Objectives
This is a protocol for a Cochrane Review (prognosis). The objectives are as follows:
To identify and meta‐analyse prognostic factors for efficacy and safety of CAR T‐cell therapy for treating aggressive large B‐cell lymphomas before therapy has started.
| Population |
|
| Index factor(s) |
|
| Comparator(s) |
|
| Outcome(s) | Efficacy outcomes
Safety outcomes
|
| Timing | Timing of prediction
Prediction horizon
|
| Setting |
|
| CAR: chimeric antigen receptor; CRS: cytokine release syndrome; DLBCL: diffuse large B‐cell lymphoma; FL: follicular lymphoma; HGBCL: high‐grade B‐cell lymphoma; ICANS: immune effector cell‐associated neurotoxicity syndrome; PMBCL: primary mediastinal large B‐cell lymphoma | |
Background
Description of the health condition and context
Aggressive large B‐cell lymphomas
Aggressive large B‐cell lymphomas are a category of fast‐growing lymphomas that affect mature B cells. B cells, a type of white blood cell, are part of the adaptive immune system. The malignant cells of aggressive B‐cell lymphomas present as large neoplastic cells, which may spread at early stages throughout the body (Swerdlow 2017). When left untreated they are fatal, and thus, usually need intensive treatment.
With approximately one‐third of cases, diffuse large B‐cell lymphoma (DLBCL) is the most common form of non‐Hodgkin lymphoma (Swerdlow 2017), and the most common aggressive B‐cell lymphoma. It has an age‐adjusted incidence rate of 5.3 per 100,000 population in the USA (between 2008 and 2017) with a five‐year relative survival across all ages of 63.8% (2008 to 2016; Howlader 2020). Mediastinal large B‐cell lymphoma has a better prognosis than DLBCL not otherwise specified, and other subtypes. The neoplasm is more common in men than women and affects mostly the elderly (the median age of diagnosis is 66 years). However, it occasionally occurs in younger adults and even children (Howlader 2020).
The tumour may initially present in the lymph nodes or extranodal sites (in 40% of cases); for example, in the gastrointestinal tract, bone, testes, spleen, Waldeyer ring, salivary glands, thyroid, liver, kidneys, and adrenal glands. DLBCL often presents as asymptomatic or with unspecific symptoms and swelling of affected lymph nodes or spleen. Typical B symptoms such as fever, night sweats, and weight reduction can occur but are infrequent (Swerdlow 2017). DLBCL is staged with the Lugano Classification into four tumour stages based on physical exams, imaging (e.g. positron emission (PET) or computed tomography (CT)), biopsy, blood tests, or bone marrow biopsy. Stage I is assigned when only one lymph node or lymphoid organ is affected, and stage II describes two or more affected lymph node areas on the same side of the diaphragm; both are considered limited stages. The label E is added for an affected non‐lymphoid organ. Stage III describes lymphoma in lymph node areas both above and below the diaphragm or the spleen, and stage IV is assigned if multiple lymph nodes and other body parts such as lungs, bones, or liver are affected (Cheson 2014).
For this review, we will also include primary mediastinal large B‐cell lymphoma (PMBCL), follicular lymphoma (FL) grade 3b, and high‐grade B‐cell lymphoma (HGBCL) in adults.
CAR T‐cell therapy
The usual treatment for aggressive large B‐cell lymphomas is a chemotherapy regimen (cyclophosphamide, doxorubicin, vincristine, and prednisone; 'CHOP') in combination with a monoclonal antibody (rituximab, polatuzumab vedotin) and, if needed, radiotherapy (Kubuschok 2015; Pfreundschuh 2006). After first‐line treatment, approximately 10% to 15% of patients will relapse within one year, and 30% to 40% will relapse overall, resulting in a considerably worse prognosis. In such cases, high‐dose chemotherapy followed by autologous stem cell transplantation (ASCT) is the therapy of choice, but individuals must meet the eligibility requirements (age < 65 years without major comorbidities) and respond to a conditioning chemotherapy regimen. Hence, for more than 60% of this group, other options such as dose or regimen alterations of chemotherapy have to be considered (Sarkozy 2018).
One of the new treatment options that has shown initial encouraging results is chimeric antigen receptor (CAR) T‐cell therapy (Neelapu 2017; Schuster 2018), a therapy based on genetically engineered T‐cells harvested from the individual to be treated, or in exceptional cases, a healthy donor. Outside the body, T‐cells are equipped with CARs, synthetic receptors consisting of an antigen binding domain, the hinge region, a transmembrane domain, and intracellular signalling domains (Sterner 2021; St‐Pierre 2022). The antigen binding domain redirects the CAR‐equipped T‐cells to recognise cancer cells that express a specific target antigen, which for B‐cells typically is the target cluster of differentiation (CD) 19. The antigen‐binding domain is connected via a spacer domain to the transmembrane domain. The intracellular signalling domain is the T‐cell activating region, consisting of a CD3 signalling domain and an additional co‐stimulatory domain in second‐generation CAR T‐cells, and is responsible for an effective T‐cell response and persistence (Sterner 2021). After reinfusion to the patient, the cells bind to tumour cells and are activated, which leads to natural multiplication and cytotoxic action against the cancer cells. Even after the elimination of the cancer cells, CAR T‐cells may persist and reactivate upon tumour recurrence (Hartmann 2017).
Multiple challenges are associated with CAR T‐cell therapy. Cytokine release syndrome (CRS) and immune effector cell‐associated neurotoxicity syndrome (ICANS) constitute particularly characteristic and potentially life‐threatening toxicities associated with CAR T‐cell therapy. Also related to the CAR T‐cells' target‐receptor are antigen escape and 'on target‐off tumour effects'. Antigen escape refers to the situation where the target antigen is no longer expressed on cancerous cells, which, over time, can lead to therapy‐resistant disease. 'On target‐off tumour effects' describes the issue that the target antigen can also be presented on non‐malignant, physiological tissue cells, which results in adverse events, although primarily relevant for solid tumours. (Sterner 2021).
In a recent report prepared for the application of CD19‐directed CAR T‐cell therapy for the World Health Organization (WHO) Model List of Essential Medicines 2023, CAR T‐cell therapy in early relapsed or primary refractory large B‐cell lymphoma was systematically reviewed and evaluated (Csenar 2022). The authors identified three multicentric randomised controlled trials (RCTs) reporting data on 865 participants (Bishop 2022; Kamdar 2022; ZUMA‐7). The evidence suggests that CAR T‐cell therapy likely improves progression‐free survival, may increase event‐free survival, likely increases overall response rate, and may improve overall survival when compared to standard care including ASCT. However, the data are still considered immature, with results based on interim analyses. In terms of safety, CAR T‐cell therapy may result in little to no difference in the occurrence of serious adverse events compared with ASCT, but patients treated with CAR T‐cells can experience potentially life‐threatening adverse events such as CRS or ICANS.
Description of the prognostic factors
For this systematic review, we are interested in prognostic factors, thus characteristics that are associated with changes in the outcome in participants with aggressive large B‐cell lymphoma receiving CD19‐directed CAR T‐cell therapy. These characteristics, which may identify those patients with aggressive large B‐cell lymphomas who are likely to benefit from CAR T‐cell therapy or those who are highly at risk for treatment failure or severe toxicities may, at a later stage, inform trials that explore treatment adaptation for individual patients (predictive factor research). Currently, we aim to identify the patients who benefit most, and thus, spare patients with potentially severe, life‐threatening adverse events with CAR T‐cell therapy without anticipated benefit. Several studies have explored whether specific patient characteristics or biomarkers can predict the response to CAR T‐cell therapy (Kittai 2021; Nastoupil 2020; Vercellino 2020). These characteristics, when used to predict an outcome, are termed prognostic factors. Some prognostic factors that have been explored in individual studies are, for example, comorbidities, age (Kittai 2021), sex, Eastern Cooperative Oncology Group (ECOG) performance status (Oken 1982), the baseline total metabolic tumour volume (TMTV), elevated lactate dehydrogenase (LDH), number of extranodal sites, high International Prognostic Index (IPI; Nastoupil 2020; Vercellino 2020; Westin 2019), and other biomarkers (e.g. Schuster 2018).
For this systematic review, we are interested in prognostic factors that can be assessed at any time point before or at the start of CAR T‐cell therapy, to avoid a high‐priced form of therapy, where a lack of response or severe adverse events are possible. As the effect may vary between different CAR T‐cell products and medical conditions, we will explore these in a subgroup analysis.
Health outcomes
The most relevant health outcomes are, on the one hand, efficacy outcomes such as overall survival, response outcomes (complete response, partial response), treatment‐related mortality, progression‐free survival, and quality of life, and on the other hand, the most severe safety outcomes, which are ICANS and CRS as well as immune parameters such as grade 3 to 4 infections, B‐cell aplasia, and hypogammaglobulinaemia. For this review, we will consider any available follow‐up time horizon.
Why it is important to do this review
The WHO Cancer Medicines Working Group for the Model Lists of Essential Medicines recognises the high therapeutic relevance of CAR T‐cell therapy and agrees that the evidence base on this novel therapy should continue to be monitored (WHO 2023). As CAR T‐cell therapy may lead to severe and potentially life‐threatening adverse events, such as CRS, neurotoxicity (CAR T‐cell therapy‐related encephalopathy syndrome, including for example, encephalopathy, delirium, coma, somnolence, agitation, and more), cytopenias, and infections, among others for certain groups of individuals (Abramson 2020; Schuster 2018), it is essential to anticipate who may benefit and who may be at high risk of developing severe adverse events. The limitations of the available evidence on the efficacy of CAR T‐cell therapy, the economic costs with a potentially high burden on healthcare systems, and the complexity of the preparation of and treatment with CAR T‐cells underline this need to understand the prognosis of aggressive B‐cell lymphomas after CAR T‐cell therapy in adults.
Multiple studies have investigated the association between several different factors and clinical outcomes after CAR T‐cell therapy. However, these have not yet been systematically summarised using a systematic review. Complications are that often multiple prognostic factors are analysed simultaneously and tested hypothesis‐free (Riley 2013). Furthermore, significant results are more likely to be reported, and gain visibility and further citation. It is therefore important to collect information on the replicability of previously reported prognostic factors and create an overview of all analyses, including a prognostic factor.
Objectives
To identify and meta‐analyse prognostic factors for efficacy and safety of CAR T‐cell therapy for treating aggressive large B‐cell lymphomas before therapy has started.
| Population |
|
| Index factor(s) |
|
| Comparator(s) |
|
| Outcome(s) | Efficacy outcomes
Safety outcomes
|
| Timing | Timing of prediction
Prediction horizon
|
| Setting |
|
| CAR: chimeric antigen receptor; CRS: cytokine release syndrome; DLBCL: diffuse large B‐cell lymphoma; FL: follicular lymphoma; HGBCL: high‐grade B‐cell lymphoma; ICANS: immune effector cell‐associated neurotoxicity syndrome; PMBCL: primary mediastinal large B‐cell lymphoma | |
Methods
Criteria for considering studies for this review
Types of studies
Study designs
To assess the association between a prognostic factor before the start of treatment and an outcome, any cohort of at least 20 patients with a confirmed diagnosis of a relapsed or refractory aggressive large B‐cell lymphoma who receive CAR T‐cell therapy is appropriate for inclusion without the need for a comparator arm. We set the cut‐off at 20 participants, which indicates a maximum of 10 participants per factor level, because effect estimates may not be meaningful below this number.
The following study designs are eligible for inclusion:
RCTs, CAR T‐cell arm only; and
prospective or retrospective cohort studies.
We will exclude case reports, case series, and case‐control or case‐only studies, as they cannot provide accurate effect estimates of factor‐outcome relationships, due to selective sampling based on the outcome.
Publication formats
We will include the following publication formats:
full‐text publications; and
conference abstracts and preprints, if they provide sufficient information for analysis, otherwise we will list them as 'studies awaiting classification'.
Targeted population
We will include studies on adults with a confirmed diagnosis of a relapsed or refractory aggressive large B‐cell lymphoma, specifically, DLBCL, PMBCL, FL grade 3b, and HGBCL, who received CD19‐directed CAR T‐cell therapy, regardless of sex, ethnicity, or previous treatment. Mixed populations including other cancers or haematological malignancies are eligible if the study authors provide data separately for participants with relapsed or refractory aggressive large B‐cell lymphomas, or more than 80% of participants fit this inclusion criterion. We chose this cut‐off to be more inclusive, as many studies on CAR T‐cell therapy include B‐cell lymphomas that are not the target of the current review, but still obtain a homogeneous sample.
Studies conducted in any setting are eligible for inclusion.
Types of prognostic factors
We will include all prognostic factors that were explored concerning one of our predefined outcomes in the population of interest; that is, adult patients with aggressive large B‐cell lymphomas who received CAR T‐cell therapy.
Types of outcomes to be predicted
We will focus on prognostic factors associated with the following outcomes.
Efficacy
Overall survival is defined as the time from the start of the prediction up to the last follow‐up that a person is alive, considering events such as study dropout or closure as a censoring event.
Response, usually assessed shortly following CAR T‐cell therapy by Lugano Classification (Cheson 2014; Younes 2017), by using fluorodeoxyglucose (FDG)‐PET/CT.
Treatment‐related mortality is defined as death due to treatment toxicity, usually measured as part of serious adverse event monitoring, which is rated for its potential relationship with treatment in clinical trials.
Progression‐free survival is defined as time point of prediction to the date of documentation of progression or death from any cause, whichever occurs first. Patients alive and free of progression are censored at the date of the last follow‐up.
Quality of life, a concept to measure the well‐being of a person; measured with scales such as the European Organisation for Research and Treatment of Cancer Quality of Life (EORTC‐QL) questionnaires (EORTC 2023). We do not limit the tools used to assess quality of life and will summarise data as standardised mean differences where necessary.
Safety outcomes
ICANS, defined as grade 3 to 4 neurotoxicity, consists of a range of symptoms such as encephalopathy, delirium, headache, dizziness, aphasia, ataxia, motor dysfunction, amongst others (Castaneda‐Puglianini 2021). In trials, these symptoms are collected via case report forms.
Drug treatment for ICANS is defined as dexamethasone, prendis(ol)on, methylprednisolone as a treatment for ICANS.
CRS is a combination of symptoms such as fevers, hypotension, hypoxemia, tachycardia and organ dysfunction (Castaneda‐Puglianini 2021).
Drug treatment for CRS, defined as tocilizumab to treat CRS
Infections, grade 3‐4
B‐cell aplasia
Prolonged neutropenia grade 3‐4 (> 6 months)
Hypogammaglobulinaemia
ICANS and CRS are frequent CAR T‐cell therapy‐associated adverse events. They may result in prolonged hospitalisation or death, and thus, are highly relevant to people potentially receiving this treatment. The definition of ICANS and CRS may pose difficulties in comparability across studies, as ICANS and CRS can be measured using various scales that may not be mappable to each other. Therefore, we will interpret the use of medications for the treatment of these adverse events as surrogates (i.e. dexamethasone, prendis(ol)on, methylprednisolone for ICANS, and tocilizumab for CRS; Santomasso 2019).
Timing
We are interested in factors that predict treatment outcome events at the time of the decision to treat with CAR T‐cell treatment and at the time of treatment. CAR T‐cell treatment is a personalised therapy that has to be produced individually for each patient, therefore, the time between decision and therapy can be around 41 days (range 36 to 48 days; Le Gouill 2021). At any of these time points, the decision for CAR T‐cell therapy could still be withdrawn in order to avoid severe adverse events and high costs.
We will not limit inclusion to a specific prediction horizon. Instead, we will include factors that predict our planned outcomes at any future time point. However, longer follow‐up, especially for overall survival and progression‐free survival, is more informative, and will therefore be preferred. Where necessary (i.e. for dichotomous and continuous outcomes), we will classify outcomes into short‐term (up to 3 months), medium‐term (3 to 12 months) and long‐term (after 12 months) follow‐ups.
Search methods for identification of studies
Electronic searches
We will combine the search terms for aggressive large B‐cell lymphomas, CAR T‐cell therapy, and a filter for prognostic factors (Geersing 2012), terms for outcomes, the Cochrane filter for RCTs (Lefebvre 2022) and Phase III Filter (Cooper 2019).
We will not apply any restrictions regarding language. However, the search strategy will be developed in English, based on the assumption that at least the abstracts of studies in this emerging field will be available in English. We will involve Cochrane Engage for translations of full‐text publications in languages that lie outside the skill set of our team.
The first identified clinical trials in a systematic review by Ernst 2021 started in 2014. However, we will search the sources below from 2010 onwards, to capture registered and published trials.
-
Databases of medical literature:
Central Register of Controlled Trials (CENTRAL) (via the Cochrane Library) (from 2010 onwards);
MEDLINE (via OVID) (from 2010 onwards); and
Embase (via OVID) (from 2010 onwards).
-
Preprint server:
MedRxiv and BioRxiv (medrxiv.org).
-
Clinical trials registries:
Clinicaltrials.gov (clinicaltrials.gov); and
WHO International Clinical Trials registry Platform (ICTRP) (trialsearch.who.int).
-
Conference proceedings (the most recent volumes, if not included in CENTRAL):
American Society of Hematology;
American Society of Clinical Oncology;
European Hematology Association; and
European Society for Medical Oncology.
The search strategy for MEDLINE can be found in Appendix 1.
Searching other resources
Handsearching (reference list of included studies, relevant review articles, citing articles, and current treatment guidelines).
Data collection
Selection of studies
Two review authors will independently screen the records retrieved by the electronic search strategy in a two‐stage process for eligibility. First, we will screen titles and abstracts, and obtain eligible full texts for further, more detailed examination. We will solve disagreements by discussion. If the disagreement persists, we will add the record to the second stage, the full‐text screening. We will pilot our eligibility criteria by involving three review authors for screening and discussion of at least 100 records.
We will screen preselected full‐text publications independently and in duplicate. We will solve disagreements by consulting at least one additional review author. We will document the process in a PRISMA flow diagram, which shows the number of records or studies at each step (Page 2021). We will list those studies that may be expected to be included in this review but that we exclude after full‐text screening in the 'Characteristics of excluded studies', with their respective reason for exclusion.
Data extraction and management
Two experienced review authors will extract data from the eligible studies independently and in duplicate, according to Cochrane methods (Li 2023), into a standardised and piloted data extraction form. We will solve discrepancies by discussion or by involving a third review author. We will extract the following items.
Study level
General information: author, publication year, study identifier, country, language
Study characteristics: study design, recruitment method, dates, setting, eligibility criteria, and handling of missing data
Participants: age, sex, ECOG performance status, number of participants, participants lost to follow‐up, condition, disease stage, extranodal sites, LDH, number and type of previous treatments, and additional comorbidities
Treatment characteristics: type of CAR T‐cell therapy, duration, concomitant therapy
Factor/outcome level
Prognostic factor: prognostic factor, factor definition, factor measurement tool/method if applicable, type of factor (binary, continuous, dichotomised) and cut‐off when dichotomised, the timing of factor measurement
Outcomes: outcome, outcome definition, method of outcome measurement, time point/duration of outcome measurement (eligible outcomes are detailed under 'Types of outcomes to be predicted')
Effect measures: number of events, sample size, type of effect measure (e.g. risk ratio (RR), hazard ratio (HR), odds ratio (OR)), adjustment of effect measure, adjustment factors, effect measure, 95% confidence interval (CI) or standard error (SE)
Assessment of risk of bias in included studies
Currently, the recommended tool for assessing the risk of bias in prognostic factor studies is the Quality in Prognostic Studies (QUIPS) tool (Hayden 2013). This tool consists of the following six domains.
Study participation: does the study sample adequately represent the population of interest?
Study attrition: do the study data available (i.e. participants not lost to follow‐up) adequately represent the study sample?
Prognostic factor measurement: is the prognostic factor measured in a similar way for all participants?
Outcome measurement: is the outcome of interest measured in a similar way for all participants?
Other prognostic factors (covariates): are other important potential covariates appropriately accounted for?
Statistical analysis and reporting: is the statistical analysis appropriate, and are all primary outcomes reported?
We will use the QUIPS tool to assess the risk of bias in each study across the aforementioned domains; however, in keeping with Hayden 2019 and Westby 2018, we will tailor the prompting items to the topic of this review (Appendix 2).
Two review authors will independently judge the risk of bias for each domain as low, moderate, or high. If a conflict arises that the review authors cannot solve by discussion, we will consult a third review author to reach a consensus. We will pilot assessments on five studies to ensure consistency of ratings between review authors.
We will combine these ratings to provide an overall risk of bias rating for the factor‐outcome relationship for each study. If all domains are rated as low risk of bias, we will consider the factor‐outcome relationship to be at low risk of bias for the respective study. If one or more domains are rated high risk of bias, we will rate the factor‐outcome relationship for the study at high risk of bias. If any domain is rated as moderate risk of bias, and none is rated as high risk, we will consider the relationship to be at moderate risk of bias. We will examine differences in risk of bias in sensitivity analyses (see Sensitivity analyses).
We are aware of ongoing efforts to improve the QUIPS tool. If the new tool is available before we commence the risk of bias assessment, we will use the updated version; if not, we will use the version in Appendix 2.
Measures of association
For overall survival, we will collect HR and their 95% CIs, where available, because survival analyses carry more information on the time component compared to RRs, which can only be measured at one time point each. If survival analyses are not commonly reported, we will select a patient‐ or clinically relevant time point and extract RRs instead. For the response outcome, we will extract RRs and their respective CIs. For the adverse events of neurotoxicity and CRS, we would prefer to extract HRs and their respective 95% CIs. However, if not commonly reported, we will collect RRs instead. If only ORs are reported, we will extract them and attempt to convert them to RRs.
Where we decide to summarise survival analyses and no HRs are reported, we will attempt to recalculate the HR based on methods suggested by Tierney and colleagues (Tierney 2007); that is, based on log‐rank P values and the number of events/participants, extraction of survival curves, etc. For all factors and outcomes, we aim to summarise adjusted associations. We expect adjustment for the following clinically relevant variables: age, clinical stage, ECOG performance status, LDH (> 1 x normal), and extranodal site (> 1).
For continuous outcomes, we will extract the mean difference (MD) with 95% CIs, if studies use the same scale. If studies use multiple scales, we will calculate the standardised mean difference (SMD) with 95% CIs.
Dealing with missing data
In case of missing data regarding data extraction or quality assessment, we will first try to recalculate the effect measures from available data (Tierney 2007). Should this not be possible, we will contact the primary study authors for additional information. If we cannot obtain additional data, we will use only the available data.
Assessment of heterogeneity
We will judge clinical heterogeneity based on study, participant, and intervention characteristics. Heterogeneity due to participant characteristics would be covered by using adjusted effect estimates. However, if these are not available, we will look at the similarities between studies before pooling. We will explore intervention characteristics (i.e. CAR T‐cell type and disease) in subgroup analysis. We will judge methodological heterogeneity based on study design (i.e. prospective and retrospective cohorts) and risk of bias assessments.
To assess statistical heterogeneity, we will use the Chi² test at a significance level of P < 0.10, indicative of heterogeneity (Deeks 2023). The resulting I² statistic (Higgins 2003), will quantify heterogeneity according to the guidance in the Cochrane Handbook for Systematic Reviews of Interventions (I² statistic > 30% to signify moderate heterogeneity, I² statistic > 75% to signify considerable heterogeneity; Deeks 2023). If heterogeneity is considerable, we will conduct sensitivity analyses, as described below. Sources of heterogeneity depend on the prognostic factor under exploration but may include patient characteristics and treatment characteristics.
Assessment of reporting deficiencies and publication bias
Prognostic factor research is at high risk of publication bias, as studies with significant or large effects are more likely to be reported in a study in sufficient detail, or even published, than those with small effects. Adjusted effects may only be reported for those factors that retain statistical significance in both univariable and multivariable analysis (Riley 2019).
Although most tests of small‐study effect suffer from very low power (Debray 2018), if a meta‐analysis contains more than 10 studies, we will examine the small‐study effect by visually exploring the symmetry of funnel plots, and with Egger's test at the 10% level (which is indicative of asymmetry; Sterne 2011). For meta‐analyses of overall survival and progression‐free survival analysed as HRs, we will use the method described by Debray 2018, to create funnel plots and test for asymmetry (Debray funnel inverse variance, which is based on the inverse of the event rate instead of the total sample size due to potential dropout).
Data synthesis
Data synthesis and meta‐analysis approaches
We will analyse each prognostic factor and each outcome separately for factor‐outcome combinations for which a minimum of three eligible studies provide quantitative data, using RevMan Web 2023 for analysis. We will consider only aggregate data, and not attempt to collect individual participant data, as this would require separate data use agreements with primary data holders and may take considerably more time and resources than available. If the clinical and methodological characteristics of primary studies are sufficiently homogeneous, we will pool the effect measures across trials. We aim to pool adjusted analyses, which may include a slightly different set of covariates per study. If we cannot obtain an adjusted analysis, we will use unadjusted effect estimates in the same meta‐analysis, but remove them in sensitivity analysis to explore their effect on the pooled result.
If studies report events per group for homogeneous time points, we will pool RRs and their corresponding 95% CIs using the Mantel‐Haenszel method in a random‐effects model because we assume that different populations would have varying true relationships between factors and outcomes. To meta‐analyse HRs, we will use the inverse variance methods and a random‐effects model. We will invert HRs where necessary (Deeks 2023).
In case data of some factor‐outcome relationships do not allow for meta‐analysis, we will present data in tabular form, narratively.
Subgroup analysis and investigation of heterogeneity
We will explore the following characteristics as subgroup analyses if the number of studies and participants allows.
Type of CAR T‐cell therapy type (e.g. axicabtagene ciloleucel, tisagenlecleucel, lisocabtagene ciloleucel)
Type of lymphoma (DLBCL versus PMBCL versus FL grade 3b versus HGBCL)
Further subgroup analyses are not applicable to this review, as the relevant subgroups would be explored either as prognostic factors or as adjustment factors.
Sensitivity analysis
We plan to conduct sensitivity analyses to examine the effect of the following.
Risk of bias (low and moderate versus high risk)
Study design (prospective study/clinical trial versus retrospective data collection/'real‐world data')
Adjustment (adjusted versus unadjusted analysis)
Conclusions and summary of findings
We will use the modified GRADE framework for prognostic factors, rating the five standard domains of risk of bias, inconsistency, imprecision, indirectness, and publication bias for prognostic factor‐outcome combinations. The overall certainty in the evidence can be rated as high, moderate, low, or very low. Following the adaptation for prognostic factors, we will start at 'high certainty in the evidence' (for prognosis research, RCTs are not the gold standard) and rate down based on the five domains, but consider a large effect, dose‐response relationship and plausible confounding for rating up (Foroutan 2020; Iorio 2015). Two review authors will rate the combinations independently and discuss discrepancies.
We will present the main results for the following outcomes in summary of findings tables.
Overall survival, time‐to‐event outcome, for the longest follow‐up time frame
Best response
Treatment‐related mortality, medium‐term (up to 12 months)
ICANS within 3 months
CRS within 3 months
Prioritised prognostic factors for the summary of findings tables are age, sex, previous therapy, number of previous therapies, LDH, and changes in the cellular tumour antigen p53.
Acknowledgements
We thank our patients and patient representatives, Bernhard Jochheim, Maria Magdalena Geider, Christa Knebel, and Dr Ulrike Holtkamp for their valuable contributions during a patient workshop.
Editorial support: Cochrane Haematology supported the authors in the development of this systematic review protocol.
The following people conducted the editorial process for this article.
Sign‐off Editors: (reviewed collated reviewers' comments and provided guidance to authors) Lise Estcourt, Haematology/Transfusion Medicine, NHS Blood and Transplant, Oxford; (final editorial decision) Toby Lasserson, Deputy Editor in Chief, Cochrane
Managing Editor (selected peer reviewers, collated peer‐reviewer comments, provided editorial guidance to authors, edited the article): Dr Joanne Duffield, Cochrane Central Editorial Service
Editorial Assistant (conducted editorial policy checks and supported editorial team): Leticia Rodrigues, Cochrane Central Editorial Service
Copy Editor (copy‐editing and production): Denise Mitchell, Cochrane Central Production Service
Peer reviewers (provided comments and recommended an editorial decision): Jennifer Hilgart, Cochrane (methods), Maria‐Inti Metzendorf, Institute of General Practice, Medical Faculty of the Heinrich Heine University, Düsseldorf, Germany (search), Dr Robin Sanderson, King's College Hospital, London (clinical), Pasqualina Santaguida PhD PT, Department of Health Research Methods, Evidence and Impact (HEI) McMaster University (clinical), Jessica D'Urbano (consumer). One additional peer reviewer provided clinical peer review but chose not to be publicly acknowledged.
Appendices
Appendix 1. Preliminary MEDLINE search strategy
# Searches
1 Lymphoma, B‐Cell/
2 ((b‐cell adj3 lymphom*) or (bcell adj3 lymphom*) or (b‐cell adj3 malignanc*) or (bcell adj3 malignanc*)).tw,kf.
3 (large B‐cell* or high‐grade B‐cell* or highgrade B‐cell* or high‐grade Bcell* or highgrade Bcell*).tw,kf.
4 (malignant* lymphom* or hematological malignant or haematological malignant).tw,kf.
5 exp Lymphoma, Large B‐Cell, Diffuse/
6 ((large cell* adj3 lymphom*) or (lymphoid* adj3 lymphom*) or (histiocytic* adj3 lymphom*) or (plasmablastic* adj3 lymphom*)).tw,kf.
7 (DLBCL or LBCL).tw,kf.
8 Mediastinal Neoplasms/ and Lymphoma, B‐Cell/
9 (mediastin* lymphom* or PMBCL or PMBL).tw,kf.
10 Lymphoma, Follicular/
11 (brill symmers or (nodular adj2 lymphom*) or (follicular* adj3 lymphom*)).tw,kf.
12 HGBCL.tw.
13 ((aggressive adj3 lymphom*) or (aggressive adj3 malignanc*)).tw,kf.
14 or/1‐13
15 Immunotherapy, Adoptive/
16 ((cell therap* or cellular* therap* or immune therap* or immuno‐therap* or immunotherap*) adj3 adoptiv*).tw,kf,nm.
17 ((t‐cell* adj1 therap*) or (tcell* adj1 therap*) or (car‐t adj2 therap*)).tw.
18 Antigens, CD19/
19 receptors, antigen, t‐cell/ or receptors, chimeric antigen/
20 ((artificial* adj3 T cell receptor*) or (artificial* adj3 Tcell receptor*)).tw,kf,nm.
21 ((chimeric* adj3 antigen receptor*) or (chimeric* adj3 (immune receptor* or immuno‐receptor* or immunoreceptor*)) or (chimeric* adj3 T cell*) or (chimeric* adj3 Tcell*)).tw,kf,nm.
22 ((chimaeric* adj3 antigen receptor*) or (chimaeric* adj3 (immune receptor* or immuno‐receptor* or immunoreceptor*)) or (chimaeric* adj3 T cell*) or (chimaeric* adj3 Tcell*)).tw,kf,nm.
23 (CD19 or antiCD19).tw,kf,nm.
24 ("CART2019.1" or UCART or universalCART or "duoCAR‐T" or "CD19‐TriCART‐T").tw,kf.
25 (axicabtagene* or yescarta* or axi‐cel* or KTE‐C19*).tw,kf,nm.
26 (tisagenlecleucel* or kymriah* or tisa‐cel* or CART‐19 or CART19 or "CTL 019").tw,kf,nm.
27 (lisocabtagene* or liso‐cel* or JCAR017*).tw,kf,nm.
28 (relmacabtagene* or relma‐cel* or carteyva*).tw,kf,nm.
29 or/15‐28
30 14 and 29
31 Validat$.tw. or Predict$.ti. or Rule$.tw. or (Predict$ and (Outcome$ or Risk$ or Model$)).tw. or ((History or Variable$ or Criteria or Scor$ or Characteristic$ or Finding$ or Factor$) and (Predict$ or Model$ or Decision$ or Identif$ or Prognos$)).tw. or (Decision$.tw. and ((Model$ or Clinical$).tw. or logistic models/)) or (Prognostic and (History or Variable$ or Criteria or Scor$ or Characteristic$ or Finding$ or Factor$ or Model$)).tw. or ("Stratification" or "Discrimination" or "Discriminate" or c‐statistic or "Area under the curve" or "AUC" or "Calibration" or "Indices" or "Algorithm" or "Multivariable").tw.
32 "Predictive Value of Tests"/ or Observer Variation/ or (predict* or scor* or observ*).tw.
33 exp animals/ not humans/
34 (31 or 32) not 33
35 exp Neurotoxicity Syndromes/
36 (poisoning adj2 nervous system*).tw,kf.
37 (toxic adj2 (encephalopath* or encephaliti*)).tw,kf.
38 (neurotoxic disorder* or neurotoxicity syndrome*).tw,kf.
39 (neurotoxi* adj1 (disorder* or disease*)).tw,kf.
40 ICANS.tw,kf.
41 Cytokine Release Syndrome/
42 (cytokine release syndrom* or cytokine storm* or hypercytokinemia* or hyper‐cytokinemia* or hypercytokinaemia* or hyper‐cytokinaemia*).tw,kf.
43 exp Survival Rate/ or Disease‐Free Survival/ or Survival/ or Progression‐Free Survival/ or exp Survival Analysis/
44 (suvival* or OS or PFS).tw,kf.
45 (response* or remission*).tw,kf.
46 exp Neutropenia/
47 (neutropeni* or neutrophil*).tw,kf.
48 ((treatment* adj2 related) and (death* or mortality*)).tw,kf.
49 ((hematological or haematological) adj3 toxicit*).tw,kf.
50 (cytopeni* or cytopaeni* or pancytopeni* or pancytopaeni*).tw,kf.
51 "Drug‐Related Side Effects and Adverse Reactions"/ or exp poisoning/ or exp drug toxicity/ or exp drug monitoring/
52 (safe or safety or side effect* or toxicit* or complications or noxious or tolerability or (treatment* adj2 emergent*)).tw.
53 (ae or to or po or co or mo).fs.
54 ((adverse or undesirable or harms* or serious or toxic) adj3 (effect* or reaction* or event*)).ti,ab.
55 exp treatment outcome/
56 (outcome* or (efficacy adj1 (clinical or treatment or therap*))).tw,kf.
57 or/35‐56
58 randomized controlled trial.pt.
59 controlled clinical trial.pt.
60 randomi?ed.ab.
61 placebo.ab.
62 drug therapy.fs.
63 randomly.ab.
64 trial.ab.
65 groups.ab.
66 or/58‐65
67 exp animals/ not humans/
68 66 not 67
69 clinical trial, phase iii/
70 ("Phase 3" or "phase3" or "phase III" or P3 or "PIII").ti,ab,kw.
71 (69 or 70) not 67
72 68 or 71
73 30 and (34 or 57 or 72)
74 limit 73 to yr="2010 ‐ current"
Appendix 2. Quality In Prognosis Studies (QUIPS): risk of bias assessment
The here‐presented adaptation of the QUIPS tool (Hayden 2013), is based on two Cochrane Reviews (Hayden 2019; Westby 2018).
We will assess the risk of bias for each prognostic factor–outcome combination, although certain subitems may be rated at the study level (e.g. selective reporting, attrition bias), if possible.
Domain 1: study participation
Selection bias: is the relationship between a prognostic factor and an outcome different for participants and eligible non‐participants?
Low risk of bias
All of the following:
the sample was recruited randomly or consecutively from the source population; and
the study sample represents the population of interest on critical characteristics sufficiently, to limit potential bias of the observed relationship between prognostic factors and outcome.
Moderate risk of bias
There was insufficient information to judge whether the prognostic factor‐outcome association differed between participants included in the study and eligible non‐participants.
High risk of bias
The sample was recruited selectively.
Domain 2: study attrition
Attrition bias: is it likely that the relationship between prognostic factor and outcome is different for completing and non‐completing participants (in other words, did participants who completed the study represent the baseline sample)?
Low risk of bias
Any of the following:
there were no or only few (< 5%) missing outcome data; or
-
reasons for missing outcome data indicate that missing data were unrelated to the actual outcome event (missing at random ‐ it is implausible that patients who were particularly well or sick were lost to follow‐up):
the number of missing outcome data was balanced across categories, and the reasons were similar across groups; or
appropriate methods for missing data imputation were used.
Moderate risk of bias
There was insufficient information to judge whether the prognostic factor‐outcome association differed for completing and non‐completing participants.
High risk of bias
Any of the following:
the reasons why data were missing were probably related to the outcome value or the prognostic factor (i.e. imbalance of numbers or reasons for missing data);
high percentages of outcome data missing; or
inappropriate use of imputation (i.e. simple imputation).
Domain 3: prognostic factor measurement
Measurement bias: was the prognostic factor measured or defined differently in relation to the level of outcome?
Low risk of bias
All of the following apply:
the prognostic factor is clearly defined and measured similarly for all participants (method and setting), irrespective of outcome level;
the measurement is valid and reliable;
continuous variables were either used as such, or the proper cut points were used if they were employed; and
the proportion of participants with incomplete data was low or unrelated to the outcome; if imputation techniques were used, missing data were imputed using appropriate methods.
Moderate risk of bias
There was insufficient information to judge whether the prognostic factor's measurement varied according to the outcome level.
High risk of bias
One of the following:
the prognostic factor was measured differently for participants with different outcome levels;
an unreliable prognostic factor measurement method was used (non‐differential misclassification bias);
cut‐off points for dichotomous prognostic variables were chosen improperly;
the proportion of participants with incomplete data was high and potentially related to the outcome;
details about the prognostic factor were based on recall or potentially unreliable participant records (reporting bias); or
prognostic factor measurement was unreliable due to the potential for reporting bias or participant recollection.
Domain 4: outcome measurement
Risk of bias related to the measurement of outcome: was the outcome measured or obtained in a way that would allow differential measurement of outcome related to the baseline level of prognostic factor(s)?
Low risk of bias
All of the below:
the outcome measure was clearly defined and described, and the measure valid and reliable; and
the outcome was measured consistently across participants, independent from prognostic factor information (blinding of outcome assessor or extractor is an advantage).
Moderate risk of bias
There was insufficient information to judge whether varied baseline levels of the prognostic factor led to different measurements of the outcome.
High risk of bias
Any of the following:
an unreliable or non‐validated outcome measure was used;
the outcome was measured as participant or caretaker recall; or
the outcome was measured differently for participants with varying prognostic factor values, where outcome measurement or extraction was not blinded.
Domain 5: adjustment/covariates
Risk of bias due to confounding: could the relationship between a prognostic factor and outcome be distorted by another factor/covariate that is related to both the examined prognostic factor and the outcome?
The following key adjustment factors should be considered in adjusted analyses, if not explored as the prognostic factor.
Age
Sex
Histology
Lactate dehydrogenase (LDH)
Eastern Cooperative Oncology Group (ECOG) performance status
Extranodal region involved (≤ 1/> 1)
Number of previous lines of therapy
Chimeric antigen receptor (CAR) T‐cell therapy
Bridging therapy (yes/no)
Response to bridging therapy
Bulky tumour
Low risk of bias
Important adjustment factors were considered when designing the study (e.g. matching, stratification), conducting the analysis (e.g. multivariable regression), and comparing the results of the adjusted and unadjusted analyses to determine any discrepancies.
-
Key adjustment factors should be:
measured validly and reliably;
measured in the same way for all participants independent of target prognostic factor level or outcome;
imputed appropriately for missing values, where necessary; and
if applicable, therapies were taken into consideration during the analysis.
Moderate risk of bias
There was insufficient information to judge the potential distortion of the prognostic factor–outcome relationship (e.g. adjustment factors were unclear).
High risk of bias
Any of the following:
in the design or analysis, important adjustment factors were not taken into consideration;
adjustment factors were measured inadequately;
adjustment factor measurement was different across prognostic factor levels or outcome levels; or
interventions were different for different prognostic factor levels or outcome levels.
Domain 6: statistical analysis and reporting
Risk of bias related to the statistical analyses and completeness of results: were the study’s statistical analyses inappropriate, or were there indications that results were presented selectively?
Low risk of bias
All of the following:
the statistical analysis was adequate (e.g. taking into account the study design, unit of analysis issues, etc.);
the strategy for model building was appropriate (the inclusion of variables in the statistical model); and
results were not reported selectively.
Moderate risk of bias
There was insufficient evidence to make a judgement on whether the reported results were biased, either in relation to the analysis or the reporting of results.
High risk of bias
Any of the following:
the statistical model was not adequate;
the model‐building approach was inappropriate (e.g. regarding variable or interaction selection in multivariable analysis, or ignoring relevant intervention effects); or
results were selectively reported (i.e. based on statistical significance ‐ selection of certain associations of prognostic factors and outcomes over others, only selected time points).
Contributions of authors
| CH | Conceptualisation of the protocol ‐ writing of manuscript, methods |
| MG | Conceptualisation of the protocol ‐ background |
| ME | Conceptualisation of the protocol ‐ background |
| IM | Development of search strategy and definition of databases |
| BvT | Conceptualisation of the protocol ‐ clinical expertise, definition of the PICO, eligibility criteria |
| NS | Conceptualisation of the protocol ‐ clinical and methodological input |
| NK | Conceptualisation of the protocol ‐ writing of manuscript, background, methods; methodological input |
Sources of support
Internal sources
-
University Hospital of Cologne, Germany
Cochrane Haematology, Department I of Internal Medicine
External sources
-
German Ministry for Education and Research (BMBF), Germany
Grant (number 01KG2204) to develop this protocol and systematic review. The funder had no role in the design, conduct or publication of the protocol.
Declarations of interest
| CH | None known. CH is the Managing Editor of Cochrane Haematology. She was not involved in the editorial process for this protocol. |
| MG | None known. MG is a member of staff at Cochrane Haematology. He was not involved in the editorial process for this protocol. |
| ME | None known. ME is a member of staff at Cochrane Haematology. He was not involved in the editorial process for this protocol. |
| IM | None known. IM is the Information Specialist at Cochrane Haematology. She was not involved in the editorial process for this protocol. |
| BvT | BvT declares grant/research funding from Esteve, Merck Sharp & Dohme Corporation, Novartis and Takeda Pharmaceutical Company (all paid to institution). He also declares (1) consulting fees from Allogene, Amgen, Bristol‐Myers Squibb, Cerus Corporation, Gilead Sciences Inc, Incyte Corporation, IQVIA Ltd, Lilly Deutschland, Merck Sharp and Dohme, Miltenyi Biotec, Noscendo GmbH, Novartis Pharma, PentixaPharm GmbH, Pfizer Canada Inc, Pierre Fabre Pharmaceuticals, Inc, Qualworld, Quintiles Transnational Corp, Roche, Sobi, Inc and Takeda Pharmaceutical Company; (2) honoraria for speaking engagements from AbbVie, AstraZeneca, Bristol‐Myers Squibb, Gilead Sciences Inc, Incyte Corporation, Lilly Deutschland, Merck Sharp and Dohme, Novartis, Roche and Takeda Pharmaceutical Company; and (3) travel support from AbbVie, AstraZeneca, Gilead Sciences Inc, Lilly Deutschland, Merck Sharp and Dohme, Novartis, Pierre Fabre Pharmaceuticals, Inc, Roche and Takeda Pharmaceutical Company (all personal payments). BvT is a health professional in Haematology and stem cell transplantation at University Hospital Essen and an Editor with Cochrane Haematology. He was not involved in the editorial process for this protocol. |
| NS | None known. She is the Co‐ordinating Editor of Cochrane Haematology. She was not involved in the editorial process for this protocol. |
| NK | None known. NK is a member of staff at Cochrane Haematology. She was not involved in the editorial process for this protocol. |
The authors CH, MG, ME, IM, NS, and NK are affiliated with Cochrane Haematology but are not otherwise involved with the editorial process.
New
References
Additional references
Abramson 2020
- Abramson JS, Palomba ML, Gordon LI, Lunning MA, Wang M, Arnason J, et al. Lisocabtagene maraleucel for patients with relapsed or refractory large B-cell lymphomas (TRANSCEND NHL 001): a multicentre seamless design study. Lancet 2020;396(10254):839-52. [DOI: 10.1016/S0140-6736(20)31366-0] [DOI] [PubMed] [Google Scholar]
Bishop 2022
- Bishop MR, Dickinson M, Purtill D, Barba P, Santoro A, Hamad N, et al. Second-line tisagenlecleucel or standard care in aggressive B-cell lymphoma. New England Journal of Medicine 2022;386(7):629-39. [DOI: 10.1056/NEJMoa2116596] [DOI] [PubMed] [Google Scholar]
Castaneda‐Puglianini 2021
- Castaneda-Puglianini O, Chavez JC. Assessing and management of neurotoxicity after CAR-T therapy in diffuse large B-Cell lymphoma. Journal of Blood Medicine 2021;24(12):775-83. [DOI: 10.2147/JBM.S281247] [DOI] [PMC free article] [PubMed] [Google Scholar]
Cheson 2014
- Cheson BD, Fisher RI, Barrington SF, Cavalli F, Schwartz LH, Zucca E, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. Journal of Clinical Oncology 2014;32(27):3059-67. [DOI: ] [DOI] [PMC free article] [PubMed] [Google Scholar]
Cooper 2019
- Cooper C, Varley-Campbell J, Carter P. Established search filters may miss studies when identifying randomized controlled trials. Journal of Clinical Epidemiology 2019;112:12-9. [DOI: 10.1016/j.jclinepi.2019.04.002] [DOI] [PubMed] [Google Scholar]
Csenar 2022
- Csenar M, Oeser A, Bora A-M, Gippert T, Borchmann P, Kron F, et al. Application of CD19-directed CAR T-cell therapy for the WHO Model List of Essential Medicines 2023. cdn.who.int/media/docs/default-source/essential-medicines/2023-eml-expert-committee/applications-for-addition-of-new-medicines/a8_cd19-car-t-cells.pdf?sfvrsn=724db817_2 (accessed prior to 14 September 2023).
Debray 2018
- Debray TP, Moons KG, Riley RD. Detecting small-study effects and funnel plot asymmetry in meta-analysis of survival data: a comparison of new and existing tests. Research Synthesis Methods 2018;9(1):41-50. [DOI: ] [DOI] [PMC free article] [PubMed] [Google Scholar]
Deeks 2023
- Deeks JJ, Higgins JP, Altman DG, editor(s). Chapter 10: Analysing data and undertaking meta-analyses. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.4 (updated August 2023). Cochrane, 2023. Available from www.training.cochrane.org/handbook.
EORTC 2023
- EORTC quality of life questionnaires. qol.eortc.org/questionnaires/ (accessed 13 September 2023).
Ernst 2021
- Ernst M, Abd El Aziz M, Besiroglu B, Caro-Valenzuela C, Oeser A, Estcourt LJ, et al. Chimeric antigen receptor (CAR) T-cell therapy for people with relapsed or refractory diffuse large B-cell lymphoma. Cochrane Database of Systematic Reviews 2021, Issue 9. Art. No: CD013365. [DOI: 10.1002/14651858.CD013365.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]
Foroutan 2020
- Foroutan F, Guyatt G, Zuk V, Vandvik PO, Alba AC, Mustafa R, et al. Guidelines 28: use of GRADE for the assessment of evidence about prognostic factors: rating certainty in identification of groups of patients with different absolute risks. Journal of Clinical Epidemiology 2020;121:62-70. [DOI: 10.1016/j.jclinepi.2019.12.023] [DOI] [PubMed] [Google Scholar]
Geersing 2012
- Geersing GJ, Bouwmeester W, Zuithoff P, Spijker R, Leeflang M, Moons KG. Search filters for finding prognostic and diagnostic prediction studies in Medline to enhance systematic reviews. PLoS One 2012;7(2):e32844. [DOI: 10.1371/journal.pone.0032844] [DOI] [PMC free article] [PubMed] [Google Scholar]
Hartmann 2017
- Hartmann J, Schüssler-Lenz M, Bondanza A, Buchholz CJ. Clinical development of CAR T cells—challenges and opportunities in translating innovative treatment concepts. EMBO Molecular Medicine 2017;9(9):1183-97. [DOI: 10.15252/emmm.201607485] [DOI] [PMC free article] [PubMed] [Google Scholar]
Hayden 2013
- Hayden JA, Van der Windt DA, Cartwright JL, Côté P. Assessing bias in studies of prognostic factors. Annals of Internal Medicine 2013;158(4):280-6. [DOI: 10.7326/0003-4819-158-4-201302190-00009] [DOI] [PubMed] [Google Scholar]
Hayden 2019
- Hayden JA, Wilson MN, Riley RD, Iles R, Pincus T, Ogilvie R. Individual recovery expectations and prognosis of outcomes in non‐specific low back pain: prognostic factor review. Cochrane Database of Systematic Reviews 2019, Issue 11. Art. No: CD011284. [DOI: 10.1002/14651858.CD011284.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]
Higgins 2003
- Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. [DOI] [PMC free article] [PubMed] [Google Scholar]
Howlader 2020
- Howlader N, Noone AM, Krapcho M, Miller D, Brest A, Yu M, et al. SEER Cancer Statistics Review, 1975-2017. seer.cancer.gov/archive/csr/1975_2017/ (accessed prior to 14 September 2023).
Iorio 2015
- Iorio A, Spencer FA, Falavigna M, Alba C, Lang E, Burnand B, et al. Use of GRADE for assessment of evidence about prognosis rating confidence in estimates of event rates in broad categories of patients. BMJ 2015;350:h870. [DOI: 10.1136/bmj.h870] [DOI] [PubMed] [Google Scholar]
Kamdar 2022
- Kamdar M, Solomon SR, Arnason J, Johnston PB, Glass B, Bachanova V, et al. Lisocabtagene maraleucel versus standard of care with salvage chemotherapy followed by autologous stem cell transplantation as second-line treatment in patients with relapsed or refractory large B-cell lymphoma (TRANSFORM): results from an interim analysis of an open-label, randomised, phase 3 trial. Lancet 2022;399(10343):2294-308. [DOI: 10.1016/S0140-6736(22)00662-6] [DOI] [PubMed] [Google Scholar]
Kittai 2021
- Kittai AS, Huang Y, Gordon M, Denlinger N, Mian A, Fitzgerald L, et al. Comorbidities predict inferior survival in patients receiving chimeric antigen receptor T cell therapy for diffuse large B cell lymphoma: a multicenter analysis. Transplantation and Cellular Therapy 2021;27(1):46-52. [DOI: 10.1016/j.bbmt.2020.09.028] [DOI] [PubMed] [Google Scholar]
Kubuschok 2015
- Kubuschok B, Held G, Pfreundschuh M. Management of diffuse large B-cell lymphoma (DLBCL). In: Evens AM, Blum KA, editor(s), editors(s). Non-Hodgkin Lymphoma: Pathology, Imaging, and Current Therapy. Cham: Springer International Publishing, 2015:271-88. [DOI: 10.1007/978-3-319-13150-4_11] [DOI] [PubMed] [Google Scholar]
Le Gouill 2021
- Le Gouill S, Bachy E, Di Blasi R, Cartron G, Beauvais D, Le Bras F, et al. First results of DLBCL patients treated with CAR-T cells and enrolled in DESCAR-T Registry, a French real-life database for CAR-T cells in hematologic malignancies. Hematological Oncology 2021;39(Suppl 2):138-9. [ABSTRACT NUMBER: 084] [DOI: 10.1002/hon.84_2879] [DOI] [Google Scholar]
Lefebvre 2022
- Lefebvre C, Glanville J, Briscoe S, Featherstone R, Littlewood A, Marshall C, et al. Technical supplement to Chapter 4: Searching for and selecting studies. In: Higgins JP, Thomas J, Chandler J, Cumpston MS, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.3 (updated February 2022). Cochrane, 2022. Available from: www.training.cochrane.org/handbook.
Li 2023
- Li T, Higgins JP, Deeks JJ, editor(s). Chapter 5: Collecting data. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.4 (updated August 2023). Cochrane, 2023. Available from www.training.cochrane.org/handbook.
Nastoupil 2020
- Nastoupil LJ, Jain MD, Feng L, Spiegel JY, Ghobadi A, Lin Y, et al. Standard-of-care axicabtagene ciloleucel for relapsed or refractory large B-cell lymphoma: results from the US lymphoma CAR T consortium. Journal of Clinical Oncology 2020;38(27):3119-28. [DOI: 10.1200/JCO.19.02104] [DOI] [PMC free article] [PubMed] [Google Scholar]
Neelapu 2017
- Neelapu SS, Locke FL, Bartlett NL, Lekakis LJ, Miklos DB, Jacobson CA, et al. Axicabtagene ciloleucel CAR T-cell therapy in refractory large B-cell lymphoma. New England Journal of Medicine 2017;377(26):2531-44. [DOI: 10.1056/NEJMoa1707447] [DOI] [PMC free article] [PubMed] [Google Scholar]
Oken 1982
- Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. American Journal of Clinical Oncology 1982;5(6):649–55. [PMID: ] [PubMed] [Google Scholar]
Page 2021
- Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [DOI: 10.1136/bmj.n71] [DOI] [PMC free article] [PubMed] [Google Scholar]
Pfreundschuh 2006
- Pfreundschuh M, Trümper L, Österborg A, Pettengell R, Trneny M, Imrie K, et al. CHOP-like chemotherapy plus rituximab versus CHOP-like chemotherapy alone in young patients with good-prognosis diffuse large-B-cell lymphoma: a randomised controlled trial by the MabThera International Trial (MInT) Group. Lancet. Oncology 2006;7(5):379-91. [DOI: 10.1016/S1470-2045(06)70664-7] [DOI] [PubMed] [Google Scholar]
RevMan Web 2023 [Computer program]
- Review Manager Web (RevMan Web). Version 5.5.0. The Cochrane Collaboration, 2023. Available at revman.cochrane.org.
Riley 2013
- Riley RD, Hayden JA, Steyerberg EW, Moons KG, Abrams K, Kyzas PA, et al. Prognosis research strategy (PROGRESS) 2: prognostic factor research. PLOS Medicine 2013;10(2):e1001380. [DOI: 10.1371/journal.pmed.1001380] [DOI] [PMC free article] [PubMed] [Google Scholar]
Riley 2019
- Riley RD, Moons KG, Snell KI, Ensor J, Hooft L, Altman DG, et al. A guide to systematic review and meta-analysis of prognostic factor studies. BMJ 2019;364:k4597. [DOI: 10.1136/bmj.k4597] [DOI] [PubMed] [Google Scholar]
Santomasso 2019
- Santomasso B, Bachier C, Westin J, Rezvani K, Shpall EJ. The other side of CAR T-cell therapy: cytokine release syndrome, neurologic toxicity, and financial burden. American Society of Clinical Oncology Educational Book 2019;39:433-44. [DOI: 10.1200/EDBK_238691] [DOI] [PubMed] [Google Scholar]
Sarkozy 2018
- Sarkozy C, Sehn LH. Management of relapsed/refractory DLBCL. Best Practice & Research Clinical Haematology 2018;31(3):209-16. [DOI: 10.1016/j.beha.2018.07.014] [DOI] [PubMed] [Google Scholar]
Schuster 2018
- Schuster SJ, Bishop MR, Tam CS, Waller EK, Borchmann P, McGuirk JP, et al. Tisagenlecleucel in adult relapsed or refractory diffuse large B-Cell lymphoma. New England Journal of Medicine 2018;380(1):45-56. [DOI: 10.1056/NEJMoa1804980] [DOI] [PubMed] [Google Scholar]
St‐Pierre 2022
- St-Pierre F, Gordon LI. CAR T-cell therapy for relapsed/refractory non-Hodgkin's lymphoma: a comprehensive review. Clinical Advances in Hematology & Oncology 2022;20(5):309-18. [PMID: ] [PubMed] [Google Scholar]
Sterne 2011
- Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ 2011;343:d4002. [DOI: 10.1136/bmj.d4002] [DOI] [PubMed] [Google Scholar]
Sterner 2021
- Sterner RC, Sterner RM. CAR-T cell therapy: current limitations and potential strategies. Blood Cancer Journal 2021;11(4):69. [DOI: 10.1038/s41408-021-00459-7] [DOI] [PMC free article] [PubMed] [Google Scholar]
Swerdlow 2017
- Swerdlow SH, Camp E, Harris NL, Jaffe ES, Pileri SA, Stein H, et al. WHO classification of tumours of haematopoietic and lymphoid tissues. Lyon, France: World Health Organization, 2017. [URL: archive.org/details/whoclassificatio00swer/page/292/mode/2up] [Google Scholar]
Tierney 2007
- Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials 2007;8(1):16. [DOI: 10.1186/1745-6215-8-16] [DOI] [PMC free article] [PubMed] [Google Scholar]
Vercellino 2020
- Vercellino L, Di Blasi R, Kanoun S, Tessoulin B, Rossi C, D'Aveni-Piney M, et al. Predictive factors of early progression after CAR T-cell therapy in relapsed/refractory diffuse large B-cell lymphoma. Blood Advances 2020;4(22):5607-15. [DOI: 10.1182/bloodadvances.2020003001] [DOI] [PMC free article] [PubMed] [Google Scholar]
Westby 2018
- Westby MJ, Dumville JC, Stubbs N, Norman G, Wong JK, Cullum N, et al. Protease activity as a prognostic factor for wound healing in venous leg ulcers. Cochrane Database of Systematic Reviews 2018, Issue 9. Art. No: CD012841. [DOI: 10.1002/14651858.CD012841.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]
Westin 2019
- Westin JR, Tam CS, Borchmann P, Jaeger U, McGuirk JP, Holte H, et al. Correlative analyses of patient and clinical characteristics associated with efficacy in tisagenlecleucel-treated relapsed/refractory diffuse large B-cell lymphoma patients in the Juliet trial. Blood 2019;134:4103. [DOI: 10.1182/blood-2019-129107] [DOI] [Google Scholar]
WHO 2023
- WHO Model Lists of Essential Medicines Cancer Medicines Working Group. Advice for the expert committee on selection and use of essential medicines. tinyurl.com/2p8pb955 (accessed 2 January 2023).
Younes 2017
- Younes A, Hilden P, Coiffier B, Hagenbeek A, Salles G, Wilson W, et al. International Working Group consensus response evaluation criteria in lymphoma (RECIL 2017). Annals of Oncology 2017;28(7):1436-47. [DOI: 10.1093/annonc/mdx097] [DOI] [PMC free article] [PubMed] [Google Scholar]
ZUMA‐7
- Locke FL, Miklos DB, Jacobson CA, Perales M-A, Kersten M-J, Oluwole OO, et al. Axicabtagene ciloleucel as second-line therapy for large B-cell lymphoma. New England Journal of Medicine 2022;386(7):640-54. [DOI: 10.1056/NEJMoa2116133] [DOI] [PubMed] [Google Scholar]