Comparison of Efficacy in Abnormal Cervical Cell Detection between Liquid-based Cytology and Conventional Cytology

It is well known that cervical cancer is the second most common cancer affecting women after breast cancer in Thailand. Cervical cancer age-standardized incidence rate is 16.7 in 2007-2009 (Khuhaprema et al., 2013). As known, the goal for cervical cancer screening program is to reduce invasive cervical cancer. The incidence of cervical cancer did not reduce satisfactory due to insufficient screening. Evidence has shown that cervical cytology can increase detect rate of precancerous lesion. Conventional Pap smear has been used in Thailand national cervical cancer screening policy. A systemic review demonstrated that the sensitivity and specificity of conventional cytology (CC) range from 79 to100% (mean 95%) and 30 to 80% (mean 47%), respectively (Nanda et al., 2000). It is widely acknowledged that two thirds of the overall falsenegative rate of CC can be attributed to sampling errors. There are two major components to sampling errors: clinician derived, in which the representative diagnostic cells simply are not collected by the device, and sample preparation derived, in which the representative cells on the collection device are not transferred to the slide. To overcome these problems, liquid-based cytology (LBC), the new collection device and the monolayer or thin-layer


Introduction
It is well known that cervical cancer is the second most common cancer affecting women after breast cancer in Thailand. Cervical cancer age-standardized incidence rate is 16.7 in 200716.7 in -200916.7 in (Khuhaprema et al., 2013. As known, the goal for cervical cancer screening program is to reduce invasive cervical cancer. The incidence of cervical cancer did not reduce satisfactory due to insufficient screening. Evidence has shown that cervical cytology can increase detect rate of precancerous lesion. Conventional Pap smear has been used in Thailand national cervical cancer screening policy. A systemic review demonstrated that the sensitivity and specificity of conventional cytology (CC) range from 79 to100% (mean 95%) and 30 to 80% (mean 47%), respectively (Nanda et al., 2000). It is widely acknowledged that two thirds of the overall falsenegative rate of CC can be attributed to sampling errors. There are two major components to sampling errors: clinician derived, in which the representative diagnostic cells simply are not collected by the device, and sample preparation derived, in which the representative cells on the collection device are not transferred to the slide. To overcome these problems, liquid-based cytology (LBC), the new collection device and the monolayer or thin-layer
Many reports have shown the superiority of LBC over CC (Roberts et al., 1997, Austin et al., 1998, Corkill et al., 1998, Fremont-Smith et al., 2004. LBC was approved by the United States Food and Drug Administration for use in the primary screening of cervical specimens. Nowadays, a number of different LBC techniques are available in Thailand, but it is not widely available and still costly. We use Pathtezt™ because it has never been tested for the accuracy, sensitivity, and specificity compared to the CC.
To evaluate effectiveness of LBC cervical screening test, we compared the efficacy of abnormal epithelial cells detection and quality of smears of LBC to CC.

Materials and Methods
From November 2012 to December 2013, the crosssectional study was performed at Gynecologic Outpatient Unit, Cytology Unit, Chonburi Cancer Hospital (CCH) and Department of pathology, National Cancer Institute (NCI).
Women between 25 and 65 years old who attended for pelvic examination or cervical cancer screening program, gave written informed consent after reviewing a summary of this study plan. Women who had previously undergone any cervical surgical procedures, were pregnant, used any kinds of vaginal preparations or pelvic examination within previous 48 hours were excluded. Endo-ectocervical cell was collected by rotating 5 times in the same direction with cervical broom. This is the split-sample, matchedpair study. After directly smearing cell onto a single glass slide for CC, the leftover cell on broom was rinsed into a vial of specific preservative solution. The CC slides were immediately immersed in 95% ethanol for fixation. The LBC vials were processed by a semi-automated sample preparation instrument. A rotary drive mechanism gently disperses samples. A Pneumatic/Fluidic system was controlled by a microprocessor, monitored cell collection. After the cells are collected on the exterior surface of the membrane, the filter is inverted and gently pressed against the PathTeztTM Microscope Slide. Natural attraction and slight positive air pressure cause the cell to adhere to the slide resulting in an even distribution of cells in a defined circular area. The slides from both techniques were prepared, screened, interpreted by two cytotechnologists at CCH and the results were confirmed by pathologist at NCI using the Bethesda 2001 system. The interpretations of the results from one technique were made without knowledge of those from the other technique. Abnormal cytology was defined as atypical squamous cells of undetermined significance (ASCUS) or higher.
The women who had abnormal cytology would undergo colposcopically directed cervical biopsy and/ or endocervical curettage. Then some treatments were performed later (loop electrosurgical excision procedure, conization and hysterectomy). Hysterectomy was performed in some normal cytological women due to other gynecologic indication not cancer i.e. leiomyoma, endometriotic cyst. The tissues were transported to NCI for pathological diagnosis that was used as the gold standard. This study used the worst histological results.
The results from both techniques were compared quality of slide, accuracy, sensitivity and specificity. Data were analyzed using PASW Statistics 18. Test for difference by McNemar's test. In all tests, results with P-value of less than 0.05 were considered statistically significant.

Results and Discussion
There were 1206 women recruited during the study period. Table 1 demonstrates the comparison of quality and quantity of CC and LBC. There is no difference of squamous metaplasia cell and endocervical cell in both techniques. The LBC significantly decreased RBC, WBC, mucous and thick smear on slides. This results in the decrease of unsatisfactory smear incidence. Bishop et al (1998) showed more than 15% of slide by Pap smear is obscured by RBC, WBC, mucous and thick smear that relate to false negative results.
The comparison of cytological diagnoses between CC and LBS showed that the cytological diagnoses from both techniques agreed exactly in 1154 of the 1206 women (95.7%) and the detection rate of abnormal cytology in both techniques is not different. In this study we chose the spilt-sample study so it is difficult to ensure that the both cytology specimens are comparable, because the    DOI:http://dx.doi.org/10.7314/APJCP.2015.16.16.7381 Comparison of Efficacy in Abnormal Cervical Cell Detection between Liquid-based Cytology and Conventional Cytology smear than CC (Sieber et al., 2008). Table 4 shows the correlation of histologic and cytological diagnoses for each technique in 103 women. 2 AIS, 1 AGC and 3 HSIL women by CC is negative by histology may be due to contaminate with abnormal cell from other slide or missed interpretation on obscured cell, Thus CC false positive is higher than LBC as shown in Table 5. False negative is important. Three women with CIN II-III by histology are negative finding by CC that may be associated with thick smear, obscure factor, and inadequate transfer of the cell onto the slide or the abnormal cell detached from slide in fixation process. Table 5 shows the sensitivity and specificity of the histologic and cytological results for CC and LBC at ASC-US cut off and the comparison of efficacy. CC false positive and false negative are higher than LBC. LBC has shown higher sensitivity, specificity, PPV, NPV and accuracy but no statistical significance. Our LBC result is different from the Siriraj-LBC. The study showed that the Siriraj-LBC increase detection rate of abnormal cervicovaginal cells. This probably caused by decrease false negative but increase false positive from the baseline value by CC. Siriraj-LBC reveal a PPV of 71.1% and a NPV of 42.2% (Laiwejpithaya et al., 2008). Our result was compatible with many previous reports. ThinPrep-LBC (TP) and the CC sensitivity, specificity, PPV and NPV at LSIL cut off were 82.8, 83.0, 88.9, 74.6%, respectively and 89.6, 69.8, 83.0, 80.4%, respectively. TP was significantly more specific for diagnosing lesion than CC. The sensitivity of the two methods was equivalent (Park et al., 2000). Arunratsamee and Siwadune (2012) evaluated the efficacy between LBC and Pap. Accuracy, sensitivity, specificity, PPV and NPV of Cytoneph-LBC and Pap, when using ASC-US as cut off were 76.4, 87.5, 56.7, 59.8, 86.0%, respectively and 70.7,90.4, 37.7, 51.6, 84.3%, respectively. The study found no difference in diagnostic performance on both methods. A systemic review and meta-analysis study showed there was not a significant difference in specificity and sensitivity between LBC and CC  In this study, not all of the study population undergoes the gold standard test. The efficacy parameters were the represent of 103 women who has pathologic results to compare to the new tool. The screening test with higher sensitivity, specificity and cheaper should be used for cervical cancer screening. Larger studies to evaluate the real efficacy parameters and cost-effective analysis of this LBC test is required to improve cervical cancer screening program in Thailand.
In conclusion Pathtezt™ LBC method can improve  specimen quality in a thin layer with relatively free of inflammatory cells and clean background. LBC has shown higher sensitivity, specificity and accuracy at ASC-US cut off in our setting. Therefore the screening performance of LBC is not inferior to the Conventional Pap smear and may be used as an alternative screening method for cervical cancer.