When words first fail: Predicting the emergence of primary progressive aphasia variants from unclassifiable anomic performance in early disease.
Melissa D Stockbridge, Donna C Tippett, Bonnie L Breining and 1 others
PMID 37377938WHAT IT FOUND
In early unclassifiable primary progressive aphasia, a Boston Naming Test score of 18 or lower was associated with later semantic variant, and 28/30 or higher with later nonfluent variant.
Key findings
01Among patients who were initially unclassifiable and later received a variant, the Boston Naming Test was the only single task whose scores significantly predicted the eventual variant.
02BNT cut-points separated variants: 28/30 or higher identified later nfvPPA with 91% accuracy, 18 or lower identified later svPPA with 91% accuracy, and 20/30 or higher identified later lvPPA with 70% accuracy.
03Low scores on written noun naming, BNT, HANA, and Berndt verb verification were most useful for predicting later svPPA, while high Berndt verb scores predicted later lvPPA and high BNT or HANA scores predicted later nfvPPA.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
Only 19 patients were analysed, and the authors say the small number likely made the models learn the quirks of these patients rather than a stable rule. The sample included only English-speaking patients who returned for follow-up and later received a recognized variant, so it does not describe patients who never developed a variant or who were unclassifiable for other reasons. The study was retrospective and single-centre, and referral patterns may have favoured patients with later disease progression. Tasks with fewer than 15 scores were dropped, and the four key naming tasks were so closely related that they could not be combined into one model. The cut-points were exploratory and need prospective replication before clinical use.
Declared interests
The authors report no competing interests to declare.
The easy way to misread this
Do not use a Boston Naming Test score alone to diagnose a PPA variant or promise a patient how the disease will progress. The prediction came from a very small retrospective sample, and the authors say the models may have learned the quirks of these patients rather than a stable rule.