Pre- or Perioperative Immunotherapy Combined with Chemotherapy Versus Chemotherapy Alone in Resectable Non-Small Cell Lung Cancer (NSCLC): A Systematic Literature Review

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript addresses a highly relevant and clinically important topic in thoracic oncology and provides a comprehensive overview of perioperative or neoadjuvant chemoimmunotherapy in resectable NSCLC.

The inclusion of recent pivotal phase II and III trials, the structured presentation of survival, pathological response, and safety outcomes, and the use of PRISMA methodology represent important strengths of the study. However, several minor issues should be addressed before publication.

The authors should explicitly state adherence to PRISMA 2020 guidelines and clarify whether the review protocol was prospectively registered. In addition, the rationale for choosing a narrative synthesis instead of a meta-analysis should be expanded by more clearly explaining the sources of heterogeneity, including differences in immunotherapeutic agents, treatment strategies, chemotherapy regimens, endpoint definitions, follow-up durations, and study designs.

The implications of the risk-of-bias assessment should also be discussed in greater detail, particularly regarding the susceptibility of retrospective cohort studies to selection bias and residual confounding. Statistical reporting across the tables should be standardized, with clarification of “NR” entries and more consistent reporting of HRs, ORs, confidence intervals, and p-values.

Furthermore, the discussion of treatment-related adverse events (TRAEs) should be interpreted more cautiously, as some conclusions regarding toxicity differences between specific immune checkpoint inhibitors appear speculative due to the absence of direct comparative analyses. The manuscript would also benefit from a stronger discussion of the heterogeneity in adverse-event reporting across studies, including differences in grading systems, attribution of immune-mediated toxicity, and definitions of serious TRAEs. Surgical outcome data, such as R0 resection rates, minimally invasive surgery eligibility, and operative outcomes, are clinically relevant and should either be summarized more systematically in the discussion or presented in a supplementary table.

The limitations section should additionally acknowledge publication bias, immature overall survival data, heterogeneous PD-L1 assessment criteria, varying disease stages, and the inclusion of retrospective studies. Finally, several minor language and typographical issues should be corrected, including terms such as “subtyp,” “integrate part,” and “other regiments,” and professional English language editing is recommended. Overall, the manuscript is scientifically valuable and suitable for publication after minor revision.

Author Response

Comment 1: "The authors should explicitly state adherence to PRISMA 2020 guidelines and clarify whether the review protocol was prospectively registered."

Response 1: Thank you for pointing this out - we have added the statements "This systematic review was conducted in accordance with the PRISMA 2020 guidelines. Further information is provided via the PRISMA 2020 flow diagram of Study Selection and the PRISMA 2020 Checklist in the supplementary materials [16]. The study protocol of this review was not registered." in the 2.6. Data Synthesis section, page 4, line 161-165. 

Comment 2: "In addition, the rationale for choosing a narrative synthesis instead of a meta-analysis should be expanded by more clearly explaining the sources of heterogeneity, including differences in immunotherapeutic agents, treatment strategies, chemotherapy regimens, endpoint definitions, follow-up durations, and study designs."

Response 2: Thank you for this comment, we agree and have altered the rationale to the following: "The evidence synthesis was done in a narrative way due to heterogeneity in study design, outcome measures, and data reporting. The individual studies differed in the type(s) of immunotherapeutic and chemotherapeutic agents applied, the corresponding treatment schedule, endpoints of interest, and follow-up duration." in the 2.6. Data Synthesis section, page 4, line 158-161.

Comment 3: "The implications of the risk-of-bias assessment should also be discussed in greater detail, particularly regarding the susceptibility of retrospective cohort studies to selection bias and residual confounding."

Response 3: Thank you for this comment, we agree and have therefore added the corresponding statements in 3.2. Risk of Bias, page 5, line 186-188 "In the case of prospective and retrospective cohort studies, their susceptibility to selection bias and confounding should be taken into consideration." as well as a statement in the discussion section, page 19, line 468-471 "Another limitation to consider is the inclusion of prospective and retrospective cohort studies: The non-randomized nature of these studies makes them susceptible to selection bias and confounding, potentially reducing the certainty and strength of evidence provided."

Comment 4: "Statistical reporting across the tables should be standardized, with clarification of “NR” entries and more consistent reporting of HRs, ORs, confidence intervals, and p-values."

Response 4: We thank the reviewer for this comment. There is a lack of consistency in statistical reporting across the individual studies, with some unfortunately not providing any statistical data at all, which we therefore marked with NR (not reported). If statistical data was provided, we extracted the data correspondingly and tried to make the list as continuous as possible. Overall, we respectfully believe that the varying extent of statistical data available makes it difficult to make the reporting more consistent. 

Comment 5: "Furthermore, the discussion of treatment-related adverse events (TRAEs) should be interpreted more cautiously, as some conclusions regarding toxicity differences between specific immune checkpoint inhibitors appear speculative due to the absence of direct comparative analyses."

Response 5: Thank you for pointing this out. We agree and have therefore altered the interpretation to the following: "Moreover, these interpretations are limited by the lack of comparative analyses and should therefore be interpreted cautiously. Further research comparing toxicity between the respective groups is required to allow for confident conclusions." in the discussion section, page 17, line 375-378.

Comment 6: "The manuscript would also benefit from a stronger discussion of the heterogeneity in adverse-event reporting across studies, including differences in grading systems, attribution of immune-mediated toxicity, and definitions of serious TRAEs."

Response 6: We thank the reviewer for this comment. We agree and have therefore added the following statements to the discussion section, page 19, line 486-490 "Moreover, the reporting of TRAEs was marked by heterogeneity with varying extent of data availability. Also, while most studies explicitly reported assessing TRAEs according to the CTCAE, some did not specify the grading system applied. In the context of TRAEs, another limitation is that only some studies provided information on the proportion of immune-mediated adverse events."

Comment 7: "Surgical outcome data, such as R0 resection rates, minimally invasive surgery eligibility, and operative outcomes, are clinically relevant and should either be summarized more systematically in the discussion or presented in a supplementary table."

Response 7: Thank you very much for pointing this out. We agree and have summarized data on surgical outcomes from the CheckMate 816 study in the discussion section, page 17-18, line 397-410: "Patients in the intervention arm were more likely to undergo definite surgery (83.2% vs. 75.4%) and to be eligible for a minimally invasive approach (29.5% vs. 21.5%), with the rate of conversion to thoracotomy after starting with minimally invasive surgery being lower in this group (11.4% vs. 15.6%). Furthermore, the need for a pneumonectomy was lower in patients treated with Nivolumab compared to the control group (16.8% vs. 25.2%). Moreover, the success of completing tumor resection was higher in the intervention arm, with higher R0 rates (83.2% vs. 77.8%) and lower R1 rates (10.7% vs. 15.6%). R2 rates were similar between the two groups (3.4% vs. 3.0%) Also, the duration of surgery was notably shorter in the intervention arm with a median of 185.0 minutes compared to a median of 213.5 minutes in the comparison group. However, despite the differences in surgical approach rates, the median length of hospital stay was identical for both arms with 10.0 days. Nevertheless, the incidence of surgery-related adverse events was lower in the immunochemotherapy group for both any grade (41.6% vs. 46.7%) and grade 3 or 4 (11.4% vs. 14.8%)."

Comment 8: "The limitations section should additionally acknowledge publication bias, immature overall survival data, heterogeneous PD-L1 assessment criteria, varying disease stages, and the inclusion of retrospective studies."

Response 8: Thank you for this comment, we agree and have therefore altered the following statements of the limitation section: "The literature search was limited to studies published in English or German and there is a possibility of publication bias. Another limitation to consider is the inclusion of retrospective cohort studies. Also, besides meeting the eligibility criteria of this review, the included studies differed in terms of study design, population, intervention, and outcome characteristics as well as in endpoints of interest and mechanism of conduct. Sample sizes varied considerably across the individual studies and study designs ranged from single-center to multi-center, multinational studies. The disease stages of participants varied across the individual studies and the immunotherapeutic regimens differed as various types of agents were used, with some studies focusing on one immune checkpoint inhibitor and others analyzing outcomes of four to seven different agents. Another limitation of this review relates to the variability in extent and measures of reported outcomes, the immaturity of overall survival data, and heterogenous PD-L1 assessment criteria." in the discussion section, page 19, line 467-481.

Comment 9: "Finally, several minor language and typographical issues should be corrected, including terms such as “subtyp,” “integrate part,” and “other regiments,” and professional English language editing is recommended."

Response 9: Thank you for pointing this out, we have corrected the mistakes mentioned and have re-checked the manuscript for language and grammar. 

Reviewer 2 Report

Comments and Suggestions for Authors

This is a well researched and inclusive analysis of published and presented data.  From the studies included, can you address the concerns of many Oncologists that some patients do not respond to "neo-adjuvant" therapy and may lose the opportunity for surgery with this approach?  Tabularized information regarding the PD-L1 expression tests utilized in the studies and correlation of response/outcome related to PD-L1 expression would be of interest to readers, also some commentary regarding applicability to older patients?  I find no issues with your methods and conclusions which are consistent with published literature.

Author Response

Comment 1: "From the studies included, can you address the concerns of many Oncologists that some patients do not respond to "neo-adjuvant" therapy and may lose the opportunity for surgery with this approach?"

Response 1: Thank you for this comment. We have added the following statements in the discussion section, page 18, line 411-428: "A common concern regarding neoadjuvant immunotherapy is that patients might not respond and potentially lose eligibility for curative surgery due to disease progression. However, data of the CheckMate 816 trial [5,32] demonstrate the opposite: Not only did more patients in the intervention group undergo definite surgery but also fewer cancellations of surgery occurred (15.6% vs. 20.7%). In fact, cancellations due to disease progression were recorded in 6.7% in the intervention arm compared to 9.5% in the comparison arm. In the KEYNOTE-671 study [6,29], findings are similar, with more patients in the intervention group undergoing surgery than in the comparison group (82.1% vs. 79.4%). Also, patients treated with immunochemotherapy were less likely to have their surgery cancelled (17.9% vs. 20.5%) and cancellation rates caused by disease progression were also lower in the intervention arm (4.1% vs. 8.2%). In the AEGEAN trial [18, 37], a slightly lower percentage of patients in the intervention arm underwent surgery (81.0% vs. 81.3%) and the overall rate of surgery cancellation due to any cause was slightly higher in patients treated with immunochemotherapy (19.0% vs. 18.7%). However, cancellations due to disease progression were slightly less likely in the intervention group than in the comparison group (6.8% vs. 7.5%). These findings highlight the need for a predictive biomarker to reduce the number of patients progressing beyond surgical resectability due to failed neoadjuvant therapy."

Comment 2: "Tabularized information regarding the PD-L1 expression tests utilized in the studies and correlation of response/outcome related to PD-L1 expression would be of interest to readers"

Response 2: Thank you very much for this comment. We looked at the tests used to determine PD-L1 expression levels and added the types of assays applied in the RCTs included in this review. As only 3 different types of assays were applied overall, we politely believe that a table might not be necessary in this case. Other than that, we pointed out the correlation of PD-L1 expression and response by using the CheckMate 816 study as an example. The following statements were added in the discussion section, page 18, line 429-437: "PD-L1 expression levels are commonly used as an indicator for responsiveness. The RCTs included in this review determined PD-L1 expression using the VENTANA PD-L1 (SP263), PD-L1 IHC 28-8 pharmDx, or the PD-L1 IHC 22C3 pharmDx assay [5, 6, 17, 18, 22,23, 26, 27, 29, 30, 32, 33]. However, PD-L1 expression is not a reliable parameter to determine eligibility. In fact, the CheckMate 816 study [5,32] demonstrated that patients across all PD-L1 expression levels had beneficial outcomes. However, higher PD-L1 expression levels correlated with higher outcome benefits. Therefore, determining a patient’s eligibility for immunotherapy requires a multi-disciplinary approach, taking the overall clinical context into consideration."

Comment 3: "also some commentary regarding applicability to older patients"

Response 3: Thank you for pointing this out. We have added the following statement in the discussion section, page 18, line 443-454: "The retrospective cohort study of Shuai et al. [25] focused on elderly patients (≥65 years) and their findings suggest higher survival and pathological response rates in those patients treated with immunochemotherapy. However, grade ≥3TRAEs were more likely to occur in the intervention group. RCTs such as the CheckMate 816 study[5,32] provided subgroup analyses, also comparing outcomes between patients under the age of 65 and patients with 65 years and older: The data indicate that elderly patients undergoing immunochemotherapy achieved superior outcomes compared to elderly patients in the control group, though the magnitude of benefit was slightly lower compared to patients under the age of 65 receiving immunochemotherapy. However, as the RCTs included in this review did not primarily focus on the effect of immunochemotherapy in the elderly population, future RCTs with this specific primary endpoint are needed for further conclusions."

Reviewer 3 Report

Comments and Suggestions for Authors

In this review, the authors compared survival, response, and adverse events in NSCLC patients treated with the combination of chemo- and immune-therapy vs. chemotherapy alone. Overall, this study presents some interesting results about the efficacy of immunotherapy and potential concerns related to patient selection. The following questions should be addressed:

 

1. In the survival outcomes session, patients receiving immunotherapy show superior survival. However, the data should be visualized in curves of DFS, EFS, or OS instead of presenting in the current table format.
2. In the TRAEs section, patients receiving immunotherapy show more cases. Are there correlations between immune-related adverse events and other adverse events? Do patients receiving immunotherapy have more chance of immune-related adverse events than the comparison group but similar chance of other adverse events?
3. In the discussion section, the authors claim “Beyond inducing tumor cell death, chemotherapy also increases immunotherapeutic efficacy by priming the immune system.” More literature should be cited here to support the claim, including pathways like immunogenic cell death and activation of immune-activation pathways, considering different chemotherapeutics have different mechanisms and efficacy in synergistic effects with immunotherapy.

Author Response

Comment 1: "In the survival outcomes session, patients receiving immunotherapy show superior survival. However, the data should be visualized in curves of DFS, EFS, or OS instead of presenting in the current table format."

Response 1: The authors want to thank the reviewer for this excellent suggestion. We tried to summarize the survival data in one curve, but the different studies unfortunately used different outcome parameters (OS, PFS, DFS, EFS) and different observation periods and some e.g. reported 2 year other 30 & 18 months and others only median survival. Since we have performed a review and no meta analysis, we have not requested the survival data and events from the corresponding studies. Thus, the only way to summarize the reported and available outcome of the heterogeneous studies is table nr 5. This weakness of our study that results from the heterogeneity of the analyzed literature regarding outcome reporting is now also discussed in the study limitation section (page 19, line 479-483).

Comment 2: "In the TRAEs section, patients receiving immunotherapy show more cases. Are there correlations between immune-related adverse events and other adverse events? Do patients receiving immunotherapy have more chance of immune-related adverse events than the comparison group but similar chance of other adverse events?"

Response 2: Thank you for this comment. We have added the following in the discussion section, page 17, line 381-391: "In fact, data of the CheckMate 816 trial [5,32] indicate that while the addition of immunotherapy slightly increased immune-mediated adverse events, it did not put the intervention group at a higher risk of experiencing other adverse events. In fact, the overall incidence of TRAEs was lower in patients treated with immunochemotherapy. In the KEYNOTE-671 study [6,29], patients in the intervention group were also more susceptible to immune-mediated adverse events and data also show a slightly higher total rate of TRAEs in this group. This was also the case for the AEGEAN study [18, 37], in which the intervention arm was more prone to immune-mediated adverse events, though rates of general adverse events were not significantly higher in patients treated with immunochemotherapy. In fact, the occurrence of ≥3 TRAEs was slightly lower in the intervention group."

Comment 3: "In the discussion section, the authors claim “Beyond inducing tumor cell death, chemotherapy also increases immunotherapeutic efficacy by priming the immune system.” More literature should be cited here to support the claim, including pathways like immunogenic cell death and activation of immune-activation pathways, considering different chemotherapeutics have different mechanisms and efficacy in synergistic effects with immunotherapy."

Response 3: Thank you very much for this comment. We agree and have therefore added the following statements into the discussion section, page 16, line 324-341:"These superior outcomes associated with chemoimmunotherapy could be explained by the synergistic effect: Beyond inducing tumor cell death, chemotherapy also increases immunotherapeutic efficacy by priming the immune system. This challenges the longstanding classification of chemotherapy as an immunosuppressive agent. In fact, under certain conditions, chemotherapy modulates the patient’s immune system through a variety of mechanisms, such as targeting immunosuppressive cells, inducing immunogenic cell death (ICD), thereby stimulating a T-cell response, activating natural killer (NK) cells, as well as increasing antigen presentation and immune cell infiltration. Moreover, different chemotherapeutic agents exhibit different immunomodulatory effects: unlike many other platinum compounds, Cisplatin is not considered a true inducer of ICD, however it stimulates the immune system by targeting immunosuppressive cells, modulating antigen presentation and immune cell infiltration. Other platinum-based agents and taxanes, on the other hand, are established inducers of ICD, thereby potentially enhancing the efficacy of PD-1/PD-L1 inhibition. Gemcitabine, for example, has been shown to be highly effective at NK cell activation when administered at low doses. Moreover, the immunomodulatory capability not only varies between different chemotherapeutic agents but is also influenced by the dosage and timing of the respective agent. However, further research is needed to determine optimal treatment schedules [35, 36]."

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors addressed all my comments appropriately. I recommend the acceptance of this manuscript.