Background
There are 2 known pathways for tumorigenesis of vulvar squamous cell carcinoma—a human papillomavirus–dependent pathway characterized by p16 overexpression and a human papillomavirus–independent pathway linked to lichen sclerosus, characterized by TP53 mutation. A correlation of human papillomavirus dependency with a favorable prognosis has been proposed.
Objective
The objective of the study was to further understand the role of human papillomavirus and p53 status in vulvar squamous cell carcinoma and characterize its clinical relevance.
Study Design
The Arbeitsgemeinschaft Gynaecological Oncology Chemo and Radiotherapy in Epithelial Vulvar Cancer-1 study is a retrospective cohort study of 1618 patients with primary vulvar squamous cell carcinoma Fédération Internationale de Gynécologie et d’Obstétrique stage ≥1B treated at 29 gynecologic cancer centers in Germany between 1998 and 2008. For this translational substudy, formalin-fixed paraffin-embedded tissue was collected. A tissue microarray was constructed (n=652 samples); p16 and p53 expression was determined by immunohistochemistry. Human papillomavirus status and subtype were analyzed by polymerase chain reaction.
Results
p16 immunohistochemistry was positive in 166 of 550 tumors (30.2%); p53 staining in 187 of 597 tumors (31.3%). Only tumors with available information regarding p16 and p53 immunohistochemistry and without p53 silent expression pattern were further analyzed (n=411); 3 groups were defined: p53+ (n=163), p16+/p53− (n=132), and p16−/p53− (n=116). Human papillomavirus DNA was detected in 85.6% of p16+/p53− tumors; human papillomavirus-16 was the most common subtype (86.3%) . Patients with p16+ tumors were younger (64 vs 72 years for p53+, respectively, 69 years for p16−/p53− tumors; P <.0001) and showed lower rates of lymph-node involvement (28.0% vs 42.3% for p53+, respectively, 30.2% for p16−/p53− tumors; P =.050). Notably, 2-year-disease-free and overall survival rates were significantly different among the groups: disease-free survival, 47.1% (p53+), 60.2% (p16−/p53−), and 63.9% (p16+/p53−) ( P<. 001); overall survival, 70.4% (p53+), 75.4% (p16−/p53−), and 82.5% (p16+/p53−) ( P=. 002). In multivariate analysis, the p16+/p53− phenotype showed a consistently improved prognosis compared with the other groups (hazard ratio, 0.66; 95% confidence interval, 0.44–0.99; P =.042).
Conclusion
p16 overexpression is associated with an improved prognosis whereas p53 positivity is linked to an adverse outcome. Our data support the hypothesis of a clinically relevant third subgroup of vulvar squamous cell carcinoma with a p53−/p16− phenotype showing an intermediate prognosis that needs to be further characterized.
Introduction
Over the last 15 years, the incidence of vulvar squamous cell cancer (VSCC) has almost doubled in Germany; in particular, the proportion of younger females at the age between 30 and 49 years affected by the disease has been increasing. , However, most women diagnosed are still at the age of 70 to 75 years, and worldwide, VSCC is regarded as a rare disease. According to currently available evidence, there are 2 major pathways for tumorigenesis of VSCC: a human papillomavirus (HPV) dependent with p16 overexpression as a potential surrogate for HPV-associated transformation and an HPV-independent route linked to lichen sclerosus, characterized by TP53 mutation. The understanding of the further molecular landscape underlying VSCC development is only just evolving. HPV association was proposed in 20% to 40% of all VSCC cases with HPV-16 being the most commonly identified subtype (75% of all HPV-related cases). In head and neck cancers, which show a similar biologic behavior compared with vulvar cancer, HPV-related disease has repeatedly been demonstrated to have a much more favorable prognosis and a better response to chemoradiation. , In contrast, no consistent data with regard to an improved prognosis for patients with high-risk HPV-related disease are available for VSCC. Reasons for the ambiguous results of the role and frequency of HPV in VSCC might be different detection methods, small and heterogeneous patient cohorts, and the retrospective character of the available studies. Information on the prognostic relevance of TP53 mutation and consecutive p53 expression in VSCC in view of the increasing knowledge on HPV is similarly restricted. , In preinvasive disease, there is robust evidence of a higher progression rate to VSCC in case of non-HPV-related differentiated vulvar intraepithelial neoplasia (VIN) with p53 overexpression, whereas the usual type VIN lesions show lower progression rates and lesser potential for recurrence. In a systematic review from Sand et al, p53 expression (n=310) was also associated with a poorer prognosis (hazard ratio [HR], 1.81 for overall survival [OS]) in invasive disease. Unfortunately, there are very few and small studies investigating all markers (p16, HPV, p53) in the same patient cohort to allow for a valuation with regard to distribution and prognostic differences. , ,
Why was this study conducted?
For vulvar squamous cell cancer (VSCC), data regarding a more favorable prognosis for patients with human papillomavirus (HPV)–related tumors have been inconsistent. This study is the translational part of the exceptionally large and clinically well characterized Arbeitsgemeinschaft Gynäkologische Onkologie Chemo and Radiotherapy in Epithelial Vulvar Cancer-1 study of 1618 patients with VSCC that aims to further understand the role of HPV and p53 in VSCC.
Key findings
In this analysis, we show that HPV driveness/p16 overexpression is associated with an improved prognosis in VSCC, whereas p53 overexpression is linked to an adverse outcome with lower 2-year disease-free and overall survival rates.
What does this add to what is known?
Our study supports a clinically relevant third subgroup of tumors showing neither p16 nor p53 overexpression that has an intermediate prognosis.
Therefore, to further understand the role of HPV and p53 in VSCC, we analyzed tumor samples from the Arbeitsgemeinschaft Gynäkologische Onkologie Chemo and Radiotherapy in Epithelial Vulvar Cancer-1 (AGO-CaRE-1) study, an exceptionally large and clinically well characterized cohort from 29 German cancer centers.
Materials and Methods
The AGO-CaRE translational study is a substudy of the AGO-CaRE-1 study. AGO-CaRE-1 is a large retrospective study, evaluating treatment patterns and prognostic factors in vulvar cancer. Participating institutions included all patients with the diagnosis of invasive vulvar cancer stage >pT1a independent of the mode and initial place of treatment. Detailed information about the recruitment and data collection were published by Mahner et al . In short, 1618 adult patients with stage IB–IV VSCC (Union for International Cancer Control version 6 ), being treated between 1998 and 2008 at 29 AGO cancer centers in Germany, were included. Patient data collection was performed retrospectively between February and December 2011. Documentation and analysis were done through a specifically designed centralized database by the AGO study group. In the database, tumor characteristics and aspects of surgical and nonsurgical treatment were collected including: tumor, nodes, and metastatis stage, tumor size, depth of invasion, grade, number and localization of lymph nodes involved, surgical therapy of the vulva and nodes, pathologic resection margin, total dosage and fields of irradiation, and, if applicable, agent and dosage of chemotherapy and date and treatment of recurrent disease and/or date of last contact or death. Furthermore, patient characteristics such as Eastern Cooperative Oncology Group (ECOG) performance status and relevant comorbidities were documented. To account for possible bias from informative missing values, we introduced the category “unknown” for each variable to keep all patients in the analysis.
For this CaRE translational substudy, available formalin-fixed paraffin-embedded (FFPE) tissue from the patients documented in the AGO-CaRE-1 database was collected centrally (n=807).
The AGO-CaRE-1 study and the translational substudy were approved by each local ethics committee (leading vote: Hamburg [reference number PV3658]) and registered with ClinicalTrials.gov ( NCT01304667 ).
DNA isolation and human papillomavirus polymerase chain reaction
From each FFPE tissue block, 3×10 μm thick sections were cut and DNA isolation was performed using the NucleoSpin DNA FFPE XS kit (Macherey-Nagel GmbH & Co. KG, Düren, Germany) following the manufacturer’s instructions. To minimize the potential for polymerase chain reaction (PCR) contamination, the microtome blade was cleaned between each block. In addition, negative control blocks were regularly cut and analyzed in between. DNA quality was first proven by PCR using actin-specific primers. HPV status was analyzed by PCR using GP5+/GP6+ primers as previously described, and each HPV-positive result was validated by genotyping with direct DNA sequencing. p16+, HPV negative cases (n=28) were tested for amplification of human beta globin; if positive (n=20), HPV PCR was repeated with 2 primer sets. GP5+/GP6+ (tttgttactgtggtagatactac/gaaaaataaactgtaaatcatattc) and MY09/MY11 (cgtccaaaaggaaactgagccgtccmarrggawactgagc/gcacagggacataacaatgggcmcagggwcataayaatgg).
Tissue microarray
The tissue microarray (TMA) manufacturing process was described in detail before (PMID:9662379). In brief, 807 vulva study samples were validated by a pathologist with a special focus on gynecologic pathology (E.B.). Because of too small volume of unequivocal cancer for reliable punching, 155 cases were not usable for TMA construction. From the other 652 samples, 1 tissue core measuring 0.6 mm in diameter was taken for TMA construction resulting in 2 TMA blocks. The presence or absence of tumor tissue was validated by visual inspection of the hematoxylin and eosin–stained TMA slides.
Immunohistochemistry
Freshly cut TMA sections were immunostained at 1 day and in 1 experiment. Slides were deparaffinized and exposed to heat-induced antigen retrieval for 5 minutes in an autoclave at pH 7.8 Tris–ethylenediaminetetraacetic acid–citrate buffer. Primary antibody specific for p16 (dilution 1:150, monoclonal antibody, Cat#DIA-P16−OD; dianova GmbH, Hamburg, Germany) and p53 (dilution 1:3600, monoclonal antibody, DO-7; DAKO RTU, Glostrup, Denmark) was applied at 37°C for 60 minutes. Bound antibody was then visualized using the EnVision Kit (DAKO RTU, Glostrup, Denmark) according to the manufacturer’s directions. Immunohistochemistry (IHC) expression levels were determined by using a 4-step scoring system: negative, no staining at all; weak, 1+ staining intensity in ≤70% positive tumor cells or 2+ staining intensity in ≤30% positive tumor cells; moderate, 1+ staining intensity in >70% tumor cells, 2+ in >30% but ≤70% positive tumor cells or 3+ in ≤30% positive tumor cells; and strong, 2+ >70% or 3+ >30% positive tumor cells. In the case of p53, only cases in the category “strong” were regarded as positive and therefore TP53 abnormal. Most TP53 mutations (>70%) have been described as missense mutations leading to an exceptionally stable p53 protein, resulting in strong p53 expression; the rate of TP53 knockout mutations resulting in a completely negative expression pattern of p53 varies between 50% and 80% in the currently available sparse literature. , Furthermore, the proportion of p53 IHC completely negative cases could be overrated owing to array-based evaluation in the current study. Therefore, the mutational status of those cases remains unclear. Consequently, they were not added to the p53 abnormal group but excluded from further analysis (n=124). For p16, both categories “strong” and “moderate” were considered positive because of the small number of samples in the category “moderate” (n=23) and detection of HPV DNA in the majority of these cases (68%).
Statistical analysis
Analyses were performed using Stata (StataCorp LP, version 14.2; StataCorp LLC, College Station, TX). Quantitative variables were summarized using means and standard deviations, and categorical variables are summarized using absolute and relative frequencies. For the determination of significance, we calculated P values using 2-sided tests with a 5% level for significance. Tumor and patient characteristics were compared across groups using analysis of variance (quantitative variables), Pearson’s chi-square test, or Fisher’s exact test (categorical variables). Disease-free survival (DFS) was defined as the time interval between primary diagnosis and disease progression or death of any cause, and OS was the period resulting from primary diagnosis to death of any cause. Cox regression analyses were conducted to determine prognostic factors in (multivariate) survival analysis. Kaplan-Meier curves were calculated to describe (disease-free) survival in subgroups.
Results
Immunohistochemistry
Of 807 collected tumor samples, 652 had enough tumor volume for being punchable for the TMA and were confirmed as VSCC in central review (University Medical Center Hamburg-Eppendorf). p16 staining was evaluable in 550 TMA spots. Notably, 166 of 550 tumors (30.2%) showed a positive (strong [n=143]/moderate [n=23]) p16 expression, whereas 327 spots were classified as negative and 57 as weak. p53 staining was interpretable in 597 spots; among them, 187 of 597 (31.3%) were classified as positive (strong p53 expression), 174 as negative, 178 as weak, and 58 as moderate (wild-type pattern) ( Figure 1 ). The combined expression profiles of p53 and p16 of all tumors with available information regarding p16 and p53 IHC staining (n=535) are presented in Table 1 . In view of their unclear TP53 mutational status, we excluded all p16−/p53 completely negative cases (n=105) and the p16+/p53 completely negative cases without HPV detection in the PCR (n=19) from further analyses ( Table 2 ). Only 12 tumors (2.9%) showed a coexpression of p16 and p53. Because TP53 mutations with strong p53 overexpression is a good explanation for non-HPV–related p16 overexpression, these cases were included in the p53+ group for further analysis. Table 3 shows the tumor and patient characteristics of the 411 cases with regard to the expression subgroups: p53+, p16+/p53−, and p16−/p53−. Interestingly, there was a relevant number of tumors with neither p16 nor p53 overexpression (116 of 411). Compared with the other groups, patients with p16+/p53− tumors were significantly younger at diagnosis (64 vs 72 years for p53+, respectively, 69 years for p16−/p53− tumors; P <.0001) and showed lower rates of lymph-node involvement (28.0% vs 42.3% for p53+, respectively, 30.2% for p16−/p53− tumors; P =.050) ( Table 3 ). In correspondence with the more advanced age of patients with p53+ tumors, the ECOG performance status of these patients was significantly worse than the other subgroups (ECOG 2, 22.1% vs 15.9% for p16+/p53−, respectively, 15.5% for p16−/p53− tumors; P =.009).
p16 | p53 | ||||
|---|---|---|---|---|---|
| Negative | Weak | Moderate | Strong | Total | |
| Negative | 84 | 69 | 32 | 130 | 315 |
| Weak | 21 | 11 | 4 | 21 | 57 |
| Moderate | 9 | 4 | 4 | 6 | 23 pos |
| Strong | 45 | 76 | 13 | 6 | 140 pos |
| Total | 159 | 160 | 53 | 163 pos | 535 |
p16 | p53 | ||||
|---|---|---|---|---|---|
| Negative | Weak | Moderate | Strong | Total | |
| Negative | 0 | 69 | 32 | 130 | 231 |
| Weak | 0 | 11 | 4 | 21 | 36 |
| Moderate | 6 | 4 | 4 | 6 | 20 pos |
| Strong | 29 | 76 | 13 | 6 | 124 pos |
| Total | 35 | 160 | 53 | 163 pos | 411 |
| Total (n=411) n | p16+/p53− (n=132) | p53+ (n=163) | p16−/p53− (n=116) | P value | ||||
|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | |||
| Age, y (mean, SD) | 68.3±14.0 | 63.5±16.1 | 71.6±12.1 | 68.9±12.3 | <.0001 a | |||
| Tumor stage b pT1b | 116 | 44 | 33.3 | 39 | 23.9 | 33 | 28.4 | .154 c |
| pT2 | 202 | 56 | 42.4 | 90 | 55.2 | 56 | 48.3 | |
| pT3/4 | 51 | 22 | 16.7 | 15 | 9.2 | 14 | 12.1 | |
| Unknown | 42 | 10 | 7.6 | 19 | 11.7 | 13 | 11.2 | |
| Nodal status pN0 | 235 | 86 | 65.2 | 80 | 49.1 | 69 | 59.5 | .050 c |
| pN1 | 141 | 37 | 28.0 | 69 | 42.3 | 35 | 30.2 | |
| Unknown | 35 | 9 | 6.8 | 14 | 8.6 | 12 | 10.3 | |
| hrHPV negative | 154 | 6 | 4.6 | 94 | 57.7 | 55 | 47.4 | <.001 d |
| Unknown e | 53 | 13 | 9.8 | 23 | 14.1 | 17 | 14.7 | |
| Positive | 204 | 113 | 85.6 | 46 | 28.2 | 44 | 37.9 | |
| Type 16 | 176 | 101 | 76.5 | 37 | 22.7 | 38 | 32.7 | |
| Type 33 | 15 | 10 | 7.6 | 2 | 1.2 | 3 | 2.6 | |
| Type 18 | 6 | 0 | 0.0 | 4 | 2.4 | 2 | 1.7 | |
| Type 45 | 4 | 1 | 0.8 | 2 | 1.2 | 1 | 0.86 | |
| Other | 2 | 1 | 0.8 | 1 | 0.6 | 0 | 0.0 | |
| Tumor diameter, mm (mean, SD), n=423 | 36.3±30.1 | 39.2±41.5 | 34.8±21.8 | 35.3±24.7 | .512 a | |||
| Depth of invasion, mm (mean, SD), n=312 | 8.5±8.7 | 6.9±5.6 | 9.7±10.8 | 8.4±8.2 | .135 a | |||
| Grading G1 | 42 | 8 | 6.1 | 16 | 9.8 | 18 | 15.5 | .121 c |
| G2 | 232 | 79 | 59.8 | 95 | 58.3 | 58 | 50.0 | |
| G3 | 95 | 35 | 26.5 | 36 | 22.1 | 24 | 23.1 | |
| Unknown | 42 | 10 | 7.6 | 16 | 9.8 | 16 | 10.2 | |
| ECOG 0 | 106 | 42 | 31.8 | 30 | 18.4 | 34 | 29.3 | .009 c |
| 1 | 75 | 21 | 15.9 | 36 | 22.1 | 18 | 15.5 | |
| 2 | 48 | 15 | 11.4 | 28 | 17.2 | 5 | 4.3 | |
| 3 | 17 | 6 | 4.6 | 6 | 3.7 | 5 | 4.3 | |
| 4 | 2 | 1 | 0.8 | 0 | 0.0 | 1 | 0.9 | |
| Unknown | 163 | 47 | 35.6 | 63 | 38.6 | 53 | 45.7 | |
| Type of vulvar surgery | .638 c | |||||||
| Wide excision | 39 | 11 | 8.3 | 14 | 8.6 | 14 | 12.1 | |
| Partial vulvectomy | 130 | 47 | 35.6 | 52 | 31.9 | 31 | 26.7 | |
| Complete vulvectomy | 194 | 59 | 44.7 | 78 | 47.6 | 57 | 49.1 | |
| Exenteration | 3 | 2 | 1.5 | 0 | 0.0 | 1 | 0.9 | |
| Surgery type unknown | 35 | 9 | 6.8 | 14 | 8.6 | 13 | 10.3 | |
| No surgery | 10 | 4 | 3.0 | 5 | 3.1 | 1 | 0.9 | |
| Resection status R0 | 281 | 95 | 72.0 | 108 | 66.2 | 78 | 67.2 | .100 c |
| R1 | 52 | 18 | 13.6 | 24 | 14.7 | 10 | 8.6 | |
| Rx (including no surgery, n=10) | 68 | 19 | 14.4 | 31 | 19.0 | 28 | 24.1 | |
| Type of groin surgery | .397 c | |||||||
| LAE performed | 304 | 99 | 75.0 | 115 | 70.5 | 90 | 77.6 | |
| SLN procedure | 81 | 19 | 14.4 | 40 | 24.5 | 22 | 18.9 | |
| No LAE/unknown | 107 | 33 | 25.0 | 48 | 29.4 | 26 | 22.4 | |
| radiotherapy during primary treatment | .496 c | |||||||
| Yes | 126 | 38 | 28.8 | 57 | 34.9 | 32 | 27.5 | |
| Including vulva | 98 | 34 | 25.7 | 40 | 24.5 | 25 | 21.5 | |
| No | 250 | 86 | 65.2 | 89 | 54.6 | 72 | 62.1 | |
| Unknown | 35 | 8 | 6.1 | 17 | 10.4 | 12 | 10.3 | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
