Page 135 - Screening for Cervical Cancer: Systematic Evidence Review

Appendix C. Evidence Tables

Evidence Table 2. New Methods for Preparing or Evaluating Cervical Cytology (cont'd)

Patients & Methods Outcomes Measured Study Results & Limitations* Quality Considerations
ThinPrep 2000
47 patients with known No blind interpretation Independent reference Quality Score=3.5
cervical cancer pathologist only for samples Ref. Std: 2
Ability of ThinPrep from ThinPrep 2000 study only Blind: 0
and conventional Verification: 0
smear to identify Study only evaluated cancer Consecutive: 0
cancerous cells as patients, therefore unable to Spectrum: 0
cancer estimate specificity Publication: 1
Industry: .5
ThinPrep identified 21/22
smears as cancer (1 was
ASCUS)

Conventional smear identified
19/22 as cancer (1 was
ASCUS, 2 were
unsatisfactory)

Uncertain sample Specimen adequacy No reference standard; Quality Score=1.5
collection. and yield of abnormal compared only to historical Ref. Std: 0
Spectrum not described diagnoses cohort of conventional smear Blind: 0
samples. Prevalence of Verification: 0
Thresholds for disease unknown Consecutive: 0
ThinPrep and Spectrum: 0
conventional smear of Histologic verification of most Publication: 1 Industry: .5
ASCUS/AGUS, LSIL, ThinPrep HSIL
SIL NOS, HSIL
ThinPrep resulted in more
abnormal diagnoses
(ASCUS+) than historical
conventional smear, and less
unsatisfactory or limited
smears

PPV of ThinPrep for
HSIL=93.2%
Historical Positive Predictive
Value of conventional smear
for HSIL=78.8%

C-41