Early Stage Cervical Cancers Containing Human Papillomavirus Type 18 DNA Have More Nodal Metastasis and Deeper Stromal Invasion¹

Cervical cancer is one of the leading causes of cancer in women, worldwide. It accounts for >200,000 deaths/year (1, 2). Papanicolauou smear testing has been an effective screening method for this cancer. However, in many developing countries, cervical cancer remains the leading cause of morbidity and mortality from cancer in women.

A sexually transmitted agent has long been implicated in the cause of cervical cancer. In the mid-1970s, HPV³ emerged as a potential carcinogenic cofactor. Subsequent studies showed that cervical cancer is highly associated with HPV infection. In fact, HPV is detected in >99% of cervical cancers and is now considered the primary etiologic agent in cervical cancer tumorigenesis (3). To date, >100 different HPV genotypes have been cloned and sequenced (4). Among them, ∼35 HPV types infect the female genital tract (5). HPV types 6 and 11 are usually associated with common genital condyloma and thus appear to pose low risk for invasive cervical carcinoma. HPV types 16, 18, 31, 33, and 45 as well as other less frequent types carry a higher risk for cervical cancer.

Different genotypes appear to pose differing prognosis for cervical cancer patients. Some genotypes appear to be associated with more frequent cancer recurrence and thus have a poorer prognosis. We have demonstrated previously the independently poor prognostic significance of HPV type 18 DNA among cervical cancers treated with RH-BPLND. This association with outcome was not statistically significant among women with advanced disease treated with radiation (6).

Along with our previous report, there is growing evidence which supports HPV type 18’s association with a more aggressive form of cervical cancer compared with other HPV types (7, 8, 9). This seems especially significant when the studies examine the early stage cervical cancer patients (6, 9). However, an explanation for this poor outcome is not yet known. Because women with early lesions (Stage IA1 and IA2) are almost always cured and because advanced disease (Stage IIB and greater) represents a very heterogeneous group of patients, we decided to study stage IB cancers because treatment was standardized, and complete pathological data were available. In our present study, we report a complete analysis of the data from our stage IB cervical cancer subgroup treated with radical hysterectomy and lymphadenectomy. In addition to demographic and histological information, treatment and survival data are reviewed.

From April 1983 through August 1993, fresh tumor tissue was collected from patients with cervical cancer who underwent RH-BPLND at the University of California, Irvine Medical Center, Long Beach Memorial Medical Center or at the City of Hope National Medical Center. The study was approved by the institutional review boards at all institutions in accordance with assurances filed with and approved by the United States Department of Health and Human Services.

Previously, based on 291 women with all stages of cervical cancer, we reported that along with lymph node status and depth of stromal invasion, the presence of HPV 18 DNA was found to be an independent prognostic factor. The survival of women with cervical cancers containing HPV 18 was significantly worse when compared with those with non-HPV 18-containing tumors. This appeared especially true in early stage cervical cancers. Since then, others have also noted that HPV 18-containing cervical carcinomas were associated with poor prognosis, especially when diagnosed at an early stage. In the current study, to determine the significance of HPV 18 in early stage cervical cancer, a subset of 144 patients with stage IB were sorted from our earlier dataset. In addition to demographic and histological information, this study differs from our initial report in that it contains subsequent treatment and recurrence data.

Briefly, HPV was typed from fresh frozen tumor tissue samples collected from our stage IB patient group. The tumor tissue was collected from radical hysterectomy surgical specimen and was rapidly frozen at −70°C. Subsequently, high molecular weight DNA was extracted and purified from frozen tumor tissue by homogenization, proteinase K treatment, phenol extraction, ammonium acetate salt displacement, ethanol precipitation, and RNase treatment, as described previously (6). In each case, the presence of invasive cancer in the tissue used for DNA extraction was histologically confirmed by light microscopy on H&E-stained paraffin sections of immediately adjacent tissue. The HPV nucleotide sequences were amplified from purified tumor DNA samples by PCR. Consensus L1 primers (MY09/MY11) and modifications of type-specific primers were used for PCR amplification. HPV typing was then performed by methods described previously (10, 11).

Clinical data for our patient group were obtained from hospital, office, and tumor registry records. The collected data were then analyzed using NCSS 2001 and PASS 2000 by NCSS Statistical Software (Kaysville, UT). Two-tailed Student’s t test was used in comparing continuous variables, such as age and tumor volume. Nominal variables, including demographic information, histological data, and tumor descriptions, were evaluated by the use of χ² tests. Survival curves were estimated by the Kaplan-Meier method, and two-tailed Log-rank test determined the differences. Patients who died of causes unrelated to cervical cancer were censored. By use of Cox proportional hazards model, multivariate analysis was performed to determine the independent impact on survival of HPV type, adjuvant therapy, lymph node status, and depth of stromal invasion.

Among the 144 stage IB tumor samples, 33 cervical tumor specimens contained HPV 18 DNA. Seventy-four samples contained HPV 16 DNA. The remaining 37 tumor specimens were found to have other HPV types or no HPV DNA. The demographics of the stage IB cervical cancer patients were compared in relation to their HPV type. Variables such as age, race, and marital status were compared. None of the demographic variables was found to be statistically significant (Table 1).

The histological factors were then examined. As predicted, HPV 18-containing tumors were more likely to be adenocarcinoma or adenosquamous carcinoma (Table 2). However, no difference was found among different HPV types when evaluating for tumor grade or size. Two important variables were found to be statistically significant. HPV type 18 was associated with deeper cervical stromal involvement when compared with HPV 16 and other HPV types (P < 0.009). In addition, HPV 18-containing cervical cancer patients had higher incidence of pelvic lymph node metastasis (P < 0.03).

After primary surgical management for stage IB cervical cancer, 67% of HPV 18-containing and 49% of non-HPV 18-containing cervical cancer patients received pelvic irradiation as an adjuvant therapy (Table 3). The larger proportion of the HPV 18-containing cervical cancer patients receiving pelvic irradiation compared with the non-HPV 18 patients reflects an increase in intermediate/high-risk pathological findings among HPV 18-positive cancers (12). This difference was not statistically significant. However, cervical cancer recurrences were more common in the HPV 18-containing population than in the non-HPV 18-containing patient group. Eleven of HPV 18 patients (33%) recurred as compared with 18 of non-HPV 18 patients (16%). This difference was statistically significant (P < 0.03). The HPV 18 group had five patients with pelvic recurrences and four with distant recurrences, whereas non-HPV 18 group had 9 and 5, respectively.

Life table analysis of 144 patients with stage IB cervical cancer who were surgically managed with RH-BPLND was done next. Among the variables which were analyzed, HPV type, depth of cervical invasion, and pelvic lymph node involvement had statistically significant 5-year survival differences (Table 4). In agreement with their well-known prognostic importance, deeper cervical stromal involvement of the tumor and pelvic lymph node metastasis were associated with poor 5-year survival for stage IB cervical cancer patients. In addition, the presence of HPV 18 DNA showed worse 5-year survival when compared with those with other HPV types or with no detectable HPV DNA. When multivariate analysis was performed, the lymph node status of stage IB cervical cancer patients was the only variable which was a statistically significant predictor of survival, independent of HPV type, adjuvant therapy, and depth of stromal invasion.

During the mid-1970s, a possible association between HPV infection and the development of cervical cancer was proposed. Since then, continuing research on cervical cancer has solidified HPV’s importance as a major factor in tumorigenesis. To date, >100 different genotypes of HPV have been identified. And many of these HPV types pose a significant risk for development of cervical cancer.

It appears that prognosis may be related to the HPV genotypes, which cause these cancers, resulting in significant survival differences among afflicted patients. We have reported previously a worse prognosis associated with HPV type 18 in early stage cervical cancer patients (6). Similarly, others have also noted higher recurrence rates and more aggressive disease among HPV 18-containing tumors when compared with non-HPV 18 cervical cancers.

The scientific literature investigating the prognostic significance of HPV in cervical carcinomas has yielded mixed results (Table 5). Earlier studies by Barnes and Walker (13, 14) questioned the possibility of differing prognosis among different genotypes of HPV. HPV 18 appeared to have more nodal metastasis and higher recurrence when compared with other HPV types, although the differences were not statistically significant. Others presented data that cervical cancer with no detectable HPV were more aggressive and had a worse prognosis (15, 16). Nonetheless, when reviewing the literature, it becomes apparent that the majority of the reports show higher recurrence rates among the women with HPV 18-containing cervical cancers and an associated worse prognosis (6, 7, 8, 9, 14, 18, 19). This seems especially true in early stages of disease.

On the basis of these recent reports, we decided to perform a subset analysis of stage IB cervical cancer patients treated at our institution and its affiliated sites, as well as a review of the literature. As hypothesized, our data show that in stage IB patients who had undergone radical surgery, HPV 18-containing tumors are associated with a worse prognosis and a higher rate of nodal metastases. As reported in our previous studies, HPV 18 type was also strongly associated with adenocarcinoma histology in this subgroup of stage IB tumors. There were, however, no differences between HPV 18 and non-HPV 18 patients with respect to demographics and tumor grade and size. Of note, a separate statistical analysis between HPV 18 and HPV 16 patients yielded similar results as when the non-HPV 18 patients were used for comparison (data not shown).

The explanation for the worse prognosis among HPV 18-containing cervical cancer patients is not known. A number of molecular explanations have been proposed, including higher frequencies of integration events and enhanced E7 phosphorylation by HPV 18 (22, 23, 24, 25). There has been a long-held view that HPV 18-associated cancers have shorter preclinical detectable phase and rapidly progress through the preinvasive stages of neoplasia (26). In a recent report by Woodman et al. (27), this seemingly shorter precancerous stage may simply be attributable to less severe cytologic changes associated with HPV 18 infection and not with more aggressive natural history. However, our results seem to indicate that the worse prognosis could be secondary to more invasive disease, resulting in increased nodal metastases among HPV 18-containing cancers. Importantly, one must note that the overall size of the tumors was not different between HPV 18 and non-HPV 18 cases. Perhaps, depth is simply a more accurate reflection of disease volume, although all tumor measurements, including size, were based on pathological not clinical assessments. Others have suggested a similar increase in nodal disease among HPV 18-containing cancers, e.g., Barnes et al. had noted that HPV 18 was associated with more nodal involvement, but their result was not statistically significant. Our data represent a much larger group of stage IB patients and show a significant association (P < 0.04). Interestingly, when data from past studies are combined, a statistically significant (P < 0.01) association between HPV 18-containing cancers and increased nodal metastases is demonstrated similar to the current series (Table 6).

Our data showed that a significantly larger proportion of the HPV 18-containing cancers were adenocarcinomas. Although there has long been a debate with regards to possible poorer prognosis for the adenocarcinomas as compared with squamous carcinomas in cervical cancer, it is generally accepted that there is no difference in survival when corrected for stage. The 1998 FIGO Annual Report noted no difference in survival for squamous carcinomas and adenocarcinomas in stage I cancers in its multivariant analysis (28). Thus, histology does not appear to be an important factor in terms of prognosis and survival.

For our current analysis, we were also interested in postsurgical treatment and outcome results of our early stage cancer patients. We hoped to determine whether the HPV 18’s poor survival data resulted from differences in postsurgical treatment of these patients. Our results show that HPV 18 patients had actually received more adjuvant treatment than non-HPV 18 group. However, there were more recurrences among HPV 18 patients. One could argue that HPV 18 patients received more radiation because of more nodal involvement and deeper cervical invasion. When subsets of the patients who did not receive adjuvant therapy were compared, there were still more recurrences among the HPV 18 group. HPV 18 tumors appear to be more aggressive with or without current treatment regimens. The sites of recurrence for the two groups were similar and did not favor pelvic versus distant region.

For HPV 18-containing early stage cervical cancers, radiation does not appear to provide adequate local control when current protocols are followed (29). Higher rates of recurrences are seen for HPV 18-associated cancers whether or not patients were treated with adjuvant radiation. The influence of adjuvant chemotherapy with radiation deserves more study in this setting and may help overcome the adverse affect of HPV 18 (30). Hopefully, as more genomic information about the virus and its association with cervical cancer becomes available in the future, more novel therapies may be discovered for particular types of HPV-associated tumors to improve patient outcome.