Correlation of pathological types and imaging features in pancreatic cancer

Abstract

The study by Luo et al published in the World Journal of Gastrointestinal Oncology presents a thorough and scientific methodology. Pancreatic cancer is the most challenging malignancy in the digestive system, exhibiting one of the highest mortality rates associated with cancer globally. The delayed onset of symptoms and diagnosis often results in metastasis or local progression of the cancer, thereby constraining treatment options and outcomes. For these patients, prompt tumour identification and treatment strategising are crucial. The present objective of pancreatic cancer research is to examine the correlation between various pathological types and imaging data to facilitate therapeutic decision-making. This study aims to clarify the correlation between diverse pathological markers and imaging in pancreatic cancer patients, with prospective longitudinal studies potentially providing novel insights into the diagnosis and treatment of pancreatic cancer.

Key Words: Pancreatic cancer; Pathological types; Imaging features; Association; Non-invasive tests

Core Tip: This study validates and extends previous findings regarding the imaging characteristics of pancreatic cancer. In a larger cohort, comprehensive imaging assessments using diverse modalities and rigorous statistical analysis facilitate more precise quantification of these features. It aids in the development of more accurate image-based diagnostic algorithms for pancreatic cancer, which can guide treatment decisions and enhance screening and monitoring protocols for high-risk patients.

TO THE EDITOR

We are gratified by the publication of Luo et al’s high-caliber article in the World Journal of Gastrointestinal Oncology, which addresses clinical issues pertaining to the relationship between pathological types and imaging characteristics of pancreatic cancer, thereby informing future clinical treatment strategies[1]. This study included 500 eligible pancreatic cancer patients from a single center over the past decade, with pathological samples collected via surgical resection or image-guided biopsy. Meanwhile, computed tomography (CT), magnetic resonance imaging (MRI), and endoscopic ultrasound (EUS) were employed to meticulously examine the imaging characteristics of patients and to delineate the imaging features associated with various pathological types. Statistical analysis was conducted on the baseline data of all patients. Multinomial logistic regression analysis is employed to illustrate various imaging characteristics significantly correlated with a particular pathological type. Hypo enhancement [odds ratio (OR) = 22.1, 95% confidence interval (CI): 13.5-36.2] and indistinct margins (OR = 15.3, 95%CI: 9.2-25.4) exhibited a strong correlation with pancreatic ductal adenocarcinoma (PDAC). Intraductal papillary mucinous neoplasms (IPMN) was strongly associated with cystic lesions that connected to the pancreatic duct (OR = 185.7, 95%CI: 78.3-440.2). Neuroendocrine tumors (NET) exhibited a significant correlation with both well-defined borders (OR = 12.4, 95%CI: 6.7-22.9) and arterial phase hyperenhancement (OR = 28.6, 95%CI: 14.9-54.8). The Cohen’s kappa coefficient for categorical variables and the intraclass correlation coefficient for continuous variables indicated substantial agreement among observers regarding the imaging assessment, particularly concerning tumor size and vascular invasion. The thorough analysis offers significant insights, and this study may enhance early diagnosis, inform personalized treatment strategies, and ultimately improve outcomes for patients with pancreatic cancer by refining our capacity to characterize pancreatic lesions.

Pancreatic cancer ranks as the seventh foremost cause of cancer-related mortality worldwide. Although its prevalence differs significantly among various countries, global patterns suggest an upward trajectory, with pancreatic cancer increasingly emerging as the second foremost cause of cancer mortality in Western nations. Unfortunately, the majority of patients exhibit advanced-stage disease, leading to diminished survival rates and restricted treatment alternatives. The pancreas, a crucial organ for secretion, can develop various tumor types. PDAC is the predominant type, constituting roughly 90% of all pancreatic cancers. Less prevalent types, including IPMN and NET, account for the remaining instances[2-4]. Historically, biomarker screening and diverse pathological biopsy methods have been used to ascertain the etiology of pancreatic cancer. Furthermore, imaging techniques such as CT, MRI, and EUS have been employed for the diagnosis and staging of pancreatic lesions. These instruments enable the detection of pancreatic cancer lesions and assist in directing targeted diagnostic and therapeutic approaches[5-7]. However, owing to nuanced anatomical and pathological distinctions among pancreatic lesions, existing research is deficient in thorough examinations of the correlation between pancreatic cancer subtypes and non-invasive diagnostic methods, thereby impeding enhancements in diagnostic precision and the advancement of early detection[8].

SITUATION ON IMAGING FEARURES OF PANCREATIC CANCER

CT is widely used in clinical environments and provides numerous benefits, including reduced radiation exposure relative to alternative imaging modalities and the mitigation of risks linked to contrast agents. CT imaging, being the most frequently utilized modality, exhibits characteristics that are closely linked to different pathological types of pancreatic cancer. Recent studies indicate that PDAC typically appears as a hypodense mass with irregular margins on CT imaging, while IPMNs frequently present as cystic lesions with mural nodules. In addition, NETs often demonstrate hypervascular enhancement on contrast-enhanced imaging[9-11]. Advancements in medical technology have rendered multi-slice spiral CT (MDCT) and MRI increasingly vital for the detection of pancreatic cancer, especially in instances exhibiting atypical imaging characteristics. The concurrent application of CT and MRI has shown better diagnostic effectiveness in identifying pancreatic masses and malignancies. MRI provides enhanced soft tissue resolution and diverse imaging modalities, rendering it particularly effective for identifying liver metastases, potentially surpassing MDCT in sensitivity[12,13]. EUS is a non-invasive modality that provides high sensitivity for identifying small lesions and facilitates histological sample acquisition. EUS is especially effective in staging pancreatic cancer, offering high specificity for lymph node involvement and the extent of vascular invasion. Nonetheless, EUS is significantly dependent on the operator’s expertise and is not accessible in all areas[14,15]. Moreover, positron emission tomography (PET)/CT has proven effective in assessing the response of pancreatic tumors to chemotherapy and radiotherapy, functioning as a complement to other imaging techniques[16,17]. Contemporary multimodal imaging techniques, CT, MRI, and EUS, offer significant diagnostic insights into pancreatic cancer. Comprehending the unique imaging features of various pancreatic cancer subtypes is essential for distinguishing pancreatic malignancies from other pancreatic lesions, thus enabling precise diagnosis and suitable treatment planning.

FUTURE PROSPECT: BETWEEN PATHOLOGICAL TYPES AND IMAGING FEATURES

PDAC, the predominant type of pancreatic cancer, is typically distinguished by indistinct margins and hypoenhancement throughout all imaging phases. Previous research has validated these imaging characteristics[1], which were additionally corroborated by our study. Our study revealed that the predominant imaging features of PDAC were a hypodense mass on CT and hypointensity on T1-weighted MRI. These characteristics are likely ascribed to stromal desmoplasia in PDAC, leading to hypoenhancement. PDAC lesions generally range from 4 cm to 5 cm in size, predominantly situated in the head or neck of the pancreas, with a lesser percentage located in the body or tail. Initial indicators of PDAC, including parenchymal heterogeneity and the depletion of normal adipose tissue surrounding the pancreas, were observed in certain patients. Additionally, quantitative imaging characteristics such as intensity, shape, size, volume, and texture can be derived through imaging omics, providing enhanced diagnostic value[18]. In addition, the imaging features of IPMN predominantly consist of cystic lesions that connect with the pancreatic duct, a characteristic that aids in differentiating IPMNs from other intricate cystic pancreatic lesions. The prevalent imaging characteristics of NET encompass arterial-phase hyperenhancement and distinct margins, corroborating prior research and reinforcing the correlation between these imaging traits and NET[19,20].

This study presented novel insights, particularly concerning the “double-duct sign”, which is more frequently observed in PDAC than in inflammatory conditions and is not exclusive to PDAC. Cancer-induced stenosis generally presents as a sudden obstruction of the common bile duct, along with a corresponding sudden narrowing of the main pancreatic duct at the same level. Our results highlight the significant pancreatic duct dilation in PDAC, reinforcing the “double-duct sign” as a possible marker for pancreatic tumors[21,22]. Mural nodules measuring ≥ 5 mm were identified as a risk factor for IPMN based on CT/MRI imaging. In certain instances, 20% of NETs may display varying levels of calcification, potentially linked to local vascular infiltration resulting in dystrophic calcification. These calcifications are typically situated at the core of the mass and manifest as irregularly shaped deposits. Nonetheless, additional research is necessary to ascertain whether these calcifications can consistently distinguish NETs from other tumor types or inflammatory conditions[23,24]. In our study, wall nodules were identified in 60% of IPMN cases, and calcification was noted in 25% of NET cases, both of which were more prevalent than in prior reports. The potential of these features as supplementary diagnostic markers in instances of inconclusive radiographic findings requires further examination.

This study offers valuable insights, although it has limitations. Additional multicenter prospective studies utilizing more standardized imaging protocols and pathological assessments are required to validate the identified correlations between imaging characteristics and pathological types. Furthermore, independent external cohorts are essential to corroborate these associations. The imaging parameters employed in this study were suboptimal due to technical constraints, and the subjectivity intrinsic to statistical methods for assessing imaging characteristics presents an additional challenge.

This study revealed significant correlations between particular imaging characteristics and pancreatic cancer subtypes, potentially aiding in longitudinal dynamic monitoring and the creation of image-based diagnostic models for pancreatic cancer. A diagnostic algorithm integrating various imaging characteristics, such as “cystic lesions + mural nodules-IPMN”, may be proposed as an innovative diagnostic strategy based on these data. In addition, the capacity to forecast pathological subtypes through imaging characteristics could guide treatment choices. However, additional research is necessary to ascertain if these imaging characteristics can inform targeted therapies, and precise survival and outcome data are essential to validate their clinical applicability. Progress in artificial intelligence may improve imaging feature extraction, providing opportunities for extensive screening of asymptomatic populations. Nonetheless, obstacles such as dataset heterogeneity and variability in clinical practice persist[9]. Future research should focus on establishing quantitative correlations between imaging characteristics and clinical outcomes, thereby advancing precision medicine for pancreatic cancer.

CONCLUSION

This study analyzed the correlation between pathological classifications and imaging characteristics of pancreatic cancer, yielding consistent and significant insights to enhance non-invasive screening for pancreatic disorders. However, prospective studies are required to further substantiate the findings and assess their clinical implications. Future research could integrate artificial intelligence technology into longitudinal dynamic tracking studies to monitor alterations in imaging features, offering novel approaches for early diagnosis, treatment strategy development, and enhancement of patient outcomes.