AUTHORS
Jingyi Yang, PhD, Hao Guo, MD, PhD2, Qianying Pan, MD3, Xingchen Liu, MD4, Qing Zhang, MD5, Fenghua Gao6, quande Lin, PhD7, Hu Zhou, MD8, Qinglan Zhang, MS9, Qian Li10, Xueli Jiao, MD11, Ran Zhao, MD12, Jianwei Du13, Lihua Dong14, Jinxin Dong15, Qingxin Xia16, Yanmei Li17, Qiuhong Wang18, shuhao mei19, Ping Wang20, Lu-Gui Qiu, PhD21, Keshu Zhou, PhD22
AFFILIATIONS
1 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
2 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
3 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
4 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
5 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
6 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
7 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
8 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
9 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
10 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
11 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
12 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
13 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
14 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
15 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
16 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
17 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
18 Jiyuan People's Hospital, Jiyuan, China
19 Xuchang Central Hospital?, Xuchang, China
20 Zhumadian Central Hospital?, Zhumadian, China
21 State Key Laboratory of Experimental Hematology, National Clinical Research Center for Blood Diseases, Haihe Laboratory of Cell Ecosystem, Institute of Hematology & Blood Diseases Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, China
22 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, Henan, China
ABSTRACT
Background
POLARIX established Pola-R-CHP as a first-line standard for DLBCL. Whether real-world (RW) first-line pola-based therapy in Chinese patients matches trial efficacy is unknown. We performed a matching-adjusted indirect comparison (MAIC) against POLARIX. A methodological caution in RW response reporting is that restricting complete-response (CR) analysis to patients who complete a fixed number of cycles introduces immortal-time bias, because completing therapy requires surviving and remaining progression-free long enough to do so. We therefore report CR using a mature-cohort estimand (all patients reaching end-of-treatment assessment, with early progression or death counted as non-CR) and provide the completer-restricted estimate only as a bias-illustrating appendix.
Methods
RW patients receiving first-line pola-based therapy (n=602; 402 with complete matching covariates) were compared with POLARIX (Tilly 2022 NEJM; 28.2-month follow-up, ITT) using unanchored MAIC with entropy-balancing weights (Signorovitch 2010). Six matching covariates were balanced: age >60, female sex, IPI 3–5 (a force-retained balancing covariate, not an eligibility criterion), stage III–IV, elevated LDH, and cell-of-origin (GCB). The primary endpoints were the overall response rate (ORR) and complete response (CR) rate, compared between the weighted RW pola-based group and each POLARIX arm (R-CHOP the primary comparison; Pola-R-CHP a consistency check). CR was analysed under a prespecified mature-cohort estimand (all patients reaching end-of-treatment [EOT] assessment); a completer-restricted (≥6 cycles) estimate is reported separately in the appendix and flagged as subject to immortal-time bias. Confidence intervals for the weighted CR were obtained from the MAIC robust (sandwich) standard error; an earlier center-clustered bootstrap gave concordant limits but was unstable at this sample size owing to occasional degenerate entropy-balancing weights. Post-weighting covariate balance and effective sample size (ESS) were assessed.
Results
After weighting, all six covariates matched the POLARIX target to within 0.00 percentage points (pp); ESS=268 (67% retained). In the full (ITT-like) RW population, weighted ORR was 95.2% and CR 65.3%. Under the mature-cohort estimand, weighted CR was 65.3% (robust-SE 95% CI 59.0–71.6), matching the full-population estimate.
ORR (robust across subsets): RW ORR exceeded R-CHOP (83.8%) by +11.4pp (95% CI 6.9–15.8; P<0.0001) and Pola-R-CHP (85.5%) by +9.7pp (5.4–14.1; P<0.0001). ORR was stable at ~95% across all subsets (full/prospective-only/original-source), indicating a consistent real-world response advantage.
CR (mature-cohort estimand): weighted RW CR was 65.3%, lower than R-CHOP (74.0%; RD −8.7pp, 95% CI −16.3 to −1.2; P=0.023) and Pola-R-CHP (78.0%; RD −12.7pp, 95% CI −20.1 to −5.3; P<0.001). This difference was statistically significant against Pola-R-CHP and is not explained by follow-up maturity: the mature-cohort CR (65.3%) equals the full-population CR (65.3%).
Immortal-time-bias check (appendix). Restricting to patients completing ≥6 cycles raised the weighted CR from 65.3% (mature cohort) to 79.3%. This increase is a known consequence of immortal-time bias—patients who progress or die early cannot enter the completer group—and does not represent higher efficacy. Accordingly, the completer-restricted CR (79.3%) is reported only to quantify the magnitude of the bias and is not used for comparison with POLARIX. The unbiased mature-cohort estimand (CR 65.3%) is the basis for all CR comparisons above; the ORR advantage is robust and does not depend on the estimand.
Conclusions
Real-world first-line pola-based therapy achieved ORR significantly higher than both POLARIX arms (+11.4pp vs R-CHOP; +9.7pp vs Pola-R-CHP), robust across all sensitivity subsets. Under the unbiased mature-cohort estimand, weighted CR was 65.3% (95% CI 59.0–71.9), numerically lower than the POLARIX Pola-R-CHP arm (78.0%; RD −12.7pp, 95% CI −20.1 to −5.3). The higher CR seen when analysis is restricted to treatment completers (79.3%) reflects immortal-time bias and should not be interpreted as efficacy. Because this is an unanchored MAIC in which residual confounding cannot be excluded, these findings are hypothesis-generating: they support a robust real-world ORR advantage while indicating that CR may be lower than in POLARIX, warranting confirmation in adjusted or prospective comparisons.
DISCLOSURES
Jingyi Yang, PhD: No relevant conflicts of interest to declare.