Effect of Lenalidomide Maintenance in Chronic Lymphocytic Leukemia: A Meta-Analysis and Trial-Sequential Analysis
Round 1
Reviewer 1 Report
This manuscript, written by Dr. TY Yu, original research, with the title of “Effect of lenalidomide maintenance in chronic lymphocytic leukemia: a meta-analysis and trial-sequential analysis” made a meta-analysis evaluating the effect of lenalidomide in CLL. The manuscript is well written, it is easy to read and to understand.
Not all potential readers may be familiar with meta-analysis, but as clinicians may be interested in lenalidomide and CLL. Therefore, the authors may add figure legends explaining how to interpret the data and results shown in the figures.
Additional comments:
(1) Line 16-17. Could you please confirm the sentence “disease that cannot be treated with standard therapy”?
(2) Line 40. Could you please confirm that the use of “obviously” is correct?
(3) Line 47. Could you please add some information regarding the predictive markers? For example, del(17p) or TP53 mutations; IGHV mutation status, ZAP-70, and CD38; NOTCH1 mutations; BTK, PLCg2, and BCL2 mutations; Rai, Binet, lymphocyte double time, B2 microglobulin, del 11q, del 13q, trisomy 12; etc.
(4) Lines 48-59. If you agree, you could mention that the initial management of CLL depends on
- 1 or more disease-related complications
- del17p or TP53 mutation
- IGHV mutational status
- Assessment of fitness
With those factors, an algorithm can be created. Other drugs used are Obinutuzumab, acalabrutinib, venetoclax (BCL2 inhibitor), and combinations.
(5) Line 76. If you agree, could you please mention that “Lenalidomide, an immunomodulatory agent, has been used in the treatment of patients with multiple myeloma, certain subtypes of non-Hodgkin lymphoma such as mantle cell lymphoma, and CLL, the last as part of clinical trials. Based upon review of the trial data, it appears that lenalidomide therapy causes no specific immune dysfunction that would increase the risk of opportunistic infections in these patients. As such, no routine antimicrobial prophylaxis is generally provided to patients receiving this agent. However, this agent is not approved for therapy of CLL patients outside clinical trials.” (please paraphrase it).
Additional useful sentences are “Lenalidomide is a 4-amino-glutamyl analogue of thalidomide that lacks the neurologic side effects of sedation and neuropathy and has emerged as a drug with activity against various hematological and solid malignancies. It is approved by FDA for clinical use in myelodysplastic syndromes with deletion of chromosome 5q and multiple myeloma. Lenalidomide has been shown to be an immunomodulator, affecting both the cellular and humoral limbs of the immune system. It has also been shown to have anti-angiogenic properties (https://doi.org/10.1186/1756-8722-2-36)”.
(6) In line 106, it is stated that “the diagnosis of CLL was determined on the basis of the criteria established by the International Workshop on Chronic Lymphocytic Leukemia”. Could you please confirm that the diagnosis of CLL was also made according to the 2016 revision of the World Health Organization classification of lymphoid neoplasms?
(7) Could you please add a list of abbreviations at the end of the manuscript?
(8) In line 151, the acronym “TSA” appears. Not all authors may know the meaning (trial-sequential analysis?).
(9) Regarding the characteristics of the included studies. Only 4 series were included. I understand that this is a limitation, due to the available scientific evidence. In the authors opinion, is it adequate? Is the number of patients enough?
(10) I^2 indicates the level of heterogeneity. It can take values from 0% to 100%. If I^2 ≤ 50%, studies are considered homogeneous, and a fixed effect model of meta-analysis can be used. If I^2 > 50%, the heterogeneity is high, and one should use a random effect model for meta-analysis. In Figure 2A and 2B, the I^2 > 50%, so in the text, the authors report the random effect model. Should the “fixed” data be eliminated from the Figure because it is not useful for this case? This may create confusion in the reader. Or at least, add a comment in the figure legend.
(11) Could you please explain in Figure 2 what TE and seTE means?
(12) In a forest plot, usually the intervention group and the control group have a column indicating the number of patients (n/N).
(13) In Figure 3. Could you please add a legend explaining how to interpret a TSA Figure? (https://doi.org/10.1186/s12874-017-0315-7). The potential readers may be hematologists or hematopathologists without any previous knowledge of meta-analysis.
(14) Could you please explain why the SAE and FAE did not use a fixed effect model?
(15) Since there is no limitation of space in the mdpi format, why not include the forest plots for the data of Table 2. It could be added to the appendix or as supplementary information.
Author Response
Reviewer 1
This manuscript, written by Dr. TY Yu, original research, with the title of “Effect of lenalidomide maintenance in chronic lymphocytic leukemia: a meta-analysis and trial-sequential analysis” made a meta-analysis evaluating the effect of lenalidomide in CLL. The manuscript is well written, it is easy to read and to understand.
Not all potential readers may be familiar with meta-analysis, but as clinicians may be interested in lenalidomide and CLL. Therefore, the authors may add figure legends explaining how to interpret the data and results shown in the figures.
Response:
Many thanks for your suggestions on our manuscript. We have added figure legends explaining how to interpret the data and results shown in the figures in detail. Please refer to the descriptions on each figure.
Revision in manuscript:
Figure Legends
Figure 1. Flow diagram of the identification process for eligible studies
Figure 2. Meta-analysis of primary outcomes regarding the progression free survival (A), overall survival (B)
PFS, progression-free survival; OS, overall survival; HR, hazard ratio; CI, confidence interval
Figure 2A, Meta-analysis of progression free survival. The analysis included data from 5 comparisons with a total of 385 patients receiving lenalidomide maintenance therapy and 348 patients receiving observation alone. The p value for heterogeneity was 0.06 and I square for heterogeneity was 57%, indicating moderate heterogeneity. We adopt random-effects model and lenalidomide maintenance therapy was associated with a statistically significant effect in prolongation of PFS. (HR, 0.43; 95% CI, 0.28–0.68).
Figure 2B, Meta-analysis of overall survival. The analysis included data from 4 comparisons with a total of 360 patients receiving lenalidomide maintenance therapy and 333 patients receiving observation alone. The p value for heterogeneity was 0.10 and I square for heterogeneity was 52%, indicating moderate heterogeneity. We adopt random-effects model and lenalidomide maintenance showed no difference in OS, compared with observation (HR, 0.62; 95% CI, 0.29–1.30).
Figure 3. Trial sequential analysis (TSA) of primary outcomes regarding the progression free survival (A), overall survival (B)
X-axis, the number of patients (information size); Y-axis, cumulative Z-score; Horizontal green lines, conventional boundaries; Sloping red lines at the top left-hand corners (trial sequential boundaries) represent the TSA threshold for statistical significance. Vertical red full line represents the required information size (RIS). Solid blue line is the cumulative Z-curve.
Figure 3A, Trial sequential analysis of progression free survival. The cumulative number of patients did not exceed the required information size of 923, but the Z-curves surpassed the significance boundary in favor of lenalidomide maintenance therapy, suggesting a conclusive result.
Figure 3B, Trial sequential analysis of overall survival. The cumulative number of patients did not exceed the required information size of 769, and the Z-curves did not surpass any significance boundary either, suggesting an in-conclusive result.
Additional comments:
(1) Line 16-17. Could you please confirm the sentence “disease that cannot be treated with standard therapy”?
Response:
Many thanks for your precise comments on our manuscript. We will carefully revise the text according to your suggestions and make the text more clear and concise. Please refer to the clean version of the manuscript for more details.
Revision in manuscript:
Abstract
Chronic lymphocytic leukemia (CLL) is the most common lymphoproliferative disease in adults. Despite durable responses to front-line therapy and sustained remission rates in patients with CLL, CLL is still incurable within standard therapy and relapses eventually.
(2) Line 40. Could you please confirm that the use of “obviously” is correct?
Response:
Many thanks for your precise comments on our manuscript. We will carefully revise the text according to your suggestions and make the text more clear and concise. Please refer to the clean version of the manuscript for more details.
Revision in manuscript:
Introduction
The annual incidence of CLL in Western countries is 4.2 per 100,000 and gradually increases with age.
(3) Line 47. Could you please add some information regarding the predictive markers? For example, del(17p) or TP53 mutations; IGHV mutation status, ZAP-70, and CD38; NOTCH1 mutations; BTK, PLCg2, and BCL2 mutations; Rai, Binet, lymphocyte double time, B2 microglobulin, del 11q, del 13q, trisomy 12; etc.
Response:
Many thanks for your valuable comments on our manuscript. We agree that the introduction should add the information of predictive markers of CLL. Thus, we will carefully revise the introduction.
Revision in manuscript:
Introduction
Several predictive markers are associated with prognosis in CLL. Immunoglobulin heavy chain variable region (IGHV) gene mutation is associated with high response rates and improved overall survival (OS) in patients receiving FCR (fludarabine, cyclophosphamide, and rituximab). Cytogenetic abnormalities detected by fluorescence in situ hybridization (FISH) are present in more than 80% of patients with previously untreated CLL. Del(13q) is associated with good prognosis and longest median survival (133 months). Del(11q) is generally associated with disease progression and a shorter median survival (79 months). Del(17p), whitch reflects deletion of the TP53 gene, with poor response to chemotherapy, short treatment-free intervals, and short median survival (32 months). Among the cell surface markers (CD38 and ZAP-70) detected by flow cytometry or immunohistochemistry CD38 expression and/or ZAP-70 were associated with shorter PFS and OS outcomes. An integrative prognostic model, including NOTCH1, SF3B1, and BIRC3 mutations with remission of cytogenetic abnormalities detected by FISH, has been proposed to classify patients into four distinct prognostic subgroups: high-risk (TP53 and/or BIRC3 abnormalities); intermediate risk [NOTCH and/or SF3B1 mutations and/or del(11q)]; low risk (trisomy 12 and wild-type for all genetic lesions) and very low risk [del(13q) only]. The 10-year survival rates for the four subgroups were 29%, 37%, 57%, and 69%, respectively."
(4) Lines 48-59. If you agree, you could mention that the initial management of CLL depends on
1 or more disease-related complications
del17p or TP53 mutation
IGHV mutational status
Assessment of fitness
With those factors, an algorithm can be created. Other drugs used are Obinutuzumab, acalabrutinib, venetoclax (BCL2 inhibitor), and combinations.
Response:
Many thanks for your valuable comments on our manuscript. We agree that the introduction should add the initial management of CLL with markers such as del(17p)/TP53 and IGHV mutation. Thus, we will carefully revise the introduction.
Revision in manuscript:
Introduction
Chemoimmunotherapy is not recommended because del(17p)/TP53 mutations are associated with low response rates. The PFS benefit of acalabrutinib +-obinutuzumab was observed in CLL patients with del(17p) or TP53 mutations, IGHV-unmutated, and IGHV-mutated CLL. In patients with del(17p) or TP53 mutations, the undetectable-MRD rate and PFS rate were significantly higher with venetoclax + obinutuzumab than with chlorambucil + obinutuzumab.
(5) Line 76. If you agree, could you please mention that “Lenalidomide, an immunomodulatory agent, has been used in the treatment of patients with multiple myeloma, certain subtypes of non-Hodgkin lymphoma such as mantle cell lymphoma, and CLL, the last as part of clinical trials. Based upon review of the trial data, it appears that lenalidomide therapy causes no specific immune dysfunction that would increase the risk of opportunistic infections in these patients. As such, no routine antimicrobial prophylaxis is generally provided to patients receiving this agent. However, this agent is not approved for therapy of CLL patients outside clinical trials.” (please paraphrase it).
Additional useful sentences are “Lenalidomide is a 4-amino-glutamyl analogue of thalidomide that lacks the neurologic side effects of sedation and neuropathy and has emerged as a drug with activity against various hematological and solid malignancies. It is approved by FDA for clinical use in myelodysplastic syndromes with deletion of chromosome 5q and multiple myeloma. Lenalidomide has been shown to be an immunomodulator, affecting both the cellular and humoral limbs of the immune system. It has also been shown to have anti-angiogenic properties (https://doi.org/10.1186/1756-8722-2-36)”.
Response:
Many thanks for your valuable comments on our manuscript. We will revise the manuscript according to your suggestion.
Revision in manuscript:
Introduction
Lenalidomide, an immunomodulatory agent, has been used to treat multiple myeloma, certain subtypes of non-Hodgkin lymphoma such as mantle cell lymphoma, and CLL, the last as part of clinical trials. Based on a review of the trial data, it appears that lenalidomide therapy causes no specific immune dysfunction that would increase the risk of opportunistic infections in these patients. Therefore, routine antimicrobial prophylaxis is not usually offered to patients receiving this agent. However, this agent is not approved for the treatment of CLL patients outside clinical trials.
Lenalidomide is a 4-amino-glutamyl analogue of thalidomide that lacks the neurologic side effects of sedation and neuropathy, and has emerged as a drug with activity against various hematological and solid malignancies. It is FDA-approved for the clinical treatment of myelodysplastic syndromes with chromosome 5q deletion and multiple myeloma. Lenalidomide has been shown to be an immunomodulator, affecting both the cellular and humoral limbs of the immune system. It has also been shown to have anti-angiogenic properties.
Lenalidomide has a special mechanism of action: it can target cancer cells and modulate or interrupt several interactions of CLL cells and elements in their microenvironment [23-25] that can stimulate leukemia or have an impact on survival [26, 27]. The efficacy and safety of lenalidomide as maintenance therapy for CLL are inconsistent within individual studies and still under debate. Moreover, it remains unknown whether previous meta-analyses established sufficient statistical power to draw a firm conclusion. Therefore, to assess the relative efficacy and safety of lenalidomide maintenance therapies in patients with CLL, we performed a systematic review and meta-analysis of all available phase II and III randomized controlled trials (RCTs) and applied trial sequential analysis (TSA) to examine the statistical power of the meta-analysis.
(6) In line 106, it is stated that “the diagnosis of CLL was determined on the basis of the criteria established by the International Workshop on Chronic Lymphocytic Leukemia”. Could you please confirm that the diagnosis of CLL was also made according to the 2016 revision of the World Health Organization classification of lymphoid neoplasms?
Response:
Many thanks for your valuable comments on our manuscript. We will revise the manuscript according to your suggestion.
Revision in manuscript:
Methods
The diagnosis of CLL was determined on the basis of the criteria established by the International Workshop on Chronic Lymphocytic Leukemia and also made according to the 2016 revision of the World Health Organization classification of lymphoid neoplasms.
(7) Could you please add a list of abbreviations at the end of the manuscript?
Response:
Many thanks for your valuable comments on our manuscript. We will add a list of abbreviations according to your suggestion.
Revision in manuscript:
Abbrevations
BCL-2: B-cell lymphoma-2 protein
BR: Bendamustine and Rituximab
BTK: Bruton’s tyrosine kinase
CIs: confidence intervals
CLL: chronic lymphocytic leukemia
CR: Complete response
DB: Double blind
FAE: fatal adverse events
FCR: Fludarabine, Cyclophosphamide, and Rituximab
FCRB: Fludarabine, Cyclophosphamide, Rituximab, and Bendamustine
FDA: Food and Drug Administration
FISH: fluorescence in situ hybridization
FR: Fludarabine and Rituximab
TD: treatment discontinuation
HR: Hazard ratio
IGHV: immunoglobulin heavy chain variable region
ITT: intention-to-treat
MC: Multiple centers
MRD: minimal residual disease
OP: Open label
ORs: odds ratios
OS: overall survival
PFS: progression-free survival
PI3K: phosphoinositide 3-kinase
PR: Partial response
R-Chlorambucil: Rituximab and Chlorambucil
RCTs: randomized controlled trials
SAE: serious adverse events
SC: Single center
TSA: trial sequential analysis
(8) In line 151, the acronym “TSA” appears. Not all authors may know the meaning (trial-sequential analysis?).
Response:
We appreciate your kind reminder. The acronym “TSA” would be trial sequential analysis and we will revise the manuscript.
Revision in manuscript:
Methods
Trial sequential analysis (TSA)
(9) Regarding the characteristics of the included studies. Only 4 series were included. I understand that this is a limitation, due to the available scientific evidence. In the authors opinion, is it adequate? Is the number of patients enough?
Response:
Thanks for your kind comments. Indeed, there are only four studies included in our meta-analyses, that would be comparatively small numbers. As a result, to examine the statistical power, we conducted trial sequential analysis (TSA) to provide more information on the precision and uncertainty of meta-analysis results in the two primary outcomes.
In the TSA of PFS, the cumulative number of patients did not exceed the required information size of 923, but the Z-curves surpassed the significance boundary in favor of lenalido-mide maintenance therapy, suggesting a conclusive result and providing convincing statistical evidence to our results. The result indicates that the number of patients in PFS is sufficient, adequate and enough.
In the TSA of OS, the cumulative number of patients did not exceed the required information size of 769, and the Z-curves did not surpass any significance boundary either, suggesting an inconclusive result. The result indicates that the number of patients in OS is insufficient, inadequate and not enough. Consequently, further studies are warranted to provide convincing statistical evidence.
(10) I^2 indicates the level of heterogeneity. It can take values from 0% to 100%. If I^2 ≤ 50%, studies are considered homogeneous, and a fixed effect model of meta-analysis can be used. If I^2 > 50%, the heterogeneity is high, and one should use a random effect model for meta-analysis. In Figure 2A and 2B, the I^2 > 50%, so in the text, the authors report the random effect model. Should the “fixed” data be eliminated from the Figure because it is not useful for this case? This may create confusion in the reader. Or at least, add a comment in the figure legend.
Response:
Thank you for your suggestion, we renew the figures and eliminate the fixed effect model.
Revision in manuscript:
Please refer to the new figure 2.
(11) Could you please explain in Figure 2 what TE and seTE means?
Response:
Thank you for your comments. To avoid confusion, we eliminate the statistical term “TE and seTE”.
Revision in manuscript:
Please refer to the new figure 2.
(12) In a forest plot, usually the intervention group and the control group have a column indicating the number of patients (n/N).
Response:
Thank you for your suggestion, we add the columns indicating the number of patients (n/N).
Revision in manuscript:
Please refer to the new figure 2.
(13) In Figure 3. Could you please add a legend explaining how to interpret a TSA Figure? (https://doi.org/10.1186/s12874-017-0315-7). The potential readers may be hematologists or hematopathologists without any previous knowledge of meta-analysis.
Response:
Thank you for your suggestion, we rephrase the figure legends and attach the explanation. Please refer to the descriptions on each figure.
Revision in manuscript:
Figure 3. Trial sequential analysis (TSA) of primary outcomes regarding the progression free survival (A), overall survival (B)
X-axis, the number of patients (information size); Y-axis, cumulative Z-score; Horizontal green lines, conventional boundaries; Sloping red lines at the top left-hand corners (trial sequential boundaries) represent the TSA threshold for statistical significance. Vertical red full line represents the required information size (RIS). Solid blue line is the cumulative Z-curve.
Figure 3A, Trial sequential analysis of progression free survival. The cumulative number of patients did not exceed the required information size of 923, but the Z-curves surpassed the significance boundary in favor of lenalidomide maintenance therapy, suggesting a conclusive result.
Figure 3B, Trial sequential analysis of overall survival. The cumulative number of patients did not exceed the required information size of 769, and the Z-curves did not surpass any significance boundary either, suggesting an in-conclusive result.
(14) Could you please explain why the SAE and FAE did not use a fixed effect model?
Response:
Thank you for your precise comments. The SAE and FAE should use a fixed effect model. We add the descriptions in the method, and revise Table 2.
Revision in manuscript:
Method - Statistical analysis
Administration of fixed-effect or random-effects was interpreted with statistical heterogeneity by authors. If an estimate of the between-study variance, known as tau-squared, was low (or zero), then we would choose a fixed-effect model; otherwise, we would choose random-effects.
Table 2
Serious adverse events
fixed-effect, OR, 4.64 (2.96 to 7.26)
Fatal adverse events
fixed-effect, OR, 0.86 (0.28 to 2.63)
(15) Since there is no limitation of space in the mdpi format, why not include the forest plots for the data of Table 2. It could be added to the appendix or as supplementary information.
Response:
Thank you for your suggestion. We will include the forest plots for the data of Table 2 in the Supplemental Information 4.
Revision in manuscript:
Please refer to the Supplemental Information 4.
Reviewer 2 Report
The meta-analysis is well written.
Introduction: smooth flow of thoughts
Methods:
Line 112, for this section, what was the raw data input you have found in the articles. e.g was it HR or binary events?
Line 134-135: what was the formula used or cite a reference.
Figures:
Fig3. the legend need to be complete, explaining what is A and B?
Table 2 suggest swab the last two columns because the p value for the pooled estimate is not shown and might be confused with the heterogeneity p value.
Results: Glad that you mentioned Line 229
Discussion
Disapprove using uncommon abbreviation Line 274 and is confusing.
Author Response
Reviewer 2
The meta-analysis is well written.
Introduction: smooth flow of thoughts
Response:
Thank you for your kind comments.
Methods:
Line 112, for this section, what was the raw data input you have found in the articles. e.g was it HR or binary events?
Response:
Thank you for your comments. We input the Hazard Ratio and Confidence Interval (Upper limit and lower limit at 95%) to calculate estimated treatment effect and standard error of individual studies in the Time-to-event outcomes. We input the number of patients analyzed, and number of events per arm in the binary/dichotomous outcomes.
Line 134-135: what was the formula used or cite a reference.
Response:
Thank you for your comments. In Line 134-135 stated the per-protocol and intention-to-treat (ITT) design, we cited a reference to explain the Intention to treat and per protocol analysis in clinical trials.
Reference:
Tripepi G, Chesnaye NC, Dekker FW, Zoccali C, Jager KJ. Intention to treat and per protocol analysis in clinical trials. Nephrology (Carlton). 2020;25(7):513-517. doi:10.1111/nep.13709
Figures:
Fig3. the legend need to be complete, explaining what is A and B?
Response:
Thank you for your comments. We apologize for the misunderstanding, Figure 3A would be the TSA for PFS and Figure 3B would be the TSA for OS. We add figure legends to make the figure more clear.
Table 2 suggest swab the last two columns because the p value for the pooled estimate is not shown and might be confused with the heterogeneity p value.
Response:
Thank you for your comments. To avoid confusing, we revise the column’s name to “Cochran Q p-value for heterogeneity”
Revision in manuscript:
Table 2
Cochran Q p-value for heterogeneity
Results: Glad that you mentioned Line 229
Response:
Thank you for your kind comments.
Discussion
Disapprove using uncommon abbreviation Line 274 and is confusing.
Response:
Thank you for your suggestion, we delete the abbreviation of PFS2 and directly use “The time to second objective disease progression” to avoid confusion.
Author Response File: 
Author Response.docx
