Solid pancreatic lesions: etiology, morphology and utility of macroscopic on-site evaluation in predicting good yield of endoscopic ultrasound-guided biopsy

Introduction

Solid pancreatic lesions (SPL) can be due to a variety of neoplastic and non-neoplastic causes [1]. Neoplastic causes may have benign or malignant etiologies and include pancreatic ductal adenocarcinoma (PDA), neuroendocrine tumor (NET), solid pseudopapillary tumor (SPT) of the pancreas, metastatic lesions, lymphoma, etc. Among the neoplastic lesions, PDA is the tenth most common cancer and is the fourth leading cause of cancer-related deaths worldwide [2]. Non-neoplastic pancreatic lesions may include autoimmune pancreatitis, mass-forming chronic pancreatitis, and granulomatous diseases such as tuberculosis.

The management of neoplastic SPLs depends on the grade of the tumor, size of the lesion and stage of the disease; hence, acquisition of a tissue specimen is of primary importance. The endoscopic ultrasound (EUS) scope, when passed into the stomach, can not only visualize and characterize the SPLs in all parts of the pancreas but also help obtain biopsies of these lesions. It is essential to secure a good amount of tissue specimen through the fine aspiration/biopsy needle to facilitate histologic diagnosis. In this regard, rapid on-site evaluation (ROSE) is particularly useful because it involves visualizing the adequacy of the acquired tissue under the microscope immediately after the endoscopy. However, this facility may not be available in all setups because of the cost and expertise involved. In these cases, macroscopic on-site evaluation (MOSE) can be used by the endoscopist, who can directly observe the size of tissue with the naked eye [3]. Other strategies that have been used with the endosonoscope include strain ratio (SR) and contrast harmonic EUS (CH-EUS) to differentiate benign from malignant lesions [4].

The etiologies of SPLs can vary depending on the patient’s geographic location. Various studies have been carried out worldwide to identify the frequency and characteristics of SPLs and to record their histological diagnosis. However, very little work has been done in our part of the world in this regard.

Aim

The aim of this study, therefore, was to ascertain the etiologies of SPLs, document their morphology and identify the factors associated with good yield of endoscopic ultrasound-guided biopsy, including MOSE.

Material and methods

This was a prospective observational study that was performed in the Department of Hepatogastroenterology, Sindh Institute of Urology and Transplantation, from October 2020 to December 2023. All patients of either gender or any age, with solid pancreatic lesions and referred for endoscopic ultrasound were included in the study. Informed consent was obtained from all patients participating in this research. The study was done in accordance with the 1975 Declaration of Helsinki, as revised in 2000.

Bio-data of all the patients along with indication of the procedure was recorded in the pre-structured proforma. History, clinical examination and relevant laboratory investigations were documented. EUS was performed as a day care procedure and patients were asked to come in fasting condition to our endoscopy unit on the date given. The majority of the procedures were done under conscious sedation, while general anesthesia was administered only if necessary, for instance, to children below age 16 years or those adults who could not tolerate mild sedation. Post-procedure recovery depended upon the degree of sedation they received.

The equipment used was Pentax (linear) echoendoscopes. Morphological characteristics of the solid pancreatic lesions including their size and shape were noted. Type and size of the needle used and the number of passes performed during biopsy were documented. Needles used for biopsy were fine needle aspiration (FNA) or fine needle biopsy (FNB) (Boston Scientific and Cook), with bores being 22G or 19G. ROSE was not available. Therefore, MOSE was performed by the endoscopist. Macroscopically, MOSE was considered satisfactory if at least one tissue core at least 1 inch long was secured. The secured tissue specimen was sent to the expert histopathologist for diagnosis. The tissue was placed in formalin for histopathologic analysis. Various specialized tumor markers were used by the histopathologist to identify the origin of the metastatic neoplasm. The diagnosis described on the final histopathological report was recorded. Histologically, a biopsy specimen was considered to have a good yield if it was sufficient to establish a diagnosis.

Results

Of the total 176 patients included in the study, 111 (63.1%) were male, while 65 (36.9%) were female. Mean age was 53.8 (±12.6) years, with the age range 14–80 years. The number of older patients (i.e. > 45 years) was 136 (77.3%), while that of younger patients (i.e. ≤ 45 years) was 40 (22.7%). Among the various solid pancreatic lesions evaluated by the histopathologist, 97 (67.4%) were found to be neoplastic, 47 (32.6%) non-neoplastic and 32 (18.2%) patients were found to have suboptimal tissue. At the time of endoscopy, a satisfactory MOSE was observed in 129 (73.3%) patients. At the time of histopathological evaluation, good yield biopsies (GYB) were found to be 144 (81.8%), while those that demonstrated poor yield were 32 (18.2%). The various other baseline characteristics of the patients are summarized in Table I.

The mean largest dimension of the solid pancreatic lesions was 35.6 ±14.8 mm with a range of 8 to 85 mm. Infiltration by the neoplastic lesions into the surrounding blood vessels was found in 70/97 (72.2%) patients, while in the remaining patients (27/97; 27.8%) the tumor either only caused abutment of the surrounding vessels or did not infiltrate at all. Among those diagnosed with pancreatic ductal adenocarcinoma, vascular infiltration was the commonest, accounting for 59 (77.6%) cases.

The ages of the patients affected by the different kinds of SPLs and the location of the different SPLs in different parts of the pancreas are shown in Tables II and III. As shown, the majority of the patients with PDA, pancreatic metastasis and mass-forming pancreatitis (MFCP) were of older age (> 45 years). Patients with NET and autoimmune pancreatitis (AIP) were predominantly younger (i.e. ≤ 45 years). Also, PDA and MFCP predominantly affected the head region of the pancreas, while NET and metastasis predominantly affected the body of the pancreas.

Among the 97 neoplastic SPLs, the commonest was PDA (76; 78.3%), followed by NETs (13; 13.4%), metastasis (5; 5.1%) and others (3; 3.1%). Out of the 76 PDAs, 27 (35.5%) were moderately differentiated, 22 (28.9%) well differentiated, 13 (17.1%) poorly differentiated, 9 (11.8%) undifferentiated and 5 (6.5%) had squamous differentiation of adenocarcinoma. Among the 13 patients with NETs, only two had neuroendocrine cancers, i.e. WHO grade III with Ki-67 > 20%, while all the remaining patients had neuroendocrine tumors with Ki-67 < 20%: 9 had WHO grade I and 2 patients had WHO grade II (Table IV). Of the 5 patients with metastatic lesions, 4 were those who previously had renal cell cancers (RCC), while 1 patient had metastatic breast cancer. One patient had primary B-cell lymphoproliferative disorder of the pancreas, one had spindle cell neoplasms, and one had undifferentiated cancer with osteoclast-like activity. Among the 47 non-neoplastic SPLs, the commonest was MFCP (33; 70.2%), followed by AIP (5; 10.6%), tuberculosis (2; 4.3%), abscess (1; 2.1%) and normal tissue (6; 12.8%) (Table IV).

The factors predicting good yield of biopsy are shown in Table V. On univariate analysis, the factors predicting good yield of biopsy were size of solid pancreatic lesion > 15 mm (p = 0.045), use of FNB (vs. FNA) needles (p = 0.001), more than two needle passes (p < 0.0001), and satisfactory MOSE. However, on multivariate analysis, only satisfactory MOSE (p < 0.0001) and use of FNB needles (p = 0.045) were associated with good yield of biopsy. The performance of MOSE in predicting good biopsy yield is shown in Table VI. According to our criteria, satisfactory MOSE had a sensitivity of 85.4%, specificity of 81.2%, positive predictive value (PPV) of 95.3%, negative predictive value (NPV) of 55.3%, and diagnostic accuracy of 84.7% in predicting good yield of biopsy.

Discussion

In our study, the commonest causes of neoplastic SPLs were PDA, NET, and metastasis, while the most common non-neoplastic lesions were MFCP and AIP. Similar findings have been reported by various studies done in other parts of the world. In a review article, Santo reported that the four most common SPLs were pancreatic carcinoma, pancreatic NET, solid pseudopapillary tumor, and focal chronic pancreatitis [5]. An Italian study by Fabbri et al. also found that PDA and NET are the commonest causes of SPLs [6]. In our study, among patients with NET, most of the patients had WHO grade I (Ki-67 < 3%) while those having WHO grade III (Ki-67 > 20%) were least common. Titan et al. from USA, who studied the outcomes of pancreatic NET after surgery, also found that the frequency of grade I NET was higher than that of grade II and III lesions [7]. However, they found that the most common part of the pancreas affected by NET was the tail region, followed by the body and head. By contrast, in our study the most common part of the pancreas affected was the body, followed by the head and tail. In our study, metastases in the pancreas were found in 5 cases, of which 4 were metastases from renal cell cancer, while one was from breast cancer. Surprisingly, these findings were similar to those of various studies conducted elsewhere in the world. Sperti et al. reported that the most frequent primary locations of tumors metastasizing to the pancreas were the kidney, breast, colon, skin, and lung [8]. Similarly, Pan et al. also reported that the most common primary tumors metastasizing to the pancreas are RCC, and they have better prognosis if amenable to resection [9]. Among the less frequent causes of SPLs that we found were primary pancreatic B cell lymphoma, undifferentiated cancer with osteoclast like activity, and spindle cell neoplasms. Facchinelli et al. reviewed case reports and series of patients with primary pancreatic lymphoma and found that B cell lymphoma was the commonest type of primary lymphoma arising from the pancreas [10]. Also, Sakaguchi et al. from Japan reviewed 27 case reports of patients with spindle cell type anaplastic carcinoma of the pancreas, concluding that although rare, this is an established cause of primary pancreatic cancer [11]. Furthermore, Luchini et al. from Italy presented a case series of undifferentiated carcinoma of the pancreas with osteoclast-like giant cells (UCOGC), stating that it is a morphologically and clinically distinct variant of pancreatic ductal adenocarcinoma, and showed that cases of pure UCOGC had a significantly better prognosis than did cases of UCOGC with an associated epithelial neoplasm [12]. Among the non-neoplastic SPLs, besides MFCP and AIP we also found tuberculosis (TB) as a rare cause. Since TB is very common in the India-Pakistan region and affects many organs, including the lungs, intestines, peritoneum, pericardium, brain, bone, and others, it is not surprising that it also affects the pancreas. Siddeek et al. from India has also reported similar findings that were confirmed with FNA biopsy and showed recovery with anti-tubercular therapy [13].

In our multivariate analysis, the factors predictive of good yield of biopsy were found to be use of FNB biopsy needles (compared to FNA needles) and a satisfactory macroscopic on-site evaluation (MOSE) of the secured tissue. However, the gold standard for detecting good yield biopsy in our study was an adequate specimen as stated by the histopathologist. The superiority of the FNB needles over the FNA needles has also been shown by numerous studies worldwide. Kovacevic et al. from Denmark demonstrated this superiority and found that the tissue obtained by FNB was more than 6-fold larger than that obtained by FNA needles [14]. Also, a study by Wong et al. from Thailand showed that not only was the diagnostic yield of FNB needles higher than FNA needles in acquiring tissue from SPLs without ROSE but also its use was associated with fewer needle passes and shorter total procedural time [15]. Different definitions of adequate MOSE have been adopted by various researchers. Our definition of satisfactory MOSE was visualization of at least 1 tissue core at least 1 inch long. MOSE is gaining attention day by day, and several studies have demonstrated its utility. A study by Mangiavillano et al. from Italy demonstrated that combining EUS FNB with MOSE is non-inferior to EUS FNB with 3 needle passes [16]. Furthermore, a meta-analysis showed that the pooled accuracy of FNA and/or FNB specimens in yielding a pathologic diagnosis by MOSE was 91.3% [17].

Conclusions

The commonest neoplastic solid pancreatic lesions in our study were pancreatic ductal adenocarcinomas, neuroendocrine tumors, and metastases. Among non-neoplastic solid pancreatic lesions, mass-forming chronic pancreatitis and autoimmune pancreatitis were the most common. Among the rare causes, we found tuberculosis presenting as a solid pancreatic lesion, clearly reflecting the high prevalence of this disease in our part of the world and the need to control its spread. Pancreatic ductal adenocarcinomas were commonly found in the head region of the pancreas and affecting older age, while neuroendocrine tumors had a predilection for younger patients and the body of the pancreas. MOSE can be particularly useful in those centers where ROSE is not available. We found that factors that determine a good yield of biopsy are the use of FNB needles and satisfactory MOSE. Therefore, we recommend that in the absence of ROSE, use of FNB needles together with employment of MOSE criteria may ensure that a good biopsy sample is secured for proper histopathological diagnosis.