Monday, April 22, 2024

FDA Offers Very Odd (I'm Being Polite) Assessment of LDTs

When FDA released its regulatory proposal to regulate LDTs, on September 29, 2024, it provided a very large online library of supporting documents.   I was very annoyed by the online financial analysis, which buried the all-important projected costs to industry in an obscure cloud document, and even there, in tables far in the back.

Another online document is, "Examples of IVDs Offered as Laboratory Developed Tests that Raise Public Health Concerns," part of Docket FDA 2023 N 2177 as item -0076.

To my eye, as a board-certified pathologist, the report deals strangely with false positives and false negatives.   FDA approved (cleared) tests have false positives and false negatives.   They quote that the Oncotype Dx test has a "69% false positive rate."  This is a breast cancer test where a high score indicates a 30-40% risk of recurrent breast cancer.   That does not mean a "69% false positive rate."  It means a positive score IS a 30% recurrence risk of cancer.  The Agendia Mammaprint test is FDA-cleared and has a similar function, where a positive score is a high risk of breast cancer, such as the 30% range.   FDA would say that this IVD has a "70% false positive rate"  but that is not the report, the report is a risk category, not a diagnosis of (future) cancer.

The same type of odd examples occur throughout the memo.   In other cases, single cases of a FP or FN are cited, but all FDA tests have single cases of FP and FN too.