Validity of sampling error as a cause of noncorrelation.

Area of Science:

• Gynecologic pathology
• Cervical cancer screening
• Diagnostic accuracy

Background:

• Papanicolaou (Pap) tests showing high-grade intraepithelial lesions (HSIL) can be discordant with follow-up cervical biopsies.
• Sampling error is frequently cited as the reason for this discrepancy.
• This study investigates the validity of sampling error in cases of HSIL with initially non-diagnostic biopsies.

Purpose of the Study:

• To determine the proportion of initially negative colposcopic cervical biopsies in women with HSIL that are later confirmed to have high-grade cervical intraepithelial neoplasia (HGCIN) via subsequent procedures.
• To validate sampling error as a cause of noncorrelation between Pap tests and cervical biopsies.

Main Methods:

• Retrieved 368 cases of HSIL diagnosed via Pap test between 2003 and 2005.
• Analyzed follow-up cervical biopsy results and subsequent procedures (cone biopsy/LEEP, repeat biopsy, repeat Pap test) for cytologic-histologic correlation.
• Assessed cases with discordant initial biopsy results.

Main Results:

• Of 254 HSIL cases with follow-up biopsies, 146 (57.5%) showed HGCIN.
• Among 108 cases with non-diagnostic biopsies, 47 underwent further procedures.
• Subsequent cone biopsy/LEEP confirmed HGCIN in 19 of 34 cases (56%), repeat biopsy in 5 of 9 (56%), and repeat Pap in 2 of 4 (50%).

Conclusions:

• A 43% discordant cervical biopsy rate for HGCIN was observed in women with HSIL Pap tests.
• In cases with subsequent procedures, 56% of initially negative colposcopic biopsies were followed by histology-proven HGCIN.
• These findings support sampling error as a valid explanation for noncorrelation in HSIL cases evaluated by cervical biopsy alone.