Brain metastases from pancreatic adenocarcinoma

Abstract

Brain metastases from pancreatic adenocarcinoma (PA) are extremely rare. We report the case of a patient with PA treated by chemotherapy for liver and lung metastases. In spite of efficient control of the disease and an overall survival of more than 4 years with chemotherapy, the patient developed brain metastases. Brain metastases were revealed by headaches with mental confusion, and diagnosed by cerebral computed tomography scan. Treatment by radiotherapy could not be started because of the rapid fatal outcome of the disease.

Trial registration number:

reference pubmed (PMID): 17158774

BACKGROUND

Brain metastases from pancreatic adenocarcinoma (PA) are extremely rare. PA is usually a rapidly progressive disease and the majority of affected patients die from primary lesions before exhibiting clinical signs suggestive of brain metastases. The incidence of brain metastases will probably increase in the future because treatment advances have led to longer survival periods in some patients. Treatment of those metastases at this location remains difficult and there is a need to establish a diagnostic and therapeutic strategy adapted for taking better care of these patients.

CASE PRESENTATION

We report the case of a 57-year-old patient diagnosed with brain metastases from PA. In June 2003, he developed jaundice associated with an overall deterioration of health. Biological investigations showed a liver cholestasis with high total bilirubin (204 μmol/l; N<17 μmol/l). The abdominal computed tomography (CT) scan and endoscopic ultrasound showed heterogeneous lesions in the head of the pancreas measuring 15 mm maximum diameter. The superior mesenteric vessels were not invaded and there were no lymph node or distant metastases. Curative surgery was decided upon and the patient underwent a pancreaticoduodenectomy on 27 June 2003. Pathological results showed a well differentiated pancreatic lesion classified as duct cell adenocarcinoma. R0 resection was confirmed by tumour-free surgical margins. Analysis of the 10 lymph nodes around the pancreatic lesion did not show any tumour cell infiltration.

At follow-up, 6 months after surgery, an abdominal CT scan showed eight nodules in the right of the liver that were suggestive of metastatic lesions. Serum levels of CA 19-9 were 1772 U/ml (N<33) and CEA was 2.3 ng/ml (N<5). The patient was included in a phase II clinical trial assessing the efficacy and tolerability of a chemotherapy regimen, FOLFIRI 3: leucovorin 400 mg/m² as a 2 h infusion on day 1, followed by 5-fluorouracil (5-FU) 2000 mg/m² as a 46 h infusion and irinotecan 100 mg/m² as a 60-min infusion repeated on day 1 and 3, at the beginning and the end of the 5-FU infusion (day 1=day 14).¹ Treatment began in December 2003, which led to an objective tumour response; there was a reduction of more than 90% of the sum of the longest diameter of target lesions according to RECIST (Response Evaluation Criteria In Solid Tumors) and normalisation of tumours markers. Because disease was limited to the liver and has been controlled for more than 2 years by chemotherapy, a right liver resection was decided by a multidisciplinary team in order to remove all the segments involved. Surgery was performed in January 2006. Histology confirmed that complete resection of the liver metastases from PA was achieved (R0). Chemotherapy using FOLFIRI 3 was continued for 6 months after surgery. The abdominal CT scan done after the end of chemotherapy in July 2006, showed infracentimetric and non-specific bilateral lung nodules associated with an isolated right pleural nodule. ¹⁸F-fluorodeoxyglucose positron emission tomography showed a unique location corresponding to the right pleural nodule, which was confirmed by a positive pleural biopsy. Chemotherapy by GEMOX (gemcitabine 1000 mg/m² on day 1, then oxaliplatin 100 mg/m² on day 2, day 1=day 14)² began in August 2006. The abdominal CT scan performed after four and then eight cycles of chemotherapy showed an objective tumour response greater than 50%. Oxaliplatin was discontinued after the 12th cycle for a grade 3 neurological toxicity, and gemcitabine alone was then continued. In December 2007, after 22 cycles of this treatment, all target lesions appeared under control on the last CT scan performed 1 month before; however, the patient also then presented to the emergency department with intense headaches and mental confusion without fever.

INVESTIGATIONS

A brain CT scan showed multiple spontaneously hyperdense lesions with an important surrounding oedema suggesting brain metastases from PA (fig 1). There was also a left subdural haematoma complicated by a significant mass effect.

Figure 1.

Brain computed tomography scan. Transverse axial scan showing multiple hyperdense lesions with perilesional oedema.

TREATMENT, OUTCOME AND FOLLOW-UP

Intravenous high-dose corticosteroids were administered. External radiotherapy was discussed but could not be carried out due to the rapid spread of the disease, with coma and death 3 days later.

DISCUSSION

PA is a disease with a poor prognosis. In fewer than 10% of people with PA who are diagnosed with limited local disease suitable for surgery, the 5-year relative survival rate is about 20%.³ Median survival is about 6–8 months for patients with metastatic disease.⁴ Local tumour extension or hepatic metastases from PA are common but distant metastases may occur in the lung or peritoneum. On the contrary, brain metastases from PA are extremely rare. Japanese registers of brain tumours showed that among 7498 patients with brain metastases diagnosed from 1969 to 1993, only 0.1% of these patients had brain metastases from PA.⁵ Park et al showed, in a retrospective study on 1229 patients with PA diagnosed between 1980 and 2000, only four patients (0.3%) with brain metastases.⁶ A literature review published by El Kamar et al in 2004 showed 21 publications from 1965 to 2003, reporting 40 cases of PA with central nervous system (CNS) metastases. The majority of them (27 of 40 cases) were diagnosed in a series of post-mortem examinations.⁷

Some authors have suggested that expression of certain adhesion molecules by the tumour cells could predetermine their affinity for the CNS.^8,9 For example, some variant forms of CD44 are preferentially expressed by brain metastases from colon or lung lesions than by primary brain tumours.¹⁰

Besides, in the specific case of PA, expression of these CD44 variant forms is closely correlated with the progression and metastasis potential on tumours cells.¹¹ Nevertheless, if the PA cells can theoretically express molecules with an affinity for the CNS, such as some variant forms of CD44, the extremely poor prognosis of this cancer leads most often to death before the discovery of brain metastases. On the contrary, the discovery of PA by brain metastases is possible but it remains exceptional.^6,12

A brain CT scan or magnetic resonance imaging confirm the diagnosis, showing cerebral lesions that are usually enhanced by contrast injection and sometimes associated with surrounding oedema. A cerebral stereotactic biopsy can be indicated if diagnosis is doubtful.^7,12

We can assume that the development of treatment of PA will improve patient survival in the future and, therefore, indirectly increase the incidence of brain metastases. Thus, brain metastases could become a potential diagnostic and therapeutic challenge in the next future.

Radiotherapy remains the standard treatment of brain metastases from adenocarcinomas despite the absence of studies validating this treatment option in PA. Some reported cases have shown that radiotherapy did not change the clinical course of patients with brain metastases from PA, but other cases have shown interesting results with relative good local control and prolonged survival.^6,12,13 Systemic chemotherapy is still not very efficient on brain metastases; as shown in this clinical observation where, in spite of efficient disease control and an extended survival, chemotherapy could not avoid the occurrence of brain metastases. The blood–brain barrier (BBB) is an efficient physiological barrier against the entry of many chemotherapy molecules.¹⁴ Some authors have suggested several experimental approaches to increase drug delivery to the brain such as intrathecal injections, high-dose systemic chemotherapy and use of agents that inhibit the “Multidrug resistance-associated proteins” involved in drug efflux transporters of the BBB.¹⁵ Another approach using encapsulated chemotherapy agents in nanoparticles allows the endocytosis of these particles by the endothelium of the BBB and the delivery of the drug into the brain.¹⁵ Surgery is still not an option even though there has been one case where isolated brain metastases from PA were resected and the patient has been in remission for 1 year.¹⁶ Each patient with brain metastases from PA must be dealt with by a multidisciplinary team in order to offer the best therapeutic strategy.

LEARNING POINTS

1. Brain metastases from PA are still very rare but their incidence is expected to increase in the future due to therapeutic progress and improved survival of patients.

2. We reported the case of a 57-year-old man who developed brain metastases more than 4 years after a diagnosis of PA.

3. Radiotherapy remains the standard treatment of brain metastases from adenocarcinomas in spite of the absence of studies validating this treatment option in PA.

4. Systemic chemotherapy is not very efficient due to the fact that drugs hardly cross the BBB.

5. We could assume that optimising the diffusion of chemotherapeutic drugs in the CNS, combined with the development of targeted treatments based on small molecules with the ability to cross the BBB, may improve the care of these patients in the near future.