Dear Editor,
We read with great interest the study by Sasaki et al. [1], comparing the outcomes of robot‐assisted radical cystectomy in patients aged ≥ 85 years versus 75–84 years.
Kaplan–Meier analysis of the propensity‐matched cohort (18 patients aged ≥ 85 years vs. 36 aged 75–84) found significantly worse 3‐year overall survival in the older group (59% vs. 83%, p = 0.042). However, IPTW‐adjusted Cox regression found no significant difference in overall survival between the same two age groups (HR 1.34, 95% CI 0.69–2.59, p = 0.386). These are two different statistical approaches applied to differently constructed samples (a matched 1:2 subset vs. a reweighted version of the full 26‐vs.‐193 cohort), addressing the same underlying question, and they produce opposite conclusions about statistical significance [2]. The discussion resolves this tension by emphasizing the IPTW result as the operative conclusion, stating that “there were no significant differences in OS,” while characterizing the significant Kaplan–Meier finding as an “expected” consequence of age rather than a result that requires explanation. Given that these two methods disagree specifically on the one survival endpoint most central to how the paper frames its safety conclusions, some discussion of why the matched and weighted analyses diverge, for example, whether the larger IPTW sample better accounts for residual confounding, or whether the matched cohort's smaller size explains its different result, would help readers judge which estimate to weigh more heavily rather than the paper implicitly favoring the null result [3].
The perioperative comparisons between the 18 patients aged ≥ 85 years and 36 aged 75–84, several of which reached nominal significance (urinary diversion type distribution, p = 0.047; lymph node dissection rate, p = 0.024; operative time, p = 0.041), without any correction for the number of comparisons performed [4]. At this sample size, several reported categories contained very few or zero events, for example, zero patients receiving orthotopic neobladder and zero major complications at 90 days in either group, limiting these specific comparisons to descriptive observations rather than statistically reliable estimates. The second stated limitation acknowledges the small sample size in general terms but does not address the specific risk that, with this many comparisons run on 18 patients, one or more nominally significant findings could reflect chance variation rather than a true age‐related difference in surgical practice [5].
This study addressed a genuinely underexamined question using a thoughtful combination of propensity matching and inverse probability weighting, and the multicenter Japanese cohort offered valuable data on a population with limited existing evidence. The conclusion that RARC is safe and oncologically comparable in patients aged ≥ 85 years would be strengthened by directly addressing why the matched Kaplan–Meier and IPTW‐Cox analyses disagree on overall survival, and by tempering the perioperative comparisons in light of the small matched sample and the number of comparisons performed.
Author Contributions
R.K.R. contributed to conceptualization, methodology, literature search, study selection, data extraction, quality assessment, statistical interpretation, manuscript drafting, review and editing, and final approval of the manuscript.
Funding
The author has nothing to report.
Disclosure
Generative AI tools were used solely for language refinement and formatting assistance. All scientific interpretation, literature synthesis, methodological evaluation, and manuscript preparation were independently performed by the author.
Ethics Statement
The author has nothing to report.
Conflicts of Interest
The author declares no conflicts of interest.
Acknowledgments
The author has nothing to report.
Data Availability Statement
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
References
- 1. Sasaki Y., Izumi K., Fukuta K., et al., “Robot‐Assisted Radical Cystectomy in the Very Elderly: A Propensity Score‐Matched Comparison of Outcomes Between Patients Aged ≥ 85 Versus 75–84 Years,” Geriatrics & Gerontology International 26 (2026): e70346, 10.1111/ggi.70346. [DOI] [PubMed] [Google Scholar]
- 2. Park S. Y., Ahn J., Lee J.‐H., Kwon J.‐W., and Lee H.‐N., “Overlap Weights for Binary Outcomes: A Performance Assessment,” Pharmacoepidemiology and Drug Safety 34 (2025): e70253, 10.1002/pds.70253. [DOI] [PubMed] [Google Scholar]
- 3. Cao Z., Ghazi L., Mastrogiacomo C., Forastiere L., Wilson F. P., and Fan L., “Using Overlap Weights to Address Extreme Propensity Scores in Estimating Restricted Mean Counterfactual Survival Times,” American Journal of Epidemiology 194 (2025): 2402–2411, 10.1093/aje/kwae416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Melendez C. R., “Understanding Multiplicity Issues in Statistical Analysis,” American Journal of Nursing 125 (2025): e9–e12, 10.1097/ajn.0000000000000199. [DOI] [PubMed] [Google Scholar]
- 5. Kanduri C., Mamica M., Olstad E. W., Zucknick M., Li J. J., and Sandve G. K., “Beware of Counter‐Intuitive Levels of False Discoveries in Datasets With Strong Intra‐Correlations,” Genome Biology 26 (2025): 249, 10.1186/s13059-025-03734-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.