Wolff-Parkinson-White detection at 471:1 class imbalance: a leakage-controlled study of the data bottleneck

Wolff-Parkinson-White (WPW) syndrome is a congenital cardiac pre-excitation, clinically important and often missed on the resting 12-lead electrocardiogram. Detection is hard: the signature is subtle and the condition is rare. We pool two public 12-lead corpora, PTB-XL and Chapman-Shaoxing-Ningbo, for 66,951 recordings of which 142 are WPW, a prevalence of 0.21% (about 471:1).

Under one pre-specified, leakage-controlled protocol, with a held-out fold contacted exactly once, we compare seven representations of the signal, holding the split and the evaluation fixed: beat-delineation morphology, global-statistical summaries, wavelet time-frequency localization, median-beat morphology, vectorcardiographic geometry, an on-machine commercial measurement set, and a one-dimensional convolutional network. Only the representation changes.

Within these corpora and under a modest compute budget, added diversity and capacity do not raise the ceiling: the most orthogonal detector significantly hurts, a feature-union model matches a two-member vote, a convolutional network reaches the wavelet detector without exceeding it, and self-supervised pretraining fails a pre-specified gate. A leak-free learning curve, re-selecting features at every size, is still rising at the full 115 positives for the strongest deployed detector (paired 90-to-100% difference +0.027, 95% CI [0.019, 0.033]), so it is not shown to have saturated.

An error analysis tested against independent evidence finds that the missed cases have a narrower QRS, confirmed by an on-machine measurement outside our pipeline after we show that the sign of this effect depends on which delineator measures it; that uncertain labels show no enrichment among the misses; and that some apparent false positives are recordings the corpus itself codes as pre-excited, placing part of the label problem in the negative class. We measure the optimism of non-nested selection at 0.11 to 0.13 average precision.

The deployed output is a percentile rank in a frozen reference distribution, not a probability. On the held-out fold, on 14 positives, it reaches an average precision of 0.595 and an ROC area of 0.950. It is a screening pre-filter, not a diagnostic tool.