Detection of Precancerous Cervical Lesions Is Differential by Human Papillomavirus Type

Human papillomavirus type (HPV)16 is the major carcinogenic HPV type found in cervical precancers and cancers (1–3). Although HPV18 is the second most common type in invasive cervical cancers, it is underrepresented in precancerous lesions as predicted by the prevalence of these types in cancer (1, 3–7). Whether this is due to the propensity of HPV18 lesions to progress faster from precancer to invasive cervical cancer or that HPV18-positive precancers may be preferentially missed by screening programs due to occult pathology is unknown. Using data from the atypical squamous cells of undetermined significance (ASCUS) and low-grade squamous intraepithelial lesion (LSIL) triage study (ALTS), we investigated whether diagnosis of lesions due to HPV 18 (and closely related HPV45) is delayed compared with HPV16 lesions.

ALTS (8) was a randomized clinical trial conducted to evaluate management strategies for ASCUS and LSIL, which was approved by the local and National Cancer Institute Institutional Review Boards. ALTS compared three management strategies for women with ASCUS (n = 3,488) or LSIL (n = 1,572) Pap: (a) immediate colposcopy (IC arm; referral to colposcopy regardless of enrollment test results); (b) HPV triage [HPV arm; referral to colposcopy if the enrollment HPV result was positive by hybrid capture 2 (hc2; Qiagen) or missing, or if the enrollment cytology was high-grade squamous intraepithelial lesion]; or (c) conservative management [CM arm; referral to colposcopy if the enrollment ThinPrep cytology (Cytyc Corporation) was high-grade squamous intraepithelial lesion]. Details of the pelvic examinations, 2-y follow-up, clinical management, and diagnostic procedures have been described in detail elsewhere (8–12). An exit examination with colposcopy was scheduled for all women.

Only women who underwent an enrollment colposcopy (n = 2,725; 1,825, 722, and 178 from the IC, HPV, and CM arms, respectively) were included in this analysis. The outcome was cervical intraepithelial neoplasia grade 3 or worse (CIN3)+ diagnosis by either the clinical center pathologists or Pathology Quality Control group during the 2-year duration of ALTS. The timing of the CIN3+ diagnosis was categorized as enrollment, follow-up (6, 12, and 18 mo follow-up visits), or exit. Given the limitations of colposcopy to detect small lesions typically found in young women such as those enrolled in ALTS (11, 13), we assumed that most CIN3+ were likely to be prevalent regardless of when it was detected, although a small percentage were certainly incident (14). Thus, our main exposure of interest was HPV risk group at enrollment, which was determined by testing cervical STM specimens for 27 or 38 HPV genotypes using a L1 consensus primer-based PCR assay (15, 16). Women were assigned to an a priori enrollment HPV risk group according to cervical cancer risk: positive for HPV16, else positive for HPV18 or 45 (HPV18/45; because HPV18 and 45 are genetically most similar and shown to behave similarly, we combined them), else positive for other carcinogenic types (31, 33, 35, 39, 51, 52, 56, 58, 59, 66, or 68), else positive for noncarcinogenic HPV types, or PCR negative. To ensure that the hierarchical categorization did not bias the results by favoring HPV16, we repeated our analyses restricted to women with a single HPV type infection. To confirm our findings and assumptions, we also reanalyzed the timing of diagnosis excluding probable incident diseases by limiting the analysis to women with same HPV type at enrollment and diagnosis. We used two-way tabulation to calculate Pearson's χ² test; P values of <0.05 were considered statistically significant.

Among 472 women diagnosed with CIN3+ over the 2-year duration of ALTS, we excluded 18 CIN3+ (3.9%) that were missing enrollment HPV results and 42 CIN3+ (8.9%) that were coinfected with HPV16 and HPV18 or 45 for clarity, leaving 412 CIN3+ (87.3%) in the primary analyses. Among the 412 CIN3+, 210 (51%) were HPV16 positive, 42 (10%) were HPV18/45 positive, and 121 (29%) were other carcinogenic HPV positive at enrollment.

Table 1 presents cumulative CIN3+ diagnosis by timing of diagnosis and enrollment HPV risk group. Regardless of HPV risk group, most CIN3+ were diagnosed at the enrollment visit, followed by the exit visit. However, 85% of HPV16-positive CIN3+ were diagnosed at enrollment, compared with 57% of HPV18/45 and 58% of other-carcinogenic CIN3+. In contrast, only 7% of HPV16-positive CIN3+ were diagnosed at exit, compared with 31% of HPV18/45 and 22% of other-carcinogenic CIN3+ (P < 0.001). Similar patterns were observed when stratified by study arm (P < 0.001; Table 1) and, when examined, limiting to detection of the same HPV type at enrollment and diagnosis (i.e., presumed prevalent cases, P < 0.001; Table 1) and when the analysis was restricted to single-type HPV infections (data not shown).

We present Table 2 showing the association of HPV16, HPV18/45, and other carcinogenic-HPV infections at enrollment by factors that could influence detection or progression of lesions. Other than timing of diagnosis (P = 0.001), only referral cytology (ASCUS versus LSIL; P = 0.02) and the likelihood of having a high-grade colposcopic impression (P = 0.02) were associated with HPV risk group. Stratification on referral status revealed similar patterns (data not shown). Among those who had a low-grade colposcopic impression, 38% of the HPV18/45-positive CIN3+ were diagnosed at exit compared with 10% of the HPV16-positive CIN3+ and 27% of the lesions by other carcinogenic-HPV (P < 0.001). Interestingly, among women with a high-grade colposcopic impression at enrollment, there was little difference in the timing of the CIN3+ diagnosis: 9% of the HPV18/45-positive CIN3+ were diagnosed at exit compared with 4% of HPV16-positive CIN3+, and 5% of other carcinogenic-HPV CIN3 (P = 0.7; data not shown).

Thus, we found that overall there is a deficit of HPV18/45-related CIN3+ and that timing of diagnoses differed by HPV risk group, i.e., a far greater proportion of HPV16-related CIN3+ were diagnosed at the enrollment visit compared with the other HPV risk groups. Conversely, a higher proportion of HPV18/45-related and other carcinogenic HPV–related CIN3+ were diagnosed at study exit than HPV16-related CIN3+. This report is consistent with a previous study that found the timing of CIN3+ diagnosis was delayed among HPV18-positive women (4).

These results may indicate a difficulty of detection and diagnosing HPV18/45-related precancerous lesions because HPV18/45 lesions are poorly detected by cytology and colposcopy. Although HPV18/45-related precancerous lesions were less likely to be associated with high-grade colposcopic impression as shown here and previously reported (17), when they were apparent to the colposcopist, the timing of diagnosis was similar to HPV16-related lesions with high-grade colposcopic impression.

Overall, 2-year cumulative CIN3+ was 5.0 and 1.7 times more likely to be HPV16-positive than HPV18/45 and other carcinogenic-HPV, respectively (Table 3). However, when examined cross-sectionally, we found that this ratio was 7.5 and 2.6, emphasizing the underdiagnosis of HPV18/45-related CIN3+ and other carcinogenic HPV-related CIN3+ cross-sectionally compared with HPV16, thus confirming what is observed in other studies (1, 3–7). Also, when compared with the prevalence of HPV risk groups in cancers (3), the prevalence of HPV18 in 2-year cumulative CIN3+ is still less than expected (P < 0.001), whereas the prevalence of other carcinogenic HPV genotypes is actually greater than expected (P < 0.001). Although the fraction of HPV16-positive and other carcinogenic HPV-positive CIN3+ group were comparable with literature reports of 55% and 20% of cancers attributable to HPV16 and other carcinogenic HPV types, respectively, the fraction of HPV18-positive CIN3+ was still less than expected based on the 17% of cancers attributable to HPV18 (3).

These results add to the growing literature on the natural history of HPV genotypes. HPV16 is clearly the most potent HPV carcinogen. Although HPV18/45 behaves like HPV16 in terms of its carcinogenicity (i.e., number of cancers per infection) and is the second and third most common type found in cervical cancers (3), the pattern of later detection of HPV18/45-associated CIN3 is similar to other, weaker carcinogenic HPV types.

A limitation of this study is the referral population basis of ALTS, which was restricted to women with ASCUS or LSIL cytology, and not high-grade squamous intraepithelial lesion. High-grade squamous intraepithelial lesion is preferentially caused by HPV16 and less often caused by other genotypes (6) and is an indicator of an obvious precancerous lesion. Therefore, the difference in timing of diagnosis observed between HPV risk groups in ALTS may underestimate the true differences in routine screening. A second limitation is that there were only 2 years of follow-up in ALTS. A longer follow-up may have increased our ability to find more of the HPV18/45-related CIN3+.

In conclusion, these data extend previous findings (6, 17) that HPV 18/45 lesions are underrepresented in precancers, underscoring the need for better ascertainment of HPV18/45-related lesions. Whether this can be accomplished by using HPV testing or HPV genotyping for screening and management warrants further investigation.

No potential conflicts of interest were disclosed.

Grant support: National Cancer Institute, NIH, Department of Health and Human Services contract nos. CN-55153, CN-55154, CN-55155, CN-55156, CN-55157, CN-55158, CN-55159, and CN-55105. Some of the equipment and supplies were donated or provided at reduced costs by Digene Corporation, Cytyc Corporation, Denvu, Roche Molecular Systems, and TriPath Imaging, Inc.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.