|
Are fluid-based cytologies superior to the conventional Papanicolaou test: a systematic review |
Sulik S M, Kroeger K, Schultz J K, Brown J L, Becker L A, Grant W D |
|
|
Authors' objectives To compare fluid-based cytologies (FBC) with conventional Papanicolaou (Pap) tests to determine if either is superior.
Searching MEDLINE, Best Evidence, EMBASE, Biological Abstracts/RRM and the Cochrane Library were searched from 1985 to November 1999. The search terms were reported in the paper. The manufacturers of FBC tests were contacted and reference lists were handsearched.
Study selection Study designs of evaluations included in the reviewThe studies had to compare FBC with Pap tests. The studies eligible for inclusion were: standard cohort studies in which the women received Pap or FBC; split-sample studies where both tests were applied simultaneously to the same women; and diagnostic cohort studies, where the results of the index test were compared with a reference standard.
Specific interventions included in the reviewThe studies had to include FBC (Thin Prep or AutoCyte) and conventional Pap cytology samples. Studies were excluded if they restricted their reports to a subset of Pap results.
Reference standard test against which the new test was comparedFor the assessment of diagnostic accuracy, the studies had to include colposcopy and/or biopsy systematically applied to all women, or to at least 50% of those with abnormal results (i.e. high-grade squamous intraepithelial lesions on either FBC or Pap test). In addition, studies that evaluated at least a random sub-sample of those with normal tests by colposcopy or by independent expert consensus, were also included.
Participants included in the reviewNo inclusion criteria for the participants were stated. Women with a previous abnormal Pap test and women without previous abnormal results, who came from several countries (including the USA, Germany, Costa Rica), were apparently included. However, no systematic details of the participants were reported.
Outcomes assessed in the reviewThe sensitivity, specificity, and area under the ROC curve were assessed for those tests with an appropriate reference standard. In the other studies, specimen adequacy was assessed.
How were decisions on the relevance of primary studies made?Two reviewers independently reviewed the titles and abstracts, and studies were retrieved on the opinion of either reviewer. They then independently reviewed the retrieved articles, with any differences resolved by consensus with a third author.
Assessment of study quality The authors developed a validity assessment that was systematically applied. The studies were scored out of 13, but details of the items scored were not supplied. Two reviewers independently assessed the validity of the included studies, with any differences resolved by consensus.
Data extraction Three reviewers independently extracted the data, with any differences resolved by consensus. The authors of the published papers were contacted for missing data.
The sensitivity and specificity were extracted from studies of test accuracy. In the other studies, specimen adequacy was assessed on the basis of four levels of adequacy: satisfactory, limited by the absence of endocervical cells, limited by other factors (i.e. obscuring inflammatory exudates, blood, thick tests, scant cellularity, or air-drying artifact), or unsatisfactory. The rate difference between the two tests was calculated for satisfactory and for unsatisfactory specimens.
Methods of synthesis How were the studies combined?Summary estimates of sensitivity and specificity were made using the random-effects model of DerSimonian and Laird, weighted by the inverse of the variance. The area under the receiver operating characteristic (ROC) curve was calculated for the FBC and Pap tests.
For specimen adequacy, the rate differences (RD) between the two tests were pooled for each category of adequacy, using a random-effects model if significant heterogeneity was found. Split-sample and cohort studies were also pooled separately within each category, as well as together. The method used in the absence of significant heterogeneity was not stated.
How were differences between studies investigated?For specimen adequacy, a test for statistical heterogeneity was reported to have been applied, but it was not described. The p-level to be applied was stated to be p greater than 0.05, which would appear to be a misprint.
Results of the review Five prospective diagnostic cohort studies that included appropriate reference standards were included in the analysis of test accuracy, with sample sizes totalling 14,569. It appears that 42 studies contributed data on specimen adequacy, although this is not directly stated. Details of these studies, such as study design and sample sizes, were not reported.
All of the studies included some women at high risk of an abnormal result, while two studies recruited only women with a previous abnormal result.
Sensitivity and specificity.
For the FBC test, the sensitivity was 0.90 (95% confidence interval, CI: 0.77, 0.96) and the specificity was 0.85 (95% CI: 0.74, 0.92). For the conventional Pap test, the sensitivity was 0.79 (95% CI: 0.59, 0.91) and the specificity was 0.89 (95% CI: 0.75, 0.96). There was little difference between the areas under the pooled ROC curve: 0.93 for Pap and 0.91 for FBC. The difference between them was 0.017 (95% CI: -0.33, 0.80, p=0.37).
Specimen adequacy.
FBC specimens were more likely to be satisfactory than conventional Pap specimens (RD 0.06, 95% CI: 0.03, 0.09, p<0.05). There was no difference in the proportion of unsatisfactory specimens between the two tests. The FBC tests were stated to have a 6% higher rate of specimens with an absence of endocervical cells (RD 0.06, 95% CI: 0.02, 0.10, p<0.05) and a 10% lower rate of specimens with other deficiencies (RD 0.10, 95% CI: -0.13, -0.06, p<0.05). However, this result appears to be contradicted in the graphical representation of these results (Figure 2), where the higher rate of specimens with absent endocervical cells is shown as better with FBC and the lower rate of those with other deficiencies is shown as worse with FBC.
Estimates of heterogeneity were not reported.
Authors' conclusions For most women there is no reason to replace Pap with FBC. For women at high risk of cervical cancer or who are screened infrequently, the possible increase in FBC sensitivity may outweigh the potential harms from additional false positives. Once reliable estimates of the relative sensitivity and specificity have been obtained, a decision can be made about whether the benefits derived outweigh the advantages of additional costs.
CRD commentary The inclusion criteria relating to the interventions and outcomes were well-defined, but there were no criteria for the participants and a wide variety of study designs were included. The type of participant might be an important factor in the test performance. The reference standards that were eligible were also variable. However, the authors commented on study validity issues, the lack of uniform reference standards, and the lack of data on participant characteristics. A comprehensive search was carried out for papers reported in the English language, but the exclusion of other languages may mean that some studies were missed.
The review processes were good; these should have ensured that the studies were reliably identified, data extracted and assessed for validity. Detailed validity assessments were reported to be on the journal's website, but non-subscribers cannot access these. There were few details of the individual studies in the paper, and the individual study results were not reported. The results of the heterogeneity tests were not reported and so it is impossible to assess the variability in the individual study results, although they appear to be heterogeneous from the ROC curve. Consequently, caution should be exercised when interpreting the pooled results. The authors' conclusions appear to be justified from the results presented.
Implications of the review for practice and research Practice: For most women who are not at high risk for cancer, or who have frequent screening, there is no reason to replace the Pap test with FBC.
Research: Before widespread adoption of the FBC tests for women at high risk or who are screened infrequently, it would be advisable to have additional acceptable reference standard studies, designed to avoid verification bias and to ensure that equivalent specimen collection methods are used.
Bibliographic details Sulik S M, Kroeger K, Schultz J K, Brown J L, Becker L A, Grant W D. Are fluid-based cytologies superior to the conventional Papanicolaou test: a systematic review. Journal of Family Practice 2001; 50(12): 1040-1046 Indexing Status Subject indexing assigned by NLM MeSH Confidence Intervals; Evidence-Based Medicine; Female; Histological Techniques /methods; Humans; Mass Screening /methods; Papanicolaou Test; Probability; Sensitivity and Specificity; Uterine Cervical Neoplasms /diagnosis /pathology; Vaginal Smears /methods AccessionNumber 12002000061 Date bibliographic record published 30/06/2003 Date abstract record published 30/06/2003 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|