Diagnostic value of endoscopic ultrasound in staging of pancreatic cancer

Abstract

Pancreatic cancer is recognized as one of the leading causes of cancer mortality, representing the second most common source of cancer-related deaths within the gastrointestinal domain. Surgical resection is currently the only definitive treatment; however, the subtle emergence of symptoms often leads to a diagnosis at an advanced stage, with merely 10%-15% of patients being eligible for surgical intervention. The primary obstacle to achieving a potential radical resection is the presence of distant metastatic disease or invasion of adjacent major vascular structures. This review aims to highlight the critical role of endoscopic ultrasound in the diagnosis and staging of pancreatic tumors. We systematically searched PubMed, MEDLINE and Web of Science by using ‘pancreatic cancer’ and ‘endoscopic ultrasonography’ as keywords. Relevant studies were reviewed and analyzed. Endoscopic ultrasonography (EUS) is efficient in the diagnosis and staging of pancreatic cancer, past studies reported the accuracy of EUS is 63% to 94% for T-staging and 44% to 82% for N-staging but there are still limitations that need to be comprehensively applied with other diagnostic methods to evaluation of distant metastasis for surgical resectability. Our review aims to reveal the value for the staging of pancreatic cancer.

Key Words: Endoscopic ultrasound; Pancreatic cancer; Staging; Diagnostic value

Core Tip: Pancreatic cancer is a lethal health problem. Most pancreatic cancers asymptomatic in the early stages, without noticeable clinical presentation in the early stage. Surgical resection is the only approach for pancreatic cancer. Endoscopic ultrasonography (EUS) provides precise evaluations of tumor, invasion, and lymph node involvement, but a singular imaging modality appears inadequate for the preoperative staging of pancreatic cancer. Our review aims to reveal the diagnosis and staging of EUS in pancreatic cancer.

INTRODUCTION

Pancreatic cancer is a lethal health problem, and its global burden has increased steeply in the past years[1,2]. Pancreatic cancer ranks seventh among the causes of cancer-related deaths. It is among the leading contributors to cancer mortality and holds the second position in the of fatalities associated with digestive system malignancies according to statistics[3]. Most pancreatic cancers are pathologically pancreatic ductal adenocarcinomas, which is asymptomatic in the early stages, without noticeable clinical presentation in the early stage, however pancreatic malignant tumors were diagnosed in the late stage of the disease when symptoms appear. The early diagnosis is less than 20%. Most patients are in a progressive stage at the time of diagnosis and have a poor prognosis[4]. Timely surgical resection is the only approach for pancreatic cancer. However, most patients are asymptomatic in the early stages, and only 15% of patients are eligible for surgery[5,6]. The global 5-year survival rate for pancreatic cancer is approximately 9%-12%[7].

According to recent tumor node metastases (TNM) criteria (Tables 1 and 2) and international guideline[8], the TNM staging has a direct impact on surgical resectability. The TNM staging system provides a key basis for therapeutic decision-making and prognostic assessment. Early-stage patients with radical surgery have better prognosis, whereas advanced patients usually require neoadjuvant or palliative treatment due to vascular invasion or metastasis[9], which means that the cancer has grown into nearby large blood vessels or distant organ. Accurate evaluation of resectable pancreatic cancer is essential for determining surgical indications and predicting prognosis, further emphasizing the importance of TNM staging[10]. Generally, endoscopic ultrasonography (EUS) provides precise evaluations of tumor, invasion, and lymph node involvement, but a singular imaging modality appears inadequate for the preoperative staging of pancreatic cancer. Our review aims to reveal the diagnosis and staging of EUS in pancreatic cancer.

Table 1 Pancreatic cancer tumor node metastases staging criteria.

Stage
T1	Tumor ≤ 2 cm in greatest dimension
T1a	Tumor ≤ 0.5 cm in greatest dimension
T1b	Tumor > 0.5 and < 1 cm in greatest dimension
T1c	Tumor 1 to 2 cm in greatest dimension
T2	Tumor > 2 and ≤ 4 cm in greatest dimension
T3	Tumor > 4 cm in greatest dimension
T4	Tumor involves the celiac axis, superior mesenteric artery, and/or common hepatic artery
N1	Metastasis in one to three regional lymph nodes
N2	Metastasis in four or more regional lymph nodes
M0	No distant metastasis
M1	Distant metastasis

Table 2 Prognostic staging groups 8^th edition.

T	N	M	Stage
T1	N0	M0	IA
T2	N0	M0	IB
T3	N0	M0	IIA
T1-3	N1	M0	IIB
T1-3	N2	M0	III
T4	Any N	M0	III
Any T	Any N	M1	IV

EUS IMAGING FEATURES OF PANCREATIC CANCER

EUS plays an important role in the diagnosis of pancreatic mass, particularly for small masses. Pancreatic cancer usually presents as a hypoechoic mass relative to the surrounding pancreatic parenchyma with irregular margins or poorly defined contours[11] (Figure 1A). It often shows a hypovascular pattern on contrast-enhanced EUS (Figure 1B)[12,13]. EUS-elastography (EUS-EG) can assesses tissue stiffness, which is a critical parameter in identifying malignant. Pancreatic cancer typically presents as areas of intense blue coloration on EUS-EG, indicating high stiffness. In contrast, the benign lesions were shown in a mixed coloration pattern with green to red on elastography[14] (Figure 1C). Eventually the pancreatic cancer was histopathologically confirmed with EUS-fine-needle aspiration (FNA; Figure 1D).

Figure 1 The imaging features of pancreatic cancer. A: Pancreatic cancer was showed as a hypoechoic mass on endoscopic ultrasonography (EUS); B: Pancreatic cancer was showed as hypovascular pattern on contrast-enhanced EUS; C: Pancreatic cancer was showed as areas of intense blue coloration on EUS-elastograph; D: The pancreatic cancer was diagnosed histopathologically with fine-needle aspiration.

EUS FOR STAGING

The staging accuracy of EUS reported in different studies varies widely, ranging from 63%-94% for T-staging and 44%-82% for N-staging. A meta-analysis from Li et al[15] reported the pooled sensitivity and specificity of EUS for T1-2 staging was 72% and 90%, and for T3-4 staging was 90% and 72%. This indicated that EUS is particularly more effective in vascular invasion or identifying more advanced tumors(T-staging > T3). But a study showed EUS provides effective staging of pancreatic lesions, especially high-resolution imaging and evaluation of lesions < 1 cm[16]. In patients with small pancreatic masses, the accuracy of EUS in vascular invasion may be lower. Nevertheless, the difference in sensitivity of EUS for identifying vascular invasion is significant[17]. One study also supports that, with a sensitivity of 85% and specificity of 91% for vascular invasion[18]. A meta-analysis including 1308 patients showed the pooled sensitivity was 73%, but the pooled specificity was 90.2% for vascular invasion[17]. Diameter of the pancreas lesions may affect the accuracy of EUS[19]. And the accuracy of EUS was affected by the vascular localization, studies showed that the portal or splenic vein is more accurate than the superior mesenteric artery and vein[20].

The metastasis of lymph node is one of the most important predictors for prognosis of pancreatic cancer[21], and EUS allows direct evaluation and tissue sampling of lymph nodes in the region adjacent to the pancreas, avoiding surgical intervention[16,22]. And a study from Soriano et al[23] demonstrated that EUS had the highest accuracy in N-staging than computed tomography (CT; 65%). And the study from Puli et al[24] showed the sensitivity and accuracy of EUS-FNA for retroperitoneal lymph node samples were 83% and 63%. The sensitivity and specificity of EUS-FNA in diagnosing positive para-aortic lymph nodes were 96.7% and 100%, while the sensitivity of positron emission tomography (PET)-CT was only 53.3%[16]. During the development of pancreatic cancer, tumor-associated lymphatic vessels connect the tumor to the vast lymphatic vascular system. It has been found that lymphatic vessels density (LVD) levels significantly correlate with lymph node metastasis[25]. EUS can access to tumor sample to quantify LVD by microscopic examination[26], and pancreatic cancer patients with higher levels of LVD in their tumors have a worse prognosis after radical resection compared to patients with lower levels[27].

The liver is the most common distant metastatic organ of pancreatic cancer, and the left liver can be explored by EUS through the stomach. If the lesion is appropriate, which would be confirmed by EUS-tissue acquisition (EUS-TA)[28]. CT or PET-CT is an ideal imaging modality for the detection of distant metastases, with a sensitivity of up to 97% for metastatic tumors > 1 cm but decreasing for smaller lesions. In addition, EUS-FNA has a sensitivity of 82%-94% for the diagnosis of malignant ascites or liver metastases[29], but the EUS is limited in distant place and therefore it is not completely reliable in assessing distant metastasis of tumors.

EUS is a highly sensitive and radiation-free modality for detecting pancreatic mass lesions, especially it can acquire tissue for histologic test[30]. EUS-FNA improved the accuracy of diagnosis and staging of pancreatic cancer. EUS-FNA is essential tool for obtaining a definitive cytological or histological diagnosis. EUS-FNA has a sensitivity and specificity of up to 95% and 100%[31], which is considered as a preferred method for diagnosing pancreatic masses pathologically. The EUS and its related techniques improved diagnosis and staging of pancreatic lesions. A meta-analysis including 15 studies revealed the sensitivity of EUS-FNA for pancreatic cancer was 92% and the specificity was 96%[31]. Another meta-analysis including 4766 patients, EUS-FNA diagnosed solid pancreatic mass with a sensitivity of 89% and specificity of 96%[24].

And a study including 109 patients with pancreatic lesions ≤ 10 mm showed diagnostic accuracy of EUS with rapid on-site evaluation was 87.5%, the study identified a Tis-stage patient whose lesion was no longer visible on CT scan before surgery[32]. Additionally, EUS can be complemented by advanced techniques such as contrast-enhanced EUS (CE-EUS) and EUS elastography, which improve diagnostic accuracy and aid in differentiating malignant from benign lesions.

CE-EUS improved the diagnostic rate and accuracy of EUS[33]. CE-EUS is particularly useful in cases where conventional EUS and other imaging modalities are inconclusive, and a study show that the diagnostic rate of pancreatic lesions by CE-EUS has significantly increased from 64% to 91%[34]. A meta-analysis reported pooled sensitivity and specificity of 98% and 63% for qualitative EUS-EG[35]. Moreover, other study reported that EUS-EG had a sensitivity of 91.7% and specificity of 90.0% for diagnosing vascular invasion, which was superior to CT[36].

EUS AND OTHER IMAGING MODALITY

EUS is highly effective for diagnosing and staging of pancreatic cancer, particularly when compared to other diagnostic tools such as CT and magnetic resonance imaging (MRI), but imaging modality should be selected regarding local centers’ diagnosis level[37,38]. EUS is superior for tumor detection and staging, while exhibiting comparable performance in the diagnosis of lymph node and resectability for patients with pancreatic cancer preoperatively[39-41]. A systematic review showed that EUS was more accurate than CT for staging and vascular invasion in earlier research, but similar in later research[42,43].

A retrospective study including 63 patients with surgery showed that the sensitivity of EUS and MRI for tumor resectability was 61% and 73%[44]. EUS is more sensitive to diagnose small pancreatic lesions, nevertheless MRI is more efficient to detect tumor metastasis or vascular invasion[11,45]. Combining EUS and MRI would improve the accuracy of staging. Studies have shown that when the two modalities were identical on resectability, the accuracy is reliably[23,44]. EUS with other additional technology has a sensitivity of 91%-100% and a specificity of up to 100% for detecting pancreatic cancer, which is significantly higher than CT, which has a sensitivity of 53%-91% for pancreatic cancer[18]. EUS provides superior high-resolution images of the pancreas, which is particularly useful for detecting small lesions (< 2 cm) that may be missed by CT or MRI[46].

CONCLUSION

EUS has become an essential modality for preoperative diagnosis and resectability of pancreatic cancer. EUS plays a crucial role in differentiating pancreatic lesions, preoperative staging of cancer, and identifying lymph node metastasis[47]. EUS is particularly important in the diagnosis and management of borderline resectable pancreatic cancer[48]. Since patients with borderline resectable pancreatic cancer are typically treated with neoadjuvant therapy, radical resection may become feasible if the tumor regresses following evaluation by imaging modalities such as EUS. EUS is particularly advantageous in assessing localized tumor extent and treatment response[49,50]. Those patients might still benefit from therapy[51]. EUS, in combination with other diagnostic modality and new technologies, will further enhance its role in diagnosis of pancreatic cancer. EUS-TA is the gold standard for confirming the diagnosis of pancreatic cancer because it can safely, reproducibly, and non-invasively obtain pathological specimens that can only be obtained by laparoscopic or open surgery in the past[52]. But EUS-FNA was recommended to applied to pancreatic carcinoma in situ, for that ERCP-based cytology may be helpful[53,54]. However, the criteria for TNM staging are constantly changing, and those studies we discussed were heterogeneous which were difficult to compare different T staging studies. The data from different studies is difficult due to review in the criteria for exclusion and inclusion. And the diagnostic accuracy of EUS is highly dependent on the experience of the endoscopist. EUS facilitates early diagnosis and treatment planning and improves survival rate in pancreatic cancer, but EUS is difficult to diagnose tumors over 4 cm in diameter, and ultrasound attenuation can cause the surrounding structures to be poorly visualized, affecting judgments of staging and resectability. Artificial intelligence (AI)-assisted EUS has the potential to improve the diagnostic accuracy for pancreatic cancer[55,56]. AI can assist EUS-FNA/fine-needle biopsy (FNB) by providing real-time feedback to endoscopists in selecting the appropriate needle size and type, guiding the depth and location of the puncture, and helping assess sample quality[57]. Additionally, AI is expected to reduce the number of punctures required to obtain an adequate sample, improve puncture accuracy, and reduce the risk of complications[58]. Some studies have assessed EUS-FNB specimens with the help of AI, achieving the same diagnostic accuracy as macroscopic on-site evaluation[59,60]. However, challenges remain in integrating AI into clinical practice, including the need for extensive validation and generalization of AI models[61,62]. Nowadays, quantitative parameters in CE-EUS such as relative peak enhancement and in-slope derived from time-intensity curve analysis can improve the characterization of pancreatic lesions, and many studies are exploring the diagnostic value of Harmonic CE-EUS for pancreatic masses[63]. Moreover, EUS can guide radiofrequency ablation or inject antitumor agents to treat pancreatic cancer, though which was designed for unresectable cancer[64,65]. In summary, EUS has the indispensable role in staging and therapeutic decision-making in pancreatic cancer[66,67].