Establishment of a Differential Diagnosis between benign and malignant Gastrointestinal Submucosal Tumors

1999年至2001年，在超声胃镜引导下细针吸取肿瘤42例，其中黏膜下肿瘤25例，胃肠外肿瘤17例。在黏膜下肿瘤的诊断中，细针吸取细胞学(FNAC)的年成功率依次为57.1%(8/14)、83.3%(5/6)和100%(5/5)，而细针吸取细胞学(FNAB)的可信度分别为0%(0/14)、50%(3/6)和100%(5/5)。因此，由于针和其他装置以及穿刺术的改进，组织标本的制备在技术上已经变得可行，并具有相当的可靠性。至于胃肠道外肿瘤，在研究的早期阶段，FNAC和FNAB的样本都获得了相对有利的结果。标本取自8例粘膜下肿瘤，其中7例疑似胃肠道st…。更多的是胃肠道肿瘤(GIST)，1例是异位胰腺。我们对7例疑似GIST进行了免疫组织化学染色，并将其定性为：平滑肌型(4例)、神经型(3例)、混合型(0例)和未定型(1例)。在7例中，6例(85.7%)的术前诊断与手术标本最终诊断的亚型一致。由于染色上的微小差异，在术后调查中，只有一例最初被归类为未定向型的病例被诊断为平滑肌型。这表明，使用我们的术前诊断程序对粘膜下肿瘤进行亚型分类是可能的。根据临床病程和手术结果，共25例粘膜下肿瘤分为恶性4例(含低度恶性)和良性21例。FNAB细胞异型性与FNAG细胞异型性联合诊断的准确率由单纯观察FNAB有丝分裂像诊断的60%提高到80%。这些诊断程序显示出很高的准确性。然而，考虑到这项研究中的其他因素，如目标人群中良性病例的高比例和分析的低敏感性(50%)，我们得出结论，有丝分裂特征以及细胞异型性不能完成术前良、恶性肿瘤的鉴别诊断。在整个研究过程中，我们已经建立了可靠的采样程序，目前证明可以用Ki67标记标本，Ki67被认为是显示粘膜下肿瘤的增殖能力。为了探讨Ki67标记指数在术前标本和手术标本中的相关性，以及评价其增殖能力能否作为术前粘膜下良、恶性肿瘤的预测指标，还将进行进一步的研究。较少

Endoscopic ultrasonography-guided fine needle aspiration was performed in 42 cases, 25 of which were submucosal tumors and 17 of which were extragastrointestinal tumor between 1999 and 2001. In the diagnosis of submucosal tumors the annual percentages of successful sampling for fine needle aspiration cytology (FNAC) were 57.1% (8/14), 83.3% (5/6) and 100% (5/5) in chronological order, and specimens for fine needle aspiration biopsy (FNAB) were obtained with reliability of 0% (0/14), 50% (3/6) and 100% (5/5) each year. Thus, the preparation of histologic specimens has technically become feasible with considerable reliability due to improvements in needles and other devices as well as in puncture techniques. Regarding extragastrointestinal tumors, both samples for FNAC and FNAB were obtained with comparatively favorable outcomes from the earliest stages of the study. Specimens were obtained from 8 cases with submucosal tumors, including 7 with a suspected diagnosis of gastrointestinal st … More romal tumors (GIST) and one with ectopic pancreas. We conducted immunostaining in 7 suspected cases of GIST and consequently characterized them as follows: smooth muscle type (4), neural type (3), combined type (0), and uncommitted type (1). In 6 (85.7%) of 7 cases, the preoperative diagnosis was consistent with subcategories derived from the final diagnosis on surgical specimens. Because of minor differences in staining, only one case that was initially classified into the uncommitted type was diagnosed as a smooth muscle type in the postoperative investigation. This suggests that it is possible to subcategorize submucosal tumors by using our preoperative diagnostic procedures. Based on clinical courses and surgical outcomes, a total of 25 submucosal tumor cases were differentiated into 4 malignant (including low-grade malignancy) and 21 benign cases. The differential accuracy, which was 60% when diagnosed simply by observing mitotic figures in FNAB samples, was improved to 80% when cellular atypism in FNAB and FNAG samples was concomitantly evaluated in the diagnosis. These diagnostic procedure demonstrated high accuracy. However, considering other factors in this study such as a high percentage of benign cases in the target population and a low (50%) sensitivity for analyses, we conclude that mitotic figures as well as cellular atypism will not complete a preoperative differential diagnosis between benign and malignant tumors. We have established reliable sampling procedure throughout the study and currently demonstrate that the specimens can be labeled with ki67, which is considered to indicate the proliferative capacity of submucosal tumors. Further studies will be conducted in order to investigate correlations between ki67 labeling index in a preoperative specimen and that measured in a surgical specimen, and to evaluate if the proliferative capacity can be a predictive indicator of benign or malignant submucosal tumor during the preoperative period. Less