«PacifiCare’s medical management guidelines represent the recommendation of the PacifiCare Medical Management Guideline (MMG) committee. They are ...»
Coste, 2003; Bergeron, 2001) Accuracy/performance measures: Several studies reported that thin-layer cytology had better sensitivity and/or specificity to detect cervical lesions (LSIL, high-grade squamous intraepithelial lesion [HSIL], and carcinoma) than conventional cytology (Bishop, 1998; Hutchinson, 1999; Khalbuss, 2000; Bergeron, 2001). In these studies, there was a wide range of reported sensitivities, with 53% for SpinThin, 89% for AutoCyte PREP in 2 studies, and 80% and 89% for ThinPrep, but these were still higher than those reported for conventional cytology in these studies. Reported specificities were 58% and 98% for ThinPrep, 38% and 45% for AutoCyte PREP, and 93% for SpinThin, which were higher or equivalent to specificities reported for conventional cytology.
There is substantial evidence that compared with conventional cervical cytology, thin-layer cytology improves specimen adequacy and detection of low- and high-grade squamous intraepithelial lesions (SILs), and good, but less consistent, evidence that thin-layer cytology improves sensitivity, with a relatively small loss, if any, of specificity. Thus, thin-layer preparations are suitable for use in routine cervical cytology screening (Bishop, 1998; Hutchinson, 1999; Minge, 2000;
Coste, 2003; Bergeron, 2001).
Computer-Assisted Screening For Cervical Cancer: The studies evaluated 1 of 2 uses of computer-assisted screening, (1) as a quality control rescreening device for smears previously diagnosed on manual screening as negative or abnormal, comparing it with manual screening or rescreening, or (2) for primary screening, comparing it with manual screening. While many studies used histological confirmation, for all or a subset of slides, as a reference standard, some used consensus diagnosis.
Quality control rescreening to identify false negatives: Of 8 studies that evaluated PapNet for quality control rescreening of previously diagnosed negative slides, 4 were strongly positive, 3 moderately to weakly positive, and 1 reported negative findings (Doornewaard, 1997; Koss, 1997; Mitchell, 1998; O'Leary, 1998; Bergeron, 2000; Rosenthal, 1996; Mango, 1998;
Brotzman, 1999). Three case-control studies, using histological confirmation, reported that PapNet rescreening was effective in detecting clinically significant false-negative smears. (Doornewaard, 1997; Koss,1997; Rosenthal, 1996) A large retrospective study that did not use histological verification confirmed these findings. (Mango, 1998) A retrospective study in a high-risk population reported that PapNet rescreening of smears showed a nonsignificant improvement in sensitivity over manual rescreening, but a significant improvement in specificity of the cytologic diagnoses. However, these performance measures were based on the results of a small number of discrepant cases (n=50) compared with available biopsies. Furthermore, performance measures may be affected by prevalence of the condition in the population. (Bergeron,
2000) Two retrospective studies conducted in low-prevalence study populations reported weaker results. Both identified only a small number of additional abnormalities using PapNet-assisted review of slides considered negative on two manual screenings. (Mitchell, 1998; O'Leary, 1998) A prospective study in a low-risk population reported that the addition of PapNet to routine cervical cytologic screening is unlikely to detect significant lesions with any frequency. (Brotzman, 1999) Algorithm-based rescreening or primary screening: Histological confirmation as a reference standard was lacking in many of these studies; only a few had available biopsy data for all or a subset of the study population. Additionally, in many Cervical Cancer Screening - Commercial Medical Management Guideline studies, at least one author was employed by the manufacturer of AutoPap, thus, potentially biasing results. As with PapNet, the studies selected for review evaluated 1 of 2 uses of AutoPap: (1) as a quality control rescreening device for smears previously diagnosed on manual screening as negative or abnormal, comparing it with manual screening or rescreening, (Patten, 1997; Stevens, 1997; Wilbur, 1996; Renshaw, 2001) or (2) for primary screening, comparing it with manual screening (Wilbur, 1999; Confortini, 2003; Renshaw, 2001; Vassilakos, 2002).
The evidence regarding the automated or computer-assisted screening methods is limited, but suggests that these computerassisted methods appear to be at least equivalent to manual screening/rescreening. There is some evidence that computerized screening using neural network technology, either in the quality control review mode or the primary screening mode, is effective in reducing the false-negative rate and increasing detection of clinically significant abnormalities, compared with primary manual screening and/or manual rescreening of all or a percentage of slides. Likewise, there is some evidence that algorithm-based quality control rescreening increases detection of false negatives compared with manual rescreening. There is more preliminary evidence that, used in the primary screening mode, algorithm-based screening has a high sensitivity to detect cervical abnormalities.
Hybrid Capture HPV Testing for Cervical Cancer: In clinical studies, simultaneous screening using HPV DNA testing for high-risk HPV types and Pap testing was performed using the HC2 High-Risk HPV DNA test.
The HPV DNA test is not intended to substitute for regular Pap screening. Nor is it intended to screen women under 30 who have normal Pap tests. Although the rate of HPV infection in this group is high, most infections are short-lived and not associated with cervical cancer.
An updated search of the peer-reviewed medical literature through July 2007, found a meta analysis of data from previously published North American and European studies of primary screening for cervical cancer in the general population that found that testing for HPV was more sensitive than standard cytology in detecting high-grade cervical intraepithelial neoplasia (96.1% versus 53.0%, respectively), but was also less specific (90.7% versus 96.3%). The combined analysis included 8 nonrandomized trials in which all participants underwent both routine Pap smear testing and HPV testing so that the two screening methods could be compared. Investigators compiled individual patient data (n~60,000) to calculate performance trends for each method across age groups and geographic regions. They found that the sensitivity of HPV testing for detecting CIN2+ remained high and statistically consistent across studies regardless of patient age or where testing occurred. However, the sensitivity of cytology for CIN2+ fluctuated significantly across studies (range, 18.6% to 76.7%), with highest rates noted in women aged 50 years. The specificity of both screening tests improved with increasing patient age. The findings presented in this meta-analysis support expanded use of HPV testing as a primary screening method for cervical cancer, but are weakened by differences in methodology and design between the individual trials included. In addition, these findings do not tell whether improved CIN2+ detection rates with HPV testing, compared with routine cytology, result in improved clinical outcomes for women. Therefore, the superiority of HPV testing over cytology for primary screening of cervical cancer needs to be confirmed by prospective randomized trials in which the two methods are compared head-to-head and participants are followed for an adequate duration to evaluate the impact of screening on morbidity and mortality (Cuzick, 2006).
Institute for Clinical Systems Improvement (ICSI): In a technology assessment report on human papilloma virus (HPV)
DNA testing for the screening and monitoring of cervical cancer, ICSI concluded that (ICSI, 2005):
1. In women 30 years of age or older who have an ASCUS (atypical squamous cells of undetermined significance) cervical cytology result, the HPV DNA test, when performed as a reflex test (where performance of the test is contingent on an ASCUS cytology) is safe and efficacious for use in selecting women for referral to colposcopy and biopsy (positive HPV DNA test) or for monitoring at 6 months and 12 months (negative HPV DNA test). (Conclusion Grade II)
2. In women 30 years of age or older, automatic HPV DNA testing is safe and efficacious for use as an adjunct to cervical cytology to allow less frequent screening for cervical cancer. If both the HPV DNA and the cervical cytology results are negative, the woman is considered at low risk for cervical cancer. Such women may be screened at longer intervals, such as every three years. This conclusion is not applicable for women with a recent change in sexual partner. (Conclusion Grade II) Cervical Cancer Screening - Commercial Medical Management Guideline
3. The evidence does not support the use of HPV DNA testing alone as a primary screening tool. Therefore, HPV DNA testing should not be used as a standalone test in this capacity.
4. There is not enough evidence to permit conclusions regarding the use of HPV DNA testing for the monitoring of treatment response or in selecting patients at high risk of relapse following treatment of CIN or cervical cancer.
5. The test procedure for HPV DNA is safe. The risk lies in what is done with the results. Failure to diagnose high- grade squamous intraepithelial lesions (HSIL) or invasive cervical cancer (false negative) or referral to colposcopy or biopsy when not needed (false positive) may result in morbidity or mortality.
Cervicography Nuovo et al. conducted a systematic overview of literature to assess the usefulness of cervicography (Nuovo, 1997).
Participants included women who had atypical Pap smears or condyloma, or who were referred for colposcopy after abnormal smears, or who were part of a general screening population. Seven studies met inclusion criteria and consisted of 1773 participants. Cervicography showed a high false-positive rate (8.2 to 61% for any dysplasia; 9.8 to 63.4% for highgrade cervical lesions). The positive predictive value for any dysplasia ranged from 21.9 to 44.4% and 5.4 to 31.5% for highgrade dysplasias. The negative predictive value for any dysplasia varied from 63.8 to 96.9% and 88.8 to 100% for high-grade dysplasias. It was concluded that the usefulness of cervicography depends heavily on the approach used to evaluate abnormal findings on a Pap smear.
A randomized controlled study was performed to compare cytology and cervicography as screening methods in 5550 women for the purpose of decreasing the number of pre-malignant lesions found in subsequent screenings. (Autier, 1999) The number of lesions found in subsequent screenings with cervicography decreased by 30%, but did not meet the investigator's goal of 50%. However, the authors concluded that the screening methods were complementary since most lesions found with one method on initial screening were not found by the other method.
The International Academy of Cytology Task Force conferenced on colposcopy, cervicography, speculoscopy, and endoscopy and created a Consensus Position in 1998 (van Niekerk 1998). The task force indicated that the role of cervicography, as a screening device, "remains to be defined." This decision was based on findings that the sensitivity for high grade lesions was no greater than that in cytology and specificity appeared lower. Similar conclusions were reached by Schneider et al. in a population-based study (the Guanacase Project) of cervical neoplasia that was supervised by the National Cancer Institute (Schneider, 1999). The study design involved taking cervigrams on enrollment of 8460 women and comparing results with a referent diagnosis which was determined by histologic analysis, three cytologic tests, and the performance of conventional cytologic evaluation. Eleven cancers were identified by cervicography and sensitivities for the detection of high-grade squamous intraepithelial lesions or cancer were 49.3% overall (54.6% in women younger than 50 years of age and 26.9% in women 50 years of age and older). Cytologic testing, on the other hand, yielded sensitivities of 77.2% overall (75.5% in women younger than 50 years of age and 84.6% in women 50 years of age and older). It was concluded that cytologic testing performed better than cervicography for the detection of high-grade squamous intraepithelial lesions. However, cervicography did perform marginally better than cytologic testing for the detection of invasive cervical cancer. Based on this study, cervicography is not recommended for postmenopausal women. Schneider et al. investigated the optimal performance of cervicography from the initial group of 8460 women participating in the Guanacase Project (Schneider, 2002). In this study, arbitrated cervicography demonstrated an overall sensitivity of 64% and specificity of 94% for the detection of high-grade lesions or cancer. A study published in 2003 also reported on outcomes from the Guanacase Project (Ferreccio, 2003). As part of this ongoing study, investigators evaluated the performance of single and combined visual, cytologic, and virologic tests as screening strategies. As a single test, either liquid-based cytology or human papillomavirus (HPV) DNA testing was significantly more accurate than conventional cytology or cervicography. However, a useful combination was observed between the conventional smear and cervicography.
Ferris et al. studied the efficacy of cervicography in identifying cervical cancer precursor lesions in 3134 women participating in the National Cancer Institute's multicenter atypical squamous cells of undetermined significance and lowgrade squamous intraepithelial lesion triage study (Ferris, 2001). When atypical cervicography was an indication for referral to colposcopy, the sensitivity of cervicography to identify a cervical intraepithelial neoplasia (CIN) 3 lesion was 79% and Cervical Cancer Screening - Commercial Medical Management Guideline would have required referral of 42% of women for colposcopy. The investigators concluded that cervicography performed moderately well at detecting CIN 2 or CIN 3 in study participants. The sensitivity of cervicography was better for younger women.
A retrospective study evaluated cervicography used to triage 1436 women with borderline or mildly dyskaryotic Pap smears (Mould, 2000). Cervicography has a high sensitivity but a low specificity which caused a high proportion of women to be referred for colposcopy. It was concluded that cervicography is not an efficient tool for managing women with minor Pap smear abnormalities.