Primary pancreatic lymphoma diagnosed by endoscopic ultrasound-guided fine needle biopsy

Ning-Xin Zhu, Xiao-Yan Wang, Ting Tong, Jia-Hao Xu, Yuan-Yuan Yang, Li Tian

Gastroenterology Department of the Third Xiangya Hospital, Central South University, Changsha 410013, China

TotheEditor:

Primary pancreatic lymphoma (PPL) is extremely rare, accounting for 1% of extra-nodal lymphomas and less than 0.5% of pancreatic masses [1] . Both cytological and histological analyses are required to confirm the diagnosis of PPL. The main treatments include chemotherapy and radiotherapy [2] . Pancreatic resection alone does not improve the survival rate [3] . Hence, it is necessary to differentiate PPL from other malignant tumors before surgery.To date, there is no consensus on the optimal diagnostic approach and most experience is from case reports.

We first exhibit five cases of PPL diagnosed by endoscopic ultrasound-guided fine needle biopsy (EUS-FNB). We did not observe complications of pancreatitis, surgical site bleeding, or infection after EUS-FNB procedures. All cases were recognized and treated in the Gastroenterology Department of the Third Xiangya Hospital, Central South University during January 2016 and December 2018. These EUS-FNB procedures were performed by an experienced endoscopist, utilizing a 22G core biopsy needle (Cook EchoTip ProCore, Bloomington, Indiana, USA). Clinical characteristics of all patients were consistent with the classic PPL diagnostic criteria defined by Dawson et al. and the World Health Organization framework guideline [ 4 , 5 ].

Five patients comprised two females and three males, aged 49 to 67 years. Their chief complaints were abdominal pain, distension, or mass. Patient 3 lost 5 kg weight within the last 6 months and patient 5 got a fever. No family history was reported. Patient 2 and 5 presented upper abdominal tenderness; patient 3 and 4 presented tenderness below the xiphoid; and patient 1 had no positive physical sign. White blood cell count (WBC), carcinoembryonic antigen (CEA), and carbohydrate antigen 19-9 (CA19-9) were within the normal range in all cases, while lactate dehydrogenase(LDH) was 382-886 U/L. All patients had a negative result of the serum IgG4 test. Total bilirubin (TB) of patient 1 was slightly elevated to 25 μmol/L and human immunodeficiency virus (HIV) antibody of patient 4 was positive ( Table 1 ).

Fig. 1. Enhanced CT-image of patient 2. A 9.0 × 7.2 cm mass with multiple cystic components located in the head of the pancreas. Tumor attenuation values were 38 HU at baseline, 50 HU in the arterial-pancreatic phase, and 65 HU in the portalvenous phase. Portal vein and inferior vena cava were compressed and translocated.No MPD or CBD dilation. HU: hounsfield units; MPD: main pancreatic duct; CBD:common bile duct.

Enhanced computed tomography (CT) showed a large hypodense mass lesion with heterogeneous contrast enhancement,small area of necrosis in the center, and lack of main pancreatic duct (MPD) or common bile duct (CBD) dilation for patient 2( Fig. 1 ). Patient 1 had a dilated CBD with a 1.1-cm diameter. Patient 4 had a dilated MPD with a diameter of 5 mm. No tumor necrosis or CBD/MPD dilation in other patients. No vessel infiltration.Tumor sizes and enhancement features were similar to contrastenhanced endoscopic ultrasound (CE-EUS) findings. No enlargement of mediastinal lymph nodes were observed on the chest CT scan for patient 1/2/3/4 before chemotherapy. In EUS, most pancreas neoplasms were hypoechoic. Elastography indicated that the strain ratio (SR) ranged from 4.37 to 18.83. CE-EUS scan exhibited a homogenous mass in patient 1, heterogeneous enhanced masses in patient 2/3/4, and a hypo-vascular mass in patient 5 ( Table 1 and Fig. 2 ).

Cytological analysis showed numerous hyperplastic lymphoid cells that were homogeneous and atypical. Histological analysis and immunohistochemistry results confirmed the final diagnosis of five cases as primary pancreatic diffuse large B cell lymphoma (DLBCL) ( Fig. 3 ). Three patients accepted the com-bination of cyclophosphamide, doxorubicin hydrochloride (hydroxydaunorubicin), vincristine sulfate (Oncovin), and prednisone(CHOP) chemotherapy and patient 3 received rituximab-CHOP therapy while patient 5 refused to accept any treatments. By the follow-up endpoint of June 2020, patient 1 and 3 have survived for more than 27 months and are still alive. Patient 2 was lost to follow-up. Patient 4 infected with HIV died within 2 months and patient 5 refusing any treatments carried a poor prognosis that passed away within 6 months after a definite diagnosis( Table 1 ).

Table 1 Clinical features.

Fig. 2. Images of patient 1. A: EUS scan image; B: Contrast-enhanced EUS scan image at 20 seconds; C: Elastography image [strain ratio = 12.4]. Long arrow and short arrow:the location of the mass. EUS: endoscopic ultrasound.

Fig. 3. Pathological results of patient 1. A: Hematoxylin-eosin staining of numerous small lymphocytes and few pancreatic ductal epitheliums; B: immunohistochemistry showed MUM1 positive.

Just as a large population-based study reported, male patients were slightly more than females and most patients were middleaged or elderly. The primary presenting symptom was abdominal pain, which is nonspecific and consistent with other literature reports. One interesting finding is that those common presentations of non-Hodgkin lymphoma type B symptoms like fever, drenching night sweats, unexplained weight loss of 10% or more within the last 6 months and severe itching are not so often to see in PPL. Most PPLs occur in the pancreatic head and they can also be found in other parts of the pancreas [6] . A large hypodense mass lesion with heterogeneous contrast enhancement, absence of necrosis, lack of MPD dilation, and higher attenuation values in the portal-venous phase compared to the arterial-pancreatic phase are the most suggestive enhanced CT signs of PPL that have been reported before [7] . In our cases, most tumors have no necrosis. Vessel encasement can be seen in some of the cases, but no vessel infiltration was observed, which favors the diagnosis of lymphoma rather than carcinoma. Sonographic appearances of pancreatic lymphoma are usually nonspecific. To confirm the diagnosis of PPL,the optimal method to obtain tissue samples must be highly sensitive and specific, minimum invasive, and quite practical. Some case reports use CT assisted percutaneous puncture or biopsy [8] , but their success rates and diagnostic accuracy are lower than those of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA),let alone that the injuries they caused are much more severe. Besides, EUS allows dynamic and real-time observations of the tumor,which CT cannot [9] . EUS-FNA combined with rapid onsite evaluation, immunohistochemistry, and flow cytometry is a valuable diagnostic modality for diagnosing and subtyping PPL. However,this approach has some weaknesses that can be complemented by EUS-FNB. One limitation is that EUS-FNA only provides a cytological specimen without histologic architecture that especially matters in diagnosing lymphomas [10] . Another defect is the unclear number of passes required to achieve adequate tissue sample collection without an onsite cytopathologist [11] . In a meta-analysis of comparison between EUS-FNA and EUS-FNB in sampling pancreatic masses, the forest plot showed significantly higher specimen adequacy and diagnostic accuracy in the FNB group, compared with those in the FNA group [12] . EUS-FNB combined with elastography and CE-EUS applies to diagnose pancreatic masses, which has further improved the efficacy of EUS for identifying pancreatic lesions.Elastography provides diagnostic information on elastic properties or stiffness of tissues. According to a meta-analysis, the sensitivity of EUS elastography for differentiating solid pancreatic masses was 0.97 (0.95-0.98), and the specificity was 0.76 (0.69-0.82) [13] . In our cases, the majority of SR values were high, indicating a stiff tumor and a high possibility of malignancy. CE-EUS assists in distinguishing pancreatic adenocarcinoma from other pancreatic diseases. As a non-invasive diagnostic method, it reveals the density of blood vessels in pancreatic tumor imaging [14] . What showed in our cases was that one was homogeneous enhanced, three in five patients were heterogeneously enhanced, and the last one was hypo-vascular in CE-EUS, which is distinct from pancreatic ductal adenocarcinoma that usually exhibits hypo-enhancement with CEEUS.

To sum up, we reported five rare cases of PPL diagnosed by EUS-FNB combined with elastography and CE-EUS. This approach not only obtains enough tissue samples that offer messages for differential diagnosis but also causes minimal injuries. As for now,EUS-FNB is one of the optimal techniques to diagnose PPL. On account of limited experience of PPL and chemotherapy as its firstline treatment instead of surgery, case reports and evaluations of EUS-FNB are necessary. Further large sample statistics are essential to define a more specific and accurate diagnosis procedure.

Acknowledgments

We want to thank Cheng-Hong Wang, Hematology Department of the Third Xiangya Hospital, Central South University, for her technical assistance of immunohistochemistry and diagnosis of PPL.

CRediT authorship contribution statement

Ning-Xin Zhu : Data curation, Formal analysis, Writing - original draft. Xiao-Yan Wang : Conceptualization, Funding acquisition,Supervision. Ting Tong : Data curation, Writing - review & editing.Jia-Hao Xu : Writing - review & editing. Yuan-Yuan Yang : Writing -review & editing. Li Tian : Investigation, Methodology, Supervision.

Funding

This study was supported by a grant from the Hunan Provincial Science & Technology Department of China ( 2020SK2013 ).

Ethical approval

This study was approved by the Ethics Committee of the Third Xiangya Hospital of Central South University (2019-S245). Written informed consent was obtained from all participants.

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.