College of American Pathologists
Printable Version

  Perilous times for the Pap test?


CAP Today



February 2008
PAP/NGC Programs Review

David C. Wilbur, MD

The Pap test is dead ...
(poignant pause for effect) ... in 50 years.

—Laura Koutsky, PhD, MSPH, HPV vaccine researcher;
American Society of Cytopathology Annual Meeting keynote speaker, 200
(quotation as remembered by this writer)

The report of my death was an exag¬geration.

Mark Twain, 1897 (Twain died in 1910)

When I began my career in cytology in the early 1980s, one of my mentors conveyed to me what he had been told when he entered the field in the late 1960s. He was told that Pap testing would be gone in 10 years and to “look for something else in pathology” to do with his life. The first great push for gynecologic cytology automation was fully underway at that point, and everyone expected that machines would perform the Pap test in only a few years’ time. As the story goes, automating the complex task of human cellular interpretation turned out to be more difficult than anyone expected. Fortunately, my mentor persevered in his choice of cytology as a career, and he passed his art down to me and many other followers in the ensuing years. So it was with a strong remembrance of his comments that I listened to Dr. Koutsky’s talk in 2002 about the “demise” of the Pap test—with relief that, in fact, human papillomavirus (HPV) vaccination programs alone would require a lengthy “run-in” period, during which the Pap test would be alive and well and maintain its stellar success as the mainstay of cervical cancer prevention.

The reasons for continued Pap-test–based screening, as Dr. Koutsky so elegantly presented, are self-evident. First, the vaccine covers only two specific high-risk types (16 and 18) of the virus, and though these are the most prevalent causes of cervical cancer, they are not the only viral types associated with cancer development. Second, girls 11 to 12 (with allowances for as young as nine) years of age are the primary population to be vaccinated systematically, leaving the older population covered only sporadically. Third, individuals who have already been exposed to the HPV types in the vaccine would obtain only marginal, if any, protection, and the long-term efficacy of the immune response, though promising, is not yet well known. And fourth, only females are being vaccinated at this time, which leaves a substantial reservoir of HPV in the male population. With these points in mind, it is clear that screening for cervical cancer will need to continue in the cohort of individuals not vaccinated, in any individuals within the targeted population not vaccinated (or receiving an inadequate dosage of the vaccine), or in the general population for the cancers not associated with the immunogenic viral types. Of course, further advances such as multivalent vaccines covering more HPV types and a better understanding of the overall protective effect of systematic vaccination will be forthcoming.

Suffice it to say, the vaccine alone will not kill the Pap test—yet.

When articles on HPV testing as an adjunctive technology for use in cervical cancer screening were published recently in the New England Journal of Medicine,1,2 the lay press was replete with headlines related to the incipient demise, or at a minimum the loss of importance, of the venerated Pap test. Headlines such as the following were presented:

  • “HPV Beats Pap as Cervical Cancer Test; HPV Test 40% Better at Detecting Precancerous Cells” (Salynn Boyles, WebMD, Oct. 17, 20073).

  • “Study: HPV Test More Accurate Than Pap Smear” (Rita Rubin, USA Today, Oct. 17, 20074).

  • “DNA Test for Cervical Cancer Gains Support; Accuracy Is Greater Than Pap Exams; False-Positive Fears” (Robert Tomsho, Wall Street Journal, Oct. 17, 20075).

  • “DNA Test Beats Pap Smear in Cervical Cancer Detection, Study Finds” (Raja Jagadeesan, MD, ABC News, Oct. 17, 20076).

Careful reading of the New England Journal articles, however, would not lead one to conclude that the Pap test was in as serious jeopardy as the headlines indicate. Indeed, both studies detailed testing schemes that lead to improved performance over a Pap test alone, but both included the Pap test as a part of the screening package. In the study by Mayrand, et al.,1 the Pap and HPV tests were performed on all women enrolled. HPV screening followed by Pap testing for the HPV-positive population led to the best positive predictive value for disease of any testing combination, had very high negative predictive value (99.8 percent), and had close to the lowest colposcopy referral rate (1.1 percent). Triage to colposcopy based on HPV testing alone yielded very high negative predictive values but referred more women, with lower positive predictive values for true disease. It has been known for some time that HPV testing will drive maximum sensitivity in cervical cancer screening, but because of the high prevalence of HPV in the population, predictive value for disease (and hence referral efficiency) will require a more specific test—at present the Pap test. The conclusion: The combination of the two tests, either as co-testing or in sequence, will provide for optimal screening.

In the second study, by Naucler, et al.,2 the Pap and HPV tests were performed together and compared with the performance of the Pap test alone. This study did not ask a direct question about stand-alone HPV testing. The final conclusions were as follows: 1) more high-grade dysplasias were identified by the combination of the two tests, and 2) subsequent high-grade dysplasias (so-called new or incident cases) decreased in the followup period when both tests were used. Both results lead to the conclusion that HPV testing, again, improves the sensitivity of the testing process for prevalent disease, thereby diminishing its detection in later periods. In summary, neither of these studies indicates that HPV testing should be used alone. The Pap test continues to be a mainstay in the procedure.

Such recommendations have already been incorporated into the most recent cervical cancer screening guidelines of the American Cancer Society,7 which allows co-testing for HPV in conjunction with the Pap test in women over 30 (when transient HPV infection rates decline). The 2006 consensus guidelines of the American Society of Colposcopy and Cervical Pathology8 for the management of cervical abnormalities indicate the use of HPV testing in the management of equivocal (atypical) Pap tests as an effective tool, again using the greater sensitivity of the HPV test to maximize clinical return.

It is tempting to speculate that the HPV test, given its sensitivity, should be a better test than the Pap test. But like most tests, there is an ROC curve that trades specificity for sensitivity. For now, the tried-and-true method of morphology-based Pap testing, used in conjunction with sensitive HPV molecular-detection technology, appears to be the way to go.

Suffice it to say, HPV testing will not kill the Pap test—yet.

So what about the future? In the vaccine era (when much of the screening population has already been vaccinated and assuming the vaccine works as predicted), fewer cases of high-grade dysplasias will be present in the population. Studies have clearly shown that the sensitivity of morphology-based human screening (the Pap test) will decline (see accompanying “Recommended Reading,” page 489). In addition, broad-based high-risk HPV testing will also have lower positive predictive values (there will be fewer infections with the more oncogenic HPV types 16 and 18 in the population). Both of these tests, as currently performed, will be less useful. What test will replace them: type-specific HPV testing, computerized reading of morphology-based tests, new molecular markers of disease, or any combination of the above? Testing as we now perform it will be different.

When will all this occur? Should I find a new career? I think back to the comments of my mentor and wonder.


  1. Mayrand MH, Duarte-Franco E, Rodrigues I, et al. Human papillomavirus DNA versus Papanicolaou screening tests for cervical cancer. N Engl J Med. 2007;357:1579–1588.
  2. Naucler P, Ryd W, Tornberg S, et al. Human papillomavirus and Papanicolaou tests to screen for cervical cancer. N Engl J Med. 2007;357:1589–1597.
  3. Accessed Dec. 20, 2007.
  4. Accessed Dec. 20, 2007.
  5. Accessed Dec. 20, 2007.
  6. Accessed Dec. 20, 2007.
  7. Accessed Dec. 26, 2007.
  8. Wright TC Jr, Massad LS, Dunton CJ, et al. 2006 consensus guidelines for the manage¬ment of women with abnormal cervical cancer screening tests. Am J Obstet Gynecol. 2007;197:346–355.
  9. Schiffman M. Integration of human papillomavirus vaccination, cytology, and human papillomavirus testing. Cancer (Cancer Cytopathol). 2007;111:145–153.

Dr. Wilbur, chair of the CAP Cytopathology Committee, is director of cytopathology, Massachusetts General Hospital, Boston.