Liquid-Based Cytology Has Significantly Improved Disease Detection in U.S. Cervical Cancer Screening
By R. Marshall Austin, MD, PhD
Advances in cervical cancer screening have always been associated with disagreement and controversy. After publication of his earliest 1928 monograph on the smear technique, Dr. George Papanicolaou was so discouraged by the lack of interest in his smear method by clinicians that he abandoned research for over a decade.1
Even after the Pap smear became widely utilized in the U.S. after World War II and had its effectiveness as a method capable of decreasing cervical cancer incidence, morbidity and mortality confirmed in landmark population studies in Tennessee and Kentucky,2,3 doubters remained. Many of the earliest physicians interpreting Pap smear specimens were gynecologists, since anatomic pathologists were largely skeptical of the method, and adopted cytology only slowly. As late as 1970, peer-reviewed international reports were published questioning the effectiveness of Pap screening and the precancerous character of screen-detected carcinoma-in-situ.4
Liquid-based cytology (LBC), first FDA-approved with the ThinPrep method in 1996 as "significantly more effective than the conventional Pap smear," has been associated over the years with similar debate. Controversy about LBC has come up again recently with publication of a study from Holland in the prestigious Journal of the American Medical Association (JAMA) asserting that the now-dominant LBC screening method is no better than the conventional smear for detection of significant cervical disease.5 In fact, another European study from Italy earlier had reported similar findings.6 These reports from Europe recall the post-1996 reluctance of some U.S. traditionalists to adopt the first significant methodological change in Pap screening since the advent of the smear technique. Gradually, however, the advantages of the new method with immediate wet fixation, removal of obscuring inflammation and harvesting in preservative fluid of increased numbers of cells previously being discarded with the collection devices,7 led the LBC method to largely replace the conventional smear in the U.S. Current estimates submitted by manufacturers to the Securities and Exchange Commission are that the ThinPrep method is being used for roughly 70% of U.S. Pap testing and that the alternative FDA-approved LBC SurePath method is being used for roughly 30% of LBC testing.
Numerous studies have appeared in prestigious journals reporting significantly increased detection of high-grade precancerous lesions with LBC compared to historical controls.8-10 Benchmarking data collected by the College of American Pathologists from hundreds of U.S. laboratories has documented detection rates for precancerous high-grade squamous intraepithelial lesions (HSIL) with the ThinPrep method that have consistently been twice the rates reported with the conventional Pap smear (Table 1). In 2007, Dr. Harriet Smith, a senior gynecologic oncologist at the University of New Mexico, presented data from the U.S. National Cancer Institute's Surveillance Epidemiology and End Results Program showing that since 1996 there has been an accelerated decline in incidence rates for cervical squamous carcinoma (Table 2). Dr. Smith concluded that the decline was most likely due to the widespread use of LBC in U.S. Pap screening.11
College of American Pathologists Laboratory Accreditation Program
Checklist Item CYP.07600 2006 Benchmarking Data
As with the original smear method, skeptics remain. As other developed countries have assessed adoption of the LBC method, two international meta-analyses have employed study exclusion criteria that have discarded from consideration most favorable published LBC studies from the U.S. and have concluded, similar to the recent JAMA article, that LBC is no better than the smear method after all .12,13 The lead author of at least one of these studies, an Australian researcher in the Screening and Test Evaluation Program, School of Public Health, University of Sydney, later significantly reversed course and reported in 2007 in the British Medical Journal that ThinPrep slides screened with computer-imaging significantly increased detection of high-grade cervical dysplasia compared to the conventional smear method, based on a large, well-designed, independent split-sample trial.14
More recently, discussions have focused on international studies and clinical trials comparing the ability of cytology screening with that of HPV DNA testing to detect cervical precancer. In trials comparing the conventional smear method and HPV DNA testing, sensitivity for detection of cervical precancer with HPV DNA testing using the FDA-approved Hybrid Capture 2 (HC2) method has been estimated in the range of roughly 90-95%15 compared to variable lower performance with the conventional Pap smear, ranging from as low as 20-40% in Germany to a high of 77% in England.16 In contrast, verification bias-adjusted studies comparing the ThinPrep LBC method with HC2 in England,17 Costa Rica,18 China19 and the US20 have reported sensitive detection of cervical precancer in the same range achievable as HC2 HPV DNA testing. The advantage remaining for cytology has been the positive predictive value and specificity of the cytologic method, which enables cost-effective detection of specific and clinically significant cervical disease conditions.21
Despite the evidence and extensive experience now supporting use of LBC, optimal implementation of LBC methods requires adequate training and experience for both cytotechnologists and pathologists interpreting the slides. When training and experience are insufficient, improvement with LBC over the smear technique may not occur, resulting in variable laboratory results noted over a decade ago in early split-sample studies in the U.S.22 Improvements with LBC can also be expected to be less in laboratories where conventional Pap smear performance is optimal. Despite the failure of the JAMA Dutch study to report optimal benefits with LBC achieved elsewhere, the advantages of LBC are still widely recognized in Holland, as reflected in the estimated 70% market share enjoyed there by the LBC method. Furthermore, LBC significantly decreased inadequate samples in the Dutch study and was also acknowledged as a useful platform for both computer-assisted screening and molecular testing, neither of which were evaluated in the study.
In most venues, the majority of bench-level cytotechnologists and pathologists performing testing are the first to understand the improvements and advantages of LBC. Once professionals performing the difficult work of cervical screening become experienced and familiar with the LBC method, they are usually reluctant to consider returning to the smear method. Research experts with no or very limited experience serving patients presenting for cervical screening and being personally accountable for the delivery of quality screening services (such as the authors of the accompanying editorial23) may not as easily appreciate the advantages of LBC.
Dr. Austin is professor of Pathology and Medical Director of Cytology, Magee-Womens Hospital of University of Pittsburgh Medical Center, Pittsburgh, PA.
View References on page 4...