Response to PD-1 inhibitor after progression on PD-L1 inhibitor in advanced HCC

Abstract

For hepatocellular carcinoma, the IMbrave150 trial demonstrated that combination atezolizumab and bevacizumab had significantly better overall survival compared with sorafenib in patients with unresectable disease. However, as more immune checkpoint inhibitors are approved as first-line agents in gastrointestinal cancers, there have been few reports on whether sequential PD-1/PD-L1 blockade is beneficial in the treatment of these diseases. We present a patient with hepatocellular carcinoma who had disease progression on atezolizumab, a PD-L1 inhibitor, but subsequently had a remarkable response to pembrolizumab, a PD-1 inhibitor.

Background

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death and, in the setting of unresectable disease, has traditionally been managed with tyrosine kinase inhibitors (TKIs), such as sorafenib and lenvatinib. These agents have a considerable side effect profile and prognosis remains poor as overall survival (OS) for patients on sorafenib is only 3 months longer compared with placebo.¹ Fortunately, immune checkpoint inhibitors (ICIs) targeting programmed death-ligand 1 (PD-L1) and its receptor programmed cell death protein 1 (PD-1) have been approved across various malignancies. In 2020, the IMbrave150 trial led to the Food and Drug Administration (FDA) approval of combination atezolizumab, a PD-L1 inhibitor, with bevacizumab, an anti-angiogenesis agent targeting the vascular endothelial growth factor (VEGF) ligand.² Furthermore, trials such as KEYNOTE-240 have demonstrated the efficacy of monotherapy pembrolizumab in the second-line setting for HCC.³

With the approval of ICIs in multiple cancers, and often within the same cancer type, clinicians are faced with the question of whether further PD-1/PD-L1 blockade is helpful in the setting of disease progression. There remains a paucity of data regarding the correct sequence of treatment and whether to retreat with an alternate ICI in gastrointestinal cancers. In this case report, we present a patient with advanced HCC who had disease progression on PD-L1 blockade and later had a remarkable response on treatment with a PD-1 inhibitor.

Case presentation

A woman in her 70s without underlying liver disease or cirrhosis presented to her primary care physician with fatigue. Bloodwork revealed elevated transaminases and a CT scan showed a soft tissue mass in the inferior right lobe measuring 13.6 cm. A biopsy confirmed moderately differentiated HCC. Baseline alpha-fetoprotein (AFP) was elevated to 108.9 ng/mL. A staging CT of the chest, abdomen and pelvis did not reveal metastatic spread. The patient underwent laparoscopic resection of segments 5 and 6 with final pathology pT1bNxM0 with negative margins. Seven months later, she experienced disease recurrence with peritoneal metastases. She enrolled onto a phase 3 study evaluating atezolizumab and bevacizumab vs standard of care sorafenib, a TKI, in patients with intact liver function. She was randomised to the experimental arm of 1200 mg of atezolizumab plus 15 mg/kg bevacizumab intravenously every 3 weeks. AFP prior to study initiation was 1066.7 ng/mL. Three days after cycle 1 day 1, she presented to the emergency department with worsening abdominal pain with CT showed two large abdominal and pelvic haematomas at 6.6×5.6×6.7 cm and 8.1×3.3×5.2 cm, respectively, which were attributed to bevacizumab. ASA was held, and she received one unit packed red blood cells. Three weeks later, on cycle 2 day 1, she reported subjective improvement in her abdominal pain and haemoglobin was stable. Bevacizumab was permanently discontinued and atezolizumab monotherapy was continued without a dose delay or reduction. A restaging CT scan after cycle 3 showed increase of existing peritoneal metastases but due to excellent performance status and downtrending AFP from 1067 to 446 ng/mL, she stayed on study and was treated beyond progression. However, another CT scan after cycle 5 showed continued growth of peritoneal implants and retroperitoneal nodes with an AFP increase to 1037 ng/mL (figure 1). She was taken off study at that time due to disease progression. She had a rapid clinical decline with nausea, vomiting, anorexia with weight loss, abdominal pain and severe fatigue making her wheelchair users. She was given oxycodone and antiemetic therapy as supportive care.

Figure 1.

Axial images on CT showing increasing peritoneal tumour burden (A) baseline, prior to cycle 1 of combination atezolizumab and bevacizumab (B) after cycle 3 and (C) after cycle 5.* peritoneal disease

The patient requested second-line therapy, and after extensive discussion, pembrolizumab, a PD-1 inhibitor, was recommended due to concern for intolerance to TKI therapy with her clinical symptoms. Ten weeks after the last dose of atezolizumab, she initiated treatment with pembrolizumab at a dose of 200 mg every 3 weeks. Three weeks after receiving the first dose of pembrolizumab, the patient reported marked clinical improvement with less abdominal pain and improved appetite. Restaging CT scan after 8 weeks of treatment showed response in the liver and peritoneal nodules with interval decrease in tumour burden (figure 2) and declining AFP to 9.2 ng/mL.

Figure 2.

Axial images on CT representing tumour burden before and after treatment with pembrolizumab. The first image depicts disease progression despite 3 months of atezolizumab. The second image depicts response in liver and peritoneal nodules with interval decrease in tumour burden following 8 weeks of pembrolizumab.

Outcome and follow-up

After 72 cycles, she continues to have significant response on pembrolizumab with stable disease on most recent imaging (figure 2). Her most recent AFP is 5.5 ng/mL.

Discussion

Checkpoint inhibition by PD-1/PD-L1 blockade effectively neutralises cancer cells through a robust immune response. Despite the increasing success rate of ICIs across various malignancies, currently only 20%–40% of patients respond to immunotherapy. Fortunately, next-generation sequencing technologies have been developed to help clinicians predict who will be a non-responder to treatment. By measuring a tumour’s genetic profile or a host’s immune activity, clinical decision making with ICIs has improved in sparing patients from unnecessary side effects and saving treatment costs. However clinical decision making becomes more challenging in the setting of disease progression despite initial PD-1/PD-L1 blockade. Furthermore, as more ICIs are approved, physicians will be faced with a variety of inhibitors to choose from and must decide whether second-line and beyond PD-1/PD-L1 blockade is beneficial. As treatment regimens continue to evolve, more data are needed on outcomes for patients whose disease stops responding to PD-1/PD-L1 blockade.

For HCC, ICIs have changed the treatment landscape. The IMbrave150 trial demonstrated that combination atezolizumab and bevacizumab had significantly better OS and progression-free survival outcomes compared with sorafenib in patients with unresectable HCC.² This synergistic effect may be related to reduced VEGF immunosuppression within the tumour microenvironment.⁴ Anti-VEGF therapies such as bevacizumab possibly enhance inhibition of the PD-1/PD-L1 axis and promote T-cell response to tumour cells. Despite this recent advance in treatment of HCC, there is little data on using subsequent ICIs after progression on anti-PD-L1 and anti-VEGF combination therapy. Recently, the phase 3 HIMALAYA trial showed an OS benefit of combination tremelimumab, a CTLA-4 (cytotoxic T-lymphocyte-associated protein 4) inhibitor, with durvalumab, a PD-L1 inhibitor, over sorafenib.⁵ The KEYNOTE-224 and KEYNOTE-240 trials demonstrated that HCC patients who had previously been treated with sorafenib had favourable outcomes with second-line monotherapy pembrolizumab.^{3 6}

The lack of data for second-line ICIs in patients who experience disease progression on initial PD-1/PD-L1 blockade is mostly due to second line and beyond trials excluding patients who had prior ICIs. Guidelines from the National Comprehensive Cancer Network (NCCN), American Society of Clinical Oncology (ASCO) and European Society for Medical Oncology recommend TKI or VEGF-R Antibody therapy following first-line ICI administration.^7–9 NCCN states that there is a ‘lack of data for subsequent use of immunotherapy in patients who have previously been treated with a checkpoint inhibitor.’⁹ According to ASCO Guidelines, ‘ICIs pembrolizumab or nivolumab may be especially beneficial for patients who have contraindications to or cannot tolerate TKIs,’ but also state that pembrolizumab should be considered as second-line therapy in patients who have not received prior ICI.⁷ The existing literature on drug sequencing in HCC is primarily focused on identifying superior post-TKI therapy sequences.^10–12 Authors of a meta-analyses of HCC drug sequencing discuss that further ICI after first-line ICI deserves investigation with options including switching ICI therapy or adding a TKI to existing ICI, which has a 12%–15% response rate in other diseases such a renal cell carcinoma.¹² To illustrate that point, Lowes et al present a patient whose HCC was refractory to combination pembrolizumab and lenvatinib but later responded to combination nivolumab and ipilimumab.¹³

Resistance mechanisms to initial ICI therapy include tumour intrinsic causes such as changes in $\mathrm{\beta }$-catenin signalling, downregulation of major histocompatibility complex molecules, and upregulation of additional PD-1/PD-L1 molecules. Tumour extrinsic causes include subtle changes in the tumour microenvironment, upregulation of coinhibitory molecules on lymphocytes and diminished antigen presentation and lack of T-cell activation.¹⁴ There is evidence that overcoming mechanisms of acquired resistance to first-line ICI is possible. While PD-L1 expression is dependent on the release of cytokines such as IFN-γ to trigger a signalling cascade via the JAK-STAT pathway, PD-1 expression does not rely on the same proinflammatory regulatory pathways. Therefore, switching from a PD-L1 inhibitor to a PD-1 inhibitor after a few cycles of treatment might circumvent the acquired resistance that can develop from loss of JAK1/2 function or downregulation of IFN-γ.¹⁵ Studies have already shown that sequential shifting from a PD-L1 inhibitor to its receptor, a PD-1 inhibitor, may have a benefit in patients with non-small cell lung cancer (NSCLC).¹⁶ Choosing to alternate these agents rather than combining them could also lead to less incidence of immune-related adverse events. For example, the incidence of pneumonitis with anti-PD-1/PD-L1 monotherapy is 5% compared with 10% in those with combination immunotherapy.¹⁷

More data on patient outcomes are needed to support sequential PD-1/PD-L1 blockade in the setting of disease progression. There are various ongoing prospective trials in HCC, NSCLC and melanoma that include patients with prior ICI treatment (table 1).

Table 1.

Current clinical trials that allow prior treatment with PD-1/PD-L1 blockade

ClinicalTrials.gov	Experimental arm	Phase	Tumour type
NCT03211416	Pembrolizumab+sorafenib	Ib/II	HCC
NCT03311334	Nivolumab+DSP-7888 dosing emulsion	Ib/II	HCC
NCT03539822	Durvalumab+tremelimumab+cabozantinib	I/II	HCC
NCT03475953	Avelumab+regorafenib	I/II	NSCLC
NCT05013450	Dupilumab+anti-PD-1/PD-L1	Ib/II	NSCLC
NCT04340882	Pembrolizumab+ramucirumab+docetaxel	II	NSCLC
NCT02535078	Durvalumab±tremelimumab+tebentafusp	Ib/II	Melanoma
NCT04902040	Atezolizumab or pembrolizumab+plinabulin+radiation therapy	Ib/II	Melanoma
NCT04577807	Pembrolizumab or nivolumab+lerapolturev	II	Melanoma

HCC, hepatocellular carcinoma; NSCLC, non-small cell lung cancer.

These studies could further elucidate whether sequential PD-1/PD-L1 blockade is helpful as a second-line treatment in the setting of disease progression. In addition, there are a number of adjuvant immune checkpoint inhibition-based trials in HCC (NCT 03383458, NCT 03847428, NCT 04102098). If these trials report positive results, the field will need to explore if immunotherapy based treatment in patients who experienced recurrence while on immunotherapy or soon after immunotherapy would benefit from first-line immune checkpoint inhibition. This information would also help answer the question about switch immune checkpoint inhibition therapy. However, as mentioned previously, currently there is only anecdotal evidence showing benefit of immune checkpoint inhibition switch therapy. Until more prospective clinical trials investigate anti-PD-1/PD-L1 therapy in populations with prior treatment on one of these agents, sequential PD-1/PD-L1 blockade cannot be recommended as a reliable treatment strategy at this time. Current research is focused on understanding mechanisms of PD-1/PD-L1 acquired resistance, which will help clinicians identify novel strategies to fully use the current array of ICIs.

Patient’s perspective.

I was placed on an initial immunotherapy to treat my liver cancer. There was no real progress on that immunotherapy and I was becoming more ill. As I did not want chemotherapy, my doctor gave me another immunotherapy one more try and put me on Pembrolizumab. Within a couple of months my tumour had shrunk by almost half. It was amazing how the medicine worked and continues to work. I could not be more pleased with my treatment. I suffer no side effects and tolerate the medicine well. I am very grateful for my doctors and Pembrolizumab. They saved my life.

Learning points.

• Drugs targeting the PD-1 receptor are active in hepatocellular carcinoma and can produce long-term results.

• PD-1 inhibition after prior PD-L1 inhibition failure represent a promising treatment sequence and requires further investigation.

• Immune checkpoint inhibition represents an alternative option in a patient with marginal performance status.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s)