The test was deemed valid if the positive control well had +20 SFUs and the negative control well +10 SFUsThe test positivity criteria were determined according to the negative control well SFU. If the negative control well had 0–5 SFUs, the test was accepted as positive provided either or both of the ESAT-6 or CFP-10 wells contained a mean of 6 SFUs more than the negative control. If If the negative control contained +6 SFUs the test was acccepted as positive when the antigen wells contained at least twice the SFUs in the negative well.
Oxford Immunotec now advise that the prespecified cutoff of ≥6 spots and the validity control of 0-5 spots is no longer acceptable
A Nil Control spot count in excess of 10 spots should be considered as ‘ Invalid’
Results where the highest of the Panel A or Panel B spot count is such that the (Panel minus Nil) spot count is 5, 6 or 7 spots should be considered Borderline (equivocal) and retesting by collecting another patient specimen is recommended.
This has altered their sensitivity and specificity ranges accordingly.
If all the borderline (equivocal) results are considered all positive or all negative, the estimated specificity of T-SPOT.TB was either 94.8% (290/306) or 99.0% (303/306), respectively
If all the borderline (equivocal) results are considered all positive or all negative, the estimated sensitivity of T-SPOT.TB was either 96.2% (176/183) or 90.7% (166/183), respectively.
As borderline (equivocal) results are initially rejected both sensitivity and specificity is at the lower range ie 90.7% and 94.8% respectively (click on image for FULL SIZE)
From their own trials of sensitivity and specificity (see above) they had 5.4% and 4.2% returning a borderline result. This is on top of those results termed "invalid" (or indeterminate) which were 3% for sensitivity and 1.6% for specificity. Combining borderline and invalid results we have 16/189 failure for sensitivity (8.46%) and 18/311 for specificity (5.78%)
As their clinical trials use the previous "prespecified cutoff of ≥6 spots" and parameters of validity it is doubtful if any of that data is now of relevance.
