A Primer on Chimeric Antigen Receptor T-cell Therapy: What Does It Mean for Pathologists?

Chimeric antigen receptors (CARs) are composed of an extracellular antigen recognition site, a transmembrane domain, and an intracellular signaling domain that promotes T-cell proliferation, cytolysis, and cytokine secretion.^2,3  Genetic modification plays a crucial role in the creation of CAR-Ts, with the goal of inserting the synthetic sequence of the CAR into the genome to enable cell-surface expression of the CAR. Insertion of the synthetic CAR sequence is typically achieved with the use of an inactivated lentiviral or retroviral vector. Lentiviruses are advantageous in that they infect dividing and nondividing cells and can handle larger transgenes, whereas retroviruses can only infect dividing cells.⁴ 

In current clinical practice, all CAR-T therapies are autologous, in that the starting material is the patient's own immune cells. Therefore, the first step in the manufacture of a CAR-T therapy is the harvesting procedure for collection of peripheral mononuclear cells. T cells are selected/purified from the harvesting procedure product, genetically modified so that the cells express the CAR construct, and expanded in vitro. The CAR-Ts are then infused, and the patient is monitored for efficacy and safety (Figure 1).

Engineering of CAR-Ts has developed over time (Figure 2).⁵  Initial studies that led to the development of CAR-T therapy started in the late 1980s and early 1990s, and CAR design improvement is an area of active ongoing scientific and clinical investigation in an effort to improve antitumor effects and limit toxicity. In 1989, immunoglobulin–T-cell receptor (TCR) chimeras were generated using the variable heavy-chain and variable light-chain regions of an antibody (to 2,4,6-trinitrophenyl) fused to the β or α chains of the TCR (TCR_αV_H + TCR_βV_L double-chain chimeric receptor). The constructs were introduced using a Rous sarcoma virus expression vector into a cytotoxic T cell hybridoma. The chimera recognized its targeted antigen in a non–major histocompatibility complex-restricted manner to activate the cytotoxic T cells.⁶  As the double-chain receptor was complex to engineer, transduce, and express, a simpler single-chain version that contained an artificial single-chain extracellular variable region (variable heavy-chain and variable light-chain regions joined by a linker) and the CD3ζ or Fc receptor γ chain (FcRγ) immunoreceptor tyrosine-based activation motif was developed as a first-generation CAR.⁷  This construct resulted in interleukin 2 (IL-2) secretion and target cell lysis. Although these initial constructs could increase T-cell–mediated IL-2 expression and tumor cell targeting, the T cells would eventually become anergic, as costimulatory molecules were not included in the initial constructs.

To combat T-cell anergy after antigen stimulation, a second-generation CAR (designed against CD33) incorporated an additional moiety that provided T-cell costimulation, CD28, also known as signal 2 (Figure 2).⁸  In its native form, CD28 interacts with the B7 family of costimulatory molecules on antigen-presenting cells to present a secondary activation signal in addition to the primary TCR signal. Incorporation of the costimulatory molecule resulted in increased IL-2 secretion and recruitment of the phosphatidyl-inositol 3′ kinase regulatory subunit, indicating pathway activation. With the idea that if one costimulatory molecule improves a CAR, more may be better, incorporation of more than one costimulatory molecule has been addressed (third-generational CAR-T constructs; Figure 2). For instance, it has been demonstrated that the presence of both CD28 and 4-1BB (CD137) costimulatory molecules, within an anti-CD19 single-chain variable fragment (scFv)-CD3ζ construct, resulted in a synergistic, potent antitumor response and longer survival of anti-CD19 CAR-transduced umbilical cord blood–derived T cells within a severe combined immunodeficiency mouse in vivo tumor assay.⁹  Multiple combinations of other costimulatory molecules have subsequently been incorporated into third-generational CAR-T constructs such as CD134/OX40 and CD27/TNFRSF7 (Figure 2).^10,11 

In the current clinical setting, the CAR antigen-binding domain may be composed of an antibody scFv that contains a variable heavy-chain and a variable light-chain region, which recognizes and binds the antigen of interest.¹⁰  A murine scFv against CD19 is used for the current FDA-approved therapies for B-lymphoblastic leukemia and diffuse large B-cell lymphoma (DLBCL). A “hinge” or spacer region is present extracellularly between the scFv and the membrane to allow for proper spatial orientation for antigen-scFv interaction. If the tumor cell antigen is more superficially accessible at the cell surface for CAR binding, a spacer may not be necessary.¹²  A spacer region may include common immunoglobulin domain motifs or extracellular TCR- or CD28-derived motifs. Modifications of the CAR intracellular signaling domain are the defining determinants for each generation of CAR development. Since the first generation containing only the CD3ζ region, single (second-generation) or dual (third-generation) costimulatory molecules have been added to the CAR transgene to optimize intracellular signaling and promote a robust antitumor response through T-cell activation, survival, proliferation, and persistence. Current FDA-approved therapies are second-generation CARs in that they each contain 1 costimulatory molecule (either CD28 or CD137/4-1BB; boxed in Figure 2). Variation in costimulatory motif combinations have been tested, as discussed above, for added survival and proliferative signals. Fourth-generation CARs incorporate a separate expression vector that is used to provide an additional cytokine signal, such as IL-12 (Figure 2), to remodel the tumor microenvironment.¹³  Other modifiers may include suicide switches (eg, herpes simplex virus thymidine kinase combined with ganciclovir treatment, inducible caspase 9) or switch mechanisms that allow for better control of CAR activation and CAR-T killing to reduce potentially harmful adverse effects.¹⁴ 

Whereas viral vectors are being used to introduce CARs into T cells, the use of oncolytic viruses in conjunction with CAR-T therapy is also being examined, especially in solid tumors where access of CAR-Ts is more limited.¹⁵  Oncolytic viruses have been shown to effect tumor cell lysis; however, they have also been studied as a method to deliver cytokine/chemokine support or checkpoint inhibitors, or to enhance the immune response to the tumor. CAR-Ts have also been investigated as a means of delivery of oncolytic viruses to the tumor cells.

Finally, systems other than viruses are being evaluated to deliver the CARs in a way that may reduce the risk of insertional mutagenesis and the need of complicated manufacturing, such as transposon systems known as PiggyBac and Sleeping Beauty. Although there is the potential for reduced efficiency, they may also be cheaper with simpler good manufacturing practice.¹⁶  Clustered regularly interspaced short palindromic repeat (CRISPR) gene editing is also being assessed as a method to disrupt untoward signaling within CAR-Ts that may cause unwanted apoptosis of the CAR-Ts, effect immune checkpoint regulation that can affect their response, or result in rejection of allogeneic CAR-Ts.¹⁶ 

B-cell neoplasms have been studied and assessed most comprehensively for CAR-T therapy. CD20 has been explored as a target for adoptive immunotherapy,^17,18  and CD20 and CD19 bispecific targeting CAR-Ts are under investigation.^19,20  As there are no FDA-approved CAR-T therapies targeting CD20 as of this writing, a majority of this discussion will center on CD19 targeting by CAR-Ts.

An initial study²¹  was published in 2013 in which 2 children with relapsed and refractory pre–B-cell acute lymphoblastic leukemia (ALL) were treated with CTL019 CAR-Ts (anti-CD19 attached to TCRζ and 4-1BB signaling domains transduced using lentivirus), with complete remission observed in both patients 1 month after infusion (1 patient relapsed at 2 months). Robust expansion of the CAR-Ts was reported after infusion (×1000 initial expansion; highest levels of circulating CD3-positive CAR-Ts at day 9 or 10 for each patient of 33.6% and 71.5%, respectively). There was persistence of CAR-Ts in blood, bone marrow, and cerebrospinal fluid beyond 60 days using real-time polymerase chain reaction analysis. In an expanded study²²  published in 2014, 30 children and adults with relapsed or refractory ALL (29 B-cell and 1 T-cell expressing CD19) were infused with CD19-targeted CAR-Ts. There was a 90% complete remission rate (at 1 month; 19 remained in remission) with a 67% event-free survival rate and a 78% overall survival rate. Chimeric antigen receptor T cells were identifiable in blood, bone marrow, and cerebrospinal fluid, with a 68% probability that a patient would have persistence of CAR-Ts at 6 months. In patients with sustained remissions, CAR sequences were identified by quantitative polymerase chain reaction for up to 2 years after infusion. In a report²³  from a different trial published in 2014, 16 adult patients with relapsed or refractory B-cell ALL (B-ALL) were treated with a CAR-T composed of the anti-CD19 binding site fused to TCR coactivating intracellular domains CD28 and CD3-ζ chain (19–28ζ) in a γ retroviral vector. There was a complete response rate of 88%, allowing for 44% (7 of 16) of the patients in the study²³  to become eligible for allogeneic hematopoietic stem cell transplant (as compared with 5% of relapsed B-ALL patients historically). Anti-CD19 CAR-T therapy was shown to be effective in patients with Philadelphia chromosome–positive ALL and in patients who relapsed after an allogeneic stem cell transplant. A phase 1 trial, which included 53 adult patients with relapsed B-ALL, was conducted to assess long-term outcomes and toxicity profiles (published in 2018; patients enrolled from 2010 to 2016).²⁴  Complete remission was observed in 83% of patients, with a median overall survival of 12.9 months (95% CI, 8.7–23.4 months). Patients with lower disease burden (<5% bone marrow blasts) had median overall survival of 20.1 months (95% CI 8.7–not reached). Patients with higher disease burden were more likely to have shorter long-term survival. In these studies, highest levels of CAR-Ts appeared within 7 to 20 days after infusion, dependent on compartment (peripheral blood or bone marrow). An additional phase I dose-escalation trial in 21 children and young adults with relapsed or refractory ALL or non-Hodgkin lymphoma, published in 2015, was performed.²⁵  CD19 CAR-T therapy induced a complete response in 70% and a minimal residual disease–negative complete response in 60% of patients with B-ALL with reversible toxicities. It was proposed that this therapy could be an effective bridge to stem cell transplantation.

In 2017, significant responses were reported in 28 patients with refractory diffuse large B-cell or follicular lymphomas who received CD19-directed CAR-Ts (CTL019).²⁶  Complete remission occurred in 6 of 14 patients (43%; 95% CI, 18%–71%) with DLBCL and 10 of 14 patients (71%; 95% CI, 42%–92%) with follicular lymphoma. Eighty-six percent to 89% of responders showed a durable response. A multicenter phase 2 trial, published at the same time, also demonstrated a complete response rate of 54% and an objective response rate of 82% in 101 patients with refractory DLBCL, primary mediastinal lymphoma, or transformed follicular lymphoma.²⁷  Overall survival was 52% at 18 months, with 40% demonstrating a complete response.

At the time of this writing, CAR-T therapy has been approved by the FDA only for treatment of certain B-cell neoplasms. B-cell malignancies were the initial diseases targeted for approval, as much of the successful clinical trial work thus far has been with genetically modified CAR-Ts engineered against CD19, as discussed above.

Two CAR-T therapies have been approved by the FDA. The first approved CAR-T therapy is tisagenlecleucel-T (Kymriah, Novartis Pharmaceuticals).²⁸  This therapy is indicated for patients up to 25 years old who have B-ALL that has relapsed 2 or more times or is refractory to treatment. It has also more recently been approved for relapsed or refractory large B-cell lymphoma, including DLBCL, high-grade B-cell lymphoma, and DLBCL arising from follicular lymphoma, after 2 or more lines of systemic therapy.

The second approved CAR-T therapy is axicabtagene ciloleucel intravenous infusion (Yescarta, Kite Pharma)²⁹  for certain types of large B-cell lymphoma, including DLBCL, primary mediastinal large B-cell lymphoma, high-grade B-cell lymphoma, and DLBCL arising from a follicular lymphoma. It is not approved for primary central nervous system lymphoma. Patients receiving this therapy must have failed or not responded to at least 2 other treatment modalities.

There are potentially severe and life-threatening toxicities associated with CAR-T. In the initial B-ALL study²¹  that was published in 2013, in which 2 children with B-ALL were treated, both patients exhibited cytokine release syndrome (CRS) and B-cell aplasia. In the expanded study²²  published in 2014, the probability of relapse free B-cell aplasia was 73% and CRS occurred in all patients (CRS-related mortality was 0%). Patients with severe CRS (27%) were treated with anti–IL-6 therapy (tocilizumab; n = 9). In ALL, patients with higher disease burden were more likely to have CRS and neurotoxicity and shorter long-term survival.²⁴  In the 2017 study²⁶  of 28 patients with refractory diffuse large B-cell or follicular lymphomas who received CD19-directed CAR-Ts (CTL019), neurotoxicity (39%) and CRS (57%; 18% severe) were identified as adverse events, in addition to others involving the hematopoietic (neutropenia, anemia, etc), gastrointestinal (nausea, vomiting, diarrhea, etc), and generalized (electrolyte imbalances, fatigue, chills, etc) systems. Similar toxicities were identified in a separate multicenter study.²⁷ 

These toxicities are related to the specificity of the CAR itself and the normal cytokines that are released during T-cell–mediated cytotoxicity. For example, B-cell aplasia (destruction of normal CD19⁺ B cells) with resultant hypogammaglobulinemia may be seen with CD19-specific CARs.³⁰  As a result, CAR-T recipients may require short-term or long-term intravenous immunoglobulin therapy.³⁰  However, one study³¹  has suggested that immunoglobulins formed upon exposure to vaccination or pathogens prior to CAR-T therapy exist after CAR-T therapy and remain stable for at least 6 to 12 months posttreatment, with persisting antibody-secreting plasma cells for at least 25 months. This indicates the presence of sustained CD19⁻ plasma cell populations that can maintain at least preexisting humoral immunity after CAR-T therapy, and may reduce the need for intravenous immunoglobulin, at least in some CAR-T recipients.

The most severe side effect, CRS, results from a rapid, large release of cytokines within a day to days of the CAR-T infusion, and usually starts with a fever and subsequently hypotension, coagulopathy, and capillary leak. Severe CRS can resemble more serious sequelae such as hemophagocytic lymphohistiocytosis or macrophage activation syndrome.³²  In clinical trial studies for each of the FDA-approved therapies, there were very few deaths related to CRS. In a large study of tisagenlecleucel,²²  there were 0 deaths, and in a study of axicabtagene²⁷  in lymphoma, there was 1 death due to cardiac arrest and 1 death due to hemophagocytic syndrome, attributed to CRS. An updated grading system has been put forth³³  and includes monitoring for fever, hypotension, and hypoxia to simplify the grading.

Development of CRS demonstrates that the CAR-Ts are activated, inducing an inflammatory cascade eliciting CRS-related symptoms.³⁴  Cytokine release syndrome is seen more commonly and is more severe when patients have extensive disease leading to more activation of the CAR-Ts.³²  Treatment of CRS can be managed according to grade³⁵  using supportive care (anti-inflammatories, intravenous hydration, ventilatory support, and pressor support), in addition to immune-modifying medicinal support. Tocilizumab (anti–IL-6 receptor monoclonal antibody) is important in early CRS treatment for at least grades 3 to 4. Although there was initial concern over use of corticosteroids because of potential suppression of CAR-T activity, they have since been shown to have minimal effect on CAR-T proliferation/longevity and treatment outcomes.³⁶  Corticosteroid administration is now used for at least grade 3 to 4 CRS.³⁵ 

Neurotoxicity is another side effect that can manifest after CAR-T therapy. It appears to be highly correlated with CRS, yet apparently independent of CAR-T design.³⁷  Neurotoxicity has been associated with higher disease burden, higher levels of CAR-T expansion, early and higher elevations of inflammatory cytokines, high-intensity lymphodepletion, and/or preexisting endothelial cell activation. It has been related to inflammatory cytokine elevation in the cerebrospinal fluid without a corresponding increase in cerebrospinal fluid leukocytosis or CAR-T elevation.³⁸  Fever (≥38.9°C) and inflammatory cytokine elevations within 36 hours postinfusion (IL-6, interferon [IFN] γ, monocyte chemotactic protein [MCP] 1, IL-15, IL-2 and IL-10) were higher in patients who developed grade 4 neurotoxicity.³⁷  Neurotoxicity may present with visual hallucinations, aphasia, disorientation, unresponsiveness, encephalopathy, and pain, among other symptoms. Grading criteria, for what is termed immune effector cell–associated neurotoxicity syndrome, exist based on level of consciousness, seizure activity, motor weakness, cranial nerve palsies, Cushing triad, and intracranial pressure measurements.³⁵  Management includes associated symptomatic supportive care (intracranial pressure management, seizure therapy, benzodiazepine treatment, etc), IL-6 inhibition, and corticosteroids.³⁵ 

In the future, pathology laboratories may be asked to monitor the side effects of CAR-T therapy, although exactly how remains to be seen. Laboratories in settings where CAR-T is used should at least consider whether they will need to offer assays that detect various cytokines or inflammatory mediators. In particular, it has been suggested that monitoring for CRS by measurement of IL-6 and other cytokines may be useful clinically. However, CRS can happen fairly rapidly after CAR-T infusion (1–14 days). Current laboratory methodology for cytokine measurement is not widely available clinically, nor is it rapid enough to detect CRS with the necessary turnaround time, and there has not been demand for such assays to identify CRS. Therefore, clinical signs and laboratory monitoring for organ dysfunction are currently used.

Assessment of the presence of target antigen (eg, CD19) on tumor cells was performed in many of the CAR-T studies and is often requested in the clinical setting; however, testing is not required for determining the use of CAR-T therapy, and there are no established criteria that would change patient eligibility.³⁹  Testing for CD19 is often performed routinely by flow cytometry in the workup of B-cell neoplasms and therefore would not require extra technical work in this setting. However, there may be additional effort required to interpret the percentage of positive neoplastic cells (which is not routine practice in all laboratories), and in cases where flow cytometry was not performed, CD19 immunohistochemistry would be required, which also may not be routine. In addition to the lack of standard criteria, different methodology or reagents may give variable results (similar to issues with programmed death ligand-1 [PD-L1]). Percentage and intensity of expression will be affected by whether immunohistochemistry or flow cytometry is performed, the antibody clone used, and the conjugated fluorophore or dye, and could be affected by instrument settings and calibration (Sophia Yohe, MD, oral communication, November 2018).

As CAR-T therapy is designed to target a specific antigen, it makes intuitive sense that diminished or absent expression of that antigen will affect response to therapy. CAR-T relapse may be due to antigen-positive relapse, antigen loss (eg, due to alternative splicing of CD19 messenger RNA or disrupted CD19 transport to the cell surface) or immune escape.⁴⁰  Loss of CD19 is a frequent cause of relapse among B-ALL patients treated with CD19 CAR-T therapy.³⁹  CAR-T therapy can affect follow-up immunophenotyping of the disease, similar to the effect of rituximab on CD20. Flow cytometric immunophenotyping has been a critical part of patient follow-up after various types of therapies to determine whether there is residual disease, especially for B-ALL, and CD19 is typically used in flow cytometry as part of the gating strategy to identify neoplastic cells. Laboratories that routinely perform follow-up flow cytometry for patients receiving CD19-targeted CAR-T therapy need additional ways to identify these cells, which can include alternative gating strategies and use of additional B-cell markers. It should also be recognized that detecting low levels of disease may be more difficult and the sensitivities usually associated with flow cytometry may not be achieved for patients who have received CD19 CAR-Ts.

Currently, these issues apply only to CD19 expression on the B-cell neoplasm to be treated; however, as CAR-Ts targeting different antigens are designed, there will potentially be a demand to assess these new target antigens pretreatment and in follow-up studies. As therapies are combined (for example, CD19 CAR-T, anti-CD20 rituximab, anti-CD22 therapy), it may be difficult to identify neoplastic populations in follow-up specimens.

Flow cytometry, next-generation sequencing, and quantitative polymerase chain reaction are methods that can directly measure CAR-Ts. Next-generation sequencing and quantitative polymerase chain reaction require targeting aspects unique to the CAR-Ts that are not present in normal immune cells. Such targets include the viral vector or the specific hypervariable complementary determining region of the immunoglobulin portion of the CAR construct.^41,42  Flow cytometry may similarly target a specific antigen on the CAR-Ts, but more commonly uses antibodies to identify T cells in combination with an antibody against the antibody single chain variable fragment (often denoted as the F[ab]₂ fragment), as only CAR-Ts will coexpress this antibody fragment and T-cell markers.^42,43  Functional flow cytometry assays that measure cytotoxicity or cytokine production also exist; however, these are typically reserved for development of new CAR-T therapies and not routinely used clinically.

Even when a flow cytometry assay cannot specifically detect CAR-T–specific markers, the presence of CAR-Ts may still be detected through analysis of flow cytometry data. Unusual proportions of T cells may be present, CD4:CD8 ratios may be skewed, or an unusual immunophenotype may be detected if genetic engineering has introduced or removed an antigen. Examples could include: the presence of γ-δ CAR-Ts may result in an apparent increased population of CD4 and CD8 double-negative T cells; introduction of cytokine receptors (such as IL-2 receptor CD25) or other antigens would lead to their detection; and knockout of TCRs would lead to α-β– and γ-δ–negative T cells.^44–46 

College of American Pathologists committees are assessing the need for CAR-T–related proficiency testing. Currently, the Diagnostic Immunology and Flow Cytometry Committee has proficiency testing for CD19 assessment. This committee also developed supplemental questions regarding CAR-T therapies to accompany proficiency testing in a 2019 survey to assist in information gathering regarding CAR-T monitoring. College of American Pathologists members and committees should consider CAR-T therapy when introducing or modifying CAP accreditation checklist questions for each specialty within the clinical laboratory.

During preparation of this manuscript, workgroup members asked individuals from both industry and academia, who are involved with CAR-T therapy development, how clinical laboratories may support CAR-T therapy through diagnostics, including monitoring of CAR-T efficacy and persistence, B-cell aplasia, vector-cassette integration, and immune recognition. Thus far, there are no definitive answers. Part of the reason for the uncertainty in CAR-T–related diagnostic testing is that currently approved CAR-T therapies all use an autologous cell source, which is sent for CAR construct introduction by an outside company. However, there is increasing interest in allogeneic (off-the-shelf) CAR-T therapy because of increased accessibility, improved quality control, and better efficacy through the use of T cells from a healthy donor.⁴⁷  Barriers to the use of allogeneic T cells include graft-versus-host disease and rejection, similar to hematopoietic stem cell transplantation. One can envision that testing by specialized flow cytometry laboratories, histocompatibility laboratories, transfusion medicine facilities, and molecular diagnostic laboratories will be required in order to monitor the efficacy and safety of allogeneic CAR-T therapy, as well as host immune responses to the allogeneic CAR-T product.

Administration of CAR-T therapy is complex, involving multiple steps and entities within and outside the hospital. Currently, if an institution desires to treat its patients with CAR-T therapy, the CAR-T manufacturer (as of this writing, usually a pharmaceutical company) performs an on-site evaluation and inspection to verify that all required activities to support the particular CAR-T product are in place. These include documented protocols for physician (eg, pathologist) harvesting of blood-derived T lymphocytes for development of genetically modified autologous CAR-Ts, as well as cell processing, cryopreservation, and preparation for shipping to the entity that will manufacture the CAR-T therapeutic product. Health care institutions that are currently offering or wish to offer commercial CAR-T therapies will be expected to adhere to good manufacturing practice, including those specialized facilities (which may or may not be affiliated with the institution's pathology laboratory) for cell and product handling. Specialized protocols for each institution are written that cover all the steps of mononuclear cell collection, cell counting, sterility assessment, viability assessment (by flow cytometry), identity verification (by blood typing), and storage. In order to offer FDA-approved CAR-T products, an institution must maintain records and multiple forms and updates must be filed to document product tracking/chain of custody. After submission of the mononuclear cells to the manufacturer, it takes approximately 2 to 3 weeks for the company to manufacture the CAR-Ts by introduction of the vector carrying the anti-CD19 CAR molecule and expansion of the CAR-Ts. The anti-CD19 CAR is transduced into the cells using a lentiviral vector for Kymriah and a γ-retroviral construct for Yescarta. After manufacturing, CAR-T products undergo quality checks that include visual inspection; testing to ensure the CAR is present and/or functional; viable cell count and dosing information; purity, including measures of contaminating cells; and measures of sterility. The final cryopreserved product is shipped to the site that will administer the therapy in an infusion setting within the hospital.^48,49 

Jurisdiction over accreditation and administration of CAR-T can vary within hospital settings. Directors of good manufacturing practice/good tissue practice facilities may be derived from pathology, transfusion medicine, or oncology service lines, depending on the institution. In some institutions, the oncology section may oversee the laboratory processes involved in CAR-T administration. Additionally, as the indications for these therapies can be either lymphoma or leukemia, there may be different subspecialty service lines that may oversee the same therapy administration, depending on the indication for CAR-T administration.

Serious adverse events must be reported to the company and relayed to the FDA. The CAR-T products are administered under the FDA's Risk Evaluation and Mitigation Strategies program, which is required for therapies with serious safety concerns, and is specific to the safety of the licensed product itself and not to the individual institutions or programs that partake in the administration. As FDA regulation in this area continues to evolve,⁵⁰  most institutions involved with CAR-T therapy do so under standards from the Foundation for the Accreditation of Cellular Therapy (FACT), which serves as the nonprofit accreditation body for the American Society for Blood and Marrow Transplantation. Termed the FACT Standards for Immune Effector Cells,⁵¹  the standards define an immune effector cell as “a cell that has differentiated into a form capable of modulating or effecting a specific immune response.” The FACT standards were created at the request of multiple stakeholders from industry, clinical programs, regulators, and payers. The FACT does not oversee the development of CAR-T therapy with regard to therapeutic efficacy, clinical utility, or clinical validity, but instead the safe administration of the products with stringent quality control. As of June 2018, there were 31 accredited immune effector cell programs with 25 additionally scheduled accreditation inspections, not all in academic settings.

In the clinical trial setting, under the FDA investigational new drug designation for CAR-T clinical trials, serious adverse events must be reported to both the FDA and the relevant institutional review board. The FDA has released a draft guidance for industry (July 2018) for “Long Term Follow-up After Administration of Human Gene Therapy Products.”⁵²  In this document, human gene therapy product is defined “as all products that mediate their effect by transcription or translation of transferred genetic material, or by specifically altering host (human) genetic sequences.” This includes ex vivo genetically modified human cells and would pertain to CAR-T therapy. This guidance, when finalized, will address criteria to assess potential delayed risks of gene therapy in clinical trials, including biodistribution and persistence of gene therapy products, vector persistence, integration activation and genome modifications, and preclinical evaluation of products. It will also address how long-term clinical follow-up is to be performed by industry. One such issue that highlights important considerations of viral vector delivery in patients is interference of the vectors with clinical laboratory testing. This has been illustrated in a series of 4 patients treated with lentivirus-based tisagenlecleucel who had false-positive HIV-1 nucleic acid testing–based results after receipt of CAR-T therapy.⁵³ 

The current regulatory structure relevant to CAR-T therapies is diagrammed in Figure 3. Much of the regulation relevant to production of CAR-T products falls under the FDA's Center for Biologics and Evaluation Research. In addition, clinical trials with emerging (non–FDA approved) CAR-T therapies in which CAR-T production occurs within a local good manufacturing practice facility may be regulated under FDA investigational new drug and FACT standards. Although laboratory testing in support of CAR-T therapy performed in Clinical Laboratory Improvement Amendments–certified laboratories currently falls under Centers for Medicare and Medicaid Services (CMS) regulations, the FDA may consider certain test systems to fall under the medical device regulations. Therefore, close monitoring of activities and decisions surrounding genetically modified biologic products seems prudent.

It is conceivable that the CAP Laboratory Accreditation Program may have a role in accrediting some of those laboratories that will be processing and testing clinical specimens from patients undergoing CAR-T therapy. Partnerships between different accrediting agencies will help to ensure consistency of standards across the CAR-T procurement and delivery process. For example, the CAP has been approved by the FACT as an accrediting organization providing histocompatibility services appropriate for hematopoietic cellular therapy transplant.

College of American Pathologists advocacy is continuously evaluating and actively monitoring how the CMS is developing coverage and reimbursement policies for CAR-T products. As described above, CAR-T therapies are now FDA approved and are increasingly used for patient care. A diagram of the CAP's actions surrounding CAR-T therapies is shown in Figure 4. Until recently, CAR-T therapies were available only through clinical trials, and reimbursement for some biological gene products/therapeutic agents was tied to Health Common Procedure Coding System level II Q codes developed by the CMS. In June 2018, the American Medical Association Current Procedural Terminology (CPT) Editorial Panel accepted a multispecialty society proposal (including the CAP) to establish 4 new category III CPT codes related to CAR-T therapy. These codes capture physician services and other CAR-T–related services beyond the biological gene product/therapeutic agent and are effective January 1, 2019 (Figure 5).⁵⁴  The associated descriptions of procedure for each of these codes were carefully vetted by the CAP Economic Affairs Committee to ensure that all pathology-related services were specifically identified and recognized. Presumably, as CAR-T therapies are more widely adopted, these CPT category III codes may be considered by the American Medical Association CPT editorial panel for conversion to new category I CPT codes, which would be advanced for appropriate Relative Value Scale Update Committee and CMS review and valuation of services provided by pathologists or other physicians.

In May 2018, the CMS initiated a national coverage determination analysis for CAR-T therapies. The CAP submitted comments to the CMS emphasizing that:

Noting that approaching CAR-T therapy reimbursement as if it were a drug therapy was inappropriate, the CAP urged the CMS to pursue established mechanisms for recognizing physician services and facility reimbursement through the development of American Medical Association CPT codes and/or Health Common Procedure Coding System level II G codes and subsequent valuation.

On August 22, 2018, the CMS convened a meeting of the Medicare Evidence Development & Coverage Advisory Committee (MEDCAC), which advises the CMS on national coverage issues.⁵⁵  The Medicare Evidence Development & Coverage Advisory Committee does not make coverage determinations, but reviews the state of evidence and makes recommendations to the CMS, typically by voting on a series of questions posed to the panel by the CMS. Specifically, the MEDCAC Panel was asked to assess whether scientific evidence supports a specific number of outcome assessments, study design characteristics, study duration, and suitable controls for applying patient-reported outcomes (PROs) to health outcomes research as applicable to CAR-T therapy. The meeting also explored the challenges regarding the validity, reliability, and generalizability of existing PRO assessments. The committee panel heard from CAR-T therapy drug makers, health researchers, and policy makers. Presentations before the panel included a review of trial data on existing CAR-T treatments and the PROs under consideration. Many of the presenters expressed concern that PROs are not yet ready for real-world coverage decisions because PRO data relating to CAR-T therapy, although valuable, can be difficult to understand for both patients and providers, owing to high variability, and can be challenging to interpret. There is no standardization in how they are scaled, scored, or presented. In response to the questions posed by the CMS, the MEDCAC Panel mostly endorsed PROs for CAR-T therapies and expressed interest in seeing their further development but urged caution to ensure that patient concerns about CAR-T therapies are heeded. The CAP administered a comment letter to the CMS in March 2019 requesting (1) a flexible process for routinely extending coverage as newer CAR therapies become available for reasonable patient access, (2) recognition for all provider services as separate and distinct processes from the manufacturing step (harvesting, transportation, storage, administration, etc), and (3) allowance of Medicare Administrative Contractors to determine coverage for new therapies/technologies at the local level as they are available.

Because of the CAP's advocacy, the CMS will increase reimbursements for CAR-T therapies in 2020. On August 2, 2019, the CMS released the 2020 Hospital Inpatient Prospective Payment System final regulation, which included a new technology add-on payment for CAR-T therapy. The CAP advocated for increasing the add-on payment of the new technology, which brings the maximum add-on payment for CAR T-cell therapies to $242 450, up from $186 500.⁸⁹  In its final coverage regulation published August 7, 2019, the CMS announced it will cover CAR-T therapies nationwide when administered in inpatient or outpatient facilities that are enrolled in the FDA's Risk Evaluation and Mitigation Strategies drug safety program for FDA-approved indications. Medicare also covers approved CAR-T therapies for off-label uses that are recommended by CMS-approved compendia. Additionally, the agency agreed not to proceed with coverage with evidence development, a model that collects additional data and allows for coverage only within the context of a research setting.

The list price of CAR-T therapy is $373 000 to treat lymphoma and $425 000 for B-ALL.⁵⁶  Not included in this figure are supportive care measures related to therapy administration and aforementioned side effects. Additionally, because of the relatively rapid approval, long-term costs and survival benefits are not fully understood. One study⁵⁷  suggested an estimated lifetime cost of $553 000 for treatment with axicabtagene ciloleucel for lymphoma (compared with $173 000 for chemotherapy). Economic modeling studies^57,58  have been done in B-cell lymphoma that suggest that CAR-T therapy, although more expensive than chemotherapy, improves life years and quality-adjusted life years, with cost-effectiveness per quality-adjusted life year of CAR-T associated with long-term survival. In an economic study⁵⁹  in B-ALL, tisagenlecleucel also showed gain in quality-adjusted life years over clofarabine (9.28 versus 2.10 quality-adjusted life years) associated with gains in survival, which suggested that the price was in alignment with the benefit of the therapy.

CAR-T therapy is in clinical trials for other hematologic diseases. CD19-targeted CAR-T therapy has induced sustained remission in refractory chronic lymphocytic leukemia with an overall response rate of 57% in one study⁶⁰  of 14 patients, and no relapse in patients with a complete response. Interestingly, a recent study⁶¹  showed that chronic lymphocytic leukemia patients with “healthier” early memory T cells (as measured by CD8 and CD27 expression and absence of CD45RO) were more likely to respond. In multiple myeloma, CD19 is typically absent; however smaller, less-differentiated myeloma propagating subsets may have CD19 expression. A subset of such patients with refractory multiple myeloma treated previously with autologous hematopoietic cell transplant were treated with anti-CD19 CAR-T with concomitant melphalan and autologous hematopoietic cell transplant. Twenty percent of patients demonstrated superior progression-free survival compared with previous autologous hematopoietic cell transplant alone.⁶²  In a study⁶³  of 16 patients with refractory multiple myeloma receiving CAR-Ts (with preconditioning chemotherapy), transduced with a retroviral vector encoding CAR–anti-BCMA (B-cell maturation antigen, expressed on multiple myeloma cells and normal plasma cells), there was an overall response rate of 81%. Preclinical evidence in cell line and mouse work suggests that CAR-Ts targeting FLT3 may be effective in treating acute myeloid leukemia.^64,65 

Whereas there have been many studies investigating CAR-T therapy in various solid tumor malignancies (for some examples, see the Table), CAR-T therapy of solid tumors has been less successful.⁶⁶  This is due, in part, to the lack of specific targetable antigens (eg, various cytokeratins are found on multiple epithelial cell types) and difficulty in overcoming the protective barriers of solid tumor cell malignancies. These barriers include fibrosis and extracellular matrix, immune suppression and cytokine/chemokine imbalance favoring the tumor, tumor endothelial barriers preventing extravasation of T cells, and poor metabolic and hypoxic states inhibiting CAR-T growth. Manipulation of CAR-T therapy (such as CARs engineered to produce IL-12 to mitigate the tumor microenvironment—so-called armored CARs) may be necessary to overcome some of these barriers.

Currently approved CAR-T therapies are expensive ($400 000–500 000 per dose) and require centralized manufacture and an extensive infrastructure in specialized tertiary care centers. As the safety of CAR-T therapies improve, these treatments may be offered to patients in community hospitals or even outpatient clinics. To enable CAR-T therapy at reduced cost, regional or local CAR-T manufacturing is being explored, such as producing CAR-Ts using the CliniMACS Prodigy as a closed-system processor for enrichment, transduction, washing and expansion.^67,68  This system allows for a relatively short processing time (13–14 days) to create sufficient functional CAR-Ts without the infrastructure required for centralized manufacturing.

Currently, the FDA-approved therapies require removal of autologous T cells from the patient, shipping to the manufacturer for CAR-T production, and return of the cells to the patient. This process typically requires at least 2 weeks for a product to be returned to the patient. In some instances, the patient's T cells may not be the best starting material for CAR-T therapy because of prior exposure to chemotherapy or other treatments. As a result, companies are developing off-the-shelf allogeneic CAR-T therapies in which T cells from other individuals could be manufactured with CARs and be given as a universal cellular product, eliminating the need for patient cell collection and the lengthy manufacturing time for autologous CAR-T preparation.⁶⁹  Barriers to this approach could include rejection of the infused product by the host immune system and the risk of graft-versus-host disease. Whereas α-β T cells have been primarily used for CAR-T construction, γ-δ T cells are currently being examined as a way to prevent graft-versus-host disease development. Removal of major histocompatibility complex class I proteins from donor T cells is also being considered to decrease the chance of rejection of the allogeneic CAR-T product. However, this in turn would also remove the ligand for killer immunoglobulin-like receptor (KIR), which normally inhibits natural killer (NK) cells, allowing for NK-mediated cell killing of infused cells. Trials are also underway to manipulate a patient's hematopoietic stem cell transplantation donor T cells to be infused back into the recipient to improve killing of the disease.⁷⁰ 

One of the first allogeneic CAR-T therapy clinical trials is underway for unresectable colorectal cancer. In this trial, allogeneic CAR-Ts are engineered to express a chimeric receptor based on NK group 2D (NKG2D), a receptor normally expressed on NK cells and certain T-cell populations, that also recognizes multiple expressed tumor proteins.^71,72 

Autologous NK cells have shown limited efficacy for autologous cell killing. However, alloreactive NK cells are being studied as an off-the-shelf cell type that can use a CAR.⁴  Natural killer cells mediate anticancer effects, but should not induce graft-versus-host disease or autoimmune toxicity. In the allogeneic setting, they can be rapidly available (off the shelf) and have a short-term lifespan in vivo, reducing the potential for long-term adverse events (cytopenias, aplasia). However, they can also be difficult to obtain and harvest, can be difficult to transduce, may be infused with contaminating T or B cells, and may require cytokine support for persistence (IL-2 or IL-15), leading to unwanted side effects.⁴  Although NK cells can be obtained from peripheral blood, bone marrow, or stem cells, cord blood has been shown to be a good source of allo-NK cells (15%–20% NK cells) with tolerance to human leukocyte antigen mismatches and better transduction efficiency.⁴  Incorporation of cytokine genes (IL2 or IL15) to assist in NK cytokine support and and/or suicide genes (such as inducible caspase, iCasp9), to inactivate unexpectedly developed toxicity, are also being studied in order to improve therapeutic efficacy and reduce potential harm.⁴ 

CAR-T therapy is rapidly becoming available to treat various hematologic malignancies at many larger health care institutions. The expansion of clinical trials in the use of CAR-T therapy is ongoing, and additional indications for CAR-T therapy in oncology are currently being pursued. Multiple avenues are being considered to improve the use of CAR-T or alternative CAR cell therapy. However, CAR-based cellular therapies are expensive and not without serious side effects. Currently, the administration of CAR-T is complex, requiring close coordination among multiple hospital services and pharmaceutical companies. There are still many questions as to the roles pathologists and clinical laboratories will play in CAR-T production, administration, and monitoring of cell persistence and adverse events. As diagnosticians and laboratorians charged with assay development and quality assurance oversight in cellular therapy administration and disease monitoring, it is important that pathologists be included as a stakeholder in CAR-T therapy to ensure that the patients receiving these therapies achieve the best possible outcome.