Irish Society of Gastroenterology Summer Meeting 2018

Outcomes for pancreatic cancer still strikingly poor

Patient outcomes for pancreatic cancer (PC) remain strikingly poor and are out of line with other solid tumours, the Irish Society of Gastroenterology heard at its summer meeting in Killarney. Consultant Gastroenterologist Prof Dermot O’Toole gave members a presentation on ‘Pancreatic Cancer Trends and Staging’.

While the poor prognosis — due to its aggressive biology and resistance to chemotherapy and radiotherapy — has remained unchanged in three decades, the total number of PC-related deaths is currently increasing and this increase affects all subtypes of the disease. There are 333,000 cases globally and, worryingly, it is predicted to be the second-leading cause of cancer death in the US by 2030. Surgery is the only possible cure but is possible in less than 20 per cent of cases. Just one-in-five of those patients are alive at five years.

In the UK population in 2011, the incidence of pancreatic cancer was 8,773, or 15.7 per 100,000. In 2012 there were 8,662 deaths “so incidence equals mortality,” he said.

While there are no accurate recent incidence rates of PC in Ireland, according to a study, Global Trends in Pancreatic Cancer Mortality From 1980 Through 2013 and Predictions for 2017, Ireland has seen an increase in PC mortality.

However, there have been some positives too. Pancreatic cancer mortality in middle-aged and younger individuals appeared to be declining, supporting a reversal of previous increased mortality trends seen in prior decades.

PC mortality began to decline in countries that introduced strict tobacco consumption control. The increase in obesity rates in some Western countries may explain why PC has not decreased as significantly as other tobacco-associated cancer, such as lung cancer.

The rise of obesity, which Prof O’Toole said was closely linked with PC, could see this positive trend reversing.

Prof Dermot O’Toole

In terms of causation, 90 per cent of cases are said to be sporadic and it can be argued that mucinous cystic neoplasms (MCNs) are represented in this group. Tobacco accounts for one-third of this group, with a relative risk of two-to-three.

In the non-sporadic group, hereditary cases account for just 1 per cent but the relative risk is very high, at 50-to-70 per cent.

Patients with familial forms, which are non-genetically inherited and means two or more first-degree relatives with pancreatic cancer, are being seen more often by clinicians, said Prof O’Toole.

There are also other predispositions, with 5-to-10 per cent germline mutations as well as non-genetically defined groups, or “genes that still haven’t been identified,” he said. Other risk factors include diabetes, heavy alcohol consumption and chronic pancreatitis.

He also asked where cystic precursor lesions fit into the overall picture.

“We see a lot of patients with pancreatic cysts and the mucinous cystic neoplasms, whether intraductal papillary mucinous neoplasms or mucinous cystadenoma; they are now really part of the surveillance strategy and there are strict guidelines in terms of how to deal with them.”

Some patients also have the non-visible form, which are not cysts but human pancreatic intraepithelial neoplasia (PanINs).

“The rates of mucinous adenocarcinoma and cystic mucinous adenocarcinoma have declined since 1992,” he said. “This has been paralleled by the surveillance protocols of incidental tumours that are discovered in MRIs or CT scans for other reasons. And I think this will ultimately influence the incidence in the epidemiology as well.”

PanIN sequence can be seen in both the sporadic and non-sporadic forms and the morphological abnormalities that occur through the PanIN sequence can result in some atrophy or lobular obstruction.

“You can see this sometimes, as small areas that might manifest as chronic pancreatitis-type changes in patients when endoscopic ultrasound is performed, or one would occasionally see it in MRI examination. But it is very, very subtle and hard to distinguish, making it difficult to identify lesions that are at a precursor stage.”

The ongoing challenge for clinicians is patients presenting too late.

Other challenges include that the tests for early detection are usually very invasive and there are a lack of biomarkers. Therefore, he said, clinicians should be focusing on high-risk groups, such as patients with precursor cystic tumours, diabetes and familial and genetic forms.

“High-risk groups are defined currently as having a risk of pancreatic ductal adenocarcinoma (PDAC) six-to-eight higher than control,” he explained.

Current tests used, such as the CA 19-9, lack sensitivity and specificity, the meeting heard and it is recommended to carry out CA 19-9 and CEA. The meeting also heard that CA 125 and CA 19-9 and/or LAMC2 are other markers that slightly increase the sensitivity for the detection of early cancer.

“EUS and axial imaging are useful for symptomatic patients but what about the precursor lesions and in familial cases?” the professor asked.

Clinicians tend to use EUS and MRIs, which are considered by many to be the most appropriate tests to identify these small precursor lesions. However, over-diagnosis is a concern and some cysts are benign or of a very low grade in type. Biopsying some of these very small lesions can be problematic.

CT is probably not a useful tool for screening, the professor said. It exposes patients to radiation and has low detection rates, while ERCP is not a useful tool in daily practice.

Target population

For a successful screening programme, clinicians should detect and treat early tumours, T1N0MO margin-negative pancreatic cancer, and high-grade dysplastic precursor lesions.

There is agreement that first-degree relatives (FDRs) of patients with PC from a familial PC kindred with at least two affected FDRs is a reasonable target population, as are patients with Peutz- Jeghers syndrome and patients with the genetic abnormalities p16, BRCA2 and HNPCC mutation carriers with ≥FDR.

“And that’s not an insignificant number of the population,” said Prof O’Toole. Screening should be carried out with EUS, MRI and not CT or ERCP, he said.

He also asked what clinicians should do when lesions are detected, as there is a “morbidity of surgery versus the yield”. The yield varies from about 2 per cent to 50 per cent in terms of a significant pathology when these patients are operated on. There is also a lack of prospective data on the longitudinal nature of the evolution of detected lesions and he asked if it is possible to use Sendai criteria for cysts in high-risk groups.

The questions of what age should clinicians start screening at and is it cost-effective were also asked.

There is a lack of consensus around the need for EUS fine-needle aspiration to evaluate cysts in high-risk groups, and what the optimal screening modalities and intervals were for follow-up imaging.

There is much disagreement around which screening abnormalities were of sufficient concern for surgery to be recommended, he added. However, the agreement on the need for smoking cessation was unanimous.

“We tell them all to stop smoking because we know the accumulated risk of having familial kindreds and smoking is absolutely detrimental. And in obesity as well, we’re tending to send patients to the dietician if obesity is a problem,” said Prof O’Toole.

In terms of making a diagnosis, clinical suspicion is still very important. Biopsies are important, as surgeons now want to have histological proof, the meeting heard. Small tissue samples, smears usually, are also important and point to the development of personalised medicine.

The role of EUS-guided core biopsy is really important for precision medicine, he added, as there is more and more work on organoid and xenograft development. “This will allow us to tailor exactly what patients should be treated with or shouldn’t be treated with.”

The feasibility of whole-exome sequencing of tumour and biomarkers was also mentioned, as was the challenge of getting tissue, even with the latest technology.

Regarding the use of PET/CT, Prof O’Toole cited a recently-published retrospective study of 550 patients. This found that PET/CT significantly improved diagnostic accuracy in all scenarios looked at.

PET/CT changed the staging of pancreatic cancer in 56 (p= 0.001) and influenced management in 250 patients (45 per cent). Furthermore, PET/CT stopped resection in 58 (20 per cent) due to have surgery.

However, the benefit of PET/CT was limited in patients with chronic pancreatitis or other pancreatic tumours.

The most cost-effective use of PET/CT was in the subgroup of patients who were suspected of having pancreatic cancer on MDCT and who were planned for surgery.

In terms of palliative care, there have been improvements, but the use of aggressive combination therapies is dependent on the performance status of the patient.

In conclusion, Prof O’Toole said that there have been some interesting trends in PC incidence, including a decrease in younger age groups. However, he was concerned that the rise of obesity may reverse this positive trend over time.

He concluded that the outcome remains strikingly poor and is out of line with other solid tumours, “but we can make a difference by targeting high-risk groups and identifying reliable genetic and epigenetic markers”.