Data from the NCDB collected between January 1, 2010 and December 31, 2017 were utilized for the analyses. A total of 12 063 cases were identified during this period (median age 64.3 ± 14.4 years). Patients were predominantly non‐Hispanic white (85.9%), non‐Hispanic black (8.0%), and Hispanic (4.3%). These patients received therapy for vulvar cancer at American College of Surgeons’ Commission on Cancer (CoC) accredited hospitals and were treated per the prevailing standard of care. It is therefore safe to assume that the disease course was determined by stage at diagnosis and not significantly impacted by variations in therapy. Data selection criteria are shown in Table 1 . Stage distribution of included cases is shown in Table 2 .
Statistical analysis was performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA). For each stage category, many possible tumor characteristics were included. Tumor characteristics that were evaluated included tumor size, depth of invasion, extent of local spread, lymph node status, and regional and distant spread of disease. Log‐rank and Wilcoxon tests were used to analyze overall survival similarities between and within groups of tumor characteristics resulting in the new (2021) FIGO staging for carcinoma of the vulva (Table3 ).
Groups with overlapping survivals were then combined into individual stages and substages. These analyses were repeated multiple times. A secondary analysis was then performed to check whether each higher stage carries a worse prognosis than the preceding stage or substage. Starting from Stage IA the analysis confirmed that each higher stage or substage indeed has a worse prognosis than the preceding stage or substage; furthermore, that the associated prognosis with each stage or substage is unchanged by including tumor characteristics from any or all of the preceding stages or substages. As an example, Stage IIIB is defined by lymph node metastasis >5 mm, and combining this with disease extension to upper two‐thirds of perineal structures did not change the prognosis. Similarly, Stage IIIC is defined by lymph node metastasis with extracapsular extension, and combining this with lymph node metastases that are >5 mm without extracapsular extension and tumor extension to upper two‐thirds of adjacent perineal structures did not alter the prognosis.
Kaplan–Meier curves were generated for all of the stages and substages (Figure1 ). The method used for computing confidence intervals requires an estimate of the survival distribution at the value nth percentile (n = 50). This means that if the survival function does not reach n–5 (0.45), a standard error or confidence interval bounds for the median will not be estimated. This is while the median overall survival for each stage or substage is not given.
The NCDB data utilized for these analyses may not cover the entire global experience, but the large size of the data compensates for this shortcoming and makes the findings generalizable.
Statistical analysis was performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA). For each stage category, many possible tumor characteristics were included. Tumor characteristics that were evaluated included tumor size, depth of invasion, extent of local spread, lymph node status, and regional and distant spread of disease. Log‐rank and Wilcoxon tests were used to analyze overall survival similarities between and within groups of tumor characteristics resulting in the new (2021) FIGO staging for carcinoma of the vulva (Table
Groups with overlapping survivals were then combined into individual stages and substages. These analyses were repeated multiple times. A secondary analysis was then performed to check whether each higher stage carries a worse prognosis than the preceding stage or substage. Starting from Stage IA the analysis confirmed that each higher stage or substage indeed has a worse prognosis than the preceding stage or substage; furthermore, that the associated prognosis with each stage or substage is unchanged by including tumor characteristics from any or all of the preceding stages or substages. As an example, Stage IIIB is defined by lymph node metastasis >5 mm, and combining this with disease extension to upper two‐thirds of perineal structures did not change the prognosis. Similarly, Stage IIIC is defined by lymph node metastasis with extracapsular extension, and combining this with lymph node metastases that are >5 mm without extracapsular extension and tumor extension to upper two‐thirds of adjacent perineal structures did not alter the prognosis.
Kaplan–Meier curves were generated for all of the stages and substages (Figure
The NCDB data utilized for these analyses may not cover the entire global experience, but the large size of the data compensates for this shortcoming and makes the findings generalizable.