Six cohort studies (312 patients) were included in the review. All of these studies were considered to have a representative sample, a short enough period between reference standard and index test, an accurate description of the index test and reported uninterpretable/intermediate test results. Three studies met all 12 QUADAS quality criteria.
Pooled sensitivity of anti-p155 for diagnosing cancer-associated dermatomyositis was 78% (95% CI 45 to 94). Pooled specificity was 89% (95% CI 82 to 93). The diagnostic odds ratio was 27.3 (95% CI 6.6 to 112.8). The positive likelihood ratio was 6.79 (95 CI 4.11 to 11.23) and the negative likelihood ratio was 0.25 (95% CI 0.08 to 0.76). There was substantial statistical heterogeneity for all of these outcomes except the positive likelihood ratios. The area under the SROC curve was 0.91 (95% CI 0.88 to 0.93).
Using a pooled prevalence of 17%, anti-p155 had a positive predictive value of 58% and a negative predictive value of 95%.
There was no evidence of publication or small sample biases. Sensitivity analyses were also reported in the paper.