Patterns of failure and prognostic factor analyses in locally advanced cervical cancer patients staged by magnetic resonance imaging and treated with curative intent
透過磁振造影分期並接受根治性治療的局部晚期子宮頸癌患者的失敗模式和預後因素分析

Earlier we had shown that tumor volume and corpus invasion were important prognostic factors in cervical cancer and that corpus invasion was associated with nodal metastases. In view of these findings, we wanted to examine the factors associated with the patterns of relapse in cervical cancer patients who were staged by magnetic resonance imaging (MRI) and treated with curative intent. This was a retrospective study of locoregionally advanced cervical cancer patients treated with curative intent. All patients had examination under anesthesia and pretreatment MRI. Potential prognostics examined were FIGO stage, clinical diameter, histology, corpus invasion, tumor volume, and age. Outcome measures examined were times to failure, local failure, nodal failure, and distant failure. There were 249 eligible patients. The median age of the patients was 58 years, had squamous histology, and of tumors exhibited corpus invasion. Median tumor volume was (range 1-628). The mean follow-up was 4.5 years. Eighty-five patients had relapsed and 89 died ( 70 following failure and 19 otherwise). At 5 years, for all patients, the failure-free rate was , the local failure-free rate , the nodal failure-free rate , and the distant failure-free rate . Corpus invasion, tumor volume, and age were all highly significantly and independently related to risk of failure at local, nodal, and distant (except tumor volume) sites. In the presence of these factors, clinical tumor diameter and FIGO stage were not significantly related to risk of any type of failure.
早些時候，我們曾發現腫瘤體積和癌腫侵犯是子宮頸癌的重要預後因素，而且癌腫侵犯與結節轉移有關。鑑於這些發現，我們希望研究與透過磁振造影（MRI）分期並接受根治性治療的子宮頸癌患者復發模式相關的因素。這是一項針對接受根治性治療的局部區域晚期子宮頸癌患者的回顧性研究。所有患者都在麻醉狀態下接受了檢查和預處理磁振造影。潛在預後指標包括FIGO分期、臨床直徑、組織學、癌腫侵犯、腫瘤體積和年齡。檢查的結果指標包括失敗時間、局部失敗、結節失敗和遠處失敗。符合條件的患者共有 249 人。患者的中位年齡為58歲， ，組織學為鱗狀組織， ，腫瘤表現為癌症病灶侵犯。腫瘤體積中位數為 （1-628）。平均追蹤時間為 4.5 年。 85例患者復發，89例死亡（70例因治療失敗，19例因其他原因）。 5年後，所有患者的無失敗率為 ，局部無失敗率為 ，結節無失敗率為 ，遠處無失敗率為 。癌灶侵犯、腫瘤體積和年齡都與局部、結節和遠端（腫瘤體積除外）的失敗風險有顯著的獨立相關性。在有這些因素的情況下，臨床腫瘤直徑和FIGO分期與任何類型的失敗風險均無明顯關係。

KEYWORDS: cervical neoplasm, corpus invasion, FIGO, patterns of failure, prognostic factors.
關鍵字：子宮頸腫瘤、子宮體受侵、FIGO、失敗模式、預後因素。

Following the use of cisplatin-based concurrent chemoradiotherapy (CCRT), the survival for patients with locally advanced cervical cancer has improved . Cumulative rates of disease-free survival at 5 years were among group of patients with bulky stage IB-IVA cervical cancer treated with CCRT and among similar patients treated with radiotherapy alone . The rates of both distant metastases and locoregional recurrences were significantly higher among patients
使用以順鉑為基礎的同步化學放射治療（CCRT）後，局部晚期子宮頸癌患者的存活率有所提高 。在接受CCRT治療的IB-IVA期子宮頸癌患者中，5年無疾病存活累積率為 ，而在接受單純放射治療的類似患者中，5年無疾病存活累積率為 。遠端轉移率 和局部復發率 都明顯高於單純放療的患者。

treated with radiotherapy alone. Advances in the treatment of epithelial tumors have been made slowly. The identification of pretreatment prognostic factors has helped in appropriate treatment selection for various groups of cancer patients, which in turn has resulted in improved locoregional control and reduction in treatment-related morbidity. The study of patterns of failure in radically treated cancer patients greatly helps in understanding the clinical biology of cancer that provides clues for identifying further prognostic factors and assists in improving cancer treatment.
單用放射治療。上皮腫瘤的治療進展緩慢。確定治療前的預後因素有助於為各類癌症患者選擇適當的治療方法，從而改善局部控制並降低與治療相關的發生率。根治性癌症患者失敗模式的研究大大有助於了解癌症的臨床生物學特性，為進一步確定預後因素提供線索，並有助於改善癌症治療。

In an earlier paper, we showed, using multifactor analysis in 179 patients with locally advanced cervical cancer, investigated by magnetic resonance imaging (MRI) and treated with concurrent cisplatin and
在早期的一篇論文中，我們利用多因素分析方法，對 179 名局部晚期子宮頸癌患者進行了磁振造影（MRI）檢查，發現他們同時接受了順鉑和雌激素治療。
radiotherapy, that tumor volume and corpus invasion provided important prognostic information, for overall survival (OS) and failure-free survival (FFS), over and above that provided by FIGO stage, clinical diameter, histology, and age . Previously, we had shown that corpus invasive tumors were associated with increased incidence of nodal metastases as investigated by pretreatment positron emission tomography . The aim of this study was, therefore, to examine how these factors (from FIGO stage, clinical diameter, tumor volume, corpus invasion, histology, and age) were related to any particular patterns of failure in a retrospective data series of locally advanced cervical cancer patients treated with curative intent using CCRT. Such information would be useful in improving current treatment of advanced cervical cancer, for example, by incorporating hypoxic sensitizers in the treatment of those patients who might predominantly fail locally or adding systemic chemotherapy in those who may fail predominantly at the distant sites. This information may also be useful in designing future clinical studies involving novel treatment strategies and stratifying patients according to these prognostic factors.
放射治療，腫瘤體積和腫瘤蒂浸潤為總生存（OS）和無失敗生存（FFS）提供了重要的預後信息，超過了FIGO 分期、臨床直徑、組織學和年齡所提供的信息 。先前，我們透過預處理正子斷層掃描研究發現，冠狀浸潤性腫瘤與結節轉移發生率增加有關 。因此，本研究的目的是在使用CCRT 進行根治性治療的局部晚期子宮頸癌患者的回顧性資料系列中，研究這些因素（包括FIGO 分期、臨床直徑、腫瘤體積、瘤體浸潤、組織學和年齡）與任何特定失敗模式的關係。這些資訊將有助於改善目前的晚期子宮頸癌治療，例如，在治療可能主要在局部失敗的患者時加入缺氧增敏劑，或在治療可能主要在遠處失敗的患者時加入全身化療。這些資訊也有助於設計未來的臨床研究，包括新型治療策略和根據這些預後因素對患者進行分層。

Patients who were previously untreated and presented to Peter MacCallum Cancer Centre, between December 1995 and December 2003, inclusive, with carcinoma of the cervix for definitive treatment or postoperative radiotherapy following definitive surgery; had squamous cell carcinoma (SCC) or adenocarcinoma (including clear cell, endometrioid, mucinous, and serous) histology; had pretreatment clinical (FIGO) staging and MRI performed; had tumors with stages IB-IVA, inclusive; and were treated by surgery, surgery and postoperative radiotherapy, or radical radiotherapy were eligible for the study. All patients satisfying these criteria were included.
在1995 年12 月至2003 年12 月（含12 月）期間，曾因子宮頸癌到Peter MacCallum 癌症中心接受明確治療或明確手術後接受術後放射治療而未接受治療的患者；組織學為鱗狀細胞癌（SCC）或腺癌（包括透明細胞、子宮內膜樣癌、黏液腺癌和漿液性癌）；進行過治療前臨床（FIGO）分期和磁振造影；腫瘤分期為IB-IVA 期（包括IB-IVA 期）；已接受手術、手術和術後放射治療或根治性放射治療。所有符合上述標準的患者均納入研究範圍。

All patients were staged according to FIGO protocol. Ninety-one patients had PET scan and 47 patients had laparoscopic nodal staging prior to commencement of their curative treatment. In all, of the patients had lymph node metastases based on structural (computed tomography , , or surgical staging prior to commencing radiotherapy. None of the patients with para-aortic nodal disease were excluded from the analysis. The information on lymph node status obtained from PET and surgical staging was used in planning of treatment fields and radiation boost dose to the nodal sites.
所有患者都按照 FIGO 方案進行了分期。 91名患者在開始根治性治療前進行了正子電腦斷層掃描，47名患者進行了腹腔鏡結節分期。根據結構（電腦斷層掃描 、 或手術分期），共有 名患者在開始放射治療前進行了淋巴結轉移分期。主動脈旁結節病 的患者均未被排除在分析之外。從 PET 和手術分期中獲得的淋巴結狀態資訊被用於規劃治療區域和結節部位的放射增強劑量。

Patients with FIGO stage IB1 disease were treated by type II or type III hysterectomy and pelvic lymphadenectomy. In a few patients, the scheduled hysterectomy was abandoned in favor of radical chemoradiation when lymphadenectomy revealed positive nodes. Other node-positive patients following radical hysterectomy received whole pelvic adjuvant radiotherapy. High-risk node-negative patients received adjuvant central pelvic radiotherapy.
FIGO IB1 期患者接受 II 型或 III 型子宮切除術和骨盆腔淋巴結切除術。少數患者在淋巴結切除術發現結節陽性時放棄了原定的子宮切除手術，轉而接受根治性化療。其他結節陽性患者 在根治性子宮切除術後接受了全盆腔輔助放射治療。結節陰性的高危險群患者 ，接受中央骨盆輔助放射治療。

During the study period, our policy was to treat nodenegative patients with central pelvic radiotherapy using a four-field box technique to a dose of 45-50.4 Gy in 1.8 Gy fractions, using an upper border of the central pelvic field above the inferior margin of the sacroiliac joint in antero-posterior projection, lateral borders limited to the true pelvis, and inferior borders at the lower margins of the obturator foramina. In lateral projection, the anterior border was at the inner aspect of the lower margin of the pubic symphysis and the posterior border intersected the S2-S3 space. A margin around the vaginal vault was further boosted to a total dose of in 30 fractions. The policy for node-positive patients was similar. Radiation portals, however, consisted of a typical pelvic field, the upper border being L5-S1, or extendedfield radiotherapy (EFRT) based on the levels of nodal involvement. In addition, postoperative node-positive patients also received weekly cis-platinum during the course of their external beam radiotherapy (EBRT). Node-positive sites were boosted to a total dose of 50.4-54 Gy.
在研究期間，我們的治療方針是對結節陰性患者進行骨盆中央放射治療，採用四野盒技術，劑量為45-50.4 Gy，分次1.8 Gy，骨盆中央野的上邊界 ，前後投影時位於骶髂關節下緣上方，側邊界限於骨盆真皮，下邊界位於閉孔下緣。側投影時，前緣位於恥骨聯合下緣的內側，後緣與 S2-S3 間隙相交。陰道穹隆周圍的 邊界進一步提高到總劑量 ，共分 30 次進行。結節陽性患者的治療方法與此類似。不過，放射門戶包括典型的骨盆腔野（上邊界為L5-S1）或根據結節受累程度進行的延伸野放射治療（EFRT）。此外，術後結節陽性患者在接受體外放射治療（EBRT）期間每週仍需接受順鉑 。結節陽性部位的總劑量為50.4-54 Gy。

For patients with FIGO stage IB1, node-positive, and all stage IB2 disease and above, the policy was to treat by radical chemoradiation. Concurrent chemotherapy was similar to that used in node-positive postoperative cases described above and was only used during EBRT component of their treatment. Patients who received both EBRT and intracavitary brachytherapy (ICRT) had ICRT within 7-10 days of completion of
對於 FIGO IB1 期、結節陽性和所有 IB2 期及以上疾病的患者，則採用根治性化放射治療。同期化療與上述結節陽性術後病例的化療類似，僅在 EBRT 治療期間使用。同時接受 EBRT 和腔內近距離放射治療（ICRT）的患者在完成 EBRT 治療後 7-10 天內接受 ICRT 治療。

EBRT. EBRT consisted of either in 20 fractions to the whole pelvis or 45 Gy in 1.8 Gy fractions when EFRT was used. ICRT consisted of in five fractions using high-dose rate brachytherapy, given twice weekly, or 40-45 Gy in two fractions using low-dose rate brachytherapy. Parametrial or nodal boost, if required, consisted of 6-10 Gy in 2 Gy fractions given in between ICRT fractions using EBRT. Patients receiving EFRT were treated supine, whereas those receiving pelvic radiotherapy were treated prone on a belly board.
EBRT。 EBRT包括 ，20次分次照射整個骨盆；或使用EFRT時，45Gy，1.8Gy分次照射。 ICRT包括使用高劑量率近距離放射療法，每週兩次，每次5分段 ，或使用低劑量率近距離放射療法，每次2分段，每次40-45 Gy。如有需要，可在兩次 ICRT 治療之間使用 EBRT，以 2 Gy 的劑量分次給予 6-10 Gy 的副房或結節增強治療。接受 EFRT 治療的患者仰臥，而接受骨盆腔放射治療的患者則俯臥在肚板上。

All patients were recommended for the use of a vaginal cylinder and estrogen cream twice daily for the first 2-3 months following completion of radiotherapy and twice a week thereafter, indefinitely, to avoid vaginal stricture and narrowing. All patients were examined once at 4 weeks and thereafter every 3 months for 1 year, every 4 months in their second and third year, and 6 monthly in the fourth and fifth year following completion of radiotherapy. After 5 years, patients were reviewed once per year, indefinitely. Patients suspected of harboring residual disease at the primary site due to a persistent bulky cervix were investigated using cytology, CT, MRI, and/or PET, as appropriate, to exclude multiple sites of residual or recurrent disease. In the initial period of the study, patients with suspected residual disease in the cervix in the absence of any metastatic disease were treated by extrafascial hysterectomy. No routine radiologic investigations were carried out during the follow-up period in asymptomatic patients. Symptomatic patients were investigated thoroughly, and if the patients found to have salvageable disease, they were treated either by radiotherapy or by surgery, or by both. A follow-up form recording the disease status and late complications was completed prospectively at each visit for every patient. Toxicities were scored according to modified World Health Organization/Radiation Therapy Oncology Group (RTOG) criteria. Six categories of toxicities consisting of bladder, small and large bowel, vagina, skin, lymphedema, and "symptoms" were recorded. The heading of symptoms included a feeling of being unwell and pain in the absence of other toxicities.
建議所有患者在放射治療結束後的前 2-3 個月每天使用兩次陰道圓筒和雌激素乳霜，此後每週使用兩次，無限期使用，以避免陰道狹窄。所有患者在放療結束後4 週接受一次檢查，此後每3 個月接受一次檢查，為期1 年，第二年和第三年每4 個月接受一次檢查，第四年和第五年每月接受6 次檢查。 5 年後，患者每年接受一次複查，無期限。對因子宮頸持續腫大而懷疑原發部位有殘留疾病的患者，請酌情使用細胞學、CT、MRI 和/或 PET 進行檢查，以排除多個部位的殘留或復發疾病。在研究初期，對子宮頸疑似殘留疾病但無任何轉移性疾病的患者進行筋膜外子宮切除術。無症狀患者在追蹤期間不進行常規放射檢查。對有症狀的患者進行徹底檢查，如果發現有可挽救的疾病，則對其進行放射治療或手術治療，或同時進行兩種治療。每位患者每次就診時都要填寫一份追蹤表，記錄疾病狀況和後期併發症。毒性根據世界衛生組織/放射治療腫瘤學組（RTOG）的標準進行評分。毒性分為膀胱、小腸和大腸、陰道、皮膚、淋巴水腫和 "症狀 "六類。症狀 "包括在無其他毒性反應的情況下感到不適和疼痛。

Failure was defined as either persistent disease following radiotherapy or recurrence of disease. The date of failure was taken as the date of sign of any of these types of failure. For persistent disease, the date of treatment failure was taken as the date of the response assessment at which persistent disease was observed.
放射治療失敗的定義是放射治療後疾病持續或疾病復發。治療失敗的日期為出現上述任何一種治療失敗跡象的日期。對於頑固性疾病，治療失敗的日期為觀察到頑固性疾病的反應評估日期。

Sites of failure were recorded as (1) local (residual or recurrent disease at cervix or uterus); (2) nodal: pelvic (parametrial below L5-S1, including vaginal failure [two patients]), abdominal (para-aortic above L5-S1), supraclavicular, and inguinal; and (3) distant.
失敗部位記錄為：(1) 局部（子宮頸或子宮殘留或復發疾病）；(2) 結節：骨盆腔（L5-S1 以下宮旁，包括陰道失敗[兩名患者]）、腹腔（L5-S1 以上腹主動脈旁）、鎖骨上和腹股溝；(3) 遠處。

Failures were analysed both as a first failure and as any failure. For first failure at multiple sites, the dominant site of failure was determined according to the hierarchy, in descending order: distant, inguinal, supraclavicular, abdominal, pelvic, and local.
失敗分析包括首次失敗和任何失敗。對於多個部位的首次失敗，主要的失敗部位根據層次結構確定，從高到低依序為：遠端、腹股溝、鎖骨上、腹部、骨盆和局部。

The start date for measuring survival, FFS, etc., was defined as follows: if the patient received radiotherapy only, the date radiotherapy started ( patients); otherwise if the date of surgery was after the date of presentation, the date of surgery ; and otherwise the date of presentation . Times to events were censored by the closeout date and failure to follow up. FFS was defined as the time to failure or death. Time to failure was defined as time to any type of failure and was censored by death without preceding failure. Times to local, nodal, and distant failure were defined as time to the failure type concerned whether it was a first or subsequent failure and was censored by death without failure. OS was defined as the time death from any cause.
測量存活期、FFS 等的起始日期定義如下：如果患者僅接受放射治療，則為放射治療開始日期（ ）；否則，如果手術日期在發病日期之後，則為手術日期 ；否則，為發病日期 。事件發生時間依結束日期及未能追蹤剔除。 FFS 定義為至失敗或死亡的時間。失敗時間定義為發生任何類型失敗的時間，並以未發生失敗前的死亡時間為剔除因素。局部失敗、結節失敗和遠處失敗的時間定義為出現相關失敗類型的時間，不論是首次失敗或後續失敗，並以未出現失敗而死亡為剔除項。 OS定義為任何原因導致的死亡時間。

The following six factors were examined in prognostic factors analyses: (1) FIGO stage (coded 1, 2, 3, and 4, and analyzed as a "continuous" variable); (2) clinical diameter , determined at examination under anesthesia [EUA]); (3) corpus invasion (no or yes); (4) tumor volume , defined as the product of three orthogonal diameters, obtained from MRI, times (ie, the volume of an ellipsoid) (analyzed as the logarithm of the volume); (5) histology (coded as SCC or other [adenocarcinoma]); and (6) age. These prognostic factors have been previously described . Tumor volume was analyzed as the logarithm of tumor volume because of the skewness of the distribution.
在預後因素分析中研究了以下六個因素：(1) FIGO 分期（編碼為1、2、3 和4，作為"連續"變數進行分析）；(2) 臨床直徑 ，在麻醉下檢查[EUA] 時確定）；(3) 腫瘤侵犯（無或有）；(4) 腫瘤體積 ，定義為核磁共振成像獲得的三個正交直徑乘以< b2> 的乘積（即橢圓體的體積）（以體積的對數進行分析）；(5) 組織學（編碼為SCC 或其他[腺癌]）；以及(6) 年齡。這些預後因素之前已有描述 。由於分佈的偏斜性，腫瘤體積以腫瘤體積的對數進行分析。

Where required, continuous variables were grouped at their median values. A closeout date for the study was defined as the earliest of the last dates of follow-up of patients alive and not lost to follow-up; all events after this date were ignored to minimize potential bias.
必要時，連續變數按其中位值分組。研究的結束日期定義為存活且未失去追蹤機會的患者的最後追蹤日期中最早的一個日期；該日期之後的所有事件均忽略不計，以盡量減少潛在的偏差。

Potential follow-up time for each patient was defined as the time from study entry to the cut-off date (or last follow-up for patients lost to follow-up). KaplanMeier curves for each of the time-to-event end points were calculated from which 5-year event-free rates were determined. Analysis of patterns of site of first failure was performed using competing risks analysis . Prognostic factor analyses were performed using Cox regression. In particular, multifactor analyses, in which the simultaneous effects on outcome of the factors considered were assessed, were performed using the backward elimination technique. Exploratory analyses were performed to examine the assumptions of the Cox model and to test for the presence of interactions. The effects of factors on outcome were summarized using 5-year rates and hazard ratios (HRs); standard errors (SE) were provided for 5-year rates and confidence intervals for HRs. Variables in multifactor analyses were coded such that HRs corresponded to FIGO-per increase of one stage level, clinical diameter-per 1-cm increase, corpus invasion-yes:no, tumor volume-per doubling of tumor volume, histology—adenocarcinoma:SCC, and age-per 10 -year increase. Multifactor analyses are presented in tables by showing the variables in the final model with independent prognostic value, in order of importance (as judged by value) and separated from the other variables and with HRs and values for these latter variables corresponding to the testing of the variable in question in the presence of the variables in the final model.
每位患者的潛在追蹤時間定義為從進入研究到截止日期的時間（或失去追蹤患者的最後一次追蹤時間）。計算每個事件終點的 KaplanMeier 曲線，並據此確定 5 年無事件發生率。使用競爭風險分析 對首次失敗部位的模式進行了分析。預後因子分析採用 Cox 迴歸法。尤其是多因素分析，即使用反向排除法評估考慮因素對結果的同時影響。進行探索性分析的目的是檢驗 Cox 模型的假設，並偵測是否有交互作用。各因素對結果的影響以 5 年死亡率和危險比 (HRs) 進行總結；5 年死亡率的標準誤差 (SE) 和 HRs 的置信區間 。多因素分析中的變數編碼為：HRs 對應於FIGO（每增加一個分期水平）、臨床直徑（每增加1 公分）、腫瘤侵犯（是：否）、腫瘤體積（每增加一倍腫瘤體積）、組織學（腺癌：SCC）和年齡（每增加10 歲）。多因素分析以表格的形式列出了最終模型中具有獨立預後價值的變量，按重要性排序（根據 值判斷），並與其他變量分開，同時列出了這些變量的HRs 和 值，對應於在最終模型中存在變數的情況下對相關變數進行的測試。

Two hundred forty-nine patients were eligible for the study. Baseline patient and tumor-related characteristics are shown in Table 1. The median patient age was 58 years. Two hundred and eleven patients (85%) had squamous histology and 156 (63%) of tumors exhibited corpus invasion. Tumor volume ranged from 1 to . The closeout date for the follow-up was December 31, 2004. The mean potential follow-up time was 4.5 years (range years). Five patients were lost to follow-up at 3 and 9 months and 1, 3.5, and 8 years.
249 名患者符合研究條件。患者基線特徵和腫瘤相關特徵見表 1。患者年齡中位數為 58 歲。 211名患者（85%）為鱗狀組織學，156名患者（63%）的腫瘤表現為冠狀溝侵犯。腫瘤體積從 1 到 不等。追蹤結束日期為 2004 年 12 月 31 日。平均潛在追蹤時間為 4.5 年（範圍為 年）。五名患者分別在追蹤 3 個月和 9 個月以及 1 年、3.5 年和 8 年後失去了追蹤機會。

There were a number of significant correlations between the factors considered. FIGO stage, corpus invasion, and tumor volume were all positively correlated with one another ( for each). Patients with FIGO stage I, II, III, and IV diseases were corpus positive in , and of cases; had median tumor volumes of , and ; had
所考慮的因素之間存在一些重要的相關性。 FIGO分期、腫瘤浸潤及腫瘤體積均呈正相關（ ）。 FIGO分期為I、II、III和IV期的患者中， 和 的腫瘤呈陽性；腫瘤體積的中位數分別為 和 ；腫瘤體積的中位數分別為1,000,000和1,000,000。
Table 1. Baseline characteristics
表 1.基線特徵

tumors with SCC histology in 74%, 89%, 93%, and ; and had mean ages of , and 57 years, respectively. Median tumor volume was 15 and for corpus-negative and corpus-positive patients, respectively. Older patients were more likely to have SCC histology (mean age 59 years for SCC versus 52 years for adenocarcinoma) and to have tumors with smaller clinical tumor diameters (mean ages 60 and 55 years for and , respectively).
腫瘤組織學為SCC的患者分別佔74%、89%、93%及 ；平均年齡分別為 、57歲及57歲。冠狀溝陰性和冠狀溝陽性患者的腫瘤體積中位數分別為15和 。年齡較大的患者更有可能具有SCC 組織學特徵（SCC 患者的平均年齡為59 歲，而腺癌患者的平均年齡為52 歲），其腫瘤的臨床直徑也較小（ 和 的平均年齡分別為60 歲和55 歲）。

Two hundred seventeen (87%) patients were treated with definitive radical radiotherapy, had postoperative radiotherapy, and only had surgery as their only treatment. Of 217 patients planned for radical RT, 35 were treated with radiotherapy alone (without concurrent chemotherapy) and the remaining 182 patients were treated with CCRT. One hundred ninety-seven patients received pelvic radiotherapy and 20 patients were treated with EFRT. Eighty-seven patients received low-dose rate brachytherapy and 122 had high-dose rate brachytherapy. Two of 87 patients with stage IB disease were treated by brachytherapy alone. Five patients completed EBRT but did not have brachytherapy and three patients did not complete EBRT.
有 217 名患者（87%）接受了明確的根治性放射治療， 接受了術後放射治療，只有 僅接受了手術治療。在 217 名計劃接受根治性 RT 治療的患者中，有 35 人接受了單純放射治療（未同時接受化療），其餘 182 人接受了 CCRT 治療。 197名患者接受了骨盆腔放射治療，20名患者接受了EFRT治療。 87名患者接受了低劑量近距離放射治療，122名患者接受了高劑量近距離放射治療。 87 位 IB 期患者中有 2 位只接受了近距離放射治療。五名患者完成了 EBRT 但未進行近距離放射治療，三名患者未完成 EBRT。

There were 89 deaths experienced by the 249 patients. The 5 -year rate was ( . One hundred five patients have either failed (85) or died without failure (19). The 5-year FFS rate was 55% ( ).
249 名患者中有 89 人死亡。 5 年的 率為 ( 。105 名患者要麼治療失敗（85 人），要麼沒有治療失敗而死亡（19 人）。 5年的FFS率為55% ( )。

Eighty-five patients failed at local, nodal, or distant sites. The 5-year (Kaplan-Meier) failure-free rate was . Summary unifactor analyses of time to failure for each of the prognostic factors examined are shown in Table 2: corpus invasion , larger tumor volume ( , increasing FIGO stage ( , and adenocarcinoma histology were all significantly related to an increased risk of failure. In multifactor analyses, corpus invasion , younger age , larger tumor volume , and adenocarcinoma histology were independently related to an increased risk of failure.
85名患者在局部、結節或遠端部位治療失敗。 5年（Kaplan-Meier）無失敗率為 。表2 顯示了對每個預後因素的失敗時間進行的單因素分析摘要：癌體侵犯 、腫瘤體積增大( 、FIGO 分期增加( 和腺癌組織學 都與失敗風險增加有顯著相關。小 、腫瘤體積較大 和腺癌組織學 與失敗風險增加有獨立關係。

A competing risk analysis of first failures or death without failure is shown in Figure 1. Five-year cumulative incidence rates were as follows: for local failure , nodal failure , distant failure , and death without failure .
首次失敗或無失敗死亡的競爭風險分析見圖 1。五年累計發生率如下：局部失敗 ，結節失敗 ，遠處失敗 ，無失敗死亡 。

The dominant site of first failure was cervix (in 8 patients), pelvis (12), abdomen (12), supraclavicular nodes (6), inguinal nodes (2), and distant sites (45). Sixty-three (74% of all failed) patients failed at multiple sites at the time of first failure. For comparability with other series where primary and pelvic failures are often grouped, in our series, primary site or pelvis was involved at first failure in 41 patients for which the (Kaplan-Meier) local pelvic failure-free rate was at 5 years.
首次失敗的主要部位是子宮頸（8 例）、骨盆腔（12 例）、腹部（12 例）、鎖骨上結節（6 例）、腹股溝結節（2 例）和遠端部位（45 例）。 63例（佔所有失敗患者的74%）患者在首次失敗時有多個部位失敗。為了與其他經常將原發和骨盆腔失敗歸為一組的系列研究進行比較，在我們的系列研究中，41 例患者的首次失敗涉及原發部位或骨盆腔，5 年後骨盆局部無失敗率（Kaplan -Meier）為 。

In the 85 patients who failed at least once, there were , and 3, local, pelvic, abdominal, distant, supraclavicular, and inguinal failures, respect-
在至少失敗過一次的 85 名患者中，分別有 、3 例、局部、骨盆、腹部、遠處、鎖骨上和腹股溝失敗，分別是

Figure 1. Cumulative incidence curves of first failure types (local, nodal, and distant) and death without failure.
圖 1.首次衰竭類型（局部、結節和遠處）和無衰竭死亡的累積發生率曲線。

ively, occurring either as a first failure or subsequently. At 5 years, the (Kaplan-Meier) failure-free rate was , the local failure-free (LFF) rate , the nodal failure-free (NFF) rate , and the distant failurefree rate .
在 5 年內，無局部失敗（LFF）率為 ，無結節失敗（NFF）率為 ，無遠處失敗率為 。 5 年後（Kaplan-Meier）無失敗率為 ，局部無失敗率（LFF）為 ，結節無失敗率（NFF）為 。

Twenty-nine patients developed local failure at a first or subsequent failure (29 as part of a first failure and 0 subsequently). The 5 -year (Kaplan-Meier) LFF rate for all patients was .
29 名患者在首次或隨後的失敗中出現局部失敗（29 人是首次失敗的一部分，0 人是隨後的失敗）。所有患者的 5 年（Kaplan-Meier）LFF 比率 。

The LFF rates according to FIGO stages I, II, III, and IV were , and , respectively
根據 FIGO I、I​​I、III 和 IV 期劃分的 LFF 率分別為 和

Table 2. Prognostic factor analyses for 5 -year failure-free rates
表 2.5 年無失敗率的預後因素分析

. The LFF rates with respect to clinical diameter less than was ( ) and for tumors of greater than or equal to was ( . Two hundred eleven patients with SCC histology had a 5-year LFF rate of (SE = , and in 38 patients with adenocarcinoma histology, the LFF rate was . Only 1 of 93 corpus noninvasive tumors had failed locally as opposed to a -year LFF rate in 28 of 156 patients whose tumors had invaded uterine corpus . With respect to tumor volume, patients with three smaller tumor ( failed locally, achieving an LFF rate of , whereas the LFF rate in 131 patients with tumor volume greater than or equal to was following local failure in 26 patients .
.臨床直徑小於的LFF 率為( ) ，腫瘤直徑大於或等於的LFF 率為( 。211 名組織學為SCC 的患者的5 年LFF 率為（SE = ，38 名組織學為腺癌的患者的LFF 率為。率為。手術失敗，其LFF 率為。 < b14>

Unifactor and multifactor (Cox regression) analyses of time to local failure are shown in Table 3. Larger tumor (volume, , younger age , and corpus invasion were significantly and independently associated with an increased risk of local failure. Clinical diameter, adenocarcinoma histology, and FIGO stage were not significantly related to risk of local failure in the presence of these factors. The HR for a doubling of tumor volume was CI 1.43-2.79), for an increase of 10 years of age was 0.65 (95% CI ), and for corpus invasion was 9.11 (95% CI 1.18-71.0).
表 3 列出了局部手術失敗時間的單一因素和多因素（Cox 回歸）分析。腫瘤體積越大（ ）、年齡越小（ ）和癌腫侵犯 ，與局部失敗的風險增加有顯著的獨立相關性。臨床直徑、腺癌組織學和FIGO分期與局部失敗的風險無明顯關係。 腫瘤體積增加一倍的HR為 CI 1.43-2.79），年齡增加10歲的HR為0.65（95% CI ），腫瘤侵犯的HR為9.11（95% CI 1.18-71.0）。

Sixty-eight patients developed nodal failure at a first or subsequent failure (63 as part of a first failure and 5
68 名患者在首次或後續治療失敗時出現結節衰竭（其中 63 人屬於首次治療失敗，5 人屬於後續治療失敗）。

Table 3. Unifactor and multifactor analyses (Cox regression) of time to local relapse
表 3.局部復發時間的單因子與多因子分析（Cox 迴歸

Variables in the upper section of the table are those that are significant in multifactor analysis; the HR and value of each other variables are those corresponding to testing that variable while adjusting for the significant variables. subsequently). The 5-year (Kaplan-Meier) NFF rate for all patients was . Five-year (KaplanMeier) NFF rates according to FIGO stages I, II, III, and IVA were , and 33% ( , respectively. Five-year NFF rates with respect to clinical diameter less than was . In larger tumors, greater than or equal to , the 5 -year NFF rate was . The 5-year NFF rates of patients with SCC and adenocarcinoma histology were and (SE =9), respectively. Only 10 of 93 (5-year NFF rate ) patients with corpus noninvasive tumors had nodal failure as opposed to a 5 -year NFF rate of in patients whose tumors had invaded the uterine corpus. With respect to tumor volume, in smaller tumors ), the 5-year NFF rate was , whereas the 5 -year NFF rate in patients with tumor volumes greater than or equal to was affecting 47 patients.
表上部分的變數為多因子分析中顯著的變數；其他各變數的 HR 和 值為在調整顯著變數的同時測試該變數所對應的值。隨後）。所有患者的5年（Kaplan-Meier）NFF率為 。根據FIGO I、I​​I、III 和IVA 期劃分的5 年（Kaplan-Meier）NFF 率分別為和33% ( 。臨床直徑小於的五年NFF 率為。對於大於或等於的較大腫瘤，5 年NFF比率。 ，而腫瘤侵犯子宮體的患者5年NFF率為。受到影響。 < b8>

Unifactor and multifactor (Cox regression) analyses of time to nodal failure are shown in Table 4. Corpus invasion, younger age, and larger tumor volume were all highly significantly and independently related to increased risk of nodal failure. The HR for corpus invasion was 3.44 (95% CI 1.62-7.30), for an increase of 10 years of age was 0.77 ( , and for a doubling of tumor volume was 1.08-1.66). Clinical diameter, adenocarcinoma histology, and FIGO stage were not significantly related to nodal failure in the presence of these factors.
表 4 列出了結節失敗時間的單因素和多因素（Cox 回歸）分析。癌灶侵犯、年齡較小、腫瘤體積較大均與結節失敗風險的增加有顯著的獨立相關性。腫瘤浸潤的HR為3.44（95% CI 1.62-7.30），年齡增加10歲的HR為0.77（ ， ，腫瘤體積增加一倍的HR為 1.08-1.66）。在有這些因素的情況下，臨床直徑、腺癌組織學和 FIGO 分期與結節失敗無明顯關係。

Fifty-two patients developed distant failure at a first or subsequent failure ( 45 as part of a first failure and
52名患者在首次或後續治療失敗時出現遠端治療失敗（其中45名患者是首次治療失敗的一部分，另1名患者是後續治療失敗的一部分）。

Table 4. Unifactor and multifactor analyses (Cox regression) of time to nodal failure
表 4.節點故障時間的單因子與多因子分析（Cox 迴歸

Variables in the upper section of the table are those that are significant in multifactor analysis; the HR and value of each other variables are those corresponding to testing that variable while adjusting for the significant variables.
表格上部的變數是在多因素分析中顯著的變數；其他各變數的 HR 值和 值是在對顯著變數進行調整後測試該變數所對應的值。

7 subsequently). The 5 -year (Kaplan-Meier) distant failure-free rate for all patients was . Unifactor and multifactor (Cox regression) analyses of time to distant failure are shown in Table 5. Only corpus invasion and younger age ( were highly significantly and independently related to an increased risk of distant failure. FIGO stage, tumor volume clinical diameter, and adenocarcinoma histology were not significantly related to distant failure in the presence of these factors.
7 隨後）。所有患者的 5 年（Kaplan-Meier）無遠端轉移失敗率為 。表5顯示了遠處失敗時間的單一因素和多因素（Cox迴歸）分析。表 5 顯示了單因子和多因素（Cox 迴歸）分析的遠端轉移失敗時間。只有冠狀溝侵犯 和年齡較小（ ）與遠端轉移失敗風險的增加有顯著的獨立相關性。 FIGO分期、腫瘤體積臨床直徑及腺癌組織學與遠端轉移失敗無明顯關係。

Fifteen extrafascial hysterectomies were performed for either confirmed or suspected residual or recurrent disease at cervix. Of these, only five hysterectomy specimens had residual or recurrent cancer and only two have remained free of disease.
有 15 例的筋膜外子宮切除術是因子宮頸確診或懷疑有殘留或復發疾病而進行。其中，只有五例子宮切除標本有殘留或復發癌症，只有兩例標本一直沒有病變。

Only eight patients had an isolated failure at either pelvic or para-aortic sites that were amenable to salvage treatment.
只有八名患者的骨盆或主動脈旁部位出現了孤立的故障，但可以進行挽救治療。

Almost all patients receiving radical CCRT developed varying degrees of diarrhea, particularly toward the end of radiotherapy. These resolved with either dietary modification or medical management. Severity of diarrhea was greater in patients receiving EFRT. None of the patients failed to complete treatment on account of treatment-related toxicities. However, a handful of patients on EFRT required treatment break or blood transfusion on account of either intestinal or hemato-
幾乎所有接受根治性 CCRT 的患者都會出現不同程度的腹瀉，尤其是在放射治療即將結束時。經過飲食調整或藥物治療後，這些腹瀉症狀都得到了緩解。接受EFRT治療的患者腹瀉程度更嚴重。沒有一名患者因治療相關毒性反應而未能完成治療。不過，也有少數接受 EFRT 治療的患者因腸道或血液毒性而需要中斷治療或輸血。

Table 5. Unifactor and multifactor analyses (Cox regression) of time to distant failure
表 5.遠期失敗時間的單因子與多因子分析（Cox 迴歸

Variables in the upper section of the table are those that are significant in multifactor analysis; the and value of each other variable are those corresponding to testing that variable while adjusting for the significant variables. logic toxicity. The patients receiving postoperative central pelvic radiotherapy for node-negative highrisk cervical cancer usually remained asymptomatic throughout the treatment.
表中上部的變數為多因素分析中顯著的變數；其他各變數的 和 值為測試該變數同時調整顯著變數的對應值。邏輯毒性。接受結節陰性高風險子宮頸癌術後骨盆中央放射治療的患者通常在整個治療過程中都沒有症狀。

Twenty-eight patients developed significant late complications. Five patients of 85 had stage I disease, 14 of 114 had stage II, and 9 of 44 had stage III disease. Of the 28,11 patients had grade 4 toxicities. Seven had vaginal vault necroses, three had rectovaginal fistulae, and two had bladder base necrosis. One patient had grade 4 toxicities both in bladder and in vagina. Rectovaginal fistula and bladder base necrosis were related to tumor invasion prior to chemoradiotherapy. Of the seven patients with vaginal vault necrosis, six patients had salvage hysterectomy performed prior to developing grade 4 toxicity. Of the remaining 17 patients, 6 had small bowel surgery for intestinal obstruction, 2 developed contracted bladder, and 9 patients (although asymptomatic) had completely fused vagina.
有 28 名患者 出現了嚴重的晚期併發症。 85 位患者中有 5 位處於 I 期，114 位患者中有 14 位處於 II 期，44 位患者中有 9 位處於 III 期。 28 名患者中有 11 名出現了 4 級毒性反應。其中 7 人陰道穹隆壞死，3 人直腸陰道瘺，2 人膀胱底部壞死。一名患者的膀胱和陰道均出現了 4 級毒性反應。直腸陰道瘺和膀胱底部壞死與化療前的腫瘤侵犯有關。在7名陰道穹隆壞死的患者中，有6名患者在出現4級毒性之前進行了挽救性子宮切除術。在其餘17名患者中，6名患者因腸阻塞接受了小腸手術，2名患者出現膀胱收縮，9名患者（雖然無症狀）的陰道完全融合。

Corpus invasion, tumor volume, and age were the dominant factors related to risks of each of the various failure types in radically treated advanced cervical cancer patients. In particular, FIGO stage and clinical diameter, in the presence of corpus invasion, tumor volume, and age, were not significantly related to risk of any of the failure types. Corpus invasion and age were independent factors for all of local, nodal, and distant failure. There was a tendency for corpus invasion to be more predictive of nodal and distant failure and tumor volume to be more predictive of local failure (and not predictive of distant failure). It has been noted in a series that younger patients have poorer prognosis, and this is demonstrated in our data.
在接受根治性治療的晚期子宮頸癌患者中，子宮體侵犯、腫瘤體積和年齡是導致各種失敗類型風險的主要因素。特別是，在存在癌體侵犯、腫瘤體積和年齡的情況下，FIGO分期和臨床直徑與任何一種失敗類型的風險都沒有顯著關係。腫瘤浸潤和年齡是所有局部、結節和遠端失敗的獨立因素。有一種趨勢是，癌體侵犯更能預測結節和遠處失敗，而腫瘤體積更能預測局部失敗（但不能預測遠處失敗）。 一系列研究指出，年輕患者的預後較差，我們的數據也證明了這一點。

The performance of FIGO stage and clinical diameter in these multifactor analyses has possible implications for the staging of advanced cervical cancer. Presently, cervical cancer is staged following clinical examination and assigned to a particular stage according to the FIGO staging system. In modern clinical practice where new imaging modalities are available and can be used, it would seem advisable to use information, such as tumor volume and corpus invasion status, to better define prognostic categories, and thereby better define the management of the patient. For example, patients with lower FIGO stage but with
在這些多因素分析中，FIGO 分期和臨床直徑的表現可能會對晚期子宮頸癌的分期產生影響。目前，子宮頸癌根據臨床檢查進行分期，並依照 FIGO 分期系統劃分為特定階段。在現代臨床實踐中，新的影像模式已經出現並且可以使用，利用腫瘤體積和癌腫侵犯狀態等資訊來更好地確定預後類別，從而更好地確定對患者的管理，似乎是明智之舉。例如，FIGO 分期較低但有以下情況的患者
corpus invasive tumors may require special consideration. Also in clinical trials, incorporation of prognostic information from these sources can help in the design of trials to test treatments in more prognostically homogeneous patient groups.
此外，在臨床試驗中，納入這些來源的預後資訊也有助於設計試驗，對預後較一致的患者群體進行治療測試。此外，在臨床試驗中，納入這些來源的預後資訊有助於設計試驗，以便在預後更一致的患者群體中測試治療方法。

Twenty-nine patients failed locally giving an LFF rate at 5 years of . This suggests that CCRT results in adequate tumor shrinkage at the primary site, enabling brachytherapy to be effectively applied. It also suggests that attempts to improve local-failure rates may not have an impact on OS. These results compared favorably with those of earlier studies, which did not use concurrent chemotherapy, such as that by Kapp et al. , where of primary failure or persistent disease was observed, and by Ferrigno et al. who reported local failure. Patient characteristics in the current series and that of chemoradiotherapy arm of RTOG 90-01 study were fairly comparable. It is not possible to compare our local failure rate with other studies where CCRT has been , for these and many other studies describe pelvic failure rates, which include local as well as pelvic failures. Allowing for this definition, our "pelvic" failure rate was , which is similar to the obtained in the chemoradiotherapy arm of RTOG and similar to that in GOG-165 . Pelvic failure without concurrent chemotherapy has been reported to be in the range of (radiation-only arm of RTOG 90-01) .
29例患者局部治療失敗，5年後的LFF率為 。這表明，CCRT 可以使原發部位的腫瘤充分縮小，使近距離治療得以有效應用。這也表明，提高局部失敗率的嘗試可能不會對手術安全性產生影響。這些結果與早期未同時使用化療的研究結果相比更為理想，如Kapp 等人的研究（ ），他們觀察到了 的原發失敗或持續性疾病，以及Ferrigno 等人的研究（ ），他們報告了 的局部失敗。本系列研究的患者特徵與RTOG 90-01研究化學放射治療組的患者特徵相當。我們無法將局部失敗率與其他使用 CCRT 的研究進行比較 ，因為這些研究和其他許多研究都描述了骨盆失敗率，其中包括局部失敗和骨盆失敗。考慮到這個定義，我們的"骨盆腔"失敗率為 ，與RTOG 化學放射治療組的 相似，也與GOG-165< b9> 相似。據報道，不同時進行化療的骨盆失敗率在 （RTOG 90-01 僅放療組） 。

Corpus invasion and younger age were independently related to an increased risk of distant failure. It is interesting that patients with tumors with larger volumes and clinical diameters are strongly associated with higher risk of distant failure in unifactor analyses, but this association disappears after allowing for corpus invasion and age. Both volume and diameter are strongly associated with corpus invasion, which may explain why only the unifactor association is significant. The overall distant failure rate in the present series was , which is higher than the reported in CCRT arm of RTOG 90-01. There is likely to have been more node-positive patients in our series, which has been observed to be associated with distant failure reported by Ohara et al. who reported the distant metastasis rate among node-negative patients at and at in patients with probable positive nodes. Also in our series, of patients had corpus inva- sion and we have shown that about of corpuspositive patients are node positive , suggesting at least node positivity in the present series compared with in the RTOG trial.
癌腫侵犯和年齡較小與遠處轉移失敗風險的增加有獨立的關係。有趣的是，在單因素分析中，腫瘤體積和臨床直徑較大的患者遠端轉移失敗的風險較高，但考慮到癌腫侵犯和年齡因素後，這種關聯就消失了。腫瘤體積和直徑都與癌腫浸潤密切相關，這或許可以解釋為什麼只有單因子相關性顯著。本研究系列的整體遠處失敗率為 ，高於 RTOG 90-01 中 CCRT 組報告的 。我們的系列研究中可能有更多的結節陽性患者，Ohara 等人的報告（ ）指出，結節陰性患者的遠端轉移率為 ，而結節可能陽性患者的遠端轉移率為 。此外，在我們的系列研究中， 的患者有冠狀溝凹陷，而我們的研究表明，約 的冠狀溝陽性患者為結節陽性 相比，本系列研究中至少 有結節陽性。

It is generally believed that increased local control results in lower incidence of distant failure and improved survival . In patients with large cervical tumors, insufficient radiation dose was thought to be the cause of increased local failure . Further, the distant metastases rate was shown to be strongly related to the clinical diameter by Perez et al. . This led to the use of extrafascial hysterectomy following radical radiotherapy in stage IB patient with tumors greater than . Extrafascial hysterectomy resulted in incidence of local failure versus when treated by radiotherapy alone . However, there was no statistical difference in survival between these groups. More importantly, the incidence of distant failure did not improve following hysterectomy when compared with those treated by radiation alone (16%). In our experience, the use of CCRT during the EBRT phase of cervical cancer treatment results in tumor shrinkage in most patients, allowing for adequate boost dose using brachytherapy.
一般認為，增加局部控制可降低遠處失敗的發生率，提高存活率 。在巨大子宮頸腫瘤患者中，放射劑量不足被認為是局部治療失敗率增加的原因 。此外，Perez 等人的研究表明，遠端轉移率與臨床直徑密切相關 。因此，在對腫瘤大於 的 IB 期患者進行根治性放射治療後，採用筋膜外子宮切除術。筋膜外子宮切除術導致 局部失敗的發生率高於 單純放療時的發生率 。然而，這兩組患者的存活率並沒有統計學差異。更重要的是，與單純放射治療相比（16%），子宮切除術後遠端失敗的發生率 並沒有改善。根據我們的經驗，在子宮頸癌治療的 EBRT 階段使用 CCRT 可使大多數患者的腫瘤縮小，從而可以使用近距離放射治療獲得足夠的增強劑量。

1 Thomas GM. Improved treatment for cervical cancer-concurrent chemotherapy and radiotherapy. Engl J Med 1999;340:1198-200.
1 Thomas GM。改善子宮頸癌的治療--同時進行化療和放療。 英國醫學雜誌，1999；340：1198-200。

2 Morris M, Eifel PJ, Lu J et al. Pelvic radiation with concurrent chemotherapy compared with pelvic and para-aortic radiation for high-risk cervical cancer [see comments]. N Engl J Med 1999;340: .
2 Morris M, Eifel PJ, Lu J et al. 骨盆腔放射治療同時化療與骨盆腔和主動脈旁放療治療高風險子宮頸癌的比較[見評論]。 N Engl J Med 1999; 340: 。

3 Narayan K, Fisher R, Bernshaw D. Significance of tumor volume and corpus uteri invasion in cervical cancer patients treated by radiotherapy. Int J Gynecol Cancer 2006;16:623-30.
3 Narayan K，Fisher R，Bernshaw D.放射治療子宮頸癌患者腫瘤體積和子宮體侵犯的意義。 Int J Gynecol Cancer 2006;16:623-30.

4 Narayan K, McKenzie AF, Hicks RJ, Fisher R, Bernshaw D, Bau S. Relation between FIGO stage, primary tumor volume, and presence of lymph node metastases in cervical cancer patients referred for radiotherapy. Int J Gynecol Cancer 2003;13:657-63.
4 Narayan K, McKenzie AF, Hicks RJ, Fisher R, Bernshaw D, Bau S. 轉診接受放射治療的子宮頸癌患者的 FIGO 分期、原發腫瘤體積和淋巴結轉移之間的關係。 Int J Gynecol Cancer 2003;13:657-63.

5 Kalbfleisch JD, Prentice RL. The statistical analysis of failure data. New York, NY: John Wiley & Sons Inc, 1980.
5 Kalbfleisch JD, Prentice RL.The statistical analysis of failure data.紐約州紐約市：John Wiley & Sons Inc, 1980.

6 Meanwell CA, Kelly KA, Wilson S et al. Young age as a prognostic factor in cervical cancer: analysis of population based data from 10,022 cases. Br Med J (Clin Res Ed) 1988;296:386-91.
6 Meanwell CA, Kelly KA, Wilson S et al. 《年輕是子宮頸癌的預後因素：對 10,022 例病例的人口資料分析》。 Br Med J (Clin Res Ed) 1988;296:386-91.

7 Kapp KS, Stuecklschweiger GF, Kapp DS et al. Prognostic factors in patients with carcinoma of the uterine cervix treated with external beam irradiation and IR-192 high-dose-rate brachytherapy. Int J Radiat Oncol Biol Phys 1998;42:531-40.
7 Kapp KS, Stuecklschweiger GF, Kapp DS et al. 接受體外照射和 IR-192 高劑量率近距離放射治療的子宮頸癌患者的預後因素。 Int J Radiat Oncol Biol Phys 1998;42:531-40.

8 Ferrigno R, dos Santos Novaes PE, Pellizzon AC et al. High-doserate brachytherapy in the treatment of uterine cervix cancer. Analysis of dose effectiveness and late complications. Int J Radiat Oncol Biol Phys 2001;50:1123-35.
8 Ferrigno R, dos Santos Novaes PE, Pellizzon AC et al. 高劑量近距離放射治療子宮頸癌。劑量有效性和晚期併發症分析。 Int J Radiat Oncol Biol Phys 2001；50：1123-35。

9 Lanciano R, Calkins A, Bundy BN et al. Randomized comparison of weekly cisplatin or protracted venous infusion of fluorouracil in combination with pelvic radiation in advanced cervix cancer: a Gynecologic Oncology Group Study. J Clin Oncol 2005;23:8289-95.
9 Lanciano R, Calkins A, Bundy BN et al. 每週順鉑或長期靜脈輸注氟尿嘧啶合併骨盆腔放射治療晚期子宮頸癌的隨機比較：婦科腫瘤組研究。 J Clin Oncol 2005;23:8289-95.

10 Eifel PJ, Winter K, Morris M et al. Pelvic irradiation with concurrent chemotherapy versus pelvic and para-aortic irradiation for high-risk cervical cancer: an update of Radiation Therapy Oncology Group Trial (RTOG) 90-01. J Clin Oncol 2004;22:872-80.
10 Eifel PJ、Winter K、Morris M 等：高風險子宮頸癌的骨盆腔照射與同期化療對比骨盆腔和主動脈旁照射：放射治療腫瘤學組試驗（RTOG）90-01 的更新。 J Clin Oncol 2004;22:872-80.

11 Petereit DG, Sarkaria JN, Potter DM, Schink JC. High-dose-rate versus low-dose-rate brachytherapy in the treatment of cervical cancer: analysis of tumor recurrence - the University of Wisconsin experience. Int J Radiat Oncol Biol Phys 1999;45:1267-74.
11 Petereit DG、Sarkaria JN、Potter DM、Schink JC。子宮頸癌治療中的高劑量率近距離放射治療與低劑量率近距離放射治療：腫瘤復發分析-威斯康辛大學的經驗。 Int J Radiat Oncol Biol Phys 1999;45:1267-74.

12 Perez CA, Grigsby PW, Chao KS, Mutch DG, Lockett MA. Tumor size, irradiation dose, and long-term outcome of carcinoma of uterine cervix. Int J Radiat Oncol Biol Phys 1998;41:307-17.
12 Perez CA、Grigsby PW、Chao KS、Mutch DG、Lockett MA。子宮頸癌的腫瘤大小、照射劑量和長期預後。 Int J Radiat Oncol Biol Phys 1998;41:307-17.

13 Ohara K, Tsunoda H, Nishida M et al. Use of small pelvic field instead of whole pelvic field in postoperative radiotherapy for node-negative, high-risk stages I and II cervical squamous cell carcinoma. Int J Gynecol Cancer 2003;13:170-6.
13 Ohara K, Tsunoda H, Nishida M et al. 在結節陰性、高危險 I 期和 II 期子宮頸鱗狀細胞癌術後放射治療中使用小盆腔野代替全盆腔野。 Int J Gynecol Cancer 2003;13:170-6.

14 Suit HD, Westgate SJ. Impact of improved local control on survival. Int J Radiat Oncol Biol Phys 1986;12:453-8.
14 Suit HD, Westgate SJ.改善局部控制對存活的影響》。 Int J Radiat Oncol Biol Phys 1986;12:453-8.

15 Durrance FY, Fletcher GH, Rutledge FN. Analysis of central recurrent disease in stages I and II squamous cell carcinomas of the cervix on intact uterus. Am J Roentgenol Radium Ther Nucl Med 1969; 106:831-8.
15 Durrance FY、Fletcher GH、Rutledge FN。完整子宮頸鱗狀細胞癌 I 期和 II 期中央復發病灶分析。 Am J Roentgenol Radium Ther Nucl Med 1969; 106:831-8.

16 Perez CA, Grigsby PW, Nene SM et al. Effect of tumor size on the prognosis of carcinoma of the uterine cervix treated with irradiation alone. Cancer 1992;69:2796-806.
16 Perez CA, Grigsby PW, Nene SM et al. 腫瘤大小對單純照射治療子宮頸癌預後的影響。癌症》1992；69：2796-806。

17 Keys HM, Bundy BN, Stehman FB et al. Radiation therapy with and without extrafascial hysterectomy for bulky stage IB cervical carcinoma: a randomized trial of the Gynecologic Oncology Group. Gynecol Oncol 2003;89:343-53.
17 Keys HM、Bundy BN、Stehman FB 等：《對大塊 IB 期子宮頸癌進行放射治療與不進行筋膜外子宮切除術：婦科腫瘤組的隨機試驗》。 Gynecol Oncol 2003;89:343-53.

Accepted for publication June 28, 2007
2007 年 6 月 28 日接受出版