Station-level 18F-FDG PET/CT for lymph node staging in gynecologic cancer: where it is reliable, where it misleads, and a two-centre analysis of discordance with histopathology

Technical Summary: Station-level 18F-FDG PET/CT for Lymph Node Staging in Gynecologic Cancer

Problem Statement
Accurate preoperative detection of lymph node metastasis is critical for staging and determining adjuvant treatment in gynecologic cancers (endometrial, ovarian, cervical, and vulvar). While 18F-fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT) is widely utilized, it exhibits a well-recognized diagnostic asymmetry: high specificity but limited sensitivity. Consequently, positive findings are generally reliable, but negative findings frequently fail to exclude metastatic disease. Previous research has largely pooled nodal stations or focused on patient-level metrics, leaving gaps in understanding where and why PET/CT fails on a station-by-station basis. Furthermore, the specific contributions of scan-day blood glucose (potentially causing false negatives via competitive inhibition) and systemic inflammation (potentially causing false positives via reactive hyperplasia) to these errors remain under-investigated in a histopathologically verified cohort.

Methodology
This retrospective, two-centre diagnostic accuracy study adhered to the STARD 2015 guidelines. The cohort consisted of 120 women with histologically confirmed gynecologic cancer who underwent preoperative FDG-PET/CT followed by lymphadenectomy.

• Data Source: Patients were drawn from two tertiary hospitals (Center 1: İzmir Katip Çelebi University Atatürk Training and Research Hospital; Center 2: University of Health Sciences, Gaziosmanpaşa Training and Research Hospital).
• Index Test: FDG-PET/CT was performed using different scanners (GE Discovery 710 and Siemens Biograph) but consistent protocols. Nodal positivity was determined based on standard clinical reports describing focal uptake above background, irrespective of size. Two independent readers classified stations, with discrepancies resolved by consensus.
• Gold Standard: Histopathological examination of the surgical specimen served as the reference. A station was considered positive if it contained at least one metastatic node.
• Analysis Scope: The study evaluated 280 nodal stations across four categories: pelvic, para-aortic, common iliac, and inguinal.
• Statistical Approach: Sensitivity, specificity, predictive values, and likelihood ratios were calculated per station with Wilson 95% confidence intervals. Patient-level clustering was addressed via bootstrap methods. The study tested two pre-specified hypotheses: (1) the association between scan-day blood glucose and false-negative findings, and (2) the association between the neutrophil-to-lymphocyte ratio (NLR) and false-positive findings. Additionally, primary-tumour SUVmax was assessed as a predictor of nodal metastasis.

Key Results

• Overall Diagnostic Performance: Across 280 stations, FDG-PET/CT demonstrated high specificity (94.0%) but only moderate sensitivity (62.5%). The false-negative rate was 37.5%, while the false-positive rate was 6.0%.
• Station-Specific Variability:
  • Pelvic Station: Highest sensitivity (71%) and specificity (95%).
  • Para-aortic Station: Lowest sensitivity (45%), despite high specificity (96%).
  • Common Iliac Station: Notable for a low positive predictive value (PPV) of 27%, indicating that most positive findings here were not confirmed by pathology. This was the only station where false positives outnumbered false negatives.
  • Inguinal Station: Evaluable in only six patients (vulvar cancer); showed 100% specificity but limited statistical power.
• Patient-Level Metrics: Per patient, sensitivity was 76.7% and specificity was 81.1%. Inter-reader agreement was high (Cohen's κ = 0.81).
• Determinants of Discordance:
  • Glucose: False-negative stations showed a trend toward higher scan-day blood glucose (median 110 mg/dL vs. 102 mg/dL; OR 1.30 per 10 mg/dL), though this did not reach statistical significance (p = 0.15).
  • Inflammation: The NLR was not associated with false-positive findings (p = 0.68).
• Predictors of Metastasis: A higher primary-tumour SUVmax independently predicted nodal metastasis (AUC 0.73). In multivariable analysis, primary-tumour SUVmax, tumour size, and tumour grade were all independently associated with nodal involvement.

Key Contributions

1. Station-Level Granularity: The study provides a detailed breakdown of diagnostic accuracy by nodal station, identifying the common iliac region as a specific "pitfall" with a high rate of false positives (low PPV) and the para-aortic region as having the lowest sensitivity.
2. Asymmetry Confirmation: It quantifies the diagnostic asymmetry, confirming that a negative PET result cannot rule out metastasis, while a positive result is highly reliable.
3. Clinical Nuance on Common Iliac Nodes: The paper highlights that pooling the common iliac station into a general "pelvic" category masks a specific diagnostic weakness, suggesting this station should be reported separately in routine practice.
4. Metabolic Context: It demonstrates that while nodal sensitivity is limited, the metabolic activity of the primary tumour (SUVmax) carries independent predictive value for nodal disease, suggesting that a negative nodal read in the context of a highly avid primary tumour warrants caution.

Significance and Claims
The authors assert that the primary clinical message is straightforward: FDG-PET/CT is best utilized to confirm suspected nodal or distant disease rather than to exclude lymph node involvement.

• Surgical Implications: A negative PET result should not justify omitting lymphadenectomy or sentinel lymph node mapping when nodal status would alter adjuvant treatment. The study reinforces that surgical staging (or sentinel lymph node biopsy with ultrastaging) remains necessary because PET misses low-volume and micrometastatic disease.
• Hypothesis-Generating Findings: The association between higher scan-day glucose and false negatives is presented as a non-significant trend requiring prospective confirmation, serving as a reminder to maintain strict normoglycaemia. Similarly, the finding that systemic inflammatory indices do not predict false positives suggests that benign FDG uptake is driven by local nodal biology rather than systemic inflammation.
• Limitations: The authors explicitly note the study's retrospective nature, the confounding of centre, surgeon, and scanner, and the restriction to surgically selected patients (verification bias). They conclude that the findings regarding glucose and the common iliac station are hypothesis-generating and require validation in larger prospective studies.

In summary, the paper argues that while PET/CT is a valuable tool for its high specificity, its limitations in sensitivity—particularly in the para-aortic region and for low-volume disease—mean it cannot replace surgical nodal staging in gynecologic oncology.