Can Intraoperative Specimen Radiograph Predict Resection Margin Status for Radioguided Occult Lesion Localisation Lumpectomy for Ductal Carcinoma In Situ Presenting with Microcalcifications?
BST Leung, AYH Wan, AKY Au, SSW Lo, WWC Wong, JLS Khoo
Hong Kong J Radiol 2015;18:11-21
DOI: 10.12809/hkjr1414265
Objective: To determine the accuracy of specimen radiograph in predicting ductal carcinoma in situ resection margin status during radioguided occult lesion localisation.
Methods: Retrospective review of cases of stereotactic radioguided occult lesion localisation for radiologically indeterminate-to-highly suspicious microcalcifications from September 2002 to May 2014 in a regional hospital was conducted. All patients diagnosed with ductal carcinoma in situ histopathologically with a specimen radiograph for review were included. An anteroposterior specimen radiograph was taken for each lumpectomy specimen. Retrospective assessments of the radiological margin, defined as the shortest distance from the outermost microcalcification to the specimen edge, in the superior, medial, inferior, and lateral borders were made. Results were correlated with histopathology findings in each border, where resection margin status was defined as positive (≤2 mm), close (2.1-5.0 mm), or clear (>5 mm). Results were analysed by Mann-Whitney U test and receiver operating characteristic curve.
Results: A total of 24 radioguided occult lesion localisation procedures revealing ductal carcinoma in situ in 23 patients were included. Among the 96 borders assessed, 12 and five had positive and close margins, respectively. Significantly smaller radiological margins were seen in borders with positive pathological margins (range, 0-17.5 mm; mean, 8.7 mm) than in those with close/clear pathological margins (range, 4.8-45.8 mm; mean, 20.1 mm; p < 0.001); and in borders with positive/close pathological margins (range, 0-17.5 mm; mean, 9.1 mm) than in those with clear pathological margins (range, 5.1-45.8 mm; mean, 20.7 mm; p < 0.001). Receiver operating characteristic curves were plotted from these results. The areas under the curve were 0.87 (95% confidence interval, 0.78-0.96) for positive margins and 0.89 (95% confidence interval, 0.82-0.96) for positive/close margins, indicating that radiological margin possessed good discriminating power for predicting resection margin status. From the receiver operating characteristic curves, a 15-mm radiological margin had the highest combination of sensitivity and specificity for predicting a positive margin (91.7% and 75.0%, respectively) and positive/close margin (94.1% and 73.4%, respectively). A 5-mm or 10-mm radiological margin resulted in higher specificity, while a 20-mm radiological margin had higher sensitivity.
Conclusion: Radiological margin assessment on specimen radiograph during radioguided occult lesion localisation showed high accuracy in predicting resection margin status in ductal carcinoma in situ presenting with microcalcifications.
中文摘要
術中標本X光片是否可預測因微鈣化而進行放射導向的隱匿性病灶定位乳房腫瘤切除術乳管原位癌病例的切緣狀態？
梁肇庭、尹宇瀚、區嘉殷、盧成瑋、黃慧中、邱麗珊
目的：在放射導向的隱匿性病灶定位過程中，研究標本X線片預測乳管原位癌環周切緣狀況的準確度。
方法：回顧2002年9月至2014年5月期間，因乳腺X線片上性質待定或高度可疑的微鈣化灶而接受立體定向放射導向隱匿性病灶定位術的病例。所有組織學上被診斷為乳管原位癌且有標本X線片的病例均被列入研究範圍。每個乳房腫瘤切除術標本都配有一張前後位X線片。放射學邊緣被定義為從最外層微鈣化至標本邊緣的最短距離，回顧性評估上方、內側、下方和外側幾個方向的邊緣。每個方向邊緣狀況與病理學結果作相關比較，病理上少於2毫米的切緣被介定為切緣陽性，2.1至5.0毫米的被介定為切緣近，多於5毫米的則為切緣清晰。最後利用Mann-Whitney U檢驗和ROC曲線分析結果。
結果：共有24項放射導向的隱匿性病灶定位術納入分析，其中確診出23名乳腺原位癌患者。評估的96個邊緣中，12個為切緣陽性，另5個為切緣近。與切緣近／清晰的病例（平均20.1毫米，介乎4.8至45.8毫米）比較，切緣陽性的病例（平均8.7毫米，介乎0至17.5毫米）有較小放射邊緣（p < 0.001）;與切緣清晰的病例（平均20.7毫米，介乎5.1至45.8毫米）比較，邊緣陽性／近的病例（平均9.1毫米，介乎0至17.5毫米）有較小放射邊緣（p < 0.001）。根據上述結果繪製出ROC曲線。切緣陽性病例的ROC曲線下面積為0.87（95%置信區間，0.78-0.96），而切緣陽性／近病例的曲線下面積為0.89（95%置信區間，0.82-0.96），表明放射學邊緣能有效預測切緣狀態。ROC曲線分析結果顯示15毫米的放射性切緣對於邊緣陽性和邊緣陽性／近病例有最高的敏感性和特異性組合，前者為91.7%和75.0%，後者為94.1%和73.4%。5毫米或10毫米的放射切緣有較高特異性，而20毫米的放射切緣則有更高敏感性。
結論：放射導向的隱匿性病灶定位乳腺腫瘤切除術中標本X線片上放射學邊緣評估可準確預測含微鈣化灶的乳管原位癌病例切緣狀態。