Oncologic end points and implication of trial design in the era of immunotherapy for resectable non–small cell lung cancer

Central MessageTargeted and immunotherapy trials in resectable NSCLC have not used overall survival primary end points. Pathologic and recurrence-based end points allow for smaller, faster and less expensive trials. Targeted and immunotherapy trials in resectable NSCLC have not used overall survival primary end points. Pathologic and recurrence-based end points allow for smaller, faster and less expensive trials. Featured Editor Note—The field of thoracic surgical oncology has never been more exciting. Within the past year, we are replacing lobectomy with sublobar resection for many patients with non–small cell lung cancer, we are offering neoadjuvant chemoimmunotherapy as standard practice for early-stage disease, and we are treating select patients with resectable non–small cell lung cancer in the adjuvant setting with targeted molecular therapy or immune checkpoint blockade. The backbone of each of these paradigm shifts is practice-changing data from high-impact randomized clinical trials, and the criteria for which we select patients for these therapies in practice are directly translated from the eligibility criteria for entry into these trials. Enhancing our understanding of the design and conduct of these important trials improves our ability to communicate with patients, increases our understanding of the limitations of these therapies, and advances our effectiveness as health care providers. The rapid pace of our field and the nuances of precision medicine require that we actively stay abreast of ground-breaking translational discovery and understand the basis for its adoption into practice. The following Feature Expert Opinion article is an easily read, compendious primer that will provide the readership with a basis for maximizing interpretation of recently published and forthcoming clinical trials in thoracic surgical oncology. Bryan M. Burt, MD, FACS Recent survival improvements in non–small cell lung cancer (NSCLC) are largely attributed to the integration of targeted therapies and immunotherapies in the metastatic setting. These exciting treatments are under investigation in numerous clinical trials in patients with resectable NSCLC. In the most basic goal of any clinical trial is to determine the effectiveness of a new therapy and whether it improves the current standard therapy. Ideal trial end points are clinically relevant, easily measured, sensitive and specific to the intervention, low-cost, and reproducible.1Prentice R.L. Surrogate endpoints in clinical trials: definitions and operational criteria.Stat Med. 1989; 8: 431-440Crossref PubMed Scopus (1521) Google Scholar Overall survival (OS) was the most common primary end point for adjuvant and neoadjuvant NSCLC trials, but that has changed in recent years. Although OS carries significant clinical relevance in resectable NSCLC populations, it falls short on several of the other ideal end-point features, which may explain the shift away from OS as the primary end point for NSCLC trials involving perioperative targeted therapies and immunotherapies. As trial design evolves, it is important to understand the impetus and clinical implications of these changes. The “gold standard” for clinical efficacy in curative intent clinical trials has been OS. In cohorts with high mortality and short life expectancy, OS is relatively easy to measure and differences are easily attributed to the novel treatment. However, in cohorts with greater cure rates and longer anticipated survival, OS becomes challenging to determine, requiring trials that are labor intensive, expensive, and with larger numbers. This was the case with the trials assessing adjuvant platinum-based chemotherapy for resected NSCLC; each contained nearly 1000 patients and took on average 10 years from initiation to determine their OS end points.2Winton T. Livingston R. Johnson D. Rigas J. Johnston M. Butts C. et al.Vinorelbine plus cisplatin versus observation in resected non–small-cell lung cancer.N Engl J Med. 2005; 352: 2589-2597Crossref PubMed Scopus (1663) Google Scholar, 3Douillard J.Y. Rosell R. De Lena M. Carpagnano F. Ramlau R. Gonzáles-Larriba J.L. et al.Adjuvant vinorelbine plus cisplatin versus observation in patients with completely resected stage IB-IIIA non–small-cell lung cancer (Adjuvant Navelbine International Trialist Association [ANITA]): a randomised controlled trial.Lancet Oncol. 2006; 7: 719-727Abstract Full Text Full Text PDF PubMed Scopus (1367) Google Scholar, 4Scagliotti G.V. Fossati R. Torri V. Crinò L. Giaccone G. Silvano G. et al.Randomized study of adjuvant chemotherapy for completely resected stage I, II, or IIIA non–small-cell lung cancer.J Natl Cancer Inst. 2003; 95: 1453-1461Crossref PubMed Scopus (581) Google Scholar Many investigators feel this is too long to wait to make changes critical to patient survival. In addition, OS benefit can be hard to verify due to increased efficacy of post-recurrence therapies. For these reasons, surrogates for OS are frequent primary end points in recent and ongoing trials involving resectable NSCLC, with the goal of making trials more efficient without negatively impacting clinical relevance. Meaningful clinical trial end points can be grouped into several broad categories, as outlined in Table 1. These include recurrence end points, pathologic markers of response, radiographic and biochemical response markers, surgical outcomes, safety, and quality of life.Table 1Common oncologic trials end points and implications in non–small cell lung cancer surgical trialsCategoryRelevance in resectable NSCLC trialsOverall survival•Ultimate goal•Challenging to measure in curable populations•Impacted by additional therapies at recurrenceRecurrence-based survival•Strong surrogates for overall survival in NSCLC•Long time to achieve•Cannot differentiate cure from delay in recurrenceResponse•Measured radiographically or by biomarker•Rapid but not fast enough for perioperative decisions•Limited accuracy with current technologyPathologic response•Unique to neoadjuvant trials•Rapid•Depth of response typically correlates with efficacy of therapy•Correlation with overall survival in NSCLC requires confirmationSurgical•Essential for neoadjuvant strategies•Frequently not incorporated into trial designSafety•Determined for novel agents in phase 1 trials•Additional assessment required for multimodality useQuality of life•Typically patient reported•Numerous validated tools•Underused in multimodality and surgical trialsNSCLC, Non–small cell lung cancer. Open table in a new tab NSCLC, Non–small cell lung cancer. Surrogate survival end points based on disease recurrence include progression-free survival (PFS), disease-free survival (DFS), event-free survival (EFS), and time to death or distant metastasis (Table 2). In NSCLC, recurrence frequently occurs sooner than death; thus, trials with recurrence end points are faster to complete and require smaller sample size. Recurrence is also agnostic to additional therapies and therefore may more closely reflect the effect of the therapy of interest. PFS is a common primary end point for metastatic NSCLC trials, with progression typically detected radiographically. In the current era of precision oncology, more clinical trials are sponsored by industry than cooperative groups and trial results are frequently used for regulatory approval,5Del Paggio J.C. Eisenhauer E.A. Booth C.M. Randomized clinical trials in the era of precision oncology—the role of end points, industry funding, and medical writing integrity—Reply.JAMA Oncol. 2021; 7: 1579-1580Crossref PubMed Scopus (1) Google Scholar which is frequently granted for treatment of stage IV disease without OS benefit. By definition, PFS implies residual disease at the end of treatment and therefore it is not a common end point for adjuvant or neoadjuvant NSCLC trials where patients are assumed disease-free at the end of therapy. DFS is a common end point in trials assessing adjuvant therapy. EFS encompasses disease recurrence and symptoms or complications specified by each individual study. Many recent and ongoing neoadjuvant immunotherapy trials use EFS end points, with the inability to undergo a planned resection as a reportable event.Table 2Commonly reported recurrence-based survival end points and their relevance in resectable non–small cell lung cancer trialsEnd pointRelevance in resectable NSCLCDisease-free survival (DFS)•Common for adjuvant trials•Faster than OS•Still requires significant time in curable cohorts•Cannot differentiate between cure and delay in recurrenceEvent-free survival (EFS)•Common for neoadjuvant trials•Events defined for each trial•Recurrence and lack of surgery common reportable eventsProgression-free survival (PFS)•Implies residual posttreatment disease•Limited utility in adjuvant and neoadjuvant trialsTime to death or distant metastasis (TTDM)•Implies residual posttreatment disease•Limited utility in adjuvant and neoadjuvant trials•Common for indolent diseasesNSCLC, Non–small cell lung cancer; OS, overall survival. Open table in a new tab NSCLC, Non–small cell lung cancer; OS, overall survival. The primary concern with surrogate recurrence-based end points surrounds the association between recurrence and OS. Do the treatments of interest truly cure patients or simply delay recurrence? This may depend on the treatment under investigation and the magnitude of the benefit. The landmark PACIFIC trial evaluated adjuvant durvalumab following definitive chemoradiotherapy in unresectable stage III NSCLC. The initial PFS data favoring durvalumab use (median 16.8 vs 5.6 months) was reported in 20176Antonia S.J. Villegas A. Daniel D. Vicente D. Murakami S. Hui R. et al.Durvalumab after chemoradiotherapy in stage III non–small-cell lung cancer.N Engl J Med. 2017; 377: 1919-1929Crossref PubMed Scopus (2631) Google Scholar and led to approval from the Food and Drug Administration (FDA) for durvalumab in that setting. This was 5 years before OS results, which also favored adding durvalumab (median 47.5 vs 29.1 months) but were not reported until 2022.7Spigel D.R. Faivre-Finn C. Gray J.E. Vicente D. Planchard D. Paz-Ares L. et al.Five-year survival outcomes from the PACIFIC trial: durvalumab after chemoradiotherapy in stage III non–small-cell lung cancer.J Clin Oncol. 2022; 40: 1301-1311Crossref PubMed Scopus (193) Google Scholar The PFS benefit was durable and translated to an OS advantage. Alternatively, the initial PFS benefit (hazard ratio, 0.66) in the IMpower010 trial led to approval from the FDA for adjuvant atezolizumab for resected stage II-IIIA NSCLC with programmed death-ligand 1 (PD-L1) staining in >1% of cells, even though the benefit was primarily driven by those tumors with high PD-L1 expression (>50% of cells).8Felip E. Altorki N. Zhou C. Csőszi T. Vynnychenko I. Goloborodko O. et al.Adjuvant atezolizumab after adjuvant chemotherapy in resected stage IB-IIIA non–small-cell lung cancer (IMpower010): a randomised, multicentre, open-label, phase 3 trial.Lancet. 2021; 398: 1344-1357Abstract Full Text Full Text PDF PubMed Scopus (339) Google Scholar Updated IMpower010 data were presented in 2022, and early OS benefit with atezolizumab was only significant in high-expressing stage II-IIIA tumors and not the broader PD-L1 >1% group. Interestingly, the European Union only ever approved adjuvant atezolizumab in the high-expressing population. Similarly, epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors have demonstrated dramatic PFS improvements and changed the landscape of stage IV care outside of clinical trials, but no randomized trial has ever demonstrated an OS benefit for EGFR tyrosine kinase inhibitors. Therefore, although there is great enthusiasm for the dramatic DFS improvement with the addition of adjuvant osimertinib in patients with completely resected EGFR-mutated NSCLC (hazard ratio, 0.17; P < .001),9Wu Y.L. Tsuboi M. He J. John T. Grohe C. Majem M. et al.Osimertinib in resected EGFR-mutated non–small-cell lung cancer.N Engl J Med. 2020; 383: 1711-1723Crossref PubMed Scopus (725) Google Scholar there are questions of how durable this will be and whether it will translate to an OS benefit. Overall response rate is common primary end point in oncology trials in the metastatic setting. It is the percent of patients with a partial or complete response to treatment as measured radiographically or by biomarker but is generally not useful in surgical trials. The lack of radiographic disease in the adjuvant setting makes overall response rate irrelevant, and in the neoadjuvant setting full measurable responses typically require several months, and surgery is recommended before time for maximal radiographic response. This was highlighted in results from CheckMate 159 (early neoadjuvant immunotherapy trial); major pathologic response (MPR) rate after 2 cycles of nivolumab was 45%, far greater than 10% partial response determined radiographically before resection.10Forde P.M. Chaft J.E. Smith K.N. Anagnostou V. Cottrell T.R. Hellmann M.D. et al.Neoadjuvant PD-1 blockade in resectable lung cancer.N Engl J Med. 2018; 378: 1976-1986Crossref PubMed Scopus (1098) Google Scholar Cell-free tumor DNA (ctDNA) is a circulating marker used clinically in patients with stage IV for mutation detection and tracking therapeutic response. ctDNA persistence following treatment in unresectable stage III is associated with poor survival.11Moding E.J. Liu Y. Nabet B.Y. Chabon J.J. Chaudhuri A.A. Hui A.B. et al.Circulating tumor DNA dynamics predict benefit from consolidation immunotherapy in locally advanced non–small cell lung cancer.Nat Cancer. 2020; 1: 176-183Crossref PubMed Scopus (128) Google Scholar ctDNA clearance was an exploratory end point in CheckMate 816, and posttreatment clearance was associated improved EFS and pathologic complete response (pCR).12Forde P.M. Spicer J. Lu S. Provencio M. Mitsudomi T. Awad M.M. et al.Neoadjuvant nivolumab plus chemotherapy in resectable lung cancer.N Engl J Med. 2022; 386: 1973-1985Crossref PubMed Scopus (323) Google Scholar Today's ctDNA assays are relatively insensitive to low disease levels, and therefore clearance currently should not serve as a marker to deescalate therapy but rather persistence of ctDNA following initial therapy could serve as a marker for earlier escalation of additional therapy. Pathologic response end points are useful primary end points in neoadjuvant trials. They are attractive because they closely reflect the magnitude of impact of the intervention and can be determined rapidly. The importance of the local responses goes far beyond what happens in the chest or local tumor control, but rather serves as important markers for the systemic response to therapy and for the eradication of micrometastatic disease, which is the survival risk for most patients with NSCLC. Common pathologic end points in NSCLC trials are outlined in Table 3 and include pCR, MPR, immune-related pathologic response, and mediastinal lymph node clearance. pCR is the absence of any active tumor cells in the resected specimen, including lymph nodes. Major pathologic response is defined as <10% of viable tumor cells in the tumor and the lymph nodes and was advocated as an important end point by NSCLC scientific community during the era of neoadjuvant chemotherapy trials, primarily because pCR was too infrequent to serve as a meaningful trial end point.13Hellmann M.D. Chaft J.E. William Jr., W.N. Rusch V. Pisters K.M. Kalhor N. et al.Pathological response after neoadjuvant chemotherapy in resectable non–small-cell lung cancers: proposal for the use of major pathological response as a surrogate endpoint.Lancet Oncol. 2014; 15: e42-e50Abstract Full Text Full Text PDF PubMed Scopus (310) Google Scholar However, in the current trials evaluating neoadjuvant immune checkpoint inhibition, pCR rates are greater, and this may negate the need for MPR as a trial end point.Table 3Commonly reported pathologic trial end points, their definitions, and relevance in resectable non–small cell lung cancer trialsEnd pointDefinitionRelevance in respectable NSCLCMediastinal nodal clearance (MNC)Absence of any active tumor cells in resected mediastinal lymph nodes•Only useful if biopsy-proven disease in mediastinal lymph nodes pretreatment and if mediastinal nodes resectedMajor pathologic response (MPR)<10% of viable tumor cells in the tumor and the lymph nodes•Important in induction chemotherapy trials because pathologic complete response (pCR) was too rare to be usefulpCRAbsence of any active tumor cells in resected specimen including lymph nodes•Strongly correlated with OS in other histologies•Reproducable association with OS needs to be established in NSCLCImmune-related pathologic response(ir-pCR)Tumor bed description increased to account for fibrosis, inflammation tumor-infiltrating lymphocytes, and tertiary lymphoid structures•Long-term follow-up needed to determine reliability as a surrogate indicator of DFS and OSNSCLC, Non–small cell lung cancer; OS, overall survival; DFS, disease-free survival. Open table in a new tab NSCLC, Non–small cell lung cancer; OS, overall survival; DFS, disease-free survival. Surgical end points are typically not considered direct surrogates of response or OS in multimodality treatment, but these end points are critical in multimodality protocols because a neoadjuvant therapy will never be clinically integrated if it poses a significant hurdle to resection (Table 4). Despite their importance, surgical outcomes are rarely even included as secondary end points in neoadjuvant and adjuvant NSCLC trials. Surgical attrition due to neoadjuvant therapy is typically included as part of EFS but could be an end point in its own right. We accept 10% to 20% attrition based upon induction chemotherapy data for stage IIIA cohorts, but it is unclear that similar rates are acceptable in earlier-stage and node negative cohorts. Similarly, R0 resection is a requirement for most adjuvant trials, but not included as a prespecified end point in many neoadjuvant trials, if it was prespecified stopping rules could halt protocols with significant differences in resection or complete resection rates. An interesting observation from CheckMate 816 was the greater rates of R1 resections in North America and greater attrition and pneumonectomy rates in Europe, highlighting important regional differences in surgical assessment and decision-making.Table 4Commonly reported surgical end points and their relevance in resectable non–small cell lung cancer trialsEnd pointRelevance in respectable NSCLCDelay•Treatment related toxicity may hinder ability to undergo surgery•10%-20% attrition reported for chemotherapy-based induction protocolsAttrition (not proceeding to planned resection)Estimated blood loss (EBL)•Indirect measure for technical challenge due to hilar fibrosisOperative time, minIndirect measure for technical challengeMinimally invasive resection rate•Indirect measure for technical challenge•Indirect marker for treatment efficacy and tumor downstaging•Impacted by regional variability in surgical decision makingPneumonectomy/lobectomy rateIndirect marker for treatment efficacy and tumor downstagingR0 resection rateIndirect marker for treatment efficacy and tumor downstagingMorbidity•Potential for additional treatments to increase surgical morbidity or decrease ability for recoveryNSCLC, Non–small cell lung cancer. Open table in a new tab NSCLC, Non–small cell lung cancer. Increased technical challenge is an important concern for neoadjuvant treatments in NSCLC due to hilar and mediastinal fibrosis, but it is hard to measure or quantify. Surrogates for technical challenge include extent of resection, minimally invasive resection rates, estimated blood loss, operative time, and completeness of resection. Other important surgical markers include postoperative morbidity and 30- and 90-day operative mortality. Equally import but rarely discussed in trial design is the significant impact of surgical quality in adjuvant and neoadjuvant NSCLC trials. Numerous quality metrics have well-documented impact on OS,14Samson P. Crabtree T. Broderick S. Kreisel D. Krupnick A.S. Patterson G.A. et al.Quality measures in clinical stage I non–small cell lung cancer: improved performance is associated with improved survival.Ann Thorac Surg. 2017; 103: 303-311Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar and therefore minimal trial data should include time to resection, extent of resection, approach, number of resected lymph nodes and nodal stations, and completeness of resection. Enhanced recovery pathways are associated with faster return to intended oncologic therapy and information on use should be considered in multi-institutional trials of adjuvant therapy.15Nelson D. Mehran R.J. Mitchell K. Correa A.M. Sepesi B. Antonoff M.B. et al.Enhanced recovery after thoracic surgery is associated with improved adjuvant chemotherapy for non–small cell lung cancer.J Thorac Cardiovasc Surg. 2019; 158: 279-286.e1Abstract Full Text Full Text PDF PubMed Scopus (24) Google Scholar Standard safety end points in oncologic trials include feasibility, safety, and toxicity. These are reported as adverse events as are the rate the interruption and discontinuation of therapy. In oncology, these issues are well documented in phase 1 and 2 trials in the metastatic setting long before therapies are combined with surgery. Adverse events are reported using Common Terminology Criteria for Adverse Events, a scale developed and updated by the Cancer Therapeutic Evaluation Program within the National Cancer Institute.16National Cancer InstituteCommon terminology criteria for adverse events (CTCAE).https://ctep.cancer.gov/protocoldevelopment/electronic_applications/ctc.htm#ctc_60Date accessed: December 12, 2022Google Scholar There is consideration for added toxicity when combining novel therapies with surgery. In adjuvant NSCLC trials, it is common to see similar rates and severity of adverse events as reported in stage IV cohorts but with increased rates of discontinuation of therapy, possibly related to decrease physical stamina and conditioning from recent surgery or to decreased tolerance for any toxicity in patients who have greater likelihood of already being cured. In neoadjuvant trials, there is concern for toxicity that prevents an operable patient from going on to resection, and the acceptance for even small rates is quite low. This is a reason that EFS end points that include the inability to undergo resection as a reportable event are important in neoadjuvant NSCLC trials. Lastly, QoL end points need to be further emphasized in oncology trials. The goal of oncologic therapy is not limited to increasing the quantity of a patient's life but also to ensuring the quality of that time. There are numerous validated tools for patient-reported QoL; the most common ones in NSCLC trials include the 36-Item Short Form Survey17RAND Corporation36-item short form survey (SF-36).https://www.rand.org/health-care/surveys_tools/mos/36-item-short-form.htmlDate accessed: December 12, 2022Google Scholar and dyspnea scores.18Mahler D.A. Mechanisms and measurement of dyspnea in chronic obstructive pulmonary disease.Proc Am Thorac Soc. 2006; 3: 234-238Crossref PubMed Scopus (55) Google Scholar The 36-Item Short Form Survey includes both physical and psychologic components but does not specifically address dyspnea or pulmonary symptoms. Lung cancer resection is an invasive procedure with associated hospitalization and recovery period, and therefore carries significant short-term impact on QoL. In the curative setting, the more pertinent questions relate to the time and extent of that impact and on a patient's ability to return to baseline function and QoL. QoL measures add labor and expense to the trial process but provide essential information for integrating novel therapies. These measures also provide insight to disparate treatment approaches, such as operative versus nonoperative treatments for patients with major comorbidity, a cohort where functional status may be as important as OS. It is an exciting time in thoracic surgical oncology, with the introduction of novel agents into the surgical populations. These represent a chance to improve survival in patients with stage I-III NSCLC. Integration of new therapies is only possible through a large series of recently completed and ongoing clinical trials, which differ in design from trials conducted 2 decades ago. OS is unlikely to serve as the primary end point for an adjuvant or neoadjuvant NSCLC trial in the near future. It is important to be mindful of the strengths and limitations of surrogates for OS that serve as the primary end point for many of these trials. Some caution is needed, as some initially positive surrogate end points may not translate to an OS benefit. Furthermore, it is important to push for a greater emphasis on surgical end points and surgical quality markers in trial design for adjuvant and neoadjuvant NSCLC.