Our case series illustrates that the clinical presentation of PPL can be difficult to differentiate from pancreatic adenocarcinoma without definitive pathological diagnosis. Reliance on symptoms, imaging and tumour markers – in the absence of definitive pathological diagnosis of suspected pancreatic adenocarcinoma – can potentially result in the misdiagnosis of a small minority of potentially curable patients. This is important as the prognosis and management of PPL differs greatly from that of adenocarcinoma.
Appearances on CT can be helpful to differentiate the two conditions, but are not definitive[9, 10]. Ca19-9 is the most useful tumour marker in pancreatic carcinoma, but can be misleading as it may also be elevated in other malignancies, particularly of the upper gastrointestinal tract, including PPL as described in case one . Without definitive pathologic diagnosis, potentially curable conditions such as PPL; as well as other malignancies with more favourable prognosis, including periampullary, distal common bile duct, duodenal and mucinous cyst adenocarcinomas, may be misdiagnosed.
Non-operative evaluation and biopsy of pancreatic masses can avoid the need for invasive surgery, if conditions such as PPL are found. The majority of cases of PPL in our literature review required laparotomy for diagnosis, which may have been avoided with successful non-operative biopsy and modern combined modality treatment. Radiological-guided percutaneous FNA of the pancreas is a very useful technique, which requires experienced radiologists and cytopathologists to obtain a diagnosis on a small amount of tissue. Endoscopic ultrasound has greatly improved the accuracy of diagnosis and obtaining diagnostic tissue [13–18]. Diagnosis of PPL may be extremely difficult on haematoxylin-eosin stains alone, and resemble poorly differentiated carcinoma and reticulum cell sarcoma, thus immuno-histochemical stains and flow cytometry are essential. It must be emphasised that cytological diagnosis may not be adequate for diagnosis and categorisation of an abdominal mass, and tissue biopsy should be considered. In some situations (including two of our patients with non-diagnostic FNA), laparotomy may be required for definitive diagnosis.
Treatment of PPL remains controversial – particularly the role of surgery and radiotherapy – and based on our findings we do not support routine pancreatectomy. All patients in our case series received multimodality treatment with both chemotherapy and radiotherapy, without surgical resection. Experience from our case series and literature review (table 2) indicates that this modern treatment regime achieves favourable outcomes, comparable or better than surgical series, without the morbidity of surgical resection. Most modern authors would not recommend surgery except when non-surgical diagnosis is unsuccessful [20–22]. Surgery is difficult in PPL because tumours are large, and often associated with an otherwise histologically normal pancreas, carrying a high risk of postoperative pancreatic fistula. Technical improvements in pancreatic surgery have led to reduced peri-operative morbidity and mortality, and in contrast Koniaris argues that pancreatectomy should be reevaluated as a method of improving local control and cure rates.
It is pertinent to consider the diminishing role of surgery in other localised extra-nodal lymphoma. Historically, the treatment of localised gastric non-Hodgkin's lymphoma, excluding MALT (mucosa-associated lymphoid tissue)-type lymphomas, was based on surgery – to ensure adequate diagnosis and staging, and maximise survival rates . Three recent large prospective studies [24–26] and two smaller randomised controlled trials [27, 28] of chemotherapy vs. combined surgery and chemotherapy have reported equivalent survival. Thus the role of surgery in gastric lymphoma may be limited to rare patients with acute complications or residual disease following non-surgical treatment [26, 29] In intestinal non-Hodgkin's lymphoma, following a prospective study there is a trend away from extensive resection . Localised extraintenstinal non-Hodgkins lymphoma has been treated without surgery for over two decades, following numerous well conducted prospective clinical trials. These changes in management illustrate the need for large cooperative prospective studies in PPL, which will better define any benefit of resection, as well as the role of therapies such as radiotherapy and rituximab.
Radiotherapy has had only limited reported use in the treatment of PPL (see table 2). Potential concerns about the safety of radiotherapy in this region may be unnecessary, as toxicity has been substantially reduced with three-dimensional treatment planning and conformal delivery of radiotherapy. With these modern techniques small bowel toxicity is minimised, and none of our patients required radiotherapy interruption. The role of radiotherapy in early stage high-grade lymphoma remains controversial in a global sense. In localised intermediate and high-grade non-Hodgkin's lymphoma, chemotherapy using the CHOP regimen plus adjuvant radiotherapy is superior to chemotherapy alone. Intensive chemotherapy regimens without radiotherapy have been shown to be superior to CHOP plus involved field radiotherapy. An important question asks whether the addition of adjuvant radiotherapy to more intensive chemotherapy regimens would further improve outcomes. We advocate combined multimodality therapy with both chemotherapy and radiotherapy for PPL.
Two of our four patients received rituximab (anti-CD20 chimeric monoclonal antibody, Genentech, CA, USA). We are not aware of its use in other series of PPL, however its use has been reported in gastric lymphoma , and evidence exists of its benefit combined with chemotherapy in other high-grade lymphoma for patients aged over 60.
Biliary sepsis is a potential problem of multimodality therapy in the setting of PPL, due to both the frequent presence of biliary stents, and the risk of neutropaenia. A study of endobiliary stents in multimodality therapy of pancreatic carcinoma reported low rates of complications, with 15 of 101 cases complicated by occlusion or migration, and no uncontrolled biliary sepsis or stent-related death. Another study of patients with malignant biliary obstruction and bile duct stents did not find increased biliary complications in patients receiving chemotherapy. Metal stents are superior to plastic stents for long-term patency [39–41]. Prompt recognition and treatment of biliary complications is important to allow ongoing therapy, as happened on several occasions to our patients.
We have examined for any possible association between polycystic kidney disease and PPL, given the existence of both conditions in case two. Although polycystic kidney disease has been associated with pancreatic cystadenoma and cystadenocarcinoma in a small number of case reports [42–44], we are not aware of any association with pancreatic lymphoma.
The study of PPL is limited by the rarity of the condition, and the consequent lack of randomised trials or large case series. The comparison of our case series with other series must be interpreted cautiously, given variable follow-up (range two – 108 months), incomplete data in some case series[2, 3, 12, 45], and the inherent publication bias within case series favouring positive results, as described by Albrecht. We strongly advocate a multi-centre prospective study of patients with PPL to improve patient outcomes.