Invasive Cervical Cancer After Positive Pap Test but Negative Human Papillomavirus Test

Precis:

The majority of cervical cancers after Pap-test–positive, human papillomavirus test–negative co-tests are symptomatic or clinically apparent.

Introduction

Recent cervical cancer screening guidelines recommend human papillomavirus (HPV) alone (HPV primary screening) over HPV and Pap test cytology cotesting.¹ The objective of this study was to investigate how often a Pap test–positive, HPV-negative result may actually have resulted in a cancer diagnosis that would not have occurred as a consequence of symptoms or clinical examination, and thereby inform clinical practice recommendations.

Methods

Individuals with a cervix aged 25–65 years who had a Pap+ HPV- cotest result within 12 months of histologically confirmed invasive cervical cancer during the years 2003–2019 were identified from the Kaiser Permanente Northern California (KPNC) Laboratory database. Individuals could undergo multiple cotests. A positive Pap test was defined as ASC-US or worse. Cotesting was performed using Hybrid Capture II (Qiagen, Germantown, MD) HPV testing and separately collected cytology as previously described.² Medical records review permitted evaluation of the clinical context in which testing took place. FIGO (2009) staging was used.^3,4 Use of the data was approved by the KPNC Institutional Review Board.

Results

Between 2003–2019, 4,055,057 Pap tests were done as part of cotests. Of these, 54 cases met the inclusion criteria. FIGO stages of these cancers included 12 Stage 1A, 21 Stage 1B1, and 21 Stage 1B2 or greater (IB2+). Of these 54 individuals, 34 (63%) were diagnosed with squamous histology and 20 (37%) with non-squamous histology. Eighteen individuals had no reported symptoms from their cancer at the time of the cotest, had no reported visible or palpable cancer on examination, and had a Pap test diagnosis of low-grade squamous intraepithelial lesion (LSIL) or worse (LSIL+) (Table 1). Two cancers which were asymptomatic and clinically inapparent followed ASCUS HPV-negative cotest results; both were discovered incidentally on pathology from hysterectomy. The remaining 34 individuals either had ASC-US Pap tests or were symptomatic or had a clinically apparent cancer on examination at the time of co-testing, as shown in Table 1. Based on 18 cases among 4,055,057 tests, the prevalence of a Pap LSIL+/HPV-negative result triggering colposcopy preceding a diagnosis of clinically inapparent and asymptomatic cancer is 1:225,281 tests (95% CI, 1:154,083 to 1:418,410).

Table 1.

Human Papillomavirus–Negative and Pap Test–Positive Cases

Pap Diagnosis	Symptomatic or Clinically Apparent	Neither Symptomatic nor Clinically Apparent	Total Cases
ASCUS	5	2	7
LSIL+	29	18	47
Total Cases	34	20	54

Discussion

In this 17-year period, approximately one tested individual per year in a population of 3 to 4 million individual KPNC Health Plan members may have had an asymptomatic, clinically inapparent invasive cervical cancer diagnosed earlier as a consequence of a Pap test result of LSIL or worse in the presence of a negative HPV test. This incidence rate is less than that of other rare cancers for which we do not screen, such as male breast cancer (approximately 7 cases per million population in the US in 2017).⁵ Evaluation of laboratory results alone may lead to substantial overestimation of the number of individuals with invasive cervical cancer who may benefit from earlier diagnosis of clinically occult cancer due to a positive Pap test at the time of a negative HPV test. Use of laboratory databases without the associated clinical information significantly overestimates the magnitude of the risk of discontinuing routine Pap test cytology.