Bruce A. Feinberg, DO; Jennifer Fillman, MBA; Justin Simoncini, MBA, MPH; and Chadi Nabhan, MD, MBA, FACP
FOR THE THIRD CONSECUTIVE YEAR, the editors of Evidence-Based Oncology (EBO)™ are dedicating the February issue to immuno-oncology (I-O), and Cardinal Health has been a part of the conversation in each of the 3 issues. We marvel at the fact that it is a mere 5 years since the FDA approval of the first I-O of the modern era, the cytotoxic T-lymphocyte associated protein 4 (CTLA-4) inhibitor ipilumumab1; 2 years since the approval of the first 2 programmed death-1 (PD-1) inhibitors (nivolumab and pembrolizumab)2,3; and just months since the approval of the first programmed death-ligand 1 (PD-L1) antagonist (atezolizumab).4 In such a relatively short time, these I-O therapies have:
• Garnered FDA approvals in 6 tumor types
• Received indications for adjuvant, first-line metastatic, and salvage disease
• Been combined for dual I-O therapy
• And, are likely to receive 5 additional tumor type approvals in the proceeding 12 to 18 months.
Key opinion leaders and subject matter experts openly conjecture the end of the chemotherapy era while the silence around precision medicine is deafening. More than half of all actively accruing cancer therapeutic trials involve I-O across 52 different malignancies as single agents, dual I-O regimens, and I-O in combination with chemotherapy and targeted therapy.
Our 2016 EBO™ article concluded with the statement: “Stakeholder adoption of I-O is no longer a question of ‘IF’ but a question of ‘WHEN.‘ Who will be treated, with ‘what’ types of cancer, in ‘which’ stage and for ‘how’ long remain unanswered questions.”5 A year later, many of these questions are already being answered as new ones emerge, as I-O is poised to become the backbone of modern cancer treatment. Chief among these questions is ‘What lies beyond CTLA-4, PD-1, and PD-L1 in the future of I-O therapeutics?’ Most I-O treatments do not directly attack the tumor, rather, they mobilize the immune system to recognize and destroy the tumor. This can be achieved using various approaches, including antibodies, peptides, proteins, small molecules, adjuvants, cytokines, oncolytic viruses, bispecific molecules, and cellular therapies.
We believe the next I-O frontier to move from the bench to the bedside is cellular therapy in the form of chimeric antigen receptor (CAR)-T cells. In the first published trial from the University of Pennsylvania (U-Penn), 27 of 30 (90%) relapsed and refractory (R/R) patients with acute lymphoblastic leukemia (ALL) experienced complete remission 1 month after CAR-T infusion—22 (73%) of them had no evidence of minimal residual disease (MRD).6 Rapid, complete, and durable responses in highly refractory patients make CAR T a potential game-changer for cancer therapy, but the issues impacting stakeholder adoption for ex-vivo activated cellular I-O are significantly more complex than anything we’ve seen before.7
History and Development of CAR-T Technology
First and foremost in the discussion is recognizing that CAR T does not represent a drug, but, rather, a complex therapeutic process. Whereas most of the previously commercialized I-O interventions—from interferons to interleukins to checkpoint inhibitors—are essentially drug therapies, CAR T is operationally more similar to hematopoietic stem cell transplantation (HSCT). In fact, the concept of CAR T has its origin in the allogeneic bone marrow transplantation (BMT) of ALL. More than a quarter century ago, observations of durable remissions in patients with ALL, post BMT, who suffered graft versus host disease (GVHD) led to the understanding that donor or grafted T-cell recognition of malignant host lymphoblasts could impart long-term disease control (graft versus leukemia effect).8 The hypothesis generated was: if autologous T cells could be conditioned/manipulated to recognize malignant cells, then tumor control could be achieved without the negative consequences of GVHD.
When mild, GVHD is a complex chronic disease in which the grafted immune system is at war with the host organs; however, GVHD is life-threatening when severe. It would be critical to solve the problem of GVHD should any immune cellular therapy be successful. One method developed to create tumor recognition without the complications of GVHD involved reprogramming autologous T cells to identify and eliminate malignant cells through tumor-specific antigen recognition. The reprogramming required harvesting T cells from the patient via apheresis, transporting the cells to a wet lab where they could be chemically modified, and then altering the cells by linking the extracellular antigen recognition domain from a monoclonal antibody fragment to the T cell’s intracellular signaling domains.9 This newly modified autologous T-cell antigen receptor complex would then be a fusion, or chimera, of 2 proteins, or CAR. While still in the lab, the newly created CAR-T cells could then be incubated, or more technically, activated, to expand their number. Once adequately expanded, the CAR-T cells could be re-infused into the patient, but only after the patient is primed with chemotherapy to deplete their own circulating lymphocytes, which might dilute the CAR-T cells’ effectiveness. T cells engineered to express such CARs engage an antigen on a tumor cell through the extracellular antibody domain, thereby activating the T cells in a major histocompatibility complex–independent manner.8 Stated less scientifically, CAR-T cells can stimulate potent cytotoxic immune responses without the negative consequences of GVHD.
Early Clinical Experience: Efficacy
In Table 1, we have summarized selective data from completed or ongoing phase I/II CAR-T trials, which have been published and presented. This early research has focused on cancers in which the malignant cells express CD19—an antigen expressed only on malignant and normal B cells. CAR-T cells expressing anti-CD19 recognize and kill CD19-expressing malignant cells.10 Among the CD19-expressing malignancies, pediatric R/R ALL has garnered the most attention. CAR-T treatment of pediatric R/R ALL has resulted in remarkable clinical benefit and fewer severe adverse events (SAEs) than adult R/R ALL treated with CAR-T cells, making pediatric ALL the leading candidate for the first FDA-approved indication.
Stakeholder adoption will likely be brisk for the same reasons, but the overall healthcare impact might be limited given the nature of this disease and the relatively small number of eligible patients; 2500 to 3500 new pediatric ALL in the United States are diagnosed annually, 80% of which are of B-cell lineage (CD19-positive); current 5-year survival is 85%, and treatment-related mortality is 1% to 3%.11 While responses are attained with blinatumomab in the small number of R/R patients, they are rarely durable and less than one-third of pediatric patients with R/R ALL are cured with allogeneic HSCT.11,12
Analogous to ALL, other CD19-expressing hematologic malignancies have been targeted with CAR-T cells. While the CAR-T cells used in these studies target anti-CD19, they are not all identical. The methodology to develop the fusion protein that makes a CAR T is unique to each manufacturer and represents 1 variable that may be responsible for differing efficacy and toxicity outcomes. The U-Penn group reported an overall response rate (ORR) of 57% (complete response [CR], 29%) in a cohort of patients with R/R chronic lymphocytic leukemia (CLL).13
The National Cancer Institute (NCI) reported on 15 patients: 9 with R/R diffuse large B-cell lymphoma (DLBCL), 2 with R/R indolent lymphoma, and 4 with R/R CLL.14 Of the entire NCI cohort of 15 patients, 8 achieved CRs; 4, partial remissions (PRs); and 1, stable disease. Investigators at the Fred Hutchinson Cancer Research Center (FHCRC) reported on 34 R/R non-Hodgkin lymphoma (NHL) patients, 18 of whom had DLBCL; 6, follicular lymphoma (FL); 6, CLL; and 4, mantle cell lymphoma (MCL).15 CRs in studied subtypes were: 38% in DLBCL, 67% in FL, and 50% in CLL.
Another study by the U-Penn group included 38 R/R patients who had either FL (14), DLBCL (21), or MCL (3).16 The median number of prior therapies was 4 (range: 1-10) and 32% of patients had prior autologous HSCT. ORR among 22 evaluable patients at 3 months was 54% in DLBCL, 100% in FL, and 50% in MCL. The 3-month ORR was 54% in DLBCL, 100% in FL, and 50% in MCL.16 The ZUMA-1 study confirmed multi-center CAR-T treatment feasibility, with the following reported results on 51 relapsed/refractory DLBCL patients treated with KTE019; Orr was 76% with CR in 47%.17 Finally, preliminary data also support activity in myeloma where studies are ongoing to explore how this strategy is best positioned among other recently approved novel antimyeloma agents.18
A critical component of the CAR-T treatment is the preinfusion conditioning regimen to reduce the circulating and competing T-cell population. Conditioning regimens varied in published studies suggesting that the conditioning program could be another factor impacting differential efficacy and toxicity. It is becoming increasingly clear that the selection of the conditioning program will be essential to optimize clinical outcomes, especially SAEs. Some patients in the trials presented were treated with fludarabine plus cyclophosphamide conditioning, while others received cyclophosphamide monotherapy. In one study, the combination regimen resulted in a higher CR (42% versus 8%), positioning it as a benchmark, if not a standard, in this nascent field of research.15
Early Clinical Experience: Toxicity
The rapidity with which CAR-T treatment has become the standard of care in pediatric R/R ALL, as well as expanded indications across R/R B-cell CD19–expressing malignancies—such as CLL, DLBCL, MCL, and others—may have more to do with managing SAEs than with efficacy. The SAE that has garnered the most attention is cytokine release syndrome (CRS) because it is the most dangerous toxicity, usually occurring shortly after infusion although it can be delayed up to 3 weeks after.26 Cytokines are immune cell signaling proteins that we associate with flu-like symptoms, which, if released into the circulation rapidly and in large amounts (cytokine storm), can be physiologically overwhelming and lead to vascular collapse and possible death.26
Grading of CRS toxicity has not been uniformly agreed upon. Severe CRS occurred in 7/16 (44%) patients in the Memorial Sloan Kettering Cancer Center (MSKCC) trial, 8/30 (27%) at U-Penn, 6/21 (29%) at the NCI, and 7/30 (23%) at FHCRC.27 Deaths attributed to CRS have been suggested in only 2 out of 97 pediatric patients with R/R ALL (2%). The Juno Rocket trial has been held twice by the FDA due to SAE-related deaths, albeit not necessarily CRS-related. Observations that CRS may be mediated by IL-6 led to tocilizumab (Genentech anti–IL-6) and etanercept (anti–TNF) use.26 IL-6 rescue appears to be an intriguing strategy to manage CRS, but its standardization and overall impact on treatment remains to be determined as patients who do not have CRS are unlikely to respond. The strongest predicting factor to developing severe CRS is the leukemia burden.26 Guidelines are being developed to manage CRS—most require that C-reactive protein (CRP) be monitored daily to identify patients becoming at risk, while tocilizumab and/or steroids are recommended for high-risk patients.26
Fevers, hypotension, and a variety of neurological symptoms (confusion, obtundation, myoclonus, and aphasia) have been frequently reported, as have fatigue, diaphoresis, anorexia, and diarrhea. Neurotoxicity deserves special mention, as it is a potentially lethal toxicity and appears unrelated to CRS. Neurotoxicity, its prevention, and its relationship to tumor response are not well understood. Tumor response is clearly the basis for tumor lysis syndrome, which can occur even 3 weeks following infusion; it was noted as late as day 22 in a CLL patient in the first published CAR-T report.10 The research into predictability, prevention, and management of SAEs without abrogating the clinical benefit of CAR-T therapy will be a critical element to the success of this I-O intervention.
An Expanding Therapeutic Platform
Although the salvage therapy of B-cell malignancies may become the earliest FDA-approved indications for CAR-T therapy, long term commercial success will require broader disease indications and labeling for use at earlier stages of disease. Methodologically, chimeric fusion protein design against antigens other than CD19, is not a limitation, nor do such antigens need to be associated with blood cells. A review of the active and accruing clinical trials provides insight into this possibility. Of the 57,889 oncology trials listed on ClinicalTrials.gov, 777 matched for checkpoint inhibitors and 121 matched for CAR T. Of these 121 CAR-T trials, 58 are being conducted in China. Of the remaining 63 trials conducted in the United States and/or the European Union, 39 are phase 1, 14 are phase 1/2 hybrid, and 5 trials are phase 2; the phase is not identified for the remaining 5 (Table 2). The phase 2 trials all focus on hematologic malignancies expressing CD19. Of the 40 trials with a projected completion date before 2019, the CAR-T platform is being tested in melanoma, sarcoma, ovarian cancer, and brain tumors based on targetable antigens other than CD19. Of particular note are 6 trials involving brain malignancy, specifically glioblastoma multiforme, a disease with limited therapeutic progress over the preceding 30 years and for which CAR-T success could lead to a fast-track FDA approval and a frontline indication.
Potential for Commercial Success
The complexities of CAR-T synthesis and delivery make clinical benefit just one aspect of commercial viability. FDA approval of CAR-T treatment is just the first barrier to overcome in their incorporation into routine cancer therapeutics. Manufacturers need to prepare to scale these products and develop wrap-around services to ensure that their unique characteristics are not a barrier to adoption. Some of these characteristics include site of care, fragility of the patient, apheresis, cryogenic transport, and cost of the therapy. Today, CAR-T treatment is only available in clinical trials through a limited network of specialized centers. This limited network will likely be maintained upon approval since the manufacturers and FDA will want to ensure that the administering center is accredited in a similar fashion to the Foundation for the Accreditation of Cellular Therapy (FACT).
Travel and boarding for very sick patients and their caregivers should be an important consideration, along with referral to these sites from regional hematologists and oncologists. Educational platforms on the referral process, side effects, and relapse management are needed for all stakeholders. The delay between apheresis and the infusion of the CAR-T therapy also represents a possible barrier. Manufacturing cannot begin until the successful completion of an apheresis process, which some patients may not tolerate, while for others the manufacturing time may be clinically impractical. Allogeneic CAR-T alternatives that avoid the logistical complexities of apheresis and patient-specific manufacturing are under active investigation but appear years behind in clinical development. Finally, only a handful of vendors offer cryogenic transport.
Despite numerous uncertainties that can impact the commercial potential of CAR-T therapy, it is noteworthy that the leading compounds are supported by well-capitalized and committed companies. A conservative industry assessment of the commercial potential of this novel therapy puts the total market for CAR-T therapy in excess of $1.5 billion by 2020. Based on projected study completions and FDA filings, the first approved indication will likely be pediatric and young adult R/R ALL, followed by R/R DLBCL. Market models also expect commercially viable approvals for R/R MCL and R/R CLL, although some of these likely depend on the postapproval success of the initial indications.
The actual cost of therapy remains to be determined, but patient-specific manufacturing is costly and not easily scalable. In determining the cost, it is important to note that all steps in the CAR-T process require a Good Manufacturing Practice (GMP) facility or similar accredited environment. The highly trained physicians and nurses, the specialized facilities for manufacturing and administration, and the ongoing research and development will dictate that the cost of this therapy will be well north of 6 figures and likely benchmarked against allogeneic HSCT with which it will compete in many indications. The additional cost of transportation and boarding during the side-effect period and care between apheresis and infusion would challenge traditional payment and reimbursement models. Once CAR-T treatment is FDA-approved and used commercially, reimbursement strategies for institutions and physicians will be among the most significant questions to impact adoption. One possibility is a bundled payment approach, wherein a limited number of institutions will be allowed to administer therapy with payments based on predefined contractual agreements that include outpatient apheresis, inpatient hospitalization, and post discharge toxicity monitoring. There is precedent for such an approach as this is standard practice for HSCT reimbursement.
Preliminary results from the Novartis Eliana global study, conducted in 25 centers across 8 countries, which were presented at the American Society of Hematology annual meeting in December 2016, illustrate the complexity, excitement and caution surrounding CAR T. The trial enrolled 87 pediatric and young adult patients with CD19-positive R/R B-ALL (average age, 12 years; range, 3-23 years). There were 5 manufacturing failures, 6 patients died before undergoing infusion, and 3 patients discontinued therapy before infusion because of AEs. A total of 62 patients underwent infusion. Efficacy data on the first 50 patients revealed that 41 patients (82%) achieved CR and were found to be negative for MRD at 3 months. AEs of grade 3 or 4, suspected to be product-related, were seen in 74% of patients in the first 8 weeks and in 10% of patients after 8 weeks. CRS was observed in 79% of patients (grade 3 in 27% and grade 4 in 27%), with average onset on day 3 (range, 1 to 22 days) and an average duration of 8 days (range, 1 to 36 days). More than half (59%) of the 49 patients who developed CRS were admitted to the intensive care unit, 20% underwent invasive ventilation, and 10% underwent dialysis. No treatment-related deaths were observed.28
CAR-T treatment, specifically Novartis CTL-019, will likely get FDA approval for R/R ALL in the pediatric population. Label restrictions requiring failing blinotumumab are unlikely, as this treatment is not “curative” and both treatments may be a bridge to an allogeneic HSCT—a known curative treatment.11 Label restrictions might also include site-of-care restrictions, (eg, FACT-accredited networks). Such site-of-care restrictions are unlikely to be a barrier in the pediatric ALL population as these patients are currently treated in academic institutions, most already FACT-accredited for HSCT; however, the majority of eligible adult patients with hematologic malignancies and solid tumors would require referrals to such centers, often from the community. Logistical, operational, and financial aspects of such referrals become significant in managing a population of fragile patients with advanced malignancy and the significant CAR-T production variables.
As with any new drug or medical device, there are a myriad, and mundane, issues related to launch. CAR-T treatment, being a complex process, complicates this considerably because the apheresis center, CAR-T manufacturing, preparatory chemotherapy for lymphocyte depletion, intensive care management post CAR-T reinfusion, and postprocedure surveillance all need to be considered in logistics and pricing. We suspect that akin to HSCT, payment bundling will be the method of choice; patients, providers, and treatment centers will need education and coordination in reimbursement from payers. Monitoring patients after completion of therapy for possible late complications will likely be part of a mandatory phase 4 research program required by the FDA given the likely fast-track status of first indications. Manufacturers might need help in disseminating information regarding “centers of excellence for CAR-T” to community oncologists so that eligible patients are identified and referred.
Although a number of the factors described will contribute to the commercial success of these compounds, including final label, pricing, and clinical data, the largest driver of market uptake may be timing to first approval and initial experience with a commercial product. Once approved, CAR-T therapy will likely continue the trend of I-O therapeutic success. On the heels of CTLA-4, PD-1, and PD-L1, activated cellular therapy in the form of CAR-T cells will further the transformation of systemic cancer care from a chemotherapy to an I-O platform. However, the expansion of I-O platforms across an increasing number of malignancies, their indication in earlier lines of treatment, the anticipated use of I-O in combination, and trial designs that treat until progression may quickly erode the enthusiasm over their clinical benefit as stakeholders become mired in the debate over their cost.