(Discovery STE, GE Medical Systems, Milwaukee, WI, USA) 所生產的儀器。檢查前病患禁食至少 4 小時，血糖不超過 120mg/dl。約 10 毫居里(mCi)或 370 百萬貝克(MBq)的氟 18-去氧葡萄糖(FDG)經靜脈注射後，受檢者被要求靜躺休息 45
14
分鐘，以便 FDG 在腫瘤內能夠充分聚積，在正常組織中能 充分排除，並經由腎臟、膀胱排泄。當 45 分鐘後，開始進 行檢查，一般自頭部掃瞄至大腿上三分之一約需一小時。
當 70 分鐘的時候，我們將再施作一次全身性正子電腦斷層 掃描。當所有影像重組後，就會由二位核子醫學科的醫師 進行判讀 (圖三)。
SUV 的定義是"mean selected region of
activity/injected dose/body weight"。FDG uptake 的 區域指的是放射性物質吸收較高的地方，一般上是針對超 過週邊背景值或組織的區域如血管，食道，淋巴結及其他 在電腦斷層上可辨識的器官(圖四)。
统計與分析:
我們參考過去的文獻記載發現，目前大多數的學者判 定腫瘤是否為惡性的 SUV 值大多介於 2.5-3.5 之間。最終 我們决定採用 Hu Q et al.於 2009 年所發表的文章，分別 以最大標準攝取值 SUVmax≥ 2.5 及 Retention index (RI)
≥10%作為本次實驗的參考值 (cutoff value)。每位病患在 術前都會接受兩次的全身正子電腦斷層掃描，因此就會取
15
得兩組不同時間內的 SUVmax 值；Early SUVmax 及 Delay SUVmax。Early SUVmax 指的是 FDG 注射到人體 45 分鐘後的 最大標準攝取值而 delay SUVmax 為 70 分鐘後的讀值;而 RI 則是 delay SUVmax 值減去 early SUVmax 值再除以 early SUVmax 值的百分比，其代表意義為判別細胞是否是惡性 (RI≥10%)或者只是發炎與感染疾病的可能性。將這 3 種讀 值收集後再組合成以下 4 組的診斷標準(Diagnostic criteria):
(1) SUVmax≥ 2.5 (2) RI ≥ 10%
(3) SUVmax 加上 RI ≥ 10%
(4) SUVmax 或 RI ≥ 10%擇一，並進行统計分析。
我們所使用的统計方法及軟體分別是 Fisher’s exact test 及 SPSS 12 版，同時採取 p 值< 0.05 為统計上有意義。
16
(表一)
17
(圖三) PET-CT scanner and procedure
(圖四) Example of an FDG PET/CT scan.
18
Appendix:
2.1 Patient Population
Twenty-six patients (all men; age range, 42 - 72 years old;
mean age, 60.4 years)who underwent preoperative FDG
PET-CT scan and subsequent surgical resection of esophageal cancer between October 2009 and April 2010 in China Medical University Hospital were retrospectively included in this study.
All patients histologically proved to have squamous cell
carcinoma and had received esophageal resection and regional lymph nodes dissection. Surgical pathology results were used to provide the final diagnosis with which the FDG PET-CT results were compared, including those with distant metastases. The visual interpretation of the FDG PET-CT and surgical pathology stage was classified according to the sixth edition of the AJCC Cancer Staging System. This study was approved by the Ethics Committee of the China Medical University Hospital
(DMR-99-IRB-010).
19
2.2 FDG PET-CT Imaging protocol, Interpretation and Calculation of Related Parameters
All patients were asked to fast for at least 4 hours before FDG PET-CT imaging. Imaging was performed with a PET-CT
scanner (Discovery STE, GE Medical Systems, Milwaukee, WI, USA). Whole - body FDG PET-CT images were acquired
approximately 45 minutes after intravenous injection of 370 MBq (10 mCi) of FDG. Delayed FDG PET-CT images were obtained approximately 70 minutes after FDG injection. PET emission images were acquired after CT scans at 2 minutes per field of view (FOV) in the 3 - dimensional acquisition mode.
The CT images were reconstructed onto a 512×512 matrix with a section thickness of 3.75 mm, then reconstructed onto a
128×128 matrix, and converted into 11-keV-equivalent attenuation factors for attenuation correction of the
corresponding PET emission images. The suspected tumoral FDG uptake was defined as focally increased radioactivity, greater than those in the surrounding background or blood pool, in the esophagus, lymph node or other recognizable
20
morphological lesional sites on the CT component of FDG PET-CT by visual interpretation. A semiquantitative parameter, standardized uptake value (SUV), was defined as “tracer activity in the tumor per unit mass divided by amount of injected
radioactivity per unit body mass”, and calculated from each region of interest with increased FDG radioactivity in the suspected tumoral region. The maximum standardized uptake value (SUVmax) of esophageal cancer and metastasis on early and delayed FDG PET-CT images were measured. The retention index (RI) based on the measured SUVmax was calculated as 100% × (delayed SUVmax－ early SUVmax) ÷ early SUVmax.
Early and delayed PET images were reviewed on the computer monitor in the trans-axial, coronal, and sagittal planes along with maximum-intensity-projection images. Two experienced nuclear medicine physicians in dependently evaluated FDG uptake both visually and semiquantitatively. The evaluating physicians were unaware of the clinical history and the PET images were compared with the corresponding CT images for accurate anatomic identification of the tumor. Any difference of
21
opinion was resolved by consensus.
2.3 Statistical analysis
We used 4 diagnostic criteria to evaluate the sensitivity, specificity and accuracy of dual-time FDG PET-CT in differentiating malignancy from a benign lesion among the primary tumor, loco-regional lymph nodes, and distant
metastases. The 4 criteria included (1) early SUVmax≧ 2.5 alone, (2) RI ≧ 10% alone, (3) combination of early SUVmax
≧ 2.5 and RI ≧ 10%, and (4) combination of early SUVmax
≧ 2.5 or RI ≧ 10%. Fisher’s exact test was applied to compare each difference in the sensitivity, specificity and
accuracy. SPSS software was used for the analysis. A p-value of less than 0.05 was considered statistically significant.
22
Chapter 3 Results
第三章 研究結果
在原發腫瘤部份，當 SUVmax≥ 2.5 或者 RI ≥ 10% 任 何一項讀值被選為判定是否為陽性時(腫瘤細胞)，其靈敏度 可提高到 96.2%，而且是统計上有意義的 (p<0.05) (表二)。
針對局部淋巴結的部份，當 SUVmax≥ 2.5 加上 RI ≥ 10%
時靈敏度會明顯的提昇到 70%，但是 p 值却只有 0.1181。至 於在其他不同組合的診斷準則方面也都無法有效提高對局 部淋巴結的偵測，所有的靈敏度，特異度及準確性均無法達 到統計學上的意義 (表三)。
最後在遠端轉移的部份，SUVmax≥ 2.5 無論是否有加 上 RI ≥10%，其靈敏度及特異度皆相同;分別是靈敏度 16.7%
而特異度則是 100%，均有達到统計學上意義(p<0.05)。至於 在準確性方面所有的診斷準大都介於 76.9-80.8%，無法達到 統計學上意義 (表四)。
23
(表二)
(表三)
(表四)
24
Appendix:
Results
3.1 Primary tumor
The sensitivity of FDG PET-CT in detecting the primary tumor site with combination of early SUVmax≧ 2.5 or RI ≧ 10 % was 96.2%. It was statistically significantly higher than the other 3 criteria. The p value of Fisher’s exact test was <0.0001 and it was statistically significant (Table 2).
3.2 Regional lymph nodes
The sensitivity of early SUVmax≧ 2.5 alone was 30.0%, but it increased to 70% when combination of early SUVmax≧ 2.5 or RI ≧ 10% was used. However, the p value was only 0.1181 and hence not significant. For the specificity, the result was reversed. When using early SUVmax≧2.5 alone, it was 93.8%
but it decreased dramatically to 56.3% when combination of early SUVmax≧ 2.5 or RI ≧ 10% was used. Similarly, the p value was not significant (p = 0.0756) (Table 3). With regard to accuracy, there was no significant correlation among the four
25
diagnostic criteria. (p>0.05 ).
3.3 Distant Metastasis
When using combination of early SUVmax≧ 2.5 and RI ≧ 10%, the sensitivity and specificity were 16.7% and 100%, respectively. The same result was found when early SUVmax≧
2.5 was used alone. However, when RI ≧ 10% was used alone, the sensitivity increased dramatically to 50% but the specificity dropped to 85%. Furthermore, a similar poor result was found when combination of early SUVmax or RI ≧ 10%, was used (Table 4). A significantly higher specificity was found when using combination of early SUVmax≧ 2.5 and RI ≧ 10% or using early SUVmax alone than using the other 2 parameters (p
= 0.0058 by using Fisher’s exact test). The sensitivity, however, was somewhat different (p = 0.4906). For the accuracies, the detection rate was between 76.9% and 80.8%, which did not reach statistical significance (p > 0.05).
26
27
斷，相對於使用雙相正子掃描(dual time PET-CT)在食道癌 的診斷的文章則只有一篇報導。除此以外還有一個更重要的 議題要討論就是對於 SUVmax 值的標準值(cutoff value)，
要取决於那一個讀值，各學者專家到目前為此也還沒有一個 定論。
在本研究中我們參考了以往的報告，以 SUVmax> 2.5 來界定良性與惡性的標準值，同時加入了 Retention Index (RI)，也即是 70 分鐘和 45 分鐘 SUVmax 值的變化百分比，
作為另一個參考值以提高診斷食道原發腫瘤，淋巴結及遠端 轉移的靈敏度，特異度及準確性。根據本研究的最後统計，
如果單獨使用 early SUVmax> 2.5 為診斷準則的話，其靈敏 度可達到 88.5%，這和其他報導過的文獻很接近。當 RI > 10%
28
70% 和 56.3% ，這和以往的 PET 報告相差不遠。檢討分析後 我們認為可能和淋巴結太小或者只有顯微的侵犯，所以葡萄 糖的吸收會比較低。同時如果局部的淋巴結與原發腫瘤太接 近或者患者的胸部呼吸太大都會影响到判讀。除此以外，如 果在手術中所切除的淋巴結數不夠或被忽略掉，淋巴結和原 腫瘤術後無法完全分陶檢查，食道複雜的淋巴系統及太微小 的淋巴結都會影响到最後的判讀結果。總括以上各項因素，
我們認為雙相正子攝影在偵測食道癌局部淋巴結轉移的角 色尚未有定論。
最後針對遠端轉移部份，我們發現如果使用雙相正子 電腦斷層攝影檢查，其特異度會高達 85%-100%，這和現有的 報導比較並不遜色。相反如果只單獨使用 SUVmax 為單一判 讀準則，其靈敏度會比同時/或加上 RI 值來得低。最後我們 的結論是在診斷食道癌的遠端轉移時建議可以把 RI 值加上 作為一項參考指標，以提高其準確度。
29
Appendix:
FDG PET-CT is widely used with cancer patients. Its role as a non –invasive imaging modality has been widely
investigated but the exact SUVmax cutoff value for esophageal cancer remains controversial. To accurately distinguish
malignant from benign lesion is challenging because FDG is taken up not only by tumor cells but by inflammatory cells as well.
Although the potential of dual-time FDG PET in evaluating various cancers has been reported, the diagnostic value of this technique for esophageal cancer has not been fully investigated. To our knowledge, only one report has
demonstrated the potential of dual–time FDG PET in evaluating the loco-regional lymph nodes in esophageal cancer.
Furthermore, there have been few reports on the sensitivity, specificity and accuracy of dual–time FDG PET-CT in evaluation of primary tumor and distant metastasis.
In our study, we assessed whether dual-time FDG
30
PET-CT would have more value than conventional FDG PET-CT imaging in differentiating between malignant and benign esophageal lesions. An arbitrary cutoff as SUVmax of 2.5 has been used in various malignancies, most in lung cancers.
Nevertheless, the frequent effective discriminability with this arbitrary value among those previous studies results infrequent citations in newly conducted studies for various tumors,
especially in the initial applications. Therefore, the current preliminary study assumed that a SUVmax of2.5 might be a potential useful cutoff for differentiation of esophageal cancer in addition to the visual interpretation. Various studies have shown that the FDG uptake in inflammatory lesions normally reached a peak at approximately 60minutes after injection. However, in some malignant lesions the uptake continuously
increased for 120 - 180 minutes. In order to increase the
detection rate, we proposed using the percentage of change in the lesion between early SUVmax and delayed SUVmax as an alternative in the diagnosis of esophageal cancer (RI > 10%).
We also hypothesized that the sensitivity, specificity and
31
accuracy would increase if the SUVmax and retention index were used simultaneously. Statistically, when using early
SUVmax≧ 2.5 alone as the diagnostic criterion, the sensitivity for the primary tumor was approximately 88.5%,which is
consistent with the finding of other recently published reports.
When combination of early SUVmax≧ 2.5 or RI ≧ 10% was used as the diagnostic criteria for imaging reading, the
sensitivity reached 96.2% and that was significantly higher than with the other three criteria. Therefore, we conclude that
combination of early SUVmax≧ 2.5or RI ≧ 10% is a reliable tool in detecting the primary site of esophageal carcinoma.
For loco-regional lymph node involvement, many studies reported that sensitivity varied, ranging from 22% and 72%. In a retrospective study, Hsu et al. reported sensitivities and
specificities rates for regional lymph node involvement of, 57.1% and 83.3%, respectively. In comparison, our data
demonstrate that the combination of early SUVmax≧ 2.5 or RI
≧ 10% produced an sensitivity of 70% and specificity of 56.3%. When compared with the usual FDG PET results, these
32
results were unsatisfactory. This could be related to the
relatively low glucose utilization of the small lymph node or a limited microscopic spread. Tracer uptake in physiologic
structures at the thoracic cage, motion or high FDG uptake in the adjacent primary tumor can lead to an underestimate of the FDG uptake of the regional lymph nodes. Furthermore, the size and morphology of the lymph nodes, a complicated lymphatic drainage network, an uneven margin between the tumor and the nodal extension, inadequate surgery and failure to detect
peri-tumoral nodes during resection may also influence the final histology finding. The number of lymph nodes examined must also be taken into account when assessing the results. In
conclusion, we believe that the ability of dual-time FDG PET-CT to detect loco-regional lymph node metastasis using different diagnostic criteria remains to be demonstrated.
In our study, high specificities were achieved in the detection of distant metastasis of esophageal carcinoma using dual-time FDG PET-CT, in a range from85% to 100%. Our results are in agreement with those of several FDG PET studies.
33
In contrast, the low sensitivities of early SUVmax≧ 2.5 alone as the single diagnostic criteria in recognizing metastatic lesions might be due to variable volumes of the metastatic lesions.
When RI is applied, significantly improvement was noted.
Therefore, based on our study, we recommend combination early SUVmax≧ 2.5 and/or RI ≧ 10% in order to increase the accuracy. One limitation of this study was the small number of cases. This may have affected the statistical calculations.
Therefore, studies with a larger number of patients should be conducted to determine the appropriate cutoff values of
SUVmax and RI for esophageal cancer. Besides, short interval of follow up may also cause some false negative results.
However, on the basis of the data reported, to some extent, the dual-time point FDG PET-CT has more value than standard PET imaging for detecting esophageal cancer. Therefore, we
recommend that FDG PET-CT be considered in routine examination prior to the treatment of esophageal cancer in orderto guide optimal clinical management for possible distal metastases to avoid unnecessary operation.
34
Chapter 5 第五章