数字 PCR 与实时荧光定量 PCR 检测慢性髓性白血病患者 BCR::ABL mRNA 水平的比较性研究

目的：

比较数字 PCR（dPCR）与实时荧光定量 PCR（qPCR）方法检测慢性髓性白血病（CML）患者外周血 BCR::ABL（P210） mRNA 水平的差异性、相关性与一致性。

方法：

本研究为非干预性、横断面研究。收集 2021 年 9 月至 2023 年 2 月在北京大学人民医院就诊的服用酪氨酸激酶抑制剂（TKI）至少获得完全细胞遗传学反应（CCyR）的 CML 患者同时采用 dPCR 和 qPCR 方法检测 BCR::ABL mRNA 水平的结果，分别采用 Wilcoxon 符号秩检验、Spearman 相关系数、Bland-Altman 分析比较两种方法的差异性、相关性、一致性。

结果：

356 例 CML 患者的 459 对外周血样本，分别采用 dPCR 与 qPCR 方法检测 BCR::ABL mRNA 水平并纳入统计分析。整体而言，两种方法检测结果差异具有统计学意义（P < 0.001）。按分子学反应分层分析显示，该差异仅存在于达到 MR4.5 及以上水平者（P < 0.001）；未达主要分子学反应（MMR）（P = 0.922）、达到 MMR（P = 0.723）及分子学反应 4（MR4）（P = 0.099）层级均无明显差异。总体分析可见，dPCR 与 qPCR 检测 BCR::ABL mRNA 水平呈中度相关（r = 0.761，P < 0.001）。随分子学反应程度加深，二者相关性逐步降低直至无明显相关：未达 MMR（r = 0.929，P < 0.001）、MMR（r = 0.815，P < 0.001）、MR4（r = 0.408，P < 0.001）、MR4.5（r = 0.176，P = 0.176）。一致性分析结果提示，MR4.5 组两种检测方法的一致性显著差于其余分层：未达 MMR（r = 0.042，P = 0.846）、MMR（r = 0.054，P = 0.229）、MR4（r = −0.020，P = 0.399）、MR4.5（r = −0.219，P < 0.001）。

结论：

CML 患者当获得深层分子学反应后，更适合采用 dPCR 方法监测 BCR::ABL（P210）mRNA 水平以提高精准性。

Objective: 

To compare digital polymerase chain reaction (dPCR) and real-time quantitative PCR (qPCR) measurements of BCR::ABL (P210) mRNA expression in patients with chronic myeloid leukemia (CML).

Methods: 

In this non-interventional, cross-sectional study, BCR::ABL (P210) mRNA was simultaneously measured by dPCR and qPCR in peripheral blood samples collected from patients with CML who underwent tyrosine kinase inhibitor therapy and who achieved at least a complete cytogenetic response from September 2021 to February 2023 at Peking University People’s Hospital. The difference, correlation, and agreement between the two methods were evaluated using the Wilcoxon signed-rank test, Spearman’s correlation, and Bland-Altman analysis, respectively.

Results: 

In total, 459 data pairs for BCR::ABL mRNA expression measured by dPCR and qPCR from 356 patients with CML were analyzed. There was a significant difference in BCR::ABL mRNA expression between the two methods (P < 0.001). When analyzed by the depth of the molecular response (MR), a significant difference only existed for patients with ≥ MR4.5 (P < 0.001). No significant difference was observed for those who did not achieve a major MR (no MMR; P = 0.922) or for those who achieved a major MR (MMR; P = 0.723) or MR4 (P = 0.099). There was a moderate correlation between the BCR::ABL mRNA expression between the two methods (r = 0.761, P < 0.001). However, the correlation gradually weakened or disappeared as the depth of the MR increased (no MMR: r = 0.929, P < 0.001; MMR: r = 0.815, P < 0.001; MR4: r = 0.408, P < 0.001; MR4.5: r = 0.176, P = 0.176). In addition, the agreement in BCR::ABL mRNA expression between the two methods in those with MR4.5 was weaker than other groups (no MMR: r = 0.042, P = 0.846; MMR: r = 0.054, P = 0.229; MR4: r = -0.020, P = 0.399; MR4.5: r = -0.219, P < 0.001) .

Conclusion: 

dPCR is more accurate than qPCR for measuring BCR::ABL (P210) mRNA expression in patients with CML who achieve a stable deep MR.