Partial response in non-resectable adenosquamous carcinoma of the pancreas with high tumour mutation burden treated with gemcitabine, nab-paclitaxel and pembrolizumab

Abstract

A previously healthy man in his 60s was diagnosed with a rare histological subtype of pancreatic cancer, adenosquamous carcinoma. After somatic mutation profiling, it was found that the tumour had microsatellite instability status high and a high tumour mutational burden. The patient was started on combination therapy with gemcitabine, nab-paclitaxel and pembrolizumab. Tumour size and biomarkers showed a dramatic response eventually leading to the patient being transitioned to maintenance therapy with pembrolizumab. The patient has demonstrated continued response since the start of the treatment. This is the first report in the literature showing a sustained response in this type of neoplasm that was treated with a checkpoint inhibitor, and thus adds to the evidence supporting universal somatic testing in all pancreatic cancers for a tailored approach to therapy.

Background

Each year, more than 60 000 people in the USA are diagnosed with pancreatic cancer.¹ It is currently the fourth leading cause of cancer-related death in male and female patients.¹ By 2030, pancreatic cancer is projected to become the second leading cause of cancer-related death.² Five-year relative survival in the USA for newly diagnosed pancreatic adenocarcinoma is 11.5%.³ However, outcomes differ depending on stage and resectability.

The most prevalent histological subtype is pancreatic ductal adenocarcinoma (PDAC).⁴ Among other subtypes, adenosquamous carcinoma of the pancreas (ASCP) is one of the rarest, presenting only in about 1%–4% of all pancreatic exocrine malignancies.^{5 6} Compared with PDAC, ASCP tends to be larger at diagnosis, more frequently metastasises to lymph nodes and more regularly shows poor differentiation. These factors lead to the conclusion that ASCP generally carries a worse prognosis.^{5 7–9}

Given that ASCP is a rare tumour, tailored treatment guidelines have not been established.⁹ Therapy is based on treatment guidelines established for PDAC.^{5 9 10} Thus, ASCP treatment depends on stage, and may involve a combination of surgery, radiation and chemotherapy. Surgery is the only curative intervention for pancreatic cancer, and it is limited to patients with localised, resectable disease. Radiation can be used for local control where surgery is not possible. Chemotherapy is given for treatment at all stages of pancreatic cancer, either in sequence with the local control measures or alone for metastatic disease. First-line chemotherapy is either a four-drug combination of leucovorin, 5-fluorouracil, irinotecan and oxaliplatin (FOLFIRINOX) or a two-drug combination of gemcitabine and nab-paclitaxel.^11–15 A recent publication based on data extracted from the National Cancer Database concluded that for unresectable pancreatic cancers, adding immunotherapies to chemotherapy, with or without radiation treatment, is associated with a significant improvement in overall survival (OS) compared with chemotherapy alone.¹⁶ Based on these findings, and keeping with prior practice patterns, one might extrapolate a benefit for immunotherapy to patients with ASCP.

With the dearth of therapies and poor prognosis for metastatic pancreatic cancer, somatic molecular profiling in pancreatic cancer has increased due to the incremental usage of marker-driven investigative treatments.^{17 18} Here, we present a case of a patient with a confirmed pathological diagnosis of ASCP that has had an exceptionally durable partial response to second-line chemotherapy with gemcitabine, nab-paclitaxel and pembrolizumab, the latter added based on the rationale of the high tumour mutational burden (TMB-H).

Case presentation

A man in his late 60s presented to the outpatient clinic with left lower rib pain. He had a history of a heart murmur, hypertension, varicose veins, elevated body mass index and a prior left hip replacement. Family history was relevant for a non-smoker father who died of unspecified lung cancer and prostate cancer, and several paternal aunts who had breast cancer. Two months later, he returned to the clinic, concerned about post-prandial epigastric discomfort, bloating and new onset of constipation. He had an unintended 13.6 kg weight loss in 8 months. Complete physical examination was normal without hepatomegaly, ascites or peripheral lymphadenopathy.

Investigations

After his second visit, a CT scan was performed, which showed an enlargement of the pancreatic body and tail measuring 10.2×4.4×4.7 cm with enlarged peripancreatic and retroperitoneal lymph nodes (figure 1). Endoscopic ultrasonography confirmed an approximate 10 cm mass in the pancreatic body. The biopsy taken reported adenosquamous carcinoma with extensive necrosis. The cancer was determined to be unresectable by the surgical team due to the presence of lymphadenopathy and soft tissue encasement of the coeliac axis, common hepatic artery and main portal vein. It was also determined that it was unlikely that the tumour would achieve resectability with neoadjuvant therapy. Further invasive procedures were, however, deferred, and a tissue diagnosis was never made to confirm stage III versus IV disease. Pathology reported a predominantly squamous differentiation (highlighted by p40 immunostain) with focal areas of glandular differentiation, confirmed by mucicarmine stain. The tumour had microsatellite instability status—high (MSI-H). A TMB showed 26 mutations per Megabase, which included ATM R1730*, BCOR S336fs*42, BCORL1 R1048*, CD70 M1I, CDKN2A A76fs*70, CDKN2A A36fs*17, CDKN2A A68V, FBXW7 R465C, IDH1 R132C, KMT2A (MLL) R2160*, KRAS G12Dm, NF1 R2812, NF2 R351H, PIK3CA A1066V, PMS2 S547FS*15 PTPRO T765FS*14, RNF43 G659FS*41, TP53 R175H.

Figure 1.

CT showing hypoattenuating heterogeneous mass in the pancreatic body and tail that measures 10.2×4.4×4.7 cm that encase the coeliac axis and the common hepatic artery, as well as the main portal vein.

Treatment

Approximately 2 weeks after his biopsy, the patient started standard treatment for unresectable pancreatic cancer with FOLFIRINOX with palliative intent to improve quality of life and reduce symptoms. After receiving two doses, he was admitted with sudden numbness and weakness of his left hand and partial facial paralysis on the left side. He was diagnosed with an ischaemic stroke and was started on enoxaparin for hypercoagulability versus paroxysmal atrial fibrillation. He resumed treatment with FOLFIRINOX. He had received 3 months of treatment when monitoring CT scans showed an increase in the size of the mass, measuring approximately 4.7×10.6×7.9 cm (previously 4.4×10.2×4.4 cm), and an increase in retroperitoneal lymphadenopathy with concomitant uptrending of carbohydrate antigen 19-9 (CA 19-9). These all signified progression of disease. Per standard of care, a decision was made to switch treatment to the second-line regimen gemcitabine and nab-paclitaxel, with the addition of pembrolizumab to his therapy. We opted for this regimen due to the presence of TMB-H. According to the latest Food and Drug Administration (FDA) approval, pembrolizumab has shown efficacy when the tumour presents with TMB-H status, and the drug has a tumour-agnostic approval in this situation.¹⁹ He showed favourable response after the first cycle, with a marked fall in CA 19-9 levels. After the third cycle, this was further supported by decreased tumour size on CT. He completed six cycles of the three-drug regimen with neuropathy in the sole of the right foot as the only complication. Given the excellent tumour response after the six cycles of combination therapy, we discontinued gemcitabine and nab-paclitaxel and he transitioned to maintenance therapy with pembrolizumab every 6 weeks (figure 2).

Figure 2.

Trend in carbohydrate antigen 19-9 (CA 19-9) and carcinoembryonic antigen (CEA) biomarker in accordance with therapeutic changes. ASCP, adenosquamous carcinoma of the pancreas; FOLFIRINOX, leucovorin, 5-fluorouracil, irinotecan and oxaliplatin.

Outcome and follow-up

As of this writing, he continues with maintenance therapy, marking 20 months since diagnosis and 15 months on pembrolizumab-containing treatments. He has had continued response to treatment, with sustained normalisation of CA 19-9 and the most recent surveillance CT showing a pancreas body/tail mass that measures 6.1×1.9 cm (figure 3). The mass still encases the mentioned vessels, which remain patent, and peripheral lymphadenopathy remains present. However, the mass has continued to decrease in size. According to RECIST criteria, the tumours’ response was determined to be a partial response. He will continue pembrolizumab for up to 2 years if the response remains stable.

Figure 3.

CT of the abdomen showing hypoattenuating mass measuring 6.1×1.9 cm that continues to decrease in size after several weeks of pembrolizumab maintenance therapy.

Discussion

The pathogenesis of ASCP and its relationship to PDAC is still unclear. Three main theories exist.^20–22 The first one states that squamous metaplasia occurs as a result of obstruction and inflammation in the milieu of a developing adenocarcinoma. The second one, the collision theory, states that two independent and distinct histological tumours arise and coalesce. This is supported by electron microscopy findings that clearly demonstrated two different cell types with adenocarcinoma having abundant endoplasmic reticulum and well-developed Golgi apparatus and secretory vesicles, and the squamous cells having scarce endoplasmic reticulum but evident bundles of tonofilaments.²¹ The final theory is that there exists a malignant differentiation by a pluripotent duct cell into two appreciable histological types.²² This is based on immunohistochemical studies that have determined that both squamous and adenomatous cells from the tumour demonstrate positive staining for CA 19-9, AT 439 and keratin.

As molecular genetic profiling techniques have become more advanced and commonplace, it is now possible in routine clinical practice to identify treatment targets independent of histology. The initial results of the Know Your Tumor Initiative conducted in 2018²³ showed that in a sample of 640 patients with pancreatic cancer, 50% had at least an alteration predictive of a potential response to targeted therapies, with a median of four per patient. In their cohort, 27% of patients had highly actionable alterations. Comparing our patient’s tumour genetic profile with this report, we find many common genes, including KRAS, TP53, ATM and IDH1. Identifying KRAS and TP53 mutations is consistent with other reports that ASCP can bear mutations similar to PDAC.²⁴ The ATM mutation is one of the most frequent DNA repair genes to be mutated.^{23 25} Studies have identified this type of mutation as highly actionable because impairment in homologous recombination repair may lead to sensitivity to DNA-strand cytotoxic agents, examples of which are platinum-based chemotherapy and poly (ADP-ribose) polymerase (PARP) inhibitors.²⁶ Notably, our experience with oxaliplatin was not favourable in this case. We did not consider using a PARP inhibitor, as this drug class is not yet FDA approved for any line of treatment of ATM mutations in pancreatic cancers. Finally, previous evidence states that IDH1 mutations bestow a better prognosis than IDH wild type for PDAC²⁷; although it should be noted that overall, their incidence in PDAC is considered extremely rare.^{23 28} Furthermore, IDH1 mutations as potential markers of targeted therapy were recently discussed in a multicentre, open-label, single-arm, dose-escalation and expansion trial.²⁹ It demonstrated the efficacy of ivosidenib, an IDH1 inhibitor, as a treatment for adult patients with relapsed or refractory acute myeloid leukaemia (R/R AML) harbouring specific IDH1 mutations such as R132C, R132G, R132H, R132L and R132S. This led to recent approval by the FDA for the use of ivosidenib as a treatment option for R/R IDH1-mutated AML.³⁰

We chose gemcitabine/nab-paclitaxel-based therapy as the second-line treatment per (National Comprehensive Cancer Network (NCCN) guidelines for PDAC. Pembrolizumab was added based on the recent approval for this medication in patients with unresectable or metastatic malignancies that are TMB-H, defined as ≥10 mutations/Megabase.¹⁹ This approval was based on findings including the phase III KEYNOTE-158 trial, which evaluated pembrolizumab 200 mg once every 3 weeks versus placebo in 1032 patients with TMB-H tumours. The overall response rate was 29%, with a 4% complete response rate and a 25% partial response rate. The duration of response for these patients also showed that 57% had response duration of ≥12 months, and 50% had response duration of ≥24 months. Furthermore, the tolerability and preliminary efficacy of adding pembrolizumab to standard gemcitabine/nab-paclitaxel therapy for PDAC was supported by a phase Ib/II clinical trial conducted by Weiss et al.³¹ However, in this trial, patients were chosen for therapy irrespective of MSI, PD-1/PD-L1 expression and TMB.

The relationships among MSI, PD-1/PD-L1 expression and TMB in different types of cancers are still under investigation, and in particular the markers predicting response to immunotherapy.³² For example, although FDA approval for some immunotherapy agents in non-small cell lung cancer includes PD-1/PD-L1 criteria, PD-L1 positivity has shown to have limited utility in predicting whether a tumour is more likely to respond to PD-1/PD-L1 blockade.³³ On the other hand, although MSI and TMB have been identified as signifiers of response to immune checkpoint inhibitors, they are imperfect.³⁴ Studies have pointed out that a higher incidence of somatic mutations in tumours can result in a higher expression of neoantigens.³⁵ These neoantigens generate a T cell-dependent immune response by CD8 cytotoxic T cells, and indeed, tumours with high TMB tend to have high numbers of infiltrating lymphocytes.¹⁹ The PD-1 blockade by pembrolizumab in the tumour microenvironment may lead to more effective T cell activation in this enriched space.

To our knowledge, this is the first reported case of the rational addition of pembrolizumab to standard chemotherapy based on TMB-H for ASCP, documenting a durable partial response. Although this is only a single case, our report should prompt research in three areas. First, there is a need for further consideration of combination chemoimmunotherapy in pancreatic tumours with TMB-H or MSI-H features. A phase Ib/II study of first-line combination chemoimmunotherapy in 12 patients with PDAC showed longer than expected OS for metastatic disease,³¹ and our patient’s relatively long survival after progressing through first-line FOLFIRINOX adds some evidence for the potential benefit of this combination. Second, this case could be used to further the research of genetic profiling as guidance of treatment selection in this more aggressive histological subtype of pancreatic cancer. Third, we would advocate for genetically profiling all unresectable pancreatic tumours, as identifying TMB-H or MSI-H could justify the addition of immunotherapy in treating of pancreatic tumours.

Patient’s perspective.

I just didn’t feel right. My belly and back felt out of sorts. My primary set me up for a CT scan diagnosis. A mass was found on the pancreas. Immediately a biopsy was done, hoping it was nothing more than a cyst. Waiting for the call back was excruciating. Then the call, not good, You Have Cancer! Ok, what to do? I had gone through a breast cancer diagnosis for my wife a few years ago; she’s fine now, and I knew exactly where I wanted to go for my treatments. Back to the oncology clinic, now it’s my turn. Stage 3 Pancreatic Cancer! I knew I would be in capable hands and where I needed to be. Time to come up with a plan to attack and beat down this dreaded disease. Two plans were presented. I chose the more aggressive, chemo and home with a pump for 2 days, cycling every 2 weeks. Due to Covid, only patients were allowed in the treatment area. My head nurse and all the staff could not have been more caring and kind. Some of the nurses and staff remembered me from being there with my wife when she was having treatment. I prepared myself for the side effects you hear so much about, however, with the anti-nausea medication, I was able to tolerate chemo with minimal side effects. After a couple of months, CT scan time. I was so looking forward to hearing, “Everything looks good, your tumor is shrinking.” Not to be, it actually increased in size. It was a gut punch I did not expect; what to do now, Plan B. Chemo every week for 3 weeks then 1 week off. The change in medication did not affect me with the dreaded side effects. If you listen to your body it will tell you what you can and can not do. Tired, take a nap, Walk, if you are able, Eat, healthy. My motto, “Positive thoughts give Positive results”. My next CT scan had the desired results- decreasing in size, not spreading, made everyone feel that we’re making progress. With continued positive results, it was time to enter the maintenance phase of treatment. Every 6 weeks, immunotherapy; blood work, oncology visit, infusion. My team devised a plan for how long to continue treatment, along with CT scans and monitoring the results. All our family and friends were so supportive and elated at my results and how I was able to tolerate the process throughout my health journey. Today I feel like my old self. So long as the process shows improvement, if any, and nothing new, my wife and I have been able to restart our lives; socializing, traveling and recreational activities.

Learning points.

• Adenosquamous carcinoma of the pancreas is a rare subtype of pancreatic cancer and carries a worse prognosis with a lower response rate to chemotherapy than the more common pancreatic ductal adenocarcinoma.

• Use of pembrolizumab added to standard of care chemotherapy based on high tumour mutational burden tumour characteristics led to a dramatic and sustained response in our patient.

• Fifty per cent of pancreatic cancers have a targetable mutation, and guidelines recommend universal somatic testing, to allow some degree of treatment selection guidance for these patients.

Ethics statements

Patient consent for publication

Obtained.