Introduction

Pancreatic cancer is the seventh leading cause of cancer-related death worldwide.1 Due to its aggressive nature, the number of deaths is almost the same as the number of cases. A diagnosis of pancreatic tumors is usually based on imaging examinations, such as ultrasound (US), computed tomography (CT), or magnetic resonance. According to the National Comprehensive Cancer Network (NCCN) guidelines,2 patients with resectable pancreatic cancer should be operated on without prior biopsy, except for patients with high-risk features, in whom neoadjuvant therapy is recommended. However, pathologic diagnosis is required in patients with nonresectable tumors that are to be treated with neoadjuvant therapy.2 Pancreatic cancer is resectable at diagnosis in only 20% of cases,3 and thus biopsy is required in a large group of patients.

The NCCN guidelines on pancreatic adenocarcinoma2 recommend endoscopic ultrasound (EUS)-guided biopsy as a method of choice when neoadjuvant therapy is considered. Fine-needle aspiration (FNA) is no longer recommended, as core biopsy provides higher diagnostic yield.4 This is especially important when genetic testing is required. CT-guided biopsy can be performed when EUS-guided biopsy is not feasible. However, there are very few studies on CT-guided core-needle biopsy of pancreatic masses.

The primary aim of this study was to present the diagnostic yield and safety of CT-guided biopsy of pancreatic masses. The secondary objective was to evaluate efficacy of CT-guided biopsy as a salvage procedure after failed EUS-guided biopsy, US-guided FNA, surgical biopsy, or endoscopic retrograde cholangiopancreatography (ERCP)-guided biopsy.

Patients and methods

The Bioethical Committee waived the need for its formal consent due to a retrospective nature of this study. The research was conducted following the Declaration of Helsinki. The retrospective analysis was done on a prospectively maintained database, including percutaneous CT-guided core biopsies of pancreatic masses in 102 consecutive patients. Each patient underwent contrast-enhanced CT or magnetic resonance imaging (MRI) (including native, arterial, and venous phases), which revealed a pancreatic mass.

CT-guided biopsy was performed in the patients in whom EUS-guided biopsy was not feasible or failed. Regarding coagulation profile, the limits that allowed us to perform the procedure included international normalized ratio below <⁠1.5 and platelet count above 50 000/ml.

One of 3 interventional radiologists (GR, DK, and KM) with at least 5 years of experience in CT-guided procedures performed the biopsies. The procedures were done using an interventional CT suite (320-row Aquilion One, Toshiba, Japan). An unenhanced CT scan was done at the beginning of the procedure and compared with previous contrast-enhanced CT or MRI scans to plan an optimal needle path. If the risk of arterial injury was high, contrast-enhanced CT was also done during the procedure. US was available at each procedure and performed just before taking a sample to ensure the needle was away from critical structures such as arteries.

An 18G biopsy gun (Max-Core, BD, Franklin Lakes, New Jersey, United States) with a 22-mm penetration depth was used to perform the procedures. The coaxial technique was applied for all biopsies (17G TruGuide coaxial needle, BD). Apart from typical sharp-tip stylet, these needles come with additional blunt-tip stylet. Whenever the coaxial needle was near a blood vessel, the blunt-tip stylet was inserted to avoid puncture of the vessel and decrease the risk of bleeding. The fat traversing route was preferred,5 but when it was unavailable, transhepatic or transgastric approach was selected. All procedures were done under local anesthesia.

After the needle removal, unenhanced CT of the upper abdomen was performed to identify possible complications. The complications were noted immediately after the procedure, and then within a 30-day period according to the Clavien–Dindo classification.6

Technical success was defined as obtaining at least 1 tissue sample. Positive diagnostic biopsy was defined as a presence of at least 1 sample that allowed for making histopathologic diagnosis according to a pathologic report. Sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) were calculated. The gold standard used for calculation of these metrics included combined data from histopathologic report, further diagnostic studies, performed surgeries, and clinical information about the course of the disease.

We collected the following data: location and size (the greatest diameter) of the tumor, number of samples taken, pathologic diagnosis, and patient radiation dose in terms of dose length product (DLP).

All data analyses were performed using statistical libraries available for Python and R (R Foundation for Statistical Computing, Vienna, Austria).

Results

Our analysis included percutaneous CT-guided core biopsies in 102 consecutive patients (55 women and 47 men) at a mean (SD) age of 65 (9) years. The masses were located in the pancreatic head (n = 48), body (n = 40), or tail (n = 14). Median (interquartile range [IQR]) size of the lesion was 39 (30–49.75) mm.

From 1 to 5 (mean 2 [0.62]) tissue samples were taken. The most frequent needle access route was the anterior one (n = 65), followed by the transgastric (n = 15), transhepatic (n = 11), posterior (n = 9), and lateral (n = 2) route.

Technical success was achieved in 100% of cases. The overall diagnostic accuracy, sensitivity, specificity, PPV, and NPV were 90.2%, 89.6%, 100%, 100%, and 37.5%, respectively. In 33 patients, CT-guided core biopsy of the pancreatic mass was done after other biopsies failed. Twenty-eight (84.8%) of such salvage CT-guided core biopsies were successful (Table 1).

Table 1. Results of computed tomography–guided core biopsy performed after a failed biopsy under different guidance

Previous failed biopsy

Patients, n

Successful salvage CT-guided core biopsy, n (%)

EUS-guided

14

11 (78.6)

ERCP

5

5 (100)

US-guided FNA

11

9 (81.8)

Surgery

3

3 (100)

Total

33

28 (84.8)

Abbreviations: CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasound; FNA, fine-needle aspiration

Median (IQR) radiation dose (DLP) was 654 (433.75–874.5) mGy/cm. Contrast-enhanced CT during the procedure was done in 16 patients.

The reported pathologic findings included adenocarcinoma (n = 79), lymphoma (n = 1), ductal carcinoma (n = 1), neuroendocrine tumor (n = 2), chronic lymphoblastic leukemia (n = 1), renal cell carcinoma metastasis (n = 1), serous cystadenoma (n = 1), and inflammation (n = 6). Samples reported as nondiagnostic (n = 10) presented with insufficient tissue amount, necrosis, or normal pancreatic tissue.

No major complications were noted. None of the patients required readmission or prolonged hospitalization. One patient experienced postprocedural abdominal pain that resolved within 2 hours after administration of paracetamol and morphine; no further analgesic treatment was necessary (Clavien–Dindo grade I). Very small hematomas (<⁠2 cm in diameter) were seen on postprocedural CT scans; however, none of them was clinically overt and they were classified as type 1 according to the Bleeding Academic Research Consortium classification.7

Discussion

Percutaneous CT-guided core biopsy is an established method of obtaining cancer tissue for pathologic diagnosis in many organs. However, it is less frequently used in pancreatic mass diagnosis than EUS-guided biopsy or aspiration, and there are very few publications regarding this topic.

This study aimed to report the effectiveness and safety of CT-guided core-needle biopsy of pancreatic masses. Technical success of 100% and clinical success of 90.2% are similar to the outcomes of previous studies. Strobl et al8 reported that 89.8% of CT-guided biopsies delivered tissue that allowed for making a definitive pathologic diagnosis. Similar results (86%) were published by Li et al.9 In some studies, the results were even better, with clinical success reaching 98.1%10 and 100%.11

Research literature on EUS-guided biopsy of pancreatic masses reports similar results in terms of effectiveness. A meta-analysis by Yang et al12 revealed pooled sensitivity of 84% (range, 43%–100%; 95% CI, 0.82–0.87). The most extensive study on EUS-guided fine-needle biopsy (FNB) involved 852 patients,13 and its overall accuracy reached 85.6% and sensitivity 83.3%. Chen et al14 obtained adequate histologic yield in 87.5% of EUS-guided FNB samples.

For 33 patients, CT-guided percutaneous core needle biopsy was a salvage procedure after other failed biopsies (EUS-guided biopsy, US-guided FNB, ERCP, or surgery). The success rate of salvage CT-guided biopsies was 84.8%. This indicates that CT-guided core-needle biopsy can often get reliable tissue samples even if other methods fail.

The NCCN guidelines2 do not recommend US-guided pancreatic biopsy. This is probably because of the advantages of CT over US guidance. CT guidance provides better visualization of the pancreas, adjacent organs, and the needle, which is crucial for the procedure’s safety. Air in the stomach or intestine impairs visibility on US images, while it does not reduce the quality of images acquired with CT. Some studies on US-guided pancreatic biopsy15,16 show promising results in terms of accuracy. However, they report a slightly higher major complication rate (2%) that should be considered.

No major complications were noted in our study and no patients required readmission or prolonged hospitalization. One patient experienced postprocedural pain that required morphine and subsided within several hours. Very small hematomas (<⁠2 cm) were seen in 2 patients but they did not present clinical significance. No other complications occurred. Regarding serious adverse events, the results are similar to other publications on CT-guided pancreatic biopsy,10,11,17 which also reported no major complications. The results of our study compare favorably with studies on EUS-FNB. For example, Thomsen et al13 reported a 5.4% complication rate with 2.3% of acute pancreatitis. In a study by Lin et al,18 the adverse event rate of EUS-guided biopsy was 5.8%, with 4.2% of patients presenting with bleeding. The transgastric and transhepatic approach was not related to increased complication risk, similarly to a paper by Hsu et al.17

These excellent safety results are associated with high-resolution CT images, which allowed for precise visualization of the pancreas and the entire needle path. The possibility to perform contrast-enhanced CT is advantageous when multiple vessels are near the tumor, and the risk of bleeding is high. A lack of bleeding episodes is most likely associated with the use of blunt-tip stylets that move the blood vessels away from the needle rather than puncture them.

In the context of precision medicine, it is worth noting that specimens obtained by CT-guided core-needle biopsies (18G) typically result in more tissue than EUS-guided biopsies (22G), thus increasing the chances for obtaining adequate specimens not only for histopathologic diagnosis but also the next-generation sequencing. Somatic profiling can identify potentially actionable genetic alterations. However, utility and impact of targeted therapies on overall survival in the case of pancreatic adenocarcinoma is still to be fully elucidated. Recent studies showed that up to 28% of advanced pancreatic cancer patients might have potentially targetable genetic variants.19 One of the most common mutations in pancreatic cancers, present in about 90% of cases, are KRAS alterations. However, targeting them still remains a challenge. In KRAS wild-type patients there are numerous alternative driver aberrations.20 In pancreatic cancer patients, somatic testing should include various genetic alterations, for example fusions (ALK, NRG1, NTRK, ROS1, FGFR2, and RET), microsatellite instability, mismatch repair deficiency or mutations in the BRAF, BRCA1/2, KRAS, and PALB2 genes.2

The study has several limitations that should be mentioned. It is a single-center study, and the analysis was retrospective. Also, no comparisons with other methods were made. Prospective comparative studies should be probably the next research step.

In summary, our study showed high effectiveness and excellent safety profile of CT-guided core-needle biopsy of pancreatic lesions.