DCE-MRI定量参数全域直方图鉴别诊断乳腺导管原位癌与乳腺导管原位癌伴微浸润

目的 探讨动态增强MRI（DCE-MRI）定量参数全域直方图鉴别诊断乳腺导管原位癌（DCIS）与 DCIS伴微浸润（DCIS-Mi）的价值。方法 收集41例DCIS-Mi和37例DCIS患者，术前行DCE-MRI，计算容积转运常数（K^trans）、速率常数（K_ep）、血浆容积分数（V_p），并记录其全域直方图参数，包括平均值、中位数及10%、25%、50%、75%、90%分位数（P₁₀、P₂₅、P₅₀、P₇₅、P₉₀）。比较DCIS-Mi和DCIS组间各参数差异，并分析最佳参数的诊断效能。结果 DCIS-Mi组K^trans参数中的平均值及各百分位数，K_ep参数中的平均值、中位数、P₅₀、P₇₅、P₉₀，V_p参数中的平均值、中位数、P₂₅、P₅₀、P₇₅均高于DCIS组（P均<0.05）。Logistic回归结果显示K^trans平均值、K^trans-P₂₅、K^trans-P₅₀、K^trans-P₇₅、K^trans-P₉₀、K_ep-P₅₀、K_ep-P₇₅、K_ep-P₉₀、V_p平均值、V_p-P₇₅为诊断DCIS-Mi的最佳变量，且其诊断DCIS-Mi的ROC曲线AUC均>0.80。联合变量Logistic回归模型具有最高的AUC（0.968），其临界值、敏感度和特异度分别为2.152、0.962和0.947。结论 DCE-MRI全域直方图分析能量化反映肿瘤特征，准确诊断及鉴别诊断乳腺DCIS与DCIS-Mi。

DCE-MRI quantitative parameter global histogram in differential diagnosis of breast ductal carcinoma in situ and breast ductal carcinoma in situ with microinvasion

Objective To explore the value of dynamic enhanced MRI (DCE-MRI) quantitative parameter global histogram in differential diagnosis of breast ductal carcinoma in situ (DCIS) and DCIS with micro-invasion (DCIS-Mi). Methods Totally 41 patients with DCIS-Mi and 37 patients with DCIS underwent DCE-MRI before operation. The volume transport constant (K^trans), rate constant (K_ep) and plasma volume fraction (V_p) were calculated, and the parameters of the global histogram were recorded, including mean, median and the quantiles of 10%, 25%, 50%, 75% and 90% (P₁₀, P₂₅, P₅₀, P₇₅ and P₉₀). The parameters were compared between DCIS-Mi and DCIS patients, and the diagnostic values were evaluated. Results The mean and percentiles of K^trans parameters, the mean, median, P₅₀, P₇₅ and P₉₀ of K_ep parameters, the mean, median, P₂₅, P₅₀ and P₇₅ of V_p parameters of DCIS-Mi patients were all higher than those of DCIS patients (all P<0.05). Logistic regression results showed that K^trans mean, K^trans-P₂₅, K^trans-P₅₀, K^trans-P₇₅, K^trans-P₉₀, K_ep-P₅₀, K_ep-P₇₅, K_ep-P₉₀, V_p mean, V_p-P₇₅ were the best variables for diagnosis of DCIS-Mi, and AUC>0.80. The joint variable Logistic regression model had the highest AUC (0.968), and its cut-off value, sensitivity and specificity was 2.152, 0.962 and 0.947, respectively. Conclusion DCE-MRI global histogram analysis can quantitatively reflect tumor characteristics, accurately differentiate breast DCIS and DCIS-Mi.