Automated tests of ANA immunofluorescence as throughput autoantibody detection technology: strengths and limitations

Anti-nuclear antibody (ANA) assay is a screening test used for almost all autoimmune rheumatic diseases, and in a number of these cases, it is a diagnostic/classification parameter. In addition, ANA is also a useful test for additional autoimmune disorders. The indirect immunofluorescence technique on monolayers of cultured epithelial cells is the current recommended method because it has higher sensitivity than solid phase assays. However, the technique is time-consuming and requires skilled operators. Automated ANA reading systems have recently been developed, which offer the advantage of faster and much easier performance as well as better harmonization in the interpretation of the results. Preliminary validation studies of these systems have given promising results in terms of analytical specificity and reproducibility. However, these techniques require further validation in clinical studies and need improvement in their recognition of mixed or less common staining patterns.


Background
Anti-nuclear antibody assay (ANA) is the screening test of choice for diagnosis of almost all systemic autoimmune rheumatic diseases (SARDs) because of its greater sensitivity compared with other assays, even though its specificity is much lower (Box 1) [1]. The gold standard method for ANA detection is still indirect immunofluorescence (IIF) on human epithelial (HEp-2) cells, as the alternative tests cannot display comparable sensitivity [2]. However, the technique is time-consuming and requires skilled operators. This fact together with the widespread increase in ANA requests and the reduction of laboratory facilities because of the budget constriction generated a strong need for advanced automated platforms as in other branches of the laboratory medicine.
These systems are based on a composition of different hardware modules combined with mathematical patternrecognition software algorithms, enabling fully automated image acquisition, analysis, and evaluation of IIF ANA tests.
The reported advantages of these systems include reduction in intra-laboratory and inter-laboratory variability, improvement in correlation between staining patterns with corresponding autoantibody reactivities, higher throughput in laboratory workflows, no requirement for a darkroom, integrated file storage, and easy retrieval of scanned wells.

Comparison of the available ANA automated reading systems
Although comparable performance between automated and conventional ANA IIF analysis for the interpretation of negative and positive samples has been reported, discrepancies between patterns have been found, especially when systems are able to detect basic patterns only, or when mixed fluorescent patterns are present in the samples [3][4][5][6][7][8][9][10][11][12][13][14].
Such IIF assays have identifed more than 50 autoantibodies against 30 different nuclear and cytoplasmic antigens [16]. The use of large cultured cells with high rates of mitosis enables adequate pattern recognition by evaluation of the fluorescence distribution during different phases of the cell cycle. In fact, identification of cell cycle dynamics (for example, interphase, mitosis) is crucial both for defining different patterns (such as the fine or large speckled patterns within a speckled staining pattern, the centromere patterns and the PCNA patterns) and for distinguishing between different patterns (for example anti-nuclear membrane from the homogeneous pattern).
Correct identification of different IIF patterns is sometimes diagnostic (for example, the centromere pattern and the PCNA pattern) or may suggest the occurrence of autoantibodies to specific antigens (Table 3). Many sera contain more than one antibody; in such cases, accurate analysis of the different patterns often requires direct evaluation of the slides to enable exact definition of the autoantibody profile in a given patient.
Systemic sclerosis (SSc) represents a paradigmatic example of an autoimmune disease that is characterized by the occurrence of ANA in virtually all patients, but for which interpretation of the patterns is complesx [17]. In fact, SSc ANA are mainly represented by four mutually exclusive specificities: anti-centromere (ACA), antitopoisomerase I, anti-nucleolar, and anti-RNA polymerase III antibodies. Anti-PM-Scl, U1-RNP and anti-Ku are  usually detected in overlap syndromes. About 60% of patients with SSc have ACA or anti-topoisomerase I antibodies as disease markers. Many other ANA that are present in SSc (for example, anti-RNA polymerase III, anti-Th/To, anti-PM/Scl, anti-Ku, anti-fibrillarin) are directed against different proteins localized in the nucleus and nucleolus. These antigen-antibody systems identify SSc subgroups with different evolution, organ involvement, and survival prognosis. The use of IIF for detection of ANA is mandatory for SSc diagnosis, displaying a sensitivity of 85% [1]. ACA and anti-topoisomerase I negative sera show strong anti-nuclear staining, featuring speckled or nucleolar (homogeneous, clumpy or speckled) patterns (Box 1). Therefore, the definition of the single nucleolar staining could address the suspect of specific autoantibodies, relevant for the diagnosis of SSc. A nucleolar ANA associated with new onset of Raynaud's phenomenon could be helpful in identifying a patient with early disease, sometimes associated with severe organ involvement. It is essential that ANA results are confirmed by more specific methods such as western blotting or immunoprecipitation assays. All these points underline the importance of correct interpretation of a given fluorescence pattern, and the need for standardization of analysis in automated systems.
There is one other important point about using automated systems for ANA reading. The ANA test was originally ordered predominantly by rheumatologists and clinical immunologists, but nowadays a broader range of clinical disciplines (including primary care, dermatology, nephrology, gastroenterology, neurology, oncology, hematology, obstetrics, gynecology, cardiology) are currently ordering the test. This change in test referral patterns affects the post-test probability for a given disease, as screening tests with limited specificity (such as IIF ANA) are strongly affected when the pre-test probability in a given population decreases [17]. A positive ANA test obtained outside of the rheumatologic setting displays poor predictive value for future development of a rheumatic disease, but it represents a significant risk factor for SLE. Taking into account that the prevalence of SLE is 1 in 2000 (0.05%), the observed frequency of 2.5% in individuals with a 1/80 positive ANA test represents a 50-fold relative risk for development of the disease [18,19]. Thus, ANA testing is a useful tool for SLE diagnosis.

Conclusions
Current evidence from preliminary study results indicates that there is good correlation between manual and automated interpretation of ANA IIF assays, at least in the ability to discriminate between positive and negative results and in recognizing the main IIF patterns. Such systems will therefore speed up routine performance of these tests and help to harmonize interpretation of the results across laboratories. However, there is a need to have their clinical diagnostic power validated by clinical studies, in addition to the analytical studies that have already been published. In addition, these new systems could be further improved if they were better able to recognize mixed fluorescent or less common fluorescent patterns. Competing interests PM has received fees as consultant for Inova and from BioRad; NB has been a paid consultant to Inova Diagnostics and has received lecture fees from A Menarini Diagnostics; and AT has received funding and reimbursements from IL, BioRad, and Thermofisher.