Table 1.
Characteristics and comparison of the studies assessing the HPV self-sampling devices’ diagnostic accuracy.
| Authors/ year | Location | Study design | Participants (N) | Age (years) |
Intervention description |
Tests performed to compare self-samples and CCS for identification of diagnostic accuracy |
Reliability of self-collected samples |
Conclusion | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Self-sampling device | Sample type |
Sampling approach | HPV-genotyping test used | Cervical cytology or histology test/interpretation | Concordance with CCS, (% and/or κ), [95 % CI] | SN (ratio or %), [95 % CI] | SP ratio or %), [95 % CI] | ||||||
| Aranda Flores et al., 2021 | Mexico/Mexico City | Randomized clinical rtial | 505 | 30–65 | 1.XytoTest medical Device; 2.Cervex-Brush |
1. Vaginal smear; 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | Abbott RealTime HR HPV test | LBC (ThinPrep medium)/Bethesda system | 78.2 %, κ = 0.34, p < 0.001 | Not reported | Not reported | Fair agreement of HPV positivity rates between the self-collected and CCS |
| Chan et al., 2023 | Hong Kong | Prospective study | 104 | 30–65 | 1. Cepillo Endocervical/Cervical Brush/Cyto-Brush + DNA sample storage card; 2. Cervex-Brush |
1. Vaginal smear; 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | 1.SentisTM HPV Assay (Sentis); 2. BD OnclarityTM HPV Assay (Onclarity) |
Not preformed | 1. 89.8 %, κ = 0.769; 2. 84.4 %, κ = 0.643 |
Not reported | Not reported | “A substantial agreement” between the self-collected and CCS |
| Chen et al., 2016 | China/ Shanghai | Case-control | 101 cases and 101 controls | 21–79 | 1.Evalyn Brush; | 1.Vaginal smear; 2.Cervical smear; |
1.Self-sampling vs. 2.Clinician-taken sample | RealTime RT PCR | Colposcopy with cervical histology/CIN system | 97.5 %, κ = 0.95 |
Not reported | Not reported | Self-sampling and CCS showed good diagnostic agreement and a very high HR-HPV positivity rate |
| Geraets et al., 2013 | Spain/Barcelona | Not reported | 182 | 17–76 | Viba brush | 1.Vaginal smear 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | HPV SPF10 PCR-DEIA-LiPA25 version | LBC (PreservCyt solution) + Colposcopy/Bethesda system | 89 % κ = 0.733 | 95.9 % | 42.9 % | HPV self-sampling might be valuable when a LBC cannot be used, but requires further investigation |
| Guan et al., 2013 | China/ Shanxi Province | Not reported | 2,500 | 30–59 | FTA Elute card | 1.Vaginal smear 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | HPV PCR, Roche HPV Linear Array | Not preformed | 91 % κ = 0.75 | Not reported | Not reported | Self-sampling wih FTA Elute cartridge showed high concordance rate with CCS |
| Enerly et al., 2016 | Norway/Oslo area | Cross-sectional | 267 | 25–69 | 1.Evalyn brush; 2. Delphi Screener |
1.Vaginal smear; 2.Cervical smear; |
1.Self-sampling vs. 2.Clinician-taken sample | CLART1HPV2 test Vs. Digene1HC2 | Not performed | 89.9 %, κ = 0.61 | Not reported | Not reported | Delphi Screener and the Evalyn brush had satisfactory samples concordance rate |
| Ertik et al., 2021 | Germany/Hannover | Prospective multicenter phase II trial (CoCoss-Trial) | 65 | 24–76 | 1.Evalyn-Brush; 2. FLOQSwab; 3. Colli-Pee FV-5000 |
1. Vaginal smear; 2. Vaginal smear; 3. Urine sample |
Self-sampling vs. Clinician-taken sample | Abbott RealTime High Risk HPV Test | Colposcopy with cervical histology/CIN system | 1. κ = 0.48 2.κ = 0.29 3.κ = 0.34 |
1.89.7 %; 2. 82.8 %; 3. 77.6 % |
1. 42.9 %; 2. 71.4 %; 3. 57.1 % |
No significant differences in SN or SP for CIN 2 + detection between the self-smears and CCS |
| Gibert et al., 2023 | Spain, Illes Balears | Cross-sectional | 120 | 40–51 | 1. Viba-Brush; 2. Mía by Xytotest; 2. Rovers Cervex-Brush |
Gibert et al., 2023 | Spain, Illes Balears | Cobas® HPV test | LBC (ThinPrep medium)/Bethesda system | 1. κ = 0.83 2.κ = 0.86 |
95.7 % | 1.88.9 % 2.91.7 % |
Agreement for HPV detection between self –collected and CCS samples was very good |
| Inturrisi et al., 2021 | The Netherlands | Case-control | 30,808 cases and 456,207 controls | 30–60 | 1.Evalyn-Brush; 2. Cervex Brush |
1. Vaginal smear; 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | Cobas HPV Test | LBC (ThinPrep medium)/CISOE-A classification | Not reported | 1. ratio = 0.88; 2. ratio = 0.94 % |
ratio = 1.02 | High accuracy of HR-HPV detection in self-collected samples compared to CCS |
| Islam et al., 2020 | Kenya/ Mombasa | Cohort | 400 | 19–66 | 1. Evalyn-Brush; 2. Viba brush |
1.Vaginal smear 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | APTIMA® HPV Assay (Hologic Inc, San Diego, USA) | CC (Aptima media, Hologic, San Diego, USA) |
86.4 % | 93 % | 66 % | Self-sampling is a “viable option” for HR-HPV mRNA testing |
| Jaworek et al.,2018 | Czech | Cross-sectional | 1,198 | 17–72 | 1.Evalyn brush | 1.Vaginal smear | 1.Self-sampling vs. 2.Clinician-taken sample | 1.Cobas 4800 HPV Test 2. LMNX Genotyping Kit HPV GP |
Not performed | 1. κ = 0.970; 2. κ = 0.906 |
1. ratio = 0.983; 2. ratio = 0.897 |
1. ratio = 0.992; 2. ratio = 0.989 |
CCS and self-samples were highly sensitive and specific for HR-HPV detection |
| Jentschke et al., 2013 | Germany/Hannover | Not reported | 140 | 16–68 | Delphi Screener | 1.Vaginal smear; 2.Cervical smear; |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | HPV DNA detection by HC2 |
Pap-test (PreservCyt solution)/Bethesda system | κ = 0.51 | 64.5–70.6 % | 31.7–38.2 % | The study shows that self-sampling with cervicovaginal lavage with ELISA is not suitable for the detection of high-grade CIN |
| Jentschke et al., 2016 | Germany/Hannover | Not reported | 136 | 17–78 | 1.Evalyn Brush; 2.Qvintip |
1.Vaginal smear; 2.Cervical smear; |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | Abbott RealTime HighRisk HPV test | Colposcopy with cervical histology/ CIN system | 1. 91.2 %, κ = 0.822; 2. 89.0 %, κ = 0.779 |
1.89.8 %; 2. 83.7 % |
1. 66.7 %; 2. 69.0 % |
Reliability of self-samples has no significant difference compared with CCS |
| Katanga et al., 2021 | Tanzania/Kilimanjaro region | CONCEPT | 464 | 35–54 | 1.Evalyn-Brush; | 1. Vaginal smear; 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | HPV detection by HC2, QIAGEN |
Not performed | 90.5 % | 61.4 % | 97.3 % | Self-sampling “seems to be a reliable alternative” to CCS |
| Ketelaars et al., 2017 | The Netherlands/Dutch population of Nijmegen | Cross-sectional | 2,460 | 30–60 | 1.Evalyn brush; 2. Rovers Cervex-Brush |
1.Cervical smear; 2.Cervical smear; |
1.Self-sampling vs. 2.Clinician-taken sample | Cobas 4800 HPV | LBC (ThinPrep medium)/ CISOE-A classification | 96.8 % | Not reported | Not reported | Self-sampling with the Evalyn Brush showed a high concordancewith CCS, |
| Klischke et al., 2021 | Germany/Hannover | Cross-sectional | 87 | >18 | 1.Evalyn-Brush; | 1. Vaginal smear; 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | GynTect® and Abbott RealTime HighRisk HPV assay | LBC (ThinPrep medium) | κ = 0.394, p < 0.001 |
26.1 % | 95.6 % | The results of the self-collected samples differed clearly in comparison to the CCS |
| Latsuzbaia et al., 2022 | Belgium/not specified | The VALHUDES framework | 486 | 31–49 | 1. Coli-Pee; 2.Multi-Collect swab 3.Evalyn Brush; 4.Qvintip; 5. Cervex-Brush |
1.Urine sample; 2. Cervical smear; 3. Vaginal smear; 4. Vaginal smear; 5. Cervical smear |
1–4. Self-sampling vs. 5.Clinician-taken sample |
Abbott RT HPV test | LBC (ThinPrep medium)/Bethesda system | 3 and 4. Hr-HPV 87.65 %, κ = 0.748; 3 and 5. 89.04 %, κ = 0.774 |
3 and 4. ratio = 0.92; 3 and 5. ratio = 0.95 |
3 and 4. ratio = 1.04; 3 and 5. ratio = 1.11 |
Self-collected samples give similarly accurate result with CCS for CIN |
| Leeman et al., 2017 | Spain | Cross-sectional | 91 | ≥18 | 1.Colli-Pee™; 2. Evalyn brush |
1.Urine sample; 2.Vaginal smear; |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | SPF10-DEIA-LiPA25 assay and GP5+/6+- EIA-LMNX |
LBC (PreservCyt solution)/Bethesda system | κ = 0.85 | 100 % | 33 % | High concordance between self-collected and CCS samples was found |
| Leinonen et al., 2018 | Norway/South East region population | Cross-sectional | 310 | 21–80 | 1.Evalyn brush; 2. FLOQSwab |
1.Vaginal smear; 2. Vaginal smear; |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | Anyplex™ II HPV28, Cobas® 4800 HPV Test and Xpert®HPV |
LBC (ThinPrep medium)/ Bethesda system | 1. 94 %, κ = 0.68; 2. 87.9 %, κ = 0.50 |
1. 91–95 %; 2. 86–88 % |
Not reported | Self-collection is comparable to CCS for detecting cervical carcinoma |
| Lichtenfels et al., 2023 | Brazil, São Paulo | Not reported | 73 | 25–65 | SelfCervix | Vaginal smear 2.Cervical smear |
Self-sampling Vs. Clinician collected |
HPV DNA detection by HC2, QIAGEN |
PreservCyt® (Hologic, MA, USA) | 87 % | 86 % | 90 % | Self-sampling using the SelfCervix® “is not inferior in HPV-DNA detection rate” compared with CCS |
| Martinelli et al., 2023 | Italy, Monza | Not reported | 245 | 17–67 | 1. Colli-pee; 2. FLOQSwab |
1.Urine sample; 2.Vaginal smear; |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | Anyplex™II HPV28 (Seegene) | LBC (ThinPrep medium)/ Bethesda system | 1. κ = 0.715; 2. κ = 0.898 |
1.90.9 % 2.95.5 % 3.95.5 % |
1.39.8 % 2.36.3 % 3.40.8 % |
High accuracy of self-collected samples confirmed in detecting HSIL |
| Nieves et al., 2013 | Mexico, Michoacán | Not reported | 2,049 | 30–50 | POI/NIH self-sampler | 1.Vaginal smear; 2.Cervical smear; |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | HPV DNA detection by HC2 |
Pap-test (PreservCyt solution)/Bethesda system | 97 % | 62.5 % | 90.5–93 % | Self-sampling applications are explored and showed a high agreement and SN with CCS |
| Othman et al., 2016 | Malaysia | Cross-sectional | 367 | 22–65 | 1.Evalyn Brush; 2.Cervex-Brush |
1.Vaginal smear; 2.Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | Not performed | LBC (PreservCyt solution)/Bethesda system | κ = 0.568,p = 0.040 | 71.9 % | 86.6 % | Self-sampling and CCS have “good diagnostic agreement” |
| Ozawa et al., 2023 | Japan | Not reported | 165 | 20–50 | 1.Home Smear Set Plus; 2. Cervex Brush |
1.Vaginal smear; 2.Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | HPV testing - Cobas 4800 HPV system (Roche Diagnostics KK) | Cytology and hystology; Bethesda system | 88.5 % κ = 0.76 | 1.81.4 % 2.89.8 % |
– | High concordance rate between self-collected and CCS |
| Pattyn et al., 2019 | Belgium/not specified | Randomized | 33 (258 samples) | 27–37 | Colli-Pee | 1. Urine sample; 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | Riatol qPCR HPV genotyping assay |
Not performed | 1.7.14 (IQR: 2.87–17.85); 2.4.5 (IQR: 1.88–9.15) ng |
Not reported | Not reported | Colli-Pee collected samples show higher HPV concentrations than cup collected samples |
| Porras et al., 2014 | Costa Rica/not specified | Costa Rica Vaccine Trial | 7,466 | 18–25 | Dacron swab | 1.Vaginal smear; 2.Cervical smear; |
1.Self-sampling vs. 2.Clinician-taken sample | HPV DNA detection by HC2 |
LBC (PreservCyt solution)/Bethesda system | κ = 0.78, McNemar χ2 = 0.62 |
88.7 % | 68.9 % | Self-collected specimens provided SN and SP comparable with CCS |
| Phoolcharoen et al., 2023 | Thailand, Bangkok | Not reported | 494 | Not available | Aptima Multitest Swab | 1.Vaginal smear; 2.Cervical smear; |
1.Self-sampling vs. 2.Clinician-taken sample | APTIMA® HPV Assay (Hologic Inc, San Diego, USA) | Colposcopy and biopsy; Bethesda system |
– | 1.87 % 2.90.2 % |
1.28.5 % 2.36.1 % |
Self-collected samples for HPV detection demonstrated good sensitivity |
| Sechi et al., 2023 | Italy/Sardinia | Cross-sectional | 185 | 34–51 | FLOQSwab; | 1. Vaginal smear; 2. Cervical smear |
1.Self-sampling vs. 2. Clinician-taken sample |
Anyplex™ II HPV HR (Seegene | – | κ = 0.98 | – | – | High reliability and accuracy of HPV-DNA tests self-collected samples via FLOQSwabs was shoen |
| Qin et al., 2016 | China/ Yunnan Province | Cross-sectional | 300 | 25–65 | FTA Elute card | 1.Vaginal smear; 2.Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | Abbott RealTime High RiskHPV assay | Colposcopy with cervical histology/CIN system | 87 %, κ = 0.731 |
100 % | 61.39 % | FTA Elute card demonstrated good performance on self-collected sample for HR-HPV |
| Sangrajrang et al., 2023 | Thailand, Bangkok | Not reported | 268 | 30–60 | Aptima Multitest Swab Specimen Collection Kit | 1. Vaginal smear; 2.Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | APTIMA® HPV Assay (Hologic Inc, San Diego, USA) | – | 92.8 % κ = 0.57 | – | – | Self-sampling is a reliable alternative to CCS |
| Saville et al., 2020 | Australia/not specified | Cross-sectional | 303 | ≥18 | 1. FLOQSwab 552C; 2. Cervex-Brush |
1.Vaginal smear 2. Cervical smear |
1.Self-sampling vs. 2.Clinician-taken sample | 1.Cobas 4800 HPV; 2. Gene Xpert HPV test; 3. BD Onclarity HPV assay; 4. Anyplex II HPV HR Detection; 5. Abbott HPV test |
Not performed | κ = 0.73 | 1. 93.8–100 %; 2. 82.15–82.4 % 3. 83.3–100 %; 4. 84.9–100 %; 5. 80–88.5 % |
1. 96.5––99 % 2. 97.7–97.5 % 3. 97.8––99.3 % 4. 98.5––99.3 % 5. 98.9––99.3 % |
Self-collection for HPV-based cervical screening shows good concordance and relative SN when compared with CCS |
| Sechi et al., 2022 | Italy/Monza | Not reported | 40 | 39.5 (mean) | 1.FLOQ Swab; 2. Evalyn Brush; 3.Her swab |
1–3. Vaginal smear; 4. Cervical smear |
1–3.Self-sampling vs. 4.Clinician-taken sample |
AnyplexI HPV28 | LBC (ThinPrep medium)/Bethesda system | 1. κ = 0.89; 2. κ = 0.79; 3. κ = 0.90 |
Not reported | Not reported | Self-collected samples showed overall high concordance with CCS |
| Shih et al., 2023 | Taiwan, Taichung | Not reported | 167 | ≥20 | 1. Urine sampler 2.Rovers Cervex-Brush 3. Digene cervical brush |
1.Urine test 2.Vaginal smear 3. Cervical smear |
1 and 2.Self-sampling vs. 3.Clinician-taken sample | 1. HPV DNA Tests by Cervista 2.HPV DNA Tests by HC II | – | κ = 0.22–0.26 | 1.75 % 2.49 % |
1.74.5 % 2.71.1 % |
HPV tests self-samples had around 60 % SN to HPV tests on CCS |
| Tiiti et al., 2021 | South Africa/Gauteng Province (black Africans) | Cross-sectional | 527 | ≥18 | 1.SelfCerv Self-Collection Cervical Health Screening Kit; 2. Cervex-Brush Combi |
1. Vaginal smear; 2.Cervical smear |
1.Self-sampling vs. 2. Clinician-taken sample |
Abbott RealTime HR-HPV and Aptima HR-HPV mRNA assays |
Not performed | 87.1 %, κ = 0.74 |
86.2 % | 88 % | Self-collected samples had good agreement with the CCS for the detection of HR-HPV |
| Tranberg et al., 2018 | Denmark/Central Region | Cross-sectional | 213 | 30–59 | Evalyn brush | Vaginal smear | 1.Self-sampling vs. 2.Clinician-taken sample | Cobas 4800 assay | LBC (not specified)/ Bethesda system | 89.2 % | 80.9 % | 91.6 % | A good concordance between self-samples and CCS in terms of HPV detection |
| Tranberg et al., 2020 | Denmark/ Central Denmark Region | Cross-sectional | 216 | 30–59 | 1. Cervex-Brush; 2. Evalyn Brush; 3. Genelock |
1.Cervical smear 2. Vaginal smear 3. Urine sample |
1–3.Self-sampling vs. 4.Clinician-taken sample | GENOMICA CLART® Vs.COBAS® 4800 assays |
Not performed | 1.κ = 0.59; 2.κ = 0.66 |
1.51.6 %; 2.63.9 % |
1.92.4 %; 2.96.5 % |
With COBAS, higher concordance between urine and vaginal self-sampling and CCS HR-HPV detection |
| Van Keer et al., 2022 | Belgium/not specified | The VALHUDES framework | 492 | 19–72 | 1. Colli-Pee 2. Cervex-Brush |
1. Urine sample; 2.Cervical smear |
1. SelSelf-sampling Clinician-taken sample | BD Onclarity HPV Assay | LBC, BD Viper LT System/ Bethesda system | κ = 0.678 | 1.90.9–91.1 %; 2. 90.9 %- 93.3 % |
1. 46.3 %; 2. 50.5 % |
BD Onclarity HPV Assay on first-void urine has similar SN and “somewhat lower” SP in self-sampling and CCS |
| Wong et al., 2018 | China, Hong Gong/sex-workers | Cross-sectional | 68 | 22–59 | 1.Dacron swab 2. Cytobrush |
1.Vaginal smear 2. Cervical sample |
Self-sampling vs. Clinician-taken sample | Genotyping assay type not psecified | Papanicolaou test/ Bethesda system | 85.3 %, κ = 0.69 | 66.7 % | 66.1 % | High concordance rate |
Table footnotes: CC – conventional cytology; CCS - clinician-collected specimens; CIN - cervical intraepithelial neoplasia; CISOE-A - composition, inflammation, squamous epithelium, other and endometrium, endocervical columnar epithelium, and adequacy of the smear; CONCEP - Comprehensive Prevention of Cervical Cancer in Tanzania; HC2 - Hybrid Capture 2; HPV – human papillomavirus; HR-HPV – high-risk human papillomavirus; HSIL – high-grade squamous intraepithelial lesion; LBC – liquid-based cytology; κ - Cohen’s Kappa; RT - real-time; SN – sensitivity; SP – specificity; POI/NIH - Preventive Oncology International/National Institutes of Health.