Pancreatic ductal adenocarcinoma (PDAC) has a poor prognosis, with a 5-year survival rate of less than 5% (1, 2) and an 80% recurrence rate after surgery (1); therefore, is considered an extremely lethal disease. The microenvironment of PDAC is closely related to its dismal prognosis, and inflammation is an essential aspect of the tumor (3). It has been demonstrated that inflammation plays a key role not only in the development of PDAC but also in determining patient prognosis (4). Therefore, understanding the inflammatory activities within the microenvironment of primary PDAC can provide meaningful information for its diagnosis, treatment, and prognosis (5). The current gold standard for the detection of inflammatory activity in PDAC remains histological evaluation with biopsy. However, this procedure has some drawbacks such as the invasiveness of the sampling, the difficulty in performing the puncture, the small size of the extracted specimen, and its contraindications, all of which compromise its clinical application to some extent.

In contrast, imaging modalities such as computed tomography (CT), magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) are noninvasive and convenient, and the inflammatory state within PDAC can be evaluated by measuring certain imaging features. For instance, correlations between pancreatic inflammatory lesions and the degree of delayed intensification were reported by Ren and Axsom (6, 7). Especially in recent years, a massive number of quantitative features can be extracted from radiological images with a variety of algorithms, and radiomics and texture analysis are expected to be feasible techniques for assessing inflammatory infiltration within PDAC (8, 9). However, to the best of our knowledge, no study has been conducted on imaging evaluation of inflammatory infiltration in PDAC. In this study, we investigated the value of traditional CT imaging by extracting texture features for the assessment of inflammatory infiltration in PDAC with the aim of providing more accurate information for the precise diagnosis and treatment of this disease.

In recent years, as an inexpensive non-invasive imaging technology, CT has become the mainstay for the diagnosis and evaluation of PDAC (11). The tumor microenvironment has a profound influence on the metabolism and biological behaviors of the tumor itself (12), especially for PDAC, and the current diagnosis and treatment challenges of PDAC are closely related to its particular microenvironment (13). With the rapid development of CT technology, especially when combined with quantitative imaging techniques such as texture analysis, the assessment of the tumor microenvironment using non-invasive imaging technology has become possible (14).

In 1863, Virchow first discovered the presence of immune cell infiltration in neoplastic tissues (15), and it is now generally accepted that chronic aseptic inflammation of the pancreas is an important risk factor for the development of PDAC, along with its initiation, progression, and metastasis (16). As an essential part of the tumor microenvironment, chronic aseptic inflammation disturbs the microenvironment and metabolism of normal pancreatic tissue by accelerating intracellular DNA damage, altering the activity of tumor-related signaling pathways, and inducing the production of inflammatory factors, thus creating favorable conditions for the malignant transformation of normal cells and the proliferation of tumor cells (3). Incio et al. demonstrated that chronic pancreatic inflammation and fibrous tissue proliferation triggered by obesity not only accelerated the progression of PDAC, but was also associated with chemotherapy resistance (17); Aziz et al. measured inflammation-related indicators such as C-reactive protein, the neutrophil-to-lymphocyte ratio, and the platelet-to-lymphocyte ratio in PDAC patients and found that the systemic-immune-inflammation index was an independent predictor of prognosis in PDAC patients (18). Furthermore, Pu et al. found that baicalein, which has anti-inflammatory effects, can slow down or even reverse inflammation-induced pancreatic acinar-to-ductal metaplasia, offering a new idea in effective treatment of PDAC in the clinic (19).

Thus, understanding inflammation in PDAC can provide important information for the diagnosis and evaluation of this lethal disease. However, at present, the assessment of inflammatory activities in tumors is mainly dependent on pathology (20). As the gold standard, pathological assessment has high accuracy, but is also traumatic, expensive, and, in certain circumstances, contraindicated. In clinical practice, on the other hand, the imaging modalities such as contrast-enhanced CT are simple, non-invasive, and repeatable. Therefore, in this study, we used pathological findings as a reference standard, to evaluate the value of contrast-enhanced CT in assessment of inflammatory infiltration in PDAC.

The present study revealed statistically significant correlations between cystic necrosis, vascular invasion, ascites and the pathological differentiation degrees in PDAC, matching the observations by Lee, Mierke, Ariake et al. (21–23); however, correlations between the aforementioned imaging features and inflammatory infiltration were not significant, indicating that conventional imaging might not be sufficient to assess the differentiation degree of PDAC. Regarding the relationship between PDAC tumor enhancement and the degree of inflammatory infiltration, this study demonstrated significant inverse correlations in the arterial and venous phases. The inverse correlations we found in this study may be explained by the fact that chronic inflammation, especially in the pancreas, raises the degree of tumor fibrosis by upregulating epithelial–mesenchymal transition (EMT) (24, 25). The fibrous tissue obstructs and delays the entry of contrast agent, resulting in less enhancement. Hattori et al. also found that the degree of fibrosis within PDAC tumors was inversely correlated with tumor enhancement on contrast-enhanced CT scans, and the correlation was more significant in the arterial and venous phases (26). However, we found that the AUCs for tumor enhancement in the arterial and venous phases were only 0.542 and 0.570, indicating that although tumor enhancement was correlated with inflammatory infiltration, it could not be used to predict inflammatory infiltration. This may be due to the limitation of naked eye observation and evaluation of CT images. In contrast, the rapid development of texture analysis in recent years can overcome limitations and extract a massive amount of quantitative information directly from image data, leading to more desirable predictive performance (8, 9). Moreover, in a retrospective study, Toft et al. demonstrated that MRI has higher accuracy, sensitivity, and specificity than CT in the diagnosis of pancreatic cancer (27), which is due to the higher soft tissue resolution of MRI. This suggests that the correlations and predictive performance of MRI scanning with indicators related to the tumor inflammatory microenvironment can be further explored in future studies.

Since the predictive value of conventional CT imaging is limited, we continued to investigate the value of contrast-enhanced CT images in assessing inflammatory infiltration within PDAC tumors using texture analysis; the Variance Threshold and SelectKBest algorithms were used to select the best texture features for inflammatory infiltration at each of contrast-enhanced CT scan phase, and logistic regression equations were established using the screened features. Comparing the ROC curves of the regression equations obtained for each phase, we found that the equation based on the optimal features from the arterial phase had the highest AUC among the phases (0.982), as well as the highest sensitivity (100%) and specificity (89.5%), demonstrating a promising predictive value; the delayed phase had the second highest AUC of 0.849, showing moderate predictive performance; and the venous phase had the lowest AUC of 0.643, showing poor predictive performance. Therefore, the arterial phase may be the optimal phase for performing texture analysis in our study, which may be related to the fact that the arterial phase magnifies the difference between tumor and normal tissue (28–30). Based on this, the inflammatory infiltration degree of PDAC can be assessed more accurately with data from this phase.

Our study had certain limitations. First, the sample size was relatively small, which may lead to bias in the results; Second, the ROIs were outlined only on images containing the maximum cross- sections of the tumor, and therefore we were unable to perform 3D whole-tumor analysis. Finally, we did not correlate imaging features with clinical symptoms, biochemical indicators, and prognostic indices.

In summary, conventional CT imaging analysis has some limitations in predicting inflammatory infiltration within PDAC tumors, while texture analysis based on contrast-enhanced CT images has better predictive performance for inflammatory infiltration, especially in the arterial phase.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee on Scientific Research of Shandong University Qilu Hospital. The ethics committee waived the requirement of written informed consent for participation.

DY contributed to conception and design of the study. FW and HG organized the database. JX performed the statistical analysis. HG wrote the first draft of the manuscript. FW wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.