急性髓细胞性白血病组织学分型与颅骨骨髓MRI分型的对照分析

目的 分析急性髓细胞性白血病（AML）颅骨骨髓MRI分型与组织学分型的关系,并探讨MRI分型的临床价值。 方法 回顾性纳入30例经骨髓细胞学检查证实为AML的病人,男19例,女11例,中位年龄46.0岁。AML组织学分型为：M1型1例,M2型14例,M3型4例,M4型6例,M5型4例,M6型1例。30例病人治疗前均行头部MRI平扫,其中13例同时行增强扫描。由2名医师分析颅骨骨髓MRI影像特征并进行分型。采用Pearson卡方检验比较不同MRI分型的AML组织学分型差异,并分析2种分型的相关性。采用Kappa检验分析2名医师阅片结果的一致性。 结果 AML颅骨骨髓MRI主要表现为T₁WI上颅骨骨髓信号减低,增强扫描显示颅骨板障周边信号明显强化。依据MRI特征将颅骨骨髓分为5型：a型5例,b型7例,c型7例,d型9例,e型2例。不同MRI分型的AML组织学分型间差异有统计学意义（P<0.05）。其中,a型中白血病M2型的占比较高（60.00%）,b型中白血病M4型占比较高（57.14%）,d型中白血病M2型占比较高（77.78%）。颅骨骨髓MRI分型与AML组织学分型呈中度相关（Cramer’s V=0.490,P=0.045）。2名医师对M1、M2、M3、M5、M6型中颅骨骨髓MRI分型的一致性均较好（均κ≥0.75）,对M4型中颅骨骨髓MRI 分型的一致性中等（0.4≤κ<0.75）。 结论 颅骨骨髓MRI分型特征可作为评估AML部分组织学分型的有效辅助手段,为临床治疗决策提供依据。

Comparison analysis between histological classification of acute myeloid leukemia and MRI classification of skull bone marrow

Objective To analyze the relationship between the classification of skull bone marrow and the histological classification of acute myeloid leukemia (AML), and to explore the clinical value of MRI classification. Methods All of 30 subjects with AML confirmed by bone marrow cytology were retrospectively analyzed, including 19 males and 11 females, with a median age of 46.0 years. AMI histological classifications: M1 1 case, M2 14 cases, M3 4 cases, M4 6 cases, M5 4 cases, and M6 1 case. All 30 patients underwent head MRI scan before treatment, and 13 of them underwent enhanced scan at the same time. We analyzed the characteristics of bone marrow on MRI and performed the imaging classification based on the MRI findings. Pearson’s chi-square test was used to compare the differences of AML histological classifications in different MRI types, and to analyze the correlation between the two classifications. Kappa test was used to analyze the consistency of the rating results of the two doctors. Result The MRI mainly showed that the signal of the skull bone marrow was decreased on T₁WI. After admission of contrast medium, bone marrow was significantly enhanced along the tables of the skull. The MRI manifestations of skull bone marrow were divided into 5 types. Type a, 5 cases; type b, 7 cases; type c, 7 case; type d, 9 cases; and type e, 2 cases. The MRI classifications between different histological types of AML were statistically different (P<0.05). Among them, leukemia M2 accounted for a high proportion in type a (60.00%), leukemia M4 accounted for a high proportion in type b (57.14%), and leukemia M2 accounted for a high proportion in type d (77.78%). The MRI classification of skull bone marrow was moderately correlated with histological classification of AML (Cramer’s V=0.490, P=0.045). The two doctors had good consistency in interpreting the MRI types of skull bone marrow (κ≥0.75) in M1, M2, M3, M5, and M6. The consistency in interpreting the MRI types of skull bone marrow in M4 was moderate (0.4≤κ<0.75). Conclusion The MRI classification of skull bone marrow can be used as an effective auxiliary method to evaluate histological classification of AML, and can provide a basis for clinical treatment decision-making.