozzi Mucelli R. Incidentally discovered benign pancreatic cys-tic neoplasms not communicating with the duc-tal system: MR/MRCP imaging appearance and evolution. Radiol Med 2013; 118: 163-180.

125) Gusmini S, Nicoletti R, Martinenghi C, Cabor-ni C, Balzano G, Zerbi A, Rocchetti SI, Arcid-iacono PG, Albarello L, De Cobelli F, Di Carlo V, Del Maschio A. Arterial vs pancreatic phase: which is the best choice in the evaluation of pancreatic endocrine tumours with multidetec-tor computed tomography (MDCT)? Radiol Med 2007; 112: 999-1012.

126) Graziani R, Manfredi R, Cicero C, Contro A, Brandalise A, Tapparelli M, Frulloni L, Vantini I, Megibow A, Pozzi Mucelli R. Role of multis-lice computed tomography in the diagnosis of gene-mutation-associated pancreatitis (GMAP). Radiol Med 2010; 115: 875-888.

127) Nardone V, Reginelli A, Guida C, Belfiore MP, Biondi M, Mormile M, Banci Buonamici F, Di Gi-orgio E, Spadafora M, Tini P, Grassi R, Pirtoli L, Correale P, Cappabianca S, Grassi R. Delta-ra-diomics increases multicentre reproducibility: a phantom study. Med Oncol 2020; 37: 38.

pancreatic cancer detection and characterization: state of the art … · 2021. 5. 28. ·...

Documents