To the Editor We were interested by a recently published article in JAMA Oncology.¹ The meta-analysis included 43 trials containing 8 randomized clinical trials with 5431 patients to compare the efficacy of neoadjuvant chemoimmunotherapy and neoadjuvant chemotherapy. They found that neoadjuvant chemoimmunotherapy was superior to neoadjuvant chemotherapy across surgical, pathological, and efficacy outcomes. This is a promising result for neoadjuvant chemoimmunotherapy, an emerging oncology treatment approach. However, we had some questions about the meta-analysis methods.

First, in the statistical analysis section, the authors did not indicate how to evaluate the heterogeneity on the I² value but mentioned that 鈥淔or each analysis, we included the 95% CIs, I² statistic, 蟿² statistic, and 蠂²; all tests were 2-sided, and a P鈥�<鈥�.05 was considered significant unless otherwise indicated.鈥�¹ To set a P value of less than .05 in the I² statistic as the heterogeneity judgment is not recommended by the Cochrane Collaboration. Chapter 10.10.2 in the Cochrane Handbook (version 6.3) said that the 蠂² test must be interpreted cautiously because it has low power in the (common) situation of a meta-analysis when studies have a small sample size or are few in number. A P value of .10, rather than the conventional level of .05, is sometimes used to determine statistical significance because a statistically significant result may indicate a problem with heterogeneity, and a nonsignificant result must not be taken as evidence of no heterogeneity.² In the meta-analysis, it may miss heterogeneity detection because some outcomes by subgroup were pooled for analysis with a small sample size.