Original article
Diagnostic value of 18F-FDG-PET or PET-CT in recurrent cervical cancer: a systematic review and meta-analysis Yanxia Chua, Ai Zhenga, Fei Wangb, Wei Lina, Xiaoyun Yanga, Ling Hana, Yali Chena and Liping Baia Objectives Accurate detection of recurrent cervical cancer remains a clinical difficulty. This study aims to assess the diagnostic value of PET or PET-computed tomography (PET-CT) using 18F-fluorodeoxyglucose (18F-FDG) in recurrent cervical cancer using a meta-analysis. Study design All published studies in English evaluating the diagnostic value of PET or PET-CT in detecting recurrent cervical cancer were collected. The methodological quality of the included studies was evaluated. Pooled sensitivity, specificity, diagnostic odds ratio, and summary receiver-operating characteristic curves were obtained using statistical software. Twenty studies were included in the meta-analysis. Results The meta-analysis showed that the pooled sensitivity and specificity of PET and PET-CT to detect distant metastasis in recurrent cervical cancer were 0.87 [95% confidence interval (CI): 0.80–0.92] and 0.97 (95% CI: 0.96–0.98), respectively. The pooled sensitivity and specificity
Introduction Despite the advances in screening and prevention, cervical cancer remains a major threat to women’s health and accounts for B150 000 deaths annually worldwide [1]. The recurrence rate of uterine cervical cancer is reported to be 6.5% after surgery and 26.2% after radiation therapy alone [2]. Once the primary treatment has failed, the opportunity for secondary cure is minimal. Morbidity of salvage therapy is usually high, and 5-year survival rates after recurrence are low [3]. Early detection of recurrence may influence survival. Therefore, attempts to improve surveillance after treatment could lead to earlier detection of relapse, and precise assessment of the recurrent status could improve outcome. Radiological techniques, such as intravenous urography, ultrasonography, computed tomography (CT), and MRI, are used to detect recurrent cervical cancer [4]. However, these imaging modalities exhibit limitations in differentiating recurrent tumor from postoperative or radiation fibrosis and in detecting metastatic normal-sized lymph nodes and extrapelvic metastases [5]. Whole-body PET using the glucose analog 18F-fluorodeoxyglucose (18F-FDG) is a unique imaging modality that is based on the differential metabolic characteristics of malignant and normal tissues. Locally recurrent disease is often difficult to detect on pelvic examination because of the thickening of soft-tissue structures after radiation and/ or surgery. Detection of recurrent disease in the pelvis using c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0143-3636
for local regional recurrence were 0.82 (95% CI: 0.72–0.90) and 0.98 (95% CI: 0.96–0.99), respectively. Conclusion 18F-FDG-PET and PET-CT are valuable methods for the assessment of recurrent cervical c 2014 Wolters cancer. Nucl Med Commun 35:144–150 Kluwer Health | Lippincott Williams & Wilkins. Nuclear Medicine Communications 2014, 35:144–150 Keywords: cervix, meta-analysis, recurrence, tomography a Department of Obstetric and Gynecology, West China Second University Hospital, Sichuan University, Chengdu, China and bDepartment of Obstetric and Gynecology, People’s Hospital of Shanxi Province, 256 Xi’an Friendship West Road, Shanxi 710068, China
Correspondence to Ai Zheng, MD, Department of Obstetric and Gynecology, West China Second University Hospital, Sichuan University, Section 3, 20S. Renmin Road, Chengdu, Sichuan 610041, China Tel: + 86 28 85503960; fax: + 86 28 85559065; e-mail: [email protected] Received 10 September 2013 Revised 27 September 2013 Accepted 1 October 2013
MRI and CT is problematic in this setting because discrete masses are usually not present. PET or PET-CT has a high diagnostic value in the detection of recurrent cervical cancer [6], but the number of cases is generally insufficient. Moreover, the results among various studies are significantly different; thus, the assessed diagnostic value of 18F-FDGPET (PET-CT) is not consistent. Meta-analysis is an important tool for precisely defining the effect of selected genetic polymorphisms on the risk of disease and identifying potentially important sources of between-study heterogeneity. Thus, the present meta-analysis aims to explore the diagnostic value of 18 F-FDG-PET or PET-CT in assessing distant metastasis and local regional recurrence in cervical cancer. The results of this study may serve as a reference for future cost-effectiveness studies.
Materials and methods Data sources
A comprehensive computer literature search of abstracts of studies on human subjects was conducted to identify articles that focused on the diagnostic performance of PET or PET-CT in detecting recurrent cervical cancer. The MEDLINE and EMBASE databases were searched from January 1995 to May 2013 using the following keywords: (‘PET’ or ‘positron emission tomography’ or ‘FDG’ or ‘fluorodeoxyglucose’), (‘cervical carcinoma’ DOI: 10.1097/MNM.0000000000000026
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Value of
or ‘cervical cancer’ or ‘carcinoma of cervix’), and (‘sensitivity’ or ‘specificity’ or ‘false negative’ or ‘false positive’ or ‘diagnosis’ or ‘detection’ or ‘accuracy’). The list of articles was supplemented with extensive crosschecking of the reference lists of all retrieved articles. Study selection
The following studies were included: (a) articles that were reported in English; (b) studies that used 18F-FDGPET or PET-CT to identify or characterize recurrent cervical cancer; (c) studies that used histopathologic analysis and/or close clinical and imaging follow-up for at least 6 months as the reference standard; (d) studies that presented sufficient data for per-patient or per-lesion statistics to calculate the true-positive and false-negative values; and (e) studies that included 10 or more patients. The following studies were excluded: (a) those that used other radiotracers; (b) those that used a small sample size; and (c) for duplicate studies, those published earlier or had a smaller sample size. Data abstraction and quality assessment
Two reviewers independently screened the titles, abstracts, and full texts to identify eligible studies. Disagreements between the reviewers were resolved by consensus. Data retrieved from the studies included a description of the study population (age), study design (whether prospective, retrospective, or unknown), patient selection (whether consecutive or nonconsecutive), and interpretation of test results (whether blinded or nonblinded). The numbers of true-positive, false-negative, false-positive, and true-negative results in the detection of recurrent cervical cancer were extracted per patient or per lesion. Table 1
18
F-FDG-PET in cervical cancer Chu et al. 145
Statistical analysis
All statistical analyses were performed using the Meta-Disc version 1.4 software package (Unidad de Bioestadı´stica Clı´nica del Hospital Ramo´n y Cajal de Madrid, Madrid, Spain). For each study, we constructed 2 2 contingency tables wherein all subjects were classified to have positive or negative 18F-FDG-PET or PET-CT results. We calculated the true-positive rate (TPR; sensitivity) and false-positive rate (FPR; 1 – specificity). We also calculated the summary receiver-operating characteristic curves and the area under the curve. A graph was plotted from the TPR and FPR points. The summary receiver-operating characteristic curve was plotted over these points to form a smooth curve, which can be achieved using the regression model proposed by Moses et al. [7]. Two statistical methods were used to precisely assess between-study heterogeneity. The w2-based Q statistic test (Cochran’s Q statistic) was used to test for heterogeneity, and the I2 statistic was calculated to quantify the proportion of total variation due to heterogeneity [8]. The I2 index was calculated to assess between-study heterogeneity. I2 values of 25, 50, and 75% were used as evidence of low, moderate, and high heterogeneity, respectively. If moderate or high heterogeneity was calculated, the random-effects model was used to pool the results; otherwise, the fixed-effects model was used when I2 was less than 50%.
Results Literature search and study design characteristics
A total of 20 publications met the inclusion criteria [6,9–27]. For three of the 20 studies, the investigators were blinded to the diagnostic findings from other modalities (e.g. CT and MRI), but the investigators for the remaining 17 studies were either not blinded or their
Main characteristics of the included studies
References Park et al. [9] Sun et al. [10] Belhocine et al. [11] Nakamoto et al. [12] Ryu et al. [13] Havrilesky et al. [14] Lai [15] Yen et al. [16] Chang et al. [17] Grisaru et al. [26] Sakurai et al. [18] Amit et al. [27] Sironi et al. [6] Chung et al. [23] Van der Veldt et al. [19] Kitajima et al. [20] Mittra et al. [21] Pallardy et al. [24] Cetina et al. [25] Lee et al. [22]
Mean age (range)
Patient selection
Blind
Evaluable patients or lesion
53 (ND) ND 52 (38–66) 52 (26–82) 51 (31–78) 42 (28–69) 51 (25–87) 51 (25–86) 54 (35–76) 56 (20–85) 56 (27–80) 50 (31–71) 28–69 53 (32–77) 41 (27–61) 58 (37–78) 50 (28–87) 46 (35–81) 47 ND
ND ND ND ND ND ND Consecutive ND Consecutive Consecutive Consecutive ND Consecutive ND ND Consecutive ND ND ND ND
ND ND ND ND ND ND ND ND ND Yes ND ND Yes ND ND Yes ND ND ND ND
36 20 60 20 249 29* 400* 550* 27 12 54 28 12 32 39 52 30 40 16 51
Recurrent number (%) 19 18 28 5 31 22 67 94 18 10 47 7 5 28 25 25 24 33 12 37
(53) (90) (47) (25) (12) (76) (17) (17) (67) (83) (87) (25) (42) (88) (64) (48) (80) (83) (75) (73)
Noninvasive modalities
Study design
PET PET PET PET PET PET PET PET PET PET/CT PET PET/CT PET/CT PET/CT PET PET/CT;PET PET/CT PET/CT PET/CT PET/CT
ND Retrospectively Retrospectively Retrospectively Retrospectively Retrospectively Retrospectively Prospective ND ND ND ND Retrospectively Retrospectively Retrospectively ND Retrospectively Retrospectively Retrospectively Retrospectively
CT, computed tomography; ND, not documented.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
146
Nuclear Medicine Communications 2014, Vol 35 No 2
blinding status was not defined. Moreover, patient recruitment was consecutive in six studies and not defined in 14 studies. The main characteristics of the included studies are presented in Table 1.
Fig. 1
(a) Park et al. [9] Sun et al. [10] Nakamoto et al. [12] Belhocine et al. [11] Ryu et al. [13] Havrilesky et al. [14] Lai et al. [15] Yen et al. [16] Chang et al. [17] Sakurai et al. [18] Kitajima et al. [20] van der Veldt et al. [19] Grisaru et al. [26] Chung et al. [23] Amit et al. [27] Sironi et al. [6] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25] Lee et al. [22]
Main results
The pooled sensitivity and specificity of PET or PET-CT were 0.92 [95% confidence interval (CI): 0.89–0.94] and 0.90 (95% CI: 0.91–0.93), respectively. The pooled sensitivity and specificity of PET or PET-CT for the assessment of distant metastasis in cervical cancer were 0.87 (0.80–0.92) and 0.97 (0.96–0.98), respectively. The pooled sensitivity and specificity for the assessment of local regional recurrence were 0.82 (0.72–0.90) and 0.98 (0.96–0.99), respectively. The results are also shown in Figs 1–3. 0
When all 12 PET studies were considered, the pooled sensitivity and specificity were 0.91 (0.88–0.94) and 0.92 (0.90–0.94), respectively. In subgroup analysis, the pooled sensitivity and specificity for the assessment of distant metastasis were 0.84 (0.74–0.91) and 0.92 (0.89–0.94), respectively. The pooled sensitivity and specificity for the assessment of local regional recurrence were 0.84 (0.74–0.91) and 0.92 (0.89–0.94), respectively (Table 2).
0.2
0.4
0.8
1
Pooled sensitivity = 0.92 (0.89 − 0.94) χ 2 = 20.49; d.f .= 19 (P = 0.3654) Inconsistency (I 2 ) = 7.3%
Sensitivity
(b) Park et al. [9] Sun et al. [10] Nakamoto et al. [12] Belhocine et al. [11] Ryu et al. [13] Havrilesky et al. [14] Lai et al. [15] Yen et al. [16] Chang et al. [17] Sakurai et al. [18] Kitajima et al. [20] van der Veldt et al. [19] Grisaru et al. [26] Chung et al. [23] Amit et al. [27] Sironi et al. [6] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25] Lee et al. [22]
When all eight PET-CT studies were considered, the pooled sensitivity and specificity were 0.94 (0.90–0.97) and 0.84 (0.75–0.91), respectively. The pooled sensitivity and specificity for the assessment of distant metastasis were 0.90 (0.80–0.96) and 0.99 (0.98–1.00), respectively. The pooled sensitivity and specificity for the assessment of local regional recurrence were 0.83 (0.69–0.93) and 0.96 (0.89–0.99), respectively (Table 2).
0
0.2
0.4
0.8
0.6
Specificity (95% CI) 0.94 (0.73 − 1.00) 1.00 (0.16 − 1.00) 0.60 (0.32 − 0.84) 0.77 (0.46 − 0.95) 0.76 (0.70 − 0.82) 0.87 (0.60 − 0.98) 0.98 (0.96 − 0.99) 0.98 (0.97 − 0.99) 0.78 (0.40 − 0.97) 0.57 (0.18 − 0.90) 0.78 (0.58 − 0.91) 0.93 (0.66 − 1.00) 1.00 (0.16 − 1.00) 0.81 (0.58 − 0.95) 1.00 (0.40 − 1.00) 1.00 (0.54 − 1.00) 0.93 (0.76 − 0.99) 0.71 (0.29 − 0.96) 0.50 (0.07 − 0.93) 0.71 (0.42 − 0.92)
Pooled specificity = 0.91 (0.90 − 0.93) χ 2 = 155.39; d.f .= 19 (P = 0.0000) 1 Inconsistency (I 2 )= 87.8 %
Specificity
(c)
SROC curve 1
Publication bias
Symmetric SROC AUC = 0.9608 SE (AUC) = 0.0080 Q∗ = 0.9057 SE (Q∗ ) = 0.0117
0.9 0.8 0.7 Sensitivity
Begg’s funnel and Egger’s test were performed to assess the publication bias in this meta-analysis. The shape of the Funnel plots did not reveal significant evidence of asymmetry, and the P value of Egger’s test was 0.243 (P > 0.05), providing statistical evidence of the symmetry of the Funnel plots (Fig. 4). Thus, the results above suggested that publication bias was not evident in this meta-analysis.
0.6
Sensitivity (95% CI) 1.00 (0.81 − 1.00) 0.89 (0.65 − 0.99) 1.00 (0.48 − 1.00) 1.00 (0.86 − 1.00) 0.90 (0.74 − 0.98) 0.86 (0.57 − 0.98) 0.91 (0.82 − 0.97) 0.89 (0.81 − 0.95) 0.94 (0.73 − 1.00) 0.91 (0.80 − 0.98) 0.80 (0.59 − 0.93) 0.92 (0.74 − 0.99) 1.00 (0.69 − 1.00) 0.90 (0.74 − 0.98) 0.86 (0.42 − 1.00) 0.83 (0.36 − 1.00) 0.92 (0.74 − 0.99) 0.96 (0.78 − 1.00) 1.00 (0.74 − 1.00) 0.97 (0.86 − 1.00)
0.6 0.5 0.4 0.3 0.2
Discussion Recent studies in patients with recurrent cervical cancers have yielded important results in terms of reduced mortality and prolonged survival with the use of multimodality treatments such as radical resection in combination with intraoperative high-dose brachytherapy [6,28]. However, these promising therapeutic methodologies entail more effective diagnostic procedures that may either help detect recurrence sufficiently early to give the patient the best chance for successful salvage therapy or confirm complete remission. Hence, unnecessary invasive treatments are avoided.
0.1 0 0
0.2
0.4
0.6
0.8
1
1−specificity
(a) Sensitivity, (b) specificity, and (c) summary receiver-operating characteristic (SROC) curves of 18F-FDG-PET. CI, confidence interval; 18 F-FDG, 18F-fluorodeoxyglucose.
Squamous cell carcinoma (SCC) antigen has been proposed as a tumor marker for the early diagnosis of recurrence in and monitoring of patients with cervical
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Value of
18
F-FDG-PET in cervical cancer Chu et al. 147
Fig. 2
(a)
Sensitivity (95% CI) Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0
0.2
0.4 0.6 Sensitivity
0.8
0.91 0.67 0.94 0.67 0.75 0.86 0.87 0.96 1.00
(0.76 − 0.98) (0.09 − 0.99) (0.70 − 1.00) (0.38 − 0.88) (0.35 − 0.97) (0.42 − 1.00) (0.70 − 0.96) (0.78 − 1.00) (0.16 − 1.00)
Pooled sensitivity = 0.87 (0.80 − 0.92) χ2 = 9.56; d.f.= 8 (P= 0.2972) 1 Inconsistency (I 2 ) = 16.3 %
(b)
Specificity (95% CI) Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0
0.2
0.4 0.6 Specificity
(c)
0.8
0.95 1.00 0.82 0.98 1.00 1.00 0.99 0.95 0.67
(0.93 − 0.96) (0.03 − 1.00) (0.48 − 0.98) (0.95 − 0.99) (0.77 − 1.00) (0.80 − 1.00) (0.98 − 1.00) (0.74 − 1.00) (0.09 − 0.99)
Pooled specificity = 0.97 (0.96 − 0.98) χ2 = 31.91; d.f .= 8 (P = 0.0001) Inconsistency (I 2 ) = 74.9 % 1
SROC curve
1
Symmetric SROC AUC = 0.9562 SE (AUC) = 0.0202 Q∗ = 0.8990 SE (Q∗ ) = 0.0285
0.9 0.8
Sensitivity
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
0.2
0.4 0.6 1−specificity
0.8
(a) Sensitivity, (b) specificity, and (c) summary receiver-operating characteristic (SROC) curves of 18 F-FDG, 18F-fluorodeoxyglucose.
1
18
F-FDG-PET. CI, confidence interval;
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
148
Nuclear Medicine Communications 2014, Vol 35 No 2
Fig. 3
(a)
Sensitivity (95% CI) Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0
0.2
0.6 0.4 Sensitivity
0.8
1
Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0.2
0.4 0.6 Specificity
(c)
0.8
(0.19 − 0.99) (0.40 − 0.97) (0.48 − 1.00) (0.05 − 0.85) (0.64 − 1.00) (0.17 − 0.77) (0.59 − 1.00) (0.66 − 1.00) (0.69 − 1.00)
Pooled sensitivty = 0.82 (0.72 − 0.90) 2 = 23.24; d.f. = 8 (P = 0.0031) Inconsistency (I 2) = 65.6 %
(b)
0
0.75 0.78 1.00 0.40 0.92 0.45 1.00 0.93 1.00
Specificity (95% CI) 1.00 (0.97 − 1.00) 0.90 (0.68 − 0.99) 1.00 (0.85 − 1.00) 1.00 (0.92 − 1.00) 0.96 (0.81 − 1.00) 0.94 (0.71 − 1.00) 1.00 (0.92 − 1.00) 0.93 (0.76 − 0.99) 0.67 (0.09 − 0.99)
Pooled specificity = 0.98 (0.96 − 0.99) 2 = 18.81; d.f. = 8 (P = 0.0159) 2 1 Inconsistency (I ) = 57.5 %
SROC curve
1
Symmetric SROC AUC = 0.9615 SE (AUC) = 0.0208 Q∗ = 0.9068 SE (Q∗ ) = 0.0308
0.9 0.8
Sensitivity
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
0.2
0.4 0.6 1−specificity
0.8
(a) Sensitivity, (b) specificity, and (c) summary receiver-operating characteristic (SROC) curves of 18 F-FDG, 18F-fluorodeoxyglucose.
cancer. The sensitivity of SCC ranges from less than 30% for stage I disease to greater than 90% for stage IV disease [29–31]. The SCC-Ag level increased in 28–88%
1
18
F-FDG-PET. CI, confidence interval;
of patients with SCC and was considered the most accurate serologic tumor marker for patients with uterine cervical carcinoma; it is also used as a diagnostic tool in
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Value of
Table 2
18
F-FDG-PET in cervical cancer Chu et al.
149
Pooled sensitivity, pooled specificity, AUC, and test for heterogeneity in the meta-analysis 95% CI
PET or PET-CT Distant metastasis Local regional recurrence PET Distant metastasis Local regional recurrence PET-CT Distant metastasis Local regional recurrence
Heterogeneity (SN)
Pooled sensitivity
Pooled specificity
0.92 0.87 0.82 0.91 0.84 0.81 0.94 0.90 0.83
0.91 0.97 0.98 0.92 0.92 0.99 0.84 0.99 0.96
(0.89–0.94) (0.80–0.92) (0.72–0.90) (0.88–0.94) (0.74–0.91) (0.64–0.92) (0.90–0.97) (0.80–0.96) (0.69–0.93)
(0.90–0.93) (0.96–0.98) (0.96–0.99) (0.90–0.94) (0.89–0.94) (0.97–1.00) (0.75–0.91) (0.98–1.00) (0.89–0.99)
Heterogeneity (SP)
AUC
I2 (%)
PH
I2 (%)
PH
0.9608 0.9562 0.9615 0.9594 0.9431 0.9622 0.9508 0.9626 0.9582
7.3 16.3 65.6 15.7 38 47.7 0 0 80.6
0.3654 0.2972 0.0031 0.2906 0.1679 0.1051 0.5405 0.6115 0.0014
87.8 74.9 57.5 92.1 57.8 54.3 34.4 65.1 58.1
0 0.001 0.0159 < 0.001 0.0503 0.0675 0.1534 0.0353 0.067
AUC, area under the curve; CI, confidence interval; CT, computed tomography; SN, sensitivity; SP, specificity.
that the presence of postoperative adhesions between loops of bowel and adjacent peritoneal surfaces complicated the imaging in the recurrent cervical cancer patients. The CT scan apparatus had difficulty in precisely detecting recurrent lesions, suggesting that interpreted CT images may have limited additional value to PET in detecting recurrent cervical cancer.
Fig. 4
Begg's funnel plot with pseudo 95% confidence limits 6
Log OR
4
2
0 −2 0
0.5
1
1.5
SE of log OR Begg’s funnel plots for assessing the publication bias (PEgger = 0.243). OR, odds ratio.
staging, a cost-effective tool for disease monitoring of therapy, and a tool for the detection of recurrence [32–35]. However, determination of the most suitable clinical regimen is difficult because of the relatively low sensitivity or lack of capacity of localization of SCC-Ag. Furthermore, conventional imaging modalities such as CT have a high diagnostic sensitivity but exhibit limitations in distinguishing between postoperative scars and relapse. PET with the radiolabeled glucose analog 18 F-FDG is a method that is based on the increased glucose metabolism of malignant tumors [16]. This method can indicate the biochemical differences between normal and malignant tissues. The addition of CT to PET provides precise anatomic information; however, whether the precise anatomic information can actually improve the overall diagnostic accuracy of PET is still unclear. In the present meta-analysis, no significant differences were found between the PET results interpreted with CT and those interpreted without the use of CT. This result could be partially attributed to the fact
The pooled estimates of the sensitivity and specificity of 18 F-FDG-PET in the detection of distant metastasis were 84% (95% CI: 74–91%) and 92% (95% CI: 89–94%), respectively. The pooled estimates of the sensitivity and specificity of PET-CT in detecting distant metastasis were 90% (95% CI: 80–96%) and 99% (95% CI: 98–100%), respectively. Therefore, PET-CT had a significantly higher diagnostic accuracy compared with PET in detecting distant metastasis in recurrent cervical cancer patients. However, the accuracy of PET-CT in detecting local regional recurrence was similar to that of 18F-FDG-PET. Nonetheless, several limitations were present in our study. First, patient characteristics are important for diagnostic accuracy; however, the effect of these variables could not be examined because of variation in data presentation or incomplete reporting of data. Second, the majority of the selected studies were retrospective. Therefore, a potential risk for subjective interpretation exists, and the diagnostic accuracy of the study can be lower. Third, selection biases may have occurred because of failure to find comprehensive, relevant articles or unpublished data. This failure could be attributed to the restriction of searching parameters. Moreover, the included articles may have poor quality. Finally, further cost-effectiveness analysis should be conducted on the surveillance techniques for cervical cancer.
Conclusion This meta-analysis is the first to evaluate the diagnostic performance of PET and PET-CT in detecting distant metastasis and local regional recurrence in cervical cancer patients. Our study showed that PET and PET-CT had
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
150
Nuclear Medicine Communications 2014, Vol 35 No 2
high diagnostic accuracy in detecting distant metastasis and local regional recurrence. These findings have clinical implication in terms of providing useful information not only to radiologists in interpreting images but also to gynecologic oncologists in choosing the imaging modality for the management of suspected recurrent cervical cancer patients.
17
Acknowledgements
20
18
19
Conflicts of interest
There are no conflicts of interest.
21
References 1 2
3 4
5
6
7
8 9
10
11
12
13
14
15 16
Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, et al. Cancer statistics, 2008. CA Cancer J Clin 2008; 58:71–96. Halpin TF, Frick HC, Munnell EW. Critical points of failure in the therapy of cancer of the cervix: a reappraisal. Am J Obstet Gynecol 1972; 114:755–764. Waggoner SE. Cervical cancer. Lancet 2003; 361:2217–2225. Walsh JW, Amendola MA, Hall DJ, Tisnado J, Goplerud DR. Recurrent carcinoma of the cervix: CT diagnosis. AJR Am J Roentgenol 1981; 136:117–122. Cunningham JJ, Fuks ZY, Castellino RA. Radiographic manifestations of carcinoma of the cervix and complications of its treatment. Radiol Clin North Am 1974; 12:93–108. Sironi S, Picchio M, Landoni C, Galimberti S, Signorelli M, Bettinardi V, et al. Post-therapy surveillance of patients with uterine cancers: value of integrated FDG PET/CT in the detection of recurrence. Eur J Nucl Med Mol Imaging 2007; 34:472–479. Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med 1993; 12:1293–1316. Higgins J, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21:1539–1558. Park DH, Kim KH, Park SY, Lee BH, Choi CW, Chin SY. Diagnosis of recurrent uterine cervical cancer: computed tomography versus positron emission tomography. Korean J Radiol 2000; 1:51–55. Sun SS, Chen TC, Yen RF, Shen YY, Changlai SP, Kao A. Value of whole body 18F-fluoro-2-deoxyglucose positron emission tomography in the evaluation of recurrent cervical cancer. Anticancer Res 2001; 21 (4B):2957–2961. Belhocine T, Thille A, Fridman V, Albert A, Seidel L, Nickers P, et al. Contribution of whole-body 18FDG PET imaging in the management of cervical cancer. Gynecol Oncol 2002; 87:90–97. Nakamoto Y, Eisbruch A, Achtyes ED, Sugawara Y, Reynolds KR, Johnston CM, et al. Prognostic value of positron emission tomography using F-18-fluorodeoxyglucose in patients with cervical cancer undergoing radiotherapy. Gynecol Oncol 2002; 84:289–295. Ryu SY, Kim MH, Choi SC, Choi CW, Lee KH. Detection of early recurrence with 18F-FDG PET in patients with cervical cancer. J Nucl Med 2003; 44:347–352. Havrilesky LJ, Wong TZ, Secord AA, Berchuck A, Clarke-Pearson DL, Jones EL. The role of PET scanning in the detection of recurrent cervical cancer. Gynecol Oncol 2003; 90:186–190. Lai CH. Management of recurrent cervical cancer. Chang Gung Med J 2004; 27:711–717. Yen TC, See LC, Chang TC, Huang KG, Ng KK, Tang SG, et al. Defining the priority of using 18F-FDG PET for recurrent cervical cancer. J Nucl Med 2004; 45:1632–1639.
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Chang TC, Law KS, Hong JH, Lai CH, Ng KK, Hsueh S, et al. Positron emission tomography for unexplained elevation of serum squamous cell carcinoma antigen levels during follow-up for patients with cervical malignancies. Cancer 2004; 101:164–171. Sakurai H, Suzuki Y, Nonaka T, Ishikawa H, Shioya M, Kiyohara H, et al. FDG-PET in the detection of recurrence of uterine cervical carcinoma following radiation therapy: tumor volume and FDG uptake value. Gynecol Oncol 2006; 100:601–607. van der Veldt AA, Buist MR, van Baal MW, Comans EF, Hoekstra OS, Molthoff CF. Clarifying the diagnosis of clinically suspected recurrence of cervical cancer: impact of 18F-FDG PET. J Nucl Med 2008; 49:1936–1943. Kitajima K, Murakami K, Yamasaki E, Domeki Y, Kaji Y, Sugimura K. Performance of FDG-PET/CT for diagnosis of recurrent uterine cervical cancer. Eur Radiol 2008; 18:2040–2047. Mittra E, El-Maghraby T, Rodriguez CA, Quon A, McDougall IR, Gambhir SS, et al. Efficacy of 18F-FDG PET/CT in the evaluation of patients with recurrent cervical carcinoma. Eur J Nucl Med Mol Imaging 2009; 36:1952–1959. Lee M, Lee Y, Hwang KH, Choe W, Park CY. Usefulness of F-18 FDG PET/ CT in assessment of recurrence of cervical cancer after treatment. Nucl Med Mol Imaging 2011; 45:111–116. Chung HH, Jo H, Kang WJ, Kim JW, Park N-H, Song Y-S, et al. Clinical impact of integrated PET/CT on the management of suspected cervical cancer recurrence. Gynecol Oncol 2007; 104:529–534. Pallardy A, Bodet-Milin C, Oudoux A, Campion L, Bourbouloux E, Sagan C, et al. Clinical and survival impact of FDG PET in patients with suspicion of recurrent cervical carcinoma. Eur J Nucl Med Mol Imaging 2010; 37:1270–1278. Cetina L, Serrano A, Cantu´-de-Leo´n D, Pe´rez-Montiel D, Estrada E, Coronel J, et al. F18-FDG-PET/CT in the evaluation of patients with suspected recurrent or persistent locally advanced cervical carcinoma. Rev Invest Clin 2011; 63:227–235. Grisaru D, Almog B, Levine C, Metser U, Fishman A, Lerman H, et al. The diagnostic accuracy of 18F-fluorodeoxyglucose PET/CT in patients with gynecological malignancies. Gynecol Oncol 2004; 94:680–684. Amit A, Beck D, Lowenstein L, Lavie O, Kedar Z, Israel O. The role of hybrid PET/CT in the evaluation of patients with cervical cancer. Gynecol Oncol 2006; 100:65–69. Kato H, Morioka H, Tsutsui H, Aramaki S, Torigoe T. Value of tumor-antigen (TA-4) of squamous cell carcinoma in predicting the extent of cervical cancer. Cancer 1982; 50:1294–1296. Duk JM, de Bruijn HW, Groenier KH, Hollema H, ten Hoor KA, Krans M, et al. Cancer of the uterine cervix: sensitivity and specificity of serum squamous cell carcinoma antigen determinations. Gynecol Oncol 1990; 39:186–194. Borras G, Molina R, Xercavins J, Ballesta A, Iglesias X. Squamous cell carcinoma antigen in cervical cancer. Eur J Gynaecol Oncol 1991; 13:414–418. Micke O, Bruns F, Scha¨fer U, Prott F-J, Willich N. The impact of squamous cell carcinoma (SCC) antigen in patients with advanced cancer of uterine cervix treated with (chemo-) radiotherapy. Anticancer Res 2005; 25 (3A):1663–1666. Micke O, Prott F, Scha¨fer U, Tangerding S, Po¨tter R, Willich N. The impact of squamous cell carcinoma (SCC) antigen in the follow-up after radiotherapy in patients with cervical cancer. Anticancer Res 2000; 20:5113–5115. Gadducci A, Tana R, Fanucchi A, Genazzani AR. Biochemical prognostic factors and risk of relapses in patients with cervical cancer. Gynecol Oncol 2007; 107:S23–S26. Hirakawa M, Nagai Y, Inamine M, Kamiyama K, Ogawa K, Toita T, et al. Predictive factor of distant recurrence in locally advanced squamous cell carcinoma of the cervix treated with concurrent chemoradiotherapy. Gynecol Oncol 2008; 108:126–129. Jeong BK, Choi DH, Huh SJ, Park W, Bae DS, Kim B-G. The role of squamous cell carcinoma antigen as a prognostic and predictive factor in carcinoma of uterine cervix. Radiat Oncol J. 2011; 29:191–198.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Diagnostic value of 18F-FDG-PET or PET-CT in recurrent cervical cancer: a systematic review and meta-analysis Yanxia Chua, Ai Zhenga, Fei Wangb, Wei Lina, Xiaoyun Yanga, Ling Hana, Yali Chena and Liping Baia Objectives Accurate detection of recurrent cervical cancer remains a clinical difficulty. This study aims to assess the diagnostic value of PET or PET-computed tomography (PET-CT) using 18F-fluorodeoxyglucose (18F-FDG) in recurrent cervical cancer using a meta-analysis. Study design All published studies in English evaluating the diagnostic value of PET or PET-CT in detecting recurrent cervical cancer were collected. The methodological quality of the included studies was evaluated. Pooled sensitivity, specificity, diagnostic odds ratio, and summary receiver-operating characteristic curves were obtained using statistical software. Twenty studies were included in the meta-analysis. Results The meta-analysis showed that the pooled sensitivity and specificity of PET and PET-CT to detect distant metastasis in recurrent cervical cancer were 0.87 [95% confidence interval (CI): 0.80–0.92] and 0.97 (95% CI: 0.96–0.98), respectively. The pooled sensitivity and specificity
Introduction Despite the advances in screening and prevention, cervical cancer remains a major threat to women’s health and accounts for B150 000 deaths annually worldwide [1]. The recurrence rate of uterine cervical cancer is reported to be 6.5% after surgery and 26.2% after radiation therapy alone [2]. Once the primary treatment has failed, the opportunity for secondary cure is minimal. Morbidity of salvage therapy is usually high, and 5-year survival rates after recurrence are low [3]. Early detection of recurrence may influence survival. Therefore, attempts to improve surveillance after treatment could lead to earlier detection of relapse, and precise assessment of the recurrent status could improve outcome. Radiological techniques, such as intravenous urography, ultrasonography, computed tomography (CT), and MRI, are used to detect recurrent cervical cancer [4]. However, these imaging modalities exhibit limitations in differentiating recurrent tumor from postoperative or radiation fibrosis and in detecting metastatic normal-sized lymph nodes and extrapelvic metastases [5]. Whole-body PET using the glucose analog 18F-fluorodeoxyglucose (18F-FDG) is a unique imaging modality that is based on the differential metabolic characteristics of malignant and normal tissues. Locally recurrent disease is often difficult to detect on pelvic examination because of the thickening of soft-tissue structures after radiation and/ or surgery. Detection of recurrent disease in the pelvis using c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0143-3636
for local regional recurrence were 0.82 (95% CI: 0.72–0.90) and 0.98 (95% CI: 0.96–0.99), respectively. Conclusion 18F-FDG-PET and PET-CT are valuable methods for the assessment of recurrent cervical c 2014 Wolters cancer. Nucl Med Commun 35:144–150 Kluwer Health | Lippincott Williams & Wilkins. Nuclear Medicine Communications 2014, 35:144–150 Keywords: cervix, meta-analysis, recurrence, tomography a Department of Obstetric and Gynecology, West China Second University Hospital, Sichuan University, Chengdu, China and bDepartment of Obstetric and Gynecology, People’s Hospital of Shanxi Province, 256 Xi’an Friendship West Road, Shanxi 710068, China
Correspondence to Ai Zheng, MD, Department of Obstetric and Gynecology, West China Second University Hospital, Sichuan University, Section 3, 20S. Renmin Road, Chengdu, Sichuan 610041, China Tel: + 86 28 85503960; fax: + 86 28 85559065; e-mail: [email protected] Received 10 September 2013 Revised 27 September 2013 Accepted 1 October 2013
MRI and CT is problematic in this setting because discrete masses are usually not present. PET or PET-CT has a high diagnostic value in the detection of recurrent cervical cancer [6], but the number of cases is generally insufficient. Moreover, the results among various studies are significantly different; thus, the assessed diagnostic value of 18F-FDGPET (PET-CT) is not consistent. Meta-analysis is an important tool for precisely defining the effect of selected genetic polymorphisms on the risk of disease and identifying potentially important sources of between-study heterogeneity. Thus, the present meta-analysis aims to explore the diagnostic value of 18 F-FDG-PET or PET-CT in assessing distant metastasis and local regional recurrence in cervical cancer. The results of this study may serve as a reference for future cost-effectiveness studies.
Materials and methods Data sources
A comprehensive computer literature search of abstracts of studies on human subjects was conducted to identify articles that focused on the diagnostic performance of PET or PET-CT in detecting recurrent cervical cancer. The MEDLINE and EMBASE databases were searched from January 1995 to May 2013 using the following keywords: (‘PET’ or ‘positron emission tomography’ or ‘FDG’ or ‘fluorodeoxyglucose’), (‘cervical carcinoma’ DOI: 10.1097/MNM.0000000000000026
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Value of
or ‘cervical cancer’ or ‘carcinoma of cervix’), and (‘sensitivity’ or ‘specificity’ or ‘false negative’ or ‘false positive’ or ‘diagnosis’ or ‘detection’ or ‘accuracy’). The list of articles was supplemented with extensive crosschecking of the reference lists of all retrieved articles. Study selection
The following studies were included: (a) articles that were reported in English; (b) studies that used 18F-FDGPET or PET-CT to identify or characterize recurrent cervical cancer; (c) studies that used histopathologic analysis and/or close clinical and imaging follow-up for at least 6 months as the reference standard; (d) studies that presented sufficient data for per-patient or per-lesion statistics to calculate the true-positive and false-negative values; and (e) studies that included 10 or more patients. The following studies were excluded: (a) those that used other radiotracers; (b) those that used a small sample size; and (c) for duplicate studies, those published earlier or had a smaller sample size. Data abstraction and quality assessment
Two reviewers independently screened the titles, abstracts, and full texts to identify eligible studies. Disagreements between the reviewers were resolved by consensus. Data retrieved from the studies included a description of the study population (age), study design (whether prospective, retrospective, or unknown), patient selection (whether consecutive or nonconsecutive), and interpretation of test results (whether blinded or nonblinded). The numbers of true-positive, false-negative, false-positive, and true-negative results in the detection of recurrent cervical cancer were extracted per patient or per lesion. Table 1
18
F-FDG-PET in cervical cancer Chu et al. 145
Statistical analysis
All statistical analyses were performed using the Meta-Disc version 1.4 software package (Unidad de Bioestadı´stica Clı´nica del Hospital Ramo´n y Cajal de Madrid, Madrid, Spain). For each study, we constructed 2 2 contingency tables wherein all subjects were classified to have positive or negative 18F-FDG-PET or PET-CT results. We calculated the true-positive rate (TPR; sensitivity) and false-positive rate (FPR; 1 – specificity). We also calculated the summary receiver-operating characteristic curves and the area under the curve. A graph was plotted from the TPR and FPR points. The summary receiver-operating characteristic curve was plotted over these points to form a smooth curve, which can be achieved using the regression model proposed by Moses et al. [7]. Two statistical methods were used to precisely assess between-study heterogeneity. The w2-based Q statistic test (Cochran’s Q statistic) was used to test for heterogeneity, and the I2 statistic was calculated to quantify the proportion of total variation due to heterogeneity [8]. The I2 index was calculated to assess between-study heterogeneity. I2 values of 25, 50, and 75% were used as evidence of low, moderate, and high heterogeneity, respectively. If moderate or high heterogeneity was calculated, the random-effects model was used to pool the results; otherwise, the fixed-effects model was used when I2 was less than 50%.
Results Literature search and study design characteristics
A total of 20 publications met the inclusion criteria [6,9–27]. For three of the 20 studies, the investigators were blinded to the diagnostic findings from other modalities (e.g. CT and MRI), but the investigators for the remaining 17 studies were either not blinded or their
Main characteristics of the included studies
References Park et al. [9] Sun et al. [10] Belhocine et al. [11] Nakamoto et al. [12] Ryu et al. [13] Havrilesky et al. [14] Lai [15] Yen et al. [16] Chang et al. [17] Grisaru et al. [26] Sakurai et al. [18] Amit et al. [27] Sironi et al. [6] Chung et al. [23] Van der Veldt et al. [19] Kitajima et al. [20] Mittra et al. [21] Pallardy et al. [24] Cetina et al. [25] Lee et al. [22]
Mean age (range)
Patient selection
Blind
Evaluable patients or lesion
53 (ND) ND 52 (38–66) 52 (26–82) 51 (31–78) 42 (28–69) 51 (25–87) 51 (25–86) 54 (35–76) 56 (20–85) 56 (27–80) 50 (31–71) 28–69 53 (32–77) 41 (27–61) 58 (37–78) 50 (28–87) 46 (35–81) 47 ND
ND ND ND ND ND ND Consecutive ND Consecutive Consecutive Consecutive ND Consecutive ND ND Consecutive ND ND ND ND
ND ND ND ND ND ND ND ND ND Yes ND ND Yes ND ND Yes ND ND ND ND
36 20 60 20 249 29* 400* 550* 27 12 54 28 12 32 39 52 30 40 16 51
Recurrent number (%) 19 18 28 5 31 22 67 94 18 10 47 7 5 28 25 25 24 33 12 37
(53) (90) (47) (25) (12) (76) (17) (17) (67) (83) (87) (25) (42) (88) (64) (48) (80) (83) (75) (73)
Noninvasive modalities
Study design
PET PET PET PET PET PET PET PET PET PET/CT PET PET/CT PET/CT PET/CT PET PET/CT;PET PET/CT PET/CT PET/CT PET/CT
ND Retrospectively Retrospectively Retrospectively Retrospectively Retrospectively Retrospectively Prospective ND ND ND ND Retrospectively Retrospectively Retrospectively ND Retrospectively Retrospectively Retrospectively Retrospectively
CT, computed tomography; ND, not documented.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
146
Nuclear Medicine Communications 2014, Vol 35 No 2
blinding status was not defined. Moreover, patient recruitment was consecutive in six studies and not defined in 14 studies. The main characteristics of the included studies are presented in Table 1.
Fig. 1
(a) Park et al. [9] Sun et al. [10] Nakamoto et al. [12] Belhocine et al. [11] Ryu et al. [13] Havrilesky et al. [14] Lai et al. [15] Yen et al. [16] Chang et al. [17] Sakurai et al. [18] Kitajima et al. [20] van der Veldt et al. [19] Grisaru et al. [26] Chung et al. [23] Amit et al. [27] Sironi et al. [6] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25] Lee et al. [22]
Main results
The pooled sensitivity and specificity of PET or PET-CT were 0.92 [95% confidence interval (CI): 0.89–0.94] and 0.90 (95% CI: 0.91–0.93), respectively. The pooled sensitivity and specificity of PET or PET-CT for the assessment of distant metastasis in cervical cancer were 0.87 (0.80–0.92) and 0.97 (0.96–0.98), respectively. The pooled sensitivity and specificity for the assessment of local regional recurrence were 0.82 (0.72–0.90) and 0.98 (0.96–0.99), respectively. The results are also shown in Figs 1–3. 0
When all 12 PET studies were considered, the pooled sensitivity and specificity were 0.91 (0.88–0.94) and 0.92 (0.90–0.94), respectively. In subgroup analysis, the pooled sensitivity and specificity for the assessment of distant metastasis were 0.84 (0.74–0.91) and 0.92 (0.89–0.94), respectively. The pooled sensitivity and specificity for the assessment of local regional recurrence were 0.84 (0.74–0.91) and 0.92 (0.89–0.94), respectively (Table 2).
0.2
0.4
0.8
1
Pooled sensitivity = 0.92 (0.89 − 0.94) χ 2 = 20.49; d.f .= 19 (P = 0.3654) Inconsistency (I 2 ) = 7.3%
Sensitivity
(b) Park et al. [9] Sun et al. [10] Nakamoto et al. [12] Belhocine et al. [11] Ryu et al. [13] Havrilesky et al. [14] Lai et al. [15] Yen et al. [16] Chang et al. [17] Sakurai et al. [18] Kitajima et al. [20] van der Veldt et al. [19] Grisaru et al. [26] Chung et al. [23] Amit et al. [27] Sironi et al. [6] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25] Lee et al. [22]
When all eight PET-CT studies were considered, the pooled sensitivity and specificity were 0.94 (0.90–0.97) and 0.84 (0.75–0.91), respectively. The pooled sensitivity and specificity for the assessment of distant metastasis were 0.90 (0.80–0.96) and 0.99 (0.98–1.00), respectively. The pooled sensitivity and specificity for the assessment of local regional recurrence were 0.83 (0.69–0.93) and 0.96 (0.89–0.99), respectively (Table 2).
0
0.2
0.4
0.8
0.6
Specificity (95% CI) 0.94 (0.73 − 1.00) 1.00 (0.16 − 1.00) 0.60 (0.32 − 0.84) 0.77 (0.46 − 0.95) 0.76 (0.70 − 0.82) 0.87 (0.60 − 0.98) 0.98 (0.96 − 0.99) 0.98 (0.97 − 0.99) 0.78 (0.40 − 0.97) 0.57 (0.18 − 0.90) 0.78 (0.58 − 0.91) 0.93 (0.66 − 1.00) 1.00 (0.16 − 1.00) 0.81 (0.58 − 0.95) 1.00 (0.40 − 1.00) 1.00 (0.54 − 1.00) 0.93 (0.76 − 0.99) 0.71 (0.29 − 0.96) 0.50 (0.07 − 0.93) 0.71 (0.42 − 0.92)
Pooled specificity = 0.91 (0.90 − 0.93) χ 2 = 155.39; d.f .= 19 (P = 0.0000) 1 Inconsistency (I 2 )= 87.8 %
Specificity
(c)
SROC curve 1
Publication bias
Symmetric SROC AUC = 0.9608 SE (AUC) = 0.0080 Q∗ = 0.9057 SE (Q∗ ) = 0.0117
0.9 0.8 0.7 Sensitivity
Begg’s funnel and Egger’s test were performed to assess the publication bias in this meta-analysis. The shape of the Funnel plots did not reveal significant evidence of asymmetry, and the P value of Egger’s test was 0.243 (P > 0.05), providing statistical evidence of the symmetry of the Funnel plots (Fig. 4). Thus, the results above suggested that publication bias was not evident in this meta-analysis.
0.6
Sensitivity (95% CI) 1.00 (0.81 − 1.00) 0.89 (0.65 − 0.99) 1.00 (0.48 − 1.00) 1.00 (0.86 − 1.00) 0.90 (0.74 − 0.98) 0.86 (0.57 − 0.98) 0.91 (0.82 − 0.97) 0.89 (0.81 − 0.95) 0.94 (0.73 − 1.00) 0.91 (0.80 − 0.98) 0.80 (0.59 − 0.93) 0.92 (0.74 − 0.99) 1.00 (0.69 − 1.00) 0.90 (0.74 − 0.98) 0.86 (0.42 − 1.00) 0.83 (0.36 − 1.00) 0.92 (0.74 − 0.99) 0.96 (0.78 − 1.00) 1.00 (0.74 − 1.00) 0.97 (0.86 − 1.00)
0.6 0.5 0.4 0.3 0.2
Discussion Recent studies in patients with recurrent cervical cancers have yielded important results in terms of reduced mortality and prolonged survival with the use of multimodality treatments such as radical resection in combination with intraoperative high-dose brachytherapy [6,28]. However, these promising therapeutic methodologies entail more effective diagnostic procedures that may either help detect recurrence sufficiently early to give the patient the best chance for successful salvage therapy or confirm complete remission. Hence, unnecessary invasive treatments are avoided.
0.1 0 0
0.2
0.4
0.6
0.8
1
1−specificity
(a) Sensitivity, (b) specificity, and (c) summary receiver-operating characteristic (SROC) curves of 18F-FDG-PET. CI, confidence interval; 18 F-FDG, 18F-fluorodeoxyglucose.
Squamous cell carcinoma (SCC) antigen has been proposed as a tumor marker for the early diagnosis of recurrence in and monitoring of patients with cervical
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Value of
18
F-FDG-PET in cervical cancer Chu et al. 147
Fig. 2
(a)
Sensitivity (95% CI) Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0
0.2
0.4 0.6 Sensitivity
0.8
0.91 0.67 0.94 0.67 0.75 0.86 0.87 0.96 1.00
(0.76 − 0.98) (0.09 − 0.99) (0.70 − 1.00) (0.38 − 0.88) (0.35 − 0.97) (0.42 − 1.00) (0.70 − 0.96) (0.78 − 1.00) (0.16 − 1.00)
Pooled sensitivity = 0.87 (0.80 − 0.92) χ2 = 9.56; d.f.= 8 (P= 0.2972) 1 Inconsistency (I 2 ) = 16.3 %
(b)
Specificity (95% CI) Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0
0.2
0.4 0.6 Specificity
(c)
0.8
0.95 1.00 0.82 0.98 1.00 1.00 0.99 0.95 0.67
(0.93 − 0.96) (0.03 − 1.00) (0.48 − 0.98) (0.95 − 0.99) (0.77 − 1.00) (0.80 − 1.00) (0.98 − 1.00) (0.74 − 1.00) (0.09 − 0.99)
Pooled specificity = 0.97 (0.96 − 0.98) χ2 = 31.91; d.f .= 8 (P = 0.0001) Inconsistency (I 2 ) = 74.9 % 1
SROC curve
1
Symmetric SROC AUC = 0.9562 SE (AUC) = 0.0202 Q∗ = 0.8990 SE (Q∗ ) = 0.0285
0.9 0.8
Sensitivity
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
0.2
0.4 0.6 1−specificity
0.8
(a) Sensitivity, (b) specificity, and (c) summary receiver-operating characteristic (SROC) curves of 18 F-FDG, 18F-fluorodeoxyglucose.
1
18
F-FDG-PET. CI, confidence interval;
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
148
Nuclear Medicine Communications 2014, Vol 35 No 2
Fig. 3
(a)
Sensitivity (95% CI) Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0
0.2
0.6 0.4 Sensitivity
0.8
1
Ryu et al. [13] Havrilesky et al. [14] Chang et al. [17] Kitajima et al. [20] van der Veldt et al. [19] Amit et al. [27] Kitajima et al. [20] Mittra et al. [21] Cetina et al. [25]
0.2
0.4 0.6 Specificity
(c)
0.8
(0.19 − 0.99) (0.40 − 0.97) (0.48 − 1.00) (0.05 − 0.85) (0.64 − 1.00) (0.17 − 0.77) (0.59 − 1.00) (0.66 − 1.00) (0.69 − 1.00)
Pooled sensitivty = 0.82 (0.72 − 0.90) 2 = 23.24; d.f. = 8 (P = 0.0031) Inconsistency (I 2) = 65.6 %
(b)
0
0.75 0.78 1.00 0.40 0.92 0.45 1.00 0.93 1.00
Specificity (95% CI) 1.00 (0.97 − 1.00) 0.90 (0.68 − 0.99) 1.00 (0.85 − 1.00) 1.00 (0.92 − 1.00) 0.96 (0.81 − 1.00) 0.94 (0.71 − 1.00) 1.00 (0.92 − 1.00) 0.93 (0.76 − 0.99) 0.67 (0.09 − 0.99)
Pooled specificity = 0.98 (0.96 − 0.99) 2 = 18.81; d.f. = 8 (P = 0.0159) 2 1 Inconsistency (I ) = 57.5 %
SROC curve
1
Symmetric SROC AUC = 0.9615 SE (AUC) = 0.0208 Q∗ = 0.9068 SE (Q∗ ) = 0.0308
0.9 0.8
Sensitivity
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
0
0.2
0.4 0.6 1−specificity
0.8
(a) Sensitivity, (b) specificity, and (c) summary receiver-operating characteristic (SROC) curves of 18 F-FDG, 18F-fluorodeoxyglucose.
cancer. The sensitivity of SCC ranges from less than 30% for stage I disease to greater than 90% for stage IV disease [29–31]. The SCC-Ag level increased in 28–88%
1
18
F-FDG-PET. CI, confidence interval;
of patients with SCC and was considered the most accurate serologic tumor marker for patients with uterine cervical carcinoma; it is also used as a diagnostic tool in
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Value of
Table 2
18
F-FDG-PET in cervical cancer Chu et al.
149
Pooled sensitivity, pooled specificity, AUC, and test for heterogeneity in the meta-analysis 95% CI
PET or PET-CT Distant metastasis Local regional recurrence PET Distant metastasis Local regional recurrence PET-CT Distant metastasis Local regional recurrence
Heterogeneity (SN)
Pooled sensitivity
Pooled specificity
0.92 0.87 0.82 0.91 0.84 0.81 0.94 0.90 0.83
0.91 0.97 0.98 0.92 0.92 0.99 0.84 0.99 0.96
(0.89–0.94) (0.80–0.92) (0.72–0.90) (0.88–0.94) (0.74–0.91) (0.64–0.92) (0.90–0.97) (0.80–0.96) (0.69–0.93)
(0.90–0.93) (0.96–0.98) (0.96–0.99) (0.90–0.94) (0.89–0.94) (0.97–1.00) (0.75–0.91) (0.98–1.00) (0.89–0.99)
Heterogeneity (SP)
AUC
I2 (%)
PH
I2 (%)
PH
0.9608 0.9562 0.9615 0.9594 0.9431 0.9622 0.9508 0.9626 0.9582
7.3 16.3 65.6 15.7 38 47.7 0 0 80.6
0.3654 0.2972 0.0031 0.2906 0.1679 0.1051 0.5405 0.6115 0.0014
87.8 74.9 57.5 92.1 57.8 54.3 34.4 65.1 58.1
0 0.001 0.0159 < 0.001 0.0503 0.0675 0.1534 0.0353 0.067
AUC, area under the curve; CI, confidence interval; CT, computed tomography; SN, sensitivity; SP, specificity.
that the presence of postoperative adhesions between loops of bowel and adjacent peritoneal surfaces complicated the imaging in the recurrent cervical cancer patients. The CT scan apparatus had difficulty in precisely detecting recurrent lesions, suggesting that interpreted CT images may have limited additional value to PET in detecting recurrent cervical cancer.
Fig. 4
Begg's funnel plot with pseudo 95% confidence limits 6
Log OR
4
2
0 −2 0
0.5
1
1.5
SE of log OR Begg’s funnel plots for assessing the publication bias (PEgger = 0.243). OR, odds ratio.
staging, a cost-effective tool for disease monitoring of therapy, and a tool for the detection of recurrence [32–35]. However, determination of the most suitable clinical regimen is difficult because of the relatively low sensitivity or lack of capacity of localization of SCC-Ag. Furthermore, conventional imaging modalities such as CT have a high diagnostic sensitivity but exhibit limitations in distinguishing between postoperative scars and relapse. PET with the radiolabeled glucose analog 18 F-FDG is a method that is based on the increased glucose metabolism of malignant tumors [16]. This method can indicate the biochemical differences between normal and malignant tissues. The addition of CT to PET provides precise anatomic information; however, whether the precise anatomic information can actually improve the overall diagnostic accuracy of PET is still unclear. In the present meta-analysis, no significant differences were found between the PET results interpreted with CT and those interpreted without the use of CT. This result could be partially attributed to the fact
The pooled estimates of the sensitivity and specificity of 18 F-FDG-PET in the detection of distant metastasis were 84% (95% CI: 74–91%) and 92% (95% CI: 89–94%), respectively. The pooled estimates of the sensitivity and specificity of PET-CT in detecting distant metastasis were 90% (95% CI: 80–96%) and 99% (95% CI: 98–100%), respectively. Therefore, PET-CT had a significantly higher diagnostic accuracy compared with PET in detecting distant metastasis in recurrent cervical cancer patients. However, the accuracy of PET-CT in detecting local regional recurrence was similar to that of 18F-FDG-PET. Nonetheless, several limitations were present in our study. First, patient characteristics are important for diagnostic accuracy; however, the effect of these variables could not be examined because of variation in data presentation or incomplete reporting of data. Second, the majority of the selected studies were retrospective. Therefore, a potential risk for subjective interpretation exists, and the diagnostic accuracy of the study can be lower. Third, selection biases may have occurred because of failure to find comprehensive, relevant articles or unpublished data. This failure could be attributed to the restriction of searching parameters. Moreover, the included articles may have poor quality. Finally, further cost-effectiveness analysis should be conducted on the surveillance techniques for cervical cancer.
Conclusion This meta-analysis is the first to evaluate the diagnostic performance of PET and PET-CT in detecting distant metastasis and local regional recurrence in cervical cancer patients. Our study showed that PET and PET-CT had
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
150
Nuclear Medicine Communications 2014, Vol 35 No 2
high diagnostic accuracy in detecting distant metastasis and local regional recurrence. These findings have clinical implication in terms of providing useful information not only to radiologists in interpreting images but also to gynecologic oncologists in choosing the imaging modality for the management of suspected recurrent cervical cancer patients.
17
Acknowledgements
20
18
19
Conflicts of interest
There are no conflicts of interest.
21
References 1 2
3 4
5
6
7
8 9
10
11
12
13
14
15 16
Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, et al. Cancer statistics, 2008. CA Cancer J Clin 2008; 58:71–96. Halpin TF, Frick HC, Munnell EW. Critical points of failure in the therapy of cancer of the cervix: a reappraisal. Am J Obstet Gynecol 1972; 114:755–764. Waggoner SE. Cervical cancer. Lancet 2003; 361:2217–2225. Walsh JW, Amendola MA, Hall DJ, Tisnado J, Goplerud DR. Recurrent carcinoma of the cervix: CT diagnosis. AJR Am J Roentgenol 1981; 136:117–122. Cunningham JJ, Fuks ZY, Castellino RA. Radiographic manifestations of carcinoma of the cervix and complications of its treatment. Radiol Clin North Am 1974; 12:93–108. Sironi S, Picchio M, Landoni C, Galimberti S, Signorelli M, Bettinardi V, et al. Post-therapy surveillance of patients with uterine cancers: value of integrated FDG PET/CT in the detection of recurrence. Eur J Nucl Med Mol Imaging 2007; 34:472–479. Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med 1993; 12:1293–1316. Higgins J, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21:1539–1558. Park DH, Kim KH, Park SY, Lee BH, Choi CW, Chin SY. Diagnosis of recurrent uterine cervical cancer: computed tomography versus positron emission tomography. Korean J Radiol 2000; 1:51–55. Sun SS, Chen TC, Yen RF, Shen YY, Changlai SP, Kao A. Value of whole body 18F-fluoro-2-deoxyglucose positron emission tomography in the evaluation of recurrent cervical cancer. Anticancer Res 2001; 21 (4B):2957–2961. Belhocine T, Thille A, Fridman V, Albert A, Seidel L, Nickers P, et al. Contribution of whole-body 18FDG PET imaging in the management of cervical cancer. Gynecol Oncol 2002; 87:90–97. Nakamoto Y, Eisbruch A, Achtyes ED, Sugawara Y, Reynolds KR, Johnston CM, et al. Prognostic value of positron emission tomography using F-18-fluorodeoxyglucose in patients with cervical cancer undergoing radiotherapy. Gynecol Oncol 2002; 84:289–295. Ryu SY, Kim MH, Choi SC, Choi CW, Lee KH. Detection of early recurrence with 18F-FDG PET in patients with cervical cancer. J Nucl Med 2003; 44:347–352. Havrilesky LJ, Wong TZ, Secord AA, Berchuck A, Clarke-Pearson DL, Jones EL. The role of PET scanning in the detection of recurrent cervical cancer. Gynecol Oncol 2003; 90:186–190. Lai CH. Management of recurrent cervical cancer. Chang Gung Med J 2004; 27:711–717. Yen TC, See LC, Chang TC, Huang KG, Ng KK, Tang SG, et al. Defining the priority of using 18F-FDG PET for recurrent cervical cancer. J Nucl Med 2004; 45:1632–1639.
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Chang TC, Law KS, Hong JH, Lai CH, Ng KK, Hsueh S, et al. Positron emission tomography for unexplained elevation of serum squamous cell carcinoma antigen levels during follow-up for patients with cervical malignancies. Cancer 2004; 101:164–171. Sakurai H, Suzuki Y, Nonaka T, Ishikawa H, Shioya M, Kiyohara H, et al. FDG-PET in the detection of recurrence of uterine cervical carcinoma following radiation therapy: tumor volume and FDG uptake value. Gynecol Oncol 2006; 100:601–607. van der Veldt AA, Buist MR, van Baal MW, Comans EF, Hoekstra OS, Molthoff CF. Clarifying the diagnosis of clinically suspected recurrence of cervical cancer: impact of 18F-FDG PET. J Nucl Med 2008; 49:1936–1943. Kitajima K, Murakami K, Yamasaki E, Domeki Y, Kaji Y, Sugimura K. Performance of FDG-PET/CT for diagnosis of recurrent uterine cervical cancer. Eur Radiol 2008; 18:2040–2047. Mittra E, El-Maghraby T, Rodriguez CA, Quon A, McDougall IR, Gambhir SS, et al. Efficacy of 18F-FDG PET/CT in the evaluation of patients with recurrent cervical carcinoma. Eur J Nucl Med Mol Imaging 2009; 36:1952–1959. Lee M, Lee Y, Hwang KH, Choe W, Park CY. Usefulness of F-18 FDG PET/ CT in assessment of recurrence of cervical cancer after treatment. Nucl Med Mol Imaging 2011; 45:111–116. Chung HH, Jo H, Kang WJ, Kim JW, Park N-H, Song Y-S, et al. Clinical impact of integrated PET/CT on the management of suspected cervical cancer recurrence. Gynecol Oncol 2007; 104:529–534. Pallardy A, Bodet-Milin C, Oudoux A, Campion L, Bourbouloux E, Sagan C, et al. Clinical and survival impact of FDG PET in patients with suspicion of recurrent cervical carcinoma. Eur J Nucl Med Mol Imaging 2010; 37:1270–1278. Cetina L, Serrano A, Cantu´-de-Leo´n D, Pe´rez-Montiel D, Estrada E, Coronel J, et al. F18-FDG-PET/CT in the evaluation of patients with suspected recurrent or persistent locally advanced cervical carcinoma. Rev Invest Clin 2011; 63:227–235. Grisaru D, Almog B, Levine C, Metser U, Fishman A, Lerman H, et al. The diagnostic accuracy of 18F-fluorodeoxyglucose PET/CT in patients with gynecological malignancies. Gynecol Oncol 2004; 94:680–684. Amit A, Beck D, Lowenstein L, Lavie O, Kedar Z, Israel O. The role of hybrid PET/CT in the evaluation of patients with cervical cancer. Gynecol Oncol 2006; 100:65–69. Kato H, Morioka H, Tsutsui H, Aramaki S, Torigoe T. Value of tumor-antigen (TA-4) of squamous cell carcinoma in predicting the extent of cervical cancer. Cancer 1982; 50:1294–1296. Duk JM, de Bruijn HW, Groenier KH, Hollema H, ten Hoor KA, Krans M, et al. Cancer of the uterine cervix: sensitivity and specificity of serum squamous cell carcinoma antigen determinations. Gynecol Oncol 1990; 39:186–194. Borras G, Molina R, Xercavins J, Ballesta A, Iglesias X. Squamous cell carcinoma antigen in cervical cancer. Eur J Gynaecol Oncol 1991; 13:414–418. Micke O, Bruns F, Scha¨fer U, Prott F-J, Willich N. The impact of squamous cell carcinoma (SCC) antigen in patients with advanced cancer of uterine cervix treated with (chemo-) radiotherapy. Anticancer Res 2005; 25 (3A):1663–1666. Micke O, Prott F, Scha¨fer U, Tangerding S, Po¨tter R, Willich N. The impact of squamous cell carcinoma (SCC) antigen in the follow-up after radiotherapy in patients with cervical cancer. Anticancer Res 2000; 20:5113–5115. Gadducci A, Tana R, Fanucchi A, Genazzani AR. Biochemical prognostic factors and risk of relapses in patients with cervical cancer. Gynecol Oncol 2007; 107:S23–S26. Hirakawa M, Nagai Y, Inamine M, Kamiyama K, Ogawa K, Toita T, et al. Predictive factor of distant recurrence in locally advanced squamous cell carcinoma of the cervix treated with concurrent chemoradiotherapy. Gynecol Oncol 2008; 108:126–129. Jeong BK, Choi DH, Huh SJ, Park W, Bae DS, Kim B-G. The role of squamous cell carcinoma antigen as a prognostic and predictive factor in carcinoma of uterine cervix. Radiat Oncol J. 2011; 29:191–198.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
