Letter to the Editor. Residual meningiomas

Rui Sui MD and Haozhe Piao MD
View More View Less
  • Cancer Hospital of China Medical University, Liaoning Cancer Hospital & Institute, Shenyang, Liaoning Province, China
Free access

If the inline PDF is not rendering correctly, you can download the PDF file here.

TO THE EDITOR: We read with great interest the research published by Materi et al.1 (Materi J, Mampre D, Ehresman J, et al. Predictors of recurrence and high growth rate of residual meningiomas after subtotal resection. J Neurosurg. Published online January 3, 2020. doi:10.3171/2019.10.JNS192466). Based on a retrospective cohort study, the authors explored the risk factors for recurrence and the high growth rate of residual meningiomas after subtotal resection. This article differentiates the risk stratification of residual meningiomas after subtotal resection and provides strong evidence for the diagnosis and treatment of meningiomas, which is worth encouraging and approving. However, after reading this article carefully, we have several issues to point out.

First, several data points need an explanation. For preoperative tumor volume in Table 2, the hazard ratio (HR) was 4.720, while the 95% confidence interval (CI) was between 1.042 and 3.146. The authors need to comment on this unlikely occurrence. Similarly, on page 3, the authors described that “the most significant benefit occurred with preoperative tumor volume < 10 cm3 (HR 0.396, 95% CI 1.446–4.399, p = 0.0004).” Here, the HR was 0.396, while the 95% CI was between 1.446 and 4.399. This represents another potential issue that requires comment. In addition, in the abstract, the p value of falcine location should be 0.021, not 0 0.021.

Second, in their statistical analysis, the authors did not clarify the representation of continuous variables (such as age or preoperative Karnofsky Performance Scale [KPS] score) or classified variables (such as sex or race). Additionally, the Kaplan-Meier (K-M)2,3 curve is mentioned twice in Figs. 1 and 2, but it is not described in the statistical analysis, which is confusing for readers. Moreover, the authors described in their statistical analysis that “stepwise multivariate proportional hazards [namely, Cox proportional hazards] regression analyses were performed to identify potential associations....” However, we disagree that this statistical analysis is appropriate. The use of Cox proportional hazards regression modeling should be in accordance with the requirement of a proportional hazard assumption.4 The authors did not verify the proportional hazard assumption in this study. Additionally, it is important to point out that there are noticeable intersections between the two survival curves in Fig. 2B and D, which indicate that the use of the K-M method and stepwise multivariate proportional hazards regression analyses is inappropriate in this study. In our opinion, a competing risk model5,6 is more suitable for the data analysis in their study.

Disclosures

The authors report no conflict of interest.

References

  • 1

    Materi J, Mampre D, Ehresman J, Predictors of recurrence and high growth rate of residual meningiomas after subtotal resection. J Neurosurg. Published online January 3, 2020. doi:10.3171/2019.10.JNS192466

    • Search Google Scholar
    • Export Citation
  • 2

    Lacny S, Wilson T, Clement F, Kaplan–Meier survival analysis overestimates cumulative incidence of health-related events in competing risk settings: a meta-analysis. J Clin Epidemiol. 2018;93:2535.

    • Search Google Scholar
    • Export Citation
  • 3

    van Walraven C, McAlister FA. Competing risk bias was common in Kaplan–Meier risk estimates published in prominent medical journals. J Clin Epidemiol. 2016;69:170173.e8.

    • Search Google Scholar
    • Export Citation
  • 4

    Wen CP, Wai JP, Tsai MK, Minimum amount of physical activity for reduced mortality and extended life expectancy: a prospective cohort study. Lancet. 2011; 378(9798):12441253.

    • Search Google Scholar
    • Export Citation
  • 5

    Shen W, Sakamoto N, Yang L. Cancer-specific mortality and competing mortality in patients with head and neck squamous cell carcinoma: a competing risk analysis. Ann Surg Oncol. 2015;22(1):264271.

    • Search Google Scholar
    • Export Citation
  • 6

    Austin PC, Lee DS, Fine JP. Introduction to the analysis of survival data in the presence of competing risks. Circulation. 2016;133(6):601609.

    • Search Google Scholar
    • Export Citation
View More View Less
  • 1 Johns Hopkins University, Baltimore, MD
  • 2 Mayo Clinic, Jacksonville, FL
Keywords:

Response

We thank Drs. Sui and Piao for their interest in our research and for their feedback on our article. On repeated statistical analysis, we corroborate the typographical errors noted in their response. Corrected values are reflected below and will be remedied via erratum:

  • • Preoperative tumor volume: HR 1.008, 95% CI 1.002–1.013
  • • “the most significant benefit occurred with preoperative tumor volume < 10 cm3 (HR 0.396, 95% CI 0.227–0.691)”
  • • Falcine location: p = 0.021

We also confirm the variable classifications raised by Sui and Piao, where age and preoperative KPS score were evaluated as continuous variables, while sex, race, and tumor location were evaluated as nominal variables. In light of these corrections and clarifications, the overall findings and conclusions of our study remain unchanged. As it pertains to the statistical analyses selected for this study, we opted to use Cox proportional hazards modeling because it is widely accepted for time-to-event data with censoring and covariate analysis.1 The identified factors were then presented using time-to-event curves in Figs. 1 and 2 from the original article. Sui and Piao are correct to point out that such statistical analyses potentiate inherent biases (e.g., lost to follow-up, death), resulting in overestimation of outcome, as well as proportionality assumptions. The Fine-Gray model was reconsidered using STATA 16.0 (StataCorp LLC) to newly create Fig. 1 and account for the following competing risks: lost to follow-up and death, totaling 36 and 3 risk events, respectively. Lost to follow-up was defined based on the standard of care at our institution, which instructs for annual MR images during the first 5 years after surgery. Therefore, patients were classified as being lost to follow-up if they had surgery more than 5 years prior to our study with less than 5 years of MR images, or if they had surgery within 5 years of our study but had no MR images within 2 years prior to the start of our study. In the multivariate competing risk analysis for tumor recurrence, falcine tumor location remained significant (HR 1.973, 95% CI 1.063–3.661, p = 0.031), as did tentorial location (HR 2.391, 95% CI 1.127–5.072, p = 0.023), while African American race (HR 1.488, 95% CI 0.852–2.601, p = 0.163) and preoperative tumor volume (HR 1.005, 95% CI 1.000–1.010, p = 0.062) trended toward significance (Fig. 1).

FIG. 1.
FIG. 1.

Competing risk curves of recurrence-free survival in all subtotally resected meningiomas dichotomized for preoperative volume < 10 cm3 (A), falcine location (B), tentorial location (C), and African American race (D). Note: new figure created based on suggested analysis from Sui and Piao.

We hope this additional analysis serves to address the comments raised by Sui and Piao. We thank them for the opportunity to bring clarity in the dissemination of our results to promote research and understanding that translates to the bedside.

References

1

Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457481.

If the inline PDF is not rendering correctly, you can download the PDF file here.

Contributor Notes

Correspondence Haozhe Piao: drpiaohaozhe@sina.com.

INCLUDE WHEN CITING Published online June 12, 2020; DOI: 10.3171/2020.2.JNS20447.

Disclosures The authors report no conflict of interest.

  • View in gallery

    Competing risk curves of recurrence-free survival in all subtotally resected meningiomas dichotomized for preoperative volume < 10 cm3 (A), falcine location (B), tentorial location (C), and African American race (D). Note: new figure created based on suggested analysis from Sui and Piao.

  • 1

    Materi J, Mampre D, Ehresman J, Predictors of recurrence and high growth rate of residual meningiomas after subtotal resection. J Neurosurg. Published online January 3, 2020. doi:10.3171/2019.10.JNS192466

    • Search Google Scholar
    • Export Citation
  • 2

    Lacny S, Wilson T, Clement F, Kaplan–Meier survival analysis overestimates cumulative incidence of health-related events in competing risk settings: a meta-analysis. J Clin Epidemiol. 2018;93:2535.

    • Search Google Scholar
    • Export Citation
  • 3

    van Walraven C, McAlister FA. Competing risk bias was common in Kaplan–Meier risk estimates published in prominent medical journals. J Clin Epidemiol. 2016;69:170173.e8.

    • Search Google Scholar
    • Export Citation
  • 4

    Wen CP, Wai JP, Tsai MK, Minimum amount of physical activity for reduced mortality and extended life expectancy: a prospective cohort study. Lancet. 2011; 378(9798):12441253.

    • Search Google Scholar
    • Export Citation
  • 5

    Shen W, Sakamoto N, Yang L. Cancer-specific mortality and competing mortality in patients with head and neck squamous cell carcinoma: a competing risk analysis. Ann Surg Oncol. 2015;22(1):264271.

    • Search Google Scholar
    • Export Citation
  • 6

    Austin PC, Lee DS, Fine JP. Introduction to the analysis of survival data in the presence of competing risks. Circulation. 2016;133(6):601609.

    • Search Google Scholar
    • Export Citation
  • 1

    Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457481.

Metrics