Calculation of diagnostic parameters of advanced serological and molecular tissue-print methods for detection of Citrus tristeza virus: a model for other plant pathogens

Citrus tristeza virus (CTV) is one of the most important virus diseases that affect citrus. Control of CTV is achieved by grafting selected virus-free citrus scions onto CTV-tolerant or -resistant rootstocks. Quarantine and certification programs are essential for avoiding the entry and propagation of severe strains of CTV. Citrus nurseries in Spain and central California (United States) maintain zero-tolerance policies for CTV that require sensitive, specific, and reliable pathogen-detection methods. Tissue-print (TP) real-time reverse-transcriptase polymerase chain reaction (RT-PCR) assay was compared with the validated TP enzyme-linked immunosorbent assay (ELISA), using the CTV-specific monoclonal antibodies 3DF1 and 3CA5, for CTV detection. In total, 1,395 samples from healthy and CTV-infected nursery and mature tree plants were analyzed with both methods. The total agreement between both detection methods was substantial (Cohen's kappa index of 0.77 ± 0.03). The diagnostic parameters of each technique (i.e., the sensitivity, specificity, and likelihood ratios) were evaluated in a second test involving 658 Citrus macrophylla nursery plants. Mexican lime indexing was used to evaluate samples with discrepant results in the analysis. For TP-ELISA, a sensitivity of 0.8015, a specificity of 0.9963, and a positive and negative likelihood ratio of 216.42 and 0.199, respectively, were estimated. For TP real-time RT-PCR, a sensitivity of 0.9820, a specificity of 0.8519, and a positive and negative likelihood ratio of 6.63 and 0.021, respectively, were estimated. These diagnostic parameters show that TP real-time RT-PCR was the most sensitive technique, whereas TP-ELISA showed the highest specificity, validating the use of the molecular technique for routine CTV-detection purposes. In addition, our results show that the combination of both techniques can accurately substitute for the conventional biological Mexican lime index for the detection of CTV. The calculation of diagnostic parameters is discussed, as a necessary tool, to validate detection or diagnostic methods in plant pathology. Furthermore, assessment of the post-test probability of disease after a diagnostic result and CTV prevalence allows selection of the best method for accurate and reliable diagnosis.