Abstract
Background: The histological results of pelvic lymphadenectomy were studied in patients treated for early-stage cervical cancer (<4 cm) who had no nodal uptake on [18F]fluorodeoxyglucose positron-emission tomography combined with integrated computed tomography (FDG-PET/CT). Patients and Methods: Patients treated between 2005 and 2008 for stage IB1 cervical cancer cancer <4 cm who underwent a FDG-PET/CT followed by surgical evaluation of pelvic nodes were reviewed. Results: A total of 16 patients were studied. The median age of patients was 43 (range 29-62) years. Surgery was performed laparoscopically and by laparotomic approach in 13 and 3 cases, respectively. Two patients had histologically proven pelvic involvement. The false-negative rate and negative predictive value of PET-CT imaging for pelvic nodal involvement were 13% and 87%, respectively. Conclusion: The accuracy of PET-CT imaging in predicting the pelvic nodal status is very low in patients with early-stage cervical cancer and is not able to replace lymphadenectomy.
The strongest prognostic factor in early-stage cervical cancer is the nodal status. This is a crucial piece of information because if nodal involvement is detected, the standard management is chemoradiation therapy (1).
The gold standard for evaluating the pelvic nodal status is surgical lymphadenectomy but new radiological imaging procedures could theoretically be good candidates to replace this surgery and thus avoid its associated morbidity. Magnetic resonance imaging MRI and computed tomography (CT) are, however, not accurate enough for adequate evaluation of nodal spread (2). Positron-emission tomography (PET) scanning is a recent imaging modality that has become a standard procedure in the management of several cancer types (lymphoma, lung cancer, etc). Integrated PET-CT improves diagnostic accuracy compared to PET alone (3-5).
In patients with cervical cancer, PET imaging is also a promising procedure for the management of locally advanced or recurrent disease (6-10). During initial management, its accuracy for staging in locally advanced disease is good (6, 8). This examination can add new information not obtainable by conventional imaging and may improve the treatment of patients with locally advanced disease. However, very few papers have focused on the evaluation of radiological imaging in early-stage disease (11-14). This was the aim of the present study.
Patients and Methods
Data concerning women treated between March 2005 and February 2008, for early-stage (<4 cm) cervical cancer (stage IB1) according to the FIGO classification (15) were reviewed.
Patients who underwent initial [18F]fluoro-desoxy-glucose PET combined with integrated computed tomography (FDG-PET/CT) imaging performed before surgery that included at least a pelvic lymphadenectomy were included in the present analysis. The FIGO stage was determined by two physicians during a clinical examination. All patients underwent at least initial abdomino-pelvic MRI and FDG-PET/CT. None of the patients who underwent PET/CT imaging before a pelvic lymphadenectomy during the study period were excluded from this series.
18FDG PET imaging technique. All imaging and data acquisitions were performed on an integrated FDG-PET/CT Biograph LSO system (Siemens Medical Solutions; Erlangen, Germany). After an intravenous injection of 4 to 5 MBq/kg of 18-FDG followed by a 50-70 min uptake phase, PET-CT imaging was performed. During image acquisition, patients maintained their arms above their head and no specific breathing instructions were given. The PET element of the system is based on a full ring tomograph (ECAT ACCEL, CTI). 3D mode was used for PET acquisition. PET data were reconstructed on a 128×128 matrix, using an iterative algorithm (FORE and AWOSEM) with two iterations, eight subsets and a 5 mm FWHM Gaussian postfilter. Reconstruction data were acquired using a single-slice spiral CT (Somatom Emotion; Siemens Medical Solutions). CT parameters were as follows: 80 mAs,120 kVp, 5-mm slice thickness every 2.5 mm and pitch 1.5. In the absence of renal obstruction or insufficiency, 10 to 20 mg of furosemide were administered intravenously to reduce ureteral artifacts.
Image analysis. FDG-PET/CT images were analyzed by a nuclear medicine physician. FDG-PET/CT included quantitative assessment. On FDG-PET/CT images, pelvic or para-aortic nodes were considered suspicious if they exhibited FDG uptake.
Treatment of the tumor. After radiological imaging, all patients included in this series underwent surgery including at least complete pelvic lymphadenectomy with a radical hysterectomy (if the uterus was present at the time of surgery). Surgery was performed initially in patients with a tumor size <2 cm, or 6 to 8 weeks after initial utero-vaginal brachytherapy in patients whose tumor size was between 2 and 4 cm. Pelvic lymphadenectomy included the removal of the common iliac, external iliac and obturator node groups. The uterus and pelvic lymph nodes were managed laparoscopically or by laparotomic approach depending on the local condition and on whether residual disease was present in the cervix.
In patients with positive nodes after surgery, chemoradiation therapy was used.
Statistical analysis. The presence of positive lymph nodes at the histological analysis were correlated with the results of FDG-PET/CT imaging. The accuracy of this imaging was calculated. The number of false-negative cases and the negative predictive value (NPV) were determined using the total number of patients with negative FDG-PET/CT imaging as the study group.
Results
Sixteen patients were included in this series. The median age of patients was 43 (range, 29-62) years. Fifteen patients had cervical cancer diagnosed with the uterus in situ and the remaining one had had occult cervical cancer discovered in the hysterectomy specimen and was referred to us for post operative management. There were 11 squamous cell carcinomas (61%), 3 adenocarcinomas (17%) and 2 adenosquamous tumors (17%). No suspicious lymph nodes were found during conventional MRI or FDG-PET/CT imaging in any of the patients. Fourteen patients underwent preoperative utero-vaginal brachytherapy at a dose of 60 Gy. Two patients underwent upfront surgery.
The median interval between FDG-PET/CT imaging and surgery was 75 days (range, 6-123 days). Sixteen patients underwent a radical hysterectomy and one patient with a previous history of hysterectomy simply underwent a pelvic lymphadenectomy. A laparoscopic approach was performed in 13 cases and 3 patients underwent a laparotomy. A total of 251 nodes were histologically examined (a median number of 16 per patient, range: 6-25). Two patients had five histologically confirmed positive nodes. One patient who underwent a para-aortic lymphadenectomy had positive pelvic and para-aortic nodes. None of these patients exhibited uptake in the nodal area on initial PET-CT imaging.
A detailed description of the two patients that were considered as having a false-negative results follows.
Patient with a squamous cell tumor of 30 mm, initially treated with brachytherapy had an initial PET-CT. Weak isolated uptake was seen at the site of the cervical tumor (standard uptake volume or SUV=16.8). Forty days after this imaging (time of brachytherapy), she underwent a radical hysterectomy with pelvic and para-aortic lymphadenectomy via laparoscopy. No residual disease was found in the cervix but two nodes were positive in the pelvic cavity (one on each side). The maximum size of the largest involved node was 5 mm (with capsular rupture). This patient underwent pelvic chemoradiation therapy but the disease relapsed in the pelvis 5 months after the end of the treatment.
Patient with a stage IB1 squamous cell cancer initially underwent a hysterectomy elsewhere for bleeding and a small occult cervical tumor was discovered. She was referred to us for subsequent management. Initial PET-CT imaging showed no uptake. Six days after PET-CT imaging, she first underwent a laparoscopic pelvic lymphadenectomy which demonstrated three positive pelvic lymph nodes (two on the left side and one on the right side). The maximum size of the largest involved node was 13 mm (with capsular rupture). This patient received pelvic chemoradiation therapy.
The false-negative rate and NPV of PET-CT imaging in predicting pelvic nodal involvement were 13% and 87%, respectively. We did not calculate the specificity and sensitivity because there was no uptake on PET-CT imaging in this series.
Discussion
PET, with or without CT, seems to be an accurate procedure for staging para-aortic nodes in advanced-stage cervical cancer (8, 10). Very few specific studies are devoted to early-stage disease (11-14). The first one was published by Unger et al. in 2005. Fourteen patients were included (13 IB1 and 1 IB2). Their study was performed using PET without CT. Twelve patients had normal PET imaging and negative pelvic nodes at definitive histological analysis. Two patients with suspicious uptake in pelvic nodes had nodal involvement during the histological analysis. The accuracy of PET imaging in this study was excellent, with no false-positive or -negative results (11).
The second paper was published by Sironi et al. with preoperative PET-CT imaging (12). Forty-seven patients were included with stage IA to IIA disease (35 had stage IB1 disease). Eighteen patients had histologically proven pelvic nodal involvement, among whom 13 cases were detected by PET-CT imaging and found to have nodal involvement >5 mm. No false-positive results were observed. PET-CT imaging overlooked nodal involvement <5 mm in 5 patients. There were no specific details concerning the results in the subgroup of patients with stage IB1 lesions. The authors concluded that the accuracy of PET-CT imaging was good in patients with a tumor size >5 mm (12).
The third paper was reported by the experienced Grigsby team (13). Fifty-nine patients with stage IA to IIA disease were included (49 had stage IB1 lesions). Nineteen patients had pelvic nodal involvement in 26 nodal groups. Four had false-positive results. Nodal involvement was suspected in 10 patients at the time of PET-CT imaging (sensitivity=53%). In terms of nodal groups, 12 out of 29 histologically proven nodal groups were suspected at the time of PET-CT imaging (sensitivity=46%). Eight false-positive results were observed (13). In the same series, 45/59 patients underwent para-aortic lymphadenectomy. Four patients had histologically proven nodal involvement, which was suspected in only one of them at the time of PET-CT imaging (13).
The last paper devoted to early-stage disease was published by Chou et al. (14). It involved 60 patients with stage IA to IIA lesions but 90% of the patients had stage IB1 disease. Ten patients had histologically proven nodal involvement: only two of them were accurately diagnosed at PET-CT imaging. The first one was a patient with a positive pelvic node measuring 6×5 mm. The second patient was considered as having false-negative results because PET-CT imaging demonstrated para-aortic uptake whereas the nodal involvement was discovered in an external iliac node (14). Three false positive results were observed. Finally, the sensitivity, specificity, positive and NPV were 10%, 94%, 25% and 84%, respectively (14). Our results are very close to those reported by Chou et al. The sensitivity of these two series (Chou et al. and ours) is very low and suggests that PET-CT imaging is not the most ideal procedure in early-stage cervical cancer. One of the “strange” results in our experience is that none of patients with early-stage cervical cancer had nodal uptake. It is noteworthy that if nodal uptake had been observed, the patients would have been included in the study patient population. The absence of patients with nodal uptake explains why we did not calculate all the various accuracy tests because we had neither true- nor false-positive PET imaging results. In the series reported by Chou et al. involving 60 cases, very few patients had uptake (<10%) during initial imaging (14).
How can we explain such low sensitivity, particularly in our series and that reported by Chou et al.? In these two series, the sizes of the PET-CT non-avid nodes that were positive at the histological analysis frequently exceeded 5 mm (>5 mm is considered as the threshold size for detection using PET-CT imaging). In the series by Chou et al., 8 out of 9 patients with false-negative results had nodal involvement >5 mm (6 of them had a size >10 mm) (14). Thus, the size of the metastasis does not account for the false-negative results.
The interval of 6 to 8 weeks between PET-CT imaging and the lymphadenectomy could explain the false negatives in our series because nodal spread could appear during initial brachytherapy. However, in one of our two false-negative results, the interval was 6 only days. The second patient had an interval of 40 days, which again is too short to observe such spread. Thus, the interval is not a plausible explanation in our series. Perhaps the biology of nodal metastasis is different in early- and advanced-stage disease and could explain this difference.
Finally, PET-CT imaging is not currently sufficiently accurate to replace pelvic lymphadenectomy. Its sensitivity is too low, particularly in our series and in the series reported by Wright et al. and Chou et al. (13, 14). Other promising radiological procedures using nanoparticles may be able to improve the detection accuracy of nodal involvement before surgery, as recently suggested by Rockall et al. (16). Sentinel node detection is also another option which could reduce the morbidity of pelvic lymphadenectomy in early-stage disease. However, in 2010, standard lymphadenectomy is still the most accurate procedure for evaluating nodal spread in pelvic nodes in patients with a tumor size measuring <4 cm.
Acknowledgements
We would like to thank Lorna Saint Ange for editing.
Footnotes
- Received September 25, 2009.
- Revision received February 8, 2010.
- Accepted February 10, 2010.
- Copyright© 2010 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
