Reporting of harms in oncological clinical study reports submitted to the European Medicines Agency compared to trial registries and publications-a methodological review.

BACKGROUND: An accurate and comprehensive assessment of harms is a fundamental part of an accurate weighing of benefits and harms of an intervention when making treatment decisions; however, harms are known to be underreported in journal publications. Therefore, we sought to compare the completeness of reporting of harm data, discrepancies in harm data reported, and the delay to access results of oncological clinical trials between three sources: clinical study reports (CSRs), clinical trial registries and journal publications. METHODS: We used the EMA clinical data website to identify all trials submitted to the EMA between 2015 and 2018. We retrieved all CSRs and included all phase II, II/III or III randomised controlled trials (RCTs) assessing targeted therapy and immunotherapy for cancer. We then identified related records in clinical trial registries and journals. We extracted harms data for eight pre-specified variables and determined the completeness of reporting of harm data in each of the three sources. RESULTS: We identified 42 RCTs evaluating 13 different drugs. Results were available on the EMA website in CSRs for 37 (88%) RCTs, ClinicalTrials.gov for 36 (86%), the European Clinical Trials Register (EUCTR) for 20 (48%) and in journal publications for 32 (76%). Harms reporting was more complete in CSRs than other sources. We identified marked discrepancies in harms data between sources, e.g. the number of patients discontinuing due to adverse events differed in CSRs and clinical trial registers for 88% of trials with data in both sources. For CSRs and publications, the corresponding number was 90%. The median (interquartile range) delay between the primary trial completion date and access to results was 4.34 (3.09-7.22) years for CSRs, 2.94 (1.16-4.52) years for ClinicalTrials.gov, 5.39 (4.18-7.33) years for EUCTR and 2.15 (0.64-5.04) years for publications. CONCLUSIONS: Harms of recently approved oncological drugs were reported more frequently and in more detail in CSRs than in trial registries and journal publications. Systematic reviews seeking to address harms of oncological treatments should ideally use CSRs as the primary source of data; however, due to problems with access, this is currently not feasible.

Results of trials assessing transarterial chemoembolization for treating hepatocellular carcinoma are critically underreported.

OBJECTIVE: We aimed to evaluate to what extent the results of registered randomized controlled trials (RCTs) assessing transarterial chemoembolization (TACE) for treating hepatocellular carcinoma (HCC) are publicly available. METHODS: We searched the Cochrane Central Register of Controlled Trials, the International Clinical Trials Registry Platform, and primary registries for RCTs assessing TACE for treating HCC, registered between January 2008 and August 2018, that had exceeded their completion date by more than 1 year. We systematically searched PubMed, EMBASE, and Google Scholar for a publication as well as the registry for results. The main outcomes were the availability of results, and the time to the first availability of results (i.e., posted on the registry or published). Secondary outcomes were the proportion of results available at 12 and 36 months after completion. RESULTS: Among 67 identified RCTs, including a total target number to 11,599 participants, 26 had publicly available results (39%; i.e., 42% of total target number of participants). Results of 25 RCTs (37%) were published, with only 3 having results posted on the registry and 2 with both published and posted results. The median (Q1-Q3) time from completion to the first public availability of results was 18 months (11-29). The cumulative percentages of RCTs with results available were 10% (95% CI, 3-17%) and 29% (95% CI, 17-39%) at 12 and 36 months, respectively, after completion. CONCLUSIONS: Despite the ethical commitments and societal expectations for disclosure of results, the availability of results of RCTs on TACE for treating HCC is very limited. KEY POINTS: • Underreporting of trial results is a major cause of wasted medical research since inaccessible research results fail to help both patients and clinicians. • Transarterial chemoembolization (TACE) is the most common treatment for hepatocellular carcinoma (HCC) and has called for considerable research efforts. • Yet, almost two-thirds of randomized controlled trials assessing TACE for treating HCC did not yield any public results, either on the registry platform or in scientific journals.

Accuracy in detecting inadequate research reporting by early career peer reviewers using an online CONSORT-based peer-review tool (COBPeer) versus the usual peer-review process: a cross-sectional diagnostic study.

BACKGROUND: The peer review process has been questioned as it may fail to allow the publication of high-quality articles. This study aimed to evaluate the accuracy in identifying inadequate reporting in RCT reports by early career researchers (ECRs) using an online CONSORT-based peer-review tool (COBPeer) versus the usual peer-review process. METHODS: We performed a cross-sectional diagnostic study of 119 manuscripts, from BMC series medical journals, BMJ, BMJ Open, and Annals of Emergency Medicine reporting the results of two-arm parallel-group RCTs. One hundred and nineteen ECRs who had never reviewed an RCT manuscript were recruited from December 2017 to January 2018. Each ECR assessed one manuscript. To assess accuracy in identifying inadequate reporting, we used two tests: (1) ECRs assessing a manuscript using the COBPeer tool (after completing an online training module) and (2) the usual peer-review process. The reference standard was the assessment of the manuscript by two systematic reviewers. Inadequate reporting was defined as incomplete reporting or a switch in primary outcome and considered nine domains: the eight most important CONSORT domains and a switch in primary outcome(s). The primary outcome was the mean number of domains accurately classified (scale from 0 to 9). RESULTS: The mean (SD) number of domains (0 to 9) accurately classified per manuscript was 6.39 (1.49) for ECRs using COBPeer versus 5.03 (1.84) for the journal's usual peer-review process, with a mean difference [95% CI] of 1.36 [0.88-1.84] (p < 0.001). Concerning secondary outcomes, the sensitivity of ECRs using COBPeer versus the usual peer-review process in detecting incompletely reported CONSORT items was 86% [95% CI 82-89] versus 20% [16-24] and in identifying a switch in primary outcome 61% [44-77] versus 11% [3-26]. The specificity of ECRs using COBPeer versus the usual process to detect incompletely reported CONSORT domains was 61% [57-65] versus 77% [74-81] and to identify a switch in primary outcome 77% [67-86] versus 98% [92-100]. CONCLUSIONS: Trained ECRs using the COBPeer tool were more likely to detect inadequate reporting in RCTs than the usual peer review processes used by journals. Implementing a two-step peer-review process could help improve the quality of reporting. TRIAL REGISTRATION: Clinical.Trials.gov NCT03119376 (Registered April, 18, 2017).

Avoidable waste of research related to outcome planning and reporting in clinical trials.

BACKGROUND: Inadequate planning, selective reporting, and incomplete reporting of outcomes in randomized controlled trials (RCTs) contribute to the problem of waste of research. We aimed to describe such a waste and to examine to what extent this waste could be avoided. METHODS: This research-on-research study was based on RCTs included in Cochrane reviews with a summary of findings (SoF) table. We considered the outcomes reported in the SoF tables as surrogates for important outcomes for patients and other decision makers. We used a three-step approach. (1) First, in each review, we identified, for each important outcome, RCTs that were excluded from the corresponding meta-analysis. (2) Then, for these RCTs, we systematically searched for registrations and protocols to distinguish between inadequate planning (an important outcome was not reported in registries or protocols), selective reporting (an important outcome was reported in registries or protocols but not in publications), and incomplete reporting (an important outcome was incompletely reported in publications). (3) Finally, we assessed, with the consensus of five experts, the feasibility and cost of measuring the important outcomes that were not planned. We considered inadequately planned or selectively or incompletely reported important outcomes as avoidable waste if the outcome could have been easily measured at no additional cost based on expert evaluation. RESULTS: Of the 2711 RCTs included in the main comparison of 290 reviews, 2115 (78%) were excluded from at least one meta-analysis of important outcomes. Every trial contributed to 55%, on average, of the meta-analyses of important outcomes. Of the 310 RCTs published in 2010 or later, 156 were registered. Inadequate planning affected 79% of these RCTs, whereas incomplete and selective reporting affected 41% and 15%, respectively. For 63% of RCTs, we found at least one missing important outcome for which the waste was avoidable and for 30%, the waste was avoidable for all important outcomes. CONCLUSIONS: Most of the RCTs included in our sample did not contribute to all the important outcomes in meta-analyses, mostly because of inadequate planning or incomplete reporting. A large part of this waste of research seemed to be avoidable.

Mapping of Crowdsourcing in Health: Systematic Review.

BACKGROUND: Crowdsourcing involves obtaining ideas, needed services, or content by soliciting Web-based contributions from a crowd. The 4 types of crowdsourced tasks (problem solving, data processing, surveillance or monitoring, and surveying) can be applied in the 3 categories of health (promotion, research, and care). OBJECTIVE: This study aimed to map the different applications of crowdsourcing in health to assess the fields of health that are using crowdsourcing and the crowdsourced tasks used. We also describe the logistics of crowdsourcing and the characteristics of crowd workers. METHODS: MEDLINE, EMBASE, and ClinicalTrials.gov were searched for available reports from inception to March 30, 2016, with no restriction on language or publication status. RESULTS: We identified 202 relevant studies that used crowdsourcing, including 9 randomized controlled trials, of which only one had posted results at ClinicalTrials.gov. Crowdsourcing was used in health promotion (91/202, 45.0%), research (73/202, 36.1%), and care (38/202, 18.8%). The 4 most frequent areas of application were public health (67/202, 33.2%), psychiatry (32/202, 15.8%), surgery (22/202, 10.9%), and oncology (14/202, 6.9%). Half of the reports (99/202, 49.0%) referred to data processing, 34.6% (70/202) referred to surveying, 10.4% (21/202) referred to surveillance or monitoring, and 5.9% (12/202) referred to problem-solving. Labor market platforms (eg, Amazon Mechanical Turk) were used in most studies (190/202, 94%). The crowd workers' characteristics were poorly reported, and crowdsourcing logistics were missing from two-thirds of the reports. When reported, the median size of the crowd was 424 (first and third quartiles: 167-802); crowd workers' median age was 34 years (32-36). Crowd workers were mainly recruited nationally, particularly in the United States. For many studies (58.9%, 119/202), previous experience in crowdsourcing was required, and passing a qualification test or training was seldom needed (11.9% of studies; 24/202). For half of the studies, monetary incentives were mentioned, with mainly less than US $1 to perform the task. The time needed to perform the task was mostly less than 10 min (58.9% of studies; 119/202). Data quality validation was used in 54/202 studies (26.7%), mainly by attention check questions or by replicating the task with several crowd workers. CONCLUSIONS: The use of crowdsourcing, which allows access to a large pool of participants as well as saving time in data collection, lowering costs, and speeding up innovations, is increasing in health promotion, research, and care. However, the description of crowdsourcing logistics and crowd workers' characteristics is frequently missing in study reports and needs to be precisely reported to better interpret the study findings and replicate them.

Comparative efficacy and safety of second-line treatments for advanced non-small cell lung cancer with wild-type or unknown status for epidermal growth factor receptor: a systematic review and network meta-analysis.

BACKGROUND: Docetaxel, pemetrexed, erlotinib, and gefitinib are recommended as second-line treatment for advanced non-small cell lung cancer (NSCLC) with wild-type or unknown status for epidermal growth factor receptor (EGFR). However, the number of published randomized clinical trials (RCTs) on this topic is increasing. Our objective was to assess the comparative effectiveness and tolerability of all second-line treatments for advanced NSCLC with wild-type or unknown status for EGFR by a systematic review and network meta-analysis. METHODS: MEDLINE, EMBASE, CENTRAL, ClinicalTrials.gov, and the US Food and Drug Administration website, as well as other sources, were searched for available reports up to June 6, 2017. Two reviewers independently selected published and unpublished reports of RCTs comparing any second-line treatments, extracted data and assessed the risk of bias of all included trials. We performed a Bayesian network meta-analysis. The primary outcomes were overall survival (OS) and progression-free survival (PFS). Secondary outcomes included objective response (ObR), the number of serious adverse events, and quality of life. RESULTS: We included 102 RCTs involving 36,058 patients (62% male, median age 61 years, 81% with stage IV cancer, 80% smokers, and 92% with performance status 0-1). We revealed a differential reporting of outcomes between efficacy and safety outcomes. Half of the trials reported safety outcomes and less than 20% quality of life. For OS, nivolumab was more effective than docetaxel (hazard ratio (HR) 0.69, 95% credible interval (CrI) 0.56-0.83), pemetrexed (0.67, 0.52-0.83), erlotinib (0.68, 0.53-0.86), and gefitinib (0.66, 0.53-0.83). Pembrolizumab, atezolizumab, and pemetrexed plus erlotinib were also significantly more effective than docetaxel, pemetrexed, erlotinib, and gefitinib. For PFS, erlotinib plus cabozantinib was more effective than docetaxel (HR 0.39, 95% CrI 0.18-0.84), pemetrexed (0.38, 0.18-0.82), erlotinib (0.37, 0.18-0.78), and gefitinib (0.38, 0.18-0.82). Cabozantinib and pemetrexed plus erlotinib were also significantly more effective than the four recommended treatments. For ObR, no treatment was significantly more effective. The effectiveness of the four recommended treatments was similar and they were ranked among the 25 less-effective treatments. For safety, evidence is insufficient to draw certain conclusions. CONCLUSIONS: Nivolumab, pembrolizumab, atezolizumab, and pemetrexed plus erlotinib may be the most effective second-line treatments for NSCLC in terms of OS. The four recommended treatments seem to have relatively poor performance. However, the impact on life expectancy of immunotherapy versus other treatments should be further explored by future analyses, and more trials comparing the novel treatments are needed to reduce uncertainty in these results. TRIAL REGISTRATION: Registration number: PROSPERO ( CRD42015017592 ).

Wasted research when systematic reviews fail to provide a complete and up-to-date evidence synthesis: the example of lung cancer.

BACKGROUND: Multiple treatments are frequently available for a given condition, and clinicians and patients need a comprehensive, up-to-date synthesis of evidence for all competing treatments. We aimed to quantify the waste of research related to the failure of systematic reviews to provide a complete and up-to-date evidence synthesis over time. METHODS: We performed a series of systematic overviews and networks of randomized trials assessing the gap between evidence covered by systematic reviews and available trials of second-line treatments for advanced non-small cell lung cancer. We searched the Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, MEDLINE, EMBASE, and other resources sequentially by year from 2009 to March 2, 2015. We sequentially compared the amount of evidence missing from systematic reviews to the randomized evidence available for inclusion each year. We constructed cumulative networks of randomized evidence over time and evaluated the proportion of trials, patients, treatments, and treatment comparisons not covered by systematic reviews on December 31 each year from 2009 to 2015. RESULTS: We identified 77 trials (28,636 patients) assessing 47 treatments with 54 comparisons and 29 systematic reviews (13 published after 2013). From 2009 to 2015, the evidence covered by existing systematic reviews was consistently incomplete: 45 % to 70 % of trials; 30 % to 58 % of patients; 40 % to 66 % of treatments; and 38 % to 71 % of comparisons were missing. In the cumulative networks of randomized evidence, 10 % to 17 % of treatment comparisons were partially covered by systematic reviews and 55 % to 85 % were partially or not covered. CONCLUSIONS: We illustrate how systematic reviews of a given condition provide a fragmented, out-of-date panorama of the evidence for all treatments. This waste of research might be reduced by the development of live cumulative network meta-analyses.

Pharmacogenomic biomarkers as inclusion criteria in clinical trials of oncology-targeted drugs: a mapping of ClinicalTrials.gov.

PURPOSE: The aim of this study was to describe pharmacogenomics-based inclusion criteria (enrichment) and the main characteristics of clinical trials involving oncology-targeted therapies. METHODS: Clinical trials of oncology-targeted therapies approved after 2005 with pharmacogenomic testing required or recommended in their label were retrieved from a mapping of the ClinicalTrials.gov database. RESULTS: We examined information for 12 drugs and 858 trials. Overall, 434 trials (51%) were enriched on the biomarker first mentioned in the label and 145 (17%) were enriched on another biomarker, whereas 270 trials (31%) included all patients. The median proportion of trials corresponding to both the drug's indication and drug's target was 35%. Of the 361 trials that tested drugs in another disease than the first one in the label, 219 (61%) were without enrichment and 87 (24%) were actually enriched but on another biomarker than the first one in the label. CONCLUSION: Several drugs have been tested in trials enriched on many different biomarkers. Nonetheless, most targeted therapies have been developed only using biomarker-positive patients; therefore, exclusion of biomarker-negative patients from treatment relies on only preclinical data and on biological understanding of the disease and target.Genet Med 18 8, 796-805.

A Multicenter Study Assessing the Real-World Use and Effectiveness of First-Line Chemotherapy Plus Immunotherapy in Advanced Small-Cell Lung Cancer (SCLC) Patients.

BACKGROUND: First-line chemotherapy plus immunotherapy (CT-IO) has recently demonstrated survival benefits over CT alone in extensive-stage small-cell lung cancer (ES-SCLC), based on randomized phase III studies. This retrospective multicenter study assessed the real-world use and effectiveness of CT-IO in ES-SCLC patients. PATIENTS AND METHODS: All newly diagnosed ES-SCLC patients from 4 French hospitals treated with CT alone or CT-IO between May 2020 and December 2021 were included. Overall survival (OS) and real-world progression-free survival (rwPFS) were estimated using the Kaplan-Meier method. Cox proportional hazard models were performed to estimate hazard ratios (HRs) with 95 % confidence intervals (CIs) in univariate and multivariate models. The aim was not to compare efficacy between groups. RESULTS: Among 104 patients, 75 (72.1%) received CT-IO. Brain metastases were diagnosed in 28.3% of patients, and 29.8% were performance status (PS) ≥ 2. At a median follow-up of 16.8 months (95%CI, 14.9-23.4), the median OS was 11.4 months (95%CI, 7.7-14.7) in the CT-IO group, and the 12-month OS rate was 43.6% (95%CI, 33.3-57.2). In the CT group, the median OS was 7.8 months (95%CI, 5.4-11.8) and the 12-month OS rate was 15.3% (95%CI, 5.7-41.0). In multivariate analyses, baseline brain and liver metastases were associated with a shorter OS for patients treated in the CT-IO group (HR, 3.80 [95%CI, 1.90-7.60] and 3.12 [95%CI, 1.60-6.08] respectively; P < 0.001 for both). CONCLUSION: We showed that clinicians have chosen to use IO beyond the specific criteria defined in guidelines. Survival data appeared promising with a median OS comparable to the one previously demonstrated in clinical trials.

Prospective Radiologic-Pathologic Correlation of Macroscopic Volume and Microscopic Extension of Nonsolid Lung Nodules on Thin-section CT Images for Sublobar Resection and Stereotactic Radiotherapy Planning.

INTRODUCTION: The objective of this study was to determine whether computed tomography (CT) could be a useful tool for nonsolid lung nodule (NSN) treatment planning, surgery or stereotactic body radiation therapy (SBRT), by assessing the macroscopic and microscopic extension of these nodules. METHODS: The study prospectively included 23 patients undergoing anatomic resection at the Foch Hospital in 2020/2021 for NSN with a ground-glass component of more than 50%. Firstly, for each patient, both the macroscopic dimensions of the NSN were assessed on CT and during pathologic analysis. Secondly, the microscopic extension was assessed during pathologic examination. Wilcoxon sign rank tests were used to compare these dimensions. Spearman correlation test and Bland-Altman analysis were used to evaluate the agreement between radiological and pathologic measurements. RESULTS: On CT, the median largest diameter and volume of NSN were 21 mm and 3780 cc, while on pathologic analysis, they were 15 mm and 1800 cc, respectively. Therefore, the largest diameter and volume of the NSN were significantly higher on CT than on pathological analysis. For microscopic extension, the median largest diameter and volume of NSN were 17 mm and 2040 cc, respectively. No significant difference was observed between the macroscopic size and the microscopic extension assessed during pathologic analysis. Moreover, correlation analysis and Bland-Altman plots showed that radiological and pathologic measurements could provide equivalent precision. CONCLUSION: Our study showed that CT did not underestimate the macroscopic size and microscopic extension of NSN and confirmed that CT can be used for NSN treatment planning.

Availability of results of academic randomized trials involving cooperative groups in oncology in France: A systematic search of clinical trial registries.

BACKGROUND: Cooperative groups' involvement is increasing in academic oncological research. We aimed to assess the impact of sponsoring by cooperative groups in France on the availability of results of academic randomized trials in oncology. METHODS: We performed a systematic search using ClinicalTrials.gov and the European Clinical Trials Register. We searched for all academic randomized trials in oncology conducted in France between January 1, 2005 and January 1, 2015. The inclusion criteria were: completed or terminated, phase 2 or 3 randomized trials with an academic (non-industry) sponsor. The main outcome was the publication of the results of trial (either as a journal article or as posting results in a registry) across each type of sponsor. RESULTS: We included 211 randomized trials, mainly phase 3 (n = 135, 64%) and evaluating pharmacological treatments (n = 149, 71%). French cooperative groups were involved in 69 trials (33%), as part of a collaboration in one third (n = 23) of instances. Seventy-one (34%) trials were run by oncologic hospitals, 50 (23%) by university hospitals, and 21 (10%) by European organizations. Seventy-seven randomized trials (36%) had available results (published n = 73, posted n = 6). Cooperative groups were involved in half of those that have been published (37/73). The cumulative probability of results availability was 57% for cooperative groups, 41% for European organizations, 32% for oncologic hospitals, and 17% for university hospital at 10 years from the beginning of trials (p = 0.0006). In the case of collaboration with cooperative groups, the cumulative probability of results availability achieved 59% for university hospitals and 74% for oncologic hospitals. CONCLUSION: The availability of results of randomized trials in oncology remains limited and almost exclusively through publications, but is higher when cooperative groups are involved. POLICY SUMMARY: Sponsoring by a cooperative group should become the rule in academic trials to increase availability of trial results.

Published trials of TACE for HCC are often not registered and subject to outcome reporting bias.

BACKGROUND & AIMS: In 2005, the registration of all randomised controlled trials (RCTs) before enrolment of participants became a condition for publication by the International Committee of Medical Journal Editors to increase transparency in trial reporting. Among RCTs on transarterial chemoembolisation (TACE) for the treatment of hepatocellular carcinoma (HCC) published after 2007, we assess the proportion that were registered and compare registered primary outcomes (PO) with those reported in publications to determine whether primary outcome reporting bias favoured significant outcomes. METHODS: We searched MEDLINE and EMBASE for reports of RCTs evaluating TACE for HCC treatment between 1 September 2007 and 31 March 2018. Registration and publication information for each included RCT was compared using a standardised data extraction form. RESULTS: Thirteen out of 53 (25%) included RCTs were correctly registered (i.e. before the starting date of the RCT), 14 (26%) were registered after the RCT starting date, and 26 (49%) were not registered. Six out of 14 of the retrospectively registered RCTs (43%) were registered after their completion date. The PO was clearly reported in the published article of all registered RCTs, whereas the report was not clear in 8/26 (31%) of the non-registered RCTs (p = 0.01). Among registered RCTs, 8/27 (30%) had major discrepancies between registered and published PO. The influence of these discrepancies could be assessed in 6 of them and was shown to statistically favour significant results in 2. CONCLUSIONS: Registration and outcome reporting in RCTs on TACE for HCC are often inadequate. Registration should be reinforced because it is a key to transparency. LAY SUMMARY: Trial registration is fundamental to our understanding and interpretation of results, as it provides information on all relevant clinical trials (to place the results in a broader context), and on the details of their associated protocols (to ensure that the scientific plan is followed). Once a randomised controlled trial (RCT) is completed, the trial results are usually publicly shared via scientific articles that are expected to thoroughly and objectively report them. This study shows that half of the RCTs evaluating transarterial chemoembolisation for hepatocellular carcinoma were not registered, and identified major discrepancies between registered and published primary outcome favouring significant results.

Future of evidence ecosystem series: 3. From an evidence synthesis ecosystem to an evidence ecosystem.

The "one-off" approach of systematic reviews is no longer sustainable; we need to move toward producing "living" evidence syntheses (i.e., comprehensive, based on rigorous methods, and up-to-date). This implies rethinking the evidence synthesis ecosystem, its infrastructure, and management. The three distinct production systems-primary research, evidence synthesis, and guideline development-should work together to allow for continuous refreshing of synthesized evidence and guidelines. A new evidence ecosystem, not just focusing on synthesis, should allow for bridging the gaps between evidence synthesis communities, primary researchers, guideline developers, health technology assessment agencies, and health policy authorities. This network of evidence synthesis stakeholders should select relevant clinical questions considered a priority topic. For each question, a multidisciplinary community including researchers, health professionals, guideline developers, policymakers, patients, and methodologists needs to be established and commit to performing the initial evidence synthesis and keeping it up-to-date. Encouraging communities to work together continuously with bidirectional interactions requires greater incentives, rewards, and the involvement of health care policy authorities to optimize resources. A better evidence ecosystem with collaborations and interactions between each partner of the network of evidence synthesis stakeholders should permit living evidence syntheses to justify their status in evidence-informed decision-making.

Future of evidence ecosystem series: 1. Introduction Evidence synthesis ecosystem needs dramatic change.

OBJECTIVES: This article presents why the planning, conduct, and reporting of systematic reviews and meta-analyses of therapeutic interventions are suboptimal. STUDY DESIGN AND SETTING: We present an overview of the limitations of the current system of evidence synthesis for therapeutic interventions. RESULTS: Systematic reviews and meta-analyses are a cornerstone of health care decisions. However, despite the increasing a number of published systematic reviews of therapeutic interventions, the current evidence synthesis ecosystem is not properly addressing stakeholders' needs. The current production process leads to a series of disparate systematic reviews because of erratic and inefficient planning with a process that is not always comprehensive and is prone to bias. Evidence synthesis depends on the quality of primary research, so primary research that is not available is biased or selectively reported raises important concerns. Moreover, the lack of interactions between the community of primary research producers and systematic reviewers impedes the optimal use of data. The context has considerably evolved, with ongoing research innovations, a new medical approach with the end of the one-size-fits-all approach, more available data, and new patient expectations. All these changes must be introduced into the future evidence ecosystem. CONCLUSION: Dramatic changes are needed to enable this future ecosystem to become user driven and user oriented and more useful for decision-making.

Future of evidence ecosystem series: 2. current opportunities and need for better tools and methods.

To become user driven and more useful for decision-making, the current evidence synthesis ecosystem requires significant changes (Paper 1. Future of evidence ecosystem series). Reviewers have access to new sources of data (clinical trial registries, protocols, and clinical study reports from regulatory agencies or pharmaceutical companies) for more information on randomized control trials. With all these newly available data, the management of multiple and scattered trial reports is even more challenging. New types of data are also becoming available: individual patient data and routinely collected data. With the increasing number of diverse sources to be searched and the amount of data to be extracted, the process needs to be rethought. New approaches and tools, such as automation technologies and crowdsourcing, should help accelerate the process. The implementation of these new approaches and methods requires a substantial rethinking and redesign of the current evidence synthesis ecosystem. The concept of a "living" evidence synthesis enterprise, with living systematic review and living network meta-analysis, has recently emerged. Such an evidence synthesis ecosystem implies conceptualizing evidence synthesis as a continuous process built around a clinical question of interest and no longer as a small team independently answering a specific clinical question at a single point in time.

Use of Late-Life Expectancy for Assessing the Long-Term Benefit of Immune Checkpoint Inhibitors.

To grade the long-term benefit of anticancer agents, the American Society of Clinical Oncology Value Framework (ASCO-VF) awards tail-of-the-curve bonus points by using milestone survival at twice the median control survival. Here, we propose an alternative, late-life expectancy that we defined as the area under the Kaplan-Meier curve from median control survival to the end of follow-up. We analyzed all indications of immune checkpoint inhibitors with survival data and found that 9 indications out of 13 (69.2%) qualified for ASCO-VF tail-of-the-curve bonus points either in progression-free or overall survival. Our proposed score recognized a long-term benefit not captured by the ASCO-VF, for example, for CHECKMATE-66 where twice the median overall survival was not reached. We found that nivolumab was associated with an increase of 65.3% (95% CI = 38.9 to 89.5) in overall survival late-life expectancy, which highlights its important long-term benefit. In conclusion, the ASCO-VF could be improved with the use of late-life expectancy.

Automatic screening using word embeddings achieved high sensitivity and workload reduction for updating living network meta-analyses.

OBJECTIVES: We aimed to develop and evaluate an algorithm for automatically screening citations when updating living network meta-analysis (NMA). STUDY DESIGN AND SETTING: Our algorithm learns from the initial screening of citations conducted when creating an NMA to automatically identify eligible citations (i.e., needing full-text consideration) when updating the NMA. We evaluated our algorithm on four NMAs from different medical domains. For each NMA we constructed sets of initially screened citations and citations to screen during an update that took place 2 years after the conduct of the NMA. We encoded free text of citations (title and abstract) using word embeddings. On top of this vectorized representation, we fitted a logistic regression model to the set of initially screened citations to predict the eligibility of citations screened during an update. RESULTS: Our algorithm achieved 100% sensitivity on two NMAs (100% [95% confidence interval 93-100] and 100% [40-100] sensitivity), and 94% (81-99) and 97% (86-100) on the remaining two others. For all NMAs, our algorithm would have spared to manually screen 1,345 of 2,530 citations, decreasing the workload by 53% (51-55), while missing 3 of 124 eligible citations (2% [1-7]), none of which were finally included in the NMAs after full-text consideration. CONCLUSION: For updating an NMA after 2 years, our algorithm considerably diminished the workload required for screening, and the number of missed eligible citations remained low.

Living network meta-analysis was feasible when considering the pace of evidence generation.

OBJECTIVES: The aim of the study was to assess the feasibility of living network meta-analysis (NMA) taking into account the pace of evidence generation across different medical areas. STUDY DESIGN AND SETTING: We performed a systematic review to identify published NMAs. For each NMA, we calculated the cumulative number of new trials. To assess the feasibility of living NMA, we considered different update frequencies (4, 6, and 12 months), then evaluated the number of new trials to be included at each update in the NMA and the workload percentage for an update relative to the initial NMA. RESULTS: We identified 77 NMAs covering 17 different medical areas; 60 (78%) had fewer than four new trials included per year, on average, and 5 (7%) had more than seven trials. With an update frequency of 4, 6, and 12 months, the median number of new trials to be included in the NMA was 0 (interquartile range, 0-1), 1 (0-2), and 2 (1-4), respectively, with mean of 4%, 5%, and 11% workload per update, respectively. CONCLUSION: The workload associated with updating a living NMA represents about one-tenth of the initial workload; therefore, living NMA is manageable.

[Lung cancer in HIV-infected patients]. / Le cancer bronchopulmonaire chez les patients infectés par le VIH.

Until 1996, AIDS was the leading cause of deaths from HIV infection. In 2010, because of introduction of powerful antiretroviral therapies, AIDS represented less than 25% of deaths. Cancer has become the leading cause of death in this population, and, because of smoking and immunosuppression, lung cancer risk is more important than in general population. Furthermore, treatment is more difficult, due to potential interactions between antiretroviral and anticancer therapies, to comorbidities and to tumor aggressiveness. Research will focus on molecular biology, immunotherapies and lung cancer screening in order to improve survival of HIV patients with lung cancer. For all these reasons, HIV patients must be included in clinical trials.

Nonsmall cell lung cancer from HIV-infected patients expressed programmed cell death-ligand 1 with marked inflammatory infiltrates.

OBJECTIVE: Immunotherapies targeting the programmed cell death-1 (PD-1)/PD-ligand 1 (PD-L1) checkpoint improved prognosis in lung cancer. PD-1/PD-L1 status, however, has not been investigated in human immunodeficiency virus (HIV)-positive patients. This study assessed PD-L1 status and tumor immune-cell infiltration in nonsmall cell lung cancer (NSCLC) in HIV patients. METHODS: Consecutive HIV patients treated between 1996 and 2014 were enrolled. PD-L1 tumor expression was assessed using immunohistochemistry with two antibodies (clones 5H1 and E1L3N), and tumor immune-cell infiltration with CD3, CD4, CD8, CD20, CD163, and MPO. PD-L1 expression and immune infiltration results were compared with those of 54 NSCLCs from unknown HIV status patients. RESULTS: Thirty-four HIV-positive patients were evaluated: predominantly men (88.2%) (median age: 51.1 years) presenting stage IV (38.2%) adenocarcinomas (76.5%). The median blood CD4 count was 480 cells/µL (86-1120) and 64% exhibited undetectable viral load. The PD-L1 score (percentage of positive cells × intensity) was higher in HIV-positive than HIV-undetermined patients with the E1L3N clone [median (range) 0 (0-150) versus 0 (0-26.7), P = 0.047], yet not with the 5H1 clone [0 (0-120) versus 0 (0-26.7) P = 0.07, respectively]. PD-L1 expression frequency did not differ between both cohorts (18.7 versus 9.3% using E1L3N and 10 versus 5.6% using 5H1 clone, respectively). There were significantly greater cytotoxic T-cell (P < 0.001), B-lymphocyte (P = 0.005), and activated macrophage (P < 0.001) infiltrations in the HIV-positive patients, but no differences for CD4 T cells. CONCLUSION: Tumors in HIV-positive patients seem to express higher PD-L1 levels with increased immune infiltration, supporting their inclusion in clinical trials assessing immune checkpoint inhibitors.