A novel framework to assess haematology and oncology registration trials: The THEOREMM project.

BACKGROUND: Methodological limitations affect a significant number of oncology and haematology trials, raising concerns about the applicability of their results. For example, a suboptimal control arm or limited access to best care upon progression may skew the trial results toward a benefit in the experimental arm. Beyond the fact that such limitations do not prevent drugs reaching the market, other assessment tools, such as those developed by professional societies-ESMO-MCBS and ASCO Value Framework-do not integrate these important shortcomings. METHODS: We propose creating a novel framework with the scope of assessing registration cancer clinical trials in haematology and oncology (randomized or single arm)-that is trials leading to a marketing authorization. The main steps of the methods are (1) assembling a scientific board; (2) defining the scope, goal and methods through pre-specified, pre-registered and protocolized methodology; (3) preregistration of the protocol; (4) conducting a scoping review of limitations and biases affecting oncology trials and assessing existing scores or methods; (5) developing a list of features to be included and assessed within the framework; (6) assessing each feature through a questionnaire sent to highly cited haematologists and oncologists involved in clinical trials; and (7) finalizing the first version of framework. RESULTS: Not applicable. CONCLUSIONS: Our proposal emerged in response to the lack of consideration for key limitations in current trial assessments. The goal is to create a framework specifically designed to assess single trials leading to marketing authorization in the field of oncology and haematogy.

[Interpreting a randomized trial in oncology: Key points for the clinician]. / Interpréter un essai randomisé en oncologie : éléments clés en clinique.

The analysis of randomized clinical trials presents a challenge for clinicians. A set of critical elements can facilitate their interpretation. One must question whether the inclusion and exclusion criteria accurately mirror clinical practice. Does the control arm align with what is currently recognized as best practice? Do patients in the control group have access to the best options when the cancer progresses or recurs? The degree of confidence with which phase II trial results can be interpreted also warrants consideration. Finally, informative censoring can be searched for by comparing early censoring rates between treatment arms. Faced with the challenges of interpreting scientific literature, these keys can help the clinician and guide the eventual integration of new results into shared medical decision-making.

L'analyse d'essais cliniques randomisés est un défi pour le clinicien. Une série d'éléments clés peuvent toutefois aider à l'interprétation. Tout d'abord, les critères d'inclusion et d'exclusion reflètent-ils la pratique quotidienne ? Ensuite, le bras contrôle correspond-il aux meilleures pratiques reconnues ? Est-ce que les patients du groupe contrôle ont un accès aux meilleures options lorsque le cancer progresse ou récidive ? Avec quelle confiance interpréter des résultats de phase II ? Enfin, la censure informative peut être recherchée en comparant les taux de censure précoce entre les bras de traitements. Face aux défis de l'interprétation de la littérature scientifique, ces clés peuvent être une aide pour le clinicien et guider l'intégration éventuelle de nouveaux résultats dans la décision médicale partagée.

Eventual success rate and predictors of success for oncology drugs tested in phase I trials.

Previous estimates of the likelihood of a drug tested in phase I trials obtaining FDA clearance are out of date. In the intervening years, newer pharmaceuticals have been developed, resulting in new delivery systems and lines of therapies. We sought to explore and update these estimates by comprehensively searching drugs tested in phase I trials and to determine the factors associated with later receiving FDA approval. In a cross-sectional analysis, we searched for anti-tumor drugs tested in phase I trials and published in scientific journals or presented at hematology/oncology conferences. For each drug, we searched PubMed for phase II and phase III studies testing the drug for the same indication tested in phase I studies. We found 51 drug approvals; four were withdrawn. The probability of a drug tested in 2015 being approved by 2021 was 6.2%. Drugs tested as monotherapy were more likely to receive approval than drugs tested in combination, and monoclonal antibodies were more likely to receive approval than drugs of other mechanisms. In adjusted models, response rates higher than 40% in phase I studies, demonstrating an improvement in overall survival (OS) in phase III studies, and drugs tested as monotherapy were associated with receiving FDA approval. When looking at all drugs tested during a single year, most drugs were not approved, and among those that are approved, almost 8% are withdrawn. Response rates higher than 40%, testing a drug as monotherapy, and demonstrating an improvement in OS were associated with receiving FDA approval.

Additional considerations before using a ctDNA-guided approach for informing adjuvant chemotherapy in colorectal cancer.

BACKGROUND: The DYNAMIC trial investigated the use of circulating tumor DNA (ctDNA) to guide adjuvant treatment decisions in stage II colon cancer. Despite the DYNAMIC trial's assertion that a ctDNA-guided approach could minimize the use of adjuvant treatment without compromising recurrence-free survival (RFS), we raised concerns regarding the trial's methodology and the practical implications of its findings in a Debate article. Here, we expand upon these concerns in a response to a correspondence by the authors of the DYNAMIC trial. MAIN BODY: We dispute the choice of a large non-inferiority margin in the DYNAMIC trial, simply because an 8.5 percentage points decrease in recurrence-free survival could result in significant harm to patients. We challenge the authors' comparisons of the DYNAMIC trial outcomes with observational studies. Such comparison is subject to selection bias and changes over time that limit their relevance. The prognostic role of ctDNA do not automatically imply that more treatment in patients with ctDNA positivity would improve outcomes, which we highlight. In real-world settings, we anticipate a potential rise in chemotherapy use due to clinicians utilizing ctDNA alongside existing clinicopathologic factors, rather than using ctDNA as an entire replacement. Lastly, a key concern in DYNAMIC was an 350% higher use of oxaliplatin in the ctDNA arm compared with standard management (9.5% versus 2.7%, respectively), which poses a risk for long-term neuropathy. CONCLUSION: We look forward improvements in patient selection in the adjuvant setting, but we maintain our reservations about the DYNAMIC trial and the real-life implementation of its results. As an alternative to exploring de-escalation strategies with large margins non-inferiority trials, we propose that superiority trials in stage II patients could be a more effective strategy.

The definition of long COVID used in interventional studies.

INTRODUCTION: There has been little consensus for a specific definition of long COVID, though several organizations have created varying ones. We sought to examine the definition of long COVID used in ongoing clinical trials. METHODS: We searched 'long COVID' and related terms on both PubMed and clinicaltrials.gov for randomized studies that either included patients with long COVID or had a persistent or long-term COVID-related outcome and abstracted long COVID definition components. RESULTS: Of the 92 studies, a laboratory-only confirmed diagnosis of COVID-19 was stipulated in 54.3% (n = 50) studies. We found eight different time durations specified for how long symptoms needed to have occurred, ranging from 4 to 52 weeks, with 12 weeks being the most common (34.8%; n = 32). 35.9% (n = 33) did not specify a time duration. There were 57 different symptoms specified in total, with a median of one symptom identified per study (range 0-32). 8.7% of trials adhered to NICE or WHO definitions. CONCLUSION: Standardized definitions of long COVID should be applied in studies assessing this condition to unify and harmonize research on this topic.

Post-progression treatment in cancer randomized trials: a cross-sectional study of trials leading to FDA approval and published trials between 2018 and 2020.

BACKGROUND: Suboptimal treatment upon progression may affect overall survival (OS) results in oncology randomized controlled trials (RCTs). We aim to assess the proportion of trials reporting post-progression treatment. METHODS: This cross-sectional analysis included two concurrent analyses. The first one examined all published RCTs of anti-cancer drugs in six high impact medical/oncology journals between January 2018 and December 2020. The second studied all US Food and Drug Administration (FDA) approved anti-cancer drugs during the same period. Included trials needed to study an anti-cancer drug in the advanced or metastatic setting. Data abstracted included the tumor type, characteristics of trials, and reporting and assessment of post-progression treatment. RESULTS: There were 275 published trials and 77 US FDA registration trials meeting inclusion criteria. Assessable post-progression data were reported in 100/275 publications (36.4%) and 37/77 approvals (48.1%). Treatment was considered substandard in 55 publications (n = 55/100, 55.0%) and 28 approvals (n = 28/37, 75.7%). Among trials with assessable post-progression data and positive OS results, a subgroup analysis identified substandard post-progression treatment in 29 publications (n = 29/42, 69.0%) and 20 approvals (n = 20/26, 76.9%). Overall, 16.4% of publications (45/275) and 11.7% of registration trials (9/77) had available post-progression data assessed as appropriate. CONCLUSION: We found that most anti-cancer RCTs do not report assessable post-progression treatment. When reported, post-progression treatment was substandard in most trials. In trials reporting positive OS results and with assessable post-progression data, the proportion of trials with subpar post-progression treatment was even higher. Discrepancies between post-progression therapy in trials and the standard of care can limit RCT results' applicability. Regulatory rules should enforce higher requirements regarding post-progression treatment access and reporting.

Umbrella review of basket trials testing a drug in tumors with actionable genetic biomarkers.

BACKGROUND: The utilization of basket trials in oncology has gained popularity because of the drive for precision medicine and the increasing ease of genetically profiling tumors. However, it is unknown if this has translated into patient benefit, either through higher response rates because of precision treatment or because of increasing options for less-common tumor types that are less represented in oncology drug trials. We sought to characterize basket studies for oncology drugs targeting a genetic biomarker, determine the responses for various tumor types and genetic biomarkers, and test for correlation between the number of participants in each tumor basket and the incidence of the respective tumor. METHODS: We conducted a retrospective cross-sectional review of oncology basket trials on Embase or clinicaltrials.gov with published data. We included studies that reported on oncology drugs that target a genetic biomarker. We examined the response for basket trial participants, stratified by tumor type and genetic biomarker and the correlation between the number of participants in each tumor basket and the incidence of the respective tumor. RESULTS: The overall response rate for all 25 included trials was 23%. The response for each genetic biomarker ranged from 0 to 69%, and for half of the genetic biomarkers, the response rate ranged from 0 to 100%, depending on tumor type. There is low correlation between the number of participants in each tumor basket and the incidence of the respective tumor (66.41 + -0.20x, R2 = 0.003, p = 0.75). CONCLUSION: While there has been an increase in the number of published basket trials and individuals included in these trials, the response rate is low, but varies widely, depending on tumor type and genetic biomarker.

Molecular testing to deliver personalized chemotherapy recommendations: risking over and undertreatment.

BACKGROUND: In the adjuvant setting of cancer treatment, de-escalation strategies have the goal of omitting or minimizing treatment in patients, without compromising outcomes. Historically, eligibility for adjuvant treatment solely relied on the patient's clinical and tumor's pathological characteristics. At the turn of the century, based on new biological understanding, molecular-based strategies were tested and sometimes implemented. MAIN BODY: However, we illustrate how molecularly based de-escalation strategies may paradoxically lead to overtreatment. This may happen when the novel approach is tested in lieu of standard management and may not yield the same results when being implemented in addition to usual practice. In the DYNAMIC trial, adjuvant chemotherapy decision in stage II colon cancer was compared between a circulating tumor DNA (ctDNA)-based approach and the standard care. We show this may result in more patients receiving oxaliplatin-based chemotherapy and may expose a similar proportion of patients to chemotherapy if the novel strategy is implemented in addition to usual practice. The other potential risk is undertreatment. We provide an illustration of early breast cancer, where the decision of adjuvant chemotherapy based on the gene expression signature MammaPrint may lead to inferior outcomes as compared with the clinico-pathologic strategy. This may also happen when non-inferiority designs have large margins. Among solutions, it should be acknowledged that clinico-pathological features, like T4 in colon cancer, may not be abandoned and replaced by novel strategies in real-life practice. Therefore, novel strategies should be tested in addition to standard of care, and not in lieu of. Second, de-escalation trials should focus on the settings where the standard of care has a widespread agreement. This would avoid the risk of testing non-inferiority against an ineffective therapy, which guarantees successes without providing informative data. CONCLUSION: Simply because a molecular test is rational does not mean it can improve patient outcomes. Here, we highlight how molecular test-based strategies may result in either overtreatment or undertreatment. In the rapidly evolving field of medicine, where technological advances may be transformative, our piece highlights scientific pitfalls to be aware of when considering running such trials or before implementing novel strategies in daily practice.

Clinical factors associated with prolonged response and survival under olaparib as maintenance therapy in BRCA mutated ovarian cancers.

OBJECTIVE: To investigate clinical factors associated with prolonged progression-free survival (PFS) and overall survival (OS) in relapsing epithelial ovarian cancer (EOC) patients with BRCA mutations and receiving olaparib as maintenance therapy in daily practice. METHODS: Multicenter (8 hospitals) European retrospective study of relapsing EOC patients having germline or somatic mutations of BRCA1/BRCA2 genes and treated with olaparib as maintenance therapy after platinum-based chemotherapy. RESULTS: One hundred and fifteen patients were included. Median age was 54 years. There were 90 BRCA1 carriers, 24 BRCA2 carriers and one patient had germline mutation of BRCA1 and BRCA2. Six patients had somatic mutations (all BRCA1) and 109 had germline mutations. Ninety percent had serous carcinomas and were platinum-sensitive. Following ultimate platinum-based chemotherapy, 69% of the patients had normalization of CA-125 levels and 87% had RECIST objective responses, either partial (53%) or complete (34%). After a median follow-up of 21 months, median PFS was 12.7 months and median OS was 35.4 months. In multivariate analysis, factors associated with prolonged PFS under olaparib were: platinum-free interval (PFI) ≥ 12 months, RECIST complete response (CR) or partial response (PR) and normalization of CA-125 upon ultimate platinum-based chemotherapy. Factors associated with prolonged OS were PFI ≥ 12 months, CR and normalization of CA-125. CONCLUSIONS: Platinum-free interval ≥ 12 months, complete response and normalized CA-125 levels after ultimate platinum-based chemotherapy are associated with prolonged PFS and OS in relapsing BRCA1/BRCA2 mutated ovarian cancer patients who received olaparib as maintenance therapy.

[Bioinformatics: a key role in oncology]. / Bioinformatique en oncologie: une discipline incontournable.

Bioinformatics is essential in clinical oncology and research. Combining biology, computer science and mathematics, bioinformatics aims to derive useful information from clinical and biological data, often poorly structured, at a large scale. Bioinformatics approaches have reclassified certain cancers based on their molecular and biological presentation, improving treatment selection. Many molecular signatures have been developed and, after validation, some are now usable in clinical practice. Other applications could facilitate daily practice, reduce the risk of error and increase the precision of medical decision-making. Bioinformatics must evolve in accordance with ethical considerations and requires multidisciplinary collaboration. Its application depends on a sound technical foundation that meets strict quality requirements.

Equal censoring but still informative: When the reasons for censoring differ between treatment arms.

In randomized controlled trials, informative censoring has been described as a potential bias, mainly affecting time-to-event composite endpoints, like progression-free survival (PFS). It is usually suspected in the presence of unequal attrition rates between arms. Early censoring occurs for different reasons: patients may withdraw from a trial because of toxicity, or because of disappointment with their allocation arm. If censoring is more frequent in one arm due to increased toxicity, this removes the frailest individuals and introduces a bias favoring this arm. Conversely, patients who withdraw because of disappointment of their allocation arm may be more affluent and healthy patients, who will seek treatment options outside the protocol. In trials with one treatment arm presenting higher toxicity rates, and the other arm potentially leading to patient disappointment, censoring can occur for different reasons in each arm however with the same rates. We modeled this hypothesis in a randomized controlled trial where modifying only 15% of censored patients' fate in each arm at early time-points made the PFS gain fade. Equal censoring but for different reasons is a hitherto unexplored form of informative censoring with potentially large implications across the cancer clinical trials landscape.

Neoadjuvant followed by adjuvant pembrolizumab in melanoma: time biases in the data analysis of the SWOG S1801 trial.

The SWOG S1801 trial investigated the role of pembrolizumab, an anti-PD1 immune checkpoint inhibitor, in the perioperative setting of stage III or IV melanoma. This phase 2 trial compared two groups: one receiving pembrolizumab both before and after surgery (neoadjuvant-adjuvant), and another receiving it only post-surgery (adjuvant-only), with event-free survival (EFS) as the primary endpoint. Neoadjuvant strategies, involving pre-surgical drug administration, potentially offer rapid tumor control and a unique opportunity to assess tumor response. However, they may expose to toxicity and delay or preclude surgery. The study met its primary endpoint, with a 72 % EFS rate in the neoadjuvant-adjuvant group, and 49 % in the adjuvant group. Here, we question the results' applicability with three potential limitations. Key concerns include an arbitrary rule in event assignment, possibly affecting the event distribution over time. Second, different rates of early censoring between groups introduce the possibility of informative censoring, which could have led to an artefactual benefit in EFS. Lastly, phase 2 trial results, by definition, carry risk of fluke results, and should be confirmed in phase 3 trial before wide adoption. Collectively, these factors must be integrated into a careful interpretation of the SWOG S1801 trial outcomes. More robust data are needed to fully appraise strengths and limitations of neoadjuvant pembrolizumab in melanoma treatment.

Progression-free survival estimates are shaped by specific censoring rules: Implications for PFS as an endpoint in cancer randomized trials.

Kaplan-Meier analysis hinges on the assumption that patients who are censored- lost to follow-up, or only recently enrolled on the study- are no different, on average, than patients who are followed. As such, censoring these patients- omitting their future information and taking the average of those who were followed- should not dramatically change the overall estimate. Yet, in a recent clinical trial, two sets of censoring rules- one favored by trialists and one favored by the US Food and Drug Administration- were applied to a progression-free survival (PFS) estimate. In response, the PFS estimate changed dramatically, increasing the median in the experimental arm from 32 to 43 months, while the control arm was essentially unchanged. In this commentary, we explore the reasons why PFS changed so dramatically. We provide a broad overview of censoring in oncology clinical trials, and suggestions to ensure that PFS is a more reliable endpoint.

Health-related quality of life in trials with high rates of early censoring: Caution advised.

Health-related quality-of-life (HRQoL) data are central to capturing the quality of patients' life, while endpoints like overall survival (OS) focus on the quantity of life. When analyzing HRQoL data gathered from patients in a randomized trial, a key consideration is the completion rate - indicating the proportion of patients remaining in the trial and with completed questionnaires. When completion rates are disproportionately low in one treatment arm, one likely explanation is that patients who did not complete questionnaires suffered more from toxicities, negatively impacting their HRQoL. This is likely the case when low completion rates occur in the more toxic arm within a randomized trial. If the HRQoL analysis is run as a complete-case analysis - only considering patients without missing data - a decrement in HRQoL can be missed. Conversely, when completion rates are high, the HRQoL data are thought to be more reliable, and informative censoring is less likely. We describe why this reasoning can be inadequate. In trials where high and imbalanced rates of early censoring affect progression-free survival or OS endpoints, the completion rates only apply to the fraction of patients remaining in the trial. In those, HRQoL results should be considered with caution, and reasons for censoring in the primary time-to-event analyses should be explored before making definite conclusions about HRQoL. This is even more relevant in trials with non-inferiority design, where a benefit in HRQoL could be used as a justification to modify practice.