Abstract. Int J Gynecol Cancer 2006, 16,

Transcription

1 Int J Gynecol Cancer 2006, 16, The distribution and differential risks of human papillomavirus genotypes in cervical preinvasive lesions: a Taiwan Cooperative Oncologic Group Study C.-A. CHEN*, C.-Y. LIUy, H.-H. CHOUz, C.-Y. CHOU, C.-M. HOk, N.-F. TWU{, Y.-Y. KAN#, M.-H. CHUANGy, T.-Y. CHU** & C.-Y. HSIEH* *Department of Obstetrics and Gynecology, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan; ydivision of Biostatistics and Bioinformatics, National Health Research Institutes, Miaoli, Taiwan; zdepartment of Obstetrics and Gynecology, Chang Gung Memorial Hospital Linkou Medical Center, Taoyuan, Taiwan; Department of Obstetrics and Gynecology, College of Medicine, National Cheng Kung University, Tainan, Taiwan; kgynecologic Cancer Center, Cathay General Hospital, Taipei, Taiwan; { Department of Obstetrics and Gynecology, Veterans General Hospital, Taipei, Taiwan; # Department of Obstetrics and Gynecology, Veterans General Hospital, Kaohsiung, Taiwan; and **Department of Obstetrics and Gynecology, Tri-Service General Hospital, Taipei, Taiwan Abstract. Chen C-A, Liu C-Y, Chou H-H, Chou C-Y, Ho C-M, Twu N-F, Kan Y-Y, Chuang M-H, Chu T-Y, Hsieh C-Y. The distribution and differential risks of human papillomavirus genotypes in cervical preinvasive lesions: a Taiwan Cooperative Oncologic Group Study. Int J Gynecol Cancer 2006;16: To clarify the distribution and relative risk of different human papillomavirus (HPV) genotypes in cervical preinvasive lesions, 1246 women with abnormal Papanicolaou smear including atypical squamous cell of unknown significance (ASCUS), atypical glandular cell of unknown significance (AGUS), low-grade squamous intraepithelial lesion (LSIL), and high-grade squamous intraepithelial lesion (HSIL) were enrolled in a multicenter, cross-sectional study. Colposcopy and HPV tests with hybrid capture 2 and polymerase chain reaction reverse line blot were performed. The prevalences of HPV in ASCUS/AGUS-negative histology, ASCUS/AGUS, LSIL, HSIL, and invasive cancer were 33.8%, 38.3%, 74.9%, 84.3% and 100%, respectively, with an overall positive rate of 68.8%. The most common HPV types were HPV 16 (18.5%), 52 (16.5%), 58 (13.2%), 33, 51, 53, 18, 39, 59, 66, MM8, and 31. In comparing the relative risk of HPV infection in different disease status, LSIL and HSIL/carcinoma had a 4.64 (95% CI: ) and (95% CI: ) folds of risk of high-risk HPV infection than the negative group. The same was true in mixed HPV infection, but not in low-risk type infection. Looking into each high-risk HPV type, the relative infection risks for LSIL and HSIL/carcinoma, in comparison with the negative group, were 1.67 ( ) and 8.67 ( ), 2017 ( ) and 3.04 ( ), and 1.40 ( ) and 5.22 ( ) for HPV type 16, 52 and 58, respectively. The study confirmed the high prevalence and risky nature of HPV 52 and 58 in Taiwanese population and conveyed the need to include these HPV types in vaccine development. KEYWORDS: cervical intraepithelial lesion, HPV genotype, HPV test. Address correspondence and reprint requests to: Dr Tang-Yuan Chu, Department of Obstetrics and Gynecology, Buddhist Tzu Chi General Hospital, Tzu Chi University, 707, Sec. 3, Chung-Yang Rd, Hualien City, Hualien, 970, Taiwan, ROC, and Dr Chang-Yao Hsieh, Department of Obstetrics and Gynecology, National Taiwan University Hospital, No. 7, Chung-Shan South Road, Taipei 100, Taiwan. tychu@mail2000.com.tw (T.Y.C.); cachen@ha.mc. ntu.edu.tw (C.Y.H.) doi: /j x Worldwide, cervical cancer is the second most common malignant disease among women. The causal link between human papillomavirus (HPV) and cervical cancer is now established beyond doubt, and the association has become an important model of viral carcinogenesis. The Papanicolaou smear (Pap smear) has been used for the screening of cervical cancer since the 1950s. It is generally accepted that Pap smear screening has played a major role in the reduction of # 2006, Copyright the Authors Journal compilation # 2006, IGCS

2 1802 C.-A. Chen et al. cervical cancer where efforts are organized and medical access is not limited by socioeconomic status. Although the cytologic screening system itself is improving by the cytologists in the reporting system (1), its limitations in the false-negative results are still present. With regard to prevention research, epidemiologists are currently evaluating opportunities to reinforce screening programs by adding HPV testing to cytology. Besides, the evaluation and introduction of HPV vaccines are promising areas for cervical cancer prevention. Recent results obtained with an HPV 16 vaccine have further raised already high expectations (2). Immune prophylaxis has been the cornerstone of preventive medicine, and the only strategy known to have eradicated or greatly reduced some major viral diseases from the human population. However, the research on HPV vaccination is only beginning and will certainly require product development and the conduct of clinical trials for years. HPVs are a heterogeneous group of viruses comprising more than 100 different genotypes, of which at least 40 are found in the genital tract (3,4). While HPV 16 is known to be the most frequent genotype detected from women with cervical cancer worldwide, a few rare genotypes have shown to be prevalent in certain geographic areas (5). In Taiwan, only limited available data described the pattern of HPV genotype distribution in the preinvasive lesions or invasive carcinoma of the cervix (6 8). More data are needed for the understanding of the epidemiology of HPV infection in the population and the implication on vaccine development. In this study, the Taiwan Cooperative Oncologic Group, its member hospitals having served more than 96% of cancer patients including those of gynecological malignancies diagnosed in Taiwan, has conducted a multicenter cross-sectional study to survey the distribution of HPV genotypes in women with cytologically preinvasive cervical lesions, including atypical squamous cell of unknown significance (ASCUS), atypical glandular cell of unknown significance (AGUS), lowgrade squamous intraepithelial lesion (LSIL), and high-grade squamous intraepithelial lesion (HSIL). Through the results of this study, the distribution of HPV genotypes and the role of HPV infection in different severities of cervical lesions can be clarified. Materials and methods From August 1999 to March 2004, women with newly discovered abnormal Pap smear (including ASCUS, AGUS, LSIL, and HSIL) within 3 months of referral from other clinics or from the screening clinics of 12 Taiwan Cooperative Oncologic Group member medical centers/hospitals were enrolled in this study. All the studied women received colposcopic examination and cervical swab to collect cervical sample for HPV tests at the same time after getting the informed consent approved by the Institutional Review Board. Cervical biopsies were not mandatory if the result of colposcopic examination was negative or LSIL. The original Pap smear and histologic slides were sent to the central pathologic committee for review; three cytopathologists and three pathologists were in charge of the review process. Sample preparations Cells from each cervical brush were swirled vigorously into 10 ml of phosphate-buffered saline, then sent to the central laboratory, and stored at 220 C until processing. After thawing, the specimens were processed by placing a 1.5-mL aliquot into a 1.5-mL Eppendorf tube and centrifuging it for 10 min at 13,0003g. The supernatant was immediately removed and discarded with a plugged Pasteur pipette. The cell pellet was dried overnight at room temperature. Then, the pellet was resuspended in 150 ll of digestion solution (10 mm Tris, 1 mm ethylenediaminetetraacetic acid, 200 lg of proteinase K per ml, and 0.1% Laureth-12) and digested at 56 C for 1 h. The digestion was followed by a 15-min incubation at 95 C to inactivate the proteinase K. Crude DNA extracts were stored at 220 C until amplification. Hybrid capture 2 Hybrid capture 2 (HC2) test (Digene Corporation, Silver Spring, MD) was used to examine the existence of HPV DNA in each specimen. The HC 2 assay included a mixture of probes for the following cervical cancer associated HPV types: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68. The US Food and Drug Administration approved threshold of 1 pg of HPV DNA/mL of test solution was used for a positive result. The low-risk probe set (types 6, 11, 42, 43, and 44) was also tested. Polymerase chain reaction methods We used MY09-MY11-HMB01 L1 consensus primers for the amplification system followed by a single hybridization with a reverse line blot detection method (the prototype of Roche linear array HPV test, Roche Molecular System Inc., Pleasanton, CA) for genotyping of the HPV (9). In brief, extracted DNA (100 ng)

3 HPV genotypes in preinvasive lesions of uterine cervix in Taiwan 1803 was amplified by polymerase chain reaction (PCR) with biotin-labeled primers for HPV (PGMY) and b2- microglobulin. Part of the amplified product was visualized by ethidium bromide staining after agarose (1.0%) gel electrophoresis. The integrity of the extracted DNA and the HPV genotype was determined by hybridization with a strip containing probes for 27 HPV types and the b2-microglobulin control and visualized by streptavidin and alkaline phosphatase staining. DNA sequencing was used to verify novel HPV types beyond the detection spectrum of the hybridization. The HPV genotyping strip contains 29 probe lines plus 1 reference ink line, detecting 27 individual HPV genotypes and two concentrations of the b-globin control probe. Two bovine serum albumin conjugated probes per HPV type, corresponding to each of two hypervariable regions within the MY09/MY11 amplicon, are deposited in a single line for each of the following HPV types: 16, 18, 26, 31, 33, 35, 39, 42, 45, 51 59, 66, 68, MM4, MM7, MM8, and MM9. HPV types 6, 11, 40, and the b-globin controls have a single probe deposited per line. Statistical analysis Unconditional multiple logistic regression was applied to estimate the odds ratios (ORs) and adjusted ORs, with 95% confidence intervals (CIs), for the potential risk of different cytologically preinvasive cervical lesions assessed at baseline and different HPV infection (as compared with HPV negative) among participants. Most variables were categorical, and nominal variables were reported as frequencies. The statistical analyses were performed using SAS software (version 8.2; SAS Institutes Inc., Cary, NC). Results A total of 1246 women with abnormal Pap smears including ASCUS, AGUS, LSIL, and HSIL were enrolled in this study. Among the 1246 women, 431 (34.6%) had cytologic ASCUS or AGUS, 442 (35.5%) had LSIL, and 373 (29.9%) had HSIL; since this study was not performed through a population-based screening, and some of the women were referred cases, the percentage of HSIL was much higher than the normal distribution. All the 1246 women had undergone colposcopic examination, and 936 of them had cervical biopsy for histologic examination. The comparison between cytologic and histologic examinations is shown in Table 1. The concordance between cytology and histology was 44.1% in LSIL and 74.7% in HSIL. In about 15% of the cytologic ASCUS/AGUS and LSIL women, colposcopic biopsy revealed HSIL lesions. The 1246 women were then reclassified into five groups according to the most serious cytologic or pathologic diagnosis. There were 148 (11.9%) women with ASCUS/AGUS cytology but histologically negative, 167 (13.4%) women with ASCUS/AGUS cytology but without biopsy, 471 (37.8%) women had LSIL, 445 (35.7%) women had HSIL, and 15 (1.2%) women had invasive cancer. The distributions of each disease in different age groups are shown in Table 2. For women younger than 30 years, the predominant lesion was LSIL (58.0%), and the distribution of cervical lesions in this age group was statistically different as compared with other age groups. Using the HC2 method, the positive rate of HPV DNA among the 1246 women was found to be 60.4% (753 positive and 493 negative), and 727 women were positive for high-risk type HPV. By the PCR method, only 709 (56.9%) women were positive for HPV DNA; there were 79.8% (605 positive and 389 negative) samples tested by HC2 and PCR that produced identical results, and 605 (70.6%) samples were positive by both HC2 and PCR methods in the 857 HPV-positive women. The agreement between HC2 and PCR methods was 0.58 (kappa value; 95% CI ¼ ). The overall HPV positive rate was 68.8% (857 in 1246 women) by either method. Among the 709 women with HPV infection detected by PCR method, 591 (83.4%) women had single Table 1. Histologic diagnosis Comparison between cytologic and histologic diagnosis in 936 women with abnormal Pap smear Cytologic diagnosis ASCUS/AGUS, LSIL, HSIL, Total Negative 148 (54.2) 124 (39.9) 45 (12.8) 317 LSIL 76 (27.8) 137 (44.1) 28 (8.0) 241 HSIL 38 (13.9) 46 (14.8) 263 (74.7) 347 Carcinoma 2 (0.7) 1 (0.3) 12 (3.4) 15 Others 9 (3.3) 3 (1.0) 4 (1.1) 16 Total

4 1804 C.-A. Chen et al. Table 2. Age group (years) The distributions of cervical lesions in different age groups N Negative a ASCUS/AGUS, Uncertain b ASCUS/AGUS, LSIL, HSIL/Carcinoma,,30* (12.5) 10 (8.9) 65 (58.0) 23 (20.5) (10.0) 32 (9.1) 156 (44.6) 127 (36.3) (11.7) 58 (13.3) 158 (36.2) 170 (38.9) (13.8) 46 (21.9) 70 (33.3) 65 (31.0) (13.9) 21 (15.3) 22 (16.1) 75 (54.7) Total (11.9) 167 (13.4) 471 (16.1) 460 (36.9) a Cytologic ASCUS/AGUS but histologically negative. b Cytologic ASCUS/AGUS but without biopsy. *P, , Chi-square test for proportions homogeneous. high-risk type HPV infection, 45 (6.3%) women had mixed high-risk type HPV infection, 39 (5.5%) women had single low-risk type HPV infection, 16 (2.3%) women were mixed high- and low-risk type HPV infected, while in 18 (2.5%) women, the infected HPV types were unclassified. The distribution of the HPV genotypes in different cervical lesions is shown in Table 3. The most prevalent single-type HPV infections were type 16, 52, and 58. According to the phylogenetics of HPV (10), group 58 families (type 33, 52, 58, and 67) were the most frequent HPV types and accounted for 40.6% of the single-type HPV infection Table 3. HPV type The distribution of HPV genotypes by different cervical lesions Negative a ASCUS/AGUS, Uncertain b ASCUS/AGUS, LSIL, HSIL, Carcinoma, Single high-risk type 16 4 (14.8) 6 (15.4) 20 (9.0) 92 (31.7) 9 (64.3) (25.9) 8 (20.5) 44 (19.9) 57 (19.7) 1 (7.1) (18.5) 10 (25.6) 18 (8.1) 58 (20.0) 3 (21.4) (7.7) 5 (2.3) 37 (12.8) (7.4) 1 (2.6) 23 (10.4) 10 (3.4) (10.3) 27 (12.2) 5 (1.7) (3.7) 1 (2.6) 13 (5.9) 5 (1.7) 1 (7.1) (3.7) 1 (2.6) 14 (6.3) 3 (1.0) (3.7) 2 (5.1) 10 (4.5) 6 (2.1) (3.7) 0 16 (7.2) 1 (0.3) (3.7) 1 (2.6) 3 (1.4) 7 (2.4) (2.3) 3 (1.0) (3.2) 1 (0.3) (2.6) 4 (1.8) 2 (0.7) (7.4) 0 4 (1.8) 1 (0.3) 0 7 MM7 1 (3.7) 0 2 (0.9) 1 (0.3) 0 4 MM4 1 (3.7) 1 (2.6) 1 (0.5) 1 (0.3) (1.4) MM (0.9) Total Single low-risk type MM8 2 (40.0) 1 (16.7) 5 (23.8) 5 (71.4) (20.0) 3 (50.0) 3 (14.3) 1 (14.3) (40.0) 1 (16.7) 5 (23.8) (19.0) 1 (14.3) (16.7) 1 (4.8) (9.5) (4.8) Total Mixed type Total 2 (3.3) 5 (8.2) 27 (44.2) 27 (44.2) 0 61 a Cytologic ASCUS/AGUS but histologically negative. b Cytologic ASCUS/AGUS but without biopsy. Total number

5 HPV genotypes in preinvasive lesions of uterine cervix in Taiwan 1805 followed by group 16 families (type 16, 31, and 35) that accounted for 23.8% of the single-type HPV infection, while group 18 family (type 18, 39, 45, 59, 68, and 70) accounted for only 12.2% of the single-type HPV infection. Among the 61 women with mixed-type HPV infections, there were 22 HPV types with 50 different combinations, and the most prevalent types were type 52 (21 women, 34.4%), 53 (16 women, 26.2%), 58 (15 women, 24.6%), 16, and 33 (13 women each, 21.3%). In women with ASCUS/AGUS cytology and negative histology, the HPV positive rate was 33.8% (50 out of 148 women); for women with ASCUS/AGUS cytology but without biopsy, the HPV positive rate was 38.3% (64 out of 167 women, OR ¼ 1.22, 95% CI ¼ ). The HPV positive rates were significantly higher in the samples of LSILs (74.9%, 353 out of 471 women; OR ¼ 5.86, 95% CI ¼ ) and HSILs (84.3%, 375 out of 445 women; OR ¼ 10.50, 95% CI ¼ ). Table 4 shows the positive rates of different risk types of HPV in different cervical diseases, the high-risk type HPV positive rate was significant in the LSILs and HSILs and also the mixed-type HPV infection was related to the LSILs and HSILs. However, the relation between the positive rates of low-risk type HPV and cervical lesions was not statistically significant. We further evaluated the relations among different types of HPV infection and different cervical lesions, the positive rates of HPV types 16 and 58 were only significantly related to HSILs, while type 52 was related to both LSILs (although weak, 95% CI ¼ ) and HSILs. The positive rates of HPV type 18 were not statistically significantly related to different cervical lesions, probably because of the insufficient case number (Table 5). In 431 cytologically ASCUS/AGUS women, 264 had undergone colposcopic biopsy for histologic diagnosis; the HPV positive rate was significantly higher in histologically LSIL and HSIL women than histologically negative women (Table 6). For cytologically LSIL women the HPV positive rates were also significantly higher in histologically LSIL and HSIL women than in histologically negative women (Table 7). Discussion The main purpose of Pap smear was to identify women with cervical lesions that confer an increased risk of cervical cancer. Such women require further follow-up and/or treatment to prevent progression to invasive disease. Consensus exists that women with high-grade cytologic lesions should be referred immediately for further examinations. However, the management of women with minor cytologic lesions such as ASCUS and LSILs remains controversial. These minor cytologic lesions are heterogeneous, as shown in Table 1; the concordance between cytology and histology of ASCUS/AGUS and LSIL was lower than HSIL. These minor cytologic lesions often regress spontaneously and do not require treatment. However, a substantial proportion of them do have histologically confirmed HSILs or worse (11,12). Even in cytologic HSIL women, the histologic concordance was 74.7%, and this means that the colposcopic biopsy was not so sensitive in detecting preinvasive cervical lesions. However, there was no information on the percentage of disease missed by initial colposcopy because this study was not designed as a longitudinal follow-up study. Epidemiologic studies have established HPV infection as the central cause of invasive cervical cancer and its precursor lesions (13). Cervical cancer typically develops over many years after initial infection with oncogenic types of HPV. The peak incidence of HPV infection is usually soon after the initiation of sexual activity and is closely followed by a peak in the incidence of low-grade cervical lesions. Most HPV infections and associated low-grade lesions regress without treatment, but for some women, they Table 4. Total Odds ratio of cervical neoplasia in single high-risk type, single low-risk type, and mixed-type HPV infections Negative b ASCUS/AGUS, Uncertain c ASCUS/AGUS, LSIL HSIL/Carcinoma, High-risk type a positive, 29 (19.6) 44 (26.3) 248 (52.7) 331 (72.0) OR (95% CI) ( ) 4.64 ( ) ( ) Low-risk type a positive, 5 (3.4) 8 (4.8) 30 (6.4) 12 (2.6) OR (95% CI) ( ) 1.81 ( ) 0.77 ( ) Mixed HPV type positive, 2 (1.4) 5 (3.0) 27 (5.7) 27 (5.9) OR (95% CI) ( ) 4.44 ( ) 4.55 ( ) a Including both single and mixed-type HPV infection. b Cytologic ASCUS/AGUS but histologically negative. c Cytologic ASCUS/AGUS but without biopsy.

6 1806 C.-A. Chen et al. Table 5. Total HPV type-specific (HPV 16, 52, and 58) odds ratio for cervical neoplasia a Negative b ASCUS/AGUS, Uncertain c ASCUS/AGUS, LSIL, HSIL/Carcinoma, HPV 16 positive, 5 (3.4) 6 (3.6) 26 (5.5) 107 (23.3) OR (95% CI) ( ) 1.67 ( ) 8.67 ( ) HPV 52 positive, 8 (5.4) 10 (6.0) 52 (11.0) 68 (14.8) OR (95% CI) ( ) 2.17 ( ) 3.04 ( ) HPV 58 positive, 5 (3.4) 11 (6.6) 22 (4.7) 71 (15.4) OR (95% CI) ( ) 1.40 ( ) 5.22 ( ) a The patient number includes both single and mixed-type HPV infections. b Cytologic ASCUS/AGUS but histologically negative. c Cytologic ASCUS/AGUS but without biopsy. ultimately progress to high-grade lesions (14). In our study, the younger women (,30 years) had a significantly higher ratio of LSILs (Table 2), and also there were no differences in the prevalence of HPV infection among different age groups. The data revealed that the incidence of high-grade lesions occurs among women in their late 20s and early 30s, a median of 5 10 years after the peak of HPV infection and occurrence of LSIL that is compatible with the previous study (15). The most common HPV types identified in this study were, in order of decreasing prevalence, HPV 16, 52, 58, 33, 51, 53, 18, 39, 59, 66, MM8, and 31. Other HPV types were detected in no more than 1% of the cervical preinvasive lesions. The HPV type 16 (131 women, 18.5%) was the most prevalent single-type HPV infection that is compatible with other studies around the world (16). HPV types 52 (16.5%) and 58 (13.2%) were the second and third most common types detected in this study; in addition, HPV types 52 and 58 were more prevalent in the LSIL and ASCUS/ AGUS lesions (Table 3), which is same as that previously reported in Taiwan (17). According to a recently published meta-analysis by the International Agency for Research on Cancer, the HPV type 52 is more frequently seen in Japan and Taiwan; however, the HPV type 58 is more prevalent in Japan, Taiwan, Hong Kong, and some southern provinces of China (18). Although HPV type 58 was thought of as an endemic type HPV in these areas, other studies also showed the high prevalence rate of HPV type 58 in Mexican and southwestern American Indian women (19,20). However, in this study, the prevalence rate of HPV type 18 was not as high as in other parts of the world, which only accounted for 5.8% of the LSIL women and 1.7% of the HSIL women. Our data showed that HPV positive rates were significantly higher in women with LSIL and HSIL than ASCUS/AGUS, and that this can also suggest the important role of HPV infection in the process of cervical preinvasive lesions. We further subgrouped the HPV types according to their oncogenic risks (Table 4); the low-risk type HPV infections were not related to different stages of cervical preinvasive lesions in contrast to the high-risk type HPVs. Similar association was found in mixed-type HPV infection since all 61 women had at least one high-risk type HPV. Due to the limitation of case numbers, the association was not as significant as that of the pure high risk HPV infection. The clinical role of multiple types of HPV infection is still unclear; Ho et al. (14), investigating the natural history of cervicovaginal HPV infections in young women, defined an OR of 4.1 (95% CI ¼ ) associated with the presence of multiple types over a 6-month period for persistent HPV infections that is considered as a major factor of progression to cervical preinvasive lesions. Two other studies showed, on the contrary, that persistence of HPV infection was independent of coinfection with other HPV types (8,21). The explanation is that not all PCR-based methods perform equally well in detecting multiple infections because of limitations in the number Table 6. result Histology Odds ratio for LSIL and HSIL/Carcinoma histology in women with ASCUS/AGUS cytology as determined by HPV Negative LSIL HSIL/Carcinoma OR (95% CI) OR (95% CI) HPV negative 98/132 (74.2) 25/132 (18.9) /132 (6.8) 1.00 HPV positive 50/132 (37.9) 51/132 (38.6) 4.00 ( ) 31/132 (23.5) 6.75 ( )

7 HPV genotypes in preinvasive lesions of uterine cervix in Taiwan 1807 Table 7. The correlation between histology and HPV status in cytologically LSIL women Histology Number of cases HPV positive, HPV negative, OR (95% CI) Negative (68.5) 39 (31.5) 1.00 LSIL (84.7) 21 (15.3) 2.53 ( ) HSIL/ Carcinoma (89.4) 5 (10.6) 3.85 ( ) of HPV types detectable and assay performance. Another problem is the difference in sensitivity for distinctive HPV types between different test systems used and the reproducibility of different HPV tests for determining the exact HPV type in the sample. This may especially be true in cases of mixed infections where one type is present in large amounts. The percentage of mixed-type HPV infection was relatively low (8.6%) in this study as compared to other published data. Whether the rate of mixed-type HPV infection in Taiwan is really low, or the reverse line blot assay to detect multiple type HPV infections was not so sensitive, would need further evaluation. However, only a fraction of precancerous lesions progress to invasive cancer. A strong candidate factor for differential progression is HPV type (22,23). Identifying HPV types that preferentially progress from HSIL to invasive cancer has implications not only for the protocols in cervical cancer screening but also for prophylactic type specific HPV vaccine trials. According to International Agency for Research on Cancer metaanalysis (18), the most common HPV types identified in invasive cervical cancer were, in order of decreasing prevalence, HPV 16, 18, 45, 31, 33, 58, 52, 35, 59, 56, 6, 51, 68, 39, 82, 73, 66, and 70. Other HPV types were detected in no more than 0.2%. There was considerable variation in HPV-specific prevalence between squamous cell carcinoma (SCC) and adenocarcinoma. HPV 16 was identified more often in SCC than in adenocarcinoma. Comparison of HPV-specific prevalence in SCC by region, HPV 16 was the predominant type in all regions studied, varying from 45.9% in Asia to 62.6% in North America and Australia. HPV 18 was found consistently in 10 14% of SCC cases. In Asia, HPV 58 (5.8%) and 52 (4.4%) were found more commonly. However, the prevalence rate of HPV types 58 and 52 in SCC might be underestimated because some of the studies included in the meta-analysis might not check all the detectable types of HPV and not all the primer sets amplify individual HPV types with the same sensitivity (24,25). It is also important to know the HPV type distribution in HSIL for ethical reasons, and the final outcome measurement in prophylactic type specific HPV vaccine trials will be the prevention of HSIL. In Table 5, we further compared the ratio of different types of HPV in different severity of cervical lesions; HPV types 16 and 58 were only significantly related to HSIL, and the finding suggested that infections by these two types of HPV were more likely to progress to HSIL. This could be interpreted in two ways: either these types have a greater potential to induce fully malignant transformation or these infections somehow preferentially evade the host immune system. Currently, two sets of prophylactic HPV vaccines undergoing clinical trials do not include HPV types 52 and 58; we suggest that these two types of HPV, especially type 58, should be considered in the strategy of the new HPV vaccine development. There is general consensus that cytologically diagnosed HSILs should be evaluated by colposcopy and biopsy. However, there is currently no consensus as to the appropriate management for women with ASCUS or LSIL. The ASCUS/LSIL Triage Study (26) demonstrated that HC2 testing for high-risk type HPV DNA is a viable option in the management of women with ASCUS. It has greater sensitivity to detect cervical intraepithelial neoplasia 3 (CIN3) or above, and the specificity was comparable to a single additional cytologic test indicating ASCUS or above. In this study, we performed colposcopy and HPV testing for every woman with cytologically ASCUS/AGUS and LSIL. There were 431 women with cytologically diagnosed ASCUS/ AGUS, 264 of them had biopsy for histologic diagnosis, the other 167 women did not have biopsy because the colposcopic examinations were normal or not worse than LSIL. Table 6 showed the significance of HPV prevalence rates according to the severity of cervical lesions. For the 76 women with biopsy-proved LSIL and 40 women with biopsy-proved HSIL or invasive cancer, the HPV positive rates were significantly higher (67.1%[51/76] and 77.5%[31/40], respectively) as compared to the 148 women with negative histology. The HPV test might help in selecting women at risk of cervical squamous epithelial lesions in this group. However, the HPV positive rate was 38.3% in women without biopsy, which was statistically insignificant with the 33.8% in the 148 women with normal histology (95% CI ¼ ). It suggested that HPV test and colposcopy may be equally effective in determining normal histology in cytologically ASCUS/AGUS women. However, the HPV test may have value in the prediction of the outcome of women with ASCUS cytology, and direct colposcopy may used in some clinical settings in the management for women with ASCUS/AGUS cytology instead of routine HPV test. The definition of cytologic LSIL subsumes cytologic features of HPV infection,

8 1808 C.-A. Chen et al. previously termed koilocytotic atypia, and mild dysplasia or cervical intraepithelial neoplasia grade 1. The results of histologic and HPV examinations of the 308 women with cytologically diagnosed LSIL who had undergone colposcopic biopsy were shown in Table 7; although the HPV positive rate was significantly higher in women with LSIL and HSIL than normal histology, the HPV positive rate was not statistically significant between histologic negative/lsil and HSIL/carcinoma (OR ¼ 1.79, 95% CI ¼ ). The HPV test might not be helpful in selecting women who were at risk of HSIL due to the high prevalence rate of HPV infection in this group. The result suggests LSIL cytology was better managed with direct colposcopy, and it is compatible with the ASCUS-LSIL Triage Study report (27). Acknowledgments This study was supported by a grant from the National Health Research Institutes of Taiwan (CRD- TCOG87SUP-02). The authors would like to thank Drs Hsin Nieh, Chiug-Ru Lai, Shih-Ming Jung, Shih-Hung Huang, Ming-Chieh Lin, and Tao-Yeuan Wang for the review of the cytologic smears and histologic sections. We also thank Ms Wei-Lien Feng, Ms Ling-Fang Lin, Ms Yung-Hsin Chin, Ms Yu-Wen Ouyang, Ms Ing-Ing Luo, Ms Te-Chih Wei, Ms Hsing-Hsun Chang, Ms Cheng-Yu Chu, Ms Pei-Shan Chung, Ms Pei-Chyi Lin, and Ms Te-Chih Wei for their help in patient registration and handling of the specimens. References 1 Solomon D, Davey D, Kurman R et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 2002;287: KoutskyLA,AultKA,WheelerCMet al. A controlled trial of a human papillomavirus type 16 vaccine. NEnglJMed2002;347: de Villiers EM. Human pathogenic papillomavirus types: an update. Curr Top Microbiol Immunol 1994;186: Franco EL. Epidemiology of anogenital warts and cancer. Obstet Gynecol Clin North Am 1996;23: Bosch FX, Manos MM, Munoz N et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group. J Natl Cancer Inst 1995;87: Chen TM, Chen CA, Wu CC, Huang SC, Chang CF, Hsieh CY. The genotypes and prognostic significance of human papillomaviruses in cervical cancer. Int J Cancer 1994;57: Lai HC, Sun CA, Yu MH, Chen HJ, Liu HS, Chu TY. Favorable clinical outcome of cervical cancers infected with human papilloma virus type 58 and related types. Int J Cancer 1999;84: Liaw KL, Hildesheim A, Burk RD et al. A prospective study of human papillomavirus (HPV) type 16 DNA detection by polymerase chain reaction and its association with acquisition and persistence of other HPV types. J Infect Dis 2001;183: Gravitt PE, Peyton CL, Apple RJ, Wheeler CM. Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a single-hybridization, reverse line blot detection method. J Clin Microbiol 1998;36: Bernard HU, Chan SY, Manos MM et al. Identification and assessment of known and novel human papillomaviruses by polymerase chain reaction amplification, restriction fragment length polymorphisms, nucleotide sequence, and phylogenetic algorithms. J Infect Dis 1994;170: Kinney WK, Manos MM, Hurley LB, Ransley JE. Where s the highgrade cervical neoplasia? The importance of minimally abnormal Papanicolaou diagnoses. Obstet Gynecol 1998;91: Wright TC, Sun XW, Koulos J. Comparison of management algorithms for the evaluation of women with low-grade cytologic abnormalities. Obstet Gynecol 1995;85: Walboomers JM, Jacobs MV, Manos MM et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189: Ho GY, Bierman R, Beardsley L, Chang CJ, Burk RD. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 1998;338: Schiffman MH. Recent progress in defining the epidemiology of human papillomavirus infection and cervical neoplasia. J Natl Cancer Inst 1992;84: Clifford GM, Smith JS, Aguado T, Franceschi S. Comparison of HPV type distribution in high-grade cervical lesions and cervical cancer: a meta-analysis. Br J Cancer 2003;89: Liaw KL, Hsing AW, Chen CJ et al. Human papillomavirus and cervical neoplasia: a case-control study in Taiwan. Int J Cancer 1995;62: Clifford GM, Smith JS, Plummer M, Munoz N, Franceschi S. Human papillomavirus types in invasive cervical cancer worldwide: a meta-analysis. Br J Cancer 2003;88: Gonzalez-Losa Mdel R, Rosado-Lopez I, Valdez-Gonzalez N, Puerto- Solis M. High prevalence of human papillomavirus type 58 in Mexican colposcopy patients. J Clin Virol 2004;29: Schiff M, Becker TM, Masuk M et al. Risk factors for cervical intraepithelial neoplasia in southwestern American Indian women. Am J Epidemiol 2000;152: Rousseau MC, Pereira JS, Prado JC, Villa LL, Rohan TE, Franco EL. Cervical coinfection with human papillomavirus (HPV) types as a predictor of acquisition and persistence of HPV infection. J Infect Dis 2001;184: Lorincz AT, Reid R, Jenson AB, Greenberg MD, Lancaster W, Kurman RJ. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogenital types. Obstet Gynecol 1992;79: Munoz N, Bosch FX, de Sanjose S et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med 2003;348: Kleter B, van Doorn LJ, Schrauwen L et al. Development and clinical evaluation of a highly sensitive PCR-reverse hybridization line probe assay for detection and identification of anogenital human papillomavirus. J Clin Microbiol 1999;37: Qu W, Jiang G, Cruz Y et al. PCR detection of human papillomavirus: comparison between MY09/MY11 and GP51/GP61 primer systems. J Clin Microbiol 1997;35: Solomon D, Schiffman M, Tarone R. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst 2001;93: ASCUS-LSIL Triage Study (ALTS) Group. A randomized trial on the management of low-grade squamous intraepithelial lesion cytology interpretations. Am J Obstet Gynecol 2003;188: Accepted for publication March 6, 2006