Relationship of late loss in lumen diameter to coronary restenosis in sirolimus-eluting stents

Area of Science:

• Cardiovascular research
• Interventional cardiology
• Biostatistics

Background:

• Drug-eluting stents (DES) have significantly reduced restenosis rates (<10%).
• Identifying robust angiographic endpoints is crucial for future clinical trials evaluating DES.
• Traditional statistical methods may struggle to accurately predict low restenosis rates.

Purpose of the Study:

• To evaluate the effectiveness of statistical transformations in predicting binary restenosis rates.
• To compare the accuracy of different statistical methods for analyzing late loss data.
• To determine the optimal angiographic endpoint for assessing restenosis in DES trials.

Main Methods:

• Analysis of late loss data from the SIRIUS (n=703) and E-SIRIUS (n=308) trials.
• Comparison of standard normal approximation with optimized power transformation for predicting restenosis.
• Assessment of in-stent late loss correlation with target-lesion revascularization.

Main Results:

• Standard normal approximation underestimated restenosis in the sirolimus arm (predicted 0.6% vs. observed 3.2%).
• Optimized power transformation improved prediction accuracy in the sirolimus arm (predicted 3.2%) and E-SIRIUS trial (predicted 4.0%).
• In-stent late loss showed a stronger correlation with target-lesion revascularization (c-statistic=0.915) than in-segment late loss.

Conclusions:

• Statistical transformation of late loss data is necessary for accurate prediction of low binary restenosis rates.
• Late loss, when analyzed with appropriate transformations, is a reliable and efficient endpoint for assessing restenosis.
• These findings support the use of transformed late loss in future clinical trials evaluating drug-eluting stents.