Table 2.
Summary of findings of included systematic reviews and meta-analysis.
| Author | Number of studies | Number of participants or lesions | Adjunctive tools with sufficient evidence | Adjunctive tools with insufficient evidence | Overall sensitivity, specificity, accuracy, positive predictive value or negative predictive value | Overall recommendation and conclusion | Publication bias | Methodological limitations |
|---|---|---|---|---|---|---|---|---|
| Patton (2003)[23] | 13 | Participants 4872 | Toluidine blue | Oral Brush cytology- OralCDx | Toluidine blue (sensitivity: 72 - 100%, specificity: 45-93%) | Fair evidence to support use of toluidine blue to aid in the diagnosis of oral cancer | Not reported | No reporting guidelines for systematic reviews (eg PRISMA) English language restriction Titles and abstracts not duplicated by multiple people Publication of bias analysis not specifically evident |
| Patton et al. (2008)[24] | 23 | subjects: 3687 lesions: 3323 |
Toluidine blue | Insufficient evidence to support or refute the use of visually based examination adjuncts. Chemiluminescence: ViziLite |
Toluidine blue (median sensitivity: 85%, median specificity: 67%, median PPV: 85%, median NPV: 83%) ViziLite (median sensitivity: 100%, median specificity: 0%, median PPV: 20%, median NPV: 0%) |
Toluidine blue is an effect diagnostic adjunct for OPMD | Not reported | No reporting guidelines for systematic reviews (eg PRISMA) English language restriction Titles and abstracts not duplicated by multiple people Publication of bias analysis not specifically evident |
| Fuller et al. (2015)[25] | 48 | Subjects: 2184 lesions: 1887 |
Oral cytology Diffuse reflectance spectroscopy Laser-induced autofluorescence |
Toluidine Blue | Cytology (sensitivity: 89.8%, specificity: 89.8%, accuracy: 85.7%) diffuse reflectance spectroscopy (sensitivity: 98.6%, specificity: 83.2%, accuracy: 96.5%) Laser-induced autofluorescence (sensitivity: 97.7%, specificity: 84.4%, accuracy: 95.9%) Toluidine blue (sensitivity: 82.1%, specificity: 52.8%, accuracy: 66.7%) |
Significant improvement in diagnostic quality of oral cytology | Not reported | English language restriction Publication of bias analysis not specifically evident |
| Rashid et al. (2015)[26] | 25 | subjects: 1133 lesions: 1182 |
Chemiluminescence and autofluorescence is better suited in specialist clinics | Limited evidence for chemiluminescence and autofluorescence in primary care | VELscope (sensitivity: 30-100%) ViziLite (sensitivity: 77.3-100%) ViziLite with toluidine blue (sensitivity: 0-59%) |
Limited evidence for chemiluminescence and autofluorescence in primary care | Modified Newcastle-Ottawa Scale (NOS) | No reporting guidelines for systematic reviews (eg PRISMA) English language restriction |
| Awan et al. (2016)[27] | 11 | subjects: 3838 | None reported | Autofluorescence: VELscope | VELscope (sensitivity: 30-100%, specificity: 15-92.3%, PPV: 6.4-58.1%, NPV: 57.1-100%) | Insufficient evidence to show direct tissue fluorescence visualization has a capability to be used as an oral cancer screening tool | Not reported | English language restriction No reporting guidelines for systematic reviews (eg PRISMA) Titles and abstracts not duplicated by multiple people |
| Nagi et al. (2016)[28] | 22 | subjects: 543 | None reported | Chemiluminescence: ViziLite Autofluorescence: VELscope |
ViziLite (sensitivity: 77.1-100%, specificity: 0-27.8%) VELscope (sensitivity: 22-100%, specificity: 16-100%) |
Both chemiluminescence and autofluorescence may aid an experienced clinician. In literature, limited discrimination in high risk lesions. | Not reported | English language restriction No reporting guidelines for systematic reviews (eg PRISMA) Titles and abstracts not duplicated by multiple people |
| Giovannacci et al. (2016)[29] | 35 | patients: 3137 | None reported | Autofluorescence Chemiluminescence Toluidine blue Chemiluminescence associated with toluidine blue |
Autofluorescence: sensitivity (mean: 72.4% range: 20-100%) specificity: (mean: 63.75%, range 15-100%), PPV: mean 55.74%, NPV: mean: 79.76% Chemiluminescence: sensitivity (mean: 86.75% range 69.6-100%). Specificity: (mean: 38.37- range: 14.2%−81.5%) PPV: mean 74.5%, NPV: mean: 63% Toluidine blue: sensitivity (mean: 72.5%, range: 56.1-95%) specificity: mean: 61.4% range: 25-74%)PPV: mean 58.16%, NPV: mean: 95.3% chemiluminescence associated with toluidine blue: sensitivity (mean: 53.93% range 0-81.8%). Specificity: (mean: 66.44- range: 37.5%−97.5%) mean 87.2%, NPV: mean: 76.1% |
Great inhomogeneity of the reported values | Not reported | English language restriction No reporting guidelines for systematic reviews (eg PRISMA) |
| Alsarraf et al. (2018)[30] | 36 | lesions: 4302 | None reported | Brush cytology | Brush cytology (sensitivity: 60-100%, specificity: 32-100%) | Meaningful evidence-based recommendations cannot be given from this study. | Not reported | English language restriction |
| Lingen et al. (2017)[31] | 37 | lesions: 5390 | Cytology | Autofluorescence vital staining Tissue reflectance Cytologic testing and vital staining together |
Cytology (sensitivity: 92%, specificity: 94%) Autofluorescence (sensitivity: 90%, specificity: 72%) vital staining (sensitivity: 87%, specificity: 71%) Tissue reflectance (sensitivity: 75, specificity: 31%) Cytologic testing and vital staining together (sensitivity: 95%, specificity: 68%) |
Cytologic testing used in suspicious lesions appears to have the highest accuracy among adjuncts | QUADAS-2 | Studies identified only covered secondary and tertiary settings |
| Bustiuc et al. (2018)[32] | 15 | Lesions: 2364 | None reported | Autofluorescence: VELscope Chemiluminescent devices |
Velscope (sensitivity: 22-100%, specificity: 8.4-96.3%) Chemiluminescent devices (sensitivity: 67-100%, specificity: 10-100%) |
Velscope, Microlux/DL, VIziLite devices can be used assupporting methods. | Not reported | English language restriction No reporting guidelines for systematic reviews (eg PRISMA) Titles and abstracts not duplicated by multiple people |
| Cicciu et al. (2019)[33] | 25 | Lesions: 1693 | None reported | Autofluorescence: VELscope | VELscope (sensitivity mean: 70.19%, sensitivity range: 8.4-100%, specificity mean: 66.95%, specificity range: 8.4-100%) | VELscope does not have the capacity to discern between a benign lesion, a malignant one, or a simple acute inflammation. | Risk of bias analysis completed | English language restriction |
| Chaitanya et al. (2019)[34] | 12 | patients: 1643 | None reported | Autofluorescence: VELscope | VELscope (sensitivity: 40%, specificity: 80%) | Autofluorescence using devices may be used as adjunct to find the exact location of the biopsy in altered mucosal conditions. | Not reported | English language restriction No reporting guidelines for systematic reviews (eg PRISMA) Titles and abstracts not duplicated by multiple people |
| Tiwari et al. (2020)[35] | 27 | sample size: 6415 | None reported | Autofluorescence - varied sensitivity and sensitivity- large range | Autofluorescence (sensitivity: 17-99.2%, specificity: 38-97.9%) | Optical fluorescence should be used as a clinical adjunct rather than a specific diagnostic adjunct | QUADAS-2 | English language restriction |
| Kim et al. (2020)[36] | 28 | sample size: 1166 | None reported | Autofluorescence Toluidine blue |
Autofluorescence (sensitivity: 79.1%, specificity: 50.9%, NPV: 59.8%) Toluidine blue (sensitivity: 75.4%, specificity: 60.3%, NPV: 68.5%) |
autofluorescence and toluidine blue staining can not reliably be used alone for screening or a diagnostic workup. | QUADAS-2 | English language restriction |
| Buenahora et al. (2021)[37] | 40 | autofluorescence sample size: 5562 chemiluminescence sample size: 1353 |
Autofluorescence | Chemiluminescent | autofluorescence (sensitivity: 86%, specificity: 72%) chemiluminescent (sensitivity: 67%, specificity: 48%) |
Autofluorescence devices displayed superior accuracy levels in the identification of premalignant lesions and early neoplastic changes | QUADAS-2 | Only 2 database searches were used |
| Mazur et al. (2021)[38] | 43 studies- qualitative synthesis, 34 papers- meta-analysis |
Vital staining: 536 | None reported | Autofluorescence High-Resolution Microendoscopy (HRME) Optical Spectroscopy Narrow Banding Imaging Vital Stain Colorants |
Autofluorescence: I2: 84.7% Q= 71.67 High-Resolution Microendoscopy (HRME): I2: 77.6% Q= 18.27 Optical Spectroscopy: I2: 80% Q= 30.18 Narrow Banding Imaging: I2: 19.7% Q= 1.24 Vital Stain Colorants: I2: 87.6% Q= 36.63 |
No technique can replace biopsy as the gold standard | Jadad Scale Newcastle-Ottawa scale Cochrane guidelines |
Language restriction to English, French, German, Spanish, Polish, Albanian, and Romanian. No excluded list of studies published |
| Moffa et al. (2021)[39] | 26 studies | 2631 oral lesions | None reported | Autofluorescence Chemiluminescence |
Autofluorescence (sensitivity: 81.3%, specificity: 52.1%) chemiluminescence (sensitivity: 84.9%, specificity: 51.8%) |
Poor specificity, and reduction of the false positive rate of both autofluorescence and chemiluminescence compared with COE. | QUADAS-2 | English language restriction Studies with a sample size under 10 patients |
| Lima et al. (2021)[40] | 45 studies | None reported | Autofluorescence: VELscope 5-Aminolevulinic acid |
VELscope (sensitivity: 33-100%, specificity: 12-88.6%) 5-Aminolevulinic acid (sensitivity: 90-100%, specificity: 51-96%) |
Autofluorescence and fluorescent probes can provide an accurate diagnosis of oral cancer, but can not replace histopathology | The Joanna Briggs Institute Critical Appraisal tools for use in JBI Systematic Reviews Checklist for Diagnostic Test Accuracy Studies | No excluded list of studies published | |
| Saraniti et al. (2021)[41] | 9 studies | number of lesions: 1507 | None reported | Narrow band imaging | Narrow band imaging- large range of sensitivities, specificity, PPV and NPV | NBI has a higher specificity, sensitivity, positive and negative predictive values and accuracy compared to white light examination | not reported | English language restriction Publication of bias analysis not specifically evident |
| Kim et al. (2021)[42] | 10 studies | patients: 1374 | Narrow Band Imaging- class III | Narrow Band Imaging - Class I or class II criteria: | Narrow band imaging- class III pooled sensitivity: 88.5%, pooled specificity: 90.1%, pooled NPV: 96.7%) Narrow Band Imaging - Class I or class II: sensitivity: 79.1% (class I), 95.6% (class II) specificity: 29.7% (class I), 19.9% (class II) |
NBI considered more accurate than white-light imaging when using the class III criteria | QUADAS-2 | No reporting guidelines for systematic reviews (eg PRISMA) Selection criteria was not detailed |
| Walsh et al. (2021) Updated paper from Macey et al. 2015[43] | 63 studies | lesions: 7942 | None reported | Vital staining Cytology Light-based detection |
vital staining (overall sensitivity: 86%, specificity: 68%) cytology (overall sensitivity: 90%, specificity: 94%) light-based detection (overall sensitivity: 87%, specificity: 50%) |
adjunctive tests cannot be recommended as replacement of gold standard scalpel biopsy and histological assessment | QUADAS-2 | no excluded list of studies published |
| Dos Santos et al. (2022)[44] | 25 studies | patients: 2249 | None reported | Autofluorescence: VELscope | VELscope (sensitivity: 74%, specificity: 57%) | Promising results regarding auto-fluorescence based methods | Joanna Briggs Institute Critical Appraisal Checklist for Diagnostic Test Accuracy Studies. | English language restriction no excluded list of studies published |
| Kim DH et al. (2022)[45] | 24 studies | patients: 1914 | Narrow band imaging | Autofluorescence Chemiluminescence Cytology Toluidine blue |
Narrow band imaging (sensitivity: 77.9%, specificity: 83.5%, NPV: 83.5%, PPV: 72.5%, accuracy: 90.8%) cytology (sensitivity: 72.1%, specificity: 86.2%, NPV: 84.9%, PPV: 75.1%, accuracy: 81.9%) Autofluorescence (sensitivity: 85.6%, specificity: 48.8%, NPV: 86.6%, PPV: 49.3%, accuracy: 66.4%) chemiluminescence (sensitivity: 87.5%, specificity: 56.8%, NPV: 83.3%, PPV: 66.4%, accuracy: 74.5%) toluidine blue (sensitivity: 71.4%, specificity: 81.1%, NPV: 70.8%, PPV: 81.1%, accuracy: 75.9%) |
narrow band imaging showed higher sensitivity, specificity, negative predictive value, positive predictive value | QUADAS-2 | no excluded list of studies published |
| Shaw et al. (2022)[46] | 24 studies | patients: 1783 | Chemiluminescence of OPMDs including: Leukoplakia Oral lichen planus Oral sub mucous fibrosis |
None reported | Chemiluminescence- accuracy of Leukoplakia (sensitivity: 75% specificity: 98%) OLP (sensitivity: 78%, specificity: 60%) Oral sub mucous fibrosis (sensitivity: 89%, specificity: 76%) |
Chemiluminescence overall had good sensitivity and specificity values for the diagnosis of OPMDs. | QUADAS-2 | no excluded list of studies published |
| Kim DH et al. (2022)[47] | 16 studies | patients: 998 | None reported | Chemiluminescence Toluidine blue |
Chemiluminescence (sensitivity: 83.1%, specificity: 41.5%, NPV: 67.4%) Toluidine blue (sensitivity: 83.2%, specificity: 42.9%, NPV: 74.7%) |
chemiluminescence is comparable or worse than toluidine blue and clinical examination | QUADAS-2 | Limited information on included studies no excluded list of studies published No reporting guidelines for systematic reviews (eg PRISMA) |
| Mendonca et al. (2022)[48] | 44 studies | lesions: 3317 | Detection of dysplastic OPMD: Autofluorescence Chemiluminescence Narrow Band Imaging Fluorescence Spectroscopy Diffuse Reflectance Spectroscopy 5-aminolevulinic acid induced protoporphyrin IX fluorescence |
Detection of OSCC: Autofluorescence Narrow Band Imaging Fluorescence Spectroscopy Diffuse Reflectance Spectroscopy |
Detection of dysplastic OPMD: Autofluorescence (sensitivity: 75%, specificity: 50%) Chemiluminescence (sensitivity: 74%, specificity: 47%) Narrow Band Imaging (sensitivity: 31%, specificity: 90%) Fluorescence Spectroscopy (sensitivity: 72%, specificity: 96%) Diffuse Reflectance Spectroscopy (sensitivity: 79%, specificity: 86%) 5aminolevulinic acid induced protoporphyrin IX fluorescence (sensitivity: 91%, specificity: 78%) Detection of OSCC: Autofluorescence (sensitivity: 0.96, specificity: 0.58) Narrow Band Imaging (sensitivity: 0.97, specificity: 0.89) Fluorescence Spectroscopy (sensitivity: 0.93, specificity: 0.97) Diffuse Reflectance Spectroscopy (sensitivity: 0.93, specificity: 0.90) |
Analysed non-invasive imaging techniques suggest higher accuracy levels in the diagnosis of OSCC when compared to dysplastic OPMDs | QUADAS-2 | English language restriction no excluded list of studies published |
| Zhang et al. (2022)[49] | 11 studies | patients: 1179 | Narrow band imaging IPCL II lesions | Narrow band imaging – IPCL II lesions: Sensitivity: 87%, specificity: 83% |
Narrow band imaging is a promising adjunctive tool for identifying malignant transformations of OPMDs. | QUADAS-2 | No excluded list of studies published |