Acceptability of Using Real-World Data to Estimate Relative Treatment Effects in Health Technology Assessments: Barriers and Future Steps.

OBJECTIVES: Evidence about the comparative effects of new treatments is typically collected in randomized controlled trials (RCTs). In some instances, RCTs are not possible, or their value is limited by an inability to capture treatment effects over the longer term or in all relevant population subgroups. In these cases, nonrandomized studies (NRS) using real-world data (RWD) are increasingly used to complement trial evidence on treatment effects for health technology assessment (HTA). However, there have been concerns over a lack of acceptability of this evidence by HTA agencies. This article aims to identify the barriers to the acceptance of NRS and steps that may facilitate increases in the acceptability of NRS in the future. METHODS: Opinions of the authorship team based on their experience in real-world evidence research in academic, HTA, and industry settings, supported by a critical assessment of existing studies. RESULTS: Barriers were identified that are applicable to key stakeholder groups, including HTA agencies (eg, the lack of comprehensive methodological guidelines for using RWD), evidence generators (eg, avoidable deviations from best practices), and external stakeholders (eg, data controllers providing timely access to high-quality RWD). Future steps that may facilitate future acceptability of NRS include improvements in the quality, integration, and accessibility of RWD, wider use of demonstration projects to highlight the value and applicability of nonrandomized designs, living, and more detailed HTA guidelines, and improvements in HTA infrastructure relating to RWD. CONCLUSION: NRS can represent a crucial source of evidence on treatment effects for use in HTA when RCT evidence is limited.

The value of innovation: association between improvements in survival of advanced and metastatic non-small cell lung cancer and targeted and immunotherapy.

BACKGROUND: Significant improvements in mortality among patients with non-small cell lung cancer (NSCLC) in the USA over the past two decades have been reported based on Surveillance, Epidemiology, and End Results (SEER) data. The timing of these improvements led to suggestions that they result from the introduction of new treatments; however, few studies have directly investigated this. The aim of this study was to investigate the extent to which population level improvements in survival of advanced and/or metastatic NSCLC (admNSCLC) patients were associated with changes in treatment patterns. METHODS: We utilized a de-identified database to select three cohorts of patients with admNSCLC: (1) patients with non-oncogene (EGFR/ALK/ROS1/BRAF) positive tumors, (2) patients with ALK-positive (ALK+) tumors, and (3) patients with EGFR-positive (EGFR+) tumors. All patients were diagnosed with admNSCLC between 2012 and 2019. Multivariable Cox models adjusting for baseline characteristics and receipt of targeted and immunotherapy were utilized to explore the relationship between these variables and changes in the hazard of death by calendar year in each cohort. RESULTS: We included 28,154 admNSCLC patients with non-oncogene positive tumors, 598 with ALK+ tumors, and 2464 with EGFR+ tumors eligible for analysis. After adjustment for differences in baseline characteristics, the hazard of death in patients who had non-oncogene positive tumors diagnosed in 2015, 2016, 2017, 2018 ,and 2019 was observed to be 12%, 11%, 17%, 20%, and 21% lower respectively than that for those diagnosed in 2012. Upon additionally adjusting for receipt of first line or second line immunotherapy, the decrease in the hazard of death by calendar year was no longer observed, suggesting improvements in survival observed over time may be explained by the introduction of these treatments. Similarly, decreases in the hazard of death were only observed in patients with ALK+ tumors diagnosed between 2017 and 2019 relative to 2012 but were no longer observed following adjustment for the use of 1st and later generation ALK inhibitors. Among patients with EGFR+ tumors, the hazard of death did not improve significantly over time. CONCLUSION: Our findings expand on the SEER data and provide additional evidence suggesting improvements in survival of patients with advanced and metastatic NSCLC over the past decade could be explained by the change in treatment patterns over this period.

Matching-adjusted indirect treatment comparison of [177Lu]Lu-DOTA-TATE, everolimus and sunitinib in advanced, unresectable gastroenteropancreatic neuroendocrine tumours: Relative effectiveness of [177Lu]Lu-DOTA-TATE in gastroenteropancreatic neuroendocrine tumours.

Head-to-head comparisons of the efficacy of treatments for gastroenteropancreatic neuroendocrine tumours (GEP-NETs) have not yet been reported. This study used a series of matching-adjusted indirect comparisons to indirectly compare the effectiveness of [177Lu]Lu-DOTA-TATE to everolimus, sunitinib and best supportive care (BSC) for extending progression-free survival and overall survival in patients with advanced, unresectable gastrointestinal (GI)-NETs and P-NETs. The results of the main analysis suggest that after accounting for differences in key prognostic variables, the hazard of progression was 62% (hazard ratio [HR], 0.38; confidence interval [CI]95 0.25-0.58) and 65% (HR 0.35 CI95 0.21-0.59) lower in patients with GI-NETs treated with [177Lu]Lu-DOTA-TATE than in those treated with everolimus and BSC, respectively. Similarly, the hazard of progression was 64% (HR 0.36 CI95 0.18-0.70), 54% (HR 0.46 CI95 0.30-0.71) and 79-87% (HR 0.21 CI95 0.13-0.32; HR 0.13 CI95 0.08-0.22) lower in patients with P-NET treated with [177Lu]Lu-DOTA-TATE than in those treated with sunitinib, everolimus and BSC, respectively. The hazard of death was 58% (HR 0.42 CI95 0.25-0.72), 47% (HR 0.53 CI95 0.33-0.87) and 44-64% (HR 0.56 CI95 0.36-0.90; HR 0.34 CI95 0.20-0.57) lower in P-patients with NET treated with [177Lu]Lu-DOTA-TATE than in those treated with sunitinib, everolimus and BSC, respectively. While our results must be interpreted with caution given the non-randomised nature of the comparisons and the potential for residual confounding, the magnitude of the effect sizes we observe and their consistency across comparators suggest that [177Lu]Lu-DOTA-TATE may be a more effective treatment option than everolimus, sunitinib and BSC in advanced, unresectable GEP-NETs.

Consistency of recording of chemotherapy cycles in the National Cancer Registration and Analysis Service Systemic Anti-Cancer Therapy database and the Hospital Episode Statistics Admitted Patient Care database.

Aim: We assessed the extent to which chemotherapy cycles recorded in Hospital Episode Statistics (HES) Admitted Patient Care (APC) were captured in National Cancer Registration & Analysis Service Systemic Anti-Cancer Therapy (SACT) for a cohort of lung cancer patients. Methods: All chemotherapy cycles recorded for linkage eligible lung cancer patients with a National Cancer Registration & Analysis Service diagnosis between 2012 and 2015 were identified in HES APC and SACT. Results: Among a population of 4070 lung cancer patients, 6076 chemotherapy cycles were observed in HES APC data. A total of 61% of cycles were recorded in SACT on the same day, 8% on a different day and 31% were not recorded in SACT. Conclusion: Our results suggest that SACT may not capture all chemotherapy cycles administered to a patient between 2012 and 2016; however, administrative changes mean data after this period may be more complete.

Factors associated with non-prescription of oral anticoagulant treatment in non-valvular atrial fibrillation patients with dementia: a CPRD-HES study.

Dementia is a common comorbidity in patients with atrial fibrillation (AF) and treatment guidelines recommend oral anticoagulant (OAC) therapy for AF patients with dementia unless concordance cannot be ensured by the caregiver. Despite this, the literature reports a low prescribing of OAC treatment in these patients. This study investigated possible factors associated with non-prescribing of OAC treatment in dementia patients newly diagnosed with non-valvular atrial fibrillation (NVAF) at age ≥ 65 years between 2013 and 2017 using the Clinical Practice Research Datalink and Hospital Episodes Statistics databases. Of 1090 dementia patients newly diagnosed with NVAF, 693 (63.6%) patients did not have a prescription for an OAC in the year following their diagnosis. The likelihood of experiencing a thromboembolic event was high, with 97% of the population having a CHA2DS2-VASc score > 2; however, little difference in the presence of stroke risk factors was observed between the prescribed and non-prescribed groups. The presence of bleeding risk factors was high; only 28 (2.6%) of patients did not have a previous fall or a HAS-BLED bleeding risk factor. A history of falls [OR = 0.76, 95% confidence intervals (CIs) (0.58, 0.98)], previous major bleed [OR = 0.56, 95% CI (0.43, 0.73)] and care home residence [OR = 0.47, 95% CI (0.30, 0.74)] were associated with not having an OAC prescription. The results suggest that dementia patients with NVAF and certain risk bleeding risk factors are less likely to be prescribed an OAC. Further work is needed to establish possible relationships between bleeding risk factors and other potential drivers of OAC prescribing.

The use of UK primary care databases in health technology assessments carried out by the National Institute for health and care excellence (NICE).

BACKGROUND: Real world evidence (RWE) is becoming more frequently used in technology appraisals (TAs). This study sought to explore the use and acceptance of evidence from primary care databases, a key source of RWE in the UK, in National Institute for Health and Care Excellence (NICE) technology assessments and to provide recommendations regarding their use in future submissions. METHODS: A keyword search was conducted relating to the main primary care databases in the UK on the NICE website. All NICE TAs identified through this search were screened, assessed for duplication and information on the data source and the way the data was used in the submission were extracted. Comments by the evidence review group (ERG) and the appraisal committee were also extracted and reviewed. All data extraction was performed by two independent reviewers and all decisions were reached by consensus with an additional third reviewer. RESULTS: A total of 52 NICE TAs were identified, 47 used the General Practice Research Database /Clinical Practice Research Datalink (GPRD/CPRD) database, 10 used The Health Improvement Network (THIN) database and 3 used the QResearch databases. Data from primary care databases were used to support arguments regarding clinical need and current treatment in 33 NICE TAs while 36 were used to inform input parameters for economic models. The databases were sometimes used for more than one purpose. The data from the three data sources were generally well received by the ERGs/committees. Criticisms of the data typically occurred where the results had been repurposed from a published study or had not been applied appropriately. CONCLUSIONS: The potential of UK primary care databases in NICE submissions is increasingly being realised, particularly in informing the parameters of economic models. Purpose conducted studies are less likely to receive criticism from ERGs/committees, particularly when providing clinical input into cost effectiveness models.

Initiation and duration of selective serotonin reuptake inhibitor prescribing over time: UK cohort study.

BACKGROUND: Recent media reports have focused on the large increase in antidepressants dispensed in England. We investigated this, focusing on selective serotonin reuptake inhibitors (SSRIs). AIMS: To examine the rate of initiation of SSRIs over time and changes over time in the duration of prescribing episodes. METHOD: We estimated initiation and duration of SSRI prescribing from 7 025 802 individuals aged over 18 years and registered with a general practice that contributed data to The Health Improvement Network. RESULTS: Rates of SSRI initiation increased from 1.03 per 100 person-years in 1995 to 2.15 in 2001, but remained stable from then to 2012. The median duration of prescribing episodes increased from 112 to 169 days for episodes starting in 1995 to 2010. CONCLUSIONS: Despite media reports describing an increasing rate of antidepressant prescribing, SSRI initiation rates have stabilised since 2001. However, our results suggest that individuals who take SSRIs are receiving treatment for longer.

Backdating of events in electronic primary health care data: should one censor at the date of last data collection.

PURPOSE: Studies using primary care databases often censor follow-up at the date data are last collected from clinical computer systems (last collection date (LCD)). We explored whether this results in the selective exclusion of events entered in the electronic health records after their date of occurrence, that is, backdated events. METHODS: We used data from The Health Improvement Network (THIN). Using two versions of the database, we identified events that were entered into a later (THIN14) but not an earlier version of the database (THIN13) and investigated how the number of entries changed as a function of time since LCD. Times between events and the dates they were recorded were plotted as a function of time since the LCD in an effort to determine appropriate points at which to censor follow-up. RESULTS: There were 356 million eligible events in THIN14 and 355 million eligible events in THIN13. When comparing the two data sets, the proportion of missing events in THIN13 was highest in the month prior to the LCD (9.6%), decreasing to 5.2% at 6 months and 3.4% at 12 months. The proportion of missing events was largest for events typically diagnosed in secondary care such as neoplasms (28% in the month prior to LCD) and negligible for events typically diagnosed in primary care such as respiratory events (2% in the month prior to LCD). CONCLUSIONS: Studies using primary care databases, particularly those investigating events typically diagnosed outside primary care, should censor follow-up prior to the LCD to avoid underestimation of event rates.

Estimating the diagnostic accuracy of rheumatoid factor in UK primary care: a study using the Clinical Practice Research Datalink.

OBJECTIVE: To investigate the diagnostic accuracy of RF as a test for RA in primary care and its impact on referral times using the Clinical Practice Research Datalink. METHODS: We identified all patients with a first RF test recorded in the Clinical Practice Research Datalink between 1 January 2000 and 31 December 2008 and those diagnosed with RA within 2 years of testing. We calculated likelihood ratios (LRs), sensitivity, specificity and predictive values of RF for a diagnosis of RA. We compared time to hospital referral in those testing positive and negative using Kaplan-Meier failure curves and log-rank tests. RESULTS: Of 62 436 first RF tests, 4679 (7.5%) were positive. There were 1753 incident cases of RA, of which 57.8% were seropositive. The positive LR for RF was 9.5 (95% CI 9.0, 10.0) and the negative LR was 0.5 (95% CI 0.4, 0.5). Sensitivity and specificity were 57.8% (95% CI 55.4%, 60.1%) and 93.9% (95% CI 93.7%, 94.1%) and the positive predictive value and negative predictive value were 21.4% (95% CI 20.3%, 22.6%) and 98.7% (95% CI 98.6%, 98.8%), respectively. Median time to first hospital contact after the first RF test in those with seropositive vs seronegative results was 54 days (95% CI 49, 58) vs 150 (95% CI 147, 152). CONCLUSION: Only 2.8% of patients undergoing RF testing were diagnosed with RA, suggesting that RF is used to screen patients with musculoskeletal symptoms rather than those with more specific features of RA. A positive RF test may be helpful in diagnosing RA in primary care but performs badly in excluding RA and may delay referral.

Missing laboratory test data in electronic general practice records: analysis of rheumatoid factor recording in the clinical practice research datalink.

PURPOSE: To investigate whether information from the literature could be used to identify periods of practice data in an electronic healthcare database during which rheumatoid factor (RF) test results are likely to be missing-not-at-random (MNAR). METHODS: RF tests recorded in the Clinical Practice Research Datalink (CPRD) were identified and defined as having a positive, negative or missing result. The proportion of positive test results was then calculated based on (i) complete-case analysis (ii) after restriction to tests from practice years with no missing test results and (iii) following multiple imputation of missing test results. The same three analyses were then carried out after excluding practice years with a proportion of positive tests incompatible with the missing completely at random (MCAR) assumption. RESULTS: We identified 127,969 RF test records, 30.4% of which did not have an associated test result. Among tests with results available, 19% were positive. Both multiple imputation of the 38,867 missing test results and restriction of the study population to the 491 practice years with complete data had little impact on the percentage of positive tests. Following exclusion of the 544 practice years in which data were likely to be MNAR the percentage of positive tests in all analyses decreased to ~7%. CONCLUSIONS: Recording of RF tests and RF test results in the CPRD is incomplete, with data likely to be MNAR in many practices. Exclusion of practice years with a high proportion of positive tests brought the distribution of positive tests in the study in line with the literature.

Heritability of spinal curvature and its relationship to disc degeneration and bone mineral density in female adult twins.

PURPOSE: To determine the heritability of spine curve using plain radiographs and to identify risk factors for spine curvature including age, body mass index, smoking, bone mineral density (BMD), and lumbar disc degeneration (LDD). METHODS: A classical twin study of 110 MZ and 136 DZ adult female twins. Demographic and clinical information obtained from long spine radiographs, lumbar spine degeneration on spine MR scan, and BMD assessed by DEXA at hip and lumbar spine were included in multiple logistic regression models to determine risk factors for spine curvature. RESULTS: Heritability estimates ranged between 41 (19-59) % for pelvic incidence to 61 (46-72) % for thoracic kyphosis; with lumbar lordosis and cervical lordosis having 59 (42-71) % and 43 (23-59) % heritability, respectively. For each spine curve, the model showing the best fit contained additive genetic and shared environmental components with no contribution from the unique environment. Significant risk factors for increased thoracic kyphosis were lumbar spine BMD, age, and cervical lordosis; for pelvic incidence were lumbar spine BMD and lumbar lordosis; for lumbar lordosis were cervical lordosis, pelvic incidence and LDD; and age alone predicted cervical lordosis (p = 0.001). CONCLUSION: In this sample of middle-aged and elderly women, there were significant genetic influences on all spine curves but particularly thoracic kyphosis and lumbar lordosis. The strongest predictor for lumbar lordosis was LDD (p < 0.0001) which is itself genetically determined in part. For thoracic kyphosis, BMD was strongly associated and remained so (for lumbar BMD) with the inclusion of age, showing BMD to be an independent risk factor. This work highlights the genetic factors influencing normal spine curvature in women.

The Clinical Practice Research Datalink for drug safety in pregnancy research: an overview.

Medicine use during pregnancy is common; however the safety of medicine use during pregnancy is largely unknown when a medicine comes to market. Electronic healthcare databases, including the Clinical Practice Research Datalink (CPRD), are increasingly being used for post-marketing surveillance in this field. The CPRD contains anonymised, longitudinal medical records routinely collected in primary care. Using CPRD data it is possible to identify medical records indicative of pregnancy, including pregnancy losses. Data on prescriptions issued can be used to determine maternal exposure and for about 80% of pregnancies it is possible to link the mother's medical record to the medical record of the child. Data in the medical records of the mother and child can then be used to identify adverse pregnancy outcomes, including congenital malformations. This paper describes some of the complexities involved in using CPRD data for pregnancy related research and discusses some of its strengths and limitations.

Transporting Comparative Effectiveness Evidence Between Countries: Considerations for Health Technology Assessments.

Internal validity is often the primary concern for health technology assessment agencies when assessing comparative effectiveness evidence. However, the increasing use of real-world data from countries other than a health technology assessment agency's target population in effectiveness research has increased concerns over the external validity, or "transportability", of this evidence, and has led to a preference for local data. Methods have been developed to enable a lack of transportability to be addressed, for example by accounting for cross-country differences in disease characteristics, but their consideration in health technology assessments is limited. This may be because of limited knowledge of the methods and/or uncertainties in how best to utilise them within existing health technology assessment frameworks. This article aims to provide an introduction to transportability, including a summary of its assumptions and the methods available for identifying and adjusting for a lack of transportability, before discussing important considerations relating to their use in health technology assessment settings, including guidance on the identification of effect modifiers, guidance on the choice of target population, estimand, study sample and methods, and how evaluations of transportability can be integrated into health technology assessment submission and decision processes.

Estimating the prevalence of diagnosed Alzheimer disease in England across deprivation groups using electronic health records: a clinical practice research datalink study.

OBJECTIVE: Estimate the prevalence of diagnosed Alzheimer's disease (AD) and early Alzheimer's disease (eAD) overall and stratified by age, sex and deprivation and combinations thereof in England on 1 January 2020. DESIGN: Cross-sectional. SETTING: Primary care electronic health record data, the Clinical Practice Research database linked with secondary care data, Hospital Episode Statistics (HES) and patient-level deprivation data, Index of Multiple Deprivation (IMD). OUTCOME MEASURES: The prevalence per 100 000 of the population and corresponding 95% CIs for both diagnosed AD and eAD overall and stratified by covariates. Sensitivity analyses were conducted to assess the sensitivity of the population definition and look-back period. RESULTS: There were 448 797 patients identified in the Clinical Practice Research Datalink that satisfied the study inclusion criteria and were eligible for HES and IMD linkage. For the main analysis of AD and eAD, 379 763 patients are eligible for inclusion in the denominator. This resulted in an estimated prevalence of diagnosed AD of 378.39 (95% CI, 359.36 to 398.44) per 100 000 and eAD of 292.81 (95% CI, 276.12 to 310.52) per 100 000. Prevalence estimates across main and sensitivity analyses for the entire AD study population were found to vary widely with estimates ranging from 137.48 (95% CI, 127.05 to 148.76) to 796.55 (95% CI, 768.77 to 825.33). There was significant variation in prevalence of diagnosed eAD when assessing the sensitivity with the look-back periods, as low as 120.54 (95% CI, 110.80 to 131.14) per 100 000, and as high as 519.01 (95% CI, 496.64 to 542.37) per 100 000. CONCLUSIONS: The study found relatively consistent patterns of prevalence across both AD and eAD populations. Generally, the prevalence of diagnosed AD increased with age and increased with deprivation for each age category. Women had a higher prevalence than men. More granular levels of stratification reduced patient numbers and increased the uncertainty of point prevalence estimates. Despite this, the study found a relationship between deprivation and prevalence of AD.

Real-world evaluation of ocrelizumab in multiple sclerosis: A systematic review.

Across its clinical development program, ocrelizumab demonstrated efficacy in improving clinical outcomes in multiple sclerosis, including annualized relapse rates and confirmed disability progression. However, as with any new treatment, it was unclear how this efficacy would translate into real-world clinical practice. The objective of this study was to systematically collate the published real-world clinical effectiveness data for ocrelizumab in relapsing remitting multiple sclerosis and primary progressive multiple sclerosis. A search strategy was developed in MEDLINE and Embase to identify articles reporting real-world evidence in people with relapsing remitting multiple sclerosis or primary progressive multiple sclerosis receiving treatment with ocrelizumab. The search focused on English language articles only but was not limited by the country in which the study was conducted or the time frame of the study. Additional manual searches of relevant websites were also performed. Fifty-two studies were identified reporting relevant evidence. Real-world effectiveness data for ocrelizumab were consistently favorable, with reductions in relapse rate and disease progression rates similar to those reported in the OPERA I/OPERA II and ORATORIO clinical trials, including in studies with more diverse patient populations not well represented in the pivotal trials. Although direct comparisons are confounded by lack of randomization of treatments, outcomes reported suggest that ocrelizumab has a similar or greater efficacy than other therapy options. Initial real-world effectiveness data for ocrelizumab appear favorable and consistent with results reported in clinical trials, providing clinicians with an efficacious option to treat patients with multiple sclerosis.

Unmeasured confounding in nonrandomized studies: quantitative bias analysis in health technology assessment.

Evidence generated from nonrandomized studies (NRS) is increasingly submitted to health technology assessment (HTA) agencies. Unmeasured confounding is a primary concern with this type of evidence, as it may result in biased treatment effect estimates, which has led to much criticism of NRS by HTA agencies. Quantitative bias analyses are a group of methods that have been developed in the epidemiological literature to quantify the impact of unmeasured confounding and adjust effect estimates from NRS. Key considerations for application in HTA proposed in this article reflect the need to balance methodological complexity with ease of application and interpretation, and the need to ensure the methods fit within the existing frameworks used to assess nonrandomized evidence by HTA bodies.

Application of quantitative bias analysis for unmeasured confounding in cost-effectiveness modelling.

Due to uncertainty regarding the potential impact of unmeasured confounding, health technology assessment (HTA) agencies often disregard evidence from nonrandomized studies when considering new technologies. Quantitative bias analysis (QBA) methods provide a means to quantify this uncertainty but have not been widely used in the HTA setting, particularly in the context of cost-effectiveness modelling (CEM). This study demonstrated the application of an aggregate and patient-level QBA approach to quantify and adjust for unmeasured confounding in a simulated nonrandomized comparison of survival outcomes. Application of the QBA output within a CEM through deterministic and probabilistic sensitivity analyses and under different scenarios of knowledge of an unmeasured confounder demonstrates the potential value of QBA in HTA.

The Predictive and Prognostic Nature of Programmed Death-Ligand 1 in Malignant Pleural Mesothelioma: A Systematic Literature Review.

Introduction: Given the emergence of combination of programmed cell death protein-1 and CTLA4 pathway blockade as effective treatment options in malignant pleural mesothelioma (MPM), there is interest in the extent to which programmed death-ligand 1 (PD-L1) expression may be prognostic of clinical outcomes and predictive of response to anti-programmed death (ligand) 1 (PD-[L]1) therapies. Methods: MEDLINE and EMBASE electronic databases were searched until November 4, 2020. English-language randomized trials and observational studies that reported clinical outcomes and PD-L1 expression in adult patients (>18 or >20 y) with MPM were included. Forest plots were used to descriptively summarize clinical outcome data across studies. Results: A total of 29 publications were identified providing data on the research question. Among the studies in which anti-PD-(L)1 therapies were not specified to have been used, 63% (10 of 16) found patients with tumors expressing PD-L1 (typically >1%) to have poorer survival than those with tumors expressing lower levels of PD-L1. Among the studies in which anti-PD-(L)1 therapies were used, 83% (five of six) did not reveal an association between survival and PD-L1 tumor expression. The single study directly comparing outcomes between those treated and untreated with anti-PD-(L)1 therapies across different PD-L1 cutoffs did not identify any differences between the groups. Conclusions: The quality and consistency of the existing evidence base are currently insufficient to draw conclusions regarding a prognostic or predictive role of PD-L1 in MPM. Furthermore, high-quality studies on this topic are required to support the use of PD-L1 as a biomarker in MPM.