Diagnostic performance of preoperative [18F]FDG-PET/CT for lymph node staging in vulvar cancer: a large single-centre study

Abstract

Purpose: This retrospective study aimed to assess the diagnostic performance of preoperative [¹⁸F]FDG-PET/CT in predicting the groin and pelvic lymph node (LN) status in a large single-centre series of vulvar cancer patients.

Methods: Between January 2013 and October 2018, among all consecutive women with proven vulvar cancer submitted to [¹⁸F]FDG-PET/CT, 160 patients were included. LNs were analysed by two qualitative methods assessing PET information (defined as visual assessment) and a combination of PET and low-dose CT information (defined as overall assessment), respectively, as well as semi-quantitative analysis (LN-SUV_max). Sensitivity, specificity, accuracy, positive and negative predictive values (PPV and NPV) in predicting the groin and pelvic LN status were calculated in the overall study population; a subset analysis of groin parameters in clinically/ultrasonography negative patients was also performed. Histopathology was the reference standard.

Results: All patients underwent vulvar and inguinofemoral LN surgery, and 35 pelvic LN surgery. Overall, 338 LN sites (296 groins and 42 pelvic sites) were histologically examined with 30.4% prevalence of metastatic groins and 28.6% for metastatic pelvic sites. In the overall study population, sensitivity (95% confidence interval, CI), specificity (95% CI), accuracy (95% CI), PPV (95% CI) and NPV (95% CI) at the groin level were 85.6% (78.3-92.8), 65.5% (59.0-72.0), 71.6% (66.5-76.8), 52.0% (44.0-60.1) and 91.2% (86.7-95.8) for visual assessment; 78.9% (70.5-87.3), 78.2% (72.5-83.8), 78.4% (73.7-83.1), 61.2% (52.3-70.1) and 89.4% (85.0-93.9) for overall assessment; and 73.3% (64.2-82.5), 85.0% (80.1-89.8), 81.4% (77.0-85.8), 68.0% (58.8-77.3) and 87.9% (83.4-92.5) for semi-quantitative analysis (SUV_max cut-off value 1.89 achieved by ROC analysis). Similar results were observed in the pelvis-based analysis.

Conclusion: In this large single-centre series of vulvar cancer patients, [¹⁸F]FDG-PET/CT showed good values of sensitivity and NPV in discriminating metastatic from non-metastatic LNs. In routine clinical practice, qualitative analysis is a reliable interpretative criterion making unnecessary commonly used semi-quantitative methods such as SUV_max.

Keywords: Lymph node staging; Personalized medicine; Vulvar cancer; [18F]FDG-PET/CT.

Fig. 1

The flowchart of study population

Fig. 2

Results of qualitative and semi-quantitative analysis in discriminating positive versus negative LNs at histopathology, for groin sites. FN, false negative; TN, true negative; TP, true positive; FP, false positive; NMHP, non-metastatic with high probability; MHP, metastatic with high probability. *Best cut-off value achieved by ROC analysis

Fig. 3

Results of qualitative and semi-quantitative analysis in discriminating positive versus negative LNs at histopathology, for pelvic sites. FN, false negative; TN, true negative; TP, true positive; FP, false positive; NMHP, non-metastatic with high probability; MHP, metastatic with high probability. *Best cut-off value achieved by ROC analysis

Fig. 4

Box plots showing distribution of groin SUV_max (a) and pelvic SUV_max (b) for non-metastatic and metastatic LNs. The boxes indicate medians with upper (Q3) and lower quartiles (Q1); the upper and lowers bars define the upper and lower adjacent values, respectively; dots indicate outliers [29, 30]

Fig. 5

The receiver operating characteristic curves of SUV_max for groin and pelvic sites

Fig. 6

A 68-year-old woman with untreated squamous cell carcinoma. Multiple intensity projection (MIP) images showing ¹⁸F-FDG uptake in two groin LNs (thin arrows) and one pelvic LN (thick arrow) (a). ¹⁸F-FDG-PET showing focal uptake both in the right (score 2 at visual assessment, SUV_max 2.5) and in the left (score 2 at visual assessment, SUV_max 6.0) groin LN (thin arrows) (b). At low-dose CT, the right LN shows short axis diameter of 11 mm and round shape, the left LN short axis of 16 mm, round shape and possible necrosis (thin arrows) (c). At overall assessment, both LNs are judged as clearly abnormal. Pathologic examination showed reactive features in all the right groin LNs removed (d) and metastasis in the largest LN among those removed in the left groin (e). The pelvic LN, which was located in the obturator region (PET and CT images not shown), was metastatic at histopathology