Abstract
This study aimed to assess the value of dual-timepoint 18F-FDG PET/CT in the prediction of lymph node (LN) status in patients with invasive vulvar cancer (VC) scheduled for inguinofemoral LN dissection. Methods: From April 2013 to July 2015, all consecutive patients with VC scheduled for inguinofemoral LN dissection were prospectively enrolled. All patients underwent a preoperative whole-body 18F-FDG PET/CT scan at 1 h (standard examination) and an additional scan from T11 to the groins at 3 h (delayed examination) after 18F-FDG injection. On both scans, each groin was visually scored 0 or 1 concerning 18F-FDG LN uptake relative to background. Semiquantitative analysis included SUVmax and the corresponding retention index of SUVmax, measured on both scans. The optimal cutoff value of these parameters was defined using a receiver-operating-characteristic analysis. Histopathology was the standard of reference. Results: Thirty-three patients were included, with a total of 57 groins dissected and histologically evaluated. At histopathology, 21 of 57 (37%) groins contained metastatic LNs. Concerning visual score, sensitivity, specificity, negative predictive value, positive predictive value, and accuracy were 95.2%, 75%, 96.4%, 69%, and 82.5% on standard scanning and 95.2%, 77.8%, 96.6%, 71.4%, and 84.2% on delayed scanning, respectively. At receiver-operating-characteristic analysis, sensitivity and specificity were 95.2% and 77.8% on standard and delayed 18F-FDG PET/CT for an SUVmax cutoff of greater than 1.32 and 1.88, respectively, and 95.2% and 80% for a retention index of SUVmax cutoff of greater than 0. Conclusion: Standard 18F-FDG PET/CT is an effective preoperative imaging method for the prediction of LN status in VC, allowing the prediction of pathologically negative groins and thus the selection of patients suitable for minimally invasive surgery. Delayed 18F-FDG PET/CT did not improve the specificity and the positive predictive value in our series. Larger studies are needed for a further validation.
Invasive vulvar carcinoma (VC) is an uncommon gynecologic tumor, with an incidence of 2.4 new cases/100,000 women per year (1). The pattern of dissemination of VC is mainly lymphatic, with prevalent involvement of the groins, whereas hematogenous spread is rare (2). Thus, the most important prognostic factor is the presence of metastatic lymph nodes (LNs) in the groins (3). In fact, the 5-y survival rate decreases from 94.7% when locoregional LNs are negative to 62% when they contain metastases (4). Therefore, accurate preoperative LN staging is critical to customize the extent of groin surgery and to select patients suitable for minimally invasive procedures, thus avoiding unnecessary inguinofemoral lymph node dissection (IFLD), which is associated with a high morbidity and worse quality of life. In recent years, PET/CT using the glucose analog 18F-FDG has been used increasingly for the evaluation of LN status in gynecologic malignancies (5), but only recently has it been recommended in VC (6). However, because of the low incidence of VC there are few studies in small series on the diagnostic accuracy of 18F-FDG PET/CT in the detection of metastatic LNs in VC; the reported sensitivity ranges from 67% to 92% and the specificity from 91% to 95% per groin (7,8). Recently, dual-timepoint (DTP) or dual-phase 18F-FDG PET/CT has been suggested as a means for detecting metastatic LNs in several gynecologic cancers (9), but its usefulness in VC has not yet been evaluated. In particular, DTP 18F-FDG PET/CT requires 2 acquisitions after a single injection of the radiotracer, that is, standard images (1 h after injection) followed by delayed images (3 h after injection) of the body region under assessment. The rationale is that malignant cells, compared with benign cells, usually show increased 18F-FDG uptake retention on delayed-timepoint imaging because of the high glycolysis activity (10).
The aim of this prospective study was to investigate the value of DTP 18F-FDG PET/CT for the assessment of LN status in patients with VC scheduled for IFLD.
MATERIALS AND METHODS
Patients and Study Design
The Institutional Review Board approved this longitudinal prospective monocenter study, and all patients signed a written informed consent form. Between April 2013 and July 2015, all consecutive patients with primary histologically proven invasive VC (i.e., depth of stromal invasion > 1 mm) referred to the Division of Gynecologic Oncology at A. Gemelli Hospital were evaluated with clinical examination and conventional imaging (6). The surgical plan was traced on the base of the disease site and extent, according to international recommendations (6,11). All patients scheduled for IFLD were considered eligible for the study and underwent preoperative 18F-FDG PET/CT scanning with DTP acquisition.
Patients with the following characteristics were excluded: prior inguinal surgery dissection; previous chemotherapy or locoregional radiotherapy within the last 5 y; contraindication to the surgery due to age or comorbidities; pregnancy or breast-feeding; blood glucose greater than 200 mg/dL; and surgery performed more than 20 d after 18F-FDG PET/CT. Pathologic results were used as the standard of reference to assess the presence of LN metastases.
18F-FDG PET/CT Acquisition
18F-FDG PET/CT scans were obtained according to the standard procedure of our center (12). All patients fasted for at least 6 h, and the glucose blood levels were less than 190 mg/dL before the 18F-FDG injection. According to body weight, 118–303 MBq of 18F-FDG were intravenously administrated. Before 18F-FDG PET/CT acquisition, patients were hydrated with 500 mL of saline solution by intravenous administration. No oral or intravenous contrast agents were used. All 18F-FDG PET/CT scans were acquired using the same PET scanner (Gemini GXL [Philips] or Biograph mCT [Siemens Medical Solutions USA, Inc.]) for each patient at 2 timepoints: 60 ± 10 min (standard 18F-FDG PET/CT scanning) and 180 ± 10 min (delayed 18F-FDG PET/CT scanning) after 18F-FDG injection. Standard 18F-FDG PET/CT scans were acquired from the skull base to mid thigh. Delayed 18F-FDG PET/CT scans were obtained from the 11th vertebra (T11) to the inguinal region. Before the 18F-FDG PET/CT acquisition, low-dose CT images (using a voltage of 110–120 kVp and tube current of 20–40 mAs, with the patient breathing normally) were acquired for anatomic reference and attenuation correction. PET images were then acquired in 3-dimensional mode, with 7–8 acquisition beds (of ∼2.5 min each) on standard scans and 1–2 acquisition beds (of ∼4 min each) on delayed scans. Matched CT and PET images were reconstructed with a field of view of 50 cm. The line-of-response row-action maximum likelihood algorithm was used for reconstruction with 144 × 144 or 256 × 256 matrix. Attenuation-corrected PET images were reviewed in transverse, sagittal, and coronal planes. PET data were also displayed in a rotating maximum-intensity projection images. To evaluate the images, PET and CT datasets were transferred to an independent computer workstation by DICOM transfer.
18F-FDG PET/CT Image Analysis
All 18F-FDG PET/CT images were interpreted and visually scored by 2 nuclear medicine physicians in consensus.
Qualitative Analysis
Qualitative analysis was performed both on standard and on delayed PET/CT, and the degree of 18F-FDG uptake in the LNs was classified as follows: normal, uptake lower than or equal to background (score 0); and abnormal, uptake higher than background (score 1). The gluteus muscle tissue was used to estimate background activity. The size of the largest LN per groin (short axis) was detected on transaxial CT images of PET/CT.
Semiquantitative Analysis
A spheric volume of interest was placed over the inguinal LN with the highest glucose uptake on the transaxial PET images, for each groin, using an isocontour threshold of 40% method (Syngo.via, MM oncology VA30; Siemens Medical Solutions) based on the SUV (13,14). SUV normalization to body weight and to injected dose was automatically assessed using the following equation:
The SUVmax within the volume of interest was measured on standard (SUVmax standard) and delayed (SUVmax delayed) PET images. Volumes of interest were carefully placed in exactly the same anatomic site, both on standard and on delayed PET/CT scans. When several hypermetabolic LNs per groin were seen on PET/CT images, the highest SUVmax was considered the representative value of that groin. When inguinal LNs did not show a significant 18F-FDG uptake, an arbitrary value of 1 for SUVmax was adopted.
Furthermore, we calculated the retention index of SUVmax (RImax) using the following formula:
The 18F-FDG PET/CT findings and histopathologic results for the inguinal LNs were compared on a groin-by-groin analysis.
Statistical Analysis
Sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and accuracy of standard and delayed PET/CT were calculated considering qualitative analysis. The receiver-operating-characteristic analysis was used to determine the optimal cutoff values of SUVmax standard, SUVmax delayed, and RImax, for differentiating benign and malignant inguinal LNs. Differences in sensitivity, specificity, and accuracy between standard and delayed PET/CT were determined using the χ2 or Fisher test. A P value of less than 0.05 was considered statistically significant. Statistical analysis was performed using MedCalc Statistical Software (version 15.11.4).
RESULTS
A total of 60 patients with primary VC were referred to the Division of Gynaecologic Oncology during the study period. Among these, 33 patients fulfilled the inclusion criteria (Fig. 1). Patients’ characteristics are reported in Table 1. All patients had a squamous cell carcinoma of the vulva. The time interval between the 18F-FDG injection and PET/CT acquisition was 60 ± 11 min for standard PET/CT and 162 ± 24 min for delayed PET/CT. Most PET/CT studies (22/33 patients, 67%) were acquired using the GXL scanner. The time interval between 18F-FDG PET/CT study and surgery was 18 ± 1 d. Fifty-seven groins (24 bilateral, 9 unilateral) in 33 patients were dissected. At pathologic examination, 21 groins contained metastatic LNs, and 36 groins were negative for metastases.
Flowchart of patients’ selection. SLNB = sentinel lymph node biopsy.
Patients’ Characteristics
The mean size of all measured LNs was 9.4 ± 3.6 mm (median, 8 mm; range, 5–21 mm). The mean size of metastatic LNs was 11.8 ± 4.6 mm (median, 10 mm; range, 6–21 mm), whereas the mean size of nonmetastatic LNs was 8.1 ± 1.7 mm (median, 8 mm; range, 5–12 mm). A significant difference was found between the size of metastatic and nonmetastatic LNs measured on low-dose CT (P < 0.002).
Qualitative Analysis
Standard 18F-FDG PET/CT was positive in 29 of 57 groins and negative in 28 of 57 groins. Visual score results are reported in Table 2. Of the 21 groins with metastatic LNs at pathologic examination, standard 18F-FDG PET/CT showed 18F-FDG uptake above the background (score 1) in 20 groins and under background (score 0) in 1 groin. Of the 36 groins with no metastatic LNs at pathologic examination, standard 18F-FDG PET/CT was negative in 27 groins and positive in 9 groins. On a groin-by-groin basis, standard 18F-FDG PET/CT yielded a sensitivity of 95.2% (95% confidence interval [CI], 85.2–99.8), specificity of 75% (95% CI, 61.5–85.1), NPV of 96.4% (95% CI, 86.7–99.4), PPV of 69% (95% CI, 55.2–80.2), and accuracy of 82.5% (95% CI, 72.7–92.3) (Table 3).
Qualitative (Visual Score) Results of Standard 18F-FDG PET/CT and Delayed 18F-FDG PET/CT
Qualitative Analysis
Delayed 18F-FDG PET/CT was positive in 28 of 57 groins and negative in 29 of 57 groins (Table 2). Of the 21 groins with metastatic LNs at pathologic examination, delayed 18F-FDG PET/CT showed abnormal 18F-FDG uptake in 20 groins, with a false-negative (FN) result occurring in 1 groin (same groin that was FN at standard imaging). Of the 36 groins with no metastatic LNs at pathologic examination, delayed 18F-FDG PET/CT was true negative in 28 groins and false-positive (FP) (Fig. 2) in 8 groins. On a groin-by-groin basis, delayed 18F-FDG PET/CT yielded a sensitivity of 95.2% (95% CI, 85.2–98.8), specificity of 77.8% (95% CI, 64.5–87.3), NPV of 96.6%, (95% CI, 86.9–99.4), PPV of 71.4% (95% CI, 57.7–82.2), and accuracy of 84.2% (95% CI, 74.8–93.6) (Table 3). No significant differences in sensitivity, specificity, and accuracy between standard and delayed PET/CT were found.
A 60-y-old woman with midline tumor. (A) Maximum-intensity projection of standard 18F-FDG PET/CT image showing focal uptake in right groin (SUVmax, 3.5) as well as in left groin (SUVmax, 2) (arrows). (B) Maximum-intensity projection of delayed scan showing increase of focal uptake in right groin (SUVmax, 6.39) and in left groin (SUVmax, 3.41) (arrows). (C) Pathologic examination showed no metastatic LNs in either groin.
Semiquantitative Analysis
Mean and median values of SUVmax standard, SUVmax delayed, and RImax for metastatic groins (group 1) and nonmetastatic groins (group 0) are shown in Table 4. SUVmax standard, SUVmax delayed, and RImax were significantly higher for group 1 than for group 0 (P < 0.0001) (Fig. 3). The area under the curve was larger in SUVmax standard, which was 0.919 (P < 0.0001; 95% CI, 81.5%–97.5%) compared with 0.899 in SUVmax delayed (P < 0.0001; 95% CI, 79%–96.3%) and 0.833 in RImax (P < 0.0001; 95% CI, 71%–92%). There was no significant difference between the areas under the curve in SUVmax standard and SUVmax delayed (P = 0.10) nor between SUVmax delayed and RImax (P = 0.055), whereas a significant difference was found between SUVmax standard and RImax (P = 0.04). At receiver-operating-characteristic analysis, the optimal cutoff values of SUVmax standard, SUVmax delayed, and RImax were >1.32, >1.88, and >0, respectively. With these cutoff values, the sensitivity and specificity, respectively, were 95.2% and 77.8% for both SUVmax standard and SUVmax delayed and 95.2% and 80% for RImax (Fig. 4).
Mean and Median of SUVmax on Standard and Delayed Scans and RI for Metastatic and Nonmetastatic Groups
Box plots showing distribution of SUVmax standard, SUVmax delayed, and RImax for metastatic and nonmetastatic groins. 0 = absence of metastases; 1 = presence of metastases.
The receiver-operating-characteristic curves of SUVmax on standard scan and delayed scan and RI.
DISCUSSION
In this prospective study, we evaluated the value of DTP 18F-FDG PET/CT for LN staging. We chose to include only patient candidates for lymphadenectomy, excluding those addressed to the sentinel node biopsy, to ensure that the reference standard (histopathologic results) could include all the inguinal LNs examined preoperatively by 18F-FDG PET/CT. Moreover, given the rapid progression of VC, we decided to include only those patients for whom a maximum 3-wk interval between the preoperative study and surgery was compiled. On the basis of criteria that considered metastatic LNs if they had 18F-FDG uptake higher than background, standard 18F-FDG PET/CT showed high sensitivity (95.2%) and relatively low specificity (75%) in detecting metastatic groins. In particular, we found 9 FP groins and 1 FN groin, the latter due to a metastatic LN of 8 mm in diameter. In 9 FP groins, pathologic examination revealed no LN metastases in 7 groins and inflammatory LNs in 2 groins. Concerning the FP results, it is widely known in the literature that inflammatory cells and activated macrophages represent a common cause of increased 18F-FDG uptake, as occurring in inguinal reactive LNs after vulvar biopsy or shaving (15). Conversely, FN results are most likely associated with PET-undetected micrometastatic foci in not enlarged nodes as well as with extensive necrosis within metastatic LNs with subsequent loss of 18F-FDG uptake (7). Previous PET or PET/CT studies in the literature showed a range of sensitivity (from 50% to 100%) and specificity (from 91% to 100%) for detecting metastatic involvement of inguinofemoral LNs in VC patients (7,8,16,17). In those studies, no cutoff values of SUVmax were determined for metastatic LNs. In our study, an SUVmax of greater than 1.32 was found to be the optimal cutoff point on standard PET/CT to provide a sensitivity similar to the one obtained from visual analysis (95.2%) but accompanied by a slightly better (77.8% vs 75%) although not statistically significant (P = 0.81) specificity in detecting metastatic groins. As already mentioned, increased 18F-FDG uptake is not specific for neoplastic involvement, because it may also be reactive to inflammation or infection (18). On the basis of our data, standard 18F-FDG PET/CT showed a high NPV (96.4%) together with a low PPV (69%) in detecting metastatic groins. Therefore, a negative 18F-FDG PET/CT scan result is highly predictive in excluding groin metastases and could potentially be used to select patient candidates for a minimal groin surgery. In contrast, a positive 18F-FDG PET/CT scan finding is not highly predictive for groin metastases and needs to be interpreted with caution. On the basis of our data, a significant difference was found between the size of metastatic and nonmetastatic LNs measured on low-dose CT. However, further studies are needed to investigate whether the combination of PET and CT criteria can better differentiate between metastatic and nonmetastatic LNs.
The rationale of DTP 18F-FDG PET/CT is that 18F-FDG uptake usually increases in malignant lesions for several hours after intravenous injection, whereas benign lesions and inflammatory cells show stable or decreasing 18F-FDG uptake over time (19,20). Such different behavior on delayed PET/CT is believed to be due to increased cell proliferation rate, enhanced expression of hexokinase type-II and glucose transporter-1 in malignant lesions, and continued clearance of background activity, thus resulting in images with improved contrast-to-noise ratio (10). Up until now, no studies have yet been performed using qualitative and semiquantitative parameters on both standard and delayed 18F-FDG PET/CT to assess LN status in VC. To our knowledge, only Lin et al. performed delayed PET scanning in VC staging, showing that it did not modify the qualitative analysis of a standard scan. However, they did not apply semiquantitative evaluation and the sample size in their study was limited (11 patients) (8). According to our qualitative analysis, delayed 18F-FDG PET/CT showed sensitivity and NPV (95.2% and 96.5%) similar to standard scanning with relatively higher specificity (77.8% vs. 75%) and PPV (71.4% vs. 69%) not reaching statistical significance (P = 0.9) in detecting metastatic groins. In particular, in the patient with an FN groin no 18F-FDG uptake, even on the delayed scanning, was seen, most likely due to the limited extent of the metastatic involvement (with micrometastases on histopathology). Moreover, on delayed PET/CT we found 8 FP groins compared with 9 FP groins on standard PET/CT. This finding suggests that the behavior of inflammatory lesions as concerning 18F-FDG uptake is not always predictable, and that delayed PET/CT appears to not reduce the rate of FP results. According to other studies, inflammatory lesions as well as infection may induce higher 18F-FDG uptake on delayed scanning mimicking malignant lesions (21–24). In the literature, the use of DTP 18F-FDG PET/CT in gynecologic malignancies is still a subject of discussion, and in our study, the delayed 8F-FDG PET/CT also was not superior to standard images in detecting LN disease. This is in accordance with the results reported in a recent metaanalysis by Shen et al. (21), which concluded that DTP 8F-FDG PET/CT had higher sensitivity but lower specificity in detecting LNs metastases on a per-patient analysis, and performed only slightly better than standard PET/CT on a per-lesion basis. On the contrary, in a retrospective study on cervical, endometrial, and ovarian cancer patients, Nogami et al. reported that DTP 18F-FDG PET/CT only significantly improved the specificity for detection of LN metastases, but also concluded that DTP scanning had an unsatisfactory impact on the overall diagnostic efficacy for LN metastasis (9). Regarding the semiquantitative analysis, an SUVmax greater than 1.88 was considered the optimal cutoff point on delayed PET/CT, providing a same sensitivity and specificity as standard 18F-FDG PET/CT images. Concerning the RI, prior studies reported that the RI might improve the accuracy of DTP 18F-FDG PET/CT in gynecologic cancer for detecting LN metastases (9,25). In the present study, we found that an RI of greater than 0% was the optimal cutoff point for nodal evaluation because it improved the specificity (80% for RI on delayed examination compared with 77.8% for SUVmax on standard PET/CT P = 0.8) but did not improve the sensitivity (95.2% in both cases).
Our study suffered some limitations. First, the population was relatively small, but in accordance with the incidence of VC (2.4/100,000 inhabitants per year) (1). Second, the patients were scanned on 2 different scanners in our department and this could have minimally affected the SUV homogeneity (26,27); however, only 11 patients were scanned on a different scanner. Third, the cutoff value of SUVmax and RI in this study was based on the data collected at our institute alone, and the absolute value of SUVmax might vary somewhat according to different imaging systems used at other institutions. However, this study has several strengths. First, it is, to our knowledge, the first prospective study to evaluate the comparison between standard and delayed scanning using qualitative and semiquantitative analysis. Second, it evaluates only patients who underwent surgery shortly after PET/CT, to compare PET/CT results with histopathologic findings.
CONCLUSION
In the light of our results, standard 18F-FDG PET/CT has high sensitivity and NPV in detecting groin lymph node metastases in VC patients. This confirms that standard 18F-FDG PET/CT represents an effective preoperative imaging method for LN staging in VC, allowing better planning of groin surgical procedures and selection of patients potentially suitable for minimally invasive surgery. However, delayed PET/CT has not significantly been able to improve the specificity and the PPV in our study. Larger studies are needed to further validate our results.
DISCLOSURE
No potential conflict of interest relevant to this article was reported.
Footnotes
Published online May 25, 2017.
- © 2017 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication April 3, 2017.
- Accepted for publication May 17, 2017.