以新式術中語言測驗精準保留術後之語言功能

Abstract

背景：腦腫瘤清醒開顱手術結合「術中電刺激功能定位」及「連續功能監測」，主要的目的是定位及監測語言功能，並且在腦腫瘤完全切除及保留術後語言功能間尋求平衡。「術中電刺激功能定位」是在以電刺激抑制局部腦功能，同時讓清醒的病患執行簡單的術中語言測驗。當電流抑制某一腦區而病患無法正確執行語言測驗時，則該腦區被視為語言功能區，在切除腦腫瘤時需要避開。而「術中連續功能監測」則為讓病患持續執行語言測驗，以監測語言功能是否因腫瘤切除過程影響而導致功能下降。術中語言測驗的種類非常多，它們因為手術環境的特殊限制，必須能被簡單執行。然而，目前沒有足夠的文獻證明，簡單的術中測驗是否能保留複雜的語言功能，而眾多測驗中，哪一種又最能保留術後的語言功能；此外，也沒有明確的定義闡明，在連續功能監測中，當病患的語言功能下降時，哪一個時機點手術應該停止，以達到最大的腦腫瘤切除率而且不影響術後的語言功能。所以，在第一篇研究中，我們提出一個簡單，但更全面的術中語言測驗，並對病患進行詳細的術後語言功能評估。我們假設這種術中測驗可以保留術後語言功能；第二篇研究中，我們定量病患術中測驗的表現，以數據化的方式監測語言功能，並分析語言功能的數值與術後的語言功能的關係。我們假設這種監測方式可以找到適當的切點數值，以預測及保留術後的語言功能。方法：我們以修改後的短版「波士頓失語症測驗」作為術前、術後的語言功能評估工具。在第一個研究中，我們提出一個更全面的術中語言測驗，它修改至短版「波士頓失語症測驗」的測驗圖卡。它包含物體、動作、顏色、數字、形狀、英文字母及中文單字的命名測驗。並以折線圖及成對樣本t檢定，來分析術後語言功能在各個細項測驗上的變化。在第二個研究中，我們以圖片命名(DO 80)及更複雜的語義關聯測驗(金字塔棕梠樹測驗，Pyramids and Palm Trees test, PPTT)當作術中連續功能監測的語言測驗，並以正確回答比率(正確率)量化術中語言功能的表現。以皮爾生相關係數及接收者操作特徵曲線(receiver operating characteristic, ROC curve)，來分析腫瘤切除末期之術中測驗的正確率與術後語言功能之關係。結果：在第一個研究中，我們提出的術中測驗可以保留病患在各個細項測驗的術後語言功能，而概念理解、詞義辨識及詞句重複這三個細項最能看出術後語言功能變化。在第二個研究中，發現術中圖片命名及語義關聯(金字塔棕梠樹)測驗的表現，都與術後一周的語言功能相關(p < 0.05)，而較複雜的「金字塔棕梠樹測驗」有較高的相關性。逐步線性回歸分析發現「金字塔棕梠樹測驗」是預測術後一周，各個語言細項測驗的唯一預測因子。當「金字塔棕梠樹測驗」的正確率下降到0.75(敏感度：80%；特異度：100%)時，術後的語言功能仍能保留。結論：我們所提出的更全面的樹中命名測驗能保留各個細項上的語言功能，而修改後的短版「波士頓失語症測驗」可以測出術後的語言功能變化。在預測術後語言功能上，「金字塔棕梠樹測驗」是個好的術中連續功能監測工具。而當它的正確率下降至0.75時，腫瘤切除應該停止，以保留術後一周的語言功能。 因此，我們建議以較全面及複雜的術中語言測驗，作為「術中電刺激功能定位」及「連續功能監測」，並以數據化的方式來監測術中語言功能。

Objective: Awake craniotomy combined with intraoperative stimulation mapping (ISM) and monitoring pursues the balance between extensive tumor resection and preserving postoperative language function. During ISM, the awake patient performs intraoperative linguistic tasks while suppressed-electrical stimulation is applying on the regional brain. However, few studies have evaluated the efficacy of intraoperative linguistic testing based on language outcomes and the dilemma that exists in patients whose tumor resection was restricted due to signs of linguistic disturbance observed during awake brain surgery. Therefore, in our first series study, we devised a novel intraoperative task and hypothesized that it can be easily applied in ISM and can preserve the language function of patients after surgery. In our second series study, we quantified the change in postoperative language function by comparing it with the intraoperative linguistic performance to understand the degree to which the recovery of language impairment caused by tumor resection can be achieved through spontaneous neuroplasticity. Methods: We used the modified short form of the Boston Diagnostic Aphasia Examination (sfBDAE) to assess preoperative and postoperative language functions. In the first series study, our novel and comprehensive naming task, which was modified using the stimulus sets of the sfBDAE, included objects, actions, colors, numbers, English letters, Chinese words, and shapes, and was employed in ISM. We analyzed the trend of the changed language function in line charts and compared the change in linguistic measurements by using a paired t test. In the second series study, a visual object naming test called the Dénomination d’objet 80 (DO 80) as well as a semantic-association test called the Pyramids and Palm Trees Test (PPTT) were used for intraoperative linguistic testing. The DO 80 and PTT were performed alternatively during the subcortical functional monitoring of tumor resection. Pearson’s correlation and multiple linear regression were used to analyze the correlation between the accuracy rates of the intraoperative linguistic tasks in the terminal stage of tumor resection and the scores of the postoperative sfBDAE subtests. A Receiver Operating Characteristic (ROC) analysis was used to define the cut point of the accuracy rates to predict postoperative language deficits. Results: In the first series study, the patient’s language function was either preserved after surgery or exhibited significant improvement over time. The trend of functional recovery varied in the complex ideational material, word discrimination, and repeating phrases subtests. In the second series study, both the intraoperative DO 80 and PPTT revealed a significant correlation with the postoperative sfBDAE domain scores (p < 0.05), with a higher correlation observed in the PPTT. A linear regression model demonstrated that only the PPTT could be included to predict the postoperative sfBDAE domain scores, with the explanatory power ranging from 0.51 to 0.89 (all p < 0.01). The ROC analysis demonstrated that the cutoff value for the accuracy rate in the PPTT was 0.75, yielding a sensitivity of 80% and specificity of 100%. Conclusion: The modified sfBDAE can indicate preoperative variations in language function, and the novel and comprehensive naming task used in this study can be applied in ISM and can preserve or improve linguistic outcomes. In addition, the PPTT is a feasible tool for intraoperative linguistic evaluations to predict postoperative language outcomes. Tumor resection should be limited when the accuracy rate of the PPTT approaches the cutoff value of 0.75. We therefore suggest using more comprehensive intraoperative tasks in ISM and intraoperative monitoring as well as the numerical accuracy rate proposed in this study for intraoperative linguistic monitoring.